Organis syntheis. Strategy and control

Organic Synthesis

Organic Synthesis: Strategy and Control Paul Wyatt Senior Lecturer and Director of Undergraduate Studies, School of Chemistry, University of Bristol, UK and

Stuart Warren Reader in Organic Chemistry, Department of Chemistry, University of Cambridge, UK

Copyright © 2007

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone (+44) 1243 779777

Email (for orders and customer service enquiries): [email protected] Visit our Home Page on www.wiley.com All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London WIT 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to [email protected], or faxed to (+44) 1243 770620. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, ONT, L5R 4J3, Canada Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Anniversary Logo Design: Richard J. Pacifico Library of Congress Cataloging-in-Publication Data Wyatt, Paul. Organic synthesis: strategy and control / Paul Wyatt and Stuart Warren. p. cm. Includes bibliographical references. ISBN: 978-0-470-48940-5 ISBN: 978-0-471-92963-5 1. Organic compounds – Synthesis. 2. Stereochemistry. I. Warren, Stuart G. II. Title. QD262.W89 2007 547´.2–dc22 2006034932 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-471-48940-5 (HB) ISBN: 978-0-471-92963-5 (PB) Typeset in 10/12pt Times by Thomson Digital Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire This book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least two trees are planted for each one used for paper production.

Contents

Preface

vii

A: Introduction: Selectivity 1. Planning Organic Syntheses: Tactics, Strategy and Control 2. Chemoselectivity 3. Regioselectivity: Controlled Aldol Reactions 4. Stereoselectivity: Stereoselective Aldol Reactions 5. Alternative Strategies for Enone Synthesis 6. Choosing a Strategy: The Synthesis of Cyclopentenones

3 9 27 43 55 71

B: Making Carbon–Carbon Bonds 7. The Ortho Strategy for Aromatic Compounds 8. σ-Complexes of Metals 9. Controlling the Michael Reaction 10. Specific Enol Equivalents 11. Extended Enolates 12. Allyl Anions 13. Homoenolates 14. Acyl Anion Equivalents

91 113 127 139 155 173 189 203

C: Carbon–Carbon Double Bonds 15. Synthesis of Double Bonds of Defined Stereochemistry 16. Stereo-Controlled Vinyl Anion Equivalents 17. Electrophilic Attack on Alkenes 18. Vinyl Cations: Palladium-Catalysed C–C Coupling 19. Allyl Alcohols: Allyl Cation Equivalents (and More)

223 255 277 307 339

D: Stereochemistry 20. Control of Stereochemistry – Introduction 21. Controlling Relative Stereochemistry 22. Resolution 23. The Chiral Pool — Asymmetric Synthesis with Natural Products as Starting Materials — 24. Asymmetric Induction I Reagent-Based Strategy 25. Asymmetric Induction II Asymmetric Catalysis: Formation of C–O and C–N Bonds 26. Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds 27. Asymmetric Induction IV Substrate-Based Strategy

371 399 435 465 505 527 567 599

vi

Contents

28. Kinetic Resolution 29. Enzymes: Biological Methods in Asymmetric Synthesis 30. New Chiral Centres from Old — Enantiomerically Pure Compounds & Sophisticated Syntheses — 31. Strategy of Asymmetric Synthesis E: Functional Group Strategy 32. Functionalisation of Pyridine 33. Oxidation of Aromatic Compounds, Enols and Enolates 34. Functionality and Pericyclic Reactions: Nitrogen Heterocycles by Cycloadditions and Sigmatropic Rearrangements 35. Synthesis and Chemistry of Azoles and other Heterocycles with Two or more Heteroatoms 36. Tandem Organic Reactions

627 651 681 717 749 777 809 835 863

General References

893

Index

895

Preface We would like to thank those who have had the greatest influence on this book, namely the undergraduates at the Universities of Bristol and Cambridge. But, particularly we would like to thank the organic chemists at Organon (Oss), AstraZeneca (Alderley Park, Avlon Works, Mölndal and Macclesfield), Lilly (Windlesham), Solvay (Weesp) and Novartis (Basel) who contributed to the way the book was written more than they might realise. These chemists will recognise material from our courses on The Disconnection Approach, Advanced Heterocyclic Chemistry, New Synthetic Methods and Asymmetric Synthesis. Additionally we would like to thank the participants at the SCI courses organised by the Young Chemists Panel. All these industrial chemists participated in our courses and allowed us to find the best way to explain concepts that are difficult to grasp. This book has changed greatly over the ten years it was being written as we became more informed over what was really needed. The book is intended for that very audience – final year undergraduates, graduate students and professional chemists in industry. PJW SGW July 2006

Section A: Introduction: Selectivity 1. Planning Organic Syntheses: Tactics, Strategy and Control ............................................. 3 2. Chemoselectivity ..................................................................................................................... 9 3. Regioselectivity: Controlled Aldol Reactions ..................................................................... 27 4. Stereoselectivity: Stereoselective Aldol Reactions ............................................................. 43 5. Alternative Strategies for Enone Synthesis ........................................................................ 55 6. Choosing a Strategy: The Synthesis of Cyclopentenones .................................................. 71

1

Planning Organic Syntheses: Tactics, Strategy and Control The roll of honour inscribed with successful modern organic syntheses is remarkable for the number, size, and complexity of the molecules made in the last few decades. Woodward and Eschenmoser’s vitamin B12 synthesis,1 completed in the 1970s, is rightly regarded as a pinnacle of achievement, but since then Kishi2 has completed the even more complex palytoxin. The smaller erythromycin and its precursors the erythronolides3 1, and the remarkably economical syntheses of the possible stereoisomers of the cockroach pheromones 2 by Still4 deal with a greater concentration of problems. O O

O HO

1 erythronolide A

OH

OH

2 periplanone-B

O O

OH

O

OH

Less applauded, but equally significant, is the general advance in synthetic methods and their industrial applications. AstraZeneca confess that it took them nearly a century to bring Victor Grignard’s methods into use, but are proud that Corey’s sulfur ylid chemistry made it in a decade. Both are used in the manufacture of the fungicide flutriafol5 3.

Me

F

S

CH2

O

F

O

F

F

Me

F

N

RCO3H

N

F

N N

N base

N

F

F

AlCl3

+

H2O

F

O

MgI

OH X

F

F

Br2

F

3 Flutriazole fungicide

OH

F

Cl

COCl Cl

Optically active and biodegradable deltamethrin6 4 has taken a large share of the insecticide market, and asymmetric hydrogenation is used in the commercial synthesis of DOPA 5 used to treat Parkinson’s disease.7 These achievements depend both on the development of new methods and on strategic planning:8 the twin themes of this book.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

4

1 Planning Organic Syntheses: Tactics, Strategy and Control

Br

HO

R R

O S

Br O

NH2

O

H

CO2H

HO

CN 4 deltamethrin

5 (S)-L-(–)-DOPA

To make any progress in this advanced area, we have to assume that you have mastered the basics of planning organic synthesis by the disconnection approach, roughly the material covered in our previous books.9 There, inspecting the target molecule, identifying the functional groups, and counting up the relationships between them usually gave reliable guidelines for a logical synthesis. All enones were tackled by some version of the aldol reaction; thus 6 would require the attack of enolate 7 on acetone. We hope you already have the critical judgement to recognise that this would need chemoselectivity in enolising 7 rather than acetone or 6, and regioselectivity in enolising 7 on the correct side. O

O

aldol

O

+ O 7

6

In this book we shall explore two new approaches to such a problem. We shall see how to make specific enol equivalents for just about any enolate you might need, and we shall see that alternative disconnections such as 6a, the acylation of a vinyl anion 8, can be put into practice. Another way to express the twin themes of this book is strategy and control: we solve problems either by finding an alternative strategy or by controlling any given strategy to make it work. This will require the introduction of many new methods - a whole chapter will be devoted to reagents for vinyl anions such as 8, and this will mean exploring modern organometallic chemistry. O

O +

6a

X

8

9

We shall also extend the scope of established reactions. We hope you would recognise the aldol disconnection in TM 10, but the necessary stereochemical control might defeat you. An early section of this book describes how to control every aspect of the aldol reaction: how to select which partner, i.e. 11 or 12, becomes an enolate (chemoselectivity), how to control which enolate of the ketone 12 is formed (regioselectivity), and how to control the stereochemistry of the product 10 (stereoselectivity). As we develop strategy, we shall repeatedly examine these three aspects of control. OH

O

O 2

R

R1

R1

H

?

10

O

aldol

11

+

O R2

R2 ?

12

The target molecules we shall tackle in this book are undoubtedly more difficult in several ways than this simple example 10. They are more complex quantitatively in that they combine functional

1 Planning Organic Syntheses: Tactics, Strategy and Control

5

groups, rings, double bonds, and chiral centres in the same target, and qualitatively in that they may have features like large rings, double bonds of fixed configuration, or relationships between functional groups or chiral centres which no standard chemistry seems to produce. Molecules 1 to 5 are examples: a quite different one is flexibilene 13, a compound from Indonesian soft coral. It has a fifteen-membered ring, one di- and three tri-substituted double bonds, all E but none conjugated, and a quaternary centre. Mercifully there are no functional groups or chiral centres. How on earth would you tackle its synthesis? One published synthesis is by McMurry.10 OMe 1. 2 x BuLi

H 2.

Cp2ZrHCl

1. PCC

HO

O

15; 78% yield

14

2. H , CH(OMe)3

17

MeO 16; 83% yield

OMe MeO

O

ZrCp2Cl 17

18 Pd(OCOCF3)2

TiCl3

OHC

O

Zn/Cu

13; flexibilene, 52% yield

19; 76% yield from 16 + 18

This short synthesis uses seven metals (Li, Cr, Zr, Pd, Ti, Zn, and Cu), only one protecting group, achieves total control over double bond geometry, remarkable regioselectivity in the Zr-Pd coupling reaction, and a very satisfactory large ring synthesis. The yield in the final step (52%) may not look very good, but this is a price worth paying for such a short synthesis. Only the first two steps use chemistry from the previous books: all the other methods were unknown only ten years before this synthesis was carried out but we shall meet them all in this book. An important reason for studying alternative strategies (other than just making the compound!) is the need to find short cheap large scale routes in the development of research lab methods into production. All possible routes must be explored, at least on paper, to find the best production method and for patent coverage. Many molecules suffer this exhaustive process each year, and some sophisticated molecules, such as Merck’s HIV protease inhibitor 20, a vital drug in the fight against AIDS, are in current production on a large scale because a good synthesis was found by this process.11 N

OH

Ph H N

N

N t-BuNH

O

OH

O

20; Crixivan

You might think that, say organometallic chemistry using Zr or Pd would never be used in manufacture. This is far from true as many of these methods are catalytic and the development of polymer-supported reagents for flow systems means that organo-metallic reagents or enzymes may be better than conventional organic reagents in solution with all the problems of by-product disposal and solvent recovery. We shall explore the chemistry of B, Si, P, S, and Se, and of metals

6

1 Planning Organic Syntheses: Tactics, Strategy and Control

such as Fe, Co, Ni, Pd, Cu, Ti, Sn, Ru and Zr because of the unique contribution each makes to synthetic methods. In the twenty years since McMurry’s flexibilene synthesis major developments have changed the face of organic synthesis. Chiral drugs must now be used as optically pure compounds and catalytic asymmetric reactions (chapters 25 and 26) have come to dominate this area, an achievement crowned by the award of the 2001 Nobel prize for Chemistry to Sharpless, Noyori and Knowles. Olefin metathesis (chapter 15) is superseding the Wittig reaction. Palladium-catalysed coupling of aromatic rings to other aromatic rings, to alkenes, and to heteroatoms (chapter 18) makes previously impossible disconnections highly favourable. These and many more important new methods make a profound impact on the strategic planning of a modern synthesis and find their place in this book.

A Modern Synthesis: Fostriecin (CI-920) The anti-cancer compound Fostriecin 21 was discovered in 1983 and its stereochemistry elucidated in 1997. Not until 2001 was it synthesised and then by two separate groups.12 Fostriecin is very different from flexibilene. It still has alkene geometry but it has the more challenging threedimensional chirality as well. It has plenty of functionality including a delicate monophosphate salt. A successful synthesis must get the structure right, the geometry of the alkenes right, the relative stereochemistry right, and it must be made as a single enantiomer. O HO O

O P

Na OH

OH

O

O OH 21; Fostriecin (CI-920)

The brief report of Jacobsen’s total synthesis starts with a detailed retrosynthetic analysis. The compound was broken into four pieces 21a after removal of the phosphate. The unsaturated lactone 24 (M is a metal) could be made by an asymmetric oxo-Diels-Alder reaction from diene 22 and ynal 23. The epoxide 25 provides a second source of asymmetry. One cis alkene comes from an alkyne 26 and the rest from a dienyl tin derivative 27. O O P HO O

21a Disconnection of Fostriecin (CI-920)

O

Na OH

OH

O OH OR

DielsAlder

RO

O O

O SiMe3 22

23

24

S

SnBu3

H

M

O

S

O 25

SiMe3 26

27

The synthesis is a catalogue of modern asymmetric catalytic methods. The epoxide 25 was resolved by a hydrolytic kinetic resolution (chapter 28) using a synthetic asymmetric cobalt complex. The asymmetric Diels-Alder reaction (chapter 26) was catalysed by a synthetic chromium

1 References

7

complex. The vinyl metal derivative 24 was made by hydrozirconation of an alkyne (this at least is similar to the flexibilene synthesis) and the secondary alcohol chiral centre was derived from the dithian 26 by hydrolysis to a ketone and asymmetric reduction with a synthetic ruthenium complex (chapter 24). The dienyl tin unit 27 was coupled to the rest of the molecule using catalytic palladium chemistry (chapter 18). Almost none of these catalytic methods was available in 1983 when flexibilene was made and such methods are a prominent feature of this book. Organic synthesis nowadays can tackle almost any problem.13 Please do not imagine that we are abandoning the systematic approach or the simpler reagents of the previous books. They are more essential than ever as new strategy can be seen for what it is only in the context of what it replaces. Anyway, no-one in his or her right mind would use an expensive, toxic, or unstable reagent unless a friendlier one fails. Who would use pyrophoric tertiary butyl-lithium in strictly dry conditions when aqueous sodium hydroxide works just as well? In most cases we shall consider the simple strategy first to see how it must be modified. The McMurry flexibilene synthesis is unusual in deploying exotic reagents in almost every step. A more common situation is a synthesis with one exotic reagent and six familiar ones. The logic of the previous books is always our point of departure.

The organisation of the book The book has five sections: A: B: C: D: E:

Introduction, selectivity, and strategy Making Carbon-Carbon bonds Carbon-Carbon double bonds Stereochemistry Functional Group Strategy

The introductory section uses aldol chemistry to present the main themes in more detail and gives an account of the three types of selectivity: chemo-, regio-, and stereo-selectivity. We shall explore alternative strategies using enones as our targets, and discuss how to choose a good route using cyclopentenones as a special case among enones. Each chapter develops strategy, new reagents, and control side-by-side. To keep the book as short as possible (like a good synthesis), each chapter in the book has a corresponding chapter in the workbook with further examples, problems, and answers. You may find that you learn more efficiently if you solve some problems as you go along.

References General references are given on page 893 1. R. B. Woodward, Pure Appl. Chem., 1973, 33, 145; A. Eschenmoser and C. E. Wintner, Science, 1977, 196, 1410; A. Eschenmoser, Angew. Chem., Int. Ed. Engl., 1988, 27, 5. 2. Y. Kishi, Tetrahedron, 2002, 58, 6239. 3. E. J. Corey, K. C. Nicolaou, and L. S. Melvin, J. Am. Chem. Soc., 1975, 97, 654; G. Stork and S. D. Rychnovsky, J. Am. Chem. Soc., 1987, 109, 1565; Pure Appl. Chem., 1987, 59, 345; A. F. Sviridov, M. S. Ermolenko, D. V. Yashunsky, V. S. Borodkin and N. K. Kochetkov, Tetrahedron Lett., 1987, 28, 3835, and references therein. 4. W. C. Still, J. Am. Chem. Soc., 1979, 101, 2493. See also S. L. Schreiber and C. Santini, J. Am. Chem. Soc., 1984, 106, 4038; T. Takahashi, Y. Kanda, H. Nemoto, K. Kitamura, J. Tsuji and Y. Fukazawa, J. Org. Chem., 1984, 51, 3393; H. Hauptmann, G. Mühlbauer and N. P. C. Walker, Tetrahedron Lett., 1986, 27, 1315; T. Kitahara, M. Mori and K. Mori, Tetrahedron, 1987, 43, 2689.

8

1 Planning Organic Syntheses: Tactics, Strategy and Control

5. P. A. Worthington, ACS Symposium 355, Synthesis and Chemistry of Agrochemicals, eds D. R. Baker, J. G. Fenyes, W. K. Moberg, and B. Cross, ACS, Washington, 1987, p 302. 6. M. Elliott, A. W. Farnham, N. F. Janes, P. H. Needham, and D. A. Pullman, Nature, 1974, 248, 710; M. Elliott, Pestic. Sci., 1980, 11, 119. 7. J. Halpern, H. B. Kagan, and K. E. Koenig, Morrison, vol 5, pp 1–101. 8. Corey, Logic; Nicolaou and Sorensen. 9. Designing Syntheses, Disconnection Textbook, and Disconnection Workbook. 10. J. McMurry, Acc. Chem. Res., 1983, 16, 405. 11. D. Askin, K. K. Eng, K. Rossen, R. M. Purick, K. M. Wells, R. P. Volante and P. J. Reider, Tetrahedron Lett., 1994, 35, 673; B. D. Dorsey, R. B. Levin, S. L. McDaniel, J. P. Vacca, J. P. Guare, P. L. Darke, J. A. Zugay, E. A. Emini, W. A. Schleif, J. C. Quintero, J. H. Lin, I.-W. Chen, M. K. Holloway, P. M. D. Fitzgerald, M. G. Axel, D. Ostovic, P. S. Anderson and J. R. Huff, J. Med. Chem., 1994, 37, 3443. 12. D. L. Boger, S. Ichikawa and W. Zhong, J. Am. Chem. Soc., 2001, 123, 4161; D. E. Chavez and E. N. Jacobsen, Angew. Chem., Int. Ed., 2001, 40, 3667. 13. D. Seebach, Angew. Chem. Int. Ed., 1990, 29, 1320; K. C. Nicolaou, E. J. Sorensen and N. Winssinger, J. Chem. Ed., 1998, 75, 1225.

2 Chemoselectivity Definitions Introduction: three types of control Chemoselectivity: simple examples and rules Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug Protection to allow a less reactive group to react When Protection is not Needed Dianions: wasting reagent to achieve selectivity Chemoselectivity by Reagent: The Pinacol Rearrangement Selectivity between secondary and tertiary alcohols by reagent Corey’s longifolene synthesis Chemoselectivity in Enol and Enolate Formation General discussion of enols and enolates Formation of specific enol equivalents Lithium enolates, enamines and silyl enol ethers Enamines Silyl enol ethers Synthesis of the ant alarm pheromone mannicone Examples of Chemoselectivity in Synthesis Synthesis of lipstatin, rubrynolide and hirsutene

Definitions Introduction: three types of control Behind all grand strategic designs in organic synthesis must lie the confidence that molecules can be compelled to combine in the ways that we require. We shall call this control and divide it into three sections by mechanistic arguments. These sections are so important that we shall devote the next three chapters to the more detailed explanation of just what the divisions mean. If you can recognise what might go wrong you are in a better position to anticipate the problem and perhaps avoid it altogether. Our three types of control are over chemoselectivity (selectivity between different functional groups), regioselectivity (control between different aspects of the same functional group), and stereoselectivity (control over stereochemistry). Examples of selectivity of all three kinds are given in The Disconnection Approach: Chemoselectivity in chapter 5, Regioselectivity in chapter 14, and Stereoselectivity in chapters 12 and 38. These aspects will not be addressed again in the present book.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

10

2 Chemoselectivity

Chemoselectivity: simple examples and rules Chemoselectivity is the most straightforward of the three types and might seem too elementary to appear in an advanced textbook. Counting the number of protecting groups in the average synthesis reveals this as a naive view. Selectivity between functional groups might involve: (a) Selective reaction of one among several functional groups of different reactivity, as in the reduction of the keto-acid 2 to give either product 1 or 3 at will. O

O

OH

?

OH

?

CO2H

1

CO2H

2

3

(b) Selective reaction of one of several identical functional groups, as in the conversion of the symmetrical diacid 5 to the half ester, half acid chloride 4, or the lactone 6 in which one of the two acids has been reduced. There is a more subtle example of this at the end of the chapter. ?

COCl

MeO2C

CO2H

HO2C

substitution

4

? reduction

O 6

5

O

(c) Selective reaction of a functional group to give a product which could itself react with the same reagent, as in the classical problem of making a ketone 8 from an acid derivative 7 without getting the alcohol 9 instead. O R1

OEt

OH

O

R2–Metal

and not R2

R1

?

8

7

R1

R2

R2

9

Organic chemists are developing ever more specific reagents to do these jobs. These reagents must carry out the reaction they are designed for and must not: (i) react with themselves. (ii) react with functional groups other than the one they are aimed at. (iii) react with the product. Proviso (ii) is obvious, but (i) and (iii) perhaps need some explanation. It seems hardly worth stating that a reagent should not react with itself, but it is only too easy to suggest using a reagent such as 11 without realising that the organo-metallic reagent will act as a base for its own hydroxyl group 12 and destroy itself. The traditional solution to this problem is protection of the OH group in 10 but ideally we should like to avoid protection altogether though this is not yet possible. Br

HO

Mg

MgBr

HO

10

HO

11

n-butanol H , H 2O

O

H 12

MgBr

O

CH3 13

work-up

OH

CH3

n-butanol

2 Definitions

11

Proviso (iii) is more obvious and yet perhaps more often catches people out. It is not always clear in exactly what form the product is produced in the reaction mixture, though a good mechanistic understanding and careful thought should reveal this. The reaction between the simple aldehyde 14 and chloral (Cl3C.CHO) looks like a straightforward route to the aldol 17, and might reasonably be carried out via the enamine 16. CHO +

HN O

14

O

15

OH

Cl3C.CHO

N

CHO

Cl3C

?

17

16

However, mixtures of 16 and chloral, in any proportion, give only the 2:1 adduct 20 which can be isolated in 83% yield.1 Obviously the immediate product 19 reacts with chloral at least as fast as does 16. Fortunately the synthesis can be rescued by acid-catalysed cleavage of 20 with HCl which gives a good yield of the target 17. CCl3

CCl3

Cl3C O

O

O

Cl3C.CHO

N N

O

Cl3C

O

N

O 18

O

19

20

Enamines are excellent at Michael additions and another plausible synthesis which “goes wrong” is the addition of acrolein to cyclohexanone mediated by the enamine 21 formed this time with pyrollidine.

O

O

CHO

N

+ HN

?

CHO

21

22

If the product is isolated by distillation, a good yield (75%) of the bicyclic ketone 23 is obtained.2 A more detailed investigation disclosed that 24 is the immediate product, that 23 is formed from it on distillation, and that the expected Michael adduct 22 can be isolated in good yield simply by the hydrolysis of 24. In other words, don’t distil! If things “go wrong” in a synthesis, this may be a blessing, as here. There are lots of ways to control Michael additions, but few ways to make bicyclic ketones like 23, and this is now a standard method.3 The moral is to make sure you know what is happening, and to be prepared to welcome the useful and unexpected result. O

N

23

O N

24

N

12

2 Chemoselectivity

Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug We need to see some of these principles in action and a proper synthesis is overdue. The anti-malarial drug amopyroquine 25 might have been derived from quinine as it has a quinoline nucleus. It also has five functional groups – three amines (all different - one aromatic, one tertiary, and one secondary), a phenol and an aryl chloride. There are four rings, three aromatic and one saturated heterocyclic. OH

OH a N

HN

b

N

Cl

N

Cl

25 Amopyroquine

N

HN

25a, b Amopyroquine disconnections

There are many possible disconnections, but we should prefer to start in the middle of the molecule to achieve the greatest simplification. Disconnection 25a would require a nucleophilic displacement (X ⫽ a leaving group) on an unactivated benzene ring 27 and looks unpromising. Disconnection 25b requires nucleophilic displacement at position 4 in a pyridine ring, an acceptable reaction because of the electron-withdrawing effect of the nitrogen atom in the ring, so this is the better route, though we may be apprehensive about controlling the chemoselectivity as there are three potential nucleophiles in 26 and two potential electrophiles in 28. X +

+

25 N

Cl

28

OH

a

b

N

H2N

N

Cl

NH2

OH

N

X

29

26

27

Further disconnections of 26 by the Mannich reaction4 and of 28 by standard heterocyclic methods give simple starting materials.5 OH

OH

Mannich

+ CH2O +

N

H2N

H2N

HN

30

26a X

O

O quinoline synthesis

FGI

+ N

Cl 28

N H

Cl 31

HO

NH2

Cl 32

Protection to allow a less reactive group to react Now the fun begins! Attempted Mannich reaction on the aminophenol 30 would be dominated by the more nucleophilic NH2 group and is no good. Acylation moderates the NH2 group by delocalisation

2 When Protection is not Needed

13

and 33 is a good choice for starting material as it is paracetamol, the common analgesic. Mannich reaction now chemoselectively gives 34 and alkaline hydrolysis of the amide gives 26. OH

CH2O

·· N H

AcOH

H2N

OH

OH

O

Ac2O

pyrrolidine

30

N

AcHN

33

34

Michael addition of acrylic acid to the chloroamine 32 is straightforward and Friedel-Crafts cyclisation of 35 gives only 31, presumably because the position next to the chlorine atom is slightly disfavoured both sterically and electronically. Chlorination and oxidation are conveniently carried out in the same step and the two halves (26 and 28) of this convergent synthesis are combined to give amopyroquine 25. Cl

CO2H CO2H

PPA

32

POCl3

26

I2

heat in aqueous ethanol

25

31 N H

Cl

Cl

N 28; X = Cl

35

In the last step we return to the original question of chemoselectivity: Only the primary amine in 26 reacts because it is more nucleophilic than OH and because the more nucleophilic tertiary amine adds reversibly – it cannot lose a hydrogen atom as it does not have one. Only the 4-chlorine atom in the pyridine 28 reacts, presumably because addition to the other position would require the disruption of both aromatic rings. Though this compound has been succeeded by better antimalarials, its synthesis illustrates the all-important principle that predictions of chemoselectivity must be based on sound mechanistic understanding. If doubt remains it is worth trying a model reaction on simpler compounds or, of course, an alternative strategy.

When Protection is not Needed Dianions: wasting reagent to achieve selectivity In that synthesis we moderated an over-reactive amino group by protection. Sometimes, protection is not necessary if we are prepared to squander some of our reagents. A trivial example is the addition of methyl Grignard to the ketoacid 36. We have already seen how acidic protons destroy Grignard reagents, but if we are prepared to waste one molecule of the Grignard, we get automatic protection of the carboxylic acid by deprotonation. Nucleophilic MeMgI will not add to the anion of a carboxylic acid but adds cleanly to the ketone to give, after workup, the alcohol 37. O

O

MeMgI

CO2H

R

MeMgI

CO2 Mg

R

2 36; Mg salt

36 O R

OH CO2 Mg

Me 37; Mg salt

2

H H2O

R

CO2H Me 37

14

2 Chemoselectivity

At first sight, the synthesis of Z-38 by the Wittig reaction seems too risky. The phosphonium salt 39 has a more acidic proton (CO2H) than the one we want to remove to make the ylid, and the aldehyde 40 not only also has an acidic proton (OH), but it prefers to remain as the cyclic hemiacetal 41 so that there is no carbonyl group at all. Wittig

Z-38 HO

Wittig CO2H reagent 39

Ph3P

CO2H

+ hemi-acetal 41

HO

OH

O

CHO

hydroxy-aldehyde 40

However, simply using a large excess of base makes the reaction work without any protection. The phosphonium salt 39 does indeed lose its first proton from the CO2H group 42, but the second molecule of base forms the ylid 43 as the two anions are far enough apart not to influence each other.6 Base also catalyses the equilibrium between the anions 44 and 45 so that 43 and 45 can react to give the target molecule. The transition state for this reaction has three partial negative charges, but they are well apart from each other and there is obviously not too much electrostatic repulsion as the reaction goes well. This case is opposite to the previous ones: careful mechanistic analysis shows that expected chemoselectivity problems do not materialise. base

base

Ph3P

39

Ph3P

CO2 42

CO2 43

base

41 O 44

O

O

Z-38

CHO 45

Chemoselectivity by Reagent: The Pinacol Rearrangement So far we have discussed chemoselectivity between different functional groups. The situation gets more complicated if the functional groups are similar, or even the same. The pinacol rearrangement is a useful route to carbonyl compounds from diols, the classical example being the rearrangement of 46 in acid solution to give the t-alkyl ketone 48. There are no chemoselectivity problems here: the two hydroxyl groups in 46 are the same so it does not matter which gets protonated and, in the rearrangement step 47, all four potential migrating groups are methyl. H

HO

OH

46 'pinacol'

Me migration

Me

HO

OH2

HO

47

O 48 'pinacolone'

Selectivity between secondary and tertiary alcohols by reagent Unsymmetrical diols provide a serious problem of chemoselectivity with an ingenious solution.7 Treatment of the diol 49 with acid leads to loss of OH from what would be the more stable t-alkyl cation and hence, by hydrogen shift, to the ketone 51.

2 Chemoselectivity by Reagent: The Pinacol Rearrangement

H

H migration

H

H

R

R

15

H

R

H

R

HO

OH

HO

49

HO

OH2

O 51

50

The alternative, more interesting rearrangement to give 53 can be initiated by tosylation of the diol 49 in weak base. It is impossible to tosylate tertiary alcohols under these conditions, as both the t-alcohol and TsCl are large, so only the secondary alcohol becomes sulfonylated and so leaves, and the rearranged ketone 53 is the only product.

H

H

TsCl

R HO

OH

Me migration

Me

pyridine

TsO

49

Me H

R

H R

OH

OH

R

52

O 53

Corey’s longifolene synthesis The question of which group migrates in a pinacol rearrangement is also a question of chemoselectivity, and usually groups that can participate because they have lone pair or π-electrons migrate best. In Corey’s longifolene synthesis,8 the 6/7 fused enone 54 was an important intermediate. Synthesis from the readily available Robinson annelation product 57 is very attractive, but this demands a ring expansion step such as the pinacol rearrangement of 55 of unknown selectivity. 1,2-Diols such as 55 normally come from the hydroxylation of an alkene, in this case the diene 56 which might be made by a Wittig reaction on the dione 57. Every step in this sequence raises a question of chemoselectivity. Which of the two ketones in 57 is more reactive? Which of the two double bonds in 56 is more easily hydroxylated? Which side of the ring migrates in the pinacol rearrangement on 55?

pinacol

O

O

O

O

?

HO

O

FGI

Wittig

?

?

HO O

54

55

56

57

One of the ketones in 57 is conjugated, and one is not. The unconjugated one is less stable and we can therefore use thermodynamic control if we protect as an acetal, a reversible process. The unconjugated ketone would also be more kinetically reactive towards the Wittig reagent. Of the two double bonds in 59, the one outside the ring is more reactive towards electrophilic reagents, again for both kinetic and thermodynamic reasons. The tosylation route ensures that the right OH group leaves in the pinacol rearrangement and because the remaining π-bond migrates better than the simple alkyl group when 60 rearranges with a weak Lewis acid, all is well. The synthesis therefore follows the route below, with all questions of chemoselectivity neatly solved. The acetal protecting group was also useful later in the synthesis.

16

2 Chemoselectivity

O

O

O

O

Ph3P HO

OsO4

OH

57 TsOH

pyridine

O

E and Z-59

58 O

O

O

O

O

HO

HO

O

LiClO4

TsCl pyridine

TsO

HO

O

CaCO3 THF

60

61

Chemoselectivity in Enol and Enolate Formation General discussion of enols and enolates We have concentrated so far on two functional groups within the same molecule. The chemoselectivity problem is just as important when we want two molecules to react together in a certain way, but, because both molecules have similar functional groups, the reaction can occur the other way round, or one of the molecules may react with itself and ignore the other. This problem is particularly acute in reactions involving enolisation. The alkylation or acylation of enols or enolates and the reaction of one carbonyl compound with another, the aldol reaction, are classical and important examples summarised in the general scheme below. We shall concentrate in this chapter on the chemoselectivity of these processes, that is we shall look at the enolisation of esters, aldehydes, and the like. OH

O

R1

O

R1 carbonyl compound

enol O

R1 enolate

O

aldol

O

O

OH R2

R1

R1

acylation O

aldol H

R2

R1

RBr

O

aldol R2

H

O

alkylation

R2

OH

O R2

R2

X H

R2

OH

O R2

R1

O R2

R1

Reaction of an ester 62 with its own alkoxide ion produces a small amount of enolate 63 that reacts with unenolised ester to give the ketoester 64. This reaction, though useful in its own right, precludes the direct alkylation of esters under these conditions.

H

H

H

EtO

R

OEt

R

R

CO2Et

OEt

R

R

O

O

62; ester

63; ester enolate

CO2Et O 64; β-keto-ester

2 Chemoselectivity in Enol and Enolate Formation

17

Formation of specific enol equivalents What is needed for the alkylation is rapid conversion of the ester into a reasonably stable enolate, so rapid in fact that there is no unenolised ester left. In other words the rate of proton removal must be faster than the rate of combination of enolate and ester. These conditions are met when lithium enolates are made from esters with lithium amide bases at low temperature, often ⫺78 ⬚C. Hindered bases must be used as otherwise nucleophilic displacement will occur at the ester carbonyl group to give an amide. Popular bases are LDA (Lithium Di-isopropyl Amide, 66), lithium hexamethyldisilazide 67, and lithium tetramethylpiperidide 68, the most hindered of all. These bases are conveniently prepared from the amine, e.g. 65 for LDA, and BuLi in dry THF solution.

Me3Si

BuLi

N

H

N

Li

SiMe3

N

N

THF

Li

Li 65 Di-isopropylamine

66; LDA 67; LHMDS Li Di-isopropyl Amide Li HexaMethylDiSilyazide

68; LTMP Li TetraMethylPiperidide

Treatment of a simple ester 62 with one of these bases at ⫺78 ⬚C leads to a stable lithium enolate 70 by initial coordination of lithium to the carbonyl group 69 and proton removal via a six-membered cyclic transition state 69a.

H

H O

R

H

H

O

LDA –78 °C THF

OEt

R2N Li H

R

R2N Li H O

R

OEt

62; ester

H OLi

R

OEt

OEt

69a

69

+ R2NH

70

Lithium enolates, enamines and silyl enol ethers Direct alkylation of lithium enolates of esters9 62 and lactones 73, via the lithium enolates 71 and 74, with alkyl halides is usually successful.

H 1. LDA

R1

O

R1

CO2Et Br

2. R 2Br

OEt

R2

62

R2

Li R1

CO2Et

71 O

O

1. LDA

72 OLi

O

O

O 2. RBr

R 73

74

75

18

2 Chemoselectivity

More impressive and more important is the performance of these lithium enolates in aldol reactions. Ester enolates are awkward things to use in reactions with enolisable aldehydes and ketones because of the very efficient self-condensation of the aldehydes and ketones. The traditional solutions involve such devices as Knoevenagel-style reactions with malonates.11 Lithium enolates of esters, e.g. 76, react cleanly with enolisable aldehydes and ketones to give high yields of aldols,12 e.g. 79 in a single step also involving a six-membered cyclic transition state 77. O LDA

MeCO2Et

Li

O

OLi +

THF –78 °C

O

OEt OEt 76

77

OLi

OH CO2Et

H

CO2Et

H2O

79, 79-90%

78

They even react cleanly with formaldehyde, thus solving the problem that the Mannich reaction is not applicable to esters. The synthesis of the exo-methylene lactone 80 can be accomplished this way. Enone disconnection13 reveals formaldehyde as the electrophilic component in a crossed aldol reaction, realised with a lithium enolate 82.14 The mono-adduct 83 of formaldehyde and the lactone 81 can be isolated and the cautious dehydration step is to avoid migration of the double bond into the ring. H

H

α,β-unsaturated carbonyl

O

O O + CH2=O

O H

H

80

81 H

H O

LDA

O

CH2O

81

1. MsCl, Et 3N

OLi

80

O 2. pyridine

H

H OH 82

83

The same technique can even be applied to carboxylic acids themselves 84 providing two molecules of base are used. The first removes the acid proton to give the lithium salt 85 and the second forms the lithium enolate 86. O

R OH 84

LDA

O

R

OLi 85

LDA

OLi

R

OLi

2 x LDA –78 ˚C, THF

84

86

These lithium derivatives are also well behaved in alkylations and aldol reactions. Krapcho’s synthesis15 of the sesquiterpene α-curcumene 92 starts with the chemoselective condensation of

2 Chemoselectivity in Enol and Enolate Formation

19

the dilithium derivative of the acid 87 with the enolisable aldehyde 89. The aldol product 90 is converted into the β-lactone 91 and hence by heating and loss of CO2 into α-curcumene 92.

2 x LDA

OLi

CO2H

OLi 88

87

1. 2 x LDA

87

CHO

89 PhSO2Cl

90 73% yield

2. 89

CO2H

pyridine

OH 90, 73% yield

140 °C

O

O

92; α-curcumene, 90% yield

91, 77% yield

You might be forgiven for thinking that lithium enolates solve all problems of enolate chemoselectivity at a stroke and wonder why they are not always used. They are very widely used, but they require strictly anhydrous conditions at low temperatures (usually ⫺78 ⬚C, the temperature of a dry ice/acetone bath) and no-one in their right mind would use these conditions if mixing the reagents in ethanol at room temperature with a catalytic amount of NaOH did nearly as well. These are the conditions of many simple aldol reactions and are preferred where practical, particularly in industrial practice. The intermediate 93 was needed in a synthesis of geiparvirin. The best aldol disconnection in the middle of the molecule gives a ketone 94, that must be enolised in the only possible position, and then react with an unenolisable and more electrophilic aldehyde 95. No selectivity problems arise and an equilibrating aldol reaction between 94 and 95 catalysed by NaOEt in EtOH gives 93 in 89% yield.16 O

O

H

aldol

Ph

HO 93

+ O

HO 94

Ph 95

Enamines Lithium enolates do not even solve all problems of chemoselectivity: most notoriously, they fail when the specific enolates of aldehydes are needed. The problem is that aldehydes self-condense so readily that the rate of the aldol reaction can be comparable with the rate of enolate formation by proton removal. Fortunately there are good alternatives. Earlier in this chapter we showed examples of what can go wrong with enamines. Now we can set the record straight by extolling the virtues of the enamines 96 of aldehydes.17 They are easily made without excessive aldol reaction as they are much less reactive than lithium enolates, they take part well in reactions such as Michael additions, a standard route to 1,5-dicarbonyl compounds, e.g. 97.18

20

2 Chemoselectivity

CHO

NR2

NR2 R2NH

OMe R

96

OMe

R

R

O

O

NR2 CHO OMe

R

H

R

H2O

O

CO2Me 97

An impressive example19 is the Robinson annelation of the unsaturated aldehyde 98 where neither aldol reaction nor double bond migration in the enamine 99 interferes. The 1,5-dicarbonyl compound 100 cyclises spontaneously to the enone 101. O

O

O

CHO

N +

Ph

OHC

N H

Ph

Ph

98

Ph 100

99

101

Silyl enol ethers For all their usefulness, enamines have now largely been superseded by silyl enol ethers. These (102-104) can be made directly with Me3SiCl from the lithium enolates of esters or acids or from aldehydes under milder conditions with a tertiary amine. The silicon atom is an excellent electrophile with a strong preference for more electronegative partners and it combines with the oxygen atom of an enolate so rapidly that no self condensation occurs even with aldehydes. Silyl Enol Ethers of Esters O LDA R R OEt

OLi

Me3SiCl

OEt 102

OEt

Silyl Enol Ethers of Acids O LDA R R OH

OLi

Me3SiCl

OSiMe3 103

OLi O

H

Et3N

OH

R H

OSiMe3

R

Silyl Enol Ethers of Aldehydes R

OSiMe3

R

Me3SiCl

OSiMe3

R H 104

The silyl enol ethers 102 and 104 are shown as single geometrical isomers for convenience: in fact they are normally formed as mixtures, though this does not usually affect their reactions. They are thermodynamically stable compounds but are easily hydrolysed with water or methanol and are usually prepared when they are needed. They are much less reactive than lithium enolates, or even enamines, and their reactions with electrophiles are best catalysed by Lewis acids, often

2 Chemoselectivity in Enol and Enolate Formation

21

TiCl4. The aldehydes 105 and 108, the one branched and the other not, are simply converted into their silyl enol ethers 106 and 109 and combined with two different enolisable aldehydes to give high yields of aldol products 107 and 110 without any self condensation of any of the four aldehydes or any cross-condensation the wrong way round.20 Ph

Me3SiCl

CHO

CHO

OSiMe3

Et3N

Ph

TiCl4

CHO OH 107; 86% yield

106

105

CHO Me3SiCl

Ph

CHO

Ph

OSiMe3

n-PrCHO

Ph

TiCl4

Et3N

OH 108

110; 78% yield

109

The silyl enol ethers of esters, e.g. 111 and lactones, e.g. 114 similarly take part in efficient aldol reactions with enolisable aldehydes and ketones with Lewis acid catalysis, again with complete regioselectivity. Example 113 is particularly impressive as the very enolisable ketone gives a high yield of an aldol product with two adjacent quaternary centres.21 Ph 1. LDA

CO2Et

OSiMe3

2. Me 3SiCl

OEt

111

O

Ph

TiCl4

113, 94% yield

112

O

Me3SiO

O

1. LDA

O

2. Me 3SiCl

CO2Et OH

+

O

TiCl4

O

O

OH

115

114

116

Synthesis of the ant alarm pheromone mannicone The synthesis of the ant alarm pheromone mannicone 117 is a good example. Enone disconnection reveals that we need a crossed aldol condensation between the symmetrical ketone 118, acting as the enol component, and the enolisable aldehyde 119. Mannicone: Analysis O

O

H

aldol

+ O enone

117

118

119

The ketone gives a mixture of geometrical isomers of the silyl enol ether 120 which condense with the aldehyde 119 to give the aldol 121 as a mixture of diastereoisomers which dehydrates to mannicone 117 in acid.22 It is particularly impressive that the optically active aldehyde 119 has its

22

2 Chemoselectivity

stereogenic centre at the enol position and yet optically active mannicone is formed by this route without racemisation. Mannicone: Synthesis OSiMe3

O

Me3SiCl

119

Et3N

TiCl4

118

OH TsOH

TM 117

120

121

We shall discuss further aspects of the aldol reaction in the next two chapters where we shall see how to control the enolisation of unsymmetrical ketones, and how to control the stereochemistry of aldol products such as 121. We shall return to a more comprehensive survey of specific enol equivalents in chapter 10. In this chapter we are concerned to establish that chemoselective enolisation of esters, acids, aldehydes, and symmetrical ketones can be accomplished with lithium enolates, enamines, or silyl enol ethers, and we shall be using all these intermediates extensively in the rest of the book.

Examples of Chemoselective Reactions in Synthesis Synthesis of lipstatin The synthesis of lipstatin 122 is too complex to discuss here in detail but an early stage in one synthesis uses a clever piece of chemoselectivity.23 Kocienski planned to make the β-lactone by a cycloaddition with the ketene 124 and to add the amino acid side chain 123 by a Mitsunobu reaction involving inversion. They therefore needed Z,Z-125 to join these pieces together. This was to be made in turn by a Wittig reaction from 126. The problem now is that 126 is symmetrical and cannot carry stereochemistry and that aldehydes are needed at both ends.

•

123

O O O

O

NHCHO

NHCHO

OH

O

124

CO2H Me3Si CHO

H

OH

H OHC

122; lipstatin

Z,Z-125

CHO 126

The way they solved the problem was this. (S)-(⫺)-Malic acid is available cheaply. Its dimethyl ester 127 could be chemoselectively reduced by borane to give 128. Normally borane does not reduce esters and clearly the borane first reacts with the OH group and then delivers hydride to the nearer carbonyl group. The primary alcohol was chemoselectively tosylated 129 and the remaining (secondary) OH protected with a silyl group 130 (TBDMS stands for t-butyldimethylsilyl and is sometimes abbreviated to TBS). Now the remaining ester can be reduced to an aldehyde 131 and protected 132. Displacement of tosylate by cyanide puts in the extra carbon atom 133 and reduction gives 134, that is the dialdehyde 126 in which one of the two aldehydes is protected. This compound was used in the successful synthesis of lipstatin.

2 Examples of Chemoselective Reactions in Synthesis

OH MeO2C

OH

BH3•SMe2 NaBH4

CO2Me

OH TsCl

HO

TsO

CO2Me

127; (S)-(–)-malic acid dimethyl ester

128; 83% yield OTBDMS

TsO

imidazole

OTBDMS TsO

CH2Cl2

130; 92% yield OTBDMS TsO

OTBDMS

DMSO

(i-PrO)3CH

CHO

TsOH

131; 78% yield

NaCN

CH(Oi-Pr)2

129; 75% yield

DIBAL

CO2Me

CO2Me

pyridine CH2Cl2

THF

t-BuMe2SiCl

23

NC

CH(Oi-Pr)2

132; 92% yield

OTBDMS

DIBAL

OHC

CH2Cl2

133; 83% yield

CH(Oi-Pr)2 134; 81% yield

The synthesis of rubrynolide The synthesis of the natural product rubrynolide from the Brazilian tree Nectandra rubra presents rather different problems. When the synthesis was planned it was supposed that rubrynolide was a trans lactone 135 but the third centre was not defined. The synthesis revealed that this structure is wrong: the lactone is actually cis and the stereochemistry24 is 2S,4R,2’S 135a.

2'

2

4

HO

HO = 'R'

O

OH

O

OH O

O

135a; rubrynolide

135; supposed rubrynolide

The synthesis was planned around the reaction of a specific enolate of ester 136 with the epoxide 137. This reaction was expected to give mainly trans 138 and is chemoselective both because of the usual enolate problem and because 137 contains a terminal alkyne. The lithium enolate was too basic and the aluminium enolate was used instead. The reaction gave an 85:15 mixture of trans and cis 138 and also an 85:15 mixture of trans and cis 139 after cyclisation. Dihydroxylation by osmylation gave a mixture of diols 140: this was deliberate so that they could determine the stereochemistry at C-2’. To the surprise of the chemists, natural rubrynolide was identical to one of the minor (i.e. cis) diols in the 15% part of the mixture. Careful NMR analysis showed that it was 135a. 1. LDA Et2AlCl

Me N O O NMO N-Methylmorpholine N-oxide

R

2.

t-BuO

O

136

R

t-BuO

O

137 R O

O 139; 95% yield

OsO4 NMO

O

TsOH

OH

138; 56% yield R

HO O

OH O

140; 63% yield

24

2 Chemoselectivity

The synthesis of hirsutene Tietze’s synthesis of hirsutene 141, an alkaloid from Uncaria rhynchophylla used in Chinese traditional medicine and with promising activity against influenza viruses, uses many chemoselective reactions of which we shall discuss just three - one at the start, one in the middle, and one at the end of the synthesis.25

aldol

N

N H H H

H

OMe

MeO2C

NH2 N H 143; tryptamine

N

N H H

O

HO2C

OMe

O

MeO2C H

141; hirsutene

CO2Me 144

142

The first reaction was the combination of the simple keto-diacid derivative 144 with tryptamine 143. How to make 144 was the problem. The diacid or its diester (the biologist’s ‘oxaloacetate’) is readily available and they decided to hydrolyse the enolate of the diester 145 with aqueous NaOH. It seems a strange decision to attack an anion with another anion but the enolate 146 is delocalised so that one ester group 146b, but not the other, shares the negative charge. The ester that does not share the negative charge is preferentially attacked by hydroxide ion.

O

O

base

CO2Me

MeO2C

O

MeO

O OMe

O

O NaOH

MeO

OMe

144

H2O

O

145

146a

146b

The second is an aldol reaction between the enolate 148 of ‘Meldrum’s acid’ 147 and the enolisable aldehyde 149. Because the enolate 148 is exceptionally stable, it can be made from 147 with a weak base and chemoselectivity (enolisation of 149) is not a problem. The unsaturated ester 151 is used immediately in a Diels-Alder reaction.

149 O

O O

NH

N H H

O

O H3N

O

NH

N H H

O

O

O

NH

N H H O

O

O

NH3

O

O

O

O

O

O

2 AcO

147 Meldrum's acid

148; enolate

150

151

At the end of the synthesis, the curious alkene, better described as an enol ether, must be introduced. The anion of the ester in 142 was prepared in base and condensed 152 with ethyl

2 References

25

formate. The chemoselectivity required is that ester 142 should react only with ethyl formate and not with itself. There is a further complication: the first product 153 has a more acidic proton than that in 142 and will form the enolate 154 under the reaction conditions. The whole system is in equilibrium and must be driven over by irreversible deprotonation by a strong base. Either LDA or Na⫹ Ph3C⫺ will do. After work-up the stable conjugated enol 155 is formed. Finally the enol is converted into the enol ether with acidic methanol to give hirsutene itself.

142

work -up

base

base

O

H

O

OMe H

O OMe

152

H

H

O

O

H

O

O

H

OH

OMe H

OMe H

OMe H

153

154

155

References General references are given on page 893 1. K. C. Brannock, R. D. Burpitt, H. E. Davis, H. S. Pridgen, and J. G. Thweatt, J. Org. Chem., 1964, 29, 2579. 2. G. Stork and H. K. Landesman, J. Am. Chem. Soc., 1956, 78, 5129. 3. A. Z. Britten and J. O’Sullivan, Tetrahedron, 1973, 29, 1331. 4. Disconnection Approach, page 158. 5. J. H. Burckhalter, F. H. Tendick, E. M. Jones, P. A. Jones, W. F. Holcomb and A. L. Rawlins, J. Am. Chem. Soc., 1948, 70, 1363; R. Joly, J. Warnant and B. Goffinet, Roussel-Uclaf, French Patent 1,514,280, Chem. Abstr., 1969, 70, 68195. 6. H.-J. Bestmann, O. Vostrowsky, H. Paulus, W. Billmann and W. Stransky, Tetrahedron Lett., 1977, 121; H.-J. Bestmann, J. Süss and O. Vostrowsky, Liebig’s Annalen, 1981, 2117. 7. D. Diederich, Houben-Weyl, 1973, 7/2a, page 958. 8. E. J. Corey, M. Ohno, R. B. Mitra, and P. A. Venkatacherry, J. Am. Chem. Soc., 1964, 88, 478. 9. R. J. Cregge, J. L. Herrmann, C. S. Lee, J. E. Richman and R. H. Schlessinger, Tetrahedron Lett., 1973, 2425. 10. G. Posner and G. L. Loomis, J. Chem. Soc., Chem. Commun., 1972, 892; J. L. Herrmann and R. H. Schlessinger, J. Chem. Soc., Chem. Commun., 1973, 711. 11. Disconnection Approach, page 160. 12. M. W. Rathke, J. Am. Chem. Soc., 1970, 92, 3222; M. W. Rathke and D. F. Sullivan, J. Am. Chem. Soc., 1973, 95, 3050; M. W. Rathke, Organic Syntheses, 1973, 53, 66. 13. Disconnection Approach, page 149, 158. 14. P. Grieco, J. Chem. Soc., Chem. Commun., 1972, 1317. 15. A. P. Krapcho and E. G. E. Jahngen, J. Org. Chem., 1974, 39, 1322. 16. P. L. Creger, J. Am. Chem. Soc., 1970, 92, 1397; P. E. Pfeiffer, L. S. Silbert and J. M. Chrinko, J. Org. Chem., 1972, 37, 451. 17. G. Stork, A. Brizzolara, H. Landesman, J. Szmuszkovicz and R. Terrell, J. Am. Chem. Soc., 1963, 85, 207. 18. Disconnection Approach, chapter 21, page 170. 19. P. Vittorelli, J. Peter-Katalinic, G. Mukherjee-Müller, H.-J. Hansen and H. Schmid, Helv. Chim. Acta, 1975, 58, 1379. 20. T. Mukaiyama, K. Banno, and K. Narasaka, J. Am. Chem. Soc., 1974, 96, 7503. 21. T. Mukaiyama, Angew. Chem., Int. Ed., 1977, 16, 817. 22. T. Mukaiyama, Chem. Lett., 1976, 279. 23. A. Pommier, J.-M. Pons and P. J. Kocienski, J. Org. Chem., 1995, 60, 7334. 24. S. K. Taylor, J. A. Hopkins, K. A. Spangenberg, D. W. McMillen and J. B. Grutzner, J. Org. Chem., 1991, 56, 5951. 25. L. F. Tietze and Y. Zhou, Angew. Chem., Int. Ed., 1999, 38, 2045; L. F. Tietze, Y. Zhou and E. Töpken, Eur. J. Org. Chem., 2000, 2247.

3

Regioselectivity: Controlled Aldol Reactions Definition Introduction and definition Regioselectivity in enol and enolate formation Regioselectivity by conditions: acid or base Specific Enol Equivalents Regioselectivity of formation of enamines Lithium enolates and silyl enol ethers Regioselective Aldol Reactions Aldol reactions with specific enol equivalents Contrast with equilibrium methods Aldols with Lewis acid catalysis: silyl enol ethers Application to the synthesis of gingerol Reaction at O or C? Silylation, Acylation and Alkylation Naked enolates Alkylation at carbon, problems with enamines Application to the synthesis of lipoic acid Alkylation with tertiary alkyl groups Acylation at Carbon Lithium enolates of carboxylic acids Enamines and silyl enol ethers Reactions with Other Electrophiles α-Halo carbonyl compounds and epoxides Michael reactions A Final Example Double Michael reactions with enamines

Definition Introduction and definition Regioselectivity means controlling different aspects of the same functional group. Classic examples are controlling direct (1,2) or conjugate (Michael or 1,4) addition to unsaturated carbonyl compounds, a subject we shall tackle in chapter 9, or controlling electrophilic substitution on unsymmetrical alkenes, which we shall meet in chapter 17. In this chapter we shall continue our study of enolisation by looking at regioselectivity in aldol and related reactions.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

28

3 Regioselectivity: Controlled Aldol Reactions

Regioselectivity in enol and enolate formation Reactions in which the enol or enolate (or equivalent) of one carbonyl compound reacts with an electrophilic carbonyl compound (usually both are aldehydes or ketones) are often loosely called aldol reactions. In the last chapter we saw how the use of lithium enolates and other specific enolate equivalents conquers the problem of chemoselectivity in enolisation of aldehydes and acid derivatives. In this chapter, we are going to use the same intermediates to solve the problem of regioselectivity in crossed aldol reactions in which the enolising component is an unsymmetrical ketone.

OH

OH and/or

R

O

acid

R

1

O

base

O and/or

R

R 2

3

R

4

5

Regioselectivity by conditions: acid or base The fundamental mechanistic distinction on which all methods ultimately depend is a precarious difference between kinetic and thermodynamic enolisation. A ketone 3, in which R is a simple alkyl group, has protons which are slightly more acidic on the less substituted side, the methyl group. This is because each alkyl group reduces the number and, by a weak electron donation, the acidity of the remaining hydrogens. An analogy is that t-BuLi is a stronger base than s-BuLi which is in turn stronger than n-BuLi. Condensation of 3 with a non-enolisable aldehyde (to avoid problems of chemoselectivity!) tends to give more aldol from 4 than from 5 when catalysed by NaOH, but the distinction is often too small to be useful. The ketone 3; R ⫽ i-Pr does condense regioselectively at the methyl group with furfuraldehyde 6 and NaOH to give the enone 7 in a useful 80% yield.1 Kinetic control favours the less substituted enolate.

NaOH

+

O

CHO

O

O

O

3; R = i-Pr

(E)-7; 80% yield

6

By contrast, the enol, and also the enolate when in combination with a metal atom, is more stable when the double bond is more substituted. Our ketone 3 will tend to produce more of enol 2 than enol 1 in acid solution, particularly if the enols are in equilibrium. Once again the distinction is usually too small to be useful. One example where it does succeed is the acidcatalysed Robinson annelation of ketone 8 with butenone 10 where reaction occurs predominantly on the more substituted enol 9: the yield of 11 is only 50%, but it is at least a one-step operation. Thermodynamic control favours the more substituted enol.2

H H2SO4

+ OH

O 8

9

O

O 10

11; 50% yield

3 Specific Enol Equivalents

29

An important application of thermodynamic control occurs in the manufacture of Viagra 12, Pfizer’s treatment for impotence.3 The NE corner of the molecule comes from the pyrazole acid 13: removal of the hydrazine portion reveals the one piece of continuous carbon skeleton 14. 1,3-Dicarbonyl disconnection gives an unenolisable diester (an oxalate) and an unsymmetrical ketone, pentan-2-one 15. O

Me N

HN

EtO

N

O S O

O

N 2 x C—N

N

RO

2 x C—N

N

N N

Me

12; Sildenafil (Viagra™)

13 O

O RO

Me

2

1 O

O

RO 1,3-diCO

OR +

O

2

Claisen ester

3

14

O

15

Condensation of 15 with diethyl oxalate in base without any control gives 14 because the true product of the reaction is the stable enolate of 14. The enolate 16 at the methyl group is preferred and the stable enolate of 14 is also preferred to the alternative because it is less substituted. This product was condensed with hydrazine to give fi rst the pyrazole 13 and then Viagra. O

15 O

O

base

(CO2Et)2

16 O

O CO2Et

14; R = Et

O CO2Et

enolate of 14; R = Et

Specific Enol Equivalents A survey of the thousands of examples of traditional aldol reactions in the back of Organic Reactions volume 16 shows how feeble this effect often is. Hardly any compounds are formed in good yield by chemoselective reactions using catalytic acid or base alone. The situation is different with modern methods.

30

3 Regioselectivity: Controlled Aldol Reactions

Regioselectivity of formation of enamines Enamines show an amplified preference for the less substituted double bond: at first this seems to contradict what we have just said, but the effect is greatest in cyclic ketones, e.g. 17, with cyclic amines.4 † It is steric in origin and arises from the eclipsing of hydrogen atoms (A1,3 strain) shown in the more substituted enamine 19. Enamines of acyclic ketones can be persuaded to give only the less substituted regio-isomer by equilibration of the immonium salt in weak base.5 H

H

A1,3

18:19 90:10 + N

N H

O

pyrrolidine

17

H

+

H

H

strain

N

18

19

Lithium enolates The most important method6 for the regioselective synthesis of less substituted enolates is kinetic enolate formation with strong irreversible bases (LDA etc). Since the lithium enolate7 20 can be converted into the silyl enol ether8 21 directly without isolation, we have access to the two most valuable specific enol equivalents for the less substituted isomer. Alkylation of the lithium enolate of 23 goes more than 99% on the less substituted side.9 O

OLi LDA

R

R

3

OSiMe3

Me3SiCl

R

20

21

Silyl enol ethers There is no such perfect method for getting enolisation to go on the more substituted side. The best is thermodynamic control in the formation of the silyl enol ether,10,11 which gives an approximate 90:10 ratio of 22:25 from 23. Silyl enol ethers can be converted into lithium enolates with MeLi (the by-product is Me4Si: useful for NMRs) and hence we can achieve alkylation on the more substituted side, e.g. 26 is benzylated with PhCH2Br to give 27; R ⫽ CH2Ph in up to 84% yield.12 O

OSiMe3

OSiMe3

OLi

Me3SiCl Et3N

Me3SiCl

LDA THF –78 °C

22

23

24

MeLi

RBr

OLi

R

O

O R

RBr

26 †

25

enamine review: Whitesell, Synthesis, 1983, 517

27

28

3 Regioselective Aldol Reactions

31

Regioselective Aldol Reactions Aldol reactions with specific enol equivalents Lithium enolates can be used directly in aldol reactions, even with enolisable aldehydes, a simple example6 being the synthesis of the enone 32. The ketone 15 forms mostly the less substituted lithium enolate which condenses 29 with butanal to give aldol 31 in reasonable yield. Elimination is usually carried out in acid solution.

O

O

Li O

O

OLi

LDA

29

15 O

30 O

OH

work-up

TsOH

enone 32

aldol 31

Contrast with equilibrium methods Traditional equilibration methods [mix 15, PrCHO and NaOH] would give the enal 33 from self-condensation of butanal. The aldehyde would ignore the less enolisable and less electrophilic ketone.

CHO

CHO +

NaOH

O

two molecules of butanal

enal 33

Aldols with Lewis acid catalysis: silyl enol ethers Silyl enol ethers also combine with aldehydes and ketones in efficient aldol reactions catalysed by Lewis acids such as SnCl4, ZnCl2, AlCl3, and TiCl4, the last being the most popular.13 Thus each of the silyl enol ethers 25 and 22 derived from the unsymmetrical ketone 23 gives a different aldol product 34 and 35 with benzaldehyde.11,14 OH

O

OSiMe3 Ph

34

O

OH

OSiMe3

PhCHO

PhCHO

TiCl4

TiCl4

25

23

22

O

Ph

35

The mechanism is a slightly more complicated example of the six-membered cyclic transition state we met in the last chapter.13 The titanium atom bonds to both oxygen atoms (of the enolate and the aldehyde) 36. As the new carbon-carbon bond is formed, one chlorine atom is transferred

32

3 Regioselectivity: Controlled Aldol Reactions

from titanium to silicon so that the immediate product is the titanium alkoxide 37. This quickly combines with Me3SiCl to give the silylated aldol as the true product 38, and hence the aldol 39 on work-up in a hydroxylic solvent.

CHO

R2CHO

R1

Cl

Cl

Me3SiO

Ti

O

Me3SiCl

Cl O

Cl

Cl

SiMe3

O

Cl

Ti

O

Me3SiCl

TiCl4

R2

R1

R1

R1

R2 36 O

37

SiMe3

OH H2O

CHO

R2

CHO

R2 R1

R1 38

39

Application to the synthesis of gingerol Mukaiyama’s synthesis14 of gingerol 40, the pungent flavouring principle of ginger, illustrates these methods.15 The obvious aldol disconnection shows that we require a specific enol of an unsymmetrical ketone 41 on the less substituted side to combine with an enolisable aldehyde 42. The ketone can be made by the FGA16 (Functional Group Addition) strategy using another aldol reaction, this time between acetone and the non-enolisable and readily available aldehyde vanillin 44, the flavouring principle of vanilla.

O

O

OH

O

MeO

MeO aldol

Bu

+ H HO

HO 40; gingerol

41 O

FGA

MeO

42 CHO

enone

MeO

O +

41 HO HO

43

44; vanillin

The first aldol can actually be carried out by traditional methods: in acid solution condensation occurs on both sides of acetone to give 45, but in alkaline solution 43 is formed in 88% yield. Mukaiyama15 preferred to use the silyl enol ether of acetone 46 with BF3 as the Lewis acid, and this worked as well (89% yield).

3 Reaction at Oxygen or Carbon? Silylation, Acylation and Alkylation

33

O MeO

OMe

O

44

44

acid

HO

43; 88% yield

base

OH

45

O O

OSiMe3

Me3SiCl

44

MeO

Et2O-BF3 CH2Cl2

Et3N

HO 46

43; 89% yield

Catalytic reduction with Raney nickel removed the double bond, and the regioselective aldol - now the silyl enol ether is essential - was carried out by isolating the silyl enol ether 47 and using TiCl4 as the Lewis acid. The yield of gingerol 40 was an impressive 92%. OSiMe3 H2

MeO

1. LDA

43

41 Raney Ni MeOH

42 Et2O-BF3 CH2Cl2

2. Me 3SiCl

40; gingerol 92% yield

HO 47; 69% yield

Reaction at Oxygen or Carbon? Silylation, Acylation and Alkylation Now that you are acquainted with the three main types of regiocontrolled specific enol equivalents (lithium enolates, silyl enol ethers, and enamines) we should consider how well they perform in the other tasks commonly required of enols. Here we shall face another problem of regioselectivity: enols, enolates, and their equivalents have two nucleophilic sites, the required carbon atom and a heteroatom (oxygen in lithium enolates and silyl enol ethers and nitrogen in enamines). We have seen this regioselectivity in action when lithium enolates reacted with aldehydes and ketones at carbon, but with Me3SiCl at oxygen. We normally want reaction at carbon, the site favoured by frontier orbital17 and thermodynamic control rather than at the heteroatom, the site favoured by charge (or Coulombic) and kinetic control. F O Ph

O

Me3SiCl

SiMe3

O TBAF

Ph

Bu4N

48

49

Ph

Bu4N

F

50

Naked enolates “Naked enolates’’ without any complexation can be made from silyl enol ethers using fluoride ion, a very selective nucleophile for silicon 49, and a non-metallic cation, usually a tetra-alkylammonium ion. The commonest reagent is “TBAF’’ (TetraButylAmmonium Fluoride Bu4N⫹ F⫺). In this way the “naked enolate’’ 50 was made. It proved to be acylated with acetic anhydride exclusively at oxygen to give the enol acetate 53 and alkylated with MeI at carbon to give the ketone 51 in 84% isolated yield.18

34

3 Regioselectivity: Controlled Aldol Reactions

O

O

O

Ph

Ph I

51; alkylated ketone 84% yield

O

O

Ph

50 CH3

O

O

O

Ph

CH3

53; enol ester 86% yield

52b; acylation (right) of a naked enolate

52a; alkylation (left) of a naked enolate

Alkylation at carbon, problems with enamines Of our three specific enolate equivalents, the lithium enolates are most similar to naked enolates: most reactive and most basic. The enamines are less reactive, only weakly basic but still nucleophilic in their own right, while the silyl enol ethers are the least reactive, not basic at all, and are rather like alkenes with slightly increased nucleophilicity because of an electron-donating substituent. We have already seen that they usually need a Lewis acid for efficient reaction, just like an alkene in, say, the Friedel-Crafts reaction. Simple alkylation works well with lithium enolates 55, but is a tight SN2 reaction 56 and works best with methyl, primary alkyl, allyl, and benzylic halides.9 Halides with acidic hydrogen atoms, such as α-halocarbonyl compounds may destroy the basic enolate by acting as an acid instead of an electrophile. C-Alkylation with Lithium Enolates O

OLi

Li

Br

O

LDA

R1

R1 54

R2

R1

R1

R2CH2Br

55

SN2 reaction works well with R2 = H, Alkyl, R2 vinyl, Ph

O

SN2

56

57

Enamines are inclined to react at nitrogen with alkyl halides, but react particularly well with α-halo carbonyl compounds 58 to give 1,4-dicarbonyl products 60, and with allylic halides by reaction at nitrogen followed by a [3,3] sigmatropic rearrangement 61 to give γ,δ-unsaturated ketones 63. C-Alkylation with Enamines O

R2N

R2NH

R1

Br

R2N R2

1

R

O R2

R1

H2O

O 54

R2

R1

O

58

Br

H

O

59

60 O

R2N

[3,3]

R1

R2N

H H2O

R1 61

62

R1 63

Silyl enol ethers 64 need Lewis acid catalysis which generates at least a partial positive charge on the alkyl group so they react best with tertiary, allylic, and benzylic halides, and reasonably

3 Reaction at Oxygen or Carbon? Silylation, Acylation and Alkylation

35

well with secondary halides. In this way all types of alkyl halide can be added regioselectively to carbonyl compounds. We have already seen examples of alkylation of the two isomeric lithium enolates of the ketone 23, giving products 27 or 28 at will with benzyl, methyl, primary alkyl, or allyl halides. C-Alkylation with Silyl Enol Ethers O

OSiMe3

R1

O

R1 54

2

64

SN1

R

R1

TiCl4

SN1 reaction works well with R2 = t-alkyl, allyl, benzyl, s-alkyl

O

R2Br

R1 65

66

Application to the synthesis of lipoic acid An example of the use of enamines comes in a synthesis19 of lipoic acid 67, the primary metabolite of redox pathways. This is an early synthesis (1957) but it has a simple and direct strategy and revises aspects of chemoselectivity as well as regioselectivity. The disulfide linkage comes from the oxidation of the two thiols in 68, and these will be made by some sort of displacement on the diol 69. There are 1,3 and 1,6 relationships in this diol acid: the 1,6 offers an opportunity to use the reconnection strategy20 via the lactone 70. Seven-membered lactones are usually made by the Baeyer-Villiger reaction21 on cyclohexanones, here 71, so we come finally to the need to alkylate cyclohexanone itself with some reagent for an a2 synthon.

Lipoic Acid: Analysis S

S

FGI

SH

SH

C–S

CO2H

CO2H

67

OH

68 O

OH

reconnect

BaeyerVilliger

OH

O

CO2H 70

69 OH

O 1,4-diCO

O

OH +

71

a2 reagent needed

d2 reagent

An epoxide would give the right oxidation level, but the efficient alkylation of enamines with α-halo carbonyl compounds evidently appealed and methyl bromoacetate was combined with the pyrrolidine enamine 72 in good yield. Chemoselective reduction of the ester inevitably required protection of the more reactive ketone The leaving group selected for eventual displacement by a sulfur nucleophile was an ester because the Baeyer-Villiger route will provide just such a leaving group at the other position.

36

3 Regioselectivity: Controlled Aldol Reactions

Lipoic Acid: Synthesis MeO2C

N H

O

Br

N

+ O

HO

OH TsOH

MeO2C 73; 71% yield

72

O

O O O

MeCO3H

1. LiAlH 4 (96%)

MeO2C

2. Ac 2O (91%) 3. TsOH (90%)

OAc

O

AcO

74; 88% yield

76; 65% yield

75

S 1. H2N

NH2

76 2. KOH

68 80% yield

FeCl3

S

S CO2H TM67; 80% yield

The Baeyer-Villiger rearrangement illustrates an important aspect of regioselectivity: which group migrates? This is not the same as the chemoselective aspect of the pinacol rearrangement discussed in the last chapter as there is only one leaving group here. The usual rule is followed: the more highly substituted group migrates. Displacement with a sulfur nucleophile is a classic problem of chemoselectivity, the problem being to prevent two alkylations at the same sulfur atom. Thiourea is used here,22 and the final oxidation is carried out with Fe(III).

Alkylation with tertiary alkyl groups Alkylation with tertiary halides is the special preserve of silyl enol ethers. Both the familiar isomers 22 and 25 give regiospecific alkylation in good yield with Lewis acid catalysis.23 The formation of 78 is remarkable as it puts two quaternary centres next to one another. OSiMe3

O

O

OSiMe3

t-BuCl

t-BuCl

TiCl4

TiCl4

77; 67% yield

25

78; 86% yield

22

Silyl enol ethers also react well with allylic halides, or with the hemiterpene fragment prenyl acetate24 79 or with secondary25 80 and tertiary26 81 benzylic halides (Zn salts are used as catalysts in these examples). In each case, regioselectivity has been demonstrated with the same two isomers 22 and 25.

Br

Br OAc O2N 79

80

81

3 Acylation at Carbon

37

Acylation at Carbon Acylation at carbon is the usual route to 1,3-dicarbonyl compounds27 and here the regioselectivity problem is more serious as enolates tend to react with acylating agents on oxygen to give enol esters. Fortunately all three specific enolate equivalents perform well with acid chlorides. The lithium enolate 83 is on the more substituted side of the ketone 82 and so must be prepared via the silyl enol ether. Acylation goes in reasonable yield28 (55%) for the acylation of a basic enolate The problem is that the acylation products are carbon acids which often “quench unreacted enolate at a rate that exceeds acylation.’’ The proton marked H in 84 is between two carbonyl groups and so is easily lost to form a stable enolate. Alternatively, oxidise the corresponding aldol.29 O

O

OLi

1. Me 3SiCl, Et 3N

O 83

H

O

Cl

2. MeLi, THF

82

83

84

Lithium enolates of carboxylic acids Even the dilithium derivatives 86 of carboxylic acids 85 are acylated at carbon by acid chlorides.30 When the product 87 is worked up, it loses CO2 in the usual way, making this a synthesis of ketones 88. Branched carboxylic acids work well in this sequence so it is the equivalent of the acylation of a secondary carbanion, i.e. the disconnection 88a is now acceptable. 2x LDA

OLi

R

R –CO2

RCOCl

CO2H

O

O OLi

85

H

CO2H 87

86

88

R C–C

O

R

+ acylation

Cl

H

H 88a

O

Enamines and silyl enol ethers The best enamines for acylation seem to be those of morpholine. The cyclohexanone 17 forms the enamine on the less substituted side and this is acylated with a simple enolisable acid chloride in reasonable yield.31 O O O

N

O

O

O

Cl

+

17

1.

N H morpholine

2. H , H 2O

enamine

1,3-diketone; 65% yield

38

3 Regioselectivity: Controlled Aldol Reactions

Silyl enol ethers react with acid chlorides and Lewis acids in what is effectively a Friedel-Crafts reaction.32 Both silyl enol ethers 22 and 25 derived from the ketone 23 give regiospecific acylation with MeCOCl and TiCl4. The yields of the two 1,3-diketones 89 and 90, formed as a mixture of diastereoisomers, are excellent. O

OSiMe3

O MeCOCl

Me3SiCl

1. LDA

Et3N

2. Me 3SiCl

TiCl4

89; 91% yield

OSiMe3

O

22

25

23 O

OSiMe3

O

MeCOCl TiCl4

90; 89% yield

25

Acylation at a methyl group in an unsymmetrical ketone, e.g. 15 usually occurs regioselectively even with traditional methods (ester as acylating agent, corresponding alkoxide as base). We shall return to this subject in chapter 10 with other specific enol equivalents, but you can see already that virtually any 1,3-diketone can be made by one of these methods.

Reactions with Other Electrophiles α-Halo carbonyl compounds and epoxides 1,4-Dicarbonyl compounds can be made from enolates by alkylation of enamines with α-halocarbonyl compounds33 or any specific enol equivalent with allylic halides as just discussed, followed in the latter case by oxidative cleavage of the double bond.34 The reactions of enolates with epoxides provides 1,4-difunctionalised compounds35 at a different oxidation level and so we should discuss another interesting question of regioselectivity. Unsymmetrical epoxides 91 are usually opened at the less substituted end 92 by strongly nucleophilic reagents to give 94 via a tight SN2 mechanism 93. O

O (–)

O

X

R

R

X

R

X 91

OH

R X (–)

92

93

94

There has been remarkably little work in this area, but the dilithium derivatives of acids, e.g. 96 do react with unsymmetrical epoxides at the less hindered end. A simple example is styrene oxide 97 which is attacked at its less substituted end to give the lactone 99 in 84% yield after cyclisation.36 O

O

OLi

2 x LDA

CO2H

97

OLi 95

OH

Ph

96

Ph

CO2H

O Ph

98

99

3 Reactions with Other Electrophiles

39

Spirocyclic lactones of structures such as 100 are interesting because of potential biological activity, and might be made from available steroids such as oestrone 101.

O

O

O H H

H

?

H

H

MeO

H

MeO 101

100

The sulfur ylid37 Me2S⫽CH2 reacts stereoselectively from the bottom face of the ketone 101 to give the epoxide 80 which in turn gives the lactone36 100a with the lithium dienolate 96.

O O Me2S

OLi

H

CH2

101 H

O H

OLi

H

96

MeO

H

H

MeO 102

100a

Michael reactions 1,5-Dicarbonyl compounds are usually made by Michael (conjugate) addition of enols to α,βunsaturated carbonyl compounds. Here too is another regioselectivity problem, the one raised at the start of the chapter concerning direct (1,2) versus conjugate (1,4 or Michael) addition to such electrophiles. This question will be more fully developed in chapter 9 but it is appropriate to examine here the performance of our three specific enol equivalents in tackling this question of regioselectivity. In general, direct addition is the kinetically controlled reaction favoured by more reactive irreversible nucleophiles and more reactive electrophiles (Michael acceptors) such as α,β-unsaturated aldehydes, while conjugate addition is the thermodynamic process favoured by less reactive or reversible nucleophiles and less reactive Michael acceptors such as esters of α,β-unsaturated acids. From this analysis we should expect the less reactive enamines and silyl enol ethers to be good at Michael additions, and we would be right. The enamine 103 of our familiar unsymmetrical ketone 23 gives a clean Michael addition with methyl acrylate on the less substituted side to give the ketoester syn-104 as the only product providing that the reaction is carried out in MeOH.38

O

O

N N H

CO2Me

CO2Me

MeOH

23

103

syn-104

40

3 Regioselectivity: Controlled Aldol Reactions

Just to make the point that chemistry has surprises even for the best informed, the rather similar Robinson annelation of dihydrocarvone 105 again using the pyrrolidine enamine 106 goes on the less substituted side 107 when one molecule of enone is used, but on the more substituted side 108 if a large excess (fivefold) of the enone is used.38 Both reactions are of course very useful but an affront to our orderly instincts: we shall meet other examples where only experiment can lay down the ground rules of selectivity. O

O 1 equivalent

N

O

107

N H O

O

105

5 equivalents

106

108

Silyl enol ethers need Lewis acid catalysis for efficient Michael reactions, such as the more substituted (and conjugated) isomer 110 forming a 1,5-diketone 111 from cyclohexenone in good yield.39 This product 111 is a mixture of diastereoisomers as have been many of the products in this chapter. We have also seen some reactions giving single diastereoisomers but without explanation. It is high time that we addressed the question of stereoselectivity and this is the subject of the next chapter. O

Me3SiCl

O

OSiMe3 Ph +

Ph

Ph

H

TiCl4

O

O 109

110

111

A Final Example Double Michael reactions with enamines But before we go, a special example40 of double regioselectivity with a doubly nucleophilic enol equivalent attacking a double electrophile. When the pyrrolidine enamine 112 of cyclohexanone attacks the unsaturated ester 113, the bicyclic ketone 114 is formed in good yield.

O O

Br

N N H

+

112

CO2Et 113

EtO2C 114

3 A Final Example

41

The first step is probably the Michael addition 115 of the enamine 112 to the unsaturated ester 113. The resulting iminium salt 116 loses a proton to regenerate an enamine. But it does so on the side away from the first reaction to avoid forcing the side chain into the same plane as the pyrrolidine ring. The new enamine is 117.

O

N

N

H

O

H

OEt

O

N OEt

Br

OEt

Br

Br

116

115

117

Now the newly formed enamine 117 can be alkylated by the allylic bromide in the side chain. To do this reaction, the molecule must put the side chain in an axial position 118. The resulting iminium salt 119 is hydrolysed on work-up to the bridged bicyclic product 114. Br

CO2Et

CO2Et

CO2Et hydrolytic work-up

117 N

N 118

O 119

114

The same treatment on 4-methoxycyclohexanone 121 derived from 4-methoxyphenol 120 by reduction and then oxidation gives compound 123.

OH

O

CO2Et

Br

N N H

CO2Et

OMe

113

O OMe

OMe

OMe 122

121

120

123

This can be cyclised in refluxing 62% HBr to an adamantane derivative, the dione 125. The mechanism of this cyclisation is by no means obvious, but with the hint that the alkene 124 is an intermediate and that the stereochemistry of the ester can change by acid-catalysed enolisation, you might see what is going on. CO2Et

CO2Et HBr

HBr

OMe O

O

O 123

O

124

Br 125

42

3 Regioselectivity: Controlled Aldol Reactions

References General references are given on page 893 1. Y. Terai and T. Tanaka, Bull. Chem. Soc. Jpn., 1956, 29, 822. 2. C. H. Heathcock, J. E. Ellis, J. E. McMurry and A. Coppolino, Tetrahedron Lett., 1971, 4995. 3. D. J. Dale, P. J. Dunn, C. Golightly, M. L. Hughes, P. C. Levett, A. K. Pearce, P. M. Searle, G. Ward and A. S. Wood, Org. Proc. Res. Dev., 2000, 4, 17. 4. G. Stork, A. Brizzolara, H. Landesman, J. Szmuszkovicz and R. Terrell, J. Am. Chem. Soc., 1963, 85, 207. 5. R. Carlson, L. Nilsson, C. Rappe, A. Babadjamian and J. Metzger, Acta Chem. Scand., 1978, B32, 85, 335, 646. 6. G. Stork, G. A. Kraus, and G. A. Garcia, J. Org. Chem., 1974, 39, 3459. 7. G. Stork, Pure Appl. Chem., 1975, 43, 553; J. d’Angelo, Tetrahedron, 1976, 32, 2979; L. M. Jackman and B. C. Lange, Tetrahedron, 1977, 33, 2737. 8. J. K. Rasmussen, Synthesis, 1977, 91; P. Brownbridge, Synthesis, 1983, 1. 9. H. O. House, M. Gall and H. D. Olmstead, J. Org. Chem., 1971, 36, 2361; G. H. Rubottom, R. C. Mott and D. S. Krueger Synth. Commun., 1977, 7, 327. 10. T. Mukaiyama, K. Banno, and K. Narasaka, J. Am. Chem. Soc., 1974, 96, 7503. 11. I. Fleming, Chimia, 1980. 34, 265. 12. M. Gall and H. O. House, Org. Synth., 1972, 52, 39. 13. T. Mukaiyama, Angew. Chem., Int. Ed., 1977, 16, 817. 14. K. Banno and T. Mukaiyama, Bull. Chem. Soc. Jpn., 1976, 49, 1453. 15. P. Denniff and D. A. Whiting, J. Chem. Soc., Chem. Commun., 1976, 712. 16. Disconnection Approach, page 205. 17. Fleming, Orbitals. 18. R. Noyori, I. Nishida and J. Sakata, Tetrahedron Lett., 1980, 21, 2085; R. Noyori, I. Nishida, J. Sakata and M. Nishizawa, J. Am. Chem. Soc., 1980, 102, 1223. 19. A. Segre, R. Viterbo, and G. Parisi, J. Am. Chem. Soc., 1957, 79, 3503. 20. Disconnection Approach, chapters 26 and 27, particularly page 223. 21. Disconnection Approach, chapters 26 and 27, particularly page 227. 22. Disconnection Approach, page 37. 23. T. H. Chan, I. Paterson, and J. Pinsonnault, Tetrahedron Lett., 1977, 4183; M. T. Reetz and W. F. Maier, Angew. Chem. Int. Ed. Engl., 1978, 17, 48; M. T. Reetz, W. F. Maier, H. Heimbach, A. Giannis and G. Anastassious, Chem. Ber., 1980, 113, 3734. 24. M. T. Reetz, S. Hüttenhain, and F. Hübner, Synth. Commun., 1981, 11, 217. 25. I. Paterson, Tetrahedron Lett., 1979, 1519. 26. M. T. Reetz and S. Hüttenhain, Synthesis, 1980, 941. 27. Disconnection Approach, page 144. 28. A. K. Beck, M. S. Hoekstra and D. Seebach, Tetrahedron Lett., 1977, 1187. 29. A. B. Smith and P. A. Levenberg, Synthesis, 1981, 567. 30. A. P. Krapcho, D. S. Kashdan, E. G. E. Jahngen and A. J. Lovey, J. Org. Chem., 1977, 42, 1189. 31. S. Hünig and M. Salzwedel, Chem. Ber., 1966, 99, 823. 32. I. Fleming, J. Iqbal and E.-P. Krebs, Tetrahedron, 1983, 39, 841. 33. Disconnection Approach, page 209. 34. Disconnection Approach, page 217. 35. Disconnection Approach, page 209. 36. P. L. Creger, J. Am. Chem. Soc., 1967, 89, 2500; J. Org. Chem., 1972, 37, 1907. 37. Disconnection Approach, page 253. 38. N. F. Firrell and P. W. Hickmott, J. Chem. Soc., Chem. Commun., 1969, 544; J. W. Huffmann, C. D. Rowe, and F. J. Matthews, J. Org. Chem., 1982, 47, 1438. 39. K. Narasaka, K. Soai, Y. Aikawa and T. Mukaiyama, Bull Chem. Soc. Jpn., 1976, 49, 779. 40. F. D. Ayres, S. I. Khan, O. L. Chapman and S. N. Kaganove, Tetrahedron Lett., 1994, 35, 7151.

4

Stereoselectivity: Stereoselective Aldol Reactions Definition Introduction and definitions Stereoselective and stereospecific Aldol Stereochemistry Introduction and stereochemical control: syn,anti and E,Z Relationship between enolate geometry and aldol stereochemistry The Zimmerman-Traxler transition state Anti-selective aldols of lithium enolates of hindered aryl esters Syn-selective aldols of boron enolates of PhS-esters Stereochemistry of aldols from enols and enolates of ketones Silyl enol ethers and the open transition state Syn selective aldols with zirconium enolates The synthesis of enones E,Z selectivity in enone formation from aldols Recent developments in stereoselective aldol reactions Stereoselectivity outside the Aldol Relationship A Synthesis of Juvabione A Note on Stereochemical Nomenclature Stereoselectivity is at first sight the easiest of the three selectivities to understand and the most difficult to exercise. It simply means control over stereochemistry. More precisely it means the control over new stereochemistry. In many reactions, whether new carbon-carbon bonds are being formed or whether some functional group is merely being altered, stereochemistry appears. It may be that a double bond is formed which can have E or Z geometry, or that a new stereogenic centre is formed, perhaps by reduction of a ketone, and therefore a relationship develops with other stereogenic centres in the molecule. If these aspects are controlled then we have stereoselectivity. One special distinction you should master from now on is that stereospecificity has a different meaning. It refers to the specific transfer of stereochemistry during a reaction because the mechanism of the reaction demands this stereochemical outcome. An SN2 reaction goes with inversion, whether the molecule likes it or not, because an SN2 reaction is stereospecific. In the reduction of a ketone the molecule may select the stereochemistry of the new OH group: this reaction is not stereospecific though it may be stereoselective. From this chapter onwards stereochemistry will be our constant companion as it is a major concern of modern organic chemistry. To start with we shall discuss diastereoselectivity alone. Soon, in chapter 15, we shall discover how to control the geometry of double bonds. Later on, in chapters 20–21 we shall discuss how to exercise stereoselectivity in general and in chapters

Organic synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

44

4 Stereoselectivity: Stereoselective Aldol Reactions

22–29, there will be a section on asymmetric synthesis, that is the synthesis of optically active compounds. In this chapter we shall continue the discussion of enolate reactions by exploring stereoselective aldol reactions.

The Stereochemistry of the Aldol Reaction Introduction and stereochemical control: syn,anti and E,Z Aldol reactions typically produce new stereogenic centres at either end of the new carboncarbon bond.1 The enolisable ketone 1 might condense with benzaldehyde to give a mixture of diastereoisomers of the aldol 2 in which the methyl and hydroxyl groups can be either on the same side (syn-2) or opposite sides (anti-2) of the carbon skeleton.2 If the aldol is dehydrated to the enone 3 there is again a question of stereoselectivity as the new double bond can be E or Z. O

O

OH

O

OH

PhCHO

R

R

base

Ph

and/or

Ph

racemic (±)-anti-2

racemic (±)-syn-2

1

R

O

O

Ph

H

R

Ph

and/or

R

E-3

Z-3

Only in modern times has aldol stereochemistry seemed a subject worth studying, or indeed even accessible to chemists. Formerly it was left to look after itself. Then Dubois carried out some simple experiments on the condensations between cyclic ketones and aldehydes in base.3 Though largely neglected at the time, these results showed that if LiOH (not NaOH) was used as the base, the anti aldol predominated. Indeed, with cyclopentanone and i-PrCHO, >95% anti-5 was formed, and syn-5 could not be detected. Later Heathcock4 showed that the lithium enolate of the open chain ketone 6 condensed with PhCHO to give >98:2 syn:anti aldol 7. O

O

O

O

OH

i-PrCHO

1. LDA

LiOH

2. PhCHO

OH Ph

4

anti-5

syn-7

6

Relationship between enolate geometry and aldol stereochemistry These apparently conflicting results can be rationalised if there is a relationship between the geometry of the enolate and the stereochemistry of the aldol product. Cyclic ketones are forced to give E enolates, thus cyclopentanone must give 8. Ketone 6 might well prefer to form the Z enolate 9 with the large t-Bu group anti to the methyl group. O

OLi

4

OLi

O 1. LDA

i-PrCHO

LiOH

anti-5 8; E-enolate

2. PhCHO

6

9; Z-enolate

syn-7

4 The Stereochemistry of the Aldol Reaction

45

The Zimmermann-Traxler transition state The relationship between enolate geometry and aldol stereochemistry has now been well established for many aldol reactions. The geometry of lithium enolates expresses itself through the so-called Zimmerman-Traxler5 transition state which is nothing more than the six-membered cyclic transition state that we met in the last chapter. When a lithium enolate reacts with an aldehyde, both the enolate and aldehyde oxygen atoms coordinate to the lithium atom 10 so that the transition state 11 is a partly unsaturated six-membered ring. O

Li

O

O

Li

O

O

Li

O

anti-5

10

11 Zimmerman-Traxler transition state

12

If we assume that the cyclic transition state 11 adopts an approximate chair conformation 13, we can explain the results. One carbon atom, marked with a spot, remains trigonal throughout the reaction and its substituent remains in a plane. All the other carbon atoms have axial and equatorial substituents. The enolate substituent is forced to occupy one of these positions because of the configuration of the enolate. The five-membered ring in 13 is equatorial because it is anti to the enolate oxygen and it no doubt enjoys this position. So far we have been dealing with stereospecificity. The substituent on the aldehyde can occupy either axial or equatorial positions, so it generally prefers the equatorial. You will see that the i-propyl group in 13 is indeed equatorial. This is stereoselectivity. The consequences are most easily appreciated by inspecting the marked hydrogen atoms: in 13: they are clearly anti and this is the relationship found in the aldol product anti-5. H is forced to occupy an axial position because it is cis to OLi in the enolate 8

H O

OH

Li H

The i-Pr group comes from the aldehyde and can choose an axial or an equatorial position. Equatorial is preferred

13

When the Z-enolate 9 reacts with benzaldehyde, one aspect of the Zimmermann-Traxler transition state 14 is quite different. The t-butyl group in 14 is forced to be axial because it is syn to the enolate oxygen even though it will not enjoy this position at all. The only alternative is to abandon the chair transition state and that would evidently lead to a worse energy penalty. The phenyl group still has a choice and it will again choose to be equatorial in 14. If we inspect the marked hydrogen atoms in 14 we can see that they are syn and so they will remain syn in the aldol product syn-7. t-Bu is forced to occupy an axial position because it is cis to OLi in the enolate 9

O

OH

Li

Ph H

H 14

The Ph group comes from the aldehyde and can choose an axial or an equatorial position. Equatorial is preferred

46

4 Stereoselectivity: Stereoselective Aldol Reactions

This is the basis behind many stereoselective aldol reactions, but in general it is not easy to choose carbonyl compounds that will reliably give E or Z enolates. We shall restrict our discussion to one or two of the more reliable. There is a problem in defining enolates as E or Z. If we follow the rules strictly, a series of enolates with the same geometry, such as the four below, would change from E to Z as we changed the metal on the enolate or as we changed from ketone to ester to thiolester. When we are discussing enolates, we shall name them E or Z according to the relationship between the substituent R and the OMetal substituent. This is the generally agreed method and takes no notice whether the enolates are strictly E or Z alkenes.

OLi

R

Me E-alkene E-enolate

OLi

R

OSiMe3

R

SMe

OMe E-alkene E-enolate

OMe Z-alkene E-enolate

OLi

R

Z-alkene E-enolate

The first two examples are based on esters, one giving the anti aldol and the other the syn. Lithium enolates of simple esters like EtCO2Me give clean aldol reactions with aldehydes but the products 17 are more or less 50:50 mixtures of syn- and anti- diastereoisomers.1 Evidently the lithium enolate 16 is also formed as a 50:50 E:Z mixture and this is hardly surprising as OMe and OLi must be about the same size.

O

OH

OLi

1. LDA

15

R

2. RCHO

OMe

OMe

O OMe

syn + anti-17

E + Z-16

Anti-selective aldols of lithium enolates of hindered aryl esters Hindered aryl esters6 such as 18 perform very much better. Only the E enolate is formed and very high ratios of anti:syn aldols 19 are produced, especially with branched aldehydes such as i-PrCHO.7

R

O

OH

R

O

1. LDA

O 18; R = H, Me, OMe

2. RCHO

R

O (±)-anti-19

The ester prefers the conformation 20 because of the anomeric effect, the bonding interaction between one of the lone pairs on the ester oxygen atom and the antibonding orbital of the sigma bond of the carbonyl group, and normal conjugation with the other lone pair of electrons on the ester oxygen atom keeps the aryl group in the plane of the carbonyl group. Perhaps this means that the deprotonation by LDA 21 can take place only when the methyl group is anti to the C=O bond.

4 The Stereochemistry of the Aldol Reaction

47

lone pair in p-orbital overlaps with C=O π* R

O

Li

R2N

R

O

Li

LDA

H

O

O

part of C–O σ*

R

O O

21

E-enolate 22

lone pair in sp2-orbital overlaps with C–O σ* 20

Alternatively one could argue that a six-membered chair transition state 23 is again involved with large axial groups (i–Pr and the THF ligand on Li) and the large aryl group enhancing the methyl group’s natural preference for an equatorial position.

R2N

Li

i-Pr

R

O

N

i-Pr

H

H

H Me

O

O O Li

E-22

R O

O

23

21

Syn-selective aldols of boron enolates of PhS-esters Most esters prefer to form E enolates, and one must turn to a different substituent and a different “metal” to get good syn selectivity. Evans and Masamune have developed the boron enolates of phenythio (PhS) esters 24 in this role.8,9 The boron enolate 25 is prepared with a dialkyl boron triflate and Hünig’s base i-Pr2NEt. The alkyl groups on boron can be butyl or cyclopentyl or else 9-BBN (9-borabicyclononane, see chapters 17 and 24) can be used.10 Triflate is trifluoromethanesulfonate, one of the best leaving groups known. Only the “Z” enolate 25 is formed, maybe because of equilibration with this relatively weak base and the lack of an anomeric effect in PhS esters. Aldol reactions now give high yields of syn aldols 26 helped by the short B–O bonds9 in the Zimmerman-Traxler transition state, cf. 23.

O

O S

R2B-OTf

BR2 S

1. RCHO

R

O S

2. oxidative work-up

i-Pr2NEt

24

OH

25 “Z ” boron enolate

26 (±)-syn aldol

48

4 Stereoselectivity: Stereoselective Aldol Reactions

Stereochemistry of aldols from enols and enolates of ketones Other than for the cyclic or t-alkyl ketones we met at the beginning of the chapter, controlling aldol reactions of ketones has been more difficult. Evans’ boron enolates work well in some cases and it is fortunate that the Z enolates are preferentially formed as these react stereoselectively with aldehydes to give syn aldols whereas the E enolates show poor stereoselectivity. Thus the symmetrical ketone 1 gives almost exclusively (>97:3) the Z boron enolate 27 and hence syn aldol 28 with the enolisable aldehyde n-PrCHO.8 Bu2B O

Bu2B-OTf

O

O

2. oxidative work-up

i-Pr2NEt

(±)-syn-28 >97:3 syn:anti

27 Z boron enolate

1

OH

1. n-PrCHO

Some unsymmetrical ketones show regioselectivity in favour of the less hindered side: 29 is an impressive example as both sides of the ketone are primary with branching occurring on one side only at the β carbon atom.8 You are advised to consult the literature before planning a stereoselective synthesis with ketone enolates. Bu2B O

O

O

Bu2B-OTf

1. PhCHO

i-Pr2NEt

2. oxidative work-up

Ph

(±)-syn-31 >97:3 syn:anti

30 Z boron enolate

29

OH

Silyl enol ethers and the open transition state A second kind of stereoselectivity is observed with silyl enol ethers. When made from cyclic ketones they must have E geometry but often react in a stereorandom fashion, as in the TiCl4 catalysed reaction of the silyl enol ether E-33 of cyclopentanone with PhCH2CH2CHO. This is totally chemoselective, but gives a 50:50 mixture of stereoisomers11 of 32. The corresponding lithium enolate is very stereoselective in the anti sense. The situation is quite different if the reactions are catalysed by fluoride ion instead of by a Lewis acid. Now the same silyl enol ether E-33 reacts stereoselectively, but in the syn sense12 to give syn-5. O

O H

OSiMe3

H

OH O

Ph

O

H

O

OH

F

TiCl4

syn and anti-32

E-33

E-34

syn -5

The Z silyl enol ether 35 from the ketone 6 also shows reversed stereoselectivity from the lithium enolate, giving the anti aldol 7 in high yield.12 Further study1 revealed that these reactions, which take place via an “open” (rather than a chelated) transition state 37, are thermodynamically

4 The Stereochemistry of the Aldol Reaction

49

controlled because they are reversible. Both geometrical isomers of many silyl enol ethers favour the same diastereoisomer of an aldol.

O O

OSiMe3

OH

O Bu4N

F

PhCHO

Ph Ph

O

Z-35

6

anti-7

36 O

H Me Ph

O H 37

Syn selective aldols with zirconium enolates Some kinds of metal enolate also give highly stereoselective reactions in the same sense whatever the geometry of the enolate. At first sight the reactions of zirconium enolates seem like lithium enolates. Using the pyrrolidine amide 38 as an example, we get the Z-enolate 39 only and this gives syn aldol products 40 with aldehydes.13

O

ZrCl2Cp2

O 1. LDA

N

N

OH

O

n-PrCHO

N

2. Cp 2ZrCl2

Z-39; >95:5 Z:E

38

syn-40; 94:6 syn:anti

However, the t-butylthiol ester 41 gives mostly the E-enolate 42 and yet still favours the syn aldol 43. Cyclopentanone, which can of course give only the E enolate 44 also favours the syn aldol14 45. In short, all zirconium enolates give syn aldols regardless of geometry. This selectivity is explained in the workbook.

O

ZrCl2Cp2

O St-Bu

OH

St-Bu

2. Cp 2ZrCl2

O

O

syn-43; 93:7 syn:anti

ZrCl2Cp2

O PhCHO

1. LDA

St-Bu

Ph

E-42; 10:90 Z:E

41

O

PhCHO

1. LDA

OH Ph

2. Cp 2ZrCl2

4

E-44

syn-45; 74:26 syn:anti

50

4 Stereoselectivity: Stereoselective Aldol Reactions

There are many more methods now developed for stereochemical control in the aldol reaction, as this is an active area of research, but we hope that we have studied enough examples in this chapter to add to the examples of chemo- and regioselectivity in the last two chapters to convince you that good control is possible over most of the types of aldol reaction that you might want to use in a synthesis. We shall leave further diastereoselectivity to chapters 20–22 and the question of making optically active aldols until chapters 23–30.

The synthesis of enones The aldol reaction is often used to make enones by dehydration of the aldol itself, a reaction which often occurs under equilibrating aldol conditions, but has to be induced in a separate step when lithium enolates or silyl enol ethers are used. In general one has to accept whatever enone geometry results from the dehydration, and this is usually controlled by thermodynamics, particularly if enone formation is reversible. Simple enones such as 46 normally form as the E isomer but the Z isomer is difficult to prepare. When the double bond is exo to a ring, e.g. 47, the E isomer is again favoured, but other trisubstituted double bonds have less certain configurations. O

O O

O PhCHO

PhCHO

Ph

base

Ph

base

E-47

E-46

Yoshikoshi’s synthesis15 of nootkatone (then supposed to be the flavouring principle of grapefruit) uses an optically active enone 52 prepared from β-pinene 48 by ozonolysis to (+)-nopinone 49 and a chemo- and regioselective aldol condensation using the silyl enol ether 50. Though the aldol reaction produces a mixture of diastereoisomers of 51, all dehydrate to the same enone E-52.

Me3SiCl

O3

O

(–)-48

Et3N

SiMe3

TiCl4

(+)-49

O

MeCHO

O

TsOH

Me

H

H OH syn and anti-51

50

O

Me

E-52

Still’s monensin synthesis16 uses the lactone 55 which inevitably has a Z double bond. This can again be made by an aldol reaction, this time using the lithium enolate of EtCO2Et and the optically active aldehyde 53. A mixture of diastereomeric aldols 54 is again produced, but all dehydrate to the Z unsaturated lactone without racemising the original stereogenic centre. Cyclisation may precede dehydration, or the E compound may equilibrate.

O

O

1. LDA

CHO

CO2Et

CO2Et

2. 53= THPO

53

TsOH

THPO CHO

OH

syn and anti-54

O 55

O

4 Stereoselectivity outside the Aldol Relationship

51

Recent Developments in Stereoselective Aldol Reactions Recent developments tend to focus on the asymmetric aspects of stereoselective aldols but the diastereoselectivity is just as important. The cyclic ester 56 must of course form an ‘E’ enolate and when the boron enolate E-57 reacts with aldehydes the anti-aldol products 58 are formed with good stereoselectivity ranging from 4:1 to >20:1. The predominate isomer is that expected from the Zimmerman-Traxler transition state. The two benzylic-O bonds can be cleaved by hydrogenation and the 2,3-dihydroxy acids anti-59 released in good yield.17 R2B

O

O

O

OH

O

OH

R2B-OTf Et3N

O O

Ph

RCHO

O

–78 °C

O

Ph

CH2Cl2

Ph

O

H2, Pd/C

HO

R

O

Ph

Ph

56

R

OH

Ph

‘E’-57

anti-58

anti-59

A particularly interesting case reported by Denmark18 reveals several types of selectivity. The methyl ester 60 of natural (S)-lactic acid forms a Weinreb amide 61 with Me3Al as catalyst. These amides are designed to react only once with Grignard reagents and organolithiums thus solving a classical chemoselectivity problem (chapter 2). Here the starting material 63 for the aldol reaction is prepared by reaction of ethyl Grignard with 61 and protection of the OH group by silylation. CO2Me

O MeHN

N

Me3Al, CH 2Cl2

OH

O

OMe

OH 60; methyl (S)-lactate

OMe

O t-BuMe2SiCl

EtMgBr

cat. DMAP THF

THF

Me

OH

61; 83% yield

62; 91% yield

OTBDMS 63; 92% yield

The silyl enol ether 64 is formed from 63 with the Z configuration because of the enormous OTBDMS protecting group and reacts with benzaldehyde to give essentially one diastereoisomer of the syn-aldol 65. The 1,3-stereochemical control across the carbonyl group is discussed in chapter 21. O

OSiMe3 Me3Si

OTf

O

OH

PhCHO

Ph

15% HMPA

Et3N

OTBDMS

OTBDMS

63; 92% yield

64; 96% yield > 20:1 Z:E

TBDMSO 65; 79% yield 30:1 diastereoisomers

Stereoselectivity outside the Aldol Relationship A Synthesis of Juvabione There are of course many kinds of stereoselectivity not involving the aldol reaction and many will appear in the rest of the book. Just for example, we give a recent synthesis of juvabione,

52

4 Stereoselectivity: Stereoselective Aldol Reactions

a pheromone that controls insect biology. It illustrates two types of pericyclic reaction and an electrophilic addition to an alkene. We are anticipating many later chapters in giving you this taste of things to come.19 Enol ester formation from crotonaldehyde gives the expected E-selectivity 66. Now the silyl enol ether is formed from this ester, also with the expected double bond geometry. The product 67 has three alkenes: each is conjugated with at least one oxygen atom.

O CHO

NaHMDS t-BuMe2SiCl

KO t-Bu, EtCOCl

O

THF, –78 °C

OTBDMS O

THF/DMPU –110 °C

66; 66% yield

67; 84% yield

The next stage occurs on heating the triene 67 with ethyl acrylate. First a Diels-Alder reaction 68 with the orientation and endo-stereoselectivity expected of this famous process gives an adduct with only two alkenes positioned so that it undergoes a [3,3]-sigmatropic (Claisen-Ireland) rearrangement 69 to give the product 70 with only one alkene.

OTBDMS CO2Me

OTBDMS

O

140 °C, 4 hours

CO2Me

DielsAlder

[3,3]

O

CO2Me

CO2Me

68

ClaisenIreland

RO2C H

69

70

Both these reactions are more or less diastereoselective and give more or less one diastereoisomer of 70. To discover the stereochemistry of the major adduct, the silyl ester (R = SiMet-Bu) was hydrolysed to the free acid 71 (note the designer chemoselectivity) and this was treated with buffered iodine in MeCN to give a crystalline iodo-lactone 72. An X-ray of this compound revealed that the structure of 71 was as shown and that the [3,3]-sigmatropic rearrangement must have had a boat transition state.

H

CO2Me H2SiF6

70

I2, NaHCO 3

HO2C

MeCN, 0 °C

H 71; 60% yield 47:29:24 diasts

O

I CO2Me

O H 72; 71% yield major diast crystallises

Now the synthesis could resume with the elaboration of intermediate 71 into juvabione 76. Acylation of diazomethane with the acid chloride of 71 gave the diazoketone and hence by rearrangement the chain extended acid 74. This new acid 74 is still a mixture of three diastereoisomers but isomerisation to the conjugated ester 75 reduces this to two in an acceptable 78:22 ratio.

4 References

53

Acylation of the acid chloride of 75 with a Grignard reagent catalysed by Fe(III) completes a synthesis of juvabione 76 with good stereochemical control.

CO2Me

1. (COCl) 2 CH2Cl2 cat DMF

HO2C

CO2Me

O

Ag

N2

2. CH 2N2 THF, Et 2O

H

H2O THF

H 73; 80% yield

71; mixture of diasts

CO2Me

CO2Me

1. (COCl) 2

DBU

HO2C

2.

CH2Cl2

HO2C

H 74; 80% yield

H

MgBr

3. mol% Fe(acac) 3

75; 87% yield; 78:22 diasts CO2Me

O

H

76; 75% yield; juvabione

References General references are given on page 893 1. T. Mukaiyama, Organic Reactions, 1982, 28, 203, C. H. Heathcock in Morrison, Volume 3, 1984, page 111. 2. S. Masamune, S. K. Ali, D. L. Snitman, and D. S. Garvey, Angew. Chem., Int. Ed., 1980, 19, 557 (see the footnote on page 558). 3. J. E. Dubois and M. Dubois, Tetrahedron Lett., 1967, 4215; J. Chem. Soc., Chem. Commun., 1968, 1567; Bull Soc. Chim. Fr., 1969, 3120 and 3553; P. Fellman and J. E. Dubois, Tetrahedron , 1978, 34, 1349. 4. C. H. Heathcock, J. Org. Chem., 1980, 45, 1066. 5. H. E. Zimmerman and M. D. Traxler, J. Am. Chem. Soc., 1957, 79, 1920. 6. C. H. Heathcock, Tetrahedron, 1981, 37, 4087. 7. General Reference for Li Enolates: C. H. Heathcock in Comp. Org. Synth. volume 2, chapter 1.6, pages 181–238. 8. D. A. Evans, J. V. Nelson, E. Vogel, and T. R. Taber, J. Am. Chem. Soc., 1981, 103, 3099. 9. M. Hirama, D. S. Garvey, L. D. Lu and S. Masamune, Tetrahedron Lett., 1979, 3937. 10. General Reference for Boron Enolates: B. M. Kim, S. F. Williams and S. Masamune in Comp. Org. Synth., volume 2, chapter 1.7, pages 239–275. 11. T. Mukaiyama, K. Banno, and K. Narasaka, J. Am. Chem. Soc., 1974, 96, 7503. 12. W. A. Kleschick, C. T. Buse and C. H. Heathcock, J. Am. Chem. Soc., 1977, 99, 247, 8109; C. H. Heathcock, C. T. Buse, W. A. Kleschick, M. C. Pirrung, J. E. Sohn and J. Lampe, J. Org. Chem., 1980, 45, 1066. 13. D. A. Evans and L. R. McGee, Tetrahedron Lett., 1980, 21, 3975. 14. Y. Yamamoto and K. Maruyama, Tetrahedron Lett., 1980, 21, 4607. 15. T. Yanami, M. Mayashita and A. Yoshikoshi, J. Org. Chem., 1980, 45, 607. 16. D. B. Collum, J. H. McDonald and W. C. Still, J. Am. Chem. Soc., 1980, 102, 2117, 2118, 2120. 17. M. B. Andrus, B. B. V. S. Sekhar, E. L. Meredith and N. K. Dalley, Org. Lett., 2000, 2, 3035.

54

4 Stereoselectivity: Stereoselective Aldol Reactions

18. S. E. Denmark and S. M. Pham, Org. Lett., 2001, 3, 2201. 19. N. Soldermann, J. Velker, O. Vallat, H. Stoeckli-Evans and R. Neier, Helv. Chim. Acta, 2000, 83, 2266.

A Note on Stereochemical Nomenclature We shall use the terms syn and anti because they have easily understood meanings and because they must refer to a diagram. Terms like erythro and threo or R*R* and R*S* or worst of all ul and lk are very difficult to understand and we shall not use them. They are of course the terms used by official bodies such as IUPAC and the Royal Society of Chemistry, probably for that very reason. Please note that the wiggly line will be used to mean that both isomers are present. It is sometimes used to mean that the stereochemistry is unknown. If you think you are interested in stereochemical nomenclature, you should consult the papers below and then read C. H. Heathcock in Asymmetric Synthesis, ed. J. D. Morrison, Academic Press, Orlando, Volume 3, 1984, page 112–115 to cheer you up again. D. Seebach and V. Prelog, Angew. Chem., Int. Ed. Engl., 1982, 21, 654. F. A. Carey and M. E. Kuehne, J. Org. Chem., 1982, 47, 3811.

5 Alternative Strategies for Enone Synthesis Introduction The Seebach nomenclature for synthons: d2, a3 etc. The Synthesis of Enones by Many Strategies Introduction to the various strategies Strategy 4a: The Aldol Route to Enones When the aldol strategy is ideal When the aldol strategy is not ideal Symmetry as a guide Wittig-style aldol methods Strategy 4b: Acylation of a Vinyl Anion Vinyl metal reagents The aliphatic Friedel-Crafts reaction Unsaturated Acyl Cations and Anions Acyl anion equivalents (d1 reagents) An example with two enones made by different strategies Strategic bonds in enone synthesis Rearrangements of allylic alcohols with Cr(VI) Looking forward to the chemistry of S and Se We have already seen how the aldol reaction may be used to prepare enones and other unsaturated carbonyl compounds. The past three chapters have introduced a lot of new methods without much analysis of strategy. This chapter and the next will redress the balance by looking at the various approaches to the synthesis of α, β-unsaturated carbonyl compounds (which we shall often loosely call enones, a term strictly reserved for α, β-unsaturated ketones) in this chapter, and then choosing which are most suitable for a particular class of enones, the cyclopentenones, in the next. One aim of this chapter in particular is to get you to think in terms of strategy. We shall be using words and phrases descriptive of strategy including a series of terms introduced by Seebach1 to describe the electronic nature of synthons.

O 4

3

2

1 X

1

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

56

5 Alternative Strategies for Enone Synthesis

The Seebach nomenclature for synthons: d2, a3 etc. Any functionalised carbon atom in the target molecule, but most commonly a carbonyl group, is numbered 1 and the atoms then numbered in turn along the chain so that the position of each relative to the C⫽O group is clear 1. A disconnection, say 1a will then require a synthon of a particular type and polarity. Thus 1a requires an alkyl halide and the synthon 3 with a nucleophilic (or donor) carbon in position 2, so we call this a d2 synthon. Similarly the disconnection 1b could be realised by adding an organometallic derivative (MeMetal) to the synthon 2 with an electrophilic (or acceptor) carbon at position 3. This becomes an a3 synthon. You will of course see that we could use an enolate for the d2 synthon, and an enone for the a3 synthon. All the enolates, and the other specific enol equivalents we have been discussing so far, are reagents for d2 synthons. Not everyone is happy about this labelling, but it is sometimes useful and we shall use it from time to time. O

O

b

1b

MeMetal +

1a

O

2;

a3

X

X 2

1a,b

synthon

O

EtBr +

X

X reagent = enone

O

a

3; d synthon

X reagent = enolate

You will perhaps realise from these particular examples that the d2 synthon will be represented in real life by an enolate or its equivalent and the a3 synthon by an α, β-unsaturated carbonyl compound, both displaying the natural polarity of these fragments. It is generally true that even numbered donor synthons (d2 , d4, etc.) and odd numbered acceptor synthons (a1, a3, etc.) have natural polarity while the odd numbered donor synthons (d1, d3, etc.) and the even numbered acceptor synthons (a2 , a4, etc.) have unnatural polarity or umpolung. This makes the numbering of such synthons a useful quick check on the type of reagent likely to be needed. O

4 R3

R1 R

2

General structure for an enone. The strategies for disconnection of 4 will be outlined first.

The Synthesis of Enones by Many Strategies Introduction to the various strategies Now we start a systematic analysis of most of the possible disconnections which might be used in planning the synthesis of enones using the generalised structure 4 as our example. The first disconnection both in importance and in order of discussion is the aldol, that is disconnection of the enone double bond 4a. You know from previous chapters that we can find a reagent for the enolate 6 when either R2 or R3 is hydrogen, providing there is branching on one side, or if R2 ⫽ R3. The Aldol Disconnection: d 2 + a1 Synthons O H R3

R1 a

a

R1

O

aldol

R2 4a

O + R2

5

O R3

R3 R2

6

6

5 The Synthesis of Enones by Many Strategies

57

Next we shall tackle the important disconnection 4b between the carbonyl group and the double bond. In many ways this is theoretically the best strategy as it is right in the centre of the molecule. This means we shall be examining the reactions between vinyl metals 7 and acylating agents 8 (where X is a leaving group such as Cl or OR). The nature of the metal and the stereochemistry of the vinyl group may be important here. Acylation of a Vinyl Anion: another a 1 Synthon O b

R3

R1

O

R1

+

R2 4b

R3

X

R2 7

8

An important family of disconnections follows. The alkylations represented by 4c(i), 4c(ii) and 4c(iii) belong to extended enolate chemistry and you will learn to call the synthons 9, 10 and 11 the α', γ and α extended enolates. Synthons 9 and 11 react in the normal d2 position but 10 has d4 reactivity. There are obvious questions of regioselectivity in this family. This chemistry will not be discussed further in this chapter as it is the subject of chapter 11. The Extended Enolate Family of Disconnections: d 2 and d4 Synthons O

O c(i)

R3

R1 i

R2 4c(i)

9 O

R1

c(ii)

R3

ii 10;

O R3

R1 R2

R2 d4

X

R2

O X

R3

R1

c(iii)

iii

R2

4c(ii) and 4c(iii)

synthon

R3

R1

11;

d2

X

synthon

The syntheses represented by disconnection 4d belong to a group where the alkene part of the enone is almost irrelevant: they involve either acylation of an organometallic compound 4d(i) or alkylation of an acyl anion equivalent (a d1 reagent) 4d(ii). Disconnection Outside the Enone: a 1 and d1 Synthons O R1

R3 +

R2 12: a 1 synthon

Metal

O d(i)

O R3

R1 R2 4d

d(ii)

R1

R3 +

R2 13: d 1 synthon

X X = leaving group

Another disconnection outside the enone system is 4e which looks at first sight like a conventional conjugate addition to an a3 Michael acceptor. However we need to get the alkene back again after the Michael addition has occurred. The a3 synthon 14 is unsaturated.

58

5 Alternative Strategies for Enone Synthesis

Michael Addition without loss of the Alkene O R3

R1

e

R1

R2 4e

O R3

+ Metal

R2 14; unsaturated a 3 synthon

We shall also explore combinations of two disconnections, that is three-component syntheses, and strategies based on the addition of functional groups (here the double bond and the carbonyl group). This strategic tour is valuable in its own right as it should give you some useful methods, but its value also lies in your grasping the possibility of many different approaches to even such a simple compound as 4. Throughout this chapter and even more so in the next chapter we shall consider the reasons for choosing one strategy or another from this perhaps rather bewildering array when faced with the synthesis of a particular compound.

Strategy 4a: The Aldol Route to Enones We have already seen several examples of this reaction in the synthesis of enones.2 It is the first disconnection you should try in planning the synthesis of any enone and you should assess the likelihood of success by the aldol reaction before going on to any other disconnection, if only to get inside the skin of the problem. Almost any enone can be made nowadays by the aldol reaction, but it is worth considering in turn the features which make it specially suitable, and those which suggest that an alternative strategy should be attempted.

When the aldol strategy is ideal The most obvious pointer to its success is if the double bond is roughly in the middle of the molecule, or if it separates two rings. This means that the disconnection will quickly lead to much simpler starting materials. We obviously want to make the enone 15 from two different aromatic starting materials and the aldol disconnection gives us an unenolisable aldehyde 16 and a very enolisable methyl ketone 17 with no problems of selectivity. Simply mixing 16 and 17 with KOH gives3 an 87% yield of the enone 15. OH

O

OH

OH Cl

O

OH

CHO

aldol

Cl +

enone

Cl 15

Cl 16

17

The double bond in enone 18 sprouts from a ring and aldol disconnection does indeed simplify the problem considerably since the starting materials are an unenolisable aldehyde 20 and a cyclic enolisable ketone 19. There is only one place for enolisation in either molecule. The condensation between 19 and 20 with HCl in MeCO2H gives the enone 18 in 75% yield.4 The stereochemistry of the enone double bond is under thermodynamic control (chapter 4).

5 Strategy 4a: The Aldol Route to Enones

59

O

O

OHC aldol

O

O

enone

O

+ O

O

O

O

O O

O

18

19

20

When the aldol strategy is not ideal Conversely if the target is a vinyl ketone, then the aldol disconnection removes only one carbon atom - not much of a simplification! Even so, the aldol and its modifications such as the Mannich reaction are often used for such syntheses. An extreme example comes in Corey’s synthesis of antheridic acid5 where an aldol disconnection of an advanced intermediate 21 removes just one carbon atom but makes the remainder 22 of this complex molecule much easier to synthesise. O

O Mannich

H

O

H

RO

O

RO O

CO2Me

O

21

CO2Me

22

If the double bond is embedded in the middle of the molecule there may be no simplification at all, indeed the proposed starting material might be much harder to make than the target itself! The simple bicyclic ketone 23 disconnects to the dione 24 which is an unsymmetrical octa-1,4-dione difficult to make and which might cyclise to give 25 if C-5 enolises and attacks C-1 instead of C-8 enolising and attacking C-4 as we require. We shall see in the next chapter how to make such enones. O

O

O 7 aldol

6

4 3

enone

5 23

8

O

8

1 2 =

1

7 6

O

4

2 3

5 O 24a

24

25

Symmetry as a guide Symmetry may guide us. The simple enone 26 disconnects to two molecules of the same symmetrical ketone 27 and is therefore easily made without selectivity problems.2 Acid or basecatalysed self-condensation of cyclopentanone gives the enone 26 in good yield. O

O aldol

+ O enone

26

27

27

60

5 Alternative Strategies for Enone Synthesis

The bicyclic compound 28 looks a poor subject for the aldol until we realise that the starting material 29 is symmetrical and as its two carbonyl groups have a 1,6-relationship, it can be made by the reconnection strategy.6 The cyclisation to give 28 goes in 96% yield.2 There is an obvious contrast with the failed synthesis of 23. There the two ketones in 24 were different so there were three possible sites for enolisation. In 29 the two ketones are the same and all four sites for enolisation are the same. There are no selectivity problems. O

O

O 8

aldol 9

enone

7

6

1 2 10

28

5 4 3

O

=

9

8

10

1

2

6

5

3

4

reconnect 1,6-diCO

O 29

6

7

29a

FGI reduction

1

30

31

The complicated heterocyclic compound 32 is symmetrical about a vertical plane and two aldol disconnections give the symmetrical ketone 34 and two molecules of the unenolisable aldehyde 33. Condensation with KOH gives7 93% of 32.

O

O

O N

2 x aldol

CHO

O N

CHO

34

symmetry NMe2

NMe2 32

NMe2

NMe2

33

33

Mechanism may help us. If the required enolate is particularly easy to make, or the electrophile particularly reactive and/or unable to enolise, these are good points. If there are problems of chemo- or regioselectivity we cannot solve, then these are bad points. Stereochemistry may be a problem too. There is not much you can do about the stereochemistry of an enone made by the aldol reaction as you tend to get the thermodynamically more stable isomer (E or Z) and if you want the other you should try a different strategy.

Wittig-style aldol methods One version of the aldol is particularly suitable for enones as it uses Wittig-style chemistry and makes alkenes in one step. As you will see in chapter 15, reactions of stabilised ylids 36 or

5 Strategy 4a: The Aldol Route to Enones

61

phosphonate anions 38 (that is those of the aldol type) with aldehydes give E-alkenes selectively. Phosphonate anions also react with ketones.

O

A Wittig-style 'aldol' reaction Ph P 3

O

R2CHO

Ph3P

R1

R1 36; stabilised ylid

35 A HornerO WadsworthEmmons (RO)2P reaction

O

base

O

O base

R1

O

O +

(RO)2P

37

R1

R2

R1

R2

R1

R2

O

38

39

In general, the aldol strategy should be your first choice. If you feel that the chemo- and regioselectivity of the aldol reaction can be controlled and that the stereochemistry is likely to be correct, you should try the aldol approach. Of course, as in all aspects of synthetic chemistry, a literature search for precedents might save a lot of time. Even the aldol fails sometimes. In a recent study of asymmetric conjugate addition, Simon Woodward and co-workers8 required a series of enones 43. Some they made with lithium enolates 41 of ketones or esters 40 added to aldehydes to give the aldols 42 that were dehydrated to the enones 43 with concentrated HCl. O

OLi

R1

O

OH

O

2. R 2CHO

1. LDA

HCl

R1

THF

40; R 1 = alkyl, OMe

R2

R1

43; enone

42; aldol

41

R2

R1

Others they made by Wittig or Horner-Wadsworth-Emmons reactions such as the methoxyketone 45 from the phosphonate ester 44. The 78% yield is of pure E-enone 45. O

O

O

K2CO3

OMe

(MeO)2P

OMe

n-C5H11CHO

45; 78% yield

44

The search for anti-HIV drugs led workers in New Zealand9 to potential peptide mimics with the central amide bond replaced by a trans alkene 47. These compounds would have roughly the same shape as natural peptides and might inhibit HIV protease. To get the right diastereoisomer two single enantiomers were coupled by the Horner-Wadsworth-Emmons reaction of phosphonate 46 and the right aldehyde. The slightly odd conditions avoided racemisation of the aldehyde. OR

Ph

O

Ph

O

Ph OHC

Ph

O

N H

P(OMe)2 O 46

O

i-Pr2NEt, LiCl MeCN

Ph

O

N H

OR Ph

O 47

62

5 Alternative Strategies for Enone Synthesis

Strategy 4b: Acylation of a Vinyl Anion Strategy 4b involves the disconnection 48 of the bond between the double bond and the carbonyl group of the enone and requires the acylation of a vinyl anion 49 or its synthetic equivalent with some a1 reagent 50, an acylating agent, usually an acid derivative. O

O strategy 4b

R2

R1

+

R1 49; vinyl anion

48

R2

X

50: acylating agent

Vinyl metal reagents Its most elementary realisation is the reaction of a vinyl Grignard reagent 52 or vinyl lithium, usually prepared from a vinyl chloride 51 in THF, with an aldehyde followed by oxidation to the enone 48. The Swern oxidation is particularly useful for the oxidation of allylic alcohols such as 53 to enones. Cl

R1

1. Mg, THF

MgCl

R1

2. R 2CHO

52

51 OH

O [O]

R2

R1

R2

R1

53

48

A more direct version is the treatment of vinyl-lithiums with carboxylic acids.10 Vinyl-lithium itself reacts with EtCO2H in DME (DiMethoxyEthane) to give the enone 56 in high yield.11 The first molecule of vinyl-lithium forms the lithium salt 54 and the second molecule adds to give the di-lithium derivative 55, an intermediate which might remind you of the lithium enolates of carboxylic acids discussed in the previous three chapters. Hydrolysis during normal work-up gives the enone 56. Li

EtCO2H

Li

O

MeO

LiO

OLi

OMe

O

H H2O

OLi 54

56; 92% yield

55

This reaction often gives much poorer yields in other cases, though the cyclic vinyl-lithium 58 prepared from the vinyl chloride 57, gives a moderate yield12 of 59. O Cl

Li Li

PhCO2H

Ph

THF

57

58

59; 42% yield

5 Strategy 4b: Acylation of a Vinyl Anion

63

The most popular method these days for the acylation of Grignard or organolithium reagents is the Weinreb amide (also discussed in chapter 8). Acylation of vinyl Grignard with the complex intermediate 60 was part of a synthesis of (⫺)-fumagillol13 62. O MeO

O

N Me

MgBr

OSiMe3

OSiMe3 OPMB

OPMB

60

61 O O OMe OH 62; (–)-fumagillol

The intramolecular acylation of the vinyl lithium derivative of 63, prepared by exchange of iodine with lithium (chapter 8) with the Weinreb amide on the side chain gives the five-membered heterocyclic ring in 64, an intermediate on the way to (⫺)-brunsvigine 65, an alkaloid of the montanine group.14 MeO

O

I

O

OH

O

O

BuLi

N

O

O

Me H

Ts

OH

O

N H

63

O

N H 65; (–)-brunsvigine

64

The aliphatic Friedel-Crafts reaction In chapter 16 we shall discover that vinyl-copper reagents can be prepared with stereochemical control over double bond geometry and acylated directly with acid chlorides. We shall also meet vinyl silanes and see how they too can be acylated with acid chlorides. In this chapter we shall consider only the acylation of alkenes themselves with acid chlorides, that is the aliphatic FriedelCrafts reaction.15 The normal Friedel-Crafts reaction 66 combines an aromatic compound with an acid chloride and a Lewis acid to give a cation 67 which loses a proton to give an aryl ketone 68. O

O O

Cl R

AlCl3

O

AlCl3

R R 66

R

H 67

68

An alkene, if it is reactive enough, can give the same sort of intermediate 70 but there is no longer the same urgency to lose a proton, as there is no aromaticity to regenerate, and a chloride ion usually adds to give the β-chloroketone 71. Base is needed to give the enone 72. By this method the enone 72 was made in 40% yield but only after extensive fractionation.12

64

5 Alternative Strategies for Enone Synthesis

O

O

O Cl

R

R

R 69

70

O

Cl base

R

H 71

72

The simple enone 74 was needed for a synthesis of trans chrysanthemic acid 73, a component of the natural insecticide pyrethrin.16 An aldol approach would require a cross-condensation between two enolisable ketones, one of them 75 unsymmetrical, and although this could easily be done an alternative was sought. O CO2H

O

aldol

O

enone

73

+

74

75

Disconnection between C⫽O and C⫽C with the aliphatic Friedel-Crafts reaction in mind would require acylation of the unsymmetrical alkene 76 with the acid chloride 77. The alkene 76 is ideal for acylation as it is tri-substituted and therefore relatively electron-rich and will react at the required (less substituted) end. The β-chloroketone 78, formed when SnCl4 was used as the Lewis acid, was treated with base without isolation to give the enone 74 in 60% overall yield. Lithium chloride may not look very basic, but in dipolar aprotic solvents like DMF (DiMethylFormamide, Me2NCHO), that do not solvate anions, chloride is a good base. O

O

SnCl4

+ 74a

O

O

FriedelCrafts

Cl

LiCl DMF

Cl

76

78

77

74

Strategy 4c: Unsaturated Acyl Cations and Anions Acyl anion equivalents (d1 reagents) Strategy d(ii) (see 4d) follows logically as similar methods are involved. The same enone 74 can be made by this strategy 74b if an acyl anion equivalent (a d1 reagent) for the synthon 79 can be alkylated at the carbon atom of the carbonyl group. O 4d(ii)

O

X

O ?

+ 74b

X = leaving group

H 79

80

The point of this is that the proposed starting material 80 can then be made by a simple Wittig reaction with the stable ylid 81, the equivalent of an enolate of MeCHO, as the extra ylid stabilisation avoids any regioselectivity problems in enolisation. The cyanohydrin derivative 82 can lose the marked proton with strong base and alkylation occurs as planned. The product 83 can be hydrolysed

5 Strategy 4c: Unsaturated Acyl Cations and Anions

65

with anions chemoselective for silicon (fluoride or chloride) to regenerate the carbonyl group with the formation of the enone17 74. We shall meet further examples of acyl anions (d1 reagents) in chapter 14.

Ph3P

CHO

O

Me2CO

H 81

OSiMe3

Me3SiCN ZnI2

H 82

80 OSiMe3

1. LDA 2. i-PrI

CN

O Et3NHF

CN

or HCl

83

74

An example with two enones made by different strategies The same strategy, but with the opposite polarity, along with other enone approaches is illustrated by the bicyclic enone 84 needed for the synthesis of the terpene cadinene. Aldol disconnection gives the 1,5-diketone 85 which we expect to make by a Michael addition.18 Hence we require a specific enol equivalent of cyclohexanone to add to the enone 86, and we have rediscovered the Robinson annelation.19

H

H 1,5-diCO

aldol

+ O

enone

O

O

O 84

O

85

86

specific enolate needed

Disconnection of the enone 86 by the strategy we have just used for 74, with the reverse polarity, gives the unsaturated acid 87 again accessible by a Wittig reaction with the protected reagent 89. We have in mind using an alkyl-lithium instead of a vinyl-lithium as a nucleophile to make 86.

4d(i)

Wittig

EtLi + O

O 86a

OH 87

CHO +

Ph3P O 88

Ph3P OH

O

OR

89

The phosphonate 90 version of the Wittig reaction, often used when there is an anion stabilising group (here CO2Me) present, gives 65% of pure E-91, easily hydrolysed to the free acid 87. Now we need to add EtLi, made from EtBr and Li metal, which gives E-86 in 92% with some recovered starting material. Finally, the Robinson annelation goes in excellent yield (83%) to give the required bicyclic enone 84. There is also reasonable selectivity in the Robinson annelation (7:3 in favour of the required syn isomer), presumably through a transition state such as 93. This is a Newman projection, often the best way to visualise the simultaneous creation of two

66

5 Alternative Strategies for Enone Synthesis

new stereogenic centres. Note that this is not aldol stereo-selectivity: it is akin to the examples discussed at the end of chapter 4 and to be discussed in chapters 20–22. O (MeO)2P

1. NaH

K2CO3

2. i-PrCHO

MeOH

OMe

O

OMe

O

90

EtLi

OH

O

91

O

87

86

H

N H

H

86

H

N O

N

O

O TM84; 83% yield, 7:3syn:anti

92

93

Strategic bonds in enone synthesis When Fétizon20 wanted to synthesise members of the cedrene and acorene classes of terpene, he chose to try the photochemical cyclisation21 of the dienone 95, hoping to get 94. The best disconnection strategically is of the bond between the two branchpoints 95. This bond is also roughly in the middle of the molecule and separates a ring from a chain and so cries out to be disconnected. We shall call such bonds strategic bonds in future. We might try an organometallic reagent 96 (M ⫽ metal) and some reagent for the synthon 97. As the enone is cyclic we cannot use the “obvious” ynone 98. Analysis: ?

hν

?

+

?

O

O 95

94

96

O

O

97

98

A more practical strategy uses the chemistry of 1,3-diones 99 which exist as an equilibrium mixture with the predominant enol form 100. Reaction with an alcohol (MeOH or i-PrOH are most popular) gives enol ethers such as 101 which react cleanly with organo-metallic reagents (such as 96; M ⫽ Li or MgBr). However, attack occurs at the carbonyl group to give 102; and not at the double bond, so the molecule is turned back to front by this reaction to give the enone 103. O

O

HO

RLi or RMgBr

i-PrOH

R

OH R H2O

H

O 99

O 100

i-PrO

i-PrO

101

102

O 103

5 Strategy 4c: Unsaturated Acyl Cations and Anions

67

When you have realised, as we hope you did from the way we drew the product 101, that enol ether formation occurs randomly at either end of the 1,3-dione system, you should also see that only symmetrical diones are suitable for this reaction. Returning to the original problem, the synthesis of 95, the ketone 104 is readily made,22 so this route is suitable using the lithium derivative made from the chloride 105 and the ethyl vinyl ether 106.

O

Cl

1. LiAlH 4

OH

2. O

1. Li, hexane

Li

H

+

H2O

2. Ph 3P, CCl 4

EtO 104

96; M = Li

105

O

OEt

106

107

95

Rearrangements of allylic alcohols with Cr(VI) A regiospecific version23 of the same strategy starts with an unsymmetrical enone, such as the Robinson annelation product 108. An organometallic compound is again added direct to the carbonyl group, but now the rearrangement is accomplished by oxidation with Cr(VI). The starting material 109 cannot be oxidised as it is a tertiary alcohol, but the product of allylic rearrangement, being a secondary alcohol 110, can, and the enone 111 must result. O

R

R

OH

RMgX

R

H

Cr(VI)

H or RLi

OH 108

109

O

110

111

This is probably a [3,3] sigmatropic rearrangement 112 of the chromate ester of the tertiary alcohol so that 109 is oxidised directly to 111 via the chromate ester 113 without any secondary alcohol 110 actually being formed. An important way to carry out this strategy appears in chapter 18 under the Heck reaction. R

OH

R

O

O Cr

Cr(VI)

OH

R

R O O

O O

109

112

113

Cr

OH

O 111

This strategy was recently applied to the synthesis of scopadulcin24 114. They had already made the enone 115 and wished to use it as starting material so means had to be found to bolt ring A onto the bottom left hand side of the enone. Addition of the lithium derivative 117, made from the chloride 116 with the aid of ultrasound (often a help in heterogeneous reactions) to the enone 115 gave the allylic alcohol 118 and this duly rearranged on Cr(VI) oxidation.

68

5 Alternative Strategies for Enone Synthesis

OH Cl

O

Li

O ?

H

A

Li, Et 2O

O

ultrasound

H

H CO2H OBz 114; scopadulcin

O

115

O

O 116

117

O

O O

117

O PCC

OH

115

O

NaOAc

H O

O

H O

O

O

119; 72% yield

118; 97% yield

Looking forward to the chemistry of S and Se Modern organic chemistry, particularly using the special reactivity of sulfur and selenium compounds, has developed regiospecific reagents for all three synthons 120, 121, and 122, and the three negative synthons of the same structure, needed for this family of strategies. We shall postpone a detailed discussion of this chemistry until chapters 16 and 18. The nucleophilic synthon corresponding to 121 will also appear as extended enolate chemistry, disconnections (7c), in chapter 11. Finally, there are some further strategies we have not even mentioned which depend on the late addition of either the carbonyl group or the double bond. These also involve sulphur and selenium chemistry and will appear in chapter 33 under the heading “FGA strategy”. You will appreciate that there are many approaches to the synthesis of enones. How you are to decide which one to adopt in a particular case is the subject of the next chapter. O

120

O

O

121

References General references are given on page 893 1. D. Seebach, Angew. Chem. Int. Ed., 1979, 18, 238. 2. Review: A. T. Nielsen and W. J. Houlihan, Org. React., 1968, 16, 1. 3. S. S. Tiwari and A. Singh, J. Indian Chem. Soc., 1961, 38, 931. 4. W. H. Perkin, A. Pollard, and R. Robinson, J. Chem. Soc., 1937, 49.

122

5 References

69

5. E. J. Corey and A. G. Myers, J. Am. Chem. Soc., 1985, 107, 5574; E. J. Corey, A. G. Myers, N. Takahashi, H. Yamane and H. Schraudolf, Tetrahedron Lett., 1986, 27, 5083. 6. Disconnection Approach, chapters 26 and 27, pages 217–228; see page 300 for the synthesis of TM 28. 7. P. I. Ittyerah and F. G. Mann, J. Chem. Soc., 1958, 467. 8. C. Börner, M. R. Dennis, E. Sinn and S. Woodward, Eur J. Org. Chem., 2001, 2435. 9. M. K. Edmonds and A. D. Abell, J. Org. Chem., 2001, 66, 3747. 10. M. J. Jorgenson, Org. React., 1970, 80, 1. 11. J. C. Floyd, Tetrahedron Lett., 1974, 2877. 12. H. E. Zimmermann, J. Org. Chem., 1955, 20, 549. 13. J. Boiteau, P. Van de Weghe and J. Eustache, Org. Lett., 2001, 3, 2737. 14. C.-K. Sha, A.-W. Hong and C.-M. Huang, Org. Lett., 2001, 3, 2177. 15. R. E. Christ and R. C. Fuson, J. Am. Chem. Soc., 1937, 59, 893. 16. J. Ficini and J. d’Angelo, Tetrahedron Lett., 1976, 2441. 17. U. Hertenstein, S. Hünig, and M. Öller, Synthesis, 1976, 416. 18. The Disconnection Approach, chapter 21, page 170. 19. E. Piers and W. M. Phillips-Johnson, Can. J. Chem., 1975, 53, 1281. 20. M. Fétizon, S. Lazare, C. Pascard, and T. Prange, J. Chem. Soc., Perkin Trans, 1, 1979, 1407; D. D. K. Manh, J. Ecoto, M. Fétizon, H. Colin, and J.-C. Diez-Masa, J. Chem. Soc., Chem. Commun., 1981, 953. 21. Disconnection Approach, chapter 32, page 268. 22. Disconnection Approach, page 1. 23. W. G. Dauben and D. M. Michno, J. Org. Chem., 1977, 42, 682. 24. S. M. A. Rahman, H. Ohno, T. Murata, H. Yoshino, N. Satoh, K. Murakami, D. Patra, C. Iwata, N. Maezaki and T. Tanaka, J. Org. Chem., 2001, 66, 4831.

6

Choosing a Strategy: The Synthesis of Cyclopentenones Strategies Based on an Aldol Reaction Using the Aliphatic Friedel-Crafts Reaction The Nazarov Reaction Cycloadditions of Fe(CO) 4 Complexes of Oxyallyl Cations The Pauson-Khand Reaction Recent Developments in the Pauson-Khand Reaction Oxidative Rearrangement of Tertiary Allylic Alcohols Other Methods

Strategies Based on an Aldol Reaction Five-membered rings occur throughout Nature and in many exciting molecules made by man. The perfumery constituent cis-jasmone 1 might claim to be the molecule synthesised in most different ways. The prostaglandins, molecules which control human physiology, contain five-membered rings and some, like PGA2 2 are cyclopentenones. Dodecahedrane 3, the highly symmetrical (CH)20 hydrocarbon, consists of nothing but five-membered rings, twelve of them in all, and has been approached from various cyclopentenones. O

O

CO2H

OH 1; cis-jasmone

2; PGA 2

3; C 20H20 dodecahedrane

We shall look at the many different strategies used to make cyclopentenones: some will be similar to the methods explored in the last chapter, some will be totally different. Some will introduce more advanced chemistry that we shall meet later in the book, but the point of this chapter is the strategy: in each case we hope you will ask yourself why that particular route was followed. When Stork1 wished to synthesise cis-jasmone 1, he considered how helpful it would be if there were an equivalent to the famous Robinson annelation2 which would make five-membered instead of six-membered rings. Analysis would start in the same way (cf. chapter 5) 1a but the starting material 4 for cyclisation would be a 1,4-diketone instead of a 1,5-diketone.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

72

6 Choosing a Strategy: The Synthesis of Cyclopentenones

cis-jasmone: Analysis I O

O aldol

O 1a

4

Disconnection 4a to an enone 5 is still possible, but we should have to add a d1 reagent, or acyl anion equivalent 6 instead of an enolate. At the time of Stork’s work, there were no d1 reagents which reliably added Michael fashion, so he invented one. It would obviously be convenient if the d1 reagent would also act as an alkylating agent in the synthesis of 6, that is if it would act as the doubly nucleophilic synthon 7. cis-jasmone: Analysis II O O

1,4-diCO

O ?

O+

C

+

O

X 4a

d1 synthon 6

5

7

8

The amino nitrile 9 fulfils all these hopes. It is easily made, it can be alkylated to give 10, and the anion from removal of its second proton gives a clean Michael reaction to form 11. Finally, the amino-nitrile functionality is easily hydrolysed at room temperature and neutral pH (unlike so many such compounds: see chapter 14) with Cu(II) catalysis to give the ene-dione 4 and hence cis-jasmone 1 by thermodynamically controlled cyclisation in weak base.1 This process of adding a new five-membered ring, not necessarily to give a cyclopentenone, is often called cyclopentannelation and there are now many such methods.3 CN

CN CH2O

HCN Et2NH

1. LDA

1. LDA

Et2N

2. 8; X = I

Et2N

9 Et2N

2. 5

10 O

CN

1% NaOH H2O, Et 2O

CuSO4 H2O EtOH

O

11

O

O

4; 72% yield from 9

1; cis-jasmone, 91% yield

This first approach kept the Michael acceptor half 5 of the Robinson annelation. The alternative would be to keep the enolate half and replace the Michael acceptor by an a 2 synthon. An example is the synthesis of the cyclopentenone 12 by Yoshikoshi.4 As before, aldol disconnection gives a 1,4-diketone 13, but this time there is a strategic bond for disconnection leading logically to a regiospecific enolate 14 and a regiospecific a2 synthon 15. From chapter 3 we know that a silyl enol ether is the best choice for a more substituted enol. It is less obvious to choose an unsaturated nitro compound 16 as the a2 reagent, but you should be aware that aliphatic nitro compounds can be converted into ketones, and that the great anion stabilising ability of the nitro group makes the synthesis of unsaturated nitro compounds a simple matter.5 Putting these two ideas together makes

6 Strategies Based on an Aldol Reaction

73

this a good strategy, particularly as the same anion-stabilisation will make the Michael addition a good reaction too. 3 aldol

O

O

O

1,4-diCO

4

2

+

1 O

O

13

12

NO2

FGI

14; regiospecific 15; a 2 synthon 16; a 2 reagent enolate

Nitropropane 17 can be condensed directly with CH2O to give the alcohol 18 which is dehydrated with phthalic anhydride to give nitroalkene 16. The silyl enol ether 20 can be made directly from the unsymmetrical ketone 19 under equilibrating conditions and reacts with the nitroalkene under Lewis acid (TiCl4 as usual) catalysis to give the adduct 21. Normally one needs quite vigorous conditions to convert nitroalkanes to ketones, and the very mild hydrolysis used in this example suggests that 21 hydrolyses unusually easily. The result is another efficient cyclopentannelation by a different strategy.4 O NO2 CH2O

NO2

HO

NO2

phthalic anhydride

phthalic anhydride

O

NaOH

17

18

O

Me3SiCl

16

Et3N

TiCl4

OSiMe3

19

O

16 NO2

H2O reflux

O 21

20

TM 12 87% yield

The first strategy one normally considers is the one which is already in the literature. In the case of the cyclopentenone 22, widely used in natural product synthesis,6 the near symmetry led to an unusual strategy via the acyloin product 23 and hence to the symmetrical diester 24 as starting material.7 O

O

FGI dehydration

MeO2C

acyloin

MeO2C

HO

22

24

23

This synthesis works well, using the silicon modification 25 of the acyloin reaction, but the final dehydration leads to a mixture of enone 22 with the rearranged enone 26. This problem was not realised until chemists8 began to use this published procedure,9 and is a warning to all chemists involved in synthesis to check the structure of their intermediates carefully by NMR.

Na

O

Me3SiO

H

H

O

O +

24 Me3SiCl

H2O

Me3SiO 25

HO 23

22

26

74

6 Choosing a Strategy: The Synthesis of Cyclopentenones

A more conventional analysis, using the aldol disconnection 22a, leads to a strategy similar to that used for 12 and we must decide which half of the 1,4-keto-aldehyde 27 should have the natural, and which the unnatural polarity. The two methyl groups help as they do not hinder chemoselective enol formation from Me2CH.CHO, but would make the a2 reagent sterically less reactive towards an SN2 reaction. O

O

aldol

O

1,4-di-CO

+ CHO

CHO 28; a 2 synthon

27

22a

29

The choice for a specific enol equivalent is easily made: we used aldehyde enamines for just these sorts of reactions in chapter 3. You may wonder why a simple α-haloketone was not chosen for the a2 reagent. This was a matter for personal preference dictated by the need to make large quantities of 22. Stevens10 and his group preferred the propargyl halide 31 even though this meant using an Hg(II) salt to catalyse the hydration of the triple bond11 to give 27. Aldol cyclisation completed the synthesis. Br

R2NH

O

Hg(OAc) 2

CHO

31

CHO

R2N 30

base

H2SO4

22

CHO 27

32

Sometimes even the aldol reaction fails to perform with its usual reliability and we must give it some help. The bicyclic enone 33 continues the contrast with Robinson annelation as it resembles the Wieland-Miescher ketone, used as a starting material for steroid syntheses,12 but with two fivemembered instead of two six-membered rings. Aldol disconnection reveals the 1,4-diketone 34 but it turns out that closing the second five-membered ring this way works rather badly. O

O

O O

aldol

O

H

1,4-diCO

+

O O 33

O

34

35

36

The solution is to make the intramolecular aldol into a Wittig reaction. There are many ways to do this using reagents such as 38 on the stable enolate 37. Note that this reagent 38 is an ylid not a phosphonium salt, so there are no acidic protons to destroy the enolate of the dione 37. This is one example of a family of reagents3 that act first as electrophiles and then as enolate equivalents. This bicyclic enone 33 has been used in the synthesis of coriolin.13 O

O H

O 2.

1. base

+ Br

O 35

33

PPh3

O 37

O

O

38

O 39

PPh3

6 Using the Aliphatic Friedel-Crafts Reaction

75

Modern versions of these strategies were needed for Robertson’s synthesis14 of the heterocyclic core of roseophilin 40. He decided to use the cyclopentenones 41 or 42 as starting materials. OMe Cl R

O

or

NH N

Cl

O

40; R = i-Pr roseophilin

Cl

O

41

42

The standard aldol disconnection on enone 42a reveals the ketoaldehyde 43 best disconnected at the branchpoint to the simple aldehyde 44 and some a2 reagent derived from acetone.

1,4-diCO

aldol

CHO Cl

O

CHO

Cl

O

Cl

O

43

42a

44

The synthesis involved the formation of an enamine 46 from the aldehyde 45 and alkylation with chloroacetone to give the ketoaldehyde 43. Then events took a twist. Aldol reaction under equilibrating conditions (KOH, THF, H2O) gave 41 rather than 42. Cl

i-Bu2NH

OHC

O

K2CO3

i-Bu2N

Cl

Cl

NaI 18-crown-6

43 58% yield from 45

KOH THF H2O

46; enamine

45

Cl O 41; 100% yield

Though milder conditions (NaOH, Et2O) gave 42 in 86% yield, this was of course racemic. The single enantiomer was made from tartaric acid 47 by conversion into the enone 48, conjugate addition of a lithium cuprate with excellent anti-selectivity 49 and elimination with DBU.14 (CH2)6Cl HO

CO2H

OSiMe2t-Bu

OSiMe2t-Bu

I(CH2)6Cl

DBU

42 HO

CO2H

47; tartaric acid

t-BuLi, CuI

O

O 48

anti-49; 88% yield

Using the Aliphatic Friedel-Crafts Reaction Strategy B of chapter 5, disconnection between the double bond and the carbonyl group 50, looks at first sight to be of little use as it would require an intramolecular aliphatic Friedel-Crafts reaction

76

6 Choosing a Strategy: The Synthesis of Cyclopentenones

on a compound of equal size. It is of value because the starting material 51 comes from some acid derivative 52 that can be made by allylation of an enolate. The Friedel-Crafts should be regioselective as cyclisation of the other end of the alkene would give a four-membered ring.

Friedel-Crafts

FGI

Cl

R

O

O

O

O

OR

OR X

R

R 50

allylation of enolate

51

R

52

This strategy has been used to make the simple cyclopentenone 54 from the unsaturated carboxylic acid 53 presumably by intramolecular aliphatic Friedel-Crafts reaction 56 and loss of a proton from the resulting cation 55. This strategy was used because 53 was available by the telomerisation of butadiene, but it would be simple to make such unsaturated acids in other ways. O

O

O 1. SOCl 2

OH

O

H R

2. AlCl 3

R 53

55

54

56

A more exciting version15 adds a three carbon unit (the d3 reagent 57) twice to the symmetrical diketone 58, oxidises the resulting tetraol 59 to a diacid which spontaneously cyclises to the dilactone 60, and generates the acylating agent and the double bond needed for the Friedel-Crafts reaction by acid-catalysed dehydration. The resulting curved row of cyclopentane rings 61 is clearly one possible starting point for the synthesis of dodecahedrane 3.

Li

O

OEt

57; d 3 reagent

Cr(VI)

1. 2 x 57

O

O

2. H , H2O

HO

OH

OH

PPA

O

O

MsOH

OH O

O 58

O

O

60

59

61

The first disconnection 50 or 62 is the same for both these approaches, giving in each case an unsaturated acid as the intermediate. In the open chain case 52, we could proceed by a Wittig disconnection of the double bond or a simple alkylation with an allylic halide 53. For the cyclic case (using a six-membered ring as the second ring) 63 these are much less attractive. The bond joining the ring to the chain is a strategic bond and we can disconnect it after FGI to the alcohol 64 reveals the need for a d3 reagent or homoenolate. Further examples of these reagents appear in chapter 13. O FriedelCrafts

FGI

CO2H

CO2H +

dehydration

62

CO2H

63

OH 64

O d3

reagent

6 The Nazarov Reaction

77

The Nazarov Reaction In addition to the application of known and established strategies, each structural type of target molecule may have special methods developed for, in this case, cyclopentenones alone. The Nazarov reaction is historically important as one of the few unexplained reactions which led Woodward towards the Woodward-Hofmann rules for pericyclic reactions.16 Treatment of a crossconjugated dienone, 65 being a minimalist example, with acid or Lewis acid leads to a cyclisation in which two electrophilic atoms react with each other to give a cyclopentenone 69. Seen as an ionic reaction, it makes no sense, but as a conrotatory 4n electrocyclic reaction17 it is entirely reasonable. Protonation gives a pentadienyl cation 66 which cyclises to a cyclopentenyl cation 67, the driving force being formation of a new σ-bond that more than compensates for the loss of conjugation. Loss of a proton and enol-keto tautomerism complete the reaction. O

OH

OH 4n conrotatory electrocyclic

H

OH

O

68

69

–H

reaction

65

66

67

Inspection of the starting material and the product reveal that the disconnection is of the bond in the ring opposite the carbonyl group. Applying this to the molecule 70 we discussed in chapter 5, and for which we rejected the aldol strategy, we discover that we need the dienone 71. An attractive disconnection corresponds to the Friedel-Crafts reaction we have just been discussing and requires in this instance cyclopentene and the unsaturated acid chloride 72. O

O Nazarov

+

70

H

O FriedelCrafts

Cl

71

75 modhephene

72

Oppolzer,18 who wanted this cyclopentenone for the synthesis of modhephene 75 followed this strategy and found that the Friedel-Crafts went well with AlCl3 as catalyst and the Nazarov cyclisation with another Lewis acid, SnCl4. You will notice the position of the double bond in the final product: there are no protons on one side of the cation 73 so one must be lost from the other side. In cases where there is a choice, the more substituted alkene results, as here. The product is ideally functionalised for development into modhephene 75. OSnCl3 72 AlCl3

71 79% yield

OSnCl3

SnCl4

H2O

H 73

70 60% overall yield from cyclopentene

74

In chapter 16 we shall see how the use of vinyl silanes has solved some of the regioselectivity problems inherent in the Nazarov reaction.19 Meanwhile we look at some recent developments on

78

6 Choosing a Strategy: The Synthesis of Cyclopentenones

the effects of substituents on the reaction. The introduction of fluorine into organic molecules is of the greatest importance for the synthesis of new drugs. The synthesis of the trifluromethyldienone 78 gives you a preview of methods we shall see later involving Sn and Pd. For the moment accept that this is a way to produce a potential substrate for a Nazarov cyclisation.20 O CF3

Br

CF3

CuI, Bu 3SnLi

SnBu3

ClOC Pd(0)

CF3

+ Cu(I)

76

76

78; 75% yield

77

The Nazarov reaction is carried out with Me3SiOTf as Lewis acid through intermediates 79 and 80. Elimination occurs exclusively away from the CF3 group to give 81 in excellent yield. OSiMe3 Me3SiOTf

78

Me3SiO

O

CF3 CF3

CH2Cl2

CF3

79

81; 95% yield

80

A silyl group, on the other hand, attracts the alkene in the product. The dienone 82 cyclises easily in acid to give only the less substituted cyclopentenone 84 in excellent yield.21 We shall discuss the stabilisation of cations such as 83 and allyl silanes such as 84 in chapter 12. OH

O SiPh2t-Bu

H SiPh2t-Bu

CF3CO2H

O SiPh2t-Bu

THF, 60 °C

H 82

H 84; 85% yield

83

Cycloadditions of Fe(CO)4 Complexes of Oxyallyl cations These Nazarov approaches add three extra carbon atoms (usually CH2CH2CO) onto a ketone to form the five-membered ring. It is also possible to add the three atoms CH2CO.CH2 onto an alkene to form a cyclopentenone of different structure. The oxyallyl cation 87, which can be made by the action of a metal, often zinc, to α,α⬘-dibromoketones 85, might provide the three carbon atoms. Unfortunately it is a two electron system and so adds to dienes 88 rather than mono-enes giving cycloheptenones 89 in good yield.22

O

O

O

O π4σ

metal

Br

Br 85

O

89a

89b

cycloaddition

Br 86

O + π2σ

87; the oxyallylcation

88

6 The Pauson-Khand Reaction

79

If iron carbonyl is used to generate the oxyallyl cation, the iron supplies two extra electrons in the complex 90 which now adds23 to mono-enes to give cyclopentanones 91. We shall see more examples of transition metal complexes altering selectivity later in the book. To get a cyclopentenone 92 we need an extra degree of unsaturation, best provided by using the enamine of a ketone instead of a simple alkene. O

O

O Fe(CO)5

R

90

×

Br

Fe(CO)4

Br 85

R

R

90

O

NR2

R 91

92

This cycloaddition works best with substituted oxyallyl cations, like the one from the dibromoketone 93 which reacts with the morpholine enamine of cyclohexanone 95 to give the cyclopentenone 96 in excellent yield.24 You should mark the difference in structure between the Nazarov products, e.g. 70, formed in an electrocyclic reaction, and these cycloadducts, e.g. 96. Each strategy has its part to play and must be chosen according to the structure of the target molecule.

O

O

O 1. Fe(CO) 5

N

+ Br

O

Fe(CO)4

Br 93

95

94

96

The Pauson-Khand Reaction The typical Nazarov strategy 51 followed by 53 disconnects the rest of the ring from the double bond, but you cannot be quite sure where the double bond will end up in the ring. Another special method, the Pauson-Khand reaction,24, 25 differs in both respects. It adds the enone portion of the ring to the rest of the molecule and you can be quite certain where the new double bond will be. The reaction is between an alkene, say cyclopentene 97, an alkyne, and Co2 (CO) 8 to form a cyclopentenone26 98 in one step. A complex 99 is first formed between the alkyne and the cobalt atoms with the two π-bonds replacing two CO molecules (both are two-electron donors). You may see this complex drawn as 99a but it really has a tricyclic ‘tetrahedrane’ structure 99b composed of three-membered rings and with a Co–Co bond. O H

H

H

H

Co(CO)3

Co2(CO)6

Co2(CO)8

H 97

H

Co2(CO)8

98

H

H 99a

Co(CO)3 99b

The π-bond of the alkene then replaces another CO on one Co atom (we hope you are alert to the difference between Co and CO!) to give the π-complex 100. A returning molecule of CO displaces the alkene from the π-complex and forces it to make the first new σ-bond, between one end

80

6 Choosing a Strategy: The Synthesis of Cyclopentenones

on the alkene and one end of the alkyne, sacrificing one C-Co σ-bond to give the σ-complex 101. Another molecule of CO returning to the Co atom forces one of the other CO molecules to insert into the Co-C bond to give the acyl σ-complex 102. This is a standard reaction of organometallic compounds which we shall meet again later. Co(CO)3

Co(CO)3

Co(CO)2

–CO

97 + 99

Co(CO)3

Co(CO)3

+CO

Co(CO)3

+CO

O

100; π-complex

101; σ-complex

102; acyl σ-complex

The cobalt-acyl σ-complex 102 is unstable and forms the third and last C–C σ-bond by addition of the carbon end (marked with a blob) of a C–Co bond to the carbonyl group 102a and decomposition of the adduct 103 with loss of another Co–C bond. The product 104 is an unstable cobalt complex of an alkene so the cobalt atoms drop out of the molecule altogether, leaving a π-bond 98 where the old alkyne used to be. Cobalt carbonyl complexes of alkynes are stable: those of alkenes are unstable. The mechanism 102a to 103 to 104 is very much an organic chemist’s view of things: inorganic chemists may well be sceptical.

Co(CO)3

(CO)3Co

Co(CO)3 (OC)3Co

Co(CO)3

Co(CO)3

O

O

O

102a

103

O

–Co2(CO)6

104

98

There is good regioselectivity with unsymmetrical alkynes, e.g. 1-heptyne 105; R ⫽ n-pentyl forms the 2-substituted cyclopentenone 106 in excellent yield26 but unsymmetrical alkenes give much poorer selectivity, e.g. 1-octene gives 21% each of the cyclopentenones 108 and 109; R1 ⫽ n-pentyl R2 ⫽ n-hexyl.27 It must be admitted that this combined yield of 40–50% yield is more typical than the 80–85% of 106, but this can be forgiven in a three-component reaction that accomplishes so much. In the formation of the first σ-bond a substituent on the alkyne evidently prefers the blobby position in 102 while substituents on the alkene are indifferent about remaining on the atom joined to cobalt.

R

O

Co2(CO)8

O R1

Co2(CO)8

R CH2=CH2

O

R1

+ R1

R2

R2

R2 105

106; 80-85%

107

108

109

The disconnections are simple 110 but one must look for a symmetrical alkene and make sure that the substituent on the alkyne is next to the ketone.

6 The Pauson-Khand Reaction

81

O O

R

Pauson-Khand

R

C

110

Intramolecular Pauson-Khand reactions are often regioselective because it is physically impossible for the molecule to cyclise any other way. Pauson-Khand disconnection of bicyclic 111 reveals an allyl ether 112 of the alcohol 113, easy to make from acetylene and cyclohexanone. In the cobalt-catalysed cyclisation, only one regioisomer is possible and this, the TM111, is formed in an excellent 80% yield.28

C–O

Pauson-Khand

O

O

+ Br

O HO

111

112

113

The synthesis of the antibiotic methylenomycin A 114 is instructive on the choice between modern and traditional methods. The key intermediate 116 has been converted into methylenomycin by several groups whose strategy differs only in the synthesis of 116. The PausonKhand disconnection is simplicity itself, requiring the symmetrical unsaturated ether 117 and symmetrical butyne. O

O

O H

reconnection

H +

O

O

PausonKhand

H

FGI

O

CO2H

H

114 methylenomycin A

O

O H

H

O

116

115

117

The traditional aldol analysis requires a 1,4-diketone 117 whose stereochemistry suggests a 2 ⫹ 2 photochemical cycloaddition29 also on 117 and oxidative cleavage of the product 118. O

O H

H

H reconnection

aldol

O H 116

2+2

O

O

O H 117

+ 117

H 118

This approach was used by Amos Smith30 starting with maleic anhydride to get an efficient photochemical reaction. Every step goes in good yield and can be carried out on a large scale, but there are five separate steps and only a 42% overall yield.

82

6 Choosing a Strategy: The Synthesis of Cyclopentenones

O

H

O 1. LiAlH 4 (96% yield)

hν

O

118

O

MeCN

H

O

1. O 3, MeOH

117 75% yield

2. Ph 3P, MeOH

2. TsCl, pyr (86% yield)

2% NaOH

116 H2O (85% yield)

O

In contrast the Pauson-Khand approach used by Billington31 is only one step, but gave less than 20% yield when first attempted. Gradually this was pushed up until it was claimed as a 70% yield, but there is a lot of recovered starting material and the conversion is lower. You must make your own choice! Improvements in the Pauson-Khand procedure using silica support,28 ultrasound,26,32 and, best of all, reaction in the presence of promoters such as N-methylmorpholine N-oxide33 (NMO) 119 have led to increased yields. Probably most of the yields quoted so far could be improved if this last method were used.

Recent Developments in the Pauson-Khand Reaction Asymmetry as well as improved procedures has been a high priority recently. The heterocyclic enyne 122 prepared from natural proline 120 gave initially only a 6% yield of the tricyclic amine 123. Addition of NMO 119 increased this to 72% but DMSO (Me2S⫽O) was the best promoter raising the yield of a single diastereoisomer of 123 to near quantitative.34 Me

NH2

NH

N O O

CO2H

121

1. Co2(CO)8 THF, 2 hours

N

N

2. 6 x DMSO 50 °C, 26 hours

CO2Me

Br NaHCO3, LiI MeCN, 60 °C

CO2Me

120; (S)-proline

119; NMO

Cl

O

CO2Me

122; 96% yield

123; 94% yield

Since the intramolecular reactions are so much the best, others have linked the alkene and the alkyne by a weak bond that can later be sacrificed. This is the ‘tether’ strategy you will meet in chapter 36. An N–O bond is ideal and with the alkyne additionally blocked with a silyl group 124, good yields of Pauson-Khand product 125 and of the amino alcohol 126 could be achieved even with an amine N-oxide as promoter. Samarium(II) iodide was used as the reducing agent.35

Me3Si

N O

124

Boc

1. Co 2(CO)8 CH2Cl2 2. Me 3N+–O–

O

N O

Me3Si 125; 76% yield

Boc SmI2 THF EtOH

O

N H

Me3Si OH 126; 83% yield

Boc

6 Oxidative Rearrangement of Tertiary Allylic Alcohols

83

Other recent developments include a high yielding and totally regioselective intermolecular reaction between an unsymmetrical alkyne 127 and propene. The cobalt complex 128 was isolated and treated with propene using NMO hydrate as promoter to give a high yield of the cyclopentenone 129. The methyl group from propene ends up next to the ketone as expected but the regioselectivity of the alkyne, evidently due to the conjugating ester group, is remarkable.36

OTBDMS

OTBDMS Co2(CO)6

CO2Et

O

OTBDMS

Co2(CO)8 petrol, 0 °C

NMO.H 2O CH2Cl2

CO2Et

CO2Et 128; 100% yield

127

129; 92% yield

Other catalysts have been developed. Cobalt deposited on charcoal is effective in a range of regioselective intramolecular reactions, though a 20 atmosphere pressure of CO must be used.37

MeO2C

20 atmospheres CO

MeO2C

MeO2C

12 weight % Co/C

MeO2C

O

130 °C, THF

131; 98% yield

130

In chapter 19 you will meet palladium allyl cations as useful reagents and Evans and Robinson38 have combined the Pauson-Khand reaction with allyl cation complexes using rhodium as a compromise between Pd and Co. The enolate 132 combines with the Rh(I) cation complex from the allylic carbonate 133 to give the enyne 134 that gives the Pauson-Khand product 135 in 87% yield with the same catalyst but at higher temperatures.

MeO2C

+

OCO2Me

MeCN, 30 °C

MeO2C 132 MeO2C MeO2C 134; 88% yield

cat Rh(I)

133 CO cat Rh(I) MeCN, reflux

MeO2C MeO2C

O

H Me 135; 87% yield

Oxidative Rearrangement of Tertiary Allylic Alcohols The oxidative rearrangement route (chapter 5) has been used in many syntheses of cyclopentenones. Büchi’s route39 to cis jasmone 1 simply involves treating the cyclopentenone 136 with MeLi and oxidatively rearranging the product 137 directly to cis jasmone 1.

84

6 Choosing a Strategy: The Synthesis of Cyclopentenones

O CrO3

MeLi

Me OH

O

1; cis-jasmone

137

136

The natural prostaglandins, like PGA2 2 are usually made by the oxidation of an allylic alcohol.40 This apparently trivial strategy is used because the stereochemistry around the fivemembered ring is easily derived from the bicyclic enone 138 by Baeyer-Villiger rearrangement.41 Interesting principles of chemoselective oxidation are involved both in the formation of 139 and in the synthesis of the starting material 138.

RO

CO2H

1. H 2O2, KOH 2. NH 3, Et 2O

138

CO2Me

OR

CHO

HO

O

RO 139

140

In the PGA synthesis, a Wittig-style reaction puts in one side chain to give 142 which has the wrong arrangement of double bond and OH group. This time oxidation cannot be used to force the rearrangement as the alcohol is secondary, so a displacement is used 143. This is stereospecific because the incoming CO2H group is physically constrained to approach from underneath to give the cis fused lactone 144. A reduction and another Wittig reaction complete the synthesis. PGAs are so easily made that they are now used as large scale intermediates in the synthesis of other PGs.42

O O

CO2Me

O

(MeO)2P

1. NaH

R

CO2H

R

2. 140

RO

O

141; R = n-pentyl

OH 2; PGA 2

142 O

O

O O

1. DIBAL 2. Wittig

142 R

R

2; PGA 2

3. Cr (VI) 4. H

RCO2

O 143

O 144

Other Methods The recent discovery that unsaturated five-membered carbonyl compounds bearing two aromatic rings such as Merck’s MK-0966 145 are new generation anti-inflammatory compounds led

6 Other Methods

85

medicinal chemists43 to make analogues such as 146. Ideal disconnections would be the removal of the two aromatic rings with the idea that they could be added as organometallic compounds to the two electrophilic sites in 147. A similar idea was used for a cyclohexenone, compound 95 in chapter 5. O O

O O

O O

S

S

S

Me

Me X

?

O

Me

Br Br

O

O

O N

145; MK-0966

146

N 148

147

The enol ether 150 was reacted with the lithium derivative of the thioether 149 to give the intermediate 151 after oxidation. A Suzuki coupling, as you will learn to call it, linked the boronic acid derivative of the pyridine to the bromoenone to give the cyclopentenone 146 in just two steps. SMe

O O 1. BuLi, THF

149

S

2. 150 3. HCl, H 2O 4. [O]

Br O

Pd(0), Ph 3P

Me

146 73% yield

(i-PrO)3B

N

150 Br

Br

O

EtO

151; 71%

Among the many new methods, one deserves special attention because it is so simple to carry out and yet includes such remarkable reactions.44 The ynoate 152 reacts with ketones to give four membered cyclic ester enolates that can cyclise 153 onto neighbouring ester groups to give bicyclic keto-lactones 154 that readily lose CO2 to give cyclopentenones 155. The origin of 152 and the mechanistic implications of each step are explored in the workbook.

O EtO

CO2Et

O

R

'2 + 2'

+

O

O R 152

intramolecular Dieckmann

O

R

R

–CO2

O

O 153

O

O

154

155

Finally another extraordinary route to nitrogen-containing cyclopentenones using interesting chemistry. de Meijere45 used chromium carbonyl complexes 160 formed from terminal alkynes 156 via addition to Cr(CO) 6, insertion of the alkyne into one of the CO ligands to give the carbene complex 158, alkylation on oxygen to give 159 and conjugate addition of a secondary amine.

86

6 Choosing a Strategy: The Synthesis of Cyclopentenones

O

O

H

CO

1. BuLi 2. Cr(CO) 6

O

O

O

O

(CO)5Cr

156

157

158 OEt

OEt Et3O

O

(CO)5Cr

O

BF4

O

Me2NH

(CO)5Cr

O

O

(CO)5Cr Me2N 160; 68% yield

159

These complexes reacted with a second alkyne 161 without carbonyl insertion to give the cyclopentadiene 161. This intermediate is also the enol ether of a cyclopentenone and hydrolyses in strong aqueous acid to give a bicyclic compound 162. EtO

OEt O

O

+ (CO)5Cr SiMe3 161

pyridine

O O

conc HCl

80 °C

Me2N 160

OH

O THF

NMe2 Me3Si 161; 40% yield

Me3Si

NMe2 162

Old and new chemistry provides many ways to make cyclopentenones. This chapter only hints at the richness of modern organic chemistry. As we progress through the book most of these methods will be explained in more detail.

References General references are given on page 893 1. G. Stork, A. A. Ozorio, and A. Y. W. Leong, Tetrahedron Lett., 1978, 5175. 2. Disconnection Approach, page 170. 3. M. Ramaiah, Synthesis, 1984, 529. 4. A. Yoshikoshi and group, J. Am. Chem. Soc., 1976, 98, 4679. 5. Disconnection Approach, chapter 22, page 179. 6. S. Danishefsky, K. Vaughan, R. C. Gadwood and K. Tsuzuki, J. Am. Chem. Soc., 1980, 102, 4262; 1981, 103, 4136. T. Hudlicky, T. M. Kutchan, S. R. Wilson and D. T. Mao, J. Am. Chem. Soc., 1980, 102, 6351; W. K. Bornack, S. S. Bhagwat, J. Ponton and P. Helquist, J. Am. Chem. Soc., 1981, 103, 4647. 7. Disconnection Approach, page 202. 8. S. Wolff, W. L. Schreiber, A. B. Smith and W. C. Agosta, J. Am. Chem. Soc., 1972, 94, 7797. 9. A. J. Bellamy, J. Chem. Soc. (B), 1969, 449. 10. R. V. Stevens, R. E. Cherpeck, B. L. Harrison, J. Lai and R. Lapalme, J. Am. Chem. Soc., 1976, 98, 6317. 11. Disconnection Approach, chapter 16, page 126. 12. Disconnection Approach, page 175. 13. B. M. Trost and D. P. Curran, J. Am. Chem. Soc., 1980, 102, 5699; 1981, 103, 7380. 14. J. Robertson, R. J. D. Hatley, and D. J. Watkin, J. Chem. Soc., Perkin Trans. 1, 2000, 3389; K. Ogura, M. Yamashita and G. Tsuchihashi, Tetrahedron Lett., 1976, 759; A. G. Myers, M. Hammond and Y. Wu, Tetrahedron Lett., 1996, 37, 3083.

6 References

87

15. P. E. Eaton and R. H. Mueller, J. Am. Chem. Soc., 1972, 94, 1014. 16. R. B. Woodward and R. Hoffmann, Angew. Chem., Int. Edn., 1969, 8, 781. 17. I. Fleming, Orbitals, page 103. 18. W. Oppolzer and K. Bättig, Helv. Chim. Acta, 1981, 64, 2489. 19. J. Mulzer, H.-J. Altenbach, M. Braun, K. Krohn, and H.-U. Reissig, Organic Synthesis Highlights, VCH, Weinheim, 1990, page 137. 20. J. Ichikawa, M. Fujiwara, T. Okauchi and T. Minami, Synlett., 1998, 927. 21. A. Barbero, P. Castreño, C. Garcia and F. J. Pulido, J. Org. Chem., 2001, 66, 7723. 22. H. M. R. Hoffman, Angew. Chem., Int. Ed., 1973, 12, 819. 23. Y. Hayakawa, K. Yokoyama and R. Noyori, J. Am. Chem. Soc., 1978, 100, 1799. 24. N. E. Schore, Org. React., 1991, 40, 1. 25. P. L. Pauson, Tetrahedron, 1985, 41, 5855. 26. D. C. Billington, I. M. Helps, P. L. Pauson, W. Thompson and D. Willison, J. Organomet. Chem., 1988, 354, 233. 27. M. E. Kraft, Tetrahedron Lett., 1988, 29, 999. 28. W. A. Smit, A. S. Gybin, A. S. Shashkov, Y. T. Strychkov, L. G. Kyz'mina, G. S. Mikaelian, R. Caple and E. D. Swanson, Tetrahedron Lett., 1986, 27, 1241. 29. Disconnection Approach, chapter 32, page 268. 30. R. M. Scarborough, B. H. Toder and A. B. Smith, J. Am. Chem. Soc., 1980, 102, 3904. 31. D. C. Billington, Tetrahedron Lett., 1983, 24, 2905. 32. P. Blaydon, P. L. Pauson, H. Brunner and R. Eder, J. Organomet. Chem., 1988, 355, 449. 33. S. Shambayati, W. E. Crowe and S. L. Schreiber, Tetrahedron Lett., 1990, 31, 5289. 34. S. Tanimori, K. Fukubayashi and M. Kirihata, Tetrahedron Lett., 2001, 42, 4013. 35. S. G. Koenig, K. A. Leonard, R. S. Löwe and D. J. Austin, Tetrahedron Lett., 2000, 41, 9393. 36. M. E. Krafft, Y. Y. Cheung and K. A. Abboud, J. Org. Chem., 2001, 66, 7443. 37. S. U. Son, S. I. Lee and Y. K. Chung, Angew. Chem., Int. Ed., 2000, 39, 4158. 38. P. A. Evans and J. E. Robinson, J. Am. Chem. Soc., 2001, 123, 4609. 39. G. Büchi and B. Egger, J. Org. Chem., 1971, 36, 2021. 40. E. J. Corey and G. Moinet, J. Am. Chem. Soc., 1973, 95, 6831. 41. Disconnection Approach, page 226. 42. E. J. Corey and X-M. Cheng, The Logic of Chemical Synthesis, chapter 11, page 250. 43. W. C. Black, C. Brideau, C.-C. Chan, S. Charleson, N. Chauret, D. Claveau, D. Ethier, R. Gordon, G. Greig, J. Guay, G. Hughes, P. Jolicoeur, Y. Leblanc, D. Nicoll-Griffith, N. Ouimet, D. Riendeau, D. Visco, Z. Wang, L. Xu and P. Prasit, J. Med. Chem., 1999, 42, 1274. 44. M. Shindo, Y. Sato and K. Shishido, J. Org. Chem., 2001, 66, 7818. 45. H. Schirmer, F. J. Funke, S. Müller, M. Noltemeyer, B. L. Flynn and A. de Meijere, Eur. J. Org. Chem., 1999, 2025.

Section B: Making Carbon–Carbon Bonds 7. The Ortho Strategy for Aromatic Compounds .................................................................. 91 8. σ-Complexes of Metals ....................................................................................................... 113 9. Controlling the Michael Reaction ..................................................................................... 127 10. Specific Enol Equivalents ................................................................................................... 139 11. Extended Enolates............................................................................................................... 155 12. Allyl Anions ......................................................................................................................... 173 13. Homoenolates ...................................................................................................................... 189 14. Acyl Anion Equivalents ..................................................................................................... 203

7

The Ortho Strategy for Aromatic Compounds Most of this chapter is dedicated to ortho-lithiation—a powerful synthetic technique. Part I looks at more traditional methods that do not include lithiation.

Introduction PART I – TRADITIONAL METHODS Friedel-Crafts Reaction and Fries Rearrangement The Claisen Rearrangement PART II – USING LITHIUM Ortho-lithiation Anisole Regioselectivity I – The ortho versus para question Directing groups Directing groups containing oxygen Directing groups containing nitrogen Regioselectivity II — ortho versus ortho Choice in ortho-lithiation Several lithiations One ortho-director leads to another Dilithiations One ortho-director rearranges into another – the anionic Fries reaction Halogens Benzyne Formation α-Lithiation Lateral Lithiation Chemoselectivity Summary

Introduction Building aromatic compounds by electrophilic substitution using ortho and para or meta directing effects is part of the early training of all organic chemists. Reactions such as the Friedel-Crafts acylation, that efficiently build carbon-carbon bonds between aromatic rings and aliphatic side chains are well known and have been assumed so far in this book. It is usually assumed that the ortho and para mixtures, that arise when such reactions are applied to activated aromatic rings, are acceptable because they can be carried out on a large scale and separation is usually simple. This may be true for very simple compounds as uses can often be found for the unwanted isomer, Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

92

7 The Ortho Strategy for Aromatic Compounds

but the modern organic chemist needs ways to control or divert such mixtures into high yields of single compounds. Making the para substituted compound is not usually difficult as large substituents direct para. Making high yields of ortho compounds regardless of the size or electronic nature of the substituents is no such easy matter and is the subject of this chapter. OMe

OMe RCOCl

1

2

R

AlCl3

O

PART I Friedel-Crafts Reaction and Fries Rearrangement The Friedel-Crafts acylation1 on simple aryl ethers such as anisole PhOMe 1 usually gives almost quantitative yields of the para ketone 2. If the acid chloride is attached to the oxygen atom by a short chain of carbon atoms (a ‘tether’) 3 the resulting intramolecular reaction is forced to give the ortho product 4, and these reactions also go in good yield.2 Cl

O

O AlCl3

3

4

O O

If the acid group is joined directly to the oxygen atom as an ester 6 we have the Fries rearrangement3 which can be controlled to give either the para or ortho products. In polar solutions such as nitrobenzene,4 the para product 5 is formed in high yield, e.g. R ⫽ Me, 75%, while in non-polar solvents or in the absence of solvent, the ortho product 7 is formed in similar yield, e.g. R ⫽ Me, 70%. OH

O

AlCl3 PhNO2

R

R O

OH

AlCl3 165 °C

R

O

O 6

5

7

This is not a concerted rearrangement mechanism of the usual cationic sort, like the pinacol (Chapter 2), but a normal ionic reaction. The Lewis acid catalyses the breakdown of the ester 8 into an acylium ion and a metal complex of the phenol which remain associated as an ion pair 9 in non-polar solvents. This naturally tends to give the ortho product as the acylium ion is held close to the ortho position by electrostatic interactions. In polar solvents, the ion pair is separated into two independent ions which show the normal Friedel-Crafts selectivity, that is high preference for the para product. AlCl3 O

R

O

AlCl3

AlCl3 R

O

R 7

O 6

O

O 8

9

7 Friedel-Crafts Reaction and Fries Rearrangement

93

Derivatives of meta cresol 10 are even more complicated. Three different sites are activated (arrows on 10), but Friedel-Crafts acylation of simple ethers 11 favours the position para to the ether group almost exclusively, anhydrides being the best acylating agents.3 Me Me

OH

Me

OMe

Ac2O

OMe

Me

AlCl3

10

O 12; 87% yield

11

The Fries rearrangement on simple esters 13 by contrast favours one of the ortho positions equally exclusively, a very high yield of the ketone 14 being obtained in the absence of solvent.3,4 In nitrobenzene, 28% of 14 and 64% of 15 are formed.5 Me

O

Me

Me

OH Me

165 °C

O

Me

OH

AlCl3

+

Me

O 14; 95% yield

13

O 15

When Rao6 wanted to make sydowic acid 16, a sesquiterpene from Aspergillus sydowi, he chose to disconnect the ether first and to make the triol 17 by addition of three methyl groups to the keto-ester 18. This is clearly a Friedel-Crafts product but how are we to get the correct position on the aromatic ring? OH OH

OH

C–C 4 x Me

C–O

O

OH

ether

HO2C

(Grignard)

HO2C 16; sydowic acid

17 OH

CO2R

OH

FriedelCrafts

+

O

O

O

O

HO2C

HO2C 18

19

20

Knowing more about the behaviour of meta cresol 10 than of the hydroxy-acid 19 with the same substitution pattern, Rao chose to work at the ‘wrong’ oxidation level through the Fries, using the cyclic anhydride 20 as the acylating agent. This gave the right isomer 21, and, by addition of four molecules of methyl Grignard (one is used up by the free OH group), and acid-catalysed cyclisation, the intermediate 23. Sadly, direct oxidation destroys the phenol, so it must be protected as an ester. This short synthesis involves regioselectivity twice: the obvious ortho Fries selectivity and regioselectivity again in the formation of a six- instead of a four- or eight-membered ring in the cyclisation of 22.

94

7 The Ortho Strategy for Aromatic Compounds

CO2Me

OH

4x MeMgI

1. 20, AlCl 3

m-cresol 10

O 2. MeOH, H

Me

21

OH OH

OH 1. Ac 2O

H2SO4

O

OH

2. KMnO 4

TM16

3. NaOH

Me

Me

23

22

The remaining site in meta cresol 10, between the OH and Me groups, is very difficult to get into, as is commonly the case when one tries to insert a largish substituent between two others which are already there. A better strategy is usually to put two next to each other and add an outside substituent last. When Wiesner7 was planning the synthesis of some delphinine alkaloids, he required the indanone 25 as a starting material in large amounts (150 g scale). The Friedel-Crafts disconnection of 25 would require an acylation of 24 between two substituents and will not give 25. b

X

Friedel -Crafts

CO2R

Me

Me

a

Fries

Nazarov

OH 24

a OH 25

b

Me

O

a

OH 26

O

Me OH 27; o-cresol

Indanones are an aromatic variety of the cyclopentenones we were considering in Chapter 6, and the Nazarov strategy looks appealing as it becomes a Friedel-Crafts reaction on ortho cresol 27 with the acid chloride of acrylic acid. The problem is to make the reaction give just that isomer, that is the one having four adjacent substituents. The Fries rearrangement is the answer as it transfers the acyl group to the next carbon atom round the ring and the alkylation step has to occur at the next ortho atom in turn. The unsaturated acid chloride is too reactive, but either of the halides 28 or 31 will do the trick.7,8 Cl Cl

AlCl3

O

Cl

ArO Fries

27

O

28

31

Me OH

Fries

ArO

Me

27

O

O OH 30

Br AlCl3

Br

Cl

29 Br

24 23% yield

Me

O 32

Br

34

O

24 35% yield

O OH 33

The yields are not wonderful, but this is the first step and it accomplishes a lot in one reaction. Presumably either haloester or the haloketones 30 or 33 derived from them can form the double

7 The Claisen Rearrangement

95

bond under the reaction conditions so that the Nazarov intermediate 34 is formed and cyclised by the same Lewis acid.

The Claisen Rearrangement One reaction much used to produce the ortho relationship is the Claisen rearrangement.9 An allyl ether 36 from a phenol and an allylic halide 35 undergoes [3,3] sigmatropic rearrangement to give an intermediate 37 that is the “keto form” of a new phenol 38 with the allyl group now in the ortho position. OH

Br

O

O [3,3]

base

+

OH

H

H phenol

35

36

37

38

The rearrangement is regiospecific with regard to the allyl portion, substituents starting next to the oxygen atom 39 finishing at the far end of the double bond 40. This example is particularly favourable as the double bond becomes more highly substituted as the reaction proceeds. OH

O reflux Et2NPh

39

40

Substituents in the middle of the allyl system 41 produce no regioselectivity problem, but the products 42 are rather inclined to cyclise to ethers 43 by protonation of the double bond.10 OH

O

heat

O H

41

42

43

Substituents which start at the end of the double bond should finish next to the ring: the formation of 45 is a particularly impressive example as the double bond moves out of conjugation. These are the most tricky type of Claisen rearrangement and other products can be formed if conditions are not carefully controlled.9 OH O 44

Ph

heat

45 Ph

There can be some regioselectivity if there are two different ortho positions. This is usually attributed to ‘double bond fixation’ as in the naphthalene 46 where the structure drawn represents

96

7 The Ortho Strategy for Aromatic Compounds

more than just a Kekulé form. The bonds marked double are shorter than those marked single and the products are 9:1 in favour of migration to the ‘better’ ring double bond in 47.

O

OH

OH

heat

+ 46

47

9:1

48

The side chain produced by the Claisen rearrangement contains a useful alkene that can be elaborated by epoxidation, for example, and the reaction is slightly more versatile than it appears at first sight.

PART II Using Lithium So far, the only metal in sight to help us form aromatic compounds has been aluminium. Now we look at a metal that activates aromatics in an entirely different way—lithium. Lithium is introduced ortho to a ring substituent that already exists 50 and the process itself called ortho- or directed lithiation. XR

XR

XR E

BuLi

[an electrophile]

Li 49

E

50

51

Ortho-lithiation Before we go any further we should be clear that ortho-lithiation is a proton-lithium exchange reaction and is not to be confused with halogen-lithium exchange. For instance, the exchange of bromine for lithium in 52 occurs, it just so happens, next to another substituent. But to call that reaction an ortho-lithiation would be misleading because we would not want to imply that the methylsulfanyl group had anything to do with it – the bromine will exchange whether the sulfide is there or not.

Anisole We have already seen how anisole 1 may be functionalised by using a Friedel-Crafts reaction. Anisole is a simple, perhaps the most simple, functionalised benzene ring that can be lithiated by ortho-lithiation. We use it here to introduce the reaction. SMe

SMe

OMe

BuLi

Br 52

OMe 1. BuLi

not ortho lithiation

2. CO 2

Li 53

1

CO2H 54; 50% yield

7 Ortho-lithiation

97

Firstly, let us consider the substituent ortho to the newly introduced lithium. The ortho directing group here is a methoxy group and within it is the oxygen atom that is doing the work. Without concerning ourselves with the mechanism of the reaction too much at this stage, the lone pair on the oxygen atom coordinates to the organolithium reagent. Notice that the oxygen is attached directly to the ring – anisole is a phenyl ether. We shall see examples later where the heteroatom is more remote. The group which directs the lithiation we shall refer to as the ‘ortho-director’ although it is called other things by different authors (DMG for Directed Metallation Group for instance).11

Regioselectivity I—The ortho versus para question Even without lithium, the reactivity of aromatic rings is nucleophilic. But the source of that nucleophilicity is very different when a lithium atom is in place. The lithium atom does more than dramatically increase the reactivity of the benzene ring, it alters its character entirely. When benzene reacts with an electrophile like NO2⫹, it is the π electrons of the ring that attack the electrophile and the aromaticity of the benzene ring is necessarily disrupted. Any one of six carbon atoms may be the focus for the attack on NO2⫹, it makes no difference. With a substituted benzene ring, we know that some sites are more reactive than others due to the influence of the substituent – usually the ortho and para positions. But with a lithiated benzene ring the situation is different. Now the attacking electrons are not π but σ. They are localised and have no choice as to which carbon atom they attack from. And it doesn’t matter if the substituent increases the π-electron density a little at this or that carbon because it will not compete with the full negative charge from the C-Li bond. A ‘full negative charge’ would imply that we have an anion. This is not the case but organolithium reagents are carbanion σ-complexes (chapter 8). Once the lithium atom is in place the reaction with electrophile is regiospecific. Anisole reacts at ortho or para positions with π-electrons OMe E+

E+ or MeO

Lithiated anisole reacts at ortho position with σ-electrons OMe E+

Li MeO

MeO Li

But what about where the lithium atom goes in the first place? If, as is usually the case, the directing substituent coordinates to the organolithium reagent, then the reaction is directed to the ortho position and the introduction of lithium is regiospecific. The organolithium reagent cannot reach any further round the ring. We shall deal with two competing ortho positions and more exceptional cases later but for now ortho-lithiation means what it says.

Directing groups The Fries and Claisen rearrangements need an oxygen atom next to the site where the new bond is to be formed. From a strategy point of view, it’s good to start with your oxygen in place because it can be difficult to introduce.12 Ortho-lithiation strategy can make use of an oxygen substituent too, but it also works extremely well with other directing groups. The nature and diversity of these other directing groups gives ortho-lithiation very broad scope and makes it a powerful reaction. After all, not all aromatic target molecules have C–O bonds of any kind, let alone Ar–O bonds next to another substituted position.

98

7 The Ortho Strategy for Aromatic Compounds

OH OH

O Fries

O OH

O OH

O Claisen

OH

OMe

OMe

ortho-lithiation 1. BuLi, 2. E

E

As we now discuss these other directing groups, it would be a good time to consider the mechanism of the reaction more closely. We should then understand why some groups work better than others and be able to predict which ones will make a successful synthesis. There are two basic aspects to the function of any directing group. It may— i) coordinate to the lithium reagent ii) increase the acidity of its neighbouring hydrogen atoms. Both aspects may be operative, as they are in the case of anisole for instance, but they need not both be present in the director. A director which does not coordinate to lithium must acidify the positions next to it considerably if it is to be effective. We shall see that fluorine is such a substituent. Conversely, if the hydrogen atoms are not activated in any way by the director then it must compensate by being a very effective donor to lithium. The consequences of having two different directing groups, one which relies upon mostly coordination and the other which relies upon mostly increased acidity, are discussed in Regioselectivity II later.

Directing groups containing oxygen We have already seen that oxygen directs lithiation in anisole 1 very effectively. Here are two more examples of aromatic compounds 55 and 56 with an oxygen substituent that can be lithiated at the marked atoms. But oxygen (unaccompanied by any other heteroatom) is a very poor director of lithiation when it is any further away from the ring. Attempts, for instance, to direct a lithiation within methoxybenzyl ether 57, fail. Deprotonation at the benzylic position occurs instead and the molecule undergoes a Wittig rearrangement. Utilising a benzyl oxygen atom alone is bad strategy – if it is used it must be done in conjunction with another directing group. Me

O

OMe O

55

BuLi

OH Me

O 56

57

58; 26% yield

There are plenty of other groups which are known to be ortho-directors. The halogens can behave as ortho-directors and fluorine is perhaps the best (which is why it is in a box in the figure). However, if you were attempting to use either bromine or iodine as an ortho-director then you’d better watch out. It would be essential not to use butyl lithium as the lithiating reagent—they

7 Ortho-lithiation

99

would simply undergo a lithium – halogen exchange instead. Sulfones 64 and phosphates 65 are also known as ortho-directors but then anything which either acidifies protons at the ortho position or would bind to an alkyl lithium reagent placing it next to an ortho proton has . . . well, at least the potential to behave as an ortho-director.

Halogens, sulfones and phosphates O O F

CF3

Cl

Br

I

S

O

t-Bu

P(OR)2 O

59

60

61

62 [not with BuLi] 63

64

65

Directing groups containing nitrogen Oxazolines 66, amides 67 and carbamates 68 are the three kings of the ortho-director world. These are the most widely used ortho directing groups and among the best. Best means that they not only give high yields but that their action is reliable. We shall also see that they are usually easy to introduce and versatile in subsequent reactions. Be careful to note the difference between an amide and a carbamate – they look very similar when their structures are not drawn out in full (CONR2 vs. OCONR2). Although benzylic ethers often fail in ortho-lithiation reactions due to the Wittig rearrangement, benzylic amines 69 are fine. They do not, however, approach the generality of the ‘three kings’.

NR2

O N

O

O

NR2

NR2

O 66; oxazoline

67; amide

68; carbamate

69

Oxazoline 71 is an intermediate in the synthesis of the azirine 70 which was used in investigations into cycloaddition reactions.13 Let’s consider how we would make the oxazoline. Clearly we can disconnect the bond between the aromatic ring and the side chain to reveal 72, the lithium derivative of oxazoline 75, which can be coupled with an electrophilic side-chain.14 O

N

O N

R 70

N R

71

Li 72

+

R Br 73

The synthesis is straightforward. The oxazoline 66 directs the ortho-lithiation and then the lithium derivative is reacted with an allyl bromide. One thing we did not consider in the analysis was the synthesis of the oxazoline itself and perhaps we should. One way to make oxazolines (and there are many) is to react an acid chloride with an amino alcohol 74 followed by thionyl chloride.

100

7 The Ortho Strategy for Aromatic Compounds

O

O

HO 1. PhCOCl

N

N

2. 73

2. SOCl 2

H2N

1. BuLi

R 74

66

71

Oxazolines are versatile functional groups. It is entirely possible for one ortho-director to direct the introduction of a new group that is then an ortho-director in its own right. This leads to all sorts of interesting possibilities and is discussed in Several Lithiations.

Regioselectivity II—ortho versus ortho What happens when there is more than one directing group on the same ring? There are of course two fundamental possibilities – either they compete or they work together. There are three basic substitution patterns because the second directing group can either be ortho, meta or para to the first. X

X

X

Y

X and Y are orthodirecting groups

Y X and Y ortho

X and Y meta

Y X and Y para

There are two ‘ortho’ positions in the first case, three in the second, and four in the third. We notice, of course, that although all four sites in the para case can react, there are only two different sorts of proton due to the plane of symmetry. The para substitution pattern is the example we look at first. Oxazoline 75 has two para related ortho-directors and two possible sites for lithiation on the ring. We can clearly see from the result that the oxazoline ring was the more potent director because the product 76 has the new methyl group ortho to it rather than ortho to the methoxy group.

O

O N

N

1. BuLi 2. MeI

MeO

75

MeO

76

Me

In general, if one of the directing groups is much more potent than the other, we would expect that directing group to direct lithiation ortho to itself. And that is indeed what we see although the situation is, in fact, more complicated than this because in some cases we can choose which group we would like to lithiate next to. This issue is discussed later. Direct comparison between an amide and other ortho-directing groups in compounds such as 77 to 80 reveal that only a carbamate is a more powerful director than an amide while oxazoline, sulfonamide and benzylamine etc. are less powerful.11,15

7 Multiple Directed Lithiations

O

Et2N

O

NEt2

Et2N

S

O O

O 77

O NEt2

Et2N

N

O

O NEt2

NEt2

O

101

78

79

80

In the same way, direct comparison with other groups15 shows that the simple methoxy director is less powerful than a sulfonamide 81 or a benzylic amine 82, but more powerful than F, CF3, NR2, or (CH2)2NR2. OMe Et2N

OMe Et2N

S

OMe

OMe

F3C

F

O O 81

82

83

84

Multiple Directed Lithiations Let’s consider how we might make compound 85. For a start three of the four groups on the aromatic core are good ortho-directors. We have just seen that an amide is a more powerful directing group than a methoxy group so we can safely disconnect the methyl group 85a in the knowledge that the lithiation of 86 will go where we want it to. The synthesis is really very easy. The reaction of 86 with sec-BuLi and TMEDA followed by methyl iodide yields the compound we want.

MeO

NEt2

O

MeO

O

NEt2

O

O 85

C–C

O

MeI +

NEt2

O O

NEt2

NEt2 Me

MeO

Me

O 85a

NEt2 O 86

Further, a carbamate – the most potent ortho-director of the three – sits in between the other two. This can be used. We will come back and look at this compound again in the next section and disconnect it further but apply another concept to its analysis – that of multiple ortho-lithiation. For now we restrict ourselves to the one disconnection. The amide 87 contains two ortho-directing groups—the amide and the fluoride. The methyl group could be introduced by reacting a lithium derivative with methyl iodide. However, this is where we run into trouble because the lithium atom will not be introduced where we want it to be. Since both fluoride and the amide are ortho-directors and since they are meta substituted then they can cooperate with one another. This was illustrated with a reaction done with benzaldehyde.

102

7 The Ortho Strategy for Aromatic Compounds

O

Me

O

H NEt2

H NEt2

C–C

MeI +

F

1. s-BuLi TMEDA

NEt2

2. PhCHO

Ph

F 87

O

F 88

OH 89

We could, of course, use a protecting group to block the position between the two ortho-directors. There is now just one thing to consider – which of the two remaining ortho positions with be lithiated. The amide is a better director than the fluoride so there is no problem – the lithiation goes the way we want it to.16 The protecting group that we use is trimethylsilyl. The lithiation is then achieved using sec-BuLi and TMEDA and the lithium derivative reacted with MeI. Deprotection of the silyl group is done using a source of fluoride which is CsF in this case. H

O

H

Protect NEt2

O

Me NEt2

1. BuLi 2. Me 3SiCl

O

1. s-BuLi TMEDA

Deprotect NEt2

DMF

2. MeI

SiMe3 F

88

SiMe3

F 88

CsF

F 90

91

The fact that the two groups cooperate with one another needs to be emphasised. From our experience of chemistry we might expect a proton between two other groups to be more difficult to get at. This is not true with ortho-lithiation and with a meta substitution pattern you can expect cooperation. Quinolines such as 92 have been widely used in the synthesis of alkaloids. We might imagine that almost any position in 92 could be lithiated, but the position between the two MeO groups is preferred. When the electrophile is Me2C⫽CH.CH2Br the product is 94, and when it is R2N.CHO the aldehyde 93 is produced. Each of these compounds was developed into γ-fagarine, from 94 starting with ozonolysis,17 and from 93 by a Wittig reaction.18 OMe

OMe

OMe CHO

1. BuLi

or N

OMe

2. E

OMe

N OMe

92

OMe

N

OMe

OMe 93

94

Choice in ortho-lithiation In some cases we can choose which group we would like to lithiate next to. This choice arises from the two mechanisms that directing groups use and is one of the more subtle aspects of directed lithiation. Perhaps the best evidence for these different mechanisms is the change in regioselectivity with competing directing groups when the reaction conditions are changed.

Reactions of Fluoroanisoles Fluoroanisoles 95 and 97 each have two different sites that could be lithiated.19 In each case, the site next to oxygen can be lithiated if a coordination mechanism is operative and the site next to fluorine can be lithiated if an acid-base mechanism is operative. If ortho-fluoroanisole 97 is

7 Several lithiations

103

reacted with BuLi followed by solid CO2, the product is exclusively the carboxylic acid 98. However, if it is reacted with a mixture of BuLi and potassium tert-butoxide followed by solid CO2 then the product is – again exclusively – the regioisomer 96. CO2H

OMe OMe

1. BuLi, t-BuOK

2. CO 2

F

3. H

CO2H

F

3. H

96

95

OMe 1. BuLi

2. CO 2

F F

OMe

98

97

The situation is, however, even more subtle than this. Not only is it possible to exploit the difference between an acid-base mechanism and a coordination mechanism, it is also possible to choose between one coordinating director and another coordinating director.20 This is the case with 100. Both the nitrogen and oxygen atoms have the potential to coordinate to lithium. In a ‘superbasic’ mixture BuLi and KOBut, the lithiation proceeds next to the most electronegative atom – oxygen 99. The reagent is not bothered about coordination but just plucks off the most acidic proton – the one next to the most electronegative atom. When BuLi is used on its own, the lithiation proceeds next to the nitrogen atom 101. Nitrogen coordinates lithium better than oxygen and so in this case the reaction goes via a coordination mechanism.20 H N

MeO2C

Ot-Bu

2. CO 2

O

MeO

H N

1. t-BuLi, t-BuOK

3. HCl 4. CH 2N2

Ot-Bu

2. CO 2

O

MeO

99

H N

1. t-BuLi,

3. HCl 4. CH 2N2

Ot-Bu

O CO2Me

MeO

100

101

Several lithiations One powerful aspect of ortho-lithiation strategy is the ability to do more than one lithiation on the same ring. Exactly how this is done varies. One might imagine a situation where – i) The new group introduced by ortho-lithiation to give 103 can itself direct a further ortho-lithiation. ii) There could be two directing groups on the ring in the first place 105 and they may be able to direct two lithiations to make a dilithiated ring. And iii) the lithium derivative reacts intramolecularly with its own directing group to yield a product 108 with a different directing group and new functionality at the position of the original directing group (W). If the last example sounds a bit complicated, all will become clear with examples of the three types.15 Case (i): R1 and R2 are directing groups OR1

OR1 R2

1. BuLi 2. R2X

Case (i): R is a directing group OR1

2. E

102

103

OR

R2

1. BuLi

E

RO

104

2. E

Case (iii):

and

E 106

W

W R2

BuLi

R2

OR

RO

105 OR1

R1

E

1. BuLi

R2

1. BuLi

are directing groups 2. E

107

108

E 109

104

7 The Ortho Strategy for Aromatic Compounds

Case (i) One ortho-director leads to another The carbamate 110 reacts with sec-butyl lithium followed by Et2NCOCl to yield the new amide 86. This reacts with sec-butyl lithium and methyl iodide to give the tetra-substituted aromatic compound21 85. OMe OMe

OMe OCONEt2

O

NEt2

1. s-BuLi

OCONEt2 1. s-BuLi

NEt2

2. ClCONEt 2

NEt2

2. MeI

O O 110

Me

86

O 85

The tetrasubstituted benzene 112 was synthesised by multiple ortho-lithiations. Three of the four groups are ortho-directors themselves. The question of which group to disconnect first is not simplified in this case by the knowledge that the group must be introduced by an adjacent orthodirector because all the groups are adjacent to at least one other. Let’s consider disconnection a. The forwards reaction from 111 would not lead to a problem of regioselectivity because the two ortho-directors direct lithiation to this site. The problem is which ortho-directing do we disconnect next? We do not have two adjacent directing groups so we cannot use one to introduce the other. This forces us to adopt an alternative strategy sometimes referred to as a ‘walk around the ring’. SiMe3

SiMe3 OCONEt2

SiMe3 OCONEt2

a

b

a CONEt2

a

OCONEt2

a CONEt2

b CONEt 2 112

CONEt2 111

SiMe3 OCONEt2

113

114

We start b at one end of the three ortho-directors and disconnect them successively. After two disconnections we are left with the carbamate 114. The silyl group will be introduced by ortho-lithiation too. The introduction of the silyl group stops the sequence of ortho-lithiations proceeding in an anti-clockwise manner and hence prevents any competition that might otherwise arise after the between the carbamate and the amide in subsequent ortho-lithiations.

c

SiMe3

SiMe3 OCONEt2

OCONEt2

c

OCONEt2 CONEt2

114c

115

CONEt2 112; no competition

OCONEt2 CONEt2 CONEt2 competition

The three lithiation reactions use sec-BuLi and TMEDA. Trimethylsilyl chloride is the first electrophile and diethyl carbamoyl chloride is used twice to give our target material 112.

7 Several lithiations

105

SiMe3 SiMe3

1. s-BuLi TMEDA

OCONEt2

1. s-BuLi, TMEDA

OCONEt2

2. ClCONEt 2

OCONEt2

2. Me 3SiCl

3. s-BuLi, TMEDA

CONEt2

4. ClCONEt 2

115

CONEt2 112

114

ii) Dilithiations True dilithiated species generated by ortho-lithiation in which there are two lithium atoms on the same ring are very rare (with dilithiated species formed by halogen/lithium exchange being more common). It is not a surprise that in those examples where dilithiated species do exist22 the lithium atoms are usually para to one another 117. Et2N

O

O

O

O

Et2N

1. 2 x s-BuLi TMEDA

NEt2

SiMe3 O

O O Me3Si

2. Me 3SiCl

O

NEt2

117; 67% yield

116

Putting two lithium atoms closer together is more difficult and while the 2,6-dilithium derivative 118 can be made, it cannot be made by double ortho-lithiation of 67; R ⫽ Et. Instead it must be made from dibromide 119 and tert-BuLi. CONEt2 H

H

CONEt2 Li

BuLi

67; R = Et

Li

CONEt2 Br

t-BuLi

Br

118

119

iii) One ortho-director rearranges into another – the anionic Fries reaction We have come across the Fries rearrangement previously in this chapter before lithium had even made its entrance. A similar rearrangement can occur in the realm of ortho-lithiated species and is sometimes call the anionic Fries rearrangement. Compare the two. O The Fries Rearrangement

R

Lewis acid

OH

OH +

O

R

R

O

O OH

O The Anionic Fries Rearrangement

NMe2 O

120

s-BuLi

O

TMEDA

Li

NMe2 NMe2

O

121

O 122

106

7 The Ortho Strategy for Aromatic Compounds

The carbamate 120 is lithiated with sec–butyl lithium to form the intermediate 121 that reacts further to form the amide 122. From the ortho-lithiation point of view we now have a different ortho-director in the molecule and, if we protect the phenol group, we can continue lithiation around the ring. This chemistry was used in the synthesis of the diene 124 needed for the antitumour compound pancratistatin23 123. Disconnection of 124 across the double bond closest to the ring reveals the aldehyde 125. This aldehyde can be introduced by the action of an ortho–lithium species (as the dialkylamide in 126 is a good ortho-director) on a formylating reagent. OH

O OH

O H

NH

O

NEt2

OH

O

O

H HO 123 pancratistatin

O

OH 124

OH OH

O

OH

O

NEt2

O

O

O

CHO

NEt2

O

125

126

The amide is a good ortho–director but so is the acetal function. The regioselectivity (actually there is the issue of chemoselectivity here too) of the lithiation reaction will depend upon which of the two is the more potent. Amides were one of our ‘three kings’ of the ortho-directors and are more powerful than acetals. Now it is time for an anionic Fries rearrangement (in the retrosynthetic direction) to give the starting material 127. OH

O OH

O O activated by acetal

NEt2 activated by amide

O anionic Fries

O

NEt2

O

126

OCONEt2 O O

126a

127

We ortho-lithiate 127 using sec-BuLi and TMEDA. The lithiation goes where we want it – ortho to the carbamate rather than ortho to the acetal. The lithiated intermediate is then allowed to warm up – we will come back to this in a moment – to yield phenol 126. This is reacted with TBDMSCl and imidazole to give the protected phenol 128. Once again, the acetal loses out and the amide does the directing during the second lithiation. Formylation is achieved using DMF. OTBDMS t-BuMe2SiCl (TBDMSCl)

1. s-BuLi TMEDA

127

2. warm to room temperature

126

CONEt2

O

OTBDMS 1. s-BuLi TMEDA

CONEt2

O

2. DMF

imidazole

O

O 128

CHO 129

7 Halogens

107

The aldehyde 129 is attacked using allyl Grignard. The alcohol 130 is converted to the mesylate and then an elimination reaction using DBU as the base gives diene 131. The protecting silyl group can easily be removed to give 124 but in fact it was retained for protection during the rest of the synthesis. OTBDMS

OTBDMS

CONEt2

O

CONEt2

O 1. MsCl, Et 3N

BrMg

129

O

O

2. DBU

OH 130

131

It is important to recognise the element of choice that exists in the anionic Fries rearrangement. After the initial lithiation of 127 we can either let it warm up and do the rearrangement or, if we are careful, we can keep it cold so that the rearrangement does not take place and we can introduce another electrophile to react with the intermediate lithiated compound instead.11

Halogens Fluorine as an ortho-director Fluorine is a good ortho-director.24 It directs the lithiation of fluorobiphenyl without the assistance of another directing group and does so very efficiently. F

F 1. s-BuLi, TMEDA

CO2H

2. CO 2

133; 79% yield

132

Fluorine as a leaving group Fluorine can also function as a leaving group from aromatic rings.25 The C!F bond is very strong and it would be misleading to describe fluorine as a good leaving group (even though many of the reactions in which it leaves are good). Nucleophiles attack because the ipso carbon is made electrophilic by the strongly electronegative fluorine. In compound 134 there are four fluorine atoms on the benzene ring: one is displaced by nucleophiles to yield the quinolone antibiotic25 135. F

F

O

F

CO2H

1.

NH

O

F

CO2H

N Me

F

N F

N

2. NaOMe, MeOH

Me 134

N

N F 135

The electron-withdrawing group on the aromatic ring is essential. Since direct SN2 displacement reactions on sp2 centres do not occur, the electron-withdrawing group gives the electrons somewhere

108

7 The Ortho Strategy for Aromatic Compounds

to go in the intermediate before expulsion of fluoride. The same process occurs when a nucleophile reacts with an acid chloride. F

F

O CO2H

F

F

O CO2H

F

O CO2H

F

F F

N

ratedetermiming step

F

N

Nu

Nu

fast

N

F

F

Nu 136

137

138

Fluoride as ortho-director and leaving group A method developed by Bridges24 enables the synthesis of the benzothiophene 139 and other similar derivatives using a combination of fluorine as both an ortho-director and a leaving group. Disconnection of the acetal will not get us very far but the double bonds of the thiophene ring also forms part of an α,β-unsaturated ester. A simple aldol disconnection here gives us aldehyde 140. The next disconnection relies on our knowledge that fluorine can behave as a leaving group from aromatic rings. The aldehyde is the necessary electron withdrawing group. We now 141 have three substituents round the ring. The acetal, although a poor ortho-director in its own right, will assist the fluorine in a lithiation. We can remove the formyl group with an ortho-lithiation in mind.

O

O

O

O

O CHO

aldol

O

O

nucleophilic aromatic substitution

O

CHO

CO2Me S

S

139

140

CO2Me

F

F

141

142

The acetal 142 is lithiated with sec-BuLi and TMEDA and formylated with DMF to give, after workup, the aldehyde 141. The next couple of reactions happen in one pot under the same set of reaction conditions. The thiol is reacted with NaH and the anion added to a solution of the aldehyde 141. After workup the product is the target material.24

O

O

O 1. s-BuLi TMEDA

O

O HS

CHO

2. DMF

NaH

3. H

F 142

O

CO2Me

F 141; 91% yield

CO2Me S 139; 56% yield

The next three sections – Benzyne Formation, α-Lithiations and Lateral Lithiations – are just a brief look at some of the other lithiation reactions that are available to the synthetic organic chemist. They are also words of warning. They show what other reactions can take place in the frame of lithiation of aromatic substrates and are intended to prime you for potential traps.

7 α-Lithiation

109

Benzyne Formation—A different aromatic strategy Benzynes are readily formed when we have a leaving group ortho to a lithium atom. Thus when ortho-bromofluorobenzene 143 is reacted with butyl lithium, the intermediate formed is ortholithiofluorobenzene 144. Lithium fluoride can eliminate from this to yield benzyne 145. Benzynes are reactive species and do not hang about. Here, ortho-bromofluorobenzene is treated with butyl lithium in the presence of furan. When the benzyne forms it rapidly reacts with furan in a cycloaddition reaction to form the adduct 146 in a 68% yield. Br

Li

BuLi, Et 2O

O

O

–30 to –25 °C

F

F

143

145; benzyne

144

146

Note the ortho relationship between the lithium and fluorine atoms on the benzene ring. We have previously seen lithium next to the fluorine without benzyne formation and yet here the very same species forms benzyne! The crucial difference between the two cases is temperature. The ortho-fluoro-aryllithium is stable at ⫺78 ⬚C and the benzyne does not form until the temperature reaches ⫺40 ⬚C to ⫺50 ⬚C. Aromatic rings with lithium adjacent to other halogens (such as bromine) are less stable and form benzynes at lower temperatures. An ortho relationship between a triflate and a silyl group is another way to generate benzynes.26 The benzyne 149 is formed when the silyl group is removed11,27 using the fluoride ion 148. OCONEt2

OCONEt2

OCONEt2

SiMe3

F

SiMe3

1. Tf 2O

OCONEt2

2. CsF

O

furan

OH

OTf

147

148

149

150

α-Lithiation An α-lithiation reaction is characterised by a lithiation occurring at an sp2 centre and α to a heteroatom28 152. Lithiation of furan with BuLi and reaction with benzaldehyde gives alcohol 154 in 98% yield. H

BuLi

Li

X

X

151

PhCHO

BuLi

152

O

–20 °C to room temperature

O

Li

O OH 154; 98% yield

153

furan

Ph

–20 °C

We might expect that the sulfonamide group in 155 would direct lithiation of the thiophene ring to its adjacent site in an ‘ortho’-lithiation. It doesn’t. An α-lithiation to give 156 occurs instead.

S

S O

155

NEt2 O

2. CO 2

1. BuLi 0 °C

Li

S

S

NEt2

O O 156

–78 °C

HO2C

S

S

NEt2

O O 157; 82% yield

110

7 The Ortho Strategy for Aromatic Compounds

One last word of warning with regard to α-lithiations. The benzofuran 158 has a methoxy group on the benzene ring and we might expect this to direct an ortho-lithiation. It would be easy to overlook the α-lithiation of the furan ring: this, in fact, happens instead. MeO

2.

MeO

O

MeO

1. BuLi

Li O

O

O

OH

157; 62% yield

156

155

Lateral Lithiation A lateral lithiation occurs at a benzylic position29 (161 reacting to 162) rather than on the benzene ring itself. These reactions are quite common so we must discuss them briefly here as the same functional groups that behave as ortho-directors are also lateral lithiation directors as in the case of the amide 161 below. The lithiated species can be represented as a lithium enolate 163 or even with chelation 164. This is not possible with an ortho-lithiation where the ‘charge’ is localised in a C-Li σ-bond. O

O LDA

NEt2

OLi

NEt2

NEt2

NEt2

O

Li

Me 161

Li

162

163

164

Chemoselectivity Lateral lithiations do not necessarily need an adjacent (ortho) director. So LDA will happily deprotonate the para methyl group of the amide 166 in THF at 0 ⬚C. Once again the ‘charge’ can delocalise into the electron withdrawing group. But reaction with a much stronger base at low temperature, secBuLi at ⫺78 ⬚C, gives the ortho-lithiation 165 we have come to expect. The lateral lithiation gives the thermodynamically more stable lithiated product whereas the ortho-product 165 is the kinetic product as complexation of the alkyl lithium reagent with the amide accelerates the lithiation.30 O

O NEt2

Me

Li

NEt2

s-BuLi TMEDA, –78 °C

165

O NEt2

LDA THF, 0 °C

Me

Li

H

H

166

167

Here is an example 168 of a reaction in which both benzyne formation and lateral lithiation are involved. What’s going on here? Try to work out the mechanism before reading the explanation. You have the clues you need – the reaction involves the formation of a benzyne and a lateral lithiation.31 OMe

OMe O

O O

MeO

2. 171 3. [O]

OMe

OMe Br

Me

1. 3 x LiTMP

Me

N OMe

MeO

Li

OMe

O 168

169; 66% yield

170 LiTMP

171

7 Summary of Reagents

111

The lactone 168 is lithiated at the benzylic site by lithium tetramethyl piperidide (LiTMP) 167 to give 172. Notice that this ‘charge’ can delocalise into the carbonyl group and that ortholithiation between the two methoxy groups does not occur. One of the reactive species then is 172. The aromatic bromide 171 is lithiated ortho to the bromide 174, eliminates lithium bromide, and generates a benzyne 173. These two species then combine. OMe

O

168

OMe

OMe

LiTMP

Me

Br

Me

OMe

Li

OMe

LiTMP

171

O MeO Li 172

173

174

The benzyne 173 is attacked regioselectively. By attacking meta to the methoxy group, the lithium atom ends up ortho to the methoxy group (175). The C-Li bond can then attack the lactone and spring it open to give 176. We are now very close, and a little oxidation will take 176 to the anthraquinone31 169. MeO

O Li O

172 + 173

OMe O

OMe

MeO

Me

Me

[O]

169 OMe

175

OMe

OMe

MeO OH 176

Summary In this chapter we have introduced many powerful reactions. We have seen that— • Ortho-lithiation turns aromatic compounds into much more powerful nucleophiles. • The range of groups that direct lithiation is extensive though some are better than others. • More than one ortho-director can lead to competition or cooperation. In some cases there is a choice over the ortho-lithiation site. • We have also seen that ortho-lithiation can be used in conjunction with other reactions – such as fluoride displacement or an anionic Fries rearrangement – that make ortho-lithiation a particularly versatile strategy.

Summary of Reagents BuLi

sec-BuLi LDA

BuLi is often sufficiently reactive to bring about an ortho-lithiation on its own. It coordinates to electronegative groups before it deprotonates in the ortho position. With tert-BuOK it forms a ‘super’ base. However, halogens will often undergo halogen/lithium exchange with BuLi. A nitrogen base (like LDA) is more appropriate in such cases. sec-BuLi is a stronger base than BuLi. It is often used in conjunction with TMEDA to ortho-lithiate. LDA prefers to deprotonate rather than react with a halogen. When an aromatic bromide reacts with BuLi a by-product is BuBr (a carbon – bromine bond). If LDA

112

7 The Ortho Strategy for Aromatic Compounds

were to lithiate an aromatic bromide the by-product would contain a week nitrogen – bromine bond and so the reaction does not occur. LTMP Lithium tetramethylpiperidide is a stronger, more hindered, and more expensive version of LDA. • Notice that ortho-lithiations frequently use BuLi or sec-BuLi but that the lateral lithiations frequently use lithium amide bases like LDA and LiTMP instead. This is not insignificant.

References General references are given on page 893 1. P. H. Gore in Friedel-Crafts and Related Reactions, Interscience, New York, 1964, Vol III, Part I, page 1. 2. F. Arndt and G. Källner, Chem. Ber., 1924, 57, 202; J. D. Loudon and R. K. Razdan, J. Chem. Soc., 1954, 4299. 3. A. H. Blatt, Chem. Rev., 1940, 27, 413; Org. React., 1942, 1, 342. 4. K. W. Rosenmund and W. Schnurr, Liebig’s Ann., 1928, 460, 56. 5. R. Baltzly, W. S. Ide and A. P. Phillips, J. Am. Chem. Soc., 1955, 77, 2523. 6. P. R. Vijayasarathy, R. B. Mane and G. S. Krishna Rao, J. Chem. Soc., Perkin Trans. 1, 1977, 34. 7. T. Y. R. Tsai, K. P. Nambiar, D. Krikorjan, M. Botta, R. Martini-Bettolo and K. Wiesner, Can. J. Chem., 1979, 57, 2124; K. Wiesner, Pure Appl. Chem., 1979, 51, 689. 8. R. E. Dean, A. Midgley, E. N. White and D. McNiel, J. Chem. Soc., 1961, 2773. 9. S. J. Rhoads and N. R. Raulins, Org. React., 1975, 22, 1. 10. A. T. Shulgin and A. W. Baker, J. Org. Chem., 1963, 28, 2468. 11. V. Snieckus, Chem. Rev., 1990, 90, 879. 12. Disconnection Approach, p.21; but see Chapter 33 below. 13. A. Padwa and H. Ku, J. Am. Chem. Soc., 1978, 100, 2181. 14. H. Reuman and A. I. Meyers, Tetrahedron, 1985, 41, 837: see page 853 in particular. 15. Clayden, Lithium, pages 28–72 and 90–96. 16. R. J. Mills, N. J. Taylor and V. Snieckus, J. Org. Chem., 1989, 54, 4372, 17. N. S. Narasimhan and R. S. Mali, Tetrahedron, 1974, 30, 4153. 18. J. F. Collins, G. A. Gray, M. F. Grundon, D. M. Harrison and C. G. Spyropoulos, J. Chem. Soc., Perkin Trans. 1, 1973, 94. 19. G. Katsoulos, S. Takagishi and M. Schlosser, Synlett, 1991, 731; M. Schlosser, F. Faigl, L. Franzini, H. Geneste, G. Katsoulos and G-F. Zhong, Pure Appl. Chem., 1994, 66, 1439. 20. R. Maggi and M. Schlosser, J. Org. Chem., 1996, 61, 5430. 21. P. Beak, A. Tse, J. Hawkins, C.-W. Chen and S. Mills, Tetrahedron, 1983, 39, 1983; Ref 11, page 895, table 12, entry 33. 22. R. J. Mills, R. F. Horvath, M. P. Sibi and V. Snieckus, Tetrahedron Lett., 1985, 24, 1145. 23. S. Danishefsky and J. Y. Lee, J. Am. Chem. Soc., 1989, 111, 4829. 24. A. J. Bridges, A. Lee, E. C. Maduakor and C. E. Schwartz, Tetrahedron Lett., 1992, 33, 7495; 7499. 25. T. Miyamoto, J. Matsumoto, K. Chiba, H. Egawa, K. Shibamori, A. Minamida, Y. Nishimura, H. Okada, M. Kataoka, M. Fujita, T. Hirose and J. Nakano, J. Med. Chem., 1990, 33, 1645. 26. Y. Himeshima, T. Sonoda and H. Kobayashi, Chem. Lett., 1983, 1211. 27. K. Shankaran and V. Snieckus, Tetrahedron Lett., 1984, 25, 2827. 28. H. W. Gschwend and H. R. Rodriguez, Org. React., 1979, 26, 1. 29. R. D. Clark and A. Jahangir, Org. React., 1995, 47, 1; Clayden, Lithium, pp 73–83. 30. P. Beak and R. A. Brown, J. Org. Chem., 1982, 47, 34. 31. C. A. Townsend, S. G. Davis, S. B. Christensen, J. C. Link and C. P. Lewis, J. Am. Chem. Soc., 1981, 103, 6885.

8 σ-Complexes of Metals Introduction The structure of organo-lithium compounds Lithium and Magnesium Complexes Making organo-metallic σ-complexes by oxidative insertion Structures of organo-lithium and magnesium reagents Reactions of organo-lithium and organo-magnesium reagents Chiral Grignard reagents Acylation at carbon with Grignard and organo-lithium reagents Transition Metal Complexes Acylation with acid chlorides Hydrometallation Carbonylation of organometallic complexes Palladium σ-complexes Recent results with Cu(I) and Pd σ-complexes

Introduction The structure of organo-lithium compounds Every reader of this book is familiar with Grignard reagents, the first useful organometallic compounds, and with lithium derivatives of organic compounds as we have already discussed these at some length in the first seven chapters. We must now widen our discussion to include all compounds with a direct metal–carbon bond and we shall call these σ-complexes in contrast to the more common η and π complexes that we shall meet in later chapters. A saturated carbon atom has no lone pair electrons or accessible empty orbitals and we can treat the metal–carbon σ-bond like the boron-carbon or silicon-carbon bond without discussing back-bonding or invoking d-orbitals. We should still recognise that the apparently simple compound MeLi actually exists as trimers and tetramers1 such as 1. In this chapter we shall be concerned with the synthesis of σ-complexes of metals such as Li, Mg, Cu, and Zr. We shall see how far simple-minded organic mechanistic ideas can be extended to these undoubtedly more complex structures, and that we need not in general worry about their polymeric nature as they often react in solution as monomers2 1b. Li Li

Li Li

1a Li atoms in MeLi

add Me group to centre of each triangle

Me Me Li

Li

Me Li

Li Me 1; MeLi

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

OR2 Me

Li

OR2 OR2

1b; MeLi in solution

8 σ-Complexes of Metals

114

We have just been considering the formation of aryl-lithiums by deprotonation with the more basic alkyl-lithiums, normally BuLi, and you are also aware that protons may be removed from many molecules to give lithium derivatives. We shall not discuss this further in this chapter, but concentrate instead on three other approaches, oxidative addition (insertion), transmetallation, and hydrometallation.

Making organo-metallic σ-complexes by oxidative insertion You are familiar with the normal method of making a Grignard reagent: an alkyl halide 2 reacts with magnesium metal in ether solution to give something we normally write as RMgBr, but which certainly contains a tetrahedral Mg atom, the remaining sites being occupied by ether molecules 4a. This is at least one of the species present in such solutions. In making these new bonds, magnesium has accepted electrons from the ether ligands. In a simple organic structure formed like this we should put charges on the atoms 4c, or draw the bonds as arrows 4b. Structure 4c is most unattractive for a metal complex as the metal ends up with a double negative charge. The best compromise is just to draw all bonds as lines 4d and leave out the charges. R Mg(0)

R

Br

R

R

Br

2

3

OEt2

Et2O

4a

Br

R

Mg

Mg Et2O

R

Br Mg

Mg

Br

OEt2

Et2O

OEt2

Et2O

4c

4b

Br Mg OEt2 4d

We shall call this reaction oxidative insertion (or oxidative addition) because the metal atom inserts itself between the carbon and halogen atoms and simultaneously undergoes oxidation from Mg(0) to Mg(II). If you are unhappy at assigning an oxidation state to a metal in a complex, then the quickest way is to remove the ligands from the metal. First remove all the real molecules (two of Et2O in this case) as you can obviously do this without changing the oxidation state of the metal. Then remove the remaining ligands as anions 3a (two in this case), leaving the same number of positive charges on the electropositive metal. The number of charges gives the oxidation state of the metal. We therefore describe 4 as a sigma complex of a carbanion with Mg(II). R

Br Mg

Et2O

OEt2 4d

remove molecular ligands

R

Br Mg

remove remaining ligands

R

Mg

Br

as anions

[2 x Et2O]

3

3a

In detail, the oxidative insertion probably begins with an interaction between the electron-rich halogen atom and the electron-poor metal 5. Back-bonding from the metal also builds up in the intermediate 6, weakening the carbon-halogen bond. The electrons of the C-Br bond then abandon the halogen atom and form a bond to the metal 6c. The carbon atom has undergone electrophilic substitution, while the metal has made two new bonds, using an empty orbital for one and a lone pair for the other. You may object that if the metal has been oxidised, then something must have – been reduced, and this is the carbon atom which is like R⫹ in RBr and is now like R in the complex. Words like might, probably, and could are used a lot in this area and question marks litter both the structures of the intermediates and the reaction pathways. We should still take pride in attempting to understand these reactions even if the details of the mechanisms are uncertain.

8 Introduction The structure of organo-lithium compounds

R

Br

Mg

5

R

Br

R

?

Mg 6a?

(–) Br

(+) Mg

R

Br Mg 7?

Mg

R

Br Mg 6c?

6b? R

115

Br

3

Magnesium probably undergoes oxidative insertion more easily than the other metals we use, but lithium, used as wire, hammered-out sheets, or even slices3 inserts quite easily into alkyl or aryl halides, particularly in THF solution.4 A one-electron donor like lithium probably follows a similar pathway to Mg but with single electron arrows. Palladium forms aryl Pd(II) derivatives in the same way,5 but the Pd must be added as a soluble complex like Pd(Ph3P) 4. Other transition metals are much less successful.

Structures of organo-lithium and magnesium reagents We may describe these species as σ-complexes of carbanions with metals but they are not carbanions. As normally prepared from alkyl halides and Mg(0) or Li(0) in ether solvents (Et2O or THF), they are soluble in organic solvents as tetrahedral complexes and are not salts. Chemists are happy to draw the monomeric species RLi or RMgX and it is likely these are the reactive entities in many reactions.

Reactions of organo-lithium and organo-magnesium reagents Grignard reagents and alkyl-lithiums, organometallic reagents of very electropositive metals with a strong affinity for oxygen, are useful reagents, as you are aware. They are very reactive, basic nucleophiles, combining well with hard electrophiles such as carbonyl compounds, but less successfully with softer electrophiles such as alkyl halides. A brief exploration of these reactions will make it clear why other organometallic σ-complexes are needed. When Cram6 was studying the stereochemistry of addition to chiral aldehydes, he compared the behaviour of i-propyl Grignard (i-PrMgBr) with i-PrLi in nucleophilic attack on the aldehyde 8. He was interested in the diastereomer ratio in the product 9, but we can note the almost identical yields from the two reagents. OH CHO Br Ph 8

1. Li ribbon pentane

1. Mg, Et 2O 2. 8

84% yield

2. 8

Ph 9

Br

88% yield

Additions of either RLi and RMgBr to most aldehydes and ketones give good yields, and there are many examples known.7,8 The only problems likely to arise are enolisation by RLi or reduction by RMgX. The latter is more serious, especially with branched Grignard reagents 10. It appears erratically8 - the very unreactive ketone Ph2CO gives9 mostly addition 13 with t-BuMgCl, but only reduction 11 with iso-butyl Grignard 12, quite the reverse of what we might expect.

8 σ-Complexes of Metals

116

H

RCHO

BrMg

Br

Mg

O

H R

Br

Mg

H

H

O

H

R

H +

H2O

HO

R

10 OH Ph

O

11; 91% yield

Ph

OH

Ph

+

+

ClMg

Ph

BrMg

Ph

Ph

13; 63% yield

12

Reaction with epoxides are generally excellent with RLi, but deprotonation leading to rearrangement can be a problem.10 Grignard reagents react well with ethylene oxide itself, but with other epoxides rearrangement can occur because of the Lewis acid effects of Mg(II). Cyclohexene oxide 14 reacts as expected with RLi to give 15, but can give the surprising product 16 with Grignard reagents, by rearrangement to the aldehyde 17 before addition. A famous example is Robinson’s reported synthesis of 15; R ⫽ PhCH2CH2- in 92% yield, later corrected to 16; R ⫽ PhCH2CH2-, and hence useless for the synthesis of steroids.11

OH

OH

RLi

RMgBr

CHO

O R

R 14

16

15

17

RMgBr

O

O

MgR

or MgBr2

O 18

14

MgBr

19

Reaction with alkyl halides is generally thwarted by the rapid exchange of Mg or Li between the two species. This can be used to advantage if the new organo-lithium reagent is more stable than the old, as in the synthesis of aryl-lithiums from aryl halides and BuLi. Direct lithiation by replacement of hydrogen generally occurs in positions ortho to heteroatoms (chapter 7) but a halogen, as in the bromo-pyrimidine 20 directs the lithium atom to that same site giving 21. Addition to the unreactive dichloro-diarylketone 22 gives the fungicide fenarimol12 23. N O N BuLi

N

Br 20

HO

N

N

21

N

Li

Cl 21

Cl 22

Cl

Cl 23

Many alkynyl, alkenyl, aryl, and functionalised alkyl-lithiums behave well in alkylation reactions because they are more stable anions than those derived from the simple unfunctionalised alkyl halides with which they are to react. Examples are 24, 25, and 26 and these will be treated in later chapters with the reactions of Grignard reagents with allyl halides.

8 Introduction The structure of organo-lithium compounds

117

S S

O

Li 24

Li

Li

N

25

26

Chiral Grignard reagents It has long been supposed that enantiomerically pure Grignard reagents cannot be made. Certainly13 ‘enantiomerically pure Grignard reagents…are not accessible from enantiomerically pure secondary alkyl halides by reaction with magnesium metal since electron transfer processes and the intervention of radicals annihilates the stereochemical information.’ Hoffmann’s recent success in this difficult area13,14 also illustrates the reactions of a simple Grignard reagent, EtMgBr. O Ar

O

NCS

S

Ph

EtMgCl

S

Ar

Ph Cl

27; 99% ee Ar = p-chlorophenyl

Ar

28; 97% ee

ClMg

O

Ph

+

S

EtMgCl

ClMg

Ph

THF –78 °C

Cl

29; 99% yield 96% ee

THF –78 °C

30; stable at –78 °C

31; stable at –78 °C

The enantiomerically pure sulfoxide 27 is chlorinated with NCS (N-chlorosuccinimide) to give the enantiomerically pure 28 easily separated from its diastereoisomer. Reaction with EtMgCl occurs in two stages. Nucleophilic attack at the sulfoxide with inversion to give 29 expels a single enantiomer of the chloro-Grignard 30. Still at low temperature an excess of EtMgCl displaces chloride with inversion from 30 to give the simple secondary Grignard reagent 31. This compound maintains its stereochemical integrity at ⫺78 ⬚C. If the Grignard reagent 31 is prepared and immediately reacted with an electrophile compatible with the conditions a product 33 can be isolated in good yield and high ee showing that 31 reacts with retention. Ph

5 x EtMgCl

27 THF –78 to –30 °C

MgCl 31; not isolated stable at –78 °C

1. N3 2. Ac 2O

Ph

SPh

PhS

N

N

NAc

32; not isolated

Ph

KOH

NHAc 33; 82% yield from 27, 92% ee

Acylation at carbon with Grignard and organo-lithium reagents Acylation at carbon is always a difficult problem because the product (usually an aldehyde or ketone) may be more reactive than the acylating agents in two ways: more electrophilic at the carbonyl group and more acidic because of enolisation. Nitriles often perform well in the acylation of RLi or RMgBr, and the reaction of lithium salts of carboxylic acids with RLi is a strategy already discussed in chapter 2. Both these methods rely on the product being released during the work-up after all RLi is quenched. One example is the cyclopropyl ketone 34 needed by

8 σ-Complexes of Metals

118

Masamune as an intermediate in his amphotericin synthesis.15 Disconnection to the hydroxyacid 36 was particularly attractive as optically active 36 was available. Cyclopropyl-lithium 35; M⫽Li is actually easier to make than other s-alkyl-lithiums as the three-membered ring stabilises the anion in the σ-complex.16 Acylation with 36 needed three molecules of 35; R ⫽ Li as two are consumed by the acidic OHs, but the reaction is an efficient acylation at carbon. OH

OH C–C

M

+

HO

O

Li

34 Br

O 34

36

35

THF

35; M=Li

37

36

Li

More commonly, acylating agents are based on tertiary amides because they are much less reactive than ketones, or on lactones because they are about as reactive than ketones. This paradox arises because tertiary amides RCONMe2 add alkyl-lithiums but the tetrahedral intermediate does not decompose while RLi is present. Hence 26 gives good yields of the ketones17 39. BuLi

N

O

1. RCONMe 2

Li

N

38

2. H +, H 2O

N

R 39

26

The “Weinreb amides” 40 are particularly efficient in acylating reactive carbon nucleophiles.18 An alkyl-lithium adds to form the chelated lithium derivative 41 which decomposes on work-up to give the ketone 42. Compounds 60 and 63 in chapter 5 are examples of Weinreb amides. O R1

O N Me 40

O

Li

Me

R1

O O Me

R2Li

N R2

H H2O

R2

R1

Me 41

42

Lactones have similar reactivity to ketones and at low temperature one molecule of an alkyllithium may give the ketone in good yield. In his synthesis of the pine saw fly sex attractant 43, Magnusson19 chose to replace the more-or-less central methyl group by a double bond 44 so that disconnection to a ketone 45 was possible. OAc

43

Reconnection to the lactone 46 then provides a means of controlling the stereochemistry using the Baeyer-Villiger rearrangement of the cyclohexanone 47. Acylation of the lactone 46 with n-octyl-lithium at low temperatures gives a 70% yield of the hydroxyketone 48, which in turn gives the pheromone 43.

8 Transition Metal Complexes

119

OAc FGA

OAc Wittig

R

43

R

acylation of RLi

O 44; R = n-octyl

45; R = n-octyl

O O

O

BaeyerVilliger

mCPBA

n-octyl lithium

46 69% yield

47 CH2Cl2

Et2O, –78 °C

Me

Me 46

47 OH

OH

2 equivs Ph3P=CH2

R

1. H2, Pd/C 100% yield

R

43 2. Ac2O pyridine

O 48; 70% yield

49; 66% yield

Transition Metal Complexes Acylation with acid chlorides These methods of acylation are devices to solve the problem that acid chlorides and organolithium or magnesium derivatives are not good partners. To get good acylation with acid chlorides, as well as good alkylation with alkyl halides, we need to turn to other metals which provide softer carbanion complexes, less basic compounds with metals showing a lower affinity for oxygen. The most important is copper. The difficulties of direct oxidative insertion with metals other than Mg or Li mean that σ-complexes are often made from organo-lithium or Grignard reagents by metal exchange. This reaction amounts to a nucleophilic substitution at the metal without a change of oxidation state so the metal is used in whatever oxidation state is finally needed. Attack of methyl lithium on a Cu(I) halide gives methyl copper 50, a σ-complex of Me⫺ and Cu(I). If an excess of MeLi is present an “ate” complex is formed, lithium dimethylcuprate 51. This is formally a compound of a copper anion 51a, just as BF4⫺ is a borate. The term “ate complex” refers to such formally anionic complex in which the metal has one extra anionic ligand. Its true structure is dimeric 51b and it exists as an equilibrium with 52 in solution.20

Cu(I)Br

MeLi

MeLi

LiBr +

MeCu

Me Me

50

Me

Me

Cu Li

Cu Me

51

Li

Li Me

51a

Cu

Me Li

Cu Me 51b

Me3Cu2Li 52

Cuprates couple efficiently with alkyl halides, as in the simple synthesis of n-undecane21 via the cuprate prepared from BuLi. When a Grignard reagent is involved, a Cu(I) compound is usually added in catalytic amounts for coupling with an alkyl halide. n-C7H15-I

Cu(I)I

BuLi THF

Bu2CuLi

n-C11H24 = n-undecane

8 σ-Complexes of Metals

120

The acetoxy-aldehyde 53 has a 1,7-relationship between the functional groups and there is no well established approach for this relationship. Disconnection to the alkyl halide 55 is appealing as it can be made from THF in one step. It might be coupled to the Grignard reagent 54 provided we protect the CHO group and exchange Mg for Cu. O

C–C

OAc

MgBr

O

H

OAc

+ I

H 53

54

55

The coupling actually needs only 0.06 moles of Cu(I) per mole of Grignard reagent from 57 and gives the product in two simple steps.22 Cuprates are often further complexed with Me2S, cyanide, or BR3 and couple well with tosylates as well as halides.23

Br

O

O HO

O

cat H

H

Br

OH

56

I

H

O

57

O

2. 0.06 equivs Cu(I)I, 48

Ac2O, HI Zn dust

O

55; 80% yield

57 1. Mg, THF

OAc

OAc

HCl, MeOH AcOH, H 2O

53 90% yield

H 58; 77% yield

Acylation of copper derivatives with acid chlorides works well, but they do not react with free carboxylic acids, unlike the more basic alkyl-lithiums. A dramatic illustration of chemoselectivity comes in the interaction of Bu2CuLi with the free acid 59 and its acid chloride 61. Reaction occurs24 at the alkyl iodide with the one to give 60 and at the acid chloride with the other to give the ketone 62. Acylation can also be achieved by many other metal complexes, from Al to Zr, but to make those we need hydrometallation, the subject of the next section. CO2H

I

Bu2CuLi

CO2H 60; n-C 14H29CO2H, 60% yield

59

O COCl

I 61

Bu2CuLi

I 62

Hydrometallation Hydrometallation is the addition of a metal hydride across a double or triple bond. It might be familiar to you through hydroboration (see chapter 17) while that other not-very-metallic metal, silicon, can be added through hydrosilylation. We shall concentrate here on hydrozirconation. Zirconium is a fairly abundant transition metal in the same group as that very useful metal titanium (see chapters 2–5) with stable oxidation states of ⫹2 and ⫹4. It is normally used in organic synthesis as bis cyclopentadienyl complexes such as Cp2ZrCl2 63 or Cp2ZrHCl 64. We met complexes like these in chapter 4 where zirconium enolates gave syn-selective aldol reactions. Hydrometallation with 64 and an alkene involves the formation 65 of an unstable 18-electron

8 Transition Metal Complexes

121

complex with the two π-electrons of the double bond and the reorganisation of this back to a 16electron σ-complex by a hydride transfer from Zr to carbon 66. The result is a carbanion-Zr(II) complex 67. Cl H

Cl Zr

Zr

Cl

Cl

Cl

Cp2Zr

Cp2Zr

H 63; ZrCp 2Cl2 16 electrons

Cp2Zr

Cl

H

H

64; ZrCp 2HCl 16 electrons

66; π-complex

65

67; σ-complex

This is a good point to address an irritating problem with σ-complexes: their tendency to decompose by β-elimination of metal hydrides. We have already met the tendency of crowded Grignard reagents to donate β-hydrogen atoms 10 and act as reducing agents. Alkyl zirconium complexes prefer to lose the same hydrogen atom by transferring it to the metal in a β-elimination. Complex 69, formed by hydrozirconation of E-hex-3-ene 68 could go back to starting material, or on, via complex 70 to a π-complex of hex-2-ene 71 and hence via 72, to the terminal σ-complex 73. ZrClCp2

H

Cp2ZrHCl

–Cp2ZrHCl

H

E-68

69

ZrCp2Cl

H

70

H H

ZrCp2Cl

ZrCp2Cl H

ZrCp2Cl 71

72

73

This means that any positional 74 or geometrical Z-68 isomer, or a mixture such as might be formed in a dehydration or Wittig reaction, gives the same terminal Zr σ-complex 73. Cp2ZrHCl

Cp2ZrHCl

ZrCp2Cl 74

73

Z-68

These organo-zirconium derivatives give good acylations with simple acid chlorides25 (and even better yields if treated with AlCl3). Thus any linear hexene gives 2-octanone 75 by zirconation and acetylation, and even branched chain alkenes such as 76 are acylated at the terminus 78 by this method. MeCOCl

73 PhH

O 75; 80% yield

Cp2ZrHCl

MeCOCl

ZrClCp2 76

77

PhH

O 78; 72% yield

8 σ-Complexes of Metals

122

Carbonylation of organometallic complexes Formylation (addition of an electrophilic CHO group) has always been a difficult operation. It cannot be accomplished with HCOCl as this decomposes to HCl and CO. Some metal complexes will react with CO and this gives us another approach to acylation at carbon. The tendency of transition metal σ-complexes to lose an H atom by β-elimination is just one symptom of the poor performance of alkyl groups as ligands. A simple σ-bond gives nowhere near the stability conferred by ligands that can simultaneously donate and accept electrons. The outstanding examples that we shall meet again and again are the phosphines, such as Ph3P, and CO. Phosphines donate their lone pair electrons and accept electrons from the metal into empty d-orbitals. On balance they are electron donors. Carbon monoxide donates its lone pair (from carbon) 79 and accepts electrons into the π* orbital 80. On balance CO is an electron acceptor: though the resulting complex can be represented with a metal-carbon double bond 81, two of these electrons come from the metal and CO is a two electron donor and two electron acceptor.

R4Zr

C

O

R4Zr

79

C

O

R4Zr

80

C 81

O

Alkyl zirconium complexes such as 67 react with CO to give an unstable 18e complex 82 that transfers the alkyl group from the metal to the CO π* orbital to give the metal acyl complex 83. This is a Zr(IV) complex of an acyl anion 83a and can be protonated to give the aldehyde 84 in excellent yield. These zirconium complexes are usually made from alkenes so that any hexene or mixture of hexenes gives 85 again in excellent yield.26

O Cl 1. Cp 2ZrHCl

C 2. CO

Cp2Zr

3. H , H 2O

Cp2Zr

H

Cp2Zr

Cl 67

O

O

Cl 82

83

84; 97% yield

O (+) Cp2Zr

(–)

linear hexenes

1. Cp 2ZrHCl

2. CO

CHO

ZrCp2Cl 3. H,

Cl

83a σ-complex of acyl anion

73

H 2O

85; 99% yield

Palladium σ-complexes The metal for carbonylation reactions, and for many other extraordinary transformations is palladium. You will be familiar with this metal’s excellent performance in hydrogenation reactions, usually as a finely divided powder on charcoal, and so you are aware that Pd(0) is stable. It also has a stable Pd(II) state, and there is little to choose between them in stability or in importance to organic chemists. Soluble reagents such as Pd(PPh3) 4, PdCl2, or Pd(OAc)2 are often used in reactions. Palladium is expensive and catalytic reactions are preferred. Mg(II), as in RMgBr, is

8 Transition Metal Complexes

123

more stable than Mg(0), as in the metal, by 2.4 eV but Pd(II) is more stable than Pd(0) by only 0.2 eV. Once Mg has been oxidised to Mg(II) it will not return to the metal without serious reduction. On the other hand, Pd(II) can easily return to Pd(0) during a chemical reaction. Pd can be used catalytically: Mg cannot. Hydrogenation works in solution because Pd forms stable π-complexes 87 with alkenes 86, easily inserts into a hydrogen molecule to give 88 and transfers H atoms one at a time to the alkene to give a σ-complex 89 and eventually the alkane 90. H PdL3

PdL4

R

R

R 86

87

PdL3

R

L

88

H

H

PdL2 H

H2

H L

H + PdL4

R 89

90

This well known reaction can deceive us into thinking that σ-complexes of Pd are stable. They are not. All the steps between 86 and 89 are reversible and in the absence of hydrogen, the reaction runs backwards. Palladium readily does the oxidative insertion into an alkyl halide to give the σ-complex 92, a carbanion complex of Pd(II), but this immediately loses hydrogen by β-elimination and the alkene 86 is formed with the loss of a PdHBr complex. Notice that this Pd(II) complex immediately reverts 93 to Pd(0).

R

oxidative insertion

Br

H

Pd(PPh3)4

Br

R

91

β-elimination

Pd(PPh3)3 92

Br Pd(PPh3)3

+

R

Ph3P

HBr + Pd(PPh3)4

H 86

93

Pd is very widely used in organic synthesis27 and we shall meet η2 alkene and η3 allyl complexes later. This chapter will end with a brief description of Pd σ-complexes. Stable Pd σ-complexes can be formed from ArX, MeX, and a few blocked alkyl compounds with no β-hydrogens. They react well with alkyl and acyl halides, but a major application is in carbonylation reactions.28 A simple example is the oxidative insertion of Pd(0) into PhI and its reaction with CO to give the complex 96. This illustrates several important points though the reaction is no great news as a synthesis. First, soluble Pd often comes quite exotically wrapped, here as the “diphos” complex 94, chiral versions of which we shall meet in chapters 25 and 26. Then the decomposition of the acyl-Pd complex 96 by a nucleophile rather than acid leads to the loss of the metal as leaving group since it reverts to Pd(0) as it goes 97. In the corresponding reaction with zirconium, the Zr left as Zr(IV) as it has no tendency to drop down to Zr(II), let alone Zr(0). Finally, the reaction is modestly catalytic in Pd, only 0.2 equivalents being required (20 mol%).

8 σ-Complexes of Metals

124

Ph2P

PPh2

Ph-I

Ph2P

Pd

Ph3P

Ph

PPh3 94

Ph2P

CO

I 95

PPh2 Pd

Ph

PPh2 Pd

Ph2P MeOH

I

PPh2

O

Pd

Ph

K2CO3

O

O

94 + PhCO2Me

I

MeO

96

MeO

O

OMe

97

98

Much more impressive examples come in intramolecular reactions where the nucleophile (usually an OH group, but it can be NH) is built into the starting material before carbonylation. In these examples, Pd is often added as PdCl2 or Pd(OAc)2 for convenience and reduced to Pd(0) (by Ph3P?) in situ. The alcohol 99 gives an excellent yield (82%) of the lactone29 102 via the complexes 100 and 101. The yield falls off when the lactone size is increased to six (70%) or seven (42%). The natural product pseudomeconin 98 has been made by this method.30 Br Br

Pd(OAc) 2, Ph 3P

OH

Br

Pd(PPh3)2 O

Pd(PPh3)2

Bu3N

O

OH

CO, Bu 3N

O O

O Pd(0)

99

100

101

102

Adaption to the synthesis of alkaloids such as 103 requires as the disconnection simply the removal of the CO group! This is good strategy as disconnecting one bond in the middle ring allows a second disconnection (Mannich or Pictet-Spengler31) to two simple starting materials 105 and 106. The Pd-catalysed carbonylation needs 5% Pd and uses a nitrogen nucleophile to decompose the intermediate Pd-acyl complex.

MeO

MeO O

N

MeO

NH

–CO

MeO

Mannich

Br

O 103

O 104

O

O

MeO NH2

MeO

5% Pd(OAc)2

105

CHO Br

105 + 106

O 106

O

104

Ph3P, Bu3N, CO

103

8 References

125

Just occasionally a Pd reaction with an alkyl σ-complex is successful. One instance is the remarkable synthesis of four-membered ring lactones 109 from styrene oxide 107 with 1.6% Pd in 63% yield.30 Probably the OH group on the same carbon as the β-Hs in 108 inhibits β-elimination. O O

Br HBr

Ph 107

1.6% PdCl2 (PPh3)2

OH

Ph

O Ph

CO, DMF

108

109

Recent results with Cu(I) and Pd σ-complexes The displacement of halides from benzene rings by nucleophiles is traditionally a reaction that requires strongly electron-withdrawing groups, such as nitro, in the ortho- or para-positions. A highly significant recent discovery is that Pd(0) and Cu(I) can catalyse such reactions in simple unactivated aryl halides with, for example, amine nucleophiles. Even aryl chlorides with electrondonating substituents such as 110 combine with simple amines providing a trace of the palladium catalyst 112 is present giving excellent yields32 of amines 111. O 1 mole % catalyst

Cl

OMe

Pd

toluene, 80 °C

HN

t-Bu t-Bu 112; catalyst

AcO

OMe morpholine

110

N

t-BuONa

O

+

111; 95% yield

The copper-catalysed reactions are not so well developed but aryl bromides combine with primary amines catalysed by Cu(I) using the salicylamide ligand 115. Free OH groups are tolerated in either partner: the amine or the aryl halide.33 Me

Br

Me H2N

H N

NEt2 OH

OH

O

ligand 115

Me 113

5 mol% CuI, K 3PO4 DMF, 90 °C

Br

Me 114; 91% yield

HexNH2, ligand 115

OH

5 mol% CuI, K 3PO4

H N

OH 115; salicylamide ligand Me

OH

DMF, 90 °C

116

117; 81% yield

References General references are given on page 893 1. T. L. Brown, Adv. Organomet. Chem., 1966, 3, 365. 2. P. G. Willard in Comp. Org. Synth., Vol 1, Chapter 1.1, page 1; W. Setzer and P. von R. Schleyer, Adv. Inorg. Chem., 1985, 24, 353. 3. M. J. Pearce, D. H. Richards and N. F. Scilly, J. Chem. Soc., Perkin Trans. 1, 1972, 1655.

126

8 σ-Complexes of Metals

4. Clayden, Lithium. 5. L. S. Hegedus, Organopalladium Chemistry in Organometallics in Synthesis II: A Manual, ed. M. Schlosser, Wiley, Chichester, 2002, pages 1125–1217; V. Farina, Transition Metal Alkyl Complexes: Oxidative Addition and Transmetallation in Comprehensive Organometallic Chemistry II, eds E. W. Abel, F. G. A. Stone and G. Wilkinson, Pergamon, Oxford, 1995, volume 12, pages 162–241. 6. D. J. Cram, F. A. A. Elhafez and H. L. Nyquist, J. Am. Chem. Soc., 1954, 76, 22. 7. H. Kropf in Houben-Weyl, VI/1a, volume 2, Alkohole II, pp 928-1304; Patai, ed, Organo-Metallic, vols 2 and 4; B. J. Wakefield in Comprehensive Organo-Metallic Chemistry, ed G. Wilkinson, Pergamon, Oxford, vol 7, 1982, chapter 44, pages 1-110; B. J. Wakefield, The Chemistry of Organolithium Compounds, Pergamon, Oxford, 1974. 8. M. S. Kharasch and O. Reinmuth, Grignard Reactions of Non-metallic Substances, Prentice-Hall, New York, 1954, pp 147-166. 9. M. S. Kharasch and S. Weinhouse, J. Org. Chem., 1936, 1, 209. 10. J. K. Crandall and M. Apparu, Organic React., 1983, 29, 345. 11. J. D. Fulton and R. Robinson, J. Chem. Soc., 1933, 1463; 1936, 71, and note on page 80. 12. H. Bredereck, F. Effenberger and E. H. Schweizer, Chem. Ber., 1962, 95, 804; J. D. Davenport, R. E. Hackler and H. M. Taylor, Eli Lilley, 1969, French Patent 1,569,940, Chem. Abstr., 1970, 72, 100745. 13. R. W. Hoffmann, B. Hälzer, O. Knopff and K. Harms, Angew Chem., Int. Ed., 2000, 39, 3072. 14. R. W. Hoffmann and P. G. Nell, Angew Chem., Int. Ed., 1999, 38, 338; R. W. Hoffmann, B. Hälzer and O. Knopff, Org. Lett., 2001, 3, 1945. 15. S. Masamune, T. Kaiho and D. S. Garvey, J. Am. Chem. Soc., 1982, 104, 5521. 16. D. Seyferth and H. M. Cohen, J. Organomet. Chem., 1963, 1, 15. 17. R. P. Cassity, L. T. Taylor and J. F. Wolfe, J. Org. Chem., 1978, 43, 2286. 18. S. Nahm and S. M. Weinreb, Tetrahedron Lett., 1981, 22, 3815. 19. G. Magnusson, Tetrahedron Lett., 1977, 2713. 20. R. G. Pearson and C. D. Gregory, J. Am. Chem. Soc., 1976, 98, 4098; B. H. Lipshutz, J. A. Kozlowski and C. M. Breneman, J. Am. Chem. Soc., 1985, 107, 3197. 21. Vogel, page 480. 22. J. C. Stowell and B. T. King, Synthesis, 1984, 278. 23. B. H. Lipshutz, R. S. Wilhelm and J. A. Kozlowski, Tetrahedron, 1984, 40, 5005. 24. G. H. Posner and C. E. Whitten, Tetrahedron Lett., 1970, 4647; G. H. Posner, C. E. Whitten and P. E. McFarland, J. Am. Chem. Soc., 1972, 94, 5106. 25. For reviews, see D. J. Cardin, M. F. Lappert, C. L. Raston, and P. I. Riley in Comprehensive Organometallic Chemistry, eds G. Wilkinson, F. G. A. Stone, and E. W. Abel, Pergamon, Oxford, 1982, vol 3, pp 549-646; E-i Negishi, and T. Takahashi, Aldrichimica Acta, 1995, 18, 31-47. 26. C. A. Bertelo and J. Schwartz, J. Am. Chem. Soc., 1975, 97, 228. 27. A. D. Ryabov, Synthesis, 1985, 233. 28. B. M. Trost, Comprehensive Organo-Metallic Chemistry, ed G. Wilkinson, Pergamon, Oxford, vol 8, chapter 57, page 800. 29. M. Mori, K. Chiba, N. Inotsume and Y. Ban, Heterocycles, 1979, 12, 921. 30. A. Cowell and J. K. Stille, J. Am. Chem. Soc., 1980, 102, 4193. 31. Disconnection Approach p. 337. 32. D. Zim and S. L. Buchwald, Org. Lett., 2002, 5, 2413. 33. F. Y. Kwong and S. L. Buchwald, Org. Lett., 2002, 5, 793.

9 Controlling the Michael Reaction Introduction: Conjugate, 1,4- or Michael addition vs direct or 1,2-addition Sulfur and phosphorus ylids Where direct (1,2-) addition is necessary Using Copper (I) to Achieve Michael Addition Stereoselectivity in Michael additions of organo-copper(I) compounds Trapping the enolate intermediate by silylation Michael Addition followed by Reaction with Electrophiles Tandem Michael/aldol reactions A Double Nucleophile: An Interlude without Copper A Michael Reaction Coupled to a Photochemical Cyclisation: Copper Again Michael Additions of Heteroatom Nucleophiles Michael Additions with and without Copper: Functionalised Michael Donors

Introduction: Conjugate, 1,4- or Michael addition vs direct or 1,2-addition HO

Nu R

1; allylic alcohol

HNu Direct or 1,2Addition

O

O

HNu

R 2; enone

Conjugate or Michael or 1,4Addition

R

Nu 3; ketone

The Michael reaction (also known as conjugate or 1,4-addition) is at its simplest the addition of a nucleophile to an electrophilic alkene, usually an unsaturated carbonyl compound 2, to give the product of 1,4-addition 3 rather than direct addition at the carbonyl group 1. It is a useful reaction in synthesis since the nucleophile can be a heteroatom (O, N, S etc) or a variety of carbon compounds (organometallic compounds, enolates, aromatic rings etc). It has been extensively reviewed1 and the ground rules2 are: Direct (1,2) attack at the C⫽O group is dominated by charge interactions and the allylic alcohol 1 is the kinetic product. Conjugate, Michael, or 1,4-addition is dominated by frontier orbital interactions and the saturated C⫽O compound 3 is the thermodynamic product. The most reactive Michael acceptors tend to give 1,2-addition (roughly CHO⬎AlkCO⬎ArCO⬎CO2R⬎CN) while the least reactive tend to give 1,4-addition (reverse order). The most reactive, most basic or hard, nucleophiles tend to give 1,2-addition: the least reactive, least basic or soft nucleophiles tend to give 1,4-addition.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

128

9 Controlling the Michael Reaction

Steric effects matter: large groups next to the C⫽O end (a1) or at the a3 position can deflect addition to the other end.

Sulfur and phosphorus ylids Sulfur and phosphorus ylids provide good illustrations. Simple sulfur ylids 5 react with aldehydes and ketones to give epoxides.3 But what if the electrophile is a conjugated aldehyde or ketone? Two reactions are possible: direct addition 6 to give the epoxide 8 and 1,4-addition 11 to give the cyclopropyl ketone4 13. The structure of the acceptor cannot generally be altered so that of the ylid must. The least stabilised, least reversible, and hardest ylids are those 5 derived from Me3S⫹ 4 and these give epoxides, while the extra anion-stabilising ability of an S⫽O bond in the sulfoxonium ylids 10 makes them softer and more reversible, and they give the cyclopropyl ketones. Me

O

Me S

base

Me

S

CH2

Me 4; sulfonium salt

Me 5; sulfonium ylid

O

O

Me2S R2

R1

O

O R2

R1

6; direct addition

R2

R1

7

8; epoxide

O Me2S

base

CH3

Me2S

O CH2

O

R1

R2

10; sulfoxonium ylid

9; sulfoxonium salt

O

Me2S

R1

R2 R1

R2

11; Michael addition

13; cyclopropane

12

A good example is the natural product eucarvone 16 mentioned in the original Corey papers.5 The simple sulfonium ylid 5 gives the epoxide 16 in 93% yield by direct addition while the sulfoxonium ylid 10 gives the cyclopropane 14 in 88% yield by conjugate addition to the nearer alkene. In an extended conjugated system the alkene nearest to the carbonyl group is normally the most electrophilic. O Me2S

Me3S

CH3

O

O t-BuOK

14

O

t-BuOK

15

16

That sulfonium ylids can add in a Michael fashion is neatly shown by an intramolecular reaction of the ylid 17. Direct attack would form an awkward 7-membered ring but conjugate attack 17 gives a five-membered ring intermediate with both new C⫺C bonds necessarily formed cis to the ylid.6 Here is regio- and stereoselectivity.

H2C

Me

Me

S

S

MeS

O O

O 17

18

9 Using Copper (I) to Achieve Michael Addition

129

Where direct (1,2-) addition is necessary For some reactions, 1,2-addition is essential. The synthesis of allylic alcohols by reduction of enones, enals, or enoate esters requires a hard reducing agent and LiAlH4 is usually suitable (chapter 19). The Wittig reaction7 also requires 1,2-addition otherwise the usual four-centred elimination of a phosphine oxide is impossible. In fact 1,2 addition is generally preferred with hard unstabilised ylids, and if the ylid is stabilised, Michael addition may be reversible. In the synthesis of α-eleostearic acid8 19 any of the double bonds could have been disconnected. The cis double bond was preferred both because unstabilised ylids usually give Z-alkenes, and because both starting materials, the dienal 20 and the ylid 21 are easily made. Bu

Wittig

CO2H

19; α-eleostearic acid Bu

CO2

CHO + PPh3 20

21

You may guess that Grignard reagents and alkyl-lithiums are also very likely to give 1,2-addition, especially with aldehydes. A classic example is the addition of methyl Grignard to crotonaldehyde 22 to give the allylic alcohol 23 in high yield.9 O

22 crotonaldehyde

HO

MeMgCl

Me H

H

23; pent-3-en-2-ol 81-86% yield

Using Copper (I) to Achieve Michael Addition That was a useful synthesis of allylic alcohols, but the greater need is to accomplish 1,4-addition reliably with hard or soft nucleophiles and reactive or unreactive Michael acceptors. The organocopper σ-complexes we met in the previous chapter are the answer. There is some discussion10 over the exact mechanism of addition of organo-copper compounds to enones, but we shall present the simplest possibility. Organo-copper reagents may add to enones by first making an η2-complex 24 with the alkene. Transfer of the ligand R from copper to the enone forms the η3 allyl (enolate) complex 25 which gives the Michael adduct 26 on protonation. R

O R

R2

R1

Cu

R2

Cu O

R R1

24; η2 alkene complex

R2

Cu

O

R

O

H

R1

25; η3 enolate complex

R1

R2 26

The addition11 may be carried out using a Cu(I) catalysed Grignard reaction, a discrete RCu species with or without other ligands such as Me2S, or with the R2CuLi complexes described in chapter 8. A typical (and very early!) result12 is the addition of MeMgBr to the enone 27 giving mostly the allylic alcohol 28 by 1,2-addition. The same reagent with catalytic Cu2Cl2 reverses the regioselectivity to a reasonable degree.

130

9 Controlling the Michael Reaction

O

Me Me

Metal

27 Me Me

O

OH

Metal = MeMgBr

28; allylic alcohol 98.5:

29; saturated ketone :1.5

17.5:

:82.5

Metal = MeMgBr + cat. Cu 2Cl2

Stereoselectivity in Michael additions of organo-copper(I) compounds Better results are achieved with stoichiometric copper reagents, either RCu [alone or as a complex with sulfides Me2S, phosphines Bu3P, or phosphites (MeO)3P] or cuprates R2CuLi. Thus Me2CuLi adds exclusively13 1,4 to the difficult enone 30, while the Cu-catalysed Grignard reaction gives mostly allylic alcohol 31. The syn stereochemistry of the 1,4-product 32 is thermodynamically controlled and appears on protonation during the acidic work-up to give the more stable cis ring junction. Me Me

Metal

Me OH O 31; allylic alcohol 30 56% Me Metal = MeMgI, Cu 2(OAc) 2 0% Me Metal = Me2CuLi

H O 32; saturated ketone 34% 100%

In other cases anti stereochemistry is preferred. An ingenious synthesis of the trans di-t-butyl cyclobutanone 34 uses Michael addition to the enone 33. Here a Cu-catalysed Grignard reaction gives total regio- and stereoselectivity as well as a quantitative yield.14 The stereochemistry again comes from protonation of an enolate during work-up, but this time it is better to have the two large groups anti on the rigid four-membered ring.

MeMgI Cu2Cl2

O

O 33

34

Similar stereoselectivities are found when the new chiral centre is formed during the addition rather than in the work-up, but the stoichiometric reagents generally perform better. Thus PhCu adds to the enone 35 to give anti-36 in both a better yield and a better stereoselectivity than a copper-catalysed Grignard reagent.15 The 1,3-stereoselection in Me2CuLi addition to 37 is also excellent,16 and results largely from axial addition rather than a preference for anti addition. These conformational aspects of stereoselectivity are discussed in chapters 20–21.

9 Using Copper (I) to Achieve Michael Addition

131

O O

PhMgBr cat Cu2Cl2

Cu2I2

35

PhCu

PhLi

Ph 75% yield 87:13 anti:syn

35

O

100% yield 96:4 anti:syn

36

O

O

Me2CuLi

+ 98% yield

:2 syn-38

anti-38, 98:

37

Open chain compounds often give good regioselectivity too.11 The simple enone 39 adds a variety of organo-copper compounds17 in up to 95% yield of the 1,4-adduct 40. Unsaturated esters generally give good results too. O

O

Me

Me-Metal

OH

+

39

40

41

Unsaturated aldehydes and enones with heavy β-substitution in particular often give poor results either by competing 1,2-addition or by aldol condensation of the enolate products. Thus the enone18 42 and most aldehydes give predominantly 1,2-adducts 44 even with Me2CuLi. These problems can be avoided if the reaction is carried out in the presence of Me3SiCl.

Me2CuLi

O

+ HO O

Me

42

43

44

Cinnamaldehyde 45 gives mostly the alcohol 47 with Me2CuLi but if Me3SiCl is added before the organo cuprate, the less reactive cyanocuprate Me(CN)CuLi can be used and the 1,4-adduct 46 isolated in 98% yield. O

Me

O

OH +

Ph

H

Ph

45 Me2CuLi 1. Me 3SiCl, 2. Me(CN)CuLi, 3. H +, H 2O

H 46 minor 98% yield

Ph

Me

47 major not detected

Trapping the enolate intermediate by silylation Cuprates react only very slowly with Me3SiCl so it can be present during the Michael addition reaction, enhancing the rate of the reaction and the proportion of 1,4-addition, and trapping the enolate product as the silyl enol ether.18–22 Thus the troublesome enone 42 gives good yields of

132

9 Controlling the Michael Reaction

48 under a variety of conditions, and the 1,4-product 43 released by acid- or fluoride-catalysed hydrolysis.

H , MeOH

Me2CuLi Me3SiCl

O

or Bu4NF, MeOH

Me3SiO

42

O

48

43

Cuprates generally perform better than RCu in additions to aldehydes. Acrolein itself 49 gives the 1,4-product 50 in 88% yield with Bu2CuLi, and even β-substitution in the crowded enal 51 does not prevent Ph2CuLi adding 1,4 to give 52 in 99% yield.20

Bu2CuLi

CHO

Ph2CuLi

OSiMe3

CHO

Me3SiCl

50; 88% yield

49

OSiMe3

Me3SiCl

Ph 52; 99% yield

51

Michael Addition followed by Reaction with Electrophiles Both these silyl enol ethers 50 and 52 could of course be hydrolysed to the saturated aldehydes, but that would be to sacrifice the useful reactivity of these intermediates in aldol and other reactions explored in chapters 2-6. A more productive development is to react the silyl enol ether with an electrophile and hence develop a synthesis from three components in two consecutive reactions.23 This approach has formed the basis of many modern syntheses as it develops the target molecule so quickly and is discussed in chapter 37 under ‘tandem reactions’. It is not necessary to trap the enolate with Me3SiCl when lithium cuprates are used with ketones as the lithium enolate is the product of 1,4-addition. You may choose the lithium enolate or the silyl enol ether, whichever is more appropriate for the next step. Cyclohexenone adds Bu2CuLi to give the lithium enolate 54 which is quenched with methyl iodide to give the anti disubstituted24 ketone 55. This could in principle be made from the ketone 56, but whatever regioselectivity 56 might show with LDA cannot be expected to favour 54.

Me O

Bu

OLi

Bu2CuLi

53

54

MeI

Bu

O

55

Bu

O

56

Tandem Michael/aldol reactions Aldol reactions with lithium enolates are improved if zinc(II) salts are present, or, if silyl enol ethers are used, TiCl4 is also helpful. We have chosen the example of Me2CuLi addition to the enone 57 as the examples in the literature23 are overwhelmingly of cyclic enones, and we wanted an acyclic example. The aldols (mixed diastereoisomers) 59 are formed in 96% yield.25

9 Michael Addition followed by Reaction with Electrophiles

O

OH

OLi

Me2CuLi

O

PhCHO ZnCl2

57

133

59; 1:2 syn:anti

Ph

58

Before looking further at syntheses developed with this strategy, we must deal with an annoying feature of cuprates of the type R2CuLi. Only one of the two R groups is transferred to the enone from the copper: the other is lost during the work-up. If this is simply Me, Bu, or Ph there is no problem, and that is why all our examples have been selected from these cuprates. Where the group attached to copper has itself been synthesised, perhaps including some stereochemical features, it is not acceptable to waste half of one’s labours. Copper prefers to retain the better complexing group, that is the one with more π-bonds, or the one which forms the more stable anion.26 There is particularly good selectivity between alkynes (retained) and alkanes (donated) so complexes such as 62 donate the alkyl group to enones and then alkynyl copper 61 can be recycled. Pent-1-yne 60 is commonly used for this sequence as it is the smallest liquid terminal alkyne (only just liquid - b.p. 40 ⬚C). RLi

1. BuLi

Pr

H

Pr

Cu

Pr

CuRLi

2. Cu(I)X

60

61 O

R

62 OLi

R

R1

R2

OSiMe3

Me3SiCl

62

R1

R2

R1

R2

63

64

The disconnection for these three-component syntheses is to remove the trans substituents from the α and β positions. Thus the ketone 65 disconnects to the enone 66, a vinyl-Cu derivative and an electrophilic bromoester. In the event, a Cu2I2 catalysed Grignard addition followed by alkylation of the magnesium enolate gives pure trans-65 in 91% yield.27 O

O

Br

O disconnect trans groups

CO2Et

CO2Et

CO2Et

+ Cu

65

66

65a

Functionalised nucleophiles normally need to be protected, as in the synthesis of the bicyclic enone 67 which requires the addition of a d3 reagent to cycloheptenone 69. The Me2S complex of the Cu(I) derivative of 70 fills this role; deprotection and cyclisation being accomplished in acid solution.28 This synthesis also illustrates that tandem Michael-aldol sequences of this kind with an intramolecular aldol are perfectly satisfactory. O

O

O CHO

aldol

CHO

1,6-diCO

+ Cu

68

67

69 OMgBr

O

1. Mg, Et 2O

O

Br 70

O

O

69; d 3 reagent O

HCl

2. Cu 2Br2, Me 2S 69, THF

71

67; 70% yield

134

9 Controlling the Michael Reaction

A Double Nucleophile: An Interlude without Copper The Erythrina alkaloids such as dihydroerythramine 72 have an awkward junction of three rings at one carbon atom. The amide 73 offers the opportunity of disconnection of one ring but this still leaves the quaternary carbon atom intact 74. O

O

O O

N

O

FGA

N

O

MeO

2 x C–C

O

MeO 72

O

H N

MeO 73

74

One successful approach imagines two C⫺C disconnections around that carbon atom 75 using a nitro group to stabilise the anion 76 and a double electrophile 77 with a Michael acceptor on one side and some suitable a2 reagent, perhaps an epoxide, on the other.29 O O

O

O

H N

C–N amide

O NO2

O

NO2

O

CO2R

and FGI

MeO

2 x C–C

76

CO2R

MeO

MeO 75

74a

77

A palladium-catalysed reaction of an allylic acetate (chapter 19) was used for the first C⫺C bond formation and the planned Michael addition for the second though the intermediate 79 was not isolated. The conformational diagram of 80 should help you to see why that diastereoisomer was favoured. CO2Me O

Cs2CO3

NO2

O

O

O N

O

78

AcO

(Ph3P)4Pd

+ Ph3P, THF 60 °C

O

76

78

O

O Cs2CO3

NO2

O

O CO2Me

Ar

NO2

CO2Me CO2Me NO2

79

80; 79% yield, 4:1 diastereoisomers major diastereoisomer of 80

A Michael Reaction Coupled to a Photochemical Cyclisation: Copper Again An intermolecular three-component step may be followed by other intramolecular reactions. In the synthesis30 of β-panasinene 81, a principal component of ginseng oil, the obvious Wittig disconnection leads to an equally obvious 2⫹2 photochemical cycloaddition31 from dienone 83. Either group may be disconnected from the β-position of 83 so it makes sense to remove the larger 84.

9 Michael Additions of Heteroatom Nucleophiles

O

135

O

Wittig

O Michael -Aldol

2+2

CH2O + Cu

81; β-panasinene

82

83

84

85

Copper-catalysed Michael addition followed by an aldol reaction with formaldehyde, gives a 1:1 mixture of diastereoisomers of the aldol 86 that can be eliminated to the enone 83. The resulting efficient photochemical cycloaddition gives ketone 82 with total regioselectivity probably because it is intramolecular. O

1.

OH

MgBr

O 1. TsCl, pyridine (82% yield)

5% Cu2Br2, Me 2S

83

hν

2. t-BuOK (78% yield)

2. CH 2O

82 67% yield

86; 90% yield, 1:1 syn:anti

85

Then came a nasty surprise. The ketone 82 is rather crowded and would not undergo a Wittig reaction with Ph3P⫽CH2, but Johnson turned this to advantage by using his asymmetric reagent 87 to resolve the compound and do the methylenation in one step. Optically active 88 can be isolated in 42% yield (out of a maximum of 50%) and gives natural (⫺)-panasinene 81. Ph O

Ph Li

O S

O

OH

S

NMe Al/Hg

NMe 87

HOAc THF

(–)-81; β-panasinene 92% yield

(+)-88; 42% yield

82

Michael Additions of Heteroatom Nucleophiles Michael additions of heteroatoms (O, N, S etc.) have been developed in modern chemistry to a high degree of regio- and stereoselectivity. Intramolecular Michael addition is particularly favoured and contributes a key reaction in Mulzer’s synthesis of the antibiotic kedomycin32 89. The intermediate 90 contains all the stereochemistry of the left hand side of kendomycin, the awkward quinone methide ring is replaced by a benzene ring, and, most significantly, a ketone has been introduced to allow Michael disconnection of a strategic C⫺O bond. OR

OH O HO

O

O Michael

O

O

OH OMe

HO HO O 89; kendomycin

OR

MeO

OMe

HO MeO

OMe 90

OMe 91

136

9 Controlling the Michael Reaction

The enone 91 was made by a Horner-Wadsworth-Emmons reaction (chapter 15) and so has the E-configuration. The cyclisation to 90 in dilute methanolic acid in 92% yield occurs with remarkable selectivity. Only one of the two free OH groups reacts, only conjugate addition occurs, and the stereoselectivity is 97:3 in favour of the isomer shown. Michael additions of heteroatoms can be coupled to reactions of the specific enolate produced. The synthesis of the modified carbonucleoside (⫺)-neoplanocin A 92 required some cyclopentane to which the heterocyclic purine could be joined. The decision was taken to close the cyclopentane by an aldol reaction. This required an aldehyde and an ester in a compound such as 94 to combine regio- and stereo-selectively without epimerisation of any chiral centres. The aldehyde is enolisable so there are serious problems.33

NH2 N

N

N

N

OH

EtO2C

H HO OH 92; (–)-neoplanocin A

HO

SR

EtO2C

SR

HO

EtO2C

aldol

Michael

OH

OH

OH

O

93

OH OH

O

94

95

A sulfur atom was introduced to supply the control. Michael addition of a thiol (PhCH2SH was actually used) to the unsaturated ester 95 would give a specific enolate that might cyclise onto the aldol fast enough to prevent enol exchange. In practice the hydroxyl groups had to be protected and the lithium thiolate was added to 96 to give the specific enolate 97 that cyclised with good stereoselectivity to give mainly 98. Elimination of the SBn group by oxidation (chapter 15) gave 99, used to complete the synthesis.

OEt EtO2C

O LiS

S

Ph

OR

OR OR

O 96

Bn

OR

O

SBn OR

EtO2C

HO

97

OR 98

OH

EtO2C

RO

OR 99

Michael Additions with and without Copper: Functionalised Michael Donors Perhaps the major disadvantage of the copper-catalysed Michael reaction is the need to protect functional groups. Michael reactions with heteroatom nucleophiles or functionalised carbon nucleophiles are usually carried out without copper. Many such nucleophiles, particularly enol equivalents, are excellent at Michael reactions, and are often to be preferred for ease of working. Silyl enol ethers, enamines, and stabilised enolates all carry out Michael reactions. In the next chapter we shall look at specific enol equivalents in more detail and note which ones are good at Michael additions.

9 References

137

Do not suppose that the regioselectivity problem is now trivial. In a study mainly aimed at achieving asymmetric reactions, Braun34 showed just how narrow is the gap between 1,2 and 1,4 addition. The silyl enol ether 101, easily made from the enantiomerically pure ester 100, gave mainly (⬎95:5) direct (1,2) addition to a simple enone under the usual conditions (TiCl4 catalysis) for Michael addition. Ph

Ph

O O

OH Ph 100

Ph

2. Me 3SiCl

Me3SiO

Ph

Ph O

O

1. LDA

OSiMe3

O

TiCl4

HO

Ph 101

Me O 102

Ph OH

Ph

A minor change in the structure of the enone: the addition of an α-methyl group to give 103, led to an almost complete reversal of the regioselectivity: the Michael (1,4) adduct 104 was formed in a 95:5 ratio to the 1,2 adducts. The only difference is a slight increase in steric hindrance for direct addition but the switch from one reaction to the other is remarkable and a warning not to be complacent about our understanding of selectivity. Ph 101

O

OH

TiCl4

O 103

O

O

Ph

Ph

104

References General references are given on page 893 1. E. D. Bergmann, D. Ginsburg, and R. Pappo, Org. React., 1959, 10, 179; J. P. Schaffer and J. J. Bloomfield, Org. React., 1967, 15, 1; the Dieckmann reaction, see table 5; G. Jones, Org. React., 1967, 15, 204; the Knoevenagel reaction, unsaturated carbonyl compounds scattered throughout the tables; G. H. Posner, Org. React., 1972, 19, 1; organo-copper reagents; E. Erdik, Tetrahedron, 1984, 40, 641; Copper (I) catalysed reactions or organolithiums and Grignard reagents; A. G. Cook, Enamines, Marcel Dekker, New York, second edition, 1988, pages 190-194 and 353-380; I. Fleming, J. Dunogues, and R. Smithers, Org. React., 1989, 37, 1; M. J. Chapdelaine and M. Hulce, Org. React., 1989, 38, 225 “tandem vicinal difunctionalisation”; D. Duval and S. Géribaldi in The Chemistry of Enones, ed S. Patai and Z. Rappoport, Wiley, Chichester, 1989, chapter 10, pages 355–469. 2. Fleming, Orbitals, pages 70-73; Disconnection Approach, page 116. 3. Disconnection Approach, page 252. 4. B. M. Trost and L. S. Melvin, Sulfur Ylids: Emerging Synthetic Intermediates, Academic Press, London, 1975. 5. E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc., 1962, 84, 3782; 1965, 87, 1353. 6. R. S. Matthews and T. E. Meteyer, Chem. Comm., 1971, 1576. 7. Disconnection Approach, page 120. 8. L. D. Bergelson and M. M. Shemyakin, Angew. Chem., Int. Ed., 1964, 3, 250 (see p. 255). 9. E. R. Coburn, Org. Synth. Coll., 1955, III, 676. 10. S. Woodward, Chem. Soc. Rev., 2000, 29, 393; see also E. J. Corey and N. W. Boaz, Tetrahedron Lett., 1985, 26, 6015; S. R. Kraus and S. G. Smith, J. Am. Chem. Soc., 1981, 103, 141; G. Hallnemo, T. Olsson and C. Ullenius, J. Organomet. Chem., 1985, 282, 133; H. O. House, Acc. Chem. Res., 1976, 9, 59. 11. G. H. Posner, Org. React., 1972, 19, 1. 12. M. S. Kharasch and P. O. Tawney, J. Am. Chem. Soc., 1941, 63, 2308.

138

9 Controlling the Michael Reaction

13. W. E. Parham and L. J. Czuba, J. Org. Chem., 1969, 34, 1899. 14. J. Salaün and J. M. Conia, Bull. Soc. Chim. France, 1968, 3730. 15. N. T. Luong-Thi and H. Riviere, Compt. Rend., 1968, 267, 776. 16. H. O. House and W. F. Fischer, J. Org. Chem., 1968, 33, 949. 17. see reference 11, table I, pages 67–68. 18. E. J. Corey and N. W. Boaz, Tetrahedron Lett., 1985, 26, 6019. 19. A. Alexakis, J. Berlan, and Y. Besace, Tetrahedron Lett., 1986, 27, 1047. 20. E. Nakamura, S. Matsuzawa, Y. Horiguchi, and I. Kuwajima, Tetrahedron Lett., 1986, 27, 4029. 21. M. Bergdahl, E.-L. Lindstedt, M. Nilsson, and T. Olsson, Tetrahedron, 1988, 44, 2055. 22. S. H. Bertz and R. A. J. Smith, Tetrahedron Lett., 1990, 31, 4091. 23. M. J. Chapdelaine and M. Hulce, Org. React., 1989, 38, 225; R. J. K. Taylor, Synthesis, 1985, 364. 24. G. H. Posner, J. Am. Chem. Soc., 1975, 97, 101. 25. K. K. Heng and R. A. J. Smith, J. Org. Chem., 1981, 46, 2932. 26. The details are more complicated: see P. A. Bartlett, J. D. Meadows and E. Ottow, J. Am. Chem. Soc., 1984, 106, 5304. 27. K. C. Nicolaou and W. E. Barnette, J. Chem. Soc., Chem. Commun., 1979, 1119. 28. A. Marfat and P. Helquist, Tetrahedron Lett., 1978, 4217. 29. C. Jousse-Karinthi, C. Riche, A. Chiaroni and D. Desmaële, Eur. J. Org. Chem., 2001, 3631. 30. C. R. Johnson and N. A. Meanwell, J. Am. Chem. Soc., 1981, 103, 7667. 31. Disconnection Approach, page 268. 32. H. J. Martin, M. Drescher, H. Kählig, S. Schneider and J. Mulzer, Angew. Chem., Int. Ed., 2001, 40, 3186. 33. M. Ono, K. Nishimura, H. Tsubouchi, Y. Nagaoka and K. Tomioka, J. Org. Chem., 2001, 66, 8199. 34. M. Braun, B. Mai, and D. Ridder, Eur. J. Org. Chem., 2001, 3155.

10 Specific Enol Equivalents Introduction: Equilibrium and specific enolates Enolates from 1,3-di-Carbonyl Compounds Enamines and Aza-Enolates The synthesis of multistriatin Metallated hydrazones Lithium Enolates and Silyl Enol Ethers Specific enolates by conjugate (Michael) addition An intermediate in Corey’s ginkgolide synthesis A synthesis of jasmine ketolactone Tables of Enol Equivalents and Specific Enolates Modern Use of Specific Enolate Equivalents Malonates, lithium enolates and silyl enol ethers in one synthesis Other metal enolates Coupling large fragments with metal enolates

Introduction: Equilibrium and Specific Enolates Almost every chapter so far has mentioned enolates in one form or another and their chemistry is central to organic synthesis. Now it is time to take stock of the various d2 reagents developed to play the role of enolates, to see what reactions they are particularly suited for, what chemo- regioand stereo-selectivity they show, and some examples of their use in total synthesis. This is only a very brief summary of simple enolates. They are discussed more fully in The Disconnection Approach, chapters 18-28, pages 140-239. Originally enolates were generated by treating carbonyl compounds with weak bases, that is bases too weak to convert the carbonyl compound entirely into its enolate. Under these conditions, self-condensation with unenolised carbonyl compound is always a risk. With aldehydes it is unavoidable. Treating an aldehyde 1 with an alkoxide produces a small amount of enolate 2 that reacts rapidly with unenolised 1 to give aldols 3, enals 4, and further condensation products.

R O

R

EtO

O

R

H 1

H 2

1

R

R CHO

OH 3; aldol

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

–H2O

R

CHO 4; enal

140

10 Specific Enol Equivalents

The less reactive ketones can be persuaded to react with other, usually non-enolisable, carbonyl compounds as in the reaction of cyclohexanone 5 with diethyl carbonate to give the β-keto-ester 7. This reaction is really under thermodynamic control as the favourable equilibrium between the product 7 and its stable enolate 8 more than compensates for the unfavourable equilibrium between 5 and its unstable enolate 6. This stable enolate 8 is a specific enolate. O

O

O EtO

EtO

O

H

O

O

O EtO

OEt

OEt

5

6

OEt

7

8

This product can be used to make 2,2,6-trimethylcyclohexanone 10 by methylation under forcing conditions to replace all the enolisable hydrogen atoms in 7a and give 9. In this reaction both the stable enolate 8 and the unstable enolates on the other side of the ketone must be methylated. The reaction1 is completed by hydrolysis and decarboxylation of the CO2Et group and the overall yield is a reasonable 41%. O H

H

O

O

Me

Me

H

OEt

1. NaH

O

Me

O Me OEt

TsOH

Me

Me

H

AcOH

2. MeI

9

7a

10

Enolates from 1,3-di-Carbonyl Compounds A simple example2 should remind you of these useful reagents. Condensation of the stable enolate 12 of the 1,3-diketone 11 with the enone 13 gives the Michael adduct 14. Elimination generates a new enone 15.

O H

NaOH EtOH Et2O

O

O

O + O

Cl 12

11 O

13 O

O

O

base

Cl

O 14

O 15

The new enone 15 condenses in turn in a second Michael addition with the keto-ester 16 under equilibrating conditions to form, presumably, intermediates like 17a and/or 17b. This gum is not isolated but hydrolysed and decarboxylated to give the steroid intermediate 18.

10 Enolates from 1,3-di-Carbonyl Compounds

141

O

O O

15

O CO2t-Bu

?

O

O

TsOH

cat t-BuOK t-BuOH

AcOH

O

O CO2t-Bu

16

O

CO2t-Bu 17a

18

17b

This sequence illustrates the use of enolates from 1,3-dicarbonyl compounds in Michael reactions: they are useful too in alkylations, aldol condensations (Knoevenagel conditions), and reactions with epoxides, as in the synthesis3 of 20. Nowadays they tend to be used if they are readily available, or if the disconnections suggest their use, as in the building of 11 into 18. Examples include the diketone 11 and the six-membered equivalent both used in steroid synthesis, acetoacetates 16 and 19 and the keto-lactones 20, malonic acid 21 and its esters, “Meldrum’s acid” 22, a very enolisable malonate derivative,4 and the keto-ester 25 formed via its stable enolate 24, by the cyclisation of the diester 23, an intermediate in nylon manufacture. The compounds 11, 16, 19, 20; R⫽H, 21, 22, and 25 are all available commercially. O O

O

O

O

EtO

CO2Et

O

OH

O

OH

R

O

Me2CO Ac2O H2SO4

O O

O

R 19

20

21

22

O CO2Et

EtO2C

O

EtO

CO2Et

CO2Et

EtOH

23

24

25

The 1,3-dicarbonyl approach is not suitable for aldehydes as the final destruction of the dicarbonyl relationship usually means that the aldehyde itself is sacrificed, as in the sequence 26 to 30 used in the alkylation of ketones.5 O

O

O

O

H

O

HCO2Et

O

O RX

base

H EtO

5

26 O

27 R

O

28 O R

NaOH

H

29

30

When the position between two carbonyl groups is blocked, attack by a nucleophile 31 occurs at the more electrophilic of the two (an aldehyde if there is one) cleaving 32 the 1,3-dicarbonyl system.

142

10 Specific Enol Equivalents

O

O

R

O H

O

O OH R

OH

R

H

31

H2O

30 + HCO2 33

32

Enamines and Aza-Enolates One specific enol equivalent that works very well for aldehydes is an enamine. Secondary amines such as Me2NH, Et2NH, pyrrolidine, piperidine, or morpholine 34 add cleanly to aldehydes to give enamines 36. These relatively stable nitrogen analogues of enols react well with reactive alkyl halides, such as α-carbonyl halides, and in Michael additions. They were considered in chapter 2 and only a few extra examples will be given here. O R

O

CHO + HN

O

N

R 34

N

R

35

36

The diketo-aldehyde 37 has four electrophilic carbon atoms (A-D) and three positions for enolisation (1-3). Cyclisation could occur in twelve different ways. Three can be regioselectively realised by different conditions.6 With morpholine catalysis, the aldehyde forms an enamine which does a Michael addition on the enone (B) to give 39. Enamine formation with PhNHMe also gives the aldehyde enamine, but this time it does an aldol condensation with the simple ketone (D) to give 38. 1 OHC

B

PhNHMe

O

TiCl4 CF3CO2H

CHO

H

O

CHO C

HN

O O

A

TsOH

D

38 enamine at 1 adds to D

O

O 39 enamine at 1 adds to B

3 2 37; a diketoaldehyde

Finally, acid catalysed reaction gives the expected thermodynamic product 40. All three products involve the aldehyde: the enamines select for d2 behaviour of the aldehyde while acid catalysis brings it in as an electrophile (a1). You cannot expect to achieve such control in all examples, but it is clear that enamines have something special to offer as aldehyde specific enol equivalents. O TsOH

37 reflux benzene

O

40 enol at 2 adds to A

10 Enamines and Aza-Enolates

143

Alkylation with simple alkyl halides is generally a poor reaction with enamines. Alkylation often takes place on nitrogen instead of carbon, and Stork and others7 have developed the aza-enolates to remedy this deficiency. A primary amine, usually cyclohexylamine, combines with an aldehyde to form an imine 42. Treatment with LDA 43 gives the lithium derivative, the analogue of a lithium enolate, known as an azaenolate 44. These intermediates are alkylated reliably at carbon 45 with most primary, and even with secondary alkyl halides, to give the alkylated aldehyde 47 after hydrolysis of the imine 46. R2N H Li H2N

R1

CHO

N

R1

LDA

N

R1

42; imine

41 Li

Br

N

R1

43 Li

R2

R2Br

N

R1

44; aza-enolate

45

R2 R1

R2 H

N

O

R1

H2O

H 46; alkylated imine

47; alkylated aldehyde

The synthesis of multistriatin Dutch elm disease is a fungal disease, carried by elm bark beetles, which devastated the elm populations of many parts of the world in the 1970s. The beetles assemble at a suitable elm tree on the call of a pheromone containing, among compounds derived from the tree, one charactacteristic of the beetle itself, multistriatin 48. The stereochemistry of this compound was deduced by the synthesis of the various possible isomers as so little natural product was available.8 We shall use three of these syntheses to illustrate the application of aza-enolates to alkylation reactions. Multistriatin 48 is an acetal formed from the keto-diol 49. Disconnection 49a next to one of the branchpoints reveals the enolate of a symmetrical ketone (Et2CO). This will have to react with an alkylating agent 50 whose 1,2-diol could be masked if it were made from an alkene 51. This is doubly attractive as the alkene 51 is available by known chemistry from 52. O

1,1-diO

O

OH

=

O

acetal

alkylation of enolate

HO OH

OH 48; multistriatin

49

O 49a

1. Mg, Et 2O FGI

2. CO 2

HO OH 50

X

dihydroxylation

X 51

OTs 51; X = OTs

Br

3. LiAlH 4 4. TsCl, pyridine

52

144

10 Specific Enol Equivalents

For the alkylation step, Silverstein9 chose the magnesium derivative 55 of the aza-enolate of 53 and cyclohexylamine 41. Reaction with the toluene sulfonate 51; X⫽OTs gave inevitably a mixture of diastereoisomers of the enone 56. It turned out not to be necessary to make the diol as the epoxide 57 could be directly transformed into multistriatin 48 with a Lewis acid catalyst (48 and 57 are isomers). This method produced all four diastereoisomers and the all equatorial compound 48 was identical with natural multistriatin.

41

O

EtMgBr

TsOH benzene Dean-Stark

1. 51; R = OTs 2. HCl, H 2O

N

N

R 54; R = Cy

53

56 O

SnCl4

mCPBA

O

O

BrMg R 55; R = Cy O

+ O

O

O

48; 34 parts

57

48a, b, c; 1, 7, and 58 parts

Synthetic optically active compound was needed to check the absolute configuration of the natural product, and disconnection of the intermediate 56a to the aldehyde 58 was chosen as 59 could be made from natural (⫹)-citronellol. C–C

C–C

H O

H

O

56a

O

58

59

For the alkylation step, the lithium derivative of the t-butyl imine 60 gave a good yield, but of course a mixture of diastereoisomers. The rest of the synthesis10 is straightforward and gave a mixture of compounds from which the correct enantiomer of multistriatin was isolated in 53% yield. It emerged from this synthesis that the chiral centre next to the ketone could be epimerised to the correct configuration during acid-catalysed cyclisation of the final intermediates. 1. LDA

H

t-BuNH2

H

O

N

59

60

2. MeI 3. H , H 2O

t-Bu

1. EtMgBr

H

2. H , H 2O

SnCl4

3. Cr(VI)

O 58a

O

4. mCPBA

56b

O O (–)-48; multistriatin

All that is needed is a more controlled synthesis of the two other chiral centres, and one such started from the available cis butenediol 61. Protection as an acetal 63, epoxidation 64, and epoxide opening with a cuprate (chapter 8) gave the correct diastereoisomer of 65. This undergoes an acid-catalysed acetal rearrangement from the less stable seven to the more stable five-membered ring 66, and a suitable leaving group (iodide) is introduced 67. Every step goes in very high yield.

10 Enamines and Aza-Enolates

HO HO

OMe 61

O

OMe

+

TsOH

CH2Cl2

O

OH

2. TsCl, pyridine (98% yield)

I O

3. NaI, Me 2CO (99% yield)

O

65; 94% yield

2. NH 4Cl, MeOH

64; 99% yield

O

OH

1. Me 2CuLi

O

63; 89% yield

1. two drops conc. HCl

O

O

mCPBA

O

62

O

145

O

66; 98% yield

67

This time the lithium enolate 68 of 53 was used to give, after equilibration of the centre next to the ketone, a stereoselective synthesis of racemic multistriatin.11 The yield in the final step was 98% of an 85:15 mixture of 48 and 70 separated by chromatography.

67

LDA

53

10% HCl (aqueous)

O

O O

O

+ O

O O

OLi 69; 84% yield, 85:15 mixture

68

48; multistriatin

70; epi-multistriatin

It would be a fair summary of the present situation to say that lithium enolates of ketones or esters are now usually preferred to aza-enolates for nucleophilic alkylation, but that lithium derivatives of cyclohexyl or t-butyl imines are preferred for the nucleophilic alkylation of aldehydes. We shall treat lithium enolates later, along with the equally popular silyl enol ethers, but there is one more nitrogen derivative which should come first.

Metallated hydrazones Various derivatives of aldehydes and ketones are used to make stable crystalline compounds for purification and identification, chiefly oximes and substituted hydrazones, and it was inevitable that they should also be tried out as sources of d2 reagents. Only the N,N-dimethylhydrazones have emerged as important reagents. Dimethyl hydrazine gives E-hydrazones, e.g. 72 from unsymmetrical ketones 71 with the NMe2 group cis to the methyl group.12 Lithium derivatives 73 can be made with LDA or BuLi and these react with electrophiles (E⫹) to give a new bond to the methyl group 74. The products are formed as stable hydrazones and must be worked up under oxidative conditions, or by alkylation and hydrolysis, to restore the carbonyl group 75. O

H2N

N

NMe2

R

Li

NMe2 LDA

R 72; hydrazone

71 N

O

NaIO4, pH7

R 2. H , H 2O

E

R

E

or 1. MeI

74

NMe2

73; aza-enolate

NMe2 E

R

N

75

146

10 Specific Enol Equivalents

Clean reaction occurs with enolisable aldehydes or ketones to give aldol products, e.g. 76 and with epoxides to give products 77 with a 1,4-relationship between the functional groups. The oxidative work-up affects neither secondary alcohol, but these can of course be oxidised with Cr(VI) if a diketone 78 is wanted. O

NMe2

N H2N

NMe2

O

1. LDA

OH

2. MeCHO 3. NaIO 4, pH 7

71; R = Et

72; R = Et

76; 99% yield

O

1. LDA

O

PCC

72; R = Et 2. O 3. NaIO 4, pH 7

O

OH 77

78

Alkylation and Michael addition also occur, and these are illustrated together in Heathcock’s synthesis of lycodine intermediates. The dimethylhydrazone of acetone 79 can be alkylated to give 80. The hydrazone is retained while the lithium derivative is again formed. Though the new hydrazone 80 must intially have the Z configuration, its lithium derivative 81 can rotate to the E configuration while lithium is exchanged for copper so that a Michael addition to the enone 82 can be carried out. The product 83 has the expected anti-1,3 relationship due to axial attack, but no stereoselectivity was found on protonation next to the ketone.13

N

NMe2

N

1. BuLi 2.

N

BuLi

Br

79

Me2N

NMe2

O

80

CN

Li

CN

1. PhSCu

81-Cu + O

O

82

81

83

Lithium Enolates and Silyl Enol Ethers Lithium enolates 85 and silyl enol ethers 86 are probably widest in application in modern organic synthesis. The basic rules of selectivity were laid down in chapters 2-4 where many examples were given. We shall simply summarise the position and add some extra versatility from chapters 5-9. These two methods must be taken together because easy interconversion means that a way of making one is a way of making another. In addition, the silyl enol ethers 86 are a source of “naked” enolates 87 when fluoride is used to remove silicon in the absence of a metal cation. Tetraalkyl ammonium fluorides such as TBAF (Bu4N⫹ F⫺) are usually used. O

THF –78 °C

R 84

O

LDA

R

Li

MeLi

85

O

Me3SiCl

SiMe3 Bu4N

(TBAF)

R 86

NBu4

O

F

R 87

10 Lithium Enolates and Silyl Enol Ethers

147

Aldehydes 88 do not form stable lithium enolates as the aldol reaction is too fast, but they form stable silyl enol ethers 89 under equilibrium conditions. Acids 90 can be transformed into either type of derivative: the dilithium enolate 91 or the ketene silyl acetals 92. Esters give predominantly E-lithium enolates and silyl enol ethers, e.g. 65:15 E:Z -95; R⫽Me. Amides rarely give satisfactory enolates. O R

Me3SiCl

OSiMe3

R

Et3N

88 R

2 x LDA

CO2H

OLi

R

Me3SiCl

LDA

OSiMe3 92

OLi

R

CO2Et

OSiMe3

R

OLi 91

90 R

89

Me3SiCl

OSiMe3

R

OEt E-94

93

OEt E-95

With ketones we come to the problem of regioselectivity, and the situation from chapter 3 is that methyl ketones 98 and ketones with one primary and one secondary alkyl group, particularly cyclic ketones such as 103 give the less substituted lithium enolate 97 or 102 by kinetically controlled deprotonation with LDA, and the more substituted silyl enol ether 99 or 104 on silylation under equilibrium conditions. Either derivative (lithium enolate or silyl enol ether) may be used to make the other, e.g. 96 and 100. OSiMe3

OLi

Me3SiCl

O LDA

R

R

OSiMe3

Me3SiCl

R

OLi MeLi

R

R

Et3N

96

97

OSiMe3

OLi

R

Me3SiCl

R

98

99

O LDA

R

100

OSiMe3 Me3SiCl

R

OLi MeLi

R

Et3N

101

102

103

104

105

Specific enolates by conjugate (Michael) addition After our discovery of Michael additions of lithium cuprates to enones in chapter 9, we are also now able to make specific enolates 106 and 108 of unsymmetrical ketones whose branchpoint is the β rather than the α atom, such as 3-alkyl cyclohexanones 107. That is, providing we don’t try and make them from the ketone itself. O

O ?

R

?

R 106

O

R 107

108

148

10 Specific Enol Equivalents

Addition of R2CuLi to cyclohexenone 109 gives the lithium enolate 110 and hence the silyl enol ether14 111. It is not so obvious how to make the alternative enolates, but reduction of enones with Li in liquid ammonia or in a simple amine such as Et2NH gives the lithium enolate by the Michael addition of hydrogen.15 Two electrons are added to the enone to give the unstable dilithium derivative 114. One molecule of a weak acid (usually t-BuOH or water) is present which protonates the more unstable of the two anions to give the lithium enolate 115. This means that enone 112 is an alternative source of 110 and 111. OLi

O

OSiMe3

O

Me3SiCl

R2CuLi

Li

110 R

R

109

R

110

111 OLi

EtNH2 or NH3(l)

R

OSiMe3 Me3SiCl

ROH

Li

R

112

OLi

O

EtNH2 or NH3(l)

R

R

113

114

115

116

Since we know many methods to make enones with complete regioselectivity (chapter 5) we are equipped to make both specific enolates of any unsymmetrical ketone, though not necessarily from the ketone itself. It is of course a strategic advantage to have methods which make carbon-carbon bonds on the way to specific enolates as these methods are more versatile and more likely to be convergent.16

An intermediate in Corey’s ginkgolide synthesis A few examples should make modern approaches plain. In Corey’s ginkgolide synthesis he needed the intermediate 117 and decided to make it by coupling two vinyl derivatives (see chapters 11 and 20), the one derived from a triple bond 118 and the other an enol derivative of a cyclic ketone 119. OHC

OH

OHC

OH

vinyl coupling

+ RO

t-Bu

t-Bu

HO2C

HO2C 117

118

119

The first 118 was made by alkylation of a malonate ester followed by hydrolysis and decarboxylation in the traditional way. With no problems of regio- or stereo-selectivity there is something to be said for methods easily carried out on a large scale. EtO2C EtO2C

NaOEt EtOH

EtO2C EtO2C

1. KOH, EtOH

+ Cl

2. 150–175 °C

EtO2C CO2Et

HO2C 118

10 Lithium Enolates and Silyl Enol Ethers

149

The second 119 clearly came from the ketone 121 and this looks like an aldol product from CH2O and a specific enolate of the ketone 121. This ketone is substituted at only one α-atom, so regioselective enolisation would be possible. But there are problems of chemoselectivity (the other CHO group) and stereoselectivity and a Michael addition approach on the enone 122 looked more promising. Aldol disconnection of 122 seems to have uncovered an excellent symmetrical intermediate 123 which can cyclise only to give 122. OHC FGI

119

OHC

OH aldol

O

Michael

O

t-Bu

OHC

OHC aldol

O

t-Bu

O

+ t-BuM

+CH2O

CHO

121

120

122

123

Corey evidently did not want to work with a keto-dialdehyde 123 and preferred to consider the aldol disconnection on the alternative enone 124 (the double bond will readily move into the ring). This reveals a symmetrical dialdehyde 125, but this is too reactive to be used and Corey preferred the half-protected compound 126. OHC

OHC FGI

O

aldol

O

122

+ CHO CHO

O

CHO

125

124

OMe

MeO

126

The morpholine enamine 127 of cyclopentanone ensured a clean aldol reaction and strong HCl put the double bond into the ring to give 128. Cuprate addition and trapping by silylation gave the specific enol equivalent 129 which reacted with formaldehyde polymer under the usual Lewis acid catalysis to give 130, which is 120 protected as an internal acetal. The electrophile (CH2O) added to the opposite face of the enol to the one occupied by the t-butyl group.17 OMe O 1. 126

N

2. 6M HCl

127

MeO

1. t-Bu2Cu(CN)Li2

O

2. Me 3SiCl

128 OMe

MeO 1. (CH 2O)3 TiCl4 CH2Cl2

MeO Me3SiO

t-Bu 129

2. MeOH

O O

t-Bu 130

A synthesis of jasmine ketolactone The synthesis of the macrocyclic lactone 131 found in jasmine flowers illustrates the use of “naked” enolates. Lactone disconnection to 132 and the usual Michael-based three-component

150

10 Specific Enol Equivalents

disconnection reveals cyclopentenone, an acetate enolate equivalent, and a Z allylic halide 133 with an OH group which will have to be protected. O

O

OH OR

CO2R O

H

C–O

OH

H

+

lactone

H

Michael disconnect trans groups

H

O 131; jasmine ketolactone

X

O 132

O 133

The t-butylthiolester 134 provides the silyl enol ether 135 as the acetate enolate equivalent and adds to cyclopentenone with fluoride catalysis to give the silyl enol ether 136. Release of the naked enolate and alkylation with the alkyne 137 gives 138 which has the correct anti stereochemistry and can simply be transformed into the hydroxy acid.18 The final cyclisation demands special methods we shall meet in chapter 43. COSt-Bu 1. LDA

O

1. Bu 4N

OSiMe3

2. Me 3SiCl

St-Bu

St-Bu

1. Bu 4N

F

2. cyclopentenone

2.

F

Br OTHP

OSiMe3 134

135

136

137 O

COSt-Bu H

1. TsOH, MeOH, H 2O

132 2. H 2, Pd/C, Lindlar

H

OTHP

3. Hg(II), H 2O

macrolactonisation

H

O

H

O

O 131

138

When the enolate is formed by reduction of an enone, the first chiral centre is created by protonation and is often under thermodynamic control. Thus the steroid intermediate 139 is reduced to the trans decalin 140. Reaction with an electrophile (here CH2O) occurs on the same side as the added proton as the group that was already there (the other ring in 140) is bigger than the added hydrogen atom. The conformational drawing 141a shows that the CH2OH group is equatorial and may also have equilibrated by enolisation.15

CH2O

Li, PhNH 2 liquid NH3

O

LiO

H

O

H

O OH

H

HO 139

140

141

141a

Tables of Enol Equivalents and Specific Enolates Table 1 lists the simplest enolates. The less reactive enol-like compounds on the left and the more reactive enolate like compounds on the right. The headings reflect this but you should be aware by

10 Modern Use of Specific Enolate Equivalents

151

now that the simple enols and enolate anions shown there cannot be prepared and isolated but are normally used in equilibrium with the carbonyl compound. In general terms the reagents on the left are good at conjugate addition while those on the right are reactive enough to be alkylated. All react with simple carbonyl compounds in aldol and Claisen ester reactions. O

OH

O

R

R

R

Carbonyl compound

Enol

Enolate anion

SiMe3

O

O

Oxygen Derivatives

R

Li

O

R

Silyl enol ether

RO

Lithium enolate Reformatsky

Nitrogen Derivatives

N

N

R Enamine

R Phosphonate ester anion

O

O

R

OEt Enol

O

(RO)2P

R

H

O

Phosphorus Derivatives

Conjugated ylide O

Li

Aza-enolate

O Ph3P

ZnBr

O

R

O OEt

R

1,3-Dicarbonyl Compound

O

OEt Enolate anion

Modern Use of Specific Enolate Equivalents So which specific enolates are in normal use today? The answer is simple - all of them! We shall end this chapter with a few examples to make the point.

Malonates, lithium enolates and silyl enol ethers in one synthesis In a study of ring closure by metathesis (chapter 15) to produce specific enol equivalents from cyclic ketones, Shibasaki19 decided to use enolates from malonates to make the many starting materials quickly. For example, alkylation followed by conjugate addition gave starting material 143 in two steps.

O MeO2C

CO2Me

1. MeO 2. Br

1. MeO

MeO2C

CO2Me 142

2. O

MeO2C

CO2Me

143

152

10 Specific Enol Equivalents

This was converted into the kinetic lithium enolate and then into the silyl enol ether 144. Metathesis with the Grubbs catalyst (you will meet this properly in chapter 15) gave the cyclic silyl enol ether 145 in 99% yield. Me3SiO OSiMe3

O

Grubbs catalyst

1. LDA 2. Me 3SiCl

MeO2C

CO2Me

MeO2C

143

CO2Me MeO2C CO2Me 145

144

If the unsymmetrical cyclic ketone 146 was treated with LDA and Me3SiCl, some of the same silyl enol ether 145 was formed but as 30% of a mixture with the alternative 147. This could also be prepared by metathesis from the unsaturated compound 148, prepared in a similar way. This is surely the best way to produce specific enolates of (though not from) unsymmetrical cyclic ketones. This study shows how an old method (malonate), modern methods (lithium enolates and silyl enol ethers) and very modern methods (metathesis) can be used in cooperation to produce excellent results. Me3SiO

O

Me3SiO

Me3SiO

1. LDA

147 only 88% yield

+ 2. Me 3SiCl

MeO2C

CO2Me

MeO2C

CO2Me MeO2C

CO2Me

MeO2C

147 (70%)

145 (30%)

146

Grubbs catalyst

CO2Me

148

Other metal enolates Later in the book, when we deal with asymmetric enolate reactions, boron enolates will be very important. A simple example20 of an aldol reaction with a boron enolate, prepared from the ester 149 and a boron triflate using an amine as base, shows why. The boron enolate 150 could be prepared with a weak base and reacts with the aldehyde without catalysis to give essentially one diastereoisomer of the aldol 151 in good yield. If the titanium enolate (prepared with TiCl4 and an amine) was used, both the yield and the stereoselectivity were worse. In other circumstances enolates of titanium and other metals are very successful.

O

O Et3N

O O

Ph Ph 149

O

Ph Ph 150

OH

O i-PrCHO

O

R2BOTf

R = cyclohexyl

BR2 O Ph

O Ph 151; 86% yield

Coupling large fragments with metal enolates Even reactive lithium enolates can be used to couple complex fragments burdened with extensive functionality and stereochemistry. Mulzer’s work on epothilones21 combined the lithium enolate

10 References

153

of the ketone 152 with an aldehyde in an aldol reaction to give a 70% yield of the epothilone analogue 153.

S

O R = SiMe2Bu-t

O

OR

1. LDA

HO

2. R*CHO

O

OR

OR 152

N

O

OR

OR

153 70% yield

The product 153 is formed in good yield and stereoselectivity though the aldehyde (whose structure you can easily deduce from that of 153) has an enolisable chiral centre and various other functional groups and stereogenic centres. Specific enolates are vital reagents for the 21st century.

References General references are given on page 893 1. C. L. Stevens and A. J. Weinheimer, J. Am. Chem. Soc., 1958, 80, 4072; H. R. Snyder, L. A. Brooks and S. H. Shapiro, Org. Synth. Coll., 1943, 2, 531. 2. S. Danishefsky and B. H. Migdalof, J. Am. Chem. Soc., 1969, 91, 2806. 3. W. L. Johnson, U. S. Patent 2,443,827 (1948); Chem. Abstr., 1949, 43, 677; G. W. Cannon, R. C. Ellis and J. R. Leal, Org. Synth. Coll., 19, 4, 597. 4. K. Pihlaja and M. Seilo, Acta Chem. Scand., 1968, 22, 3053; 1969, 23, 3003. 5. House, Modern Synthetic Reactions, chapter 9, page 492. 6. S. D. Burke, C. W. Murtiashaw and M. S. Dike, J. Org. Chem., 1982, 47, 1349. 7. G. Stork and S. R. Dowd, J. Am. Chem. Soc., 1963, 85, 2178; G. Wittig and H. Reiff, Angew. Chem. Int. Ed, 1968, 7, 7; J. K. Whitesell and M. A. Whitesell, Synthesis, 1983, 517; D. Caine in Comp. Org. Synth., vol 3, chapter 1.1. 8. W. E. Gore, G. T. Pearce and R. M. Silverstein, J. Org. Chem., 1975, 40, 1705. 9. W. E. Gore, G. T. Pearce and R. M. Silverstein, J. Org. Chem., 1976, 41, 2797. 10. G. J. Cernigliaro and P. J. Kocienski, J. Org. Chem., 1977, 42, 3622; multistriatin has also been made from glucose, K. Mori, Tetrahedron, 1976, 32, 1979; 1977, 33, 289. 11. W. J. Elliott and J. Fried, J. Org. Chem., 1976, 41, 2475. 12. E. J. Corey and D. Enders, Tetrahedron Lett., 1976, 3, 11; E. J. Corey, D. Enders and M. G. Bock, Tetrahedron Lett., 1976, 7; E. J. Corey and D. L. Boger, Tetrahedron Lett., 1978, 4597; D. Enders and P. Wenster, Tetrahedron Lett., 1978, 2853. 13. E. Kleinman and C. H. Heathcock, Tetrahedron Lett., 1979, 4125. 14. G. Posner, J. J. Sterling, C. E. Whitten, C. M. Lenz and D. J. Brunelle, J. Am. Chem. Soc., 1975, 97, 107. 15. G. Stork and J. d’Angelo, J. Am. Chem. Soc., 1974, 96, 7114. 16. Disconnection Approach, page 346. 17. E. J. Corey, M. Kang, M. C. Desai, A. K. Ghosh and I. N. Houpis, J. Am. Chem. Soc., 1988, 110, 649. 18. K. Gerlach and P. Kunzler, Helv. Chim. Acta, 1978, 61, 2505. 19. A. Okada, T. Ohshima and M. Shibasaki, Tetrahedron Lett., 2001, 42, 8023. 20. M. B. Andrus, B. B. V. S. Sekhar, E. L. Meredith and N. K. Dalley, Org. Lett., 2000, 2, 3035. 21. J. Mulzer, G. Karig and P. Pojarliev, Tetrahedron Lett., 2000, 41, 7635.

11 Extended Enolates Introduction: The extended enolate problem Kinetic and thermodynamic control Wittig and Horner-Wadsworth-Emmons Reactions Extended Aza-Enolates Extended Lithium Enolates of Aldehydes Summary: α-Alkylation of Extended Enolates Reaction in the γ-Position Extended Enolates from Unsaturated Ketones Diels-Alder Reactions Extended Enolates from Birch Reductions The Baylis-Hillman Reaction The Synthesis of Mniopetal F A Synthesis of Vertinolide Using α´-and γ-Extended Enolates Conclusion: Extended Enolate or Allyl Anion?

Introduction: The extended enolate problem The extended enolate problem1 is easily stated but not so easily solved. If an unsaturated ester 1 is treated with base, a conjugated enolate 2 is formed. The conjugated ester 3 can form the same enolate 2 by loss of one of the marked protons. This intermediate 2 is an extended enolate, extended by conjugation into a double bond and interesting because a new dimension of versatility is added to its reactions. Enolates normally have to choose between reaction at oxygen or carbon, but the extended enolate may also choose at which carbon to react. This chapter is about the reasons for that choice. O R

OEt H

H 1

O

O base

base

R γ

α

R

OEt

OEt

2; extended enolate

H

H 3

The two possibilities are traditionally known as α and γ. Reaction in the α position 4 as a d2 reagent gives unconjugated ester 5 which may become conjugated1 by α to γ proton transfer via extended enolate 6. Reaction in the γ position 9 as a d4 reagent gives conjugated 10 directly. The remaining possibility – reaction at oxygen – is realised by silylation to give the extended silyl enol ether 8.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

156

11 Extended Enolates

O

α or d2

R

OEt

O R

O R

OEt E

E 4

OEt

H 5

OSiMe3 R

O R E

6

7 O

O R

Me3SiCl

OEt

γ or d4

OEt

R

OEt

E

E

H

9

8

OEt

E

10

Kinetic and thermodynamic control The kinetic site for reactivity is the α position, where the coefficient of the HOMO is larger,2,3 but reaction at the γ position gives the more stable conjugated product. Ester 5 is the kinetic product and 7 and 10 are thermodynamic products. Early experiments4 soon revealed that reactive lithium enolates such as 12 were alkylated almost exclusively with kinetic control at the α position, making synthesis of rings 14 possible by double alkylation.5 O

OLi

O

O

LDA

1. LDA

RBr

OEt

OEt

11

OEt

3. LDA

R 11

12

OEt

2. Br(CH 2)4Br

14

13

Even Michael addition occurred at the same (α) site and the resulting specific enolate 15 can be alkylated with the stereoselectivity expected for three-component reactions from chapter 9 (anti addition of the new groups). The product 16 was used by Oppolzer in his synthesis of khusimone.6 O O

OEt

1. LDA

OEt

O OEt Br

2. cyclopentenone

11 OLi

O 16

15

The dilithium derivatives 19 of carboxylic acids 17 are much more inclined to alkylation at the γ position7 to give 20 though α-alkylation is by no means uncommon,8 while extended enolates 22 of amides such as 21 occupy a midway position, ratios of 23:24 varying from 67:33 to 98:2 for different R groups. In both these last two examples the metal can be exchanged for copper to improve γ-selectivity, as it seems copper favours conjugate addition both for nucleophiles and for electrophiles.9 O

OLi

O LDA

OH

OLi

OLi 18

17

O

LDA

NR2 22

R

OH

19

OLi

NR2 21

O RBr

LDA

H

20

O

O

RBr

NR2

+

R

NR2

R 23

24

11 Extended Aza-Enolates

157

Alkylation is essentially irreversible, so thermodynamic products are difficult to make, but the aldol reaction is reversible, and the proportion of reaction at the γ position of ester 11 and acid 17 derivatives in aldol reactions increases both with temperature and time. Hence the α-aldol 27 is the product with the extended enolate of ester 11 if the reaction is worked up at low temperatures. At higher temperatures the γ-aldols 25 and E-26 are the only products,10 the lactone 25 coming from cyclisation of Z-26. R

1. LDA

O +

O O Z-25

2. RCHO

work up at –15 °C

OEt

O

1. LDA

O

2. RCHO

OEt

OEt

work up at –78 °C

R

11

E-26

OH 27

Wittig and Horner-Wadsworth-Emmons Reactions A device for getting extended enolates of esters to combine with aldehydes and ketones in the γ position is to use a Wittig approach. The bromo-ester 28 is commercially available. Reaction with a trialkyl phosphite gives the phosphonate ester 29 that gives dienes such as 30 with aldehydes.11 Phosphonium salts can also be used.12 O

(EtO)3P

CO2Me

Br

CO2Me

1. (Me 3Si)2NLi

(EtO)2P

CO2Me

28

2. i-PrCHO

29

30; 97:3 E.E:Z,E

Other γ-bromoesters such as 31 can be made by radical bromination13 with NBS (this too is a γ reaction!) and used in similar Horner-Wadsworth-Emmons reactions. It does not matter if some reaction occurs at the γ position of 29 or 33 as the Wittig elimination cannot occur on such an intermediate and it reverses. The products of these ‘aldol’ reactions are dienes 30 and 34 and are usually made for use in Diels-Alder reactions. We shall return to this subject later in this chapter. Br

NBS

O

NaOH

11 Br

Br

CO2Et 31

base

O

O

(EtO)2P

32

O

O

O

(EtO)3P

O

RCHO

R

33

34

Extended Aza-Enolates Aldehydes present special problems as usual. For alkylations in the α-position, the aza-enolates discussed in chapter 10 are the best. The cyclohexylimine 37 can be made in 82% yield from crotonaldehyde14 and its lithium derivative 38 reacts cleanly at the α-position with alkyl halides15 to give the non-conjugated imine 39. K2CO3

CHO

LDA

+ benzene

H2N

N 37; imine

36

35

Li 38; aza-enolate

CHO

H

RBr

N

N

H2O

R

R 39

40

158

11 Extended Enolates

There is naturally a problem in hydrolysing 39 to the unconjugated aldehyde as the double bond tends to move into conjugation to give 40 unless the α-position is blocked 42 by two alkylations.15

1. LDA

N

H2O

R2

R1

R1

CHO

H

N

2. R 2Br

39

R2

R1

41

42

Aldol reactions also occur with these intermediates 38 though α,γ mixtures are often formed.16 Silylation of t-butyl imines occurs reliably in the γ-position and this is one of the best ways to control aldol reactions with extended enolates.17

CHO +

molecular sieves

H2N

1. LDA

N

0°C pentane

N

45; γ-silylated imine

44; imine

43

35

Me3Si

2. Me 3SiCl

Treatment of 45 with an aldehyde RCHO at room temperature with catalytic CsF to remove the silyl group creates low concentrations of free extended enolate that react in the γ-position with the new aldehyde to give the imine 46. This can be converted into the true γ-aldol 48 by hydrolysis and removal of the silyl group. OSiMe3

CsF, RCHO

45 DMSO

OSiMe3

ZnCl2

R

N

H2O

OH CHO

R

TBAF MeOH

46

CHO

R

47

48

Under more vigorous conditions (100 ⬚C) the dienal is formed in good yield and it is even possible to combine two enals to form a trienal, e.g. 50, regiospecifically.

45

catalytic CsF PhCH=CHCHO DMSO, 100°C

ZnCl2

Ph

N

CHO

Ph

H2O

49

50; 92% yield

Extended Lithium Enolates of Aldehydes Extended lithium enolates of conjugated aldehydes are more stable than those of simple unconjugated aldehydes, and can be used particularly where there is a branch at the α-position as in the vernolepin intermediate 53 which can be made by alkylation of the lithium derivative18 52. You will notice that the ketone in 51 is protected, an obvious precaution against enolisation, and it is impressive that the enolate-acetal in 52 does not fragment. O

OLi Br LDA

H O

O

O

O 51

CHO

H

O

O 52

53

11 Reaction in the γ-Position

159

Summary: α-Alkylation of Extended Enolates To summarise: α-alkylation of extended enolates of esters, acids, and aldehydes with alkyl halides or Michael acceptors can be accomplished reliably by kinetic methods. These reactions are particularly suited to producing quaternary centres between C⫽O and C⫽C functional groups. Aldol reactions are not so reliable, but can be controlled by temperature in some cases, and by Wittig approaches in others. We now need to look at ways of getting reaction at the γ-position.

Reaction in the γ-Position The most obvious methods of activating the γ-position would involve enamines19 and silyl enol ethers.20 These are the more stable of the various specific enol equivalents (chapter 10) and thus are more suited to thermodynamic control. Extended enamines 55 are easy to make and are excellent electron-rich dienes for the Diels-Alder reaction but react in the α-position with most alkylating agents. Me3SiCl

OSiMe3

HN

CHO +

N

Et3N

54

35

55

Silyl enol ethers from aldehydes 54 and esters 57 are reagents that give clean reaction at the γ-position with a variety of electrophiles 58. Acylation occurs under Friedel-Crafts conditions, and the crowded esterifying group (i-Pr)2CH in 59 ensures that reaction occurs entirely at the γ-position. The product is a mixture of double bond positional isomers 61 as both are conjugated with one of the carbonyl groups. OMe

OMe

OMe E

1. LDA

O

OSiMe3

2. Me 3SiCl

56 (cf. 11)

E 58

57; 93% yield

O

OSiMe3

O

1. LDA

O

O

Lewis acid

O

MeCOCl

OCHi-Pr2

2. Me 3SiCl

OCHi-Pr2

ZnBr2

60

59

61

It is more difficult to make alkylation reversible but Fleming and Paterson achieved this by using a PhS group to stabilise even a primary cation.21 An alkyl halide is converted into a sulfide 62 and then into a chlorosulfide 63 with N-chlorosuccinnimide (NCS) by the Pummerer reaction. These compounds 63 react with Lewis acids to give sulfur-stabilised cations that react in the γ-position 64 with silyl enol ethers 65 to give 66. Cl

O

SPh PhSH

N

Cl

R

NaOH

SPh NCS

R

R 62

SPh

ZnCl2

Cl

OR OSiMe3

R

63

64

O NCS N-chlorosuccinimide

OR OSiMe3 65

ZnCl2

Raney nickel

SPh

63

CO2R

R 66

CO2R

R 67

160

11 Extended Enolates

The PhS group may be removed in two ways: reductively with Raney nickel to give 67, the product of γ-alkylation of 65 with the alkyl halide, or oxidatively by thermal decomposition of the sulfoxide 68 (chapter 32) to give the new double bond in 69, the product of an extended aldol reaction, like the ester 30. Ph

SPh

[O]

CO2R

R

e.g. NaIO4

S

O heat

CO2R

R

66

CO2R

R

68

69

The true extended aldol reaction, the combination of an extended enolate in the γ-position, with an aldehyde or ketone, can best be realised by combining a silyl enol ether 54 with an acetal 70 under Lewis acid catalysis.20 The Lewis acid, usually TiCl4, catalyses the formation of the oxonium ion 71 which adds in the γ-position to the silyl enol ether [cf. 64] to give the adduct 72 from which the remaining OMe group can be removed with base to give the dienal 73, the extended aldol product. OMe R

OMe

OMe

OMe

TiCl4

R

70

OSiMe3 71

base

CHO

R

CHO

R

72

54

73

An excellent example is the synthesis of the visual pigment retinal 74. An aldol disconnection 74a is possible but four sequential aldols would be needed. But only two extended aldol disconnections 74b and 75 are needed and both require the same extended enolate22 77. b

a CHO

b

CHO

75

74; retinal

CHO

76

CHO 77 γ-extended enolate

The starting material is natural β-cyclocitral 76, converted to its methyl acetal 78. The extended enolate 77 is best realised as the silyl enol ether 79 and reaction with 78 followed by base gives the trienal 75. Repetition of the three steps gives all E retinal 74. The two extended aldol reactions go in excellent yield with total regioselectivity, only the base-catalysed eliminations with DBU are disappointing. OMe

OMe

79 76

HC(OMe)3

OMe

MeOH

CHO

OSiMe3 TiCl4

80

78; 100% yield OMe DBU

80

75 56% yield from 78

HC(OMe)3

1. 79, TiCl 4 (80% yield)

OMe MeOH

2. DBU (59% yield)

81; 95% yield

TM 74 retinal

11 Extended Enolates from Unsaturated Ketones

161

Recent developments include an asymmetric version of the aldol reaction in the γ-position. The silyl enol ether 82 from ethyl crotonate reacts with aldehydes in the presence of the Lewis acid SiCl4 and an asymmetric catalyst (see the paper if you are interested in the structure of this complex catalyst) to give a high yield of the ‘aldol’ product 83. The γ:α ratio is ⬎99:1 and the product 83 is virtually enantiomerically pure.23 OSiMe2t-Bu

O OEt

OH

PhCHO, SiCl 4

OEt

Ph

catalyst

OEt

82

ethyl crotonate

O

83; 89% yield

Extended Enolates from Unsaturated Ketones With ketones 82 a further dimension of regioselectivity is present. The extended enolate 83 can be formed in the same way by the loss of the γ-proton, and it can react in either the α or the γ positions to give conjugated 87 or 88, or unconjugated 84 products. But the ketone may also be able to enolise on the other side of the carbonyl group, usually called the α´ position, to give a new enolate 85 which can react at the α´ position to give another conjugated product 86. There are four possible products in all. The kinetic and thermodynamic factors governing the regioselectivity of the reactions of the extended enolate are as before, but the enolisation itself is now also subject to such factors. The protons in the α´ position are more acidic than those in the γ position and so 85 is the kinetic enolate, while the linearly conjugated extended enolate 83 is the thermodynamic enolate rather than the cross-conjugated and therefore less stable enolate 85.

H

removal of γ-proton

H γ

α

E

α′

γ

82

α

α

E

O

γ

O

O

O E

α′

E

86; α′ product

85

E 84; α product

83

removal of α′-proton

α′

O

O

O

E 88; α product

87; γ product

Kinetic enolisation and reaction at the α´ position are easy to control by lithium enolates. Hence cyclohexenone reacts with LDA to give the enolate 90 which is alkylated at the α´ position to give 91 but silylated on oxygen24 to give 92, much as expected for simple lithium enolates (chapter 2). O α'

OLi LDA

89

O RX

90

OLi

R

OSiMe3 Me3SiCl

91

90

92

A spectacular example is the synthesis of the bicyclic keto-ester 93. Two consecutive Michael disconnections reveal the α´ extended enolate of cyclohexanone 95 and methyl acrylate. The synthesis is even easier: kinetic enolisation of cyclohexenone with LDA and reaction with methyl

162

11 Extended Enolates

acrylate gives the cage ketone 93 directly in 90% yield.25 Reaction of cross-conjugated lithium dienolates such as 95 with Michael acceptors tends to give Diels-Alder adducts (see below) and this may be the mechanism of the addition. 5 3

2

O

O

1,5-diCO

4

1 CO2Me 93

3

= MeO2C 1 2

CO2Me 94

O

O 1,5-diCO

4 5

MeO2C

+

94a

95

Extra selectivity is needed in the synthesis of the open chain compound 96. This is clearly an aldol, but analysis reveals the need for two forms of regioselectivity in that an α´ extended enolate 97 is required to add 1,2 to a conjugated aldehyde 35. Fortunately aldehydes tend to prefer 1,2 to 1,4-addition (chapter 9), especially with hard nucleophiles such as lithium enolates, and this synthesis works as planned.26 1

2

1,3-diO

3

+

H

aldol

O

OH

O

96

O

97 LDA

H

+

O

OLi

TM 96 O

99

98

35

35

Under more equilibrating conditions such as alkoxide bases in alcohol solution or amide bases in liquid ammonia, enolisation occurs to give the extended enolate 83 which is then alkylated in the α-position by alkyl halides. At first this seems the most difficult combination to achieve: thermodynamic enolisation followed by kinetically controlled addition of an electrophile, but it is in fact a common result achieved with a variety of bases. Examples include the synthesis of pentethylcyclanone 100, an anti-tussive drug, by alkylation of the enone 103, the aldol dimer of cyclopentanone. Disconnection at the branchpoint to the available alkyl halide 102; X ⫽ Cl requires α-alkylation of the extended enolate 101 derived from the cyclopentanone aldol dimer27 103. This is easily achieved by sodium amide in toluene.28 O

X C–C

O

1,2-diX

+

O

HN O

N

alkylation of enolate

N

+

O

O 100

101

102 Cl

O

base

O

H

or acid

NaNH2

O

+

toluene

TM 100 N O

103

101

102; X = Cl

11 Extended Enolates from Unsaturated Ketones

163

Reaction in the α- or γ-positions (i.e. from the extended enolate) is again best controlled by enamines or silyl enol ethers, both being formed under equilibrating conditions.20,29 Enones such as 104 give the enamine 105 and the silyl enol ether 107 from which the lithium enolate 108 can be made. Both these intermediates give α-alkylated products 106 or 109. Direct reaction of 104 with LDA would of course give the α´ lithium enolate. NH 1. RHal

O

N

H, reflux

O

2. H 3O

R 104 Me3SiCl

105

106

Et3N, ZnCl 2

MeLi

Me3SiO

RX

LiO

O R

107

108

109

Alkylation of the lithium enolate 108 occurs in the α-position to give 109. This may remain unconjugated, but in the next example 113, became conjugated.30 Aza-enolates and hydrazones achieve α-alkylation too.31 Reaction in the γ-position of silyl enol ethers requires the same methods as were used for aldehydes and esters above. Ot-Bu

Ot-Bu NaH

+

DMSO

O

Ot-Bu O

O

MeO2C

Br

O 110

111

O

112

113 O

MeO2C O

A good example of enolate control comes in Stork’s synthesis of abietic acid.32 The first two reactions each involve regioselectivity of enolates in the formation of 115 and 117.

1. N H

NaOEt O

2. MeI 3. H, H 2O

O 114

NEt2

O 115

116

O 117

The next stage requires formation of an extended enolate 118 and its reaction with ethyl bromoacetate. Reactions of enolates with α-bromoesters are not normally recommended as the basic enolates may remove the rather acidic proton between the ester and the bromine atom. The extra conjugation in the extended enolate makes it significantly less basic and this reaction goes

164

11 Extended Enolates

cleanly in the α-position and on the opposite face to the axial methyl group. The rest of the synthesis is described in Fleming, Selected Organic Syntheses.33

KO t-Bu EtO2C

O

Br

O

O EtO2C

117

118

119

Diels-Alder Reactions The extended enolates themselves as well as the enamines and the silyl enol ethers derived from them are all good electron-rich Diels-Alder dienes showing excellent regioselectivity in reactions with electron-deficient dienophiles. Thus extended silyl enol ethers such as 54, 60, and 92 are also 1-substituted butadienes.34 A typical reaction is the cycloaddition of 120 with the alkyne 121 to give a single regioisomer of the adduct 122 in high yield. This compound was used by Schlessinger35 in his synthesis of senepoxide 123. OSiEt3

120 (cf. 54)

OAc OAc

CO2Et

80 °C

+

PhCO2

OSiEt3

CO2Et

O

121

122; 87% yield

123; senepoxide

Enamines such as 55 are even more reactive19 and 124 reacts regiospecifically with the unsaturated ester 125 to give a single regio- and stereo-isomer of 126 while no 1-RO butadiene will do so. NEt2

NEt2 CO2Me

CO2Me

+

124

125

126

The 2-silyloxybutadienes derived from the α´ enolates such as 127 are in some ways even more interesting.20 They add equally well to dienophiles such as 128 with the expected “para” regioselectivity, but the products are now specific enol equivalents of cyclic ketones: it would be difficult to make 129 from 130 in any other way. CO2Me

CO2Me

CO2Me

+ Me3SiO

Me3SiO 127

128

O 129

130

11 Extended Enolates from Birch Reductions

165

Extended Enolates from Birch Reductions Simple Birch reduction of benzoic acid 131 gives initially an intermediate that can be represented as dianion 132. Proton transfer from CO2H to the less stable of the two anions gives the enolate 133 that you will recognise as a (doubly) extended enolate with one α- and two γ-positions. Alkylation occurs at the α-position to give 134. This is a useful way to construct a quaternary centre on a six-membered ring: the two remaining alkenes can be further developed in many ways.36 O CO2H

CO2H

Li, EtNH 2

γ

α

R

CO2

RX

O

or Na, NH 3(l)

γ 131

132

133

134

Functionalised alkyl halides can be used and substitution on the benzene ring is all right. Here are two examples giving the protected ketone 136 and the diacid 137 in good yields.37 CO2H O

Li, NH 3(l)

CO2H

O

CO2H

Li, NH 3(l)

CO2H

131 O

Br

O

CO2H

Br

135

136; 94% yield

137; 79% yield

Birch reduction of aromatic heterocycles is equally rewarding if more challenging mechanistically. The pyridine diester 138 is reduced to an intermediate that could be drawn as 139. Both anions are extended enolates but one has the charge delocalised onto the nitrogen atom and so is less reactive than the other. Alkylation occurs at the α-position38 to give 140. CO2i-Pr

CO2i-Pr

CO2i-Pr

1. Na, NH 3(l)

i-PrO2C

2. RX

O

N

3. NH 4Cl

N

i-PrO2C

Oi-Pr 138

139

N R H

140; 90-99% yield

Asymmetric versions of the Birch reduction are now appearing (chapter 28) and a C2 auxiliary attached to the furoic acid 141 allows Birch reduction and alkylation between the ring oxygen atom and the carbonyl group to give, after hydrolysis, enantiomerically pure acids39 144. OMe

OMe 1. Na, NH 3

O

CO2H

N O OMe

O 141

2. RX

142

6M HCl

N O

R

O

H2O

OMe

143; 57-98% yield

O

CO2H R

144; >96% ee

166

11 Extended Enolates

The Baylis-Hillman Reaction The disconnections for the aldol reaction in the α-position on an extended enolate would be something like this: first move the alkene from conjugated 145 to unconjugated 146 so that aldol disconnection gives the extended enolate 147 derived from the enone 89. This chemistry demands at least one hydrogen atom in the γ-position of the enone 89 so that formation of the extended enolate 147 is possible. That hydrogen is of course replaced when the conjugated product 145 is formed from 146. OH

O

OH FGI

R

O

O aldol

R

145

R

O

O

147

89

H +

146

We might prefer the simpler disconnection 145a as this gives the starting materials directly. But supposing the aldol product 148 has no γ-hydrogens? No extended enolate can be formed from the enone 149, the reaction is impossible and it appears that the disconnection is meaningless. O

OH

O R

O

aldol

O

145a

O R2

R1

R

+ H

OH

89

? aldol

O

R1

148

+ H

R2

149

Fortunately this is not so and the Baylis-Hillman reaction40 allows an escape from these problems. The enone is mixed with the aldehyde and a catalyst, usually a tertiary amine or phosphine, is added. The catalyst need not be basic as its role is reversible conjugate addition 150. This gives the enolate, as we saw in chapter 10, that does the aldol addition 151. The product 152 forms a new enolate that eliminates 153 (E1cB) the catalyst to give the product 148. O

O

O

R1

O R2

R1 NR3

H

O R2

R1

NR3

150

O

OH R2

R1

NR3

151

O

OH R2

R1

NR3

152

153

148

The best catalysts are usually DABCO 154 or hydroxyquinuclidine 155 among amines and tricyclohexylphosphine 156. Efforts continue to find effective asymmetric catalysts.41 Examples with acrylates, enolisable aldehydes, and the amine catalysts are 158 and 159. OH

N =

=

N N

N

N

154; DABCO O MeO 157

O

OH

155

H

3P

155; 3-hydroxyquinuclidine

O +

OH N

MeO 158; 90% yield

156 tricyclohexylphosphine O

O 157 + H

OH

154

Pr

MeO

Pr

159; 87% yield

11 The Synthesis of Mniopetal F

167

Recent developments to make the reaction more efficient have included using stoichiometric amine in mixed organic aqueous solvents and a notable success was achieved with aromatic aldehydes to give e.g. 161 and enolisable aldehydes to give e.g. 162 in high yields under mild conditions.42 O

N

OHC

OH

OH N

1 equivalent 154

157 +

MeO

MeO2C

NHBoc

1:1 dioxan:water 1 hour, room temperature

162; 99% yield

161; 100% yield

160

Though some of these reactions use enolisable aldehydes, none uses enolisable Michael acceptors so the question of competition of this α´ reaction with extended enolisation proper has not arisen. One alternative approach uses an indium-mediated reaction of the γ-brominated ester 163 with aldehydes to give either non-conjugated 165 or conjugated 166 α´-products.43 MeO2C

MeO2C

OH

In(0)

MeCHO

OH DBU

MeO2C

MeO2C

THF/H2O

Br

InBr

163

164

165

166

One genuine case where the enone can form both normal (α´) and extended enol(ate)s is cyclohexenone 89, the compound we used to introduce this section. Recent results show that BaylisHillman reaction with this enone gives good yields of Baylis-Hillman products 167 providing stoichiometric 155 is used in solution in water or formamide (NH2CHO) with Yb(OTf)3 catalyst. The aldehydes can be aromatic 167a, aliphatic 167b or even formaldehyde 167c most easily used in aqueous solution. Under these conditions the reaction is quite fast.44 OH

O

OH

O

OH

O

O

OH

O

RCHO 1 equiv 155

R

Ph

Yb(OTf) 3 NH2CHO or H2O

89

167

167a; 82% yield

167b; 59% yield

167c; 99% yield

The Synthesis of Mniopetal F We conclude with two synthesis that include both γ- and α´-selective extended enolate chemistry.45 The first is of of mniopetal F. The Horner-Wadsworth-Emmons reaction of 168 (cf. 29) prepared by γ-bromination of an extended enol with the appropriate aldehyde gives the triene 169 with good (⬎20:1) E:Z selectivity and of course total γ-selectivity. This is converted into 170. O (MeO)2P

CO2Me 2.

168

CO2Me

1. LiHDMS

CHO

OR

CHO

169; 85% yield

170; R = SiPh 2t-Bu

168

11 Extended Enolates

A Baylis-Hillman reaction with the chiral butenolide 171 goes in excellent yield and stereoselectivity with LiSPh as the catalyst. Sulfur is of course good at conjugate additions. The product 172 is clearly destined for a Diels-Alder reaction and the product 173 is nearly ideal for conversion into mniopetal F 174. OR* O OR* O

OR*

O

O 140 °C

170

OH

PhSLi

OR

O HO

OH

O O HO

H

H

OR

xylene

CHO

O H

H 171; R* = (+)-menthyl

172; 88% yield

174; mniopetal F

173; 68% yield

A Synthesis of Vertinolide Using α ´ and γ-Extended Enolates A recent synthesis of the natural tetronic acid vertinolide 175 illutrates several aspects of extended enolate chemistry. An enone disconnection suggests an aldol reaction between the methyl ketone 176 and crotonaldehyde 35. The ketone 167 might be made by conjugate addition of a reagent for the γ-extended enolate 178 to butenone46 177. protect HO

HO

RO

aldol

O

O

O

[+ 35]

O 175; vertinolide

+

O

O 176

177

O

O

O

178; d 3 synthon γ-homoenolate

The tetronic acid 181 is easily made by bromination of the ketoester 179 and treatment of 180 with KOH. Protection with the MOM group (MeOCH2-) gives 182 from which an extended enolate might be made. O

O

HO

1. Br 2, CHCl 3

MOMO

KOH

MOMCl

2. air

CO2Et 179

Br

CO2Et 180

O

O

i-Pr2NEt

181; 67% from 167

O

O

182; 99% yield

Treatment of 182 with LDA gives the lithium enolate 183. This is an unusually interesting extended enolate. It has an obvious α-position at C2 but we want reaction to occur at the γ-position C4. The complexity arises from the other two oxygen atoms. The ring oxygen makes C2 and C3 both nucleophilic and the MOMO group makes C4 nucleophilic as all these connections are enol ethers. Finally the ring is a furan. The enolate is therefore very stable and thermodynamic control a possibility. In the event, conjugate addition occurred with butenone 177 at C4 as hoped and in good yield to give 184.

11 A Synthesis of Vertinolide Using α´ and γ-Extended Enolates

MOMO

169

MOMO

MOMO 3

2

4 O

1 OLi

O

LDA

H

O

O

O

O

177

O

182

183

184; 75% yield

The aldol reaction of 184 with crotonaldehyde 35 was rather unsatifactory as a mixture of aldol 185 and dienone 186 was formed. The aldol could be converted into the enone with basic alumina but the yield was only 61%. An alternative was needed. MOMO

MOMO

1. LDA

184 2.

O +

O

CHO

OH

35

O

O

O

O 186; 24% yield

185; 63% yield

Since the conjugate addition of the extended enolate 183 works so well, an attractive alternative is to disconnect the whole side chain and attempt conjugate addition of 183 to the trienone 187. There is an obvious problem of regioselectivity with 187. HO MOMO +

O

O

O

OLi

O

O 187

175a; vertinolide

183

The trienone 187 could be made by Lewis acid catalysed addition of the α´ silyl enol ether 188 from butenone 177 to the diethyl acetal of crotonaldehyde and elimination with basic alumina. α′

OEt

Me3SiCl

O

OSiMe3

177

basic alumina

OEt

OEt

TiCl4

188

O

187

189

Conjugate addition of the γ-extended enolate 183 to the trienone 187 went in excellent yield and gave the complete skeleton of vertinolide in one step. Deprotection gave the natural product in a remarkably short and efficient way. MOMO

MOMO

O 183

O

OLi 187

O O 186; 97% yield

O

HCl AcOH

175; 94% vertinolide

170

11 Extended Enolates

Conclusion: Extended Enolate or Allyl Anion? We have considered a variety of related intermediates in this chapter as extended versions of enolate ions 191 and as dienes. Another legitimate way to consider them is as acylated allyl anions 191c, and this leads us to the next chapter where we consider how to use allyl anions in synthesis. O

O

O

O

base

190

191a

191b

191c

References General references are given on page 893 1. M. Julia, Rec. Chem. Progr., 1963, 24, 43; House, chapter 9. 2. Fleming; Orbitals. 3. R. Gompper and H.-U. Wagner, Angew. Chem., Int. Edn. Engl., 1976, 15, 321. 4. M. W. Rathke and D. Sullivan, Tetrahedron Lett., 1972, 4249; J. L. Herrmann, G. R. Kieczykowski, and R. H. Schlessinger, Tetrahedron Lett., 1973, 2433; A. Kajikawa, M. Morisaki, and N. Ikegawa, Tetrahedron Lett., 1975, 4135. 5. A. B. Smith, S. J. Branca, and B. H. Toder, Tetrahedron Lett., 1975, 4225; T. J. Brocksom, M.. Constantino, and H. M. G. Ferraz, Tetrahedron Lett., 1977, 483. 6. W. Oppolzer and R. Pitteloud, J. Am. Chem. Soc., 1982, 104, 6478. 7. P. M. Savu and J. A. Katzenellenbogen, J. Org. Chem., 1981, 46, 239; W. Oppolzer, Tetrahedron Lett., 1982, 23, 4673. 8. S. Johnson and L. A. Bunes, J. Am. Chem. Soc., 1976, 98, 5597; M. B. Gravestock, W. S. Johnson, R. F. Myers, T. A. Bryson, D. H. Miles and B. E. Ratcliffe, J. Am. Chem. Soc., 1978, 100, 4268, W. S. Johnson, B. E. McCarry, R. L. Markezich and S. G. Boots, J. Am. Chem. Soc., 1980, 102, 352; E. J. Corey and T. Hase, Tetrahedron Lett., 1979, 335. 9. M. Majewski, G. B. Mpango, M. T. Thomas, A. Wu and V. Snieckus, J. Org. Chem., 1981, 46, 2029. 10. R. W. Dugger and C. H. Heathcock, J. Org. Chem., 1980, 45, 1181; G. Cainelli, G. Cardillo, and M. Orena, J. Chem. Soc., Perkin Trans.1, 1979, 1597; G. Cardillo, M. Orena, G. Porzi and S. Sandri, J. Org. Chem., 1981, 46, 2439. 11. W. R. Roush, J. Am. Chem. Soc., 1978, 100, 3599; W. Gärtner, D. Oesterhelt, P. Towner, H. Hopf and L. Ernst, J. Am. Chem. Soc., 1987, 103, 7642; P. J. Reider, P. Davis, D. L. Hughes and E. J. J. Grabowski, Tetrahedron Lett., 1976, 955; W. R. Roush, H. R. Gillis and S. E. Hall, Tetrahedron Lett., 1980, 21, 1023. 12. E. J. Corey and B. W. Ericson, J. Org. Chem., 1974, 39, 821. 13. R. K. Boeckmann and S. S. Ko, J. Am. Chem. Soc., 1982, 104, 1033. 14. J. K. Whitesell and M. A. Whitesell, Synthesis, 1983, 517 (see page 528). 15. G. R. Kieczykowshi, R. H. Schlessinger and R. B. Sulsky, Tetrahedron Lett., 1976, 597. 16. E. Vedejs and D. M. Gapinski, Tetrahedron Lett., 1981, 22, 4913. 17. M. Bellasoued and M. Salemkour, Tetrahedron, 1996, 52, 4607. 18. H. Ido, M. Isobe, T. Kawai and T. Goto, Tetrahedron, 1979, 35, 941. 19. P. W. Hickmott, Tetrahedron, 1984, 40, 2989. 20. P. Brownbridge, Synthesis, 1983, 85. 21. I. Fleming, J. Goldhill, and I. Paterson, Tetrahedron Lett., 1979, 3209; I. Paterson and L. G. Price, Tetrahedron Lett., 1981, 22, 2829, 2833. 22. T. Mukaiyama and A. Ishida, Chem. Lett., 1975, 1201. 23. S. E. Denmark and G. L. Beutner, J. Am. Chem. Soc., 2003, 125, 7800. 24. K. B. White and W. Reusch, Tetrahedron , 1978, 34, 2439.

11 References

171

25. R. A. Lee, Tetrahedron Lett., 1973, 3333. 26. G. Stork, G. A. Kraus and G. A. Garcia, J. Org. Chem., 1974, 39, 3459. 27. A. T. Nielsen and W. J. Houlihan, Org. React., 1968, 16, 1: see table on page 114. 28. H. Ueberwasser, CIBA, Ger. Pat., 1,059,901, (1959), Chem. Abstr., 1962, 56, 355f. 29. G. Stork, A. Brizzolara, H. Landesman and J. Szmuszkovicz, J. Am. Chem. Soc., 1963, 85, 207. 30. Z. J. Hajos, R. A. Micheli, D. R. Parrish and E. P. Oliveto, J. Org. Chem., 1967, 32, 3008; W. Nagata, T. Terasawa and T. Aoki, Tetrahedron Lett., 1963, 865; W. G. Dauben, G. Ahlgren, T. J. Leitreg, W. C. Schwatzel and M. Yoshioko, J. Am. Chem. Soc., 1972, 94, 8593; W. Oppolzer, Helv. Chim. Acta, 1979, 62, 1493. 31. G. Stork and J. Benaim, J. Am. Chem. Soc., 1971, 93, 5939. 32. G. Stork and W. Schlulenberg, J. Am. Chem. Soc., 1962, 84, 284. 33. Ian Fleming, Syntheses, pp 76–79. 34. R. H. Schlessinger and A. Lopes, J. Org. Chem., 1981, 46, 5252. 35. G. A. Berchtold, J. Ciabattoni and A. A. Tunick, J. Org. Chem., 1965, 30, 3679. 36. P. W. Rabideau and Z. Marcinow, Org. React., 1992, 42, 1. 37. C. P. Chuang and D. J. Hart, J. Org. Chem., 1983, 48, 1782. 38. T. J. Donohoe, A. J. McRiner and P. Sheldrake, Org. Lett., 2000, 2, 3861. 39. T. J. Donohoe, A. A. Calabrese, C. A. Stevenson and T. Ladduwahetty, J. Chem. Soc., Perkin Trans. 1, 2000, 3724. 40. E. Ciganek, Org. React., 1997, 51, 201. 41. P. Langer, Angew. Chem., Int. Ed., 2000, 39, 3049. 42. C. Yu, B. Liu and L. Hu, J. Org. Chem., 2001, 66, 5413. 43. J. H. Cha, A. N. Pae, K. I. I. Choi, Y. S. Choi, H. Y. Koh and E. Lee, J. Chem. Soc., Perkin Trans. 1, 2001, 2079. 44. V. K. Aggarwal, D. K. Dean, A. Mereu and R. Williams, J. Org. Chem., 2002, 67, 510. 45. J. Jauch, Eur. J. Org. Chem., 2001, 473. 46. K. Takabi, N. Mase, M. Nomoto, M. Daicho, T. Tauchi and H. Yoda, J. Chem. Soc., Perkin Trans. 1, 2002, 500.

12 Allyl Anions Introduction: Allyl Grignard Reagents The [1,3] allylic shift Allylic Lithiums and Grignard Reagents Allylic lithiums with C-2 substituents Allyl Nickel Complexes Alkylation and a cyclopentenone synthesis Allyl Silanes Synthesis: silylation, Wittig, and cycloaddition reactions The stabilisation of cations by silicon Reactions: alkylations, reactions with epoxides and aldehydes, conjugate additions Heterocyclic synthesis with allyl silanes Reactions with Co-stabilised cations An Allyl Dianion? The Role of Tin in Anion Formation Halide Exchange with Chelation: Indium Allyls Allyl Anions by Deprotonation The synthesis of all-trans dienes The synthesis of all-trans retinol

Introduction: Allyl Grignard Reagents A Grignard reagent 2 can be made from allyl bromide 1 and it reacts cleanly with electrophiles such as carbon dioxide to give the unconjugated acid 4. So also do vinyl halides such as propenyl bromide 5, giving the conjugated acid 8. It is important that you understand the difference between these two series. When the halide is directly attached to a double bond 5 we have a very unreactive vinylic halide and a well-behaved Grignard reagent 6. We have already used these intermediates in chapter 5 and we shall meet them in more detail in chapter 16. When the halide is joined, not directly to the double bond, but to the atom next door we have a reactive allylic halide 1 and a badly behaved Grignard reagent 2. Controlling the behaviour of allyl anions and metal derivatives is the subject of this chapter. Br

Et2O

1 Br 5

Mg

MgBr 2

Mg Et2O

CO2

3 MgBr

6

CO2MgBr

CO2

CO2MgBr 7

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

H H2O

CO2H 4

H H2O

CO2H 8

174

12 Allyl Anions

So what exactly is wrong with allyl Grignard reagents? Even if they are symmetrical 2, yields tend to be low as the allyl metal bond cleaves easily to give relatively stable allyl radicals that may dimerise or polymerise. When the allyl group is unsymmetrical 9 reaction often occurs at the “wrong” end of the allyl system, here to give the acid 10 and eventually the tosylate 11 needed in chapter 10 for the synthesis of multistriatin. Br

1. Mg, Et 2O

1. LiAlH 4

2. CO 2

2. TsCl, pyridine

CO2H

9

OTs 11

10

The reaction may occur through a six-membered cyclic transition state 12 that directs the electrophile to the other end of the allylic system. This is bad enough, but worse is to come.

1. Mg, Et 2O

9

2. CO 2

O

H

Mg

C

Mg Br

O 12

O

O

H2O

O

Br

OH

13

10

The [1,3] allylic shift Very few allylic derivatives are stable: only H, C, O, N, Si, and F are secure from a [1,3] shift. Halogens may move from one end to the other of the allylic system by an ionic pathway via the delocalised allylic cation 15 or by a radical chain pathway via the dibromoalkyl radical 19. These are both equilibrium reactions and the dominant component is the one with the more heavily substituted double bond 16 or 20.

Br

Br Br 14 R

hν

15 R

16 Br

R Br

Br 17

Br

R

Br

Br 18

19

20

Metals such as Li or Mg may move by way of the η3 allyl complex 22 from one end 21 to the other 23 of an allylic system. This η3 allyl complex has the metal sitting roughly at the middle of the triangle formed by the three carbon atoms of the allyl system, and at right angles to them, so that the orbitals of the metal may overlap with all three p-orbitals of the allyl group 24. This becomes a stable compound rather than an intermediate when M is a transition metal. R

R

MgBr

R

R MgBr 21

MgBr 22

23

M 24

12 Allylic Lithiums and Grignard Reagents

175

If neither the halides nor their metal derivatives have a stable structure, how can we expect to control their reactions? Things are not quite as bad as that. The bromides and chlorides are stable at low temperatures in the absence of Lewis acid catalysts or radical generators. The Grignard reagents often react cleanly to give the less substituted double bond isomer, presumably by the type of process shown in 12. We shall be seeing methods to control transition metal allyls and allyl silanes which make these intermediates valuable reagents in synthesis, and you should finish this chapter in a confident mood without fear of fluxional allyl compounds.

Allylic Lithiums and Grignard Reagents Before looking at alternative solutions, let us review the situation with allyl Grignards and allyl lithiums. Allyl Grignard 2 has a lower basicity than propyl Grignard and it reacts well with aromatic ketones to give tertiary alcohols 26 without much enolisation.1 Reactions with conjugated enones and enals occur by direct (1,2) rather than Michael (1,4) addition, as in the reaction with acrolein2 to give 25. Allyl-lithium 28 also gives reasonable yields with enolisable ketones, as in the adduct3 31. The reaction of allyl-lithium with the epoxide 29 to give the anti adduct 27 illustrates regioselectivity as well as stereospecificity.4 Reaction at the allylic end of the epoxide occurs because of the superior reactivity of allylic electrophiles in the SN2 reaction. OH

1. Mg, Et 2O

OH

1. Mg, Et 2O

Br 2. OHC

Ar

2. ArCO.Me

25; 57-59% yield

26

1

OH

Li

Li

O O

28

28

29

27

OH

30

31

Turning to unsymmetrical allyl derivatives, the halides 9 and 32 are usually used as an 80:20 mixture. Both give the same Grignard reagent that reacts with CO2 to give 10 and with aldehydes and ketones with the same regioselectivity5 to give alcohols such as 33. Another unsymmetrical allylic halide, existing only in the form 35 with the trisubstituted double bond inside the sixmembered ring, reacts with ethylene oxide at the more substituted end to give the alcohol 36. Presumably these unsymmetrical allylic Grignard reagents prefer the form with the more highly substituted double bond and react via a six-membered cyclic transition state such as 12. In some cases the regioselectivity can be reversed by exchanging Li or Mg for copper. The allylic Grignard derived from 35 gives only the alternative alcohol 34 when Cu(I) salts are added before the epoxide.6 Br 1. Mg, Et 2O

HO

Br 2. Me 2CO

9 (80 parts) OH

32 (20 parts) 1. Mg, Et 2O 2. Cu(I)

Br

1. Mg, Et 2O 2.

3.

34

33

O

35

O

OH 36

176

12 Allyl Anions

The unsymmetrical copper allyl derived from prenyl halides 38 and 39, the basic unit from which terpenes are constructed, can be alkylated at the less substituted end to give unrearranged 37, though the corresponding Grignard reagent gives a mixture of products.7 The corresponding lithium derivative, made from prenyl bromide 39, reacts with aldehydes with the reverse regioselectivity.8 The copper derivatives are probably η3 allyls and react at the less hindered end, as do the nickel allyls soon to be discussed. 1. Mg

R

1. Li, THF

Cl

2. Cu(I)I

Br

2. t-BuCHO

OH 40

3. RX

37

38

39

The extra delocalisation in an allyl anion might tempt one to make lithium allyls by removal of a proton from an alkene. This approach is not generally very successful. Either E or Z-but-2-ene 41 can be deprotonated, but t-BuLi is needed as well as the lithium chelating agent TMEDA (Me2NCH2CH2NMe2). The lithium allyl prefers the Z configuration 42 but gives a mixture of regio and stereoisomers 43 and 44 on alkylation.9 Nor is it possible to make useful lithium allyls by addition of, say BuLi, to butadiene 45 as the resulting allyl lithium 46 adds to another molecule of butadiene to give another allyl lithium 47 and polymers are formed.10 t-BuLi, hexane

R

RX

or

+ NMe2

Me2N

E-41

Li 42; 85:15 E:Z

Z-41 RLi

45

R

R Z-44; 46% yield (+ 8% E-44)

R

Li

45

43; 46% yield

Li

46

47

Allylic lithiums with C-2 substituents One way in which allyl lithiums can be made by direct deprotonation is when the allyl system has a substituent not at one end but in the middle, and when the substituent is of the kind we used in chapter 7 for directing aromatic lithiation. The simplest examples are amides such as 48 that react with two molecules of BuLi to give a dilithium derivative11 often represented as an allyl complex 50b though it is probably a lithium σ-complex 50a. These amide derivatives react with aldehydes and ketones to give lactones,12 e.g. 53 via adducts such as 52. Li

O N

R

Li

O

BuLi

N

H

R

N

N

O

N

Li

50a

Me

N

O

50b O O

OLi H H2O

2. Me 2CO

52

O

or

Li 1. two equivalents s-BuLi

51

R

Li

H Me

BuLi

H 49

48

Li

O

53

R

12 Allyl Nickel Complexes

177

Similar compounds 56 can be made13 from zinc derivatives of bromoesters 54 and even the simple alcohol14 57 gives the diol 59 and hence another exo-methylene lactone 60 by oxidation. OEt

O

OEt

O

O

1. Zn

2. RCHO

O

Zn

Br 54 OH

55 Li

t-BuOK

O

t-BuLi

Li

H 57

56 OH

Me2CO

R O [O]

O

OH 59

58

60

Allyl Nickel Complexes Perhaps the most useful of the η3 allyl complexes (cf. 24) are the π-allyls of nickel.15 The simplest type 61 are rather unstable and form the bromide-bridged complex 62 on treatment with HBr. These are stable compounds officially complexes of Ni(I) but better regarded for our purpose as dimers of η3 complexes of allyl anions and Ni(II), much as allyl Grignard reagents 2 can be regarded as σ-complexes of allyl anions and Mg(II). Direct exchange of Mg(II) for Ni(II) gives the unstable complexes 61, but the stable dimer 62 can be made by oxidative insertion of Ni(0), as its cyclo-octa-1,5-diene (COD) complex, into allyl bromide 1. MgBr

NiBr2

Br

HBr

Ni

2

Ni

Ni(0)

Ni Br 62

61

Br

COD

1

These π-allyls of nickel are prone to give allyl dimers, but the more stable complexes 62 are useful for coupling with alkyl halides, the allyl group of 62 acting as an allyl anion equivalent. That this is not an SN2 reaction, but a coupling on the nickel atom, is clear because even aryl halides react successfully. Almost all the published examples involve substituted allyl groups, particularly in the 2-position. Thus the 2-methylallyl complex 64 couples with alkyl iodides or iodobenzene to give adducts, e.g. 65 in excellent yield. There is good selectivity for iodides against other halogens: thus the complex 66, derived from 54, displaces only iodide from 67, again in excellent yield.15 Br

Ni(0)

Br

Ni COD

Ni(0)

Br 54

COD

65; 98% yield

Br EtO2C

Ph

Br 64

63 CO2Et

PhI

Ni

Ni

I

Ni Br 66

Br

CO2Et

CO2Et 67

Cl 68; 96% yield

Unsymmetrical allyl complexes react almost exclusively at the less substituted end of the allylic system. Santalene 71 has been prepared from 69 and 16 in two ways. Coupling the Grignard reagent 70; MX ⫽ MgBr with 16 gives 20% yield16 while the allyl nickel complex 70; MX ⫽ NiI couples with 16 in 95% yield.17 This is a major development from the allyl derivatives of Mg and Li, which tend to react at the more substituted end of the allyl system, and are in any case less well behaved than the nickel complexes.

178

12 Allyl Anions

X

MX metal

Br +

69

70

16

71

Alkylation and a cyclopentenone synthesis Modern developments have included allyl groups functionalised at the ends, particularly those derived from silylated enals such as 72. The application shown here creates a cyclopentenone 74 by combination of the allyl nickel 73 with an alkyne and CO rather in the style of the Pauson-Khand reaction (chapter 6).18 Me3SiO

O

MeO

1. CO,

Cl

Ni(COD)2

Ni

OMe

CO2Me

Ni

Me3SiCl

O

2. MeOH

Cl

CO2Me

OSiMe3 72 acrolein

74

73

This last example succeeds partly because nickel catalyses the union of the three reagents in the second step. Similar things happen in the combination of nickel allyls with zinc borates 76. The allyls are derived from unsymmetrical allylic acetates 75 and react at the end remote from the phenyl group (R1 is an alkyl group and R2 is an aryl or alkyl group).19 O Ph

OAc R1

Me

O

+

B O

Me

75

R2 Me

NiCl2(PPh3)2

R2 Me

Ph

77; NMI

R1 78; 76-98% yield

76

N

N Me

77; NMI

Allyl Silanes The most useful of all allyl anion equivalents are the allyl silanes.20 This is because it is easy to make them regioselectively, because they do not undergo allylic rearrangement (silicon does not do a [1,3] shift) and because their reactions with electrophiles are very well controlled: addition always occurring at the opposite end to the silicon atom. Symmetrical allyl silanes can be made from allyl-lithiums or Grignards by displacement of chloride from silicon. A useful variant is to mix the halide with a metal, e.g. sodium, and Me3SiCl in the same reaction, rather after the style of the silicon acyloin reaction,21 as in the synthesis of the acetal 80. OMe

OMe

Na, Me 3SiCl

Cl

MeO

Et2O

SiMe3

MeO

79

80

Synthesis: silylation, Wittig, and cycloaddition reactions Syntheses of unsymmetrical allyl silanes from Grignard reagents are subject to the usual regioselectivity problems. The mixture of halides 81 and 82 gives a mixture of Grignard reagents,

12 Allyl Silanes

179

see 21 and 23, from which a mixture of allyl silanes 83 and 84 can be made.22 However, this mixture can be converted23 into the thermodynamic product (83, as a mixture of E and Z isomers) by TBAF (Bu4NF) at 100 ⬚C as the fluoride ion is a nucleophilic catalyst, attacking the silicon atom with excellent chemoselectivity. More highly substituted allyl halides like 38 give only the less substituted allyl silane24 85. This may seem a disadvantage, but it is quite the reverse as you will see. Cl Cl

SiMe3

1. Mg, Et 2O

+

81

+

SiMe3

2. Me 3SiCl

83; 25% yield, 3:2 Z:E

82

84; 75% yield

1. Mg, Et 2O

Cl

SiMe3

2. Me 3SiCl

38

85

One reliable method for getting the silicon on the less substituted end of the allylic system is the Wittig approach of Seyferth.25 The unsubstituted reagent 86 reacts well with ketones to give allyl silanes of the type 87. Alternatives depend on cycloaddition reactions: 1-trimethylsiclylbutadiene 88 gives Diels-Alder adducts, e.g. 89 with maleic anhydride. Direct silylation of the cyclopentadienyl anion 91 gives 92, one of the few substituted cyclopentadienes which is useful in cycloadditions. A 2⫹2 cycloaddition with dichloroketene26 gives the cyclobutanone 93 which can be dechlorinated with zinc. Both 93 and 94 are allyl silanes.27 O SiMe3 Me3Si

1. base

Ph3P

SiMe3

H Me3Si

O

87

O

88

SiMe3

Me3Si

Cl

H

H 1. base

Cl

2. Me 3SiCl

91

H

89 Me3Si

O

COCl

H

O

O

Zn

Cl

Et3N

90

O

2. cyclohexanone

86

O

H 93

92

Cl

H 94

The stabilisation of cations by silicon Even if silicon chemistry is new to you, you should by now have a picture of stable compounds with C-Si bonds and selective reaction with fluoride. You are already familiar with silyl enol ethers as nucleophilic enolate equivalents and allyl silanes resemble these in many ways. The missing link is the β-silyl effect. A Si atom stabilises a cation in the β-position by overlap of the populated and relatively high energy C-Si σ-orbital with the empty p orbital of the cation. This overlap is already present in the preferred conformation 95a of the allyl silane 95 as an anti-bonding interaction 95b between the C-Si σ-orbital and the π orbital of the double bond. The resultant molecular orbital (the new HOMO) 95c increases the nucleophilic reactivity of the carbon atom in the γ-position. Me3Si

Me3Si α Me3Si

β

95; allyl silane

γ

Me3Si H H 95a; preferred conformation

H H 95b; σ and π bond HOMOs

σ+π

H H 95c; HOMO of molecule

180

12 Allyl Anions

Electrophilic attack on an allyl silane therefore occurs 96 away from the silicon to give the stabilised cation. A nucleophile, particularly an oxygen or halogen nucleophile such as MeOH or Cl⫺, attacks the silicon atom selectively 97 and removes it to give the allylic product 98 in which the electrophile has added regiospecifically to the far end (from Si) of the allyl silane. The β cation 97a is stabilised by a bonding overlap between the filled C–Si σ-orbital and the empty p-orbital of the cation. This is σ-conjugation or hyperconjugation. E

Me3Si

X

Me3Si

β

H

Me3Si

E

E

96

97

E

H

98

97a

Reactions: alkylations, reactions with epoxides and aldehydes, conjugate additions The range of electrophiles is very large.20 Alkylation with tertiary alkyl halides, e.g. 99, and Lewis acid catalysts allows the synthesis of molecules with two adjacent quaternary centres,28 such as 100 from 87.

SiMe3

TiCl4

+ Cl 87

100; 98% yield

99

The best alkylating agents are those which give the most stable cations: only the halide next to oxygen in 101, which can give 101a the oxonium ion 102, reacts with allyl silane29 to give 103. That the reaction involves regiospecific attack at the end of the allyl system away from silicon is shown by the oxonium ion 105 and 93 which also react stereoselectively on the outside of the folded bicyclic structure.30 Cl

Cl

Cl

Cl Me3Si

Cl

O

Cl

101

O

O

101a

102 Me3Si

MeO

Cl

MeO

H

Me3SiI

103; 90% yield H O

H 105

MeO

O

Cl

CH2 Cl

104

O

93

Cl

Cl

SnCl4

H

Cl

MeO 106; 78% yield

Epoxides generally react at the more substituted end, even in intramolecular endo-like reactions such as the cyclisation of 107, as this gives the more stable cation.31 The allyl silane must capture the epoxide in 107 as it opens and it must do so faster than the familiar rearrangement of epoxides under the same conditions (BF3).

12 Allyl Silanes

Me3Si

181

Me3Si

Et2O.BF 3

O

O

107

OH

BF3

109; 71% yield

108

There are many regiospecific reactions with aldehydes and ketones, such as the four-membered ring example 110 which gives an addition to the aldehyde 111 very like the aldol reaction with a silyl enol ether (and even uses the same catalyst, TiCl4) to give the unsaturated alcohol32 112 - presumably the key step is 113. Me3Si

Me3Si

TiCl4

+ OHC

OH 110

O

112; 86% yield

111

113

TiCl3

Additions to enones, e.g. 115 generally occur in the Michael sense giving δ,ε-unsaturated enones33 such as 116. Acetals 118 can replace aldehydes in additions with allylic transposition34 as in 117 to 119 and acylation occurs with acid chlorides as in the synthesis of the terpene artemisia ketone 122 from two C5 units with at least a passing resemblance to the biosynthesis of this irregular terpene.20,35

O O Me3Si

TiCl4

+ 114

116; 78% yield

115

CO2Me

CO2Me

Me3Si

OEt

TiCl4

+ CO2Me

Ph Ph

117

CO2Me

OEt OEt 119; 100% yield

118 O

AlCl3

O

+ Cl

SiMe3 120

121

122; 90% yield

Heterocyclic synthesis with allyl silanes Two aspects of allyl anion chemistry are combined in the synthesis of bicyclic lactams by Gramain and Remuson.36 Treatment of the unsaturated alcohol 123 with BuLi initially forms the lithium alkoxide 124 and then the allyl lithium 125, possibly better represented as 125a. The proton is removed only from the methyl group rather then the internal CH2 group. Reaction with Me3SiCl and careful acidic workup gives the allyl silane 126.

182

12 Allyl Anions

BuLi

OH

123

1. Me 3SiCl

BuLi

124

Li

OLi

OLi

125 Li

Me3Si

2. 10% H2SO4

OH

126

Li

O

125a

The allyl silane 126 is coupled with the imide 127 by a Mitsunobu procedure and one of the carbonyl groups is reduced to give the alcohol 129. Treatment with CF3CO2H now cyclises the allyl silane to give a new heterocyclic ring 130. NaBH4

126, Ph 3P DEAD

O

N H

O

CF3CO2H

MeOH

O

THF

O

N

0 ˚C

O

OH

N

N

SiMe3

SiMe3 127

O

128

130; 94% yield

129

More recent developments with allyl silanes have included the use of single enantiomers such as (R)-133 that adds in a Mannich-style reaction to give (S)-134 with about 95% enantioselectivity.37

SiMe3

ClCO2Ph

N

N

AgOTf

N

+ CO2Ph

Ph

H

CO2Ph Ph

131; quinoline

(R)-133

132

(S)-134

Reactions with Co-stabilised cations More exotic electrophiles include cobalt-stabilised cations derived the alcohol (S)-138 made by a sequence of reactions that shows the stability of allyl silanes to bases. The cuprate from Z-135 adds to a single enantiomer of the epoxide (S)-136 and the tosylate in the product (S)-137 is displaced by a Co(I) anion to give the intermediate (S)-138 as a stable orange solid.38 OH 1. Mg, THF 2. Li 2CuCl4

Cl SiMe3

Z-135

OH OTs

Co(dmgH)2py Co(dmgH)2py

3. O OTs

(S)-136

SiMe3 (S)-137

SiMe3 (S)-138

In acid solution, cobalt creates a π-stabilised cation at the site of the OH group that cyclises onto the allyl silane with the expected regioselectivity and excellent enantioselectivity. The nature of the cation stabilisation and the removal of the cobalt are discussed in the workbook. The structure of this group ‘Co(dmgH)2py’ is explained in the workbook.

12 Halide Exchange with Chelation: Indium Allyls

183

OH Co(dmgH)2py

0.3 equivs TsOH

Co(dmgH)2py

OH

THF

SiMe3 (S)-138

139

140

An Allyl Dianion? The Role of Tin in Anion Formation The prospects of inventing a reagent for the dianion synthon 141 might look remote but the reagent 142, both an allyl silane and an allyl stannane, does the job.39 Reaction with one aldehyde occurs at the end occupied by tin and the second, using a different Lewis acid, at the end occupied by silicon. The result is a tetrahydropyran 144 with 2,6-syn substituents, both being equatorial.

OH

R1CHO

Bu3Sn 141

SiMe3 142

R1

Ti(OPri)4 iPrSBEt

R2CHO

SiMe3

2

Me3SiNTf2

143

R1

O 144

R2

A commoner role for tin is to sacrifice itself in the formation of allyl anions (or allyl lithiums) with BuLi. The ‘allyl’ (it has a nitrogen atom in the backbone, so strictly an ‘aza-allyl’) tin 145 reacts with BuLi to form Bu4Sn and the allyl anion 146 that adds to the styrene 147 in a 1,3-dipolar cycloaddition to give the amide ion 148 that eliminates pyrrolidine to give the imine 149. The imine 149 is formed as a 10:1 mixture of positional isomers.40 Ph Ph

SnBu3 N

N

BuLi

N

N 147

Ph N

N

N

Bu4Sn

145

146

148

149; 88% yield 10:1 isomers

Halide Exchange with Chelation: Indium Allyls The problem with forming allyl metal derivatives by exchange of zerovalent metal with halide ions is that we often cannot be sure where the halide ion is and never sure where the metal is because of rapid [1,3] shifts, cf. compounds 21 and 23. If the allylic system has a substituent that chelates with the metal then only one derivative is likely to be formed. We need a halide at one end of the allyl group and a chelating substituent at the other. Just such compounds can be made by acylation of acrolein 72 in the presence of zinc chloride. No [1,3] shift of Br will occur as the alkene much prefers to be disubstituted and conjugated with oxygen 151 rather than unconjugated and terminal 152.

184

12 Allyl Anions

H

O O

Br

72

O

ZnCl2

+ Me

Br

O

Br

150

O O

151

152

Whichever isomer (positional or geometrical) of the indium allyl is first formed (e.g. 153) is unimportant as equilibration leads to the stable chelated version 154 that reacts with aldehydes to give, as expected, reaction at the other end of the allyl system from the metal.41 The product 155 is formed as a syn/anti mixture but the syn isomer can predominate by as much as 90:10. You may wonder why indium was used. The usual reason for indium is that its organic derivatives are stable in water but here the more important point is the low reactivity of allyl indiums. InLn

In(0)

OH

O

OH

RCHO

151

LnIn

O

O

R +

O 153

OAc syn-155

154

R OAc anti-155

Allyl Anions by Deprotonation In the next chapter we shall see that allyl anions made from allyl silanes 95 by the removal of a proton still react in the γ-position 156 but retain the silyl group in the form of a vinyl silane 157 and are reagents for the d3 synthon, the homoenolate ion. Though the Me3Si group is electrondonating it is also anion stabilising. Other electron-rich but anion-stabilising groups such as amines 158 also form anions that react in the γ-position. BuLi

Me3Si α

E

Me3Si

Me3Si

E

R2N

γ

95: allyl silane

156; reaction at C-γ

157: a vinyl silane

158: allylic amine

Allyl anions with electron-withdrawing and anion-stabilising functional groups form anions much more easily and then react (mostly) in the α-position. There are many examples of this sort of substituent including extended enolates (chapter 11), and allyl anions stabilised by phosphonium 159, phosphonate 160, sulfone 161, and cyanide 162 functional groups. All of these prefer to react in the α-position (at least kinetically) and we shall explore some of these. O Ph3P 159: allyl phosphonium ylid

Ph2P 160: allyl phosphonate anion

O O Ph

S

161: allyl sulfone anion

NC

162: allyl cyanide anion

The synthesis of all-trans dienes We shall finish this chapter by considering two ways to achieve condensation between an allyl anion and an aldehyde (or ketone) in the synthesis of dienes 163. The reagent 165 is an allyl anion stabilised by a heteroatom (Z) that can be lost with the OH group after addition to R1CHO.

12 Allyl Anions by Deprotonation

R2

R1

R1

E,E-163

185

R2

Z

+

CHO 164

E-165

The best candidates are phosphonium salts (Z ⫽ Ph3P⫹) 167, using the Wittig reaction, and sulfones (Z ⫽ PhSO2) 168, using the Julia reaction,42 as we shall see in chapter 15. Wittig reagents are easily made from allylic halides, substitution usually occurring at the less hindered end, and react with aldehydes at the site next to the Ph3P⫹ group since elimination of Ph3PO can occur only if this regioselectivity is observed (c.f. the similar solution to the γ-extended enolate problem, chapter 11).

R2

Br

Ph3P

166

1. base 2. R

167

1CHO

OH

O O Ph

R2

Ph3P

R2

S

1. base 2. R 1CHO

R2

R1

R2

R1 E,E-163 1. Ac 2O pyridine 2. Na/Hg

E,E-163

SO2Ph 169

168

These ylids are classified as “semi-stabilised” or of “intermediate” reactivity, and their stereoselectivity may be poor.43 If the stereochemistry of the double bond in the ylid (from 167) is E, this is generally retained in the product, but if it is Z, as in the ylid derived from 170, low temperatures are needed to stop rotation at the allyl ylid stage. At ⫺25 ⬚C less than 5% E-171 is formed in the synthesis of vitamin D metabolites.44 The stereochemistry of the new double bond is generally not well controlled, 1:1 ratios of E:Z are not uncommon, but Vedejs finds that phosphonium salts such as 172 with two phenyl and two allyl groups give good yields of E-dienes45 such as 173.

OH OH

1. two equivalents BuLi, –25 ˚C

PPh3

2. cyclohexanone

Z-170

Z-171 1. BuLi

2 PPh2 172

Ph

2. PhCHO

E-173

The synthesis of all-trans retinol If E,E dienes like 163 are wanted, then such Wittig reactions are ideal as the mixed products can be equilibrated to the E,E diene by addition of small amounts of radical generators such as iodine or PhSH. The commercial (BASF) synthesis of Vitamin A involves all trans retinol 174 that can be made from two different allylic ylids derived from 175 and 178 with the appropriate aldehydes 176 and 177. In both cases E,Z mixtures are formed, but equilibration with iodine gives 174 with an all E side chain.46 A different synthesis of such compounds appeared in chapter 11.

186

12 Allyl Anions

OAc

8

174a; Wittig disconnections

174; all trans- (E) retinol acetate PPh3

OAc

4

PPh3

1. PhLi

1. PhLi

174 2.

2.

OHC

OAc

OHC

176

175

OAc 177

178

Direct α-alkylation of anions of allylic sulfones 179 gives products 180 from which PhSO⫺2 can be eliminated with mild base (MeONa) to give E-dienes 181. The more usual condensation of sulfones with aldehydes is not necessary and those adducts 169 may eliminate to give vinyl sulfones instead of alkenes.47 SO2Ph 1. BuLi

SO2Ph

R

base

R

2. RCH 2Br

179

180

E-181

As part of a synthesis of sesquiterpenes, the sulfone 182 and the allylic halide 183 were coupled (allylic electrophile as well as allylic nucleophile!) to give a mixture of diastereoisomers of 184 that gave the triene 185 on elimination.48 Note that the new double bond is exclusively E. OH

OH

OH 1. two equivalents BuLi 2.

SO2Ph 182

Br

MeONa

CO2Me 183

CO2Me

CO2Me

SO2Ph 184

E-185

References General references are given on page 893 1. M. S. Kharasch and O. Reinmuth, Grignard Reactions of Non-Metallic Substances, Prentice-Hall, New York, 1954, Table VI-XVIII; J. Mazurewitsch, J. Russ. Phys. Chem. Soc., 1911, 43, 973. 2. A. J. Sisti, Org. Synth. Coll., 1973, V, 46. 3. D. Seyferth and M. A. Weiner, Org. Synth. Coll., 1973, V, 452. 4. M. Apparu and M. Barrelle, Bull. Soc. Chim. Fr., 1977, 947. 5. J. F. Lane, J. D. Roberts and W. G. Young, J. Am. Chem. Soc., 1944, 66, 543; J. D. Roberts and W. G. Young, J. Am. Chem. Soc., 1945, 67, 148. 6. G. Linstrumelle, R. Lorne and H. P. Dang, Tetrahedron Lett., 1978, 4069. 7. F. Derguini-Boumechal, R. Lorne and G. Linstrumelle, Tetrahedron Lett., 1977, 1181. 8. T. E. Stanberry, M. J. Darmon, H. A. Fry and R. S. Lenox, J. Org. Chem., 1976, 41, 2052. 9. R. B. Bates and W. A. Beavers, J. Am. Chem. Soc., 1974, 96, 5001. 10. B. J. Wakefield in Comprehensive Organometallic Chemistry, chapter 44, vol VII, page 1 (see page 8). 11. J. J. Fitt and H. W. Gschwend, J. Org. Chem., 1980, 45, 4257. 12. P. Beak and D. J. Kempf, J. Am. Chem. Soc., 1980, 102, 4550.

12 References

187

13. H. Mattes and C. Benezra, Tetrahedron Lett., 1985, 26, 5697. 14. R. M. Carlson, Tetrahedron Lett., 1978, 111. 15. M. F. Semmelhack, Org. React., 1972, 19, 115. 16. E. J. Corey, S. W. Chow, and R. A. Scherrer, J. Am. Chem. Soc., 1957, 79, 5773. 17. E. J. Corey and M. F. Semmelhack, J. Am. Chem. Soc., 1967, 89, 2755. 18. G. García-Gómez and J. M. Moretó, J. Am. Chem. Soc., 1999, 121, 878; Eur J. Org. Chem., 2001, 1359. 19. Y. Kobayashi, Y. Tokoro and K. Watatani., Eur J. Org. Chem., 2000, 3825. 20. I. Fleming, Comp. Org. Chem. vol 3, chapter 13, page 542; P. D. Magnus, T. Sarkar, and S. Djuric, Comprehensive Organometallic Chemistry, vol 7, chapter 48, page 515; E. Colvin, chapter 9, page 97; T. Chan and I. Fleming, Synthesis, 1978, 761; I. Fleming, J. Dunogues, and R. Smithers, Org. React., 1989, 37, 60. 21. I. Fleming, R. Snowden and A. Pearce, J. Chem. Soc., Chem. Commun,, 1976, 182; Disconnection Approach page 202. 22. J. Slutsky and H Kwart, J. Am. Chem. Soc., 1973, 95, 8678. 23. A. Hosomi, A. Shirahata and H. Sakurai, Chem. Lett., 1978, 901. 24. J.-P. Pillot, J. Dunogues, and R. Calas, Tetrahedron Lett., 1976, 1871. 25. D. Seyferth, K. R. Wursthorn and R. E. Mammarella, J. Org. Chem., 1977, 42, 3104. 26. Disconnection Approach, chapter 33, page 274. 27. B.-W. Au-Yeung and I. Fleming, J. Chem. Soc., Chem. Commun., 1977, 79. 28. I. Fleming and I. Paterson, Synthesis, 1979, 446. 29. H. Sakurai, Y. Sakata, and A. Hosomi, Chem. Lett., 1983, 409. 30. I. Fleming and B.-W. Au-Yeung, Tetrahedron, 1981, 37, supplement no 1, 13. 31. D. Wang and T.-H. Chan, J. Chem. Soc., Chem. Commun., 1984, 1273. 32. S. R. Wilson, L. R. Phillips, and K. J. Natalie, J. Am. Chem. Soc., 1979, 101, 3340. 33. I. Ojima, M. Kumagai, and Y. Miyazawa, Tetrahedron Lett., 1977, 1385. 34. A. Hosomi, M. Saito, and H. Sakurai, Tetrahedron Lett., 1980, 21, 355. 35. G. Déléris, J.-P. Pilot, and J.-C. Rayez, Tetrahedron, 1980, 36, 2215. 36. J.-C. Gramain and R. Remuson, Tetrahedron Lett., 1985, 26, 327. 37. R. Yamaguchi, M. Tanaka, T. Matsuda and K. Fujita, Chem. Commun., 1999, 2213. 38. G. Kettschau and G. Pattenden, Tetrahedron Lett., 1998, 39, 2027. 39. C.-M. Yu, J.-Y. Lee, B. So and J. Hong, Angew. Chem., Int. Ed., 2002, 41, 161. 40. W. H. Pearson, E. P. Stevens and A. Aponick, Tetrahedron Lett., 2001, 42, 7361. 41. M. Lombardo, R. Girotti, S. Morganti and C. Trombini, Org. Lett., 2001, 3, 2981. 42. M. Julia, Pure Appl. Chem., 1985, 57, 763. 43. B. Maryanoff and Reitz, Chem. Rev., 1989, 89, 863; I. Gosney and A. G. Rowley in Organophosphorus Reagents in Organic Synthesis, ed J. I. G. Cadogan, Academic Press, London, 1979, pages 63–77. 44. I. T. Harrison and B. Lythgoe, J. Chem. Soc., 1958, 843. 45. E. Vedejs and H. W. Fang, J. Org. Chem., 1984, 49, 210. 46. H. Pommer, Angew. Chem., 1960, 72, 811. 47. J.-L. Fabre and M. Julia, Tetrahedron Lett., 1983, 24, 4311; T. Cuvigny, C. Hervé de Penhoat and M. Julia, Tetrahedron Lett., 1983, 24, 4315; T. Cuvigny, J. L. Fabre, C. Hervé de Penhoat and M. Julia, Tetrahedron Lett., 1983, 24, 4319. 48. K. Uneyama and S. Torii, Tetrahedron Lett., 1976, 443.

13 Homoenolates Introduction: Homoenolisation and homoenolates Summary of strategies Three-Membered Ring Homoenolate Equivalents (The ‘Direct’ Strategy) Unsymmetrical cyclopropane homoenolates Conjugate addition of zinc homoenolates Zirconium homoenolates Enantioselective cyclopropane homoenolates An enantioselective amino acid homoenolate Homoenolate equivalents from three-membered rings The Defensive Strategy: d3 Reagents with Protected Carbonyl Groups Acetals Sulfones Phosphonium salts The Offensive Strategy: Heteroatom-Substituted Allyl Anions Regioselectivity Anions of allyl silanes Anions of allyl sulfides Anions of allyl ethers Anions of allyl amines Asymmetric allylic amines Allyl Carbamates Stereoselectivity in the homoaldol reaction

Introduction: Homoenolisation and homoenolates Enolisation 1 involves the removal of the α-proton from a carbonyl compound to form an enolate ion 2. Homoenolisation involves the removal of a β-proton 3 to form the homoenolate ion 4 or 5. Both the enolate and the homoenolate can be represented as carbanions, but whereas the enolate version 2b is merely a different way of representing a single delocalised structure, the homoenolate 5 is a different compound from the cyclopropane 4. No literal examples of homoenolates 5 are known so they have the status of synthons which may be represented in real life by reagents derived from cyclopropanols 4 among many other possibilities.1

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

190

13 Homoenolates

O B

O

O

base

H

R

R

1

2b

O R

H

O

O

base

B

R

2a ?

R

3

R

4

5

Direct homoenolisation of a carbonyl compound with base is not of much practical use. Ketones blocked in the α-position do exchange with deuterium in base, but after five weeks in t-BuOK/ t-BuOD at 185 ⬚C only 3.18 atoms of deuterium had been incorporated into camphor 6 with the distribution shown. The two sites for homoenolisation were responsible for only 0.65 and 0.45 atoms, in contrast to 1.27 atoms at the α and 0.81 at the γ atoms.2 0.81 t-BuOK, t-BuOH

1.27

5 weeks, 185 ˚C

O

O d3-6 (3.18 atoms D)

6

0.45 0.65

O D distribution in d3-6

Summary of strategies Useful synthetic methods therefore rely on reagents which have been devised to represent the d3 synthon or homoenolate 5 in reactions and there are three important classes of these. The cyclopropane or direct approach uses reagents such as 4. Hoppe’s ‘defensive strategy’ uses a nucleophile created at C3 that is neither stabilised nor destabilised by a protected carbonyl group as in the Grignard reagent 8. Hoppe’s ‘offensive strategy’ uses an allyl anion with a heteroatom (X in 9) at one end.1 This strategy has ‘many problems’ or rather many conditions that must be fulfilled. The heteroatom must help to stabilise the anion, the anion must react in the γ-position 10 (chapter 12), and it must be possible to hydrolyse the product 11 to a carbonyl compound.

O

O R

Br 7

O

X

R

BrMg 8

H R

9

X

X

base

E

R 10

R

E 11

Three-Membered Ring Homoenolate Equivalents (The ‘Direct’ Strategy) Perhaps the only true homoenolates used in synthesis are derived by metallation of derivatives of 3-haloacids. The acids themselves 12 give lithium 3-halocarboxylates 13 and hence by metallation the homoenolate which probably exists as 15, an analogue of the dilithium enolates of carboxylic acids (chapter 2). Reaction 16 with aldehydes or ketones gives γ-lactones 19, by a ‘homoaldol’ reaction via γ-hydroxyacids 18, common products from addition of acid homoenolates to carbonyl compounds.3

13 Three-Membered Ring Homoenolate Equivalents (The ‘Direct’ Strategy)

O

O

O BuLi

Br

OH

Br

Li

OLi

OLi

13 O

O

OLi

BuLi

12 R

191

H

R

OLi

OLi

R

O

OH

H2O

OH

OLi 16

O

O

O

Li

OLi 15

14

R

18

17

19

Ester homoenolates can be made from 3-chloroesters 20 with sodium in the presence of Me3SiCl which traps them as the cyclopropyl silyl ethers4 21, analogues of silyl enol ethers, in a step reminiscent of the acyloin condensation.5 Reaction with aldehydes and ketones again gives γ-lactones 19 and Kuwajima has shown that the titanium homoenolate 22 is a true intermediate in this reaction.6 O

OSiMe3

Na

Cl

OEt

Me3SiCl

TiCl4

19

O

OEt 21

20

RCHO

Cl3Ti OEt

22

Addition of this homoenolate to the ketone 23, a single enantiomer derived from natural proline, gives the tertiary alcohol 24 with high stereoselectivity.7 Removal of the Cbz protecting group leads to spontaneous cyclisation to give 25, an intermediate on the way to the neurotoxin pumiliotoxin 251D 26.

H

O

H 21

N

TiCl4

Cbz

H

OH CO2Et

N

H

OH

H2, Pd/C

N

N Cbz

O Bu 25; 100% yield 26; Pumiliotoxin 251D

24; 49% yield

23

OH

Unsymmetrical cyclopropane homoenolates There is an obvious difficulty if the cyclopropane ring is unsymmetrically substituted and our best guideline here is the last stage of the Favorskii reaction,8 e.g. 27 to 31, where a cyclopropoxide ion is protonated 30 at the carbon atom giving the better anion, usually at the less substituted carbon.

O

O

O Br

MeO MeOH

27

MeO

O

MeOH

28

CO2Me

MeO

Br

29

H OMe 30

31

Opening the corresponding silyl ethers 32 with TiCl4 also usually occurs at the less substituted atom to give the homoenolates6 33 and hence the lactones 34. This regioselectivity means that a chiral centre next to the carbonyl group 35 can be preserved during the formation of esters 37 by alkylation of a zinc homoenolate9 which may have the structure 36.

192

13 Homoenolates

MeO

OSiMe3 TiCl4

RCHO

Cl3Ti O 33

32

R O 34

OMe

O

H H X

Zn

H

RHal

R

Zn

CO2Me

O

(+)-35

CO2Me

OMe 36

37

Conjugate addition of zinc homoenolates The simple zinc homoenolate 39, derived from 21 with ZnCl2 and ultrasound, does a conjugate addition to the acetylenic ester with Cu(I) catalysis to give a new cyclopentenone 40, an intermediate10 on the way to gingkolide B. The stereochemistry of 38 is not affected by the reagent. O CO2Et

O

EtO2C

+ Zn O Et3SiO 38

O

CuBr.SMe 2

OEt

HMPA Et2O, THF

t-Bu

Et3SiO 39

t-Bu 40

Zirconium homoenolates Zirconium ester homoenolates 43 or 44 can be prepared from the triethyl orthoacrylate 42 with the zirconocene complex of but-1-ene 41 (Cp means cyclopentadiene). These resemble the zinc and titanium species we have been discussing but are not derived from cyclopropanes.11

Et

OEt

OEt +

Cp2Zr

OEt OEt

Cp2Zr 41

OEt

Cp2Zr

OEt 43

42

OEt OEt

OEt 44

In the presence of CuCN, these homoenolates react with acid chlorides to make γ-ketoesters but their forte is reaction with allylic phosphates. Prenyl diethylphosphate gives 45 in good yield and the more highly substituted product 46 is formed in near quantitative yield. These are impressive results: both reagents react through the γ-carbon: the homoenolate 43 where the Zr is and the allylic phosphate where the phosphate isn’t.

γ

Cp2Zr OEt 43

OEt

OEt

O

OEt

γ

P O

OEt

CO2Et

CO2Et Ph

CuCN

45; 81% yield

46; 96% yield

13 Three-Membered Ring Homoenolate Equivalents (The ‘Direct’ Strategy)

193

Enantioselective cyclopropane homoenolates Recent developments in the cyclopropane approach to homoenolates have emphasised enantioselectivity.12 Racemic mixtures of acetals of cyclopropanones can be prepared by SimmonsSmith cyclopropanation of silyl ketene acetals. Silicon-promoted conjugate addition (chapter 9) of Grignard reagents to acrylate esters 47 gives mixtures of E/Z isomers of the silyl ketene acetals 48. Cyclopropanation and exchange of the silyl group for an acetate gives a roughly 1:1 mixture of syn and anti cyclopropanes 49 and 50, easily separated by chromatography. R

RMgBr

O

Me3SiCl

OEt 47

R +

OEt OEt AcO AcO (±)-49 (±)-50 ~1:1

2. MeOH

OEt

Me3SiO

R

1. CH 2I2, Zn

3. Ac 2O, pyridine

48

Kinetic resolution (chapter 28) of each diastereoisomer with a lipase gives single enantiomers of the hemiacetals 51 and hence the zinc homoenolates 52. Various electrophiles (E in 53) gave enantiomerically pure (⬎99% ee) adducts where E = H, allyl or imine.13 R OEt AcO (±)-50

R

1. Me 3SiOTf

E

R

Zn

OEt

2. LiAlH 4

E

H

R

1. lipase from Candida antarctica

2. ZnCl 2

HO (–)-51

O OEt (–)-52

CO2Et (–)-53

An enantioselective amino acid homoenolate An important use of enantiomerically pure homoenolate equivalents is the synthesis of other α-amino acids from serine. Natural (S)-(⫹)-serine 54 is protected on both its amino and carboxyl groups and the OH group turned into a leaving group 55. Displacement with iodide gives the starting material 56 with no loss of optical activity.14 NH2

HO

1. PhCH 2OH, PhSO 2OH, CCl 4 (74%)

CO2H

NHBoc

TsO

2. (Boc) 2O, NaOH, H 2O, THF (81%)

CO2Bn

3. TsCl, pyridine (84%)

54; (S)-(+)-serine

NHBoc

NaI

I Me2CO

CO2Bn 56

55

Treatment with zinc/copper couple activated by ultrasound in dimethylacetamide gave the zinc homoenolate 57 stabilised by chelation to either the ester or the carbamate. Acylation with acid chlorides and Pd(0) catalysis gave the enantiomerically pure γ-oxo-amino acids 58 in good yield. I

NHBoc

Zn/Cu Me2NAc

CO2Bn 56

ultrasound

IZn

NHBoc CO2Bn 57

RCOCl catalytic PdCl2(PPh3)2

R

NHBoc O

CO2Bn 58

Arylation with ArI and palladium catalysis is also possible and forms the key step in an asymmetric synthesis of the antibiotic azatyrosine.15 This differentiating antibiotic has anti-cancer potential. Coupling 57 with the pyridine 59 gave 60 and removal of the methyl protecting group gave azatyrosine 61.

194

13 Homoenolates

I

NHBoc

MeO

N

catalytic PdCl2(PPh3)2

59

NH2

BBr3

57

CO2Bn

N

MeO

CH2Cl2

N

HO

60; 74% yield

CO2H

61; 90% yield

Homoenolate equivalents from three-membered rings Other three-membered ring compounds have been used to achieve the result of homoenolate attack on ketones, the most important being Trost’s cyclopropyl sulfur compounds and Corey’s cyclopropyl ethers. Attack of the sulfonium ylid16 63 on a ketone gives the usual epoxide17 even though it is an oxaspiropentane 64: this rearranges to the cyclobutanone 65 in acid. BaeyerVilliger rearrangement involves migration of the more highly substituted carbon atom18 to give the lactone19 66, the result of ester homoenolate addition to the ketone R2CO. H

Li

BuLi

Ph2S

O

R2C=O

R

O

R

RCO3H

R

R

Ph2S

62

HBF4

O

R

R

63

O

64

65

66

Corey’s method20 relies on metal exchange with the bromocyclopropane 69 prepared by carbene addition. The extra stabilisation of cyclopropyl anions (chapter 8) makes both this lithium derivative and the ylid 63 more easily handled. Addition to aldehydes or ketones gives mixtures of adducts 70 [it turns out that none of the stereochemistry of 69 or 70 matters] which fragment under Lewis acid catalysis to give the thioacetal 71. Careful hydrolysis releases the 3,4-enal E-72, the product of a homoaldol reaction with an aldehyde homoenolate and RCHO and a difficult compound to make as the double bond moves into conjugation very easily. OH CHOMe

Br

Br

OMe

68

R

1. t-BuLi

OMe

HO

2. RCHO

67

SH BF3

69

70 Hg(II)

O

R

MeCN

R

H2O CaCO3

S

CHO

71

72

The disconnections for this homoaldol are not obvious. Just as a conjugated (α,β-unsaturated) enal is derived from an aldol, we can consider deriving this unconjugated (β,γ-unsaturated) enal 72 from a hydroxyaldehyde. We might initially prefer FGI to 73 as the 1,3-relationship is easy to make. But 73 will dehydrate to the conjugated (α,β-unsaturated) enal. Instead we must set up a 1,4-diO relationship 74 so that we can disconnect to an aldehyde RCHO and the homoenolate 75 of propanal.

R

3

2

OH 73

1 CHO

FGI

FGI

R

CHO 72

OH 2 R 4

3 74

O

1,4-diO

1 CHO

CHO homoaldol

R 75; d 3 synthon

13 The Defensive Strategy: d3 Reagents with Protected Carbonyl Groups

195

The Defensive Strategy: d3 Reagents with Protected Carbonyl Groups Acetals Conceptually the simplest of all the homoenolate equivalents, organometallic reagents such as the Grignard 8 could be regarded as avoiding the problem. This is unfair and a better way to look at them is to regard their synthesis as reversing the polarity of unsaturated carbonyl compounds. Unsaturated aldehydes and ketones can be converted into bromoacetals such as 77, and Büchi and Wuest21 used the Grignard reagent derived from 77, that may have the chelated structure 78, in their synthesis of nuciferal as long ago as 1969. Br

O CHO

HO

1. Mg

OH

O

2. RCHO

O

Br

HBr

Mg

76

R

O

77

OH

H 2O

3. H,

78

CHO

79

Roush used this reagent in his synthesis of dendrobine22 for which he needed the unsaturated hydroxyaldehyde 80. Disconnection next to the OH group reveals the dienal 81 and the homoenolate synthon 75. The dienal can be made by consecutive aldol reactions or using extended enolate chemistry (chapter 11). Reaction between the Grignard reagent from 77 and the aldehyde 81 in THF gave ⬎95% of 80. Similar reagents are available at the ketone oxidation level.23 OH

O

1,4-diO

+

CHO

CHO

H

homoaldol

80

81

75

For both the alcohol and the carboxylic acid oxidation levels, the best organometallic reagent is probably the very simple protected lithium derivative 84 introduced by Eaton.24 Easily made from available bromopropanol 82, the acetal 83 gives the stable (because of chelation by one or both acetal oxygen atoms?) lithium derivative 84 by oxidative insertion. OEt O

Br

OH 82

OEt Li

Br

Cl2CH.CO2H

O

Et2O

83

Li

O 84

Addition to aldehydes, ketones, and enones occurs cleanly at the carbonyl group, e.g. to give 86, but the copper derivative adds to enones in the expected Michael sense (chapter 9) to give 1,6-dicarbonyl compounds such as 88 after oxidation. Cyclisation to 89 completes a cyclopentannelation sequence (chapter 6). OEt Li

+

O

CuI

O 85

84 O

OEt

OH

O

86

85 1. H,

O

3. H,

87

H 2O

O

OEt

O

O

2. RuO 4/NaIO4

CO2Me

base

MeOH

88

89

196

13 Homoenolates

Sulfones Possibly more popular are the protected sulfones25,26 such as 94, again easily made by Michael addition at either the sulfide or sulfone27 oxidation state as sulfur nucleophiles are excellent Michael donors. O PhSH

CHO

PhS

76

HO

CHO

OH

91

H

90

O

PhS

mCPBA

CHO

O O Ph

S

HO

PhO2S

Li

OH

O

CHO

94 H

92

O

PhO2S

93

Lithium derivatives of these sulfones can be alkylated with alkyl halides or epoxides,28 and acylated with esters to give the ketones 95 from which the sulfonyl group can be removed with aluminium amalgam. Deprotection and cyclisation provides a synthesis of cyclopentenones 98 (cf. chapter 6).26 R O

R O

Al/Hg

O

1. BuLi

R O

base

H

94 2. RCH 2CO2Me

PhO2S

O

CHO

O

O 95

R

O

97

96

98

Phosphonium salts The corresponding phosphonium salts29 99, 100, and 101, made from the bromoacetals, can be used in Wittig reactions to give Z-acetals such as 102. It is difficult to hydrolyse the acetal without isomerising the geometry or the position of the double bond, but non-cyclic acetals such as 101 satisfactorily give Z-alk-3-enals30 103.

O Ph3P

O O

Oi-Pr 101

Ph3P

100

Oi-Pr Ph3P

O

Ph3P

99

Oi-Pr Oi-Pr 101

Oi-Pr 1. BuLi

CHO

H

Oi-Pr

2. RCHO

R

Z-102

H2O

R Z-103

The Offensive Strategy: Heteroatom-Substituted Allyl Anions Regioselectivity Allyl anions with a heteroatom at one end 106; Z = OR, NR2, etc, can act as homoenolates providing certain conditions are met. The anion must be easy to make from 104 or 105, it must react reliably at the γ-position with a range of electrophiles E⫹ to give only the vinyl derivative

13 The Offensive Strategy: Heteroatom-Substituted Allyl Anions

197

107, and that vinyl derivative must be easily hydrolysed to a carbonyl compound. It turns out1 that these conditions are generally fulfilled when Z = NR2, SiR3, OR, or SR, though the substituent(s) R do also matter. If M is a transition metal, the π-allyl complex 106c may be the best representation of the allyl ‘anion.’ H base

Z

R

R γ

104

E

α 106a

Z

E

E

R

H

α

base

Z

R

R

107

E

R γ

105

Z

M CHO

Z

R

108

106c

Z 106b

Anions of allyl silanes Lithium derivatives of allyl silanes react in the γ-position with alkyl halides, epoxides, and carbonyl compounds. The lithium derivative 110 of allyl silane 109 gives only the γ-adduct 111 with ketones.31 Vinyl silanes such as 111 are usually converted into carbonyl compounds via epoxides which rearrange with Lewis acid catalysis and loss of silicon to give protected versions of ketones or aldehydes 112. Li

OH

BuLi

SiMe3

SiMe3

109

SiMe3

OMe

1. RCO 3H

cyclohexanone

O 2. BF 3 MeOH

110

111

112

The reaction may involve the formation of the β-silyl cation 114 (chapter 12) or in this case participation by the OH group may be involved. In any case the final product is a protected version 112 of the 3-hydroxyaldehyde 116.

OH

OH

O

RCO3H

SiMe3

111

BF3

113 OH

OH

OBF3 SiMe3 MeOH 114

OH

OH

CHO O

115

116

MeOH

112

117

Anions of allyl sulfides Anions of allyl sulfides have been quite widely used but their regioselectivity is unpredictable and the vinyl sulfides 107; Z = SR resulting from γ-addition are difficult to hydrolyse.32 Anions of allyl ethers are more difficult to prepare, but generally better behaved.33 A study of such anions led Still to propose that allyl anions with very anion stabilising substituents Z = COR 118, SO2Ph 119, ⫹ PPh3 120 etc, react α with all electrophiles, those with moderately anion-stabilising substituents

198

13 Homoenolates

react α (or where the metal is) with alkyl halides but γ (or where the metal isn’t) with carbonyl compounds, and those with anion destabilising substitutents Z = OR 121, NR2 122, or SiMe3 123, react γ with all electrophiles.34 A wide selection of these anions appears in Hoppe’s review.1 E

COR

SO2R

E

R

PR3

E

E

R

R

118

OR

119

NR2 E

R

120

SiR3

R

121

E

R 123

122

Anions of allyl ethers Allyl ethers 124 clearly belong in the last category and their lithium derivatives 125 react both with alkyl halides, to give 126, and carbonyl compounds, to give 128, with the correct regioselectivity for them to act as homoenolate equivalents. The Z-alkene in both products 126 and 128 suggests the chelated structure 125 for the allyl-lithium. Acidic hydrolysis releases the carbonyl group 117b, in the form of a hemiacetal 117a from the carbonyl adduct 128. These allyl ethers 124 need strong base to remove the not very acidic proton and must be lithiated at low temperature to avoid the Wittig rearrangement.33 R

RHal

s-BuLi

OMe

O

Li

–65 ˚C

124

125

H

OMe

Me

R

CHO

H2O

126

1. s-BuLi, –65 ˚C

127

H

124 2. cyclohexanone

H2O

OH

O

OMe

128

117a

OH

OH

CHO

117b

Anions of allyl amines Anions derived from enamines or allyl amines have been widely explored and are among the best homoenolate equivalents. Enamines 130 from PhNHMe and aldehydes35 or aryl ketones36 129 can be deprotonated with BuLi to give lithium allyls 131, as can the corresponding allyl amines.37 The anion of carbazole 132 reacts on nitrogen with allyl halides and the products 133 can be converted to lithium derivatives 134 with BuLi in the presence of chelating agents such as TMEDA. The representations 131a, 131b and 134 are alternatives.38 Ph

O

PhNHMe TiCl4

Ph

N

Me

Ph BuLi

Ph

129

Ph

Li

Me N Li

or Ph 131a

base Br

N

Ph 130

N

Me

131b

BuLi TMEDA

N

N

133

134

H 132

13 The Offensive Strategy: Heteroatom-Substituted Allyl Anions

199

Amino-nitriles39 136 can be made directly from unsaturated aldehydes in the first stage of the Strecker reaction and deprotonated with LDA. The same allyl-lithium derivatives 137 can be made from saturated amides40 139 via the vinyl isomer 140 of 136. All these lithium allyls, maybe 138 is the best representation, react as homoenolates. NR2

R2NH

CHO

NR2 LDA

CN

HCN

135

CN 137a

136

Li

NR2 CN

1. COCl 2

CONR2

NR2

2. Zn(CN) 2 3. Et 3N

139

138

NR2 LDA

CN

CN

140

137b

The lithium derivative from the unsymmetrical allylic carbazole 141 reacts with the unsymmetrical allyl chloride 142 to give 143. The starting material was made from carbazole 132 and 142 so this allyl electrophile has reacted twice at its less substituted end, but the nucleophilic homoenolate from 141 has reacted at its more substituted end, because that is γ-addition relative to the nitrogen atom.38,39 Hydrolysis of the resulting enamine E-143 gives the aldehyde 144.

1. BuLi, TMEDA 2. Cl

N

141

H2O

N

OHC

143

142

144

We shall see in chapter 14 that saturated amino nitriles are good d1 reagents and the unsaturated versions 140 are good d3 reagents, particularly effective at Michael addition. The lithium derivative from 140; R = Et gives conjugate addition to cyclohexenone and hence the keto-acid 146 with a 1,6 relationship between the two carbonyl groups. The reconnection strategy41 is not suitable for this compound as it is impossible to put a double bond between the two carbonyl groups. Lithium derivatives of allyl amines are homoenolates of aldehydes or ketones while lithium derivatives of unsaturated amino-nitriles are homoenolates of acids. O NEt2 CN

O

1. LDA

H

CN

2. cyclohexenone

H2O

CO2H

NEt2 140; R = Et

145

146

Asymmetric allylic amines In this style too, asymmetry has become a major concern.42 An asymmetric version 147 of this homoenolate was made from natural nor-ephedrine and could be alkylated stereoselectively using lithium tetramethylpiperidide (LiTMP) as base. The product could be hydrolysed to the bicyclic amide43 149.

200

13 Homoenolates

CN

Me

CN

Me

1. LiTMP

N

N

Me

N

Me

2. MeI

O

Li LiTMP

O

H

N

O Ph 147

Me

H2O

O Ph

Ph

148

149

An open chain example44 153 of asymmetric homoenolates from allylic amines uses the proline derivative 152 and the β-stannyl ketone 153 made by conjugate addition of Bu3SnLi to the enone 150. O

O

OMe

Bu3SnLi

N H

+ THF, –78 ˚C

OMe N

TsOH

Bu3Sn 150

151

Bu3Sn

152

153

The lithium derivative of 153 can equilibrate to the chelated isomer Z-154 that reacts cleanly in the γ-position with alkyl and allyl halides to give, after hydrolysis of the initially formed enamine 155, the ketones 156 with generally good enantiomeric excess. Structure Z-154a shows the chelation in the lithiated intermediate best. OMe N

OMe

OMe

N

N

O

1. HCl

RX

BuLi

R

153 THF –78 ˚C

2. NaOH

E-154

Z-154

156

Z-155 OMe N

Li

Z-154a

Allyl Carbamates The most sophisticated homoenolate equivalents, and the first to allow some control over stereochemistry, are the lithium derivatives 160 of allyl carbamates 159 introduced by Hoppe.1 They are particularly valuable for reaction with an aldehyde or a ketone in a homoaldol reaction. O

O

Li

O RCHO

LDA

OH

+

Cl

157 R

158

O Li

R

O O

161

O

NR2

NR2

NR2

O 160

159 O

Li O 162

R

O NR2

NR2

OCONR2 OH

R

CHO OH

163

164

13 References

201

Though anions of allyl ethers, the extra functionality allows regiospecific heteroatom-directed lithiation 160 in the manner of chapter 7. Hence reaction of 160 with an aldehyde goes through a six-membered cyclic transition state 161 to give the enol derivative 163 and hence the γ-hydroxyaldehyde 164. We described some anions of allylic amines as γ-lithiated species 131a and 154 that reacted in the γ-position. Here the position of lithiation is α 160 but that too ensures good γ-selectivity even when the γ-position is disubstituted 165 and the aldehyde is branched. Both the yield of 167 and the selectivity are excellent.40 O O O

Li

1. BuLi TMEDA

O 2. i -PrCHO

NR2

O

165

O

NR2

NR2

OH 167; 87% yield, >97% γ-addition

166

Stereoselectivity in the homoaldol reaction Because of the cyclic transition state, stereoselectivity in the manner of stereoselective aldol reactions (chapter 4) is found, that is the geometry of the double bond in 168 determines the stereochemistry of the homoaldol product 169, providing that lithium is exchanged for titanium or aluminium. That E-168 gives anti-169 and Z-168 gives syn-169 is not surprising, but the geometry of the resulting enol derivatives is. Fortunately it is irrelevant to the final product. OCONR2

1. BuLi TMEDA

OCONR2

2. (Et 2N)3TiCl

OH

3. MeCHO

E-168; R = i-Pr

anti-Z-169; R = i-Pr 1. BuLi TMEDA

OCONR2

OCONR2

2. (Et 2N)3TiCl

OH

3. MeCHO

syn-E-169; R = i-Pr

Z-168; R = i-Pr

The final hydrolysis is not trivial: an Hg(II) catalyst is necessary to give acetals 170 that can be oxidised to lactones 171. By this means both syn and anti-172, the quercus lactones from oak bark,45 can be made stereoselectively in 87% and 90% yields respectively. Me MeOH MsOH

163 cat Hg(II)

Me

RCO3H

R

O

OMe

170

BF3

R

O 171

O

Bu

O anti-172

O Bu

O syn-173

O

References General references are given on page 893 1. N. H. Werstiuk, Tetrahedron, 1983, 39, 205; D. Hoppe, Angew. Chem., Int. Edn., 1984, 23, 932. 2. A. Nickon, J. L. Lambert, J. E. Oliver, D. F. Covey and J. Morgan, J. Am. Chem. Soc., 1978, 98, 2593. 3. D. Caine and A. S. Frobese, Tetrahedron Lett., 1978, 883. 4. K. Rühlmann, Synthesis, 1971, 236.

202

13 Homoenolates

5. Disconnection Approach, page 202. 6. E. Nakamura and I. Kuwajima, J. Am. Chem. Soc., 1977, 99, 7360; 1983, 105, 651; 1984, 106, 3368. 7. A. G. M. Barrett and F. Damiani, J. Org. Chem., 1999, 64, 1410. 8. Clayden, Organic Chemistry, p. 990. 9. E. Nakamura, K. Sekiya and I. Kuwajima, Tetrahedron Lett., 1987, 28, 337. 10. M. T. Crimmins, J. M. Pace, P. G. Nantermet, A. S. Kim-Meade, J. B. Thomas, S. H. Watterson and A. S. Wagman, J. Am. Chem. Soc., 2000, 122, 8453. 11. A. Sato, H. Ito, Y. Yamaguchi and T. Taguchi, Tetrahedron Lett., 2000, 41, 10239. 12. H. Ahlbrecht and U. Beyer, Synthesis, 1999, 365. 13. B. Westermann and B. Krebs, Org. Lett., 2001, 3, 189. 14. R. F. W. Jackson, N. Wishart, A. Wood, K. James and M. J. Wythes, J. Org. Chem., 1992, 57, 3397; R. F. W. Jackson and M. Perez-Gonzalez, Org. Synth., 2005, 81, 77. 15. B. Ye and T. R. Burke, J. Org. Chem., 1995, 60, 2640. 16. B. M. Trost, Acc. Chem. Res., 1974, 7, 85. 17. Disconnection Approach, page 253. 18. Disconnection Approach, page 227. 19. B. M. Trost and M. J. Bogdanowicz, J. Am. Chem. Soc., 1972, 94, 4777; 1973, 95, 5321. 20. E. J. Corey and P. Ulrich, Tetrahedron Lett., 1975, 3685. 21. G. Büchi and H. Wuest, J. Org. Chem., 1969, 34, 1122. 22. W. R. Roush, J. Am. Chem. Soc., 1980, 102, 1390. 23. A. A. Ponaras, Tetrahedron Lett., 1976, 3105; S. A. Monti and Y.-L. Yang, J. Org. Chem., 1979, 44, 897. 24. P. E. Eaton, G. F. Cooper, R. C. Johnson and R. H. Mueller, J. Org. Chem., 1972, 37, 1947. 25. M. Julia and B. Badet, Bull. Soc. Chim. Fr., 1975, 1363. 26. K. Kondo and D. Tunemoto, Tetrahedron Lett., 1975, 1007; 1397. 27. H. W. Pinnick and M. A. Reynolds, J. Org. Chem., 1979, 44, 160. 28. K. Kondo, D. Tunemoto, and E. Saito, Tetrahedron Lett., 1975, 2275. 29. J. C. Stowell and J. R. Keith, Synthesis, 1979, 132; S. W. Baldwin and M. T. Crimmins, J. Am. Chem. Soc., 1982, 104, 1132. 30. A. R. Battersby, D. G. Buckley, J. Staunton and P. J. Williams, J. Chem. Soc., Perkin Trans. 1, 1979, 2550. 31. R. Corriu and J. Masse, J. Organomet. Chem., 1973, 57, C5; D. Ayalon-Chass, E. Ehlinger and P. Magnus, J. Chem. Soc., Chem. Commun., 1977, 772; E. Ehlinger and P. Magnus, J. Am. Chem. Soc., 1980, 102, 5004; Tetrahedron Lett., 1980, 21, 11. 32. J. F. Biellmann and J. B. Ducep, Org. React., 1982, 27, 1. 33. D. A. Evans, G. C. Andrews, and B. Buckwalter, J. Am. Chem. Soc., 1974, 96, 5560. 34. W. C. Still and T. L. Macdonald, J. Am. Chem. Soc., 1974, 96, 5561; J. Org. Chem., 1976, 41, 3620. 35. G. Rauchschwalbe and H. Ahlbrecht, Synthesis, 1974, 663. 36. H. Ahlbrecht and G. Rauchschwalbe, Synthesis, 1973, 417. 37. H. Ahlbrecht and J. Eichler, Synthesis, 1974, 672. 38. M. Julia, A. Schouteeten and M. Baillage Tetrahedron Lett., 1974, 3433; M. Julia, F. Le Goffié, and L. de Matos, Compt. Rend., 1970, 270C, 954. 39. H. Ahlbrecht and C. Vonderheid, Synthesis, 1975, 512. 40. B. Lesur, J. Toye, M. Chantrenne and L. Ghosez, Tetrahedron Lett., 1979, 2835; D. S. Grierson, M. Harris, and H.-P. Husson, J. Am. Chem. Soc., 1980, 102, 1064; B. Costisella and H. Gross, Tetrahedron, 1982, 38, 139. 41. Disconnection Approach, chapters 26 and 27, page 217. 42. H. Ahlbrecht and U. Beyer, Synthesis, 1999, 365. 43. J. B. Schwarz and A. I. Meyers, J. Org. Chem., 1998, 63, 1732; A. G. Waterson and A. I. Meyers, J. Org. Chem., 2000, 65, 7240. 44. H. Ahlbrecht, R. Schmidt, and U. Beyer, Eur. J. Org. Chem., 1998, 1371. 45. D. Hoppe and F. Lichtenberg, Angew. Chem., Int. Edn., 1984, 23, 239; D. Hoppe and A. Brönneke, Tetrahedron Lett., 1983, 24, 1687.

14 Acyl Anion Equivalents Introduction: Acyl Anions? Acyl Anion Equivalents: d1 Reagents Modified Acetals as Acyl Anion Equivalents An asymmetric synthesis of pyrenophorin The synthesis of a drug using a dithian as an acyl anion equivalent Dithioacetal monoxides Protected Cyanohydrins of Aldehydes Amino nitriles are useful for conjugate addition Acyl anion equivalents of the ester d1 synthon – CO2R Methods Based on Vinyl (Enol) Ethers and Enamines Lithium derivatives of cyclic vinyl ethers The synthesis of pederin and related anti-tumour agents Lithium derivatives of allenyl ethers Oxidative Cleavage of Allenes Vinyl Ethers and Enamines from Wittig-Style Reactions Vinyl ethers Enamines A synthesis of O-methyljoubertiamine Synthesis of enals by functionalised Wittig reagents Nitroalkanes A synthesis of strigol using nitroalkanes Synthesis of mono-protected diketones Catalytic Methods: The Stetter Reaction

Introduction: Acyl Anions? The acyl cation 2a or acylium ion 2b is a familiar intermediate in the Friedel-Crafts reaction. It is easy to make (acid chloride ⫹ Lewis acid 1) and it can be observed by NMR as it expresses the natural reactivity pattern of the acyl group. The acyl anion by contrast has umpolung or reverse polarity.1 One might imagine making it from an aldehyde by deprotonation 3 and that it would be trigonal 4a or possibly an oxy-carbene 4b. Such species are (probably) unknown and their rarity as well as their potential in synthesis has led to many synthetic equivalents for this elusive synthon. The acyl anion, the d1 synthon, is the parent of all synthons with umpolung2 and should perhaps have been treated before the homoenolates dealt with in the previous chapter.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

204

14 Acyl Anion Equivalents

O R

Cl

AlCl3

O

O

2a

R 2b

R

1

O

O

?

R

B

H

O

R

3

R 4a

4b

There are a few examples of carbonyl compounds losing a proton from the CHO group in strong base, though they are not aldehydes. The tertiary amide 5 gives a lithium derivative with t-BuLi, perhaps 6, that adds to aldehydes to give 7, the result of addition of an acyl anion to the aldehyde.3

O

O

O t-BuLi

i-Pr2N

H

R

RCHO

i-Pr2N

5

i-Pr2N

Li

OH 7

6?

There are two approaches to lithium derivatives of aldehydes, though neither starts from the aldehydes themselves. Carbonylation of primary alkyl-lithiums gives an intermediate,4 probably 8, that adds to aldehydes or ketones, e.g. 9, to give hydroxyketones 10 and can be silylated to give the acyl silane 11.

BuLi

O

O

CO

Bu

O

Me3SiCl

+ Li

8

Bu O 9

8

Bu

OH 10

SiMe3 11

Alternatively the Me3Si group may in turn be removed from the acyl silanes 11, 12 or 14 with CsF to give possibly the acyl anion itself, or at any rate a species which adds to aldehydes and ketones in the same way,5 and even does Michael reactions to give e.g. 15.

OH SiMe3

O

PhCHO CsF

O 12

Ph

O O 13

SiMe3

N Me

O 14

CsF cyclohexenone

O

N Me

O 15

Acyl Anion Equivalents: d1 Reagents The yields in these reactions are not wonderful and most syntheses planned with acyl anion or d1 synthons are realised with one of the reagents we are about to describe rather than with acyllithiums. Things may change as understanding of these rather reactive intermediates develops. There are three main types of acyl anion equivalent: reagents which can be considered as modified acetals, that is protected aldehydes, masked carbonyl compounds such a nitroalkanes, and substituted vinyl-lithiums. The rest of this chapter will be devoted to these reagents.

14 Modified Acetals as Acyl Anion Equivalents

205

Modified Acetals as Acyl Anion Equivalents Any acetal-like derivative 16 of an aldehyde can be used as an acyl anion equivalent providing the atoms X and Y stabilise an anion so that the lithium derivative 17 (or even the anion itself) can be formed, and providing that the product 18 of addition to an electrophile can be hydrolysed to a carbonyl compound. There is usually a conflict between these two requirements: thus acetals 16; X = Y = O are very easily hydrolysed but do not in general form lithium derivatives. Often one substituent is chosen to stabilise the anion (S, CN, etc) while the other is chosen to help the hydrolysis (O, N, etc). RX RCHO

YR

R

RX

BuLi

YR

R

H

RX

E

Li

16

YR

R

17

E

O

H

H2O

R

18

E 19

Dithians 20 illustrate this approach and have been remarkably popular considering the problems in their use. They can be made from aldehydes with propane-1,3-dithiol and Lewis acid catalysis. They are deprotonated with BuLi and react with alkyl halides, epoxides, and carbonyl compounds (E⫹) to give 22 and hence 19 after hydrolysis. The hydrolysis is by no means easy: there are many methods and this alone should warn us that none is very good.6

SH

SH

RCHO BF3

S

S

R

BuLi

H

S

S

R

20

S

E

Li

S

R

21

E

O Hg(II)

R

BF3

22

E 19

An asymmetric synthesis of pyrenophorin A synthesis of pyrenophorin using dithian shows how it works. The target 23 is a macrocyclic (16membered ring) dilactone and disconnection of the ester linkages reveals two identical halves 24: pyrenophorin has C2 symmetry.7 O

O O

CO2H

2 x C–O

O

O

OH

OH lactones

HO2C

O O

O

23; pyrenophorin

24

The enone 24 could be made by an aldol or Wittig approach from 25. This contains unhelpful 1,2 and 1,4 relationships and the strategy is to ignore these and disconnect either side of the ketone, imagining it as a d1 reagent used twice. The electrophiles would be one enantiomer of an alkylating agent 26, probably with the OH group protected, and a reagent for the formyl cation 28. O

O 2 × d1

aldol

CO2H

CHO OH

OH

OH 24a

25

26

X

+ S

S + 27

CHO 28

206

14 Acyl Anion Equivalents

An optically active reagent 32 for the left hand fragment can be made by asymmetric reduction of ethyl acetoacetate with ordinary baker’s yeast (see chapter 29), protection of the OH group as a mixed acetal 31, reduction of the ester and conversion of the new OH group into an iodide 32. 1. LiAlH 4

CO2Et

baker's yeast

CO2Et

OMe

CO2Et

O

O

29

OMe

O

OMe

3. NaI, acetone

CF3CO2H

OH

I

2. TsCl, pyridine

30

31

32

Alkylation of dithian itself 27 with the iodide 32 establishes the 1,4-diO relationship in 34 and this new dithian can be acylated with DMF (Me2NCHO). The carbon skeleton of half pyrenophorin 35 is completed by an E-selective Wittig reaction with a stabilised ylid. PPh3

S

S

1. BuLi

S

S

CO2Me

1. BuLi

27

H

2. 32

CHO 2. DMF

O

OMe

O

OMe

33

34

S

S CO2Me

O

OMe 35

Removal of the two protecting groups leaves the dithian in place and it is left there while the macrocyclisation is completed with a Mitsunobu reaction. Notice that this esterification occurs with inversion so only now is the correct stereochemistry in place. Finally the dithian is hydrolysed with Hg(II) and BF3.

S 1. H,

H 2O

CO2H

35 2. LiOH MeOH

S

S

OH

DEAD = DiEthyl N CO2Et N AzoDicarboxylate EtO2C

O HgO

DEAD

O

Ph3P

24

S

O

O

S

23 BF3

S

36

The synthesis of a drug using a dithian as an acyl anion equivalent The anti-antihyperlipoproteinemic drug acifran 37 is a simple heterocycle whose structure can best be understood after disconnection of the enol ether.8 The carbon framework 38 contains a

14 Modified Acetals as Acyl Anion Equivalents

207

remarkable number of relationships: two 1,2- and one 1,3-diCO relationships 38a as well as two 1,4-diCO and a 1,5-diCO. O

O C–O

Ph

O

Ph

enol ether

CO2H

1,2-diCO =

HO O

37; Acifran

1,2-diCO

O

O

Ph

CO2H HO 1,3-diCO

CO2H

38a

38

We prefer odd-numbered to even-numbered disconnections and of the two possible 1,3-diCO 38b is better as it gives an enolisable ketone 39 and the unenolisable, symmetrical and very electrophilic diethyl oxalate 41. O

O

O

O

O

1,3-diCO

Ph a

b

FGI

Ph

CO2H

+ X

b

HO

HO 39

38b

CO2H

OEt

EtO

O 41; diethyl oxatale

40

The α-hydroxyketone 39 could be made from attack of a d1 reagent on the rather enolisable ketone PhCOMe 42. The reagent chosen for the d1 synthon 43 was again a dithian 44. O 1,2-diCO

Ph

O

Ph

S

+ O

HO 39

S

H

42; a 1 reagent

43; d 1 synthon

44; d 1 reagent

The synthesis starts simply with the coupling of the lithium derivative of 44 with 39 but takes a novel turn when a dithian exchange with pyruvic acid is used to remove the dithian from 45. The keto group in pyruvic acid is unstable, being next to another carbonyl group, while that in 39 is stable. Dithian exchange removes this instability by forming the pyruvic acid derivative 46. O 1. BuLi

44

Ph

S

S

O CO2H

2. 39

+

Ph

S

S

H

CO2H

HO

OH 45

39

46

The rest of the synthesis is straightforward: no protection is needed in the favourable Claisen ester condensation to give 47 and the formation of the enol ether, being intramolecular, requires only heating with acid. This treatment also hydrolyses the remaining ester. O

O (CO2Et)2

Ph HO 39

base

O

O H

Ph

CO2Et HO 47; not isolated

Ph

O

CO2H 37; Acifran, 92% yield

208

14 Acyl Anion Equivalents

Dithioacetal monoxides Recent developments have tried to remedy the difficult hydrolysis of the dithian. The monosulfoxides of dithioacetals are better at stabilising anions and more easily hydrolysed. They can be prepared by alkylation of the parent compound 50 with a suitable alkyl halide, e.g. to give 51.9 O 1. LiAlH 4

O O Ph

O 1. BuLi

MeS OBn

I

2. Ph 3P, I 2 imidazole

48

MeS

SMe 50

49

OBn

2. 49

SMe 51

The lithium derivatives are prepared with LDA and add to most electrophiles. Only direct addition occurs to the sensitive Z-enal 52 to give a mixture of diastereoisomers 53 from which the dithioacetal is easily removed under mild conditions to give the ketone 54. The product 54 was converted into the symmetrical triether10 55. O

MeS

BnO

SMe

1. 2 x LDA

O

TFA:THF:H 2O

BnO

51 1:10:10

2. 52

BnO

HO

H

O

BnO

20 ˚C, 2 hours

HO

53

H

O

54

OBn CHO

H

H O

54 O

O

O H

52

H 55

Protected Cyanohydrins of Aldehydes Unsymmetrical d1 reagents of this kind, i.e. 16; X ≠ Y, are best represented by the various cyanohydrin derivatives. Cyanohydrins themselves have an acidic H on the free OH group, but silylated cyanohydrins11 57, formed with Lewis acid12 or crown ether1 catalysis, lack this acidic proton and give lithium derivatives 58 with LDA. O R

catalytic

H

CN

NC R

O H

56

Me3SiCN

NC

OSiMe3

R

H 57

LDA

NC

OSiMe3

R

Li 58

These lithium derivatives 58 react with many electrophiles but are particularly suited to aldehydes and ketones as the developing oxyanion captures the Me3Si group 59 with loss of the CN group making this a much easier hydrolysis than that of a dithian. Treatment with fluoride then gives the hydroxy-ketone 61. An example is the coupling of the d1 reagent 62 from benzaldehyde with cyclopentanone to give the α-hydroxyketone 64 in a respectable 78% yield from benzaldehyde.

14 Protected Cyanohydrins of Aldehydes

O

O 58

SiMe3

O

O

NC

O

59

61 O

+ Ph

O

OSiMe3

PhCHO 2. LDA

OH

R

60 O

NC

Et3N.HF

OSiMe3

R

R

1. Me 3SiCN

209

Et3N.HF

OSiMe3

Ph

OH

Ph

Li 62

64; 78% yield

63

Amino nitriles are useful for conjugate addition For d1 reagents which will add Michael fashion to any unsaturated carbonyl compound, amino-nitriles are best. They were invented by Stork13 for this very purpose and used in his cis jasmone 65 synthesis. This TM has perhaps been more synthesised than any other, though its simple structure and obvious enone disconnection to 66 make it a rather uninspiring choice these days. The 1,4-diketone 66 does indeed cyclise to 65 presumably under thermodynamic control, and the double disconnection 66 [cf. the disconnection of pyrenophorin 23] requires a d1 reagent that will react with enones 67 in the Michael sense and with alkyl halides 69. O

O

O

1,4-diCO

aldol

O+

+

O

X 68; double

65; cis-jasmone

66

67

d1 synthon

69

The amino nitrile 70 is simply and cheaply made (90% yield) and does react with alkyl halides.14 Aldehydes are easily released from 71 with aqueous oxalic acid. After a second alkylation or Michael addition an even milder method of hydrolysis, aqueous Cu(II) salts at near neutrality, releases diketones such as 57. Cyclisation in base gives cis-jasmone13 65.

CH2O

NaCN

Et2N

Et2NH NaHSO3

1. LDA

Et2N

2. 69; R=I

H

CN

Et2N

1. LDA

CN

RCHO 2. RI

H2O

R 71

70

Et2N

CN

(CO2H)2

CN

1. LDA

Cu(II)SO4

2. 67

pH 5 MeOH/H2O

70

66 72

O

73

Acyl anion equivalents of the ester d1 synthon – CO2R At the ester oxidation level the best d1 reagent is probably Yamamoto’s ROCH(CN) 2. This reagent 74 (EE = EthoxyEthoxy) is easily prepared and reacts with alkyl halides using only the mild base K2CO3. Hydrolysis of the protecting group and displacement with a secondary amine gives the amides15 77. Reaction with tosylimines gives amino acids.

210

14 Acyl Anion Equivalents

CN

K2CO3

EEO

R

RX

CN

CN

EEO

74

TFA CH2Cl2

CN

+

R

75

O

CN

HO

R

N H

CN

N

76

77

Since the anion is so stable, it reliably does conjugate additions and provides a simple route to γ-keto acid derivatives.16 The initial adducts 78 can be worked up with either alcohols or amines to give γ-keto-esters 79 or γ-keto-amides 80. NC O

O

74

O

O

CN O

1. TFA, CH 2Cl2

OEE

K2CO3

OR or

O

NR2

2. ROH or R 2NH

78

79

80

Methods Based on Vinyl (Enol) Ethers and Enamines Just as anions of allyl derivatives can be homoenolate equivalents (chapter 13) so anions of vinyl derivatives can be acyl anion equivalents. Vinyl (or enol) ethers can be lithiated reasonably easily, especially when there is no possibility of forming an allyl derivative, as with the simplest compound 81. The most acidic proton is the one marked and the vinyl-lithium derivative 82 reacts with electrophiles to give the enol ether of the product17 84. However, tertiary butyl lithium is needed and compounds with γ-CHs usually end up as the chelated allyl-lithium 85. These vinyl-lithium compounds add directly to conjugated systems but the cuprates will do conjugate addition.18 H

t-BuLi

Li

E

E

OMe

OMe

OMe

81

82

E

H H2O

83

O 84

Li

OR 85

Lithium derivatives of cyclic vinyl ethers Cyclic vinyl ethers such a dihydropyran 86 form stable lithium derivatives 87 probably because isomerisation to structures like 85 is impossible. Reaction with electrophiles such as alkyl halides gives adducts, e.g. 88, and hydrolysis under very mild conditions reveals the carbonyl group19 89.

t-BuLi

O 86

THF, 0 ˚C

7.5% HCl/H2O

RX

O

Li 87

O 88

R

THF, 25 ˚C

HO

O 89

R

These reagents, usually with further functionalisation, have been widely used. Smith’s synthesis of phyllanthoside involved the coupling of the lithium derivative of 90 with the aldehyde 92 and the exposure of one protected ketone to give 93 having two exposed ketones and one concealed as an enol ether.20

14 Methods Based on Vinyl (Enol) Ethers and Enamines

H O

O

O

t-BuLi

OBn

O H,

Swern

OHC O

90

91

O

OMEM

Li

OBn

H 2O

+ O

H

then

O

THF, 0 ˚C

211

O

92

H

OMEM

93

The hidden carbonyl group is more apparent after the removal of the MEM protecting group 94 and formation of the acetal 95 in acid solution. One diastereoisomer of the spirocyclic compound 95 is formed in 71% yield. O

H

OBn

ZnBr2

O

O

H

O

cat. H

93 CH2Cl2

O

O O

H

OBn

H

OH

95; 71% yield

94

The synthesis of pederin and related anti-tumour agents The group of antitumour agents related to pederin have a common structural feature 96 of a functionalised tetrahydropyran attached through an amide linkage to a variety of complex amines. In his successful syntheses of these compounds, Kocienski21 chose to disconnect next to the ring and use a reagent 99 for the d1 synthon 97 that would be acylated by the oxalyl derivative 98. O

OMe O

O

O +

NHR

MeO

Li

O NHR

O

OH

SePh 96; pederin antitumour agents

97

98

99

The lithium reagent 99 was prepared from the stannane 100 - one advantage being that n-BuLi is strong enough to do the job. Addition to the ester 98 gave the ketone 101 and reduction and acetal formation gave the complete structure 102 from which the alkene 96 could be revealed by oxidation and selenoxide elimination (chapter 33). This synthesis looks very easy but the complicated amines ‘RNH2’ were very difficult to make. O O

SnMe3 BuLi

NHR

O 99

98

O

–80 ˚C THF

SePh 100

O

NHR

1. LiBH( s-Bu)3 2. H,

SePh 101

OMe O

MeOH

OH SePh 102

212

14 Acyl Anion Equivalents

Lithium derivatives of allenyl ethers Allenyl ethers similarly have no γ-protons so allyl anions cannot be formed. The allenyl ethers 104 are readily made from propargyl ethers 103 and are lithiated next to oxygen 105. Reaction with lactones gives adducts 106 that decompose in acid to the hemiacetal 107a of the 1,2-diketone 107b. The conjugated alkene has the Z-configuration as the enol ether in 106 prefers protonation on the side away from the R group. The allenyl lithium 105 is a reagent for the synthon22 108. R

R

R

R

HO •

t-BuOK

H

•

BuLi

OMe 104

OMe 103

O

•

O

OMe 106

R

O R

O R

O

OMe 105

HO

H2SO4

Li

O Z-107b

O Z-107a

OH 108

O

Oxidative Cleavage of Allenes Another way that allenes can be used to provide an acyl anion equivalent is by the Lewis acid catalysed addition of propargyl silanes to electrophiles 109 followed by oxidative cleavage of the allene. The intermediate vinyl cation 110 is stabilised by the silicon β-effect (chapter 12) and loss of the Me3Si group gives the allene 111. Oxidative cleavage of the sensitive allene reveals that it has acted as a reagent for the formyl anion. H R2

Me3Si R1 109

O

R

Me3Si

LA

2

R2

•

R1 OLA

R1 OH

110

R2

1. O 3

O 2. Me 2S

R1 OH

111

112

In their synthesis of hemibrevetoxin, Nelson23 converted the centro-symmetric acetal 113 into the dialdehyde 115 by this method. Yields of the bis-allene 114 and of the final product were excellent. Note that the acetal 113 acts as a suitable electrophile as the Lewis acid (Me3SiOTf) removes the OMe group to give an oxonium ion intermediate. MeO

Me

H

O

Me

Me3Si

• Me3Si

H 113

O

O

H 1. O 3

OTf

OMe Me

H

O

•

OHC

Me

O

H

2. Me 2S

Me

114; 92% yield

H

O

CHO Me

115; 98% yield

Vinyl Ethers and Enamines from Wittig-Style Reactions Vinyl ethers More general methods depend on Wittig reactions with functionalised ylids. The ylid24 from 116 and the lithium derivatives of 117-120 all react with aldehydes and ketones to give enol ethers that can be hydrolysed to chain-extended aldehydes. Yields with the ylid from 116 are not always wonderful and the phosphonate ester 120a with a chelating substituent generally does better.25

14 Vinyl Ethers and Enamines from Wittig-Style Reactions

213

In very crowded cases the lithium derivative of the phosphine 118, though it must be made with s-BuLi, is the most reactive.26 O Ph3P

OMe

O

Ph2P

116

OMe

Ph2P

117

OMe

(Me2N)2P

118 OMe

116

OEt

(RO)2P

119

120

(EtO)2P

H2O

OR

O

O

H

O base

O

O

H

OMe

120a

Enamines Lithium derivatives of amino substituted phosphine oxides, such as the morpholine 121, give adducts 122 on reaction with aldehydes that eliminate with a potassium base to give enamines 123 in excellent yields: 90–99% for R = Ar and 63–83% for R = alkyl. These are easily hydrolysed to the homologous aldehydes27 124. O O Ph2P

O N

1. BuLi

O

Ph2P

O KH

N

H

N

R

OH

CHO

R

122; intermediate

121

R

H2O

THF

2. RCHO 3. NH 4Cl H2O/THF

124; homologous aldehyde

123; enamine

The pyrrolidine phosphonate 125 does much the same thing in one step but the real virtue of these methods is that the enamines 123 and 126 are reactive enough to combine with various electrophiles, particularly allylic halides, α-halo-carbonyl compounds and Michael acceptors (see chapter 10) to give substituted homologous aldehydes or ketones28 127. O

1. LDA

(EtO)2P

N

E

1. E

2. RCHO

N

R

2. H 2O

R

126; enamine

125

CHO 127

A synthesis of O-methyljoubertiamine This method has been applied to a synthesis of the alkaloid O-methyljoubertiamine. Reaction of the lithium derivative of 125 with the part-protected diketone 128 gave an enamine 129 that was immediately allylated to create the quaternary centre in 130. Deprotection and cyclisation completed the cyclohexenone28 131. A discussion of the strategy and further details of this synthesis are in the workbook. OMe

OMe

OMe OMe 1.

125

O

O

O 128

1. H,

Br

N

CHO

2. H 2O

LDA

O

O 129

O

O 130

H 2O

2. KOH MeOH/H2O

O 131

214

14 Acyl Anion Equivalents

Synthesis of enals by functionalised Wittig reagents If there is a leaving group in the β-position of the carbonyl starting material 132 during homologation by the methoxymethyl reagents 116 or 117, it is lost during the hydrolysis of the enol ether product 133 to form an enal 134. OMe O

CHO

enol ether hydrolysis

116 or 117

X

X

132

133

134

The leaving group X may be only an OH group, inserted by hydroxylation of a silyl enol ether 136 (chapter 33) formed by conjugate addition of a silyl cuprate (chapters 9 and 10) to an enone 135 and protected by the robust silyl group TBDMS (t-BuMe2Si-) 137 for reaction with the lithium derivative of 117. O

OSiMe3

O 1. Me 2CuLi

1. mCPBA 2. Et 3NHF

2. Me 3SiCl

3. t-BuMe2SiCl imidazole

135

136

OTBDMS 137

The lithium derivative of 117 gives an adduct from which the phosphinate can be removed with NaH. Hydrolysis of the enol ether 138 with HF in aqueous acetonitrile gave the enal, e.g. 139 in around 50% overall yield.29 The overall strategy is conjugate addition of a nucleophilic alkyl group followed by direct addition of a nucleophilic CHO group. An alternative uses a sulfur leaving group (X = SPh in 132) and the ylid from 116 to achieve the same result. The final hydrolysis requires Hg(II) to remove the sulfur.30 CHO

OMe

O

HF, H 2O

1. LDA

Ph2P

OMe

OTBDMS

2. 137 3. NaH

117

138

MeCN

139

Nitroalkanes The problem with many acetal-style d1 reagents is often that there is barely enough stabilisation for the “anion”. This problem is easily solved with the nitro group as it has enough anion-stabilising ability by itself, about the same as two carbonyl groups in fact. Using only weak bases such as tertiary amines (Hünig’s base i-Pr2NEt is popular) nitroalkanes react 141 with many electrophiles such as alkyl halides, aldehydes or ketones, and unsaturated carbonyl compounds to give products such as 142, 144, 145 or 147. For a long time the problem with the nitro group was rather the eventual conversion to a carbonyl group, e.g. 142 to 143, but there are now many reagents for this reaction at the aldehyde, ketone, or acid oxidation levels.31 The examples which follow illustrate some of these.

14 Nitroalkanes

O

N

O

O

NO2 R2

NR3

H H 140

R1

O

N

R3N

215

O R2

R1

R1

R2

R1

142 141

143

X NO2 NO2

R2CHO base

R1

NO2

R2 O

145 O

R2

R1

base

R2

R1

OH 144 NO2

+ R1

NO2

R2

R1

R2

R1

O

146

O

147

148

Alkylation solves the awkward problem of how to make aryl ketones of the pattern 151, awkward because the Friedel-Crafts reaction doesn’t work with this substitution pattern. Alkylation of a nitroalkane with the benzylic halide 149 does work well and the product 150 can be oxidised to the ketone32 151. Reactions with aldehydes work even with masked aldehydes33 like dihydropyran to give 154 and hence the ketone 155 with a 1,2-substitution pattern after hydrolysis.34 More vigorous conditions give nitroalkenes 145 from aldehydes and these will be very useful later as a2 reagents. Cl

NO2

Cl

1. DBU Me3SiCl

base

+

NO2

Cl

149

O

2. mCPBA

Cl

150

151 1. Me2NH

base

+ O

OH

NO2

OH

2. HCl H2O

O NO2

152

O

NO2

153

O

154

155

Michael additions go particularly well with nitroalkanes with catalytic base because the fi rst formed enolate is more basic than the nitroalkane. Diketones, e.g. 159, keto-aldehydes and keto-acids can all be made by this route and the two steps can be combined in one using alumina to catalyse the Michael addition (158 is not isolated) and an oxidative work-up to release the ketone.34

NO2

Al2O3

O

NO2

H2O2

+

Bu

K2CO3

O O 156

157

158

O 159; 90% yield

216

14 Acyl Anion Equivalents

A synthesis of strigol using nitroalkanes When Raphael wanted the key intermediate 162 for his synthesis of strigol 160, the aldol disconnection 162a revealed the need for the diketone 163 and hence by disconnection of a very strategic bond, the d1 synthon 164. H

O

O

O

O CO2R

H

OR 161 O

O

OR 162

OH

OH 160; strigol

O

O

O

strategic bond disconnection

aldol

+ OR

OR

162a

OR

165 natural a reagent

164; d 1 reagent?

163

3

The nitroketone 166, though at the wrong oxidation level, is an ideal reagent for this purpose as it is not necessary to protect its ketone during the Michael addition. Conversion of the nitrocompound 168 into the ketone 169 with TiCl3 illustrates the generally preferred method for this reaction.35 O

O

O

O

165

O TiCl3

H

base

H2O

NO2

O

NO2 168

NO2 167

166

169

Synthesis of mono-protected diketones The nitro group can solve the problem of making a diketone with one of the ketones protected. When Hewson wished to make a series of natural products from the key intermediate 170 using his reagent 171 he needed the diketoester 172. Disconnection of a strategic bond reveals the possibility of adding a d1 reagent to the unsaturated ketoester36 174. MeO2C MeS

O

O

PPh3

O

+ Me

H 171

MeO2C

O

+

MeS

170

O

MeO2C

172

173

174

14 Nitroalkanes

217

Attempts with various reagents showed that the ketoester was inclined to polymerise unless it was used as its acetal 176 and that acetal style reagents for the d1 synthon 173 such as 175 gave poor results. The solution was to use nitroethane to give 177 having each keto group masked in a different form. Conversion of nitro group to ketone was better with TiCl3 than other reagents. The diketoester was formed as a single (trans) isomer which gave an excellent yield of 170 with the reagent 171. O

O MeS

Me

O O

MeO2C

EtNO2

O O

MeO2C

TiCl3

O

MeO2C

CF3CO2H

O2N

SMe

171

172

170 97% yield

H2O

base

O

175; cf. 50 176

177

178

Alternative and more general approaches to partly protected diketones include the Michael addition, protection, Michael addition sequence 179 to 182 and the Michael addition reduction sequence 183 to 186. The first provides a molecule 182 containing three ketones: one free, one protected as an acetal, and one masked as a nitro group. The acetal can be removed in dilute acid without hydrolysing the nitro group and the nitro group can be converted to a ketone without disturbing the acetal,31,34 e.g. with H2O2 and K2CO3. O

O

O

MeNO2

R1

base

HO

R1

OH

O

H

R2

O base

R1

NO2 179

180 O

181

NO2

O

O

R2

R1 NO2 182

The second has been used to make prostaglandins from the intermediate aldehyde 186. Note the stereochemistry: the obvious trans arrangement of the two substituents in 184 after the Michael addition and the approach of the bulky reducing agent L-selectride® Li(s-Bu3BH) from the opposite face to the nearer and larger of the two substituents. Finally, a reductive workup (Me2S) is needed after the oxidation of nitro to ketone with ozone.37 O

O CO2Et

MeNO2

CO2Et

LiBH(s-Bu)3

base

NO2 184

183 HO CO2Et NO2 185

1. t-BuMe2SiCl imidazole, DMF

TBDMSO CO2Et

2. NaOMe, MeOH 3. O 3, MeOH 4. Me 2S

CHO 186

218

14 Acyl Anion Equivalents

Catalytic Methods: The Stetter Reaction The useful Stetter reaction38 is biomimetic - the idea and the reagent came from the coenzyme thiamine pyrophosphate 187. Removing unnecessary substituents left the core of the thiazolium species 188 where R is usually benzyl (catalyst a) but can be various other primary alkyl groups. NH2

catalyst a: R = CH2Ph (benzyl)

R N

O O O O P

HO

P

O

N

N HO

N

S

187; thiamine pyrophosphate

others (b-d) R = Me, Et (CH2)2OEt

S

188; Stetter thiazolium salts

Reaction with an aldehyde starts by the removal of a proton with quite a weak base (usually a tertiary amine) to give the ylid 189 in which the negative charge is stabilised partly by electrostatic interaction (the ylid) and partly by the sulfur atom. Nucleophilic attack on the aldehyde gives 190 which can transfer a proton from C to O to give uncharged 191. Bn Et3N

188 R = Bn

Bn N

N

N

R2 R1

Bn

R1

S

R2 S

O

O

189; ylid

R2

R1

H

S OH

190

191

This is a strange compound. The exocyclic (arrowed) alkene is an enol at one end but an enamine and a vinyl sulfide at the other. All three atoms, O, N, and S, have lone pairs making this a very electron rich alkene. The nitrogen atom dominates so that conjugate addition 192 to enones works well. The product 194 loses the catalyst 189 to give a 1,4-diketone 195. The reaction is simplicity itself: the aldehyde, enone, catalyst and weak base (usually amine or sodium acetate) are heated in ethanol. Bn

O

N 2

R1

Bn

Bn

N R

O

R1

S

R2

R3

OH 192

R3

S

O

N R1

R3

S 2

R

OH

193

O

194 O

R2

O

H

cat. 188

+ O

R3

R2

R3

Et3N

189

O 195

Examples include enolisable aldehydes adding to enolisable enones to give e.g. 197, and an impressive range of aromatic and heterocyclic compounds as either component. The pyridine 200 is formed in only 12% yield if sodium cyanide is used as catalyst.38

14 References

219

O H

Me

O

cat. 188; R = Bn

+

OMe OMe

O

Et3N, dioxane

O 197; 80% yield

196 O

O

cat. 188; R = Et

+ N 198

Et3N, dioxane

CHO 199

N O 200; 75% yield

References General references are given on page 893 1. O. W. Lever, Tetrahedron, 1976, 32, 1943. 2. T. A. Hase, Umpoled Synthons, Wiley, New York, 1987. 3. A. S. Fletcher, K. Smith and K. Swaminathan, J. Chem. Soc., Perkin Trans. 1, 1977, 1881. 4. D. Seyferth and R. M. Weinstein, J. Am. Chem. Soc., 1982, 104, 5534; D. Seyferth, R. M. Weinstein and W.-L. Wang, J. Org. Chem., 1983, 48, 1144. 5. A. Ricci, M. Fiorenza, A. Degl’Innocenti, G. Seconi, P. Dembech, K. Witzgall and H.-J. Bestmann, Angew. Chem., Int. Ed., 1985, 24, 1068; A. Ricci, A. Degl’Innocenti, S. Chimichi, M. Fiorenza, G. Rossini and H.-J. Bestmann, J. Org. Chem., 1988, 50, 130. 6. P. C. B. Page, N. B. van Niel, and J. C. Prodger, Tetrahedron, 1989, 45, 7643. 7. D. Seebach, B. Seuring, H.-O. Kalinowski, W. Lubosch and B. Renger, Angew. Chem., Int. Ed., 1977, 16, 264; B. Seuring and D. Seebach, Helv. Chim. Acta, 1977, 60, 1175. 8. I. Jirkovsky and M. N. Cayen, J. Med. Chem., 1982, 25, 1154. 9. K. Fujiwara, K. Saka, D. Takaoka and A. Murai, Synlett., 1999, 1037. 10. K. Fujiwara, D. Takaoka, K. Kusumi, K. Kawai and A. Murai, Synlett., 2001, 691. 11. S. Hünig and G. Wehner, Synthesis, 1975, 391. 12. D. A. Evans, L. K. Truesdale and G. L. Carroll, J. Chem. Soc., Chem. Commun., 1973, 55. 13. G. Stork, A. A. Ozorio and A. Y. W. Leong, Tetrahedron Lett., 1978, 5175. 14. G. Büchi, P. H. Liang and H. Wüest, Tetrahedron Lett., 1978, 2763. 15. H. Nemoto, Y. Kubota and Y. Yamamoto, J. Org. Chem., 1990, 55, 4515. 16. Y. Kubota, H. Nemoto and Y. Yamamoto, J. Org. Chem., 1991, 56, 7195. 17. J. E. Baldwin, G. A. Höfle and O. W. Lever, J. Am. Chem. Soc., 1974, 96, 7125. 18. R. K. Boeckman, K. J. Bruza, J. E. Baldwin and O. W. Lever, J. Chem. Soc., Chem. Commun., 1975, 519; C. G. Chavdarian and C. H. Heathcock, J. Am. Chem. Soc., 1975, 97, 3822. 19. R. K. Boeckman and K. J. Bruza, Tetrahedron Lett., 1977, 4187. 20. A. B. Smith and R. A. Rivero, J. Am. Chem. Soc., 1978, 109, 1272. 21. P. Kocienski, R. Narquizian, P. Raubo, C. Smith, L. J. Farrugia, K. Muir and F. T. Boyle, J. Chem. Soc., Perkin Trans. 1, 2000, 2357. 22. P. Kocienski, R. C. D. Brown, A. Pommier, M. Procter and B. Schmidt, J. Chem. Soc., Perkin Trans. 1, 1998, 9. 23. J. M. Holland, M. Lewis and A. Nelson, Angew. Chem., Int. Ed., 2001, 40, 4082. 24. S. G. Levine, J. Am. Chem. Soc., 1958, 80, 6150; G. Wittig and M. Schlosser, Chem. Ber., 1961, 94, 1373; G. Wittig, W. Böll and K.-H. Krück, Chem. Ber., 1962, 95, 2514. 25. A. F. Kluge and I. S. Cloudsdale, J. Org. Chem., 1979, 44, 4847. 26. E. J. Corey and M. A. Tius, Tetrahedron Lett., 1980, 21, 3535. 27. N. L. J. M. Broekhof, F. L. Jonkers and A. van der Gen, Tetrahedron Lett., 1978, 2433. 28. S. F. Martin and T. A. Puckette, Tetrahedron Lett., 1978, 4229; S. F. Martin, G. W. Phillips, T. A. Puckette and J. A. Colapret, J. Am. Chem. Soc., 1980, 102, 5866. 29. T. K. Jones and S. E. Denmark, J. Org. Chem., 1985, 50, 4037.

220

14 Acyl Anion Equivalents

30. J. W. Huffman, M.-J. Wu and H. H. Joyner, J. Org. Chem., 1991, 56, 5224. 31. H. W. Pinnick, Org. React., 1990, 38, 655. 32. J. M. Aizpurua, M. Oiarbide and C. Palomo, Tetrahedron Lett., 1987, 28, 5361. 33. M. Lagrenée, Compt. Rend., 1977, 284C, 153. 34. R. Ballini, M. Petrini, E. Marcantoni and G. Rosini, Synthesis, 1988, 231. 35. G. MacAlpine, R. A. Raphael, A. Shaw, A. W. Taylor, and H.-J. Wild, J. Chem. Soc., Chem. Commun., 1974, 834; J. Chem. Soc., Perkin Trans. 1, 1976, 410. 36. A. T. Hewson and D. T. MacPherson, J. Chem. Soc., Perkin Trans. 1, 1985, 2625. 37. B. P. Cho, V. K. Chadha, G. C. LeBreton, and D. L. Venton, J. Org. Chem., 1986, 51, 4279. 38. H. Stetter and H. Kuhlmann, Org. React., 1991, 40, 407.

Section C: Carbon–Carbon Double Bonds 15. Synthesis of Double Bonds of Defined Stereochemistry ................................................. 223 16. Stereo-Controlled Vinyl Anion Equivalents .................................................................... 255 17. Electrophilic Attack on Alkenes ........................................................................................ 277 18. Vinyl Cations: Palladium–Catalysed C–C Coupling ..................................................... 307 19. Allyl Alcohols: Allyl Cation Equivalents (and More) ...................................................... 339

15

Synthesis of Double Bonds of Defined Stereochemistry Introduction: Alkenes: framework or functional groups? Control of Alkene Geometry by Equilibrium Methods Some examples from other chapters A More Detailed Look at The Principles Case 1: Only one alkene is possible. This is usually a cyclic cis alkene Case 2: The two alkenes are in equilibrium and the trans alkene is formed Case 3: The cis alkene is formed stereoselectively The Wittig Reaction The last step of the Wittig reaction is stereospecific The Z-selective Wittig reaction with simple ylids Case 4: The trans alkene is formed stereoselectively The trans-Selective Wittig Reaction Crossing the Stereochemical Divide in the Wittig Reaction The Schlosser modification Conjugated Z-alkenes Stereocontrolled Reactions The Horner-Wittig reaction with phosphine oxides The Peterson reaction Stereoselective Methods for E-Alkenes: The Julia Reaction Modified Julia reactions The Kocienski modification of the Julia reaction Direct Coupling of Carbonyl Compounds and Alkenes Carbonyl compounds: the McMurry reaction Alkenes: olefin metathesis Stereoselective Methods for E-Alkenes [3,3]-Sigmatropic rearrangements [2,3]-Sigmatropic rearrangements: the Wittig rearrangement Reduction of Alkynes Stereospecific Methods for Z-Alkenes Using cyclic compounds Interconversion of E and Z Alkenes Photochemical isomerisation to the Z-isomer A synthesis of methoxatin Radical isomerisation to the E-isomer Stereospecific Interconversion of E and Z-isomers

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

224

15 Synthesis of Double Bonds of Defined Stereochemistry

Introduction: Alkenes: framework or functional groups? The carbon-carbon double bond (olefin or alkene) is a structural feature on the border of framework and functionality. Unlike the carboxylic acid in 1 it is inside the skeleton of the molecule and involves no atoms except carbon and hydrogen. It has no polarity and yet it has the potential of producing highly polar functionality at two carbon atoms in a single reaction, as in the iodolactonisation to give 2 from 1. O O

CO2H

I2

1

2

NaHCO3

I

In the same reaction stereochemistry may be generated: two stereogenic centres are in fact produced in this reaction and in the epoxidation of 3. The sense of the stereochemistry produced in these reactions depends critically on two factors: the inherent stereoselectivity of the reactions (anti in the iodolactonisation and syn in the epoxidation) and the geometry of the alkene. Inside a six-membered ring of 1 the alkene must of course by cis or Z. In the open chain allylic alcohol 3 it can be E or Z depending on the method of synthesis. This chapter explores ways to control the stereochemistry of alkenes as an essential preliminary to the control of three-dimensional stereochemistry in chapter 25 where you will meet a method to make single enantiomers of 4 from the alkenes in E- and Z-3.

R

OH

R

O

mCPBA

R

E-3

R mCPBA

OH

H

anti-4

OH

O

H

OH

Z-3

syn-4

Control of Alkene Geometry by Equilibrium Methods In many reactions there is an equilibrium between reactants and products so that only the more stable of the two alkenes is produced. In the Claisen ester condensation of cyclohexanone 8 with ethyl formate, the true product under the conditions of the reaction is the stable enolate 9 and this is reversibly protonated on workup to give the more stable H-bonded enol of the ketoaldehyde Z-10. In the aldol reaction between the same ketone and benzaldehyde, the initial product 7 gives the enolate 6 and dehydration is reversible: only the more sterically favourable E isomer of 5 is formed. Note that it is irrelevant that the aldol 7 is a mixture of diastereoisomers: all stereochemistry is lost in the formation of the enolate 6. In later parts of this chapter a more specific relationship between 3D and 2D stereochemistry will be established.

O

O

HO

O

O

O

E-5

Ph

6

Ph

7

O

H

HCO2Et

PhCHO

Ph

O

O

NaOH

EtOH

EtO

8

9

Z-10

O

15 Control of Alkene Geometry by Equilibrium Methods

225

Some examples from other chapters There are very many reactions which routinely give E-alkenes as the more stable product. You will already be aware of many of them and there are many scattered around the other chapters in this book. We shall simply give a few more examples here and a list of cross references to other examples. 1. Aldol reactions followed by dehydration, e.g. enones E-11 and E,E-12. We shall meet E, E-12; Ar ⫽ Ph later as a useful ligand for palladium.1 O

O

O

ArCHO

ArCHO

Ar

NaOH

Ar

NaOH

E-11

acetone

Ar

E,E-12

2. The base-catalysed condensation of nitromethane with aldehydes to give nitroalkenes E-13 (the Henry reaction). One example E-14 was used in a Corey prostaglandin synthesis.2 RCHO, base

CH3NO2 nitromethane

R

NO2

NO2

NC

E-13

E-14

3. Dehydration of tertiary alcohols 16 formed by the addition of organo-metallic species to ketones 15. An example is from Corey’s erythronolide synthesis where dehydration of 18 gives only the E-enyne3 19. OH

R2MgX

R1

O

or

R1

15

cat H

R2

R2Li

(TsOH etc)

R1

R2

or SOCl2, pyridine

E-17

16 OH Li

cat TsOH

O 15; R 1 = Et

E-19

18

4. The Heck reaction with α,β-unsaturated carbonyl compounds. The intermediate Pd(II) σ-complex 20 eliminates PdHI in a cis fashion and prefers to give the E-alkene 21. PdI ArI

CO2Et

Ar

syn-elimination of PdHI

Ar

CO2Et

Pd(0)

CO2Et

E-21

20

5. Nucleophilic addition to α,β-unsaturated alkynes under equilibrating conditions. This example comes from a recipe in Organic Syntheses where a good yield of crystalline E-22 is reported.4

25% aqueous Me3N

CO2Et

40:1 water : CH 2Cl2

Me3N

CO2 E-22; 76-82% yield

226

15 Synthesis of Double Bonds of Defined Stereochemistry

6. Equilibration of allylic alcohols, halides etc via an allyl cation. Whichever alcohol 23 or 26 is used as starting material, treatment with HBr gives the same equilibrium mixture of all E-24 and 25. No doubt any Z-24 has also equilibrated to E-24. Br HBr

OH

Br E-24; 80-87%

E-23

OH HBr

+ 25; 13-20%

26

7. Formation and equilibration of allylic metal derivatives such as those from the allylic bromides 24 and 25 we have just made. Reaction of the Grignard reagent with CO2 gives only 27 but displacement of bromide with cyanide leads to the isomeric acid E-28 both in good yield.5 CO2H

1. Mg, Et 2O 2. conc. HCl

CN CO2H + 28; 8.5 parts 29; 8.5 parts 70% yield

1. CuCN

E-24 + 25

2. conc. HCl

27; 75% yield

A More Detailed Look at The Principles This simple picture hides a number of principles which we shall need to explore in this chapter. In outline we can distinguish these cases: 1. 2. 3. 4. 5.

Only one alkene is possible. This is usually a cyclic cis alkene. The two alkenes are in equilibrium and the trans alkene is formed. The cis alkene is formed stereoselectively. The trans alkene is formed stereoselectively. The reaction is stereospecific. The geometry of the alkene depends on the mechanism of the reaction and the stereochemistry of the starting material.

Case 1: Only one alkene is possible. This is usually a cyclic cis alkene We shall deal with all the cases and explain the principles behind them. We need spend no time on case 1. Alkenes in 3- to 7-membered rings must be cis as the trans alkenes cannot be formed. Medium rings from 8-membered upwards can form trans alkenes but they are less stable. Only from (about) 12-membered rings upwards is the trans alkene the more stable isomer. This simple Robinson annelation (chapter 5) produces a cis cyclohexene 33 because that alone is possible. The step in which the double bond appears is the E1cB elimination 32 at the end. The enolate will always have one lobe of the p-orbital shown in the conformational diagram correctly aligned to expel the hydroxide leaving group and form the cis alkene. O

O

O

O

O HO

OHC

OHC 30

O HO

HO 31

32

33

conformation of E1cB on 32

15 A More Detailed Look at The Principles

227

Each of the following reactions gives exclusively the cis (actually E or Z as shown) alkenes 35, 36, and 38 for the same simple reason: trans alkenes cannot exist in four-, five- or sevenmembered rings. The reactions are from other chapters of the book. Ph

OSiMe3

CO2Me

O

Na

HN

OH

N

Ph H

Me3SiCl

CO2Me

OSiMe3 Z-35

34

E-36

E-38

37

The first reaction is the silicon-directed acyloin the second is simple enamine formation from chapter 2 and the third reaction is a simple E1 elimination. The same reaction will start off our next section concerning alkenes in equilibrium.

Case 2: The two alkenes are in equilibrium and the trans alkene is formed We can go back in time to some very famous work by Cram which led him to deduce his famous rule.6 Reaction of either diastereoisomer of the (racemic) tertiary alcohol 39 with acid gives the same mixture of alkenes in which the trans alkene E-40 predominates (⬎50:1) to a very great extent. The elimination is E1 via the tertiary carbocation 43. Me

Me Ph

Ph

Ph

Ph

AcOH, 75 °C

OH 39; syn alcohol

cat. TsOH AcOH, 75 °C

H 40; trans E-alkene

H Me

Ph

Ph

OH 41; anti alcohol H

H

Ph

Ph

Me

cat. TsOH

Me

Me Ph

Ph

OH2

Ph

Ph

H

42

OH2

43

44

Each alkene can be made stereospecifically (the synthesis of the Z-isomer 45 is given later in this chapter) and treatment of the cis alkene 45 under the same acidic conditions also gives the trans alkene 40. The alkenes are in equilibrium via the stable tertiary carbocation 43. The cation has two conformations 43a and 43b that can eliminate to give the E-alkene 40 and free rotation about the central σ-bond in this relatively long-lived intermediate 43 ensures that one of these conformations is accessible from all possible starting materials.

Ph

Ph

OH 39; syn alcohol

H H2O

H

H

Me

H2O

Ph 45; cis Z-alkene

H

–H

43

H

–H

Ph

Ph

H 40; trans E-alkene

Ph

Ph

Me H 43a; conformation of cation 43

Me

H H

Ph

Ph

OH 41; anti alcohol

Ph

Ph

Me Ph

H

Me

Me

Ph

Me Ph

H H 43b; conformation of cation 43

228

15 Synthesis of Double Bonds of Defined Stereochemistry

Such perfect cases of equilibration are rare. Another important one concerns equilibration by Michael addition to conjugated alkenes. A classic case is the preparation of maleate 47 and fumarate 48 esters from (the necessarily Z-) maleic anhydride 46. Simple treatment with methanol in acidic solution gives the liquid maleate 47 as expected. It is unusual for cis and trans alkenes to be given different names: maleic and fumaric acids were named before their relationship was understood. O CO2Me

MeOH

CO2Me

cat. R 2NH

O H

CO2Me

O 46; maleic anhydride

MeO2C

47; dimethyl maleate b.p. 204-205 ˚C

48; dimethyl fumarate m.p. 103-104 ˚C

Slowly in solution crystals are formed and these are the more stable fumarate ester. The process is enormously accelerated by adding a trace of an amine and must occur by reversible Michael addition of the amine or even of traces of methanol or any other stray nucleophile.7 O OMe Nu

O

O OMe Nu

CO2Me 49

CO2Me

180° rotation

OMe MeO2C

CO2Me 50

Nu

MeO2C

51

48

Once again a long-lived intermediate - the enolate ion 50 - allows free rotation about the central σ-bond and the elimination of the catalyst from a more favourable conformation 51 leading to the E-alkene 48. The same process occurs during the dehydration step following aldol reactions which occurs by an E1cB mechanism with the same sort of intermediate (e.g. formation of E-5, E-11 and E,E-12). This process is so easy that it is actually very difficult to make cis alkenes conjugated to the more reactive electrophilic groups such as aldehydes.

Case 3: The cis alkene is formed stereoselectively You might think that this case would be very rare. Why should any reaction choose to form the less stable cis alkene? There are only a few simple cases, it is true, but there is one extremely important more complex case too. So important is this case that it is the first choice alkene synthesis in most people’s mind. We shall look first at the simple Z-selective reactions. We can start with a Michael addition. This sounds crazy as we have just explained that Michael addition leads to rapid equilibration. However, Michael addition to triple bonds may not. One remarkably simple case is the addition of LiBr in weakly acidic solution to a triple bond conjugated to an ester 52.

CO2Et 52

LiBr HOAc, MeCN

Br

CO2Et

Z-53

85% yield Z-isomer only no trace of E-53

Only the Z-alkene 53 is formed. The difference here is in the structure of the enol intermediate 55. There is no free rotation. Instead the two cumulated (allene style) double bonds are held rigidly

15 A More Detailed Look at The Principles

229

at right angles to each other. Protonation occurs on the lobe of the p-orbital away from the bromine atom 55 and the Z-alkene 53 results. This is clearly kinetic control.8

O

Br

CO2Et

Br

OH OEt

•

Br H

OEt

H 54

Z-53

55

A more remarkable case is the formation of predominantly Z-enediyne 57 during the base catalysed dimerisation of the propargyl bromide 56. This is important as well as remarkable as the enediyne antibiotics contain this structural feature which has to be Z for biological activity.9 SiMe3 SiMe3 Br

Me3Si

LiHMDS 58

SiMe3

SiMe3 Z-57; 94% yield, 2:1 Z:E

Me2N Me2N

Li

HMPA 59

56

N

P

O

Me2N

58; LiHMDS Lithium HexaMethyl DiSilazide

59; HMPA HexaMethyl PhosphorAmide

You might say that the selectivity is “only” 2:1 but in truth it is remarkable that the reaction is Z-selective at all. In practice 60% Z-57 can be isolated by chromatography. The obvious mechanism is the formation of a lithium derivative, SN2 coupling 60, and base-catalysed E2 elimination 62. SiMe3

SiMe3 Br

LiHMDS Br

LiHMDS

56 Li

H

Br 60

SiMe3

SiMe3 Br 57 Li

SiMe3 61

SiMe3 62

For this mechanism to deliver Z-57 the conformation (or even configuration?) of the intermediate must favour 62 which looks rather unlikely. However, the reaction does give predominantly Z-57. Maybe the two long thin alkynes actually attract rather than repel one another 62a. The original authors offer a different (carbenoid) explanation but without much conviction. These examples both concern triple bonds. But please do not think that all reactions of alkynes are Z-selective. A salutary lesson is hidden in another volume of Organic Syntheses in a nucleophilic addition to the same alkyne as in our first example. Evidently under these polar hydrophilic conditions equilibration occurs by reversible protonation.10

25% aqueous Me3N

CO2Et

40:1 water : CH 2Cl2

Me3N

CO2 E-22; 76-82% yield

230

15 Synthesis of Double Bonds of Defined Stereochemistry

The Wittig Reaction Now to the great Z-selective alkene synthesis. The Wittig reaction. Along with the Diels-Alder and the aldol reactions this is one of the most important reactions of all time. You probably have an idea of the basic reaction:

O CH3I

base

Ph3P

Ph3P phosphine

CH3

Ph3P

phosphonium salt

CH2 +

63; ylid(e)

Wittig reaction

+ Ph3P O phosphine 64; alkene oxide

We have chosen this example to illustrate the first virtue of the Wittig reaction: you know where the alkene will be in the product 64. Even if, as here, it is less stable than some other possibility (the endocyclic alkene 66 that would be formed by dehydration of the tertiary alcohol 67), the new alkene must be between the carbon atoms formerly occupied by the phosphonium ylid and the carbonyl group. This is because the last step is the regiospecific decomposition of a fourmembered ring (an oxaphosphetane) 65.

OH

PPh3

O Ph3P

O

CH2

H

63; ylid(e)

64; alkene

65; oxaphosphetane

67; t-alcohol

66; alkene

The last step of the Wittig reaction is stereospecific This regiospecificity has stereochemical consequences too. The geometry of the alkene product 70 is not decided in the elimination of Ph3P⫽O from the oxaphosphetane 69. If this step is a concerted cis elimination then the stereochemistry of the oxaphosphetane is faithfully reproduced in the alkene product. Here is an example:

R1 R1

PPh3

1. base 2. R2

68; phosphonium salt

R1 PPh3

O

O R2 69; oxaphosphetane

+ R2

PPh3 O

70; Z-alkene

The Z-selective Wittig reaction with simple ylids If the syn oxaphosphetane 69 is favoured, the Z-alkene 70 must be formed. And it is! The formation of the oxaphosphetane is syn selective and the elimination step is stereospecific. The selectivity varies, but with R⫽Alkyl it is usually quite good. Here is an example - the synthesis of the sex attractant 75 of the Gypsy moth. This is the epoxide of a long chain cis alkene Z-74 easily made by a Wittig reaction with excellent Z-selectivity.11

15 The Wittig Reaction

231

Ph3P

Br

PPh3

MeCN

BuLi

PPh3

DMSO

71

72; 86% yield

73 mCPBA

RCHO

74; 91% yield Z-alkene isolated O 75; disparlure: sex attractant for the gypsy moth

This reaction is obviously of the greatest significance. Yet the detailed mechanism of the formation of the oxaphosphetane is not agreed. No previous intermediates have been isolated and one of the two main contenders is a one-step mechanism 76. This mechanism has the advantage of a believable explanation for the stereoselectivity. A 2⫹2 cycloaddition must have an antarafacial component - it must be a π2s ⫹ π2a reaction 76a.

2+2 cycloaddition

O

Ph3P

R2

R1 76

Ph3P

Ph

O

R1

flatten out

R

2

O R1

69

Ph

Ph P H H 69a

Ph

Ph O

R2

R1

Ph

Ph P H H 77

Ph P

Ph

O

R2

R2

R1 76a

As the two components approach each other at 90⬚ 76a, the two substituents keep out of each other’s way 77 and so, when you flatten out the four-membered ring 69a, they end up on the same side 69. Part of the evidence for this mechanism is that salt-free ylids give the highest Z-selectivity.12 Normal preparations of ylids by treatment of, say, a phosphonium bromide with BuLi, have a molecule of salt (here LiBr) inevitably present. But not this way: 1. NaNH 2, NH 3(l)

Ph3P 78

EtCHO

Ph3P

2. extract ylid with benzene

80; 97:3 E:Z

79; pure ylid - no NaBr

This procedure was used in the synthesis of the upper chain of a prostaglandin 82 (see chapter 6) where the stereochemistry is related to the biological activity.13 Note that the aldehyde is tied up as a hemiacetal in the starting material 81 (see chemoselectivity chapter) and that the “salt-free” ylid is actually a carboxylate anion made with sodium derivative of DMSO in DMSO. As you will see in the next section, not all Wittig reactions are cis selective - those of stabilised ylids are trans selective. OH HO O

Ph3P

CO2Li

CO2

ylid made with O

THPO

THPO

S

OTHP 81

Me

CH2

OTHP 82; 80% yield, all Z

232

15 Synthesis of Double Bonds of Defined Stereochemistry

Case 4: The trans alkene is formed stereoselectively Many E2 reactions fit into this category. A simple base-catalysed elimination of an alkyl bromide will normally give the E-alkene because there are two diastereotopic hydrogen atoms which could be lost and the one which gives the E-alkene is lost more quickly because that transition state has the lower energy. This is a kinetic effect on a reaction in which the transition state resembles the product. H

HA

H Ph

Ph

base

Ph

Ph

Br 83

HB

B

Br 83a

E-84; trans-stilbene

E2

Ph

Ph

antielimination

Ph

Ph

E-84; trans-stilbene

The same arguments apply to most E1cB reactions (see 32) where there are two conformations of the enolate during elimination - one leading to the Z- and the other to the E-alkene. The second is preferred in acyclic compounds as it gives the more stable transition state.

The trans-Selective Wittig Reaction Simple phosphonium ylids give Z-selective olefination. Ylids in which the carbanion is stabilised not only by the P⫹ atom but also by a conjugating group, particularly a carbonyl group, give E-selective reactions. The extra stabilisation is delocalisation of the enolate kind 85b. This simple example 85 with an aldehyde group is a stable ylid which can be prepared in and even recrystallised from water and is available from Aldrich. It reacts with aldehydes and ketones to give E-enals14 86. H Ph3P

H PhCHO

Ph3P

O

85a; ylid

O

85b; enolate

Ph

CHO

86; 99% yield, all E

A sequence using this reagent 85 followed by a Wittig reaction with an unstabilised ylid is a good route to E,Z-dienes that are often found as insect pheromones. This sequence was used in the synthesis of bombykol 90, the pheromone of the silk worm moth15 Bombyx mori. The first Wittig gives less than 5% Z-88 and the second, using a salt-free ylid, less than 3.5% E,E-89. OHC

85

CO2Me 1. (Me 3Si)2N 2. E-88

Z, E -89

CO2Me

E-88;

87 PPh3

OHC

Z, E -89

CO2Me OH

LiAlH4

Z, E -90; bombykol, 92% yield

The extra stabilisation makes the ylid rather unreactive and phosphonate esters 91 are often used instead of phosphonium salts in these reactions. Treatment with a base (NaH or RO⫺ is often used, BuLi will certainly not do) gives an inherently more reactive enolate anion 92 rather than an ylid. These Horner-Wadsworth-Emmons reagents (HWE as we shall call them, though they go under many other names) react with ketones as well as aldehydes and the product is normally the E-alkene16 93.

15 The trans-Selective Wittig Reaction

O

O

(EtO)2P

EtO or NaH

OEt

O

O

233

O

O RCHO

(EtO)2P

91

OEt

(EtO)2P

92a

CO2Et

R

OEt

E-93

92b

Here are two examples - one with an ylid and an aldehyde and one with a phosphonate ester and a ketone. Both give esters of conjugated unsaturated acids and both give the E-isomer. The first gave an excellent yield of E-95 with good selectivity (15:1) even though the new alkene is trisubstituted. It was used in the synthesis of the terpene eremophilone.17

Ph3P

CHO

O

CO2Me

CO2Me

O

O

O

O

E-95; 91% yield (15:1 E:Z)

94

eremophilone

The second gave the trisubstituted alkene E-98 (together with 9% of the Z-isomer) and this was used to make the interesting millipede compound polyzonimine.18 91

O

CO2Et +

NaH, DME

96

N

CO2Et

E-98 (9% Z-98 also formed)

97

polyzonimine

So what makes the dramatic difference between these reactions and those of the unstabilised ylids? Well, again there is no agreement, but we offer one explanation. Supposing the reaction is again stereoselective in favour of the syn oxaphosphetane 101, but this time the reaction can reverse as the starting ylid (or carbanion) is so much more stable. The more abundant syn-100 cyclises to syn-101 more slowly than does the less abundant anti-100 to anti-101 and, in turn, anti101 eliminates more rapidly than syn-101. Ph3P

CO2Me

slow

Ph3P cyclisation

Ph3P

CO2Me

R

O

O R syn-100; major adduct or betaine

Ph3P

CO2Me

elimination

O

99 cyclisation

Ph3P O

CO2Me

slow

H R Z-102; less stable cis-alkene

R syn-101; less stable oxaphosphetane

fast

O R anti-100; minor adduct or betaine

CO2Me

CO2Me

CO2Me

R anti-101; more stable oxaphosphetane

fast elimination

R H E-102; more stable trans-alkene

This is more than just an explanation of the trans selectivity. It is an alternative (many would say the correct!) mechanism for the Wittig reaction. We shall not enter the debate about whether the “adduct” (or betaine) 100 is a true intermediate. There is no doubt that the oxaphosphetane 101 is a true intermediate. This explanation is more convincing because it also allows us to explain the reactions in the next section.

234

15 Synthesis of Double Bonds of Defined Stereochemistry

Crossing the Stereochemical Divide in the Wittig Reaction The Schlosser modification The Wittig reaction is so powerful that we should like to cross into the forbidden zones - to make E-alkenes from non-stabilised ylids and to make Z-alkenes from stabilised ylids. Both are possible. The stability of the two oxaphosphetanes just considered gives us a clue as to how to accomplish the first of these. Schlosser argued that if we could equilibrate the two series without going back to the starting materials we could cross from the kinetically favoured syn betaine or oxaphosphetane to the more stable anti series. He did this by carrying out the reaction at low temperature and treating with a second molecule of base before the elimination occurred. The oxido-ylid 107 can be formed from both diastereoisomers of either the betaine 104 (if it is an intermediate) or oxaphosphetane 105. Reprotonation preferentially gives the more stable anti-104 or 105 and hence the E-alkene 106 on warming. Ph3P

H

H

Alk

slow

Ph3P cyclisation

slow elimination

O

H R Z-106; less stable cis-alkene

R syn-105; less stable oxaphosphetane

O R syn-104; major adduct or betaine Ph3P

Alk

Alk

Alk Alk 107; oxido-ylid formed reversibly with a second molecule of base R

Ph3P R

O

O

103 Ph3P

H

H Alk

fast

Ph3P

Alk

Alk fast

O

cyclisation

R anti-105; more stable oxaphosphetane

O R anti-104; minor adduct or betaine

elimination

R H E-106; more stable trans-alkene

As an example, the synthesis of oct-2-ene by the normal Wittig procedure gives an 80:20 Z:E mixture. By the Schlosser modification, set out in detail below, virtually pure (99:1) E-oct-2-ene is formed in good yield in a one-pot process. Notice the low temperature used when the aldehyde is added. This is necessary to avoid completion of the Wittig. After epimerisation with a second molecule of PhLi, the oxido ylid 107 is acidified to give the alcohol syn-109 that eliminates quickly with a potassium base.19 The Schlosser-Wittig Sequence for making simple E-Alkenes

Ph3P

Alk

PhLi

Ph3P

Alk

THF/Et2O

RCHO

Ph3P

H

Alk

warm

Alk

–78 °C

LiO R syn-104

R Z-106

108; ylid PhLi

Ph3P

Alk

HCl

Ph3P

H

Alk

Et2O

THF/Et2O

LiO R 107; oxido-ylid

Alk

t-BuOK

example:

t-BuOH

LiO R anti-109

R

E-106

70% yield, 99:1 E:Z

15 Crossing the Stereochemical Divide in the Wittig Reaction

235

A more serious example is from Johnson’s biomimetic polyolefin cyclisation. He wished to make the polyene 111 for potential cyclisation to 110, a 5-6-6-5 ring system analogous to the natural steroids. After the obvious aldol disconnection, it is best to disconnect the trans-alkene 112 in the middle of the molecule. O O H

Wittig

aldol

?

?

H O

H

H

O

110

O 111

112

This requires a Schlosser modification of the Wittig reaction and in theory the CHO and PPh3 groups could be placed on either fragment. There is, as we shall soon see, a strong reason for wishing to make the trisubstituted E-alkene by a [3,3]-sigmatropic rearrangement that gives an alk-4-enal and so this is the preferred analysis:

O

O

PPh3

Wittig

+ CHO

O

O 112

113; protect ketones

114; γ,δ-unsaturated aldehyde containing E-trisubstituted alkene

The Wittig reaction was carried out by the Schlosser method - the ylid was generated with PhLi and the aldehyde added at low temperature (⫺70 ⬚C). A second equivalent of PhLi was added and the intermediates allowed to equilibrate at ⫺30 ⬚C. Elimination of Ph3P⫽O occurred under these conditions to give the E-alkene. Deprotection and aldol condensation gave the cyclopentenone in a very impressive 46% yield over the five steps from the original aldehyde.

O

O PPh3

1. PhLi, –70 °C

O

O

1. HCl, H 2O MeOH

2. RCHO, –70 °C 3. PhLi, –30 °C

O

2. 2% NaOH in water

O O 115

111 46% yield from RCHO

O

E-116

Notice the control inherent in the aldol cyclisation. One from four possible enolates attack one of two carbonyl compounds. This is clearly thermodynamic control under these weakly basic conditions (chapter 5?). The more highly substituted alkene (here tetra-substituted) and the most stable possible ring (i.e. 5- not 3- or 4-membered) is formed. The geometry of the alkene is of course controlled by the ring.20

236

15 Synthesis of Double Bonds of Defined Stereochemistry

Conjugated Z-alkenes A typical Horner-Wadsworth-Emmons synthesis of a conjugated alkene would involve a phosphonate ester 92 and an aldehyde and would be extremely E-selective for E-93. This selectivity relies on the reversal of the reaction leading to the major adduct 100, so if we want to make this reaction Z-selective, we have to make the cyclisation of the major adduct faster. The black spot marks where the acceleration is needed. Once the syn-oxaphosphetane 101 is the major (or only) product, the Z-alkene 102 must be formed as the elimination is stereospecific. Ph3P

CO2Me

Ph3P

CO2Me

Ph3P cyclisation

O R syn-100; major adduct or betaine

R

O 99

CO2Me

CO2Me

slow

slow elimination

O

R syn-101; less stable oxaphosphetane

H R Z-102; less stable cis-alkene

This problem was solved by Still and Gennari with strongly electron-withdrawing groups (CF3CH2O-) on the phosphorus atom. Nucleophilic attack at phosphorus is speeded up and the reaction becomes Z-selective. Conditions have been chosen to get the best Z-selectivity and we can contrast the two results with the same aldehyde.21 O

CO2Me

Ph E-117; >95% yield >50:1 E:Z

O

(MeO)2P

CO2Me

(CF3CH2O)2P

PhCHO MeO, MeOH

CO2Me

(Me3Si)2NK, 18-crown-6, THF

Ph CO2Me Z-117; >95% yield >50:1 Z:E

Amazingly, this reaction works well when trisubstituted alkenes are needed - one of the few methods which does - and we can summarise that situation in a similar picture. CO2Me

Ph

O

O

Me (MeO)2P

CO2Me

Me

(CF3CH2O)2P

PhCHO MeO, MeOH

Me

Me CO2Me

(Me3Si)2NK, 18-crown-6, THF

E-118; 83% yield 22:1 E:Z

Ph

CO2Me

Z-118; >95% yield 30:1 Z:E

Stereocontrolled Reactions These two versions of the HWE are close to stereochemical control: the formation of either isomer (E or Z) at will from (more or less) the same starting materials. The next two reactions achieve this aim. By purification of Wittig-type intermediates the stereospecific elimination gives a single isomer of the alkene.

The Horner-Wittig reaction with phosphine oxides Treatment of a phosphonium salt 68 with aqueous base instead of the anhydrous bases used in the Wittig reaction leads to a stable phosphine oxide 119. The lithium derivative 120 contains a genuine C–Li bond and reacts stereoselectively with aldehydes to give stable adducts 121 that in turn gives the highly crystalline alcohols syn-122, easily purified by crystallisation and, if necessary, chromatography.22

15 Stereocontrolled Reactions

237

O O NaOH

R1

Ph3P

Ph2P

H2O

O R1

BuLi

Ph2P

THF

119 phosphine oxide

68; phosphonium salt

Li

R1

R2CHO

Ph2P

O R1

NH4Cl

R1

Ph2P

H 2O

R2 LiO 121

120

R2 HO syn-122

When one pure diastereoisomer of 122 is treated with a sodium or potassium base, a Wittigstyle elimination occurs on the anion 123 via the four-membered ring 124. The stereochemistry of 122 is not affected by this reaction so syn-122 must give Z-alkene 122. The by-product is water-soluble Ph2PO2⫺ rather than the sometimes difficult to remove Ph3PO formed in the Wittig reaction. O

O R1

Ph2P

O R1

Ph2P

NaH, DMF

R1

Ph2P

or KOH, DMSO

O

R2 HO syn-122

R2 O 123

R2

124

O R1 Ph2P O + 125; waterR2 soluble Z-126 by-product

If the E-alkene is required, reduction of the ketones 127, prepared by acylation of 119 or by oxidation of mixtures of isomers of 122, with NaBH4 in alcoholic solution gives anti-122 selectively. Stereospecific elimination then gives exclusively E-126. The reduction is controlled by Felkin-Anh selectivity (see chapter 21). O

O

R2 H

[H]

1. BuLi

119

O R1

Ph2P

2. R 2CO2R

Ph2P R1

O

[O], e.g. PDC

122

Felkin-Anh selectivity

R2 O 127

NaBH4

Ph2P

R1

R2 HO anti-122

R1 NaH, DMF or KOH, DMSO

125 + R2

E-126

A simple example is the stereochemically controlled synthesis of the two isomers of isosaffrole 130 from the phosphine oxide 128. Two steps give Z-130 in 64% yield while three steps give E-130 in 70% yield. It is a matter of judgement whether the inevitable sacrifice of yield in a twoor three-step process is worth the formation of single geometrical isomer rather than the mixture obtained from the one-step Wittig or HWE reaction.23 More complex molecules formed by the Horner-Wittig reaction include brevetoxin24 A and the immunosuppressant25 FK506. O Ph2P

O

NaH

1. BuLi

Ph2P

2. ArCHO

HO

128 1. BuLi 2. ArCO 2Me

O

O

Ph2P

Ph2P NaBH4

O O

131; 85% yield

O O Z-130; 84% yield

O syn-129; 76% yield

O

O

DMF

NaH

HO

O O anti-129; 91% yield

DMF

O O E-130; 91% yield

238

15 Synthesis of Double Bonds of Defined Stereochemistry

The intermediate 122 is a stable compound and can be made by any chemistry that gives simple alcohols. Acylation of 128 with a lactone gives the hydroxy ketone 132 and reduction gives the anti-diol 133 in excellent yield. Elimination gives pure E-134, a pheromone of the Mediterranean fruit fly. O 128

O

O

Ph2P

Me

NaBH4

BuLi

O

Ph2P

OH

O

Me OH

HO

132, 81% yield

133, 85% yield

NaH

OH

DMF

E-134, 98% yield

The γ-ketone 135 can be made by reaction of 128 with cyclohexene oxide followed by oxidation. Baeyer-Villiger rearrangement gives mostly (24:1) the lactone 136 with the right regiochemistry for elimination after hydrolysis to give26 only Z-137. O

O

O

CF3CO3H

Ph2P

CH2Cl2

Me

H

O

1. 2 x KOH H2O

O

CO2H

2. DMSO 50 ˚C

Ph2P H Me 136; 89% yield

135

Z-137; 91% yield, Z only

The Peterson reaction This silicon analogue of the Wittig reaction has the advantage of two alternative stereospecific elimination mechanisms.27 Providing you can make the silicon-containing alcohols (and there’s the problem!) as single diastereoisomers, you can get either E- or Z-alkene from either one. The best example is probably Barrett’s synthesis of trisubstituted alkenes28 The silyl-lithium derived from 138 is converted into the ketone 141 by straightforward steps. Reaction with MeLi is stereoselective (see Felkin-Anh explanation) to give the alcohol 142 with the silyl and OH groups syn. OH Cl

Si

Ph

Me Me 138

1. Li, THF 2. R 1CHO

PhSiMe2 R O 141

Si

Ph

CCl4

Me Me 139

R1

Si

Ph

Me Me 140; 71-75% yield

MeLi

R1

1. Mg, CuBr.SMe 2 2. R 2COCl

R2 H

PhSiMe2 2

R1

R1

Cl Ph3P

R2 Me OH 142

MeLi

PhMe2Si R1 O Felkin-Anh selectivity

Elimination in base goes through the Wittig-style intermediates 143 and 144 and is syn-specific to give E-145 while elimination in acid follows an E2 mechanism with an anti arrangement of silyl and OH groups to give E-145.

15 Stereoselective Methods for E-Alkenes: The Julia Reaction

Me PhSiMe2

PhSi

R1

O

R1

18-crown-6

R2

Me OH 142

Me

PhSi

KH

R2

Me

Me

239

O

R2

R1

R1 R2 Me 144

Me

143

E-alkene 145; 95:5 E:Z 50-55% yield from 138 R1

PhSiMe2

Nu TsOH

R2

R1

OH

Me

142

R2

E2

R2

R1

THF

Me

PhSiMe2

OH2

Z-alkene 145; 92:8 Z:E 50-55% yield from 138

146

Stereoselective Methods for E-Alkenes: The Julia Reaction The Julia olefin synthesis is rather like the Wittig reaction with a sulfone instead of a phosphonium salt but with one other important difference: the elimination step is stereoselective and both diastereoisomers of the intermediate can give the same isomer of the alkene. Treatment of the sulfone 147 with a strong base gives the anion 148 (or a metal derivative) that combines with an aldehyde to give a diastereomeric mixture of adducts 149. Elimination by various methods gives, in open chain compounds, mostly E-150 but, in cyclic compounds, mostly the Z-alkene.29 O O O Ar

O O R1

S

base

S

Ar

147

R1

R2CHO

O S

Ar

R1 elimination

HO 149

148

R1 R2

R2 E-150

Originally the phenylsulfone 147; Ar⫽Ph was used and the adduct was acylated then treated with dissolving metal (e.g. sodium amalgam) to bring about a reductive elimination by electron transfer. Addition of two electrons to 151 gives the dianion 152 that breaks down to the carbanion 153 or something like it - at any rate without stereochemistry - and so to the alkene. O O 1. base

O O Ph

S

R1

Ph

2. R 2CHO

Na/Hg

O

Ph

R2

S

R1

O

R2

R1

O

R2

E-150

3. R 3COCl

R3 147; Ar=Ph

S

O O R1

O 151

R3

O 152

R3

O 153

An advanced example of this simple approach was used by Danishefsky in the synthesis of the antibiotic indolizomycin.30 This example illustrates the compatibility of the Julia olefination with many sensitive functional groups. The enal 154 was combined with the lithium derivative of an allylic sulfone and the adduct acylated to give a mixture of isomers of the adduct 155. Elimination with sodium amalgam gave a good yield of the E,E,E-triene 156. [The ‘R’s are protecting groups.]

240

15 Synthesis of Double Bonds of Defined Stereochemistry

O

OR

OR

O 1. Li

N R

OR

O

SO2Ph

N H

N

R

Na/Hg

H

R

H

THF MeOH

2. Ac 2O

CHO

SO2Ph

AcO

154 156; 89% yield E,E,E-isomer only

155; 86% yield mixture of diastereoisomers

Modified Julia reactions Recent developments emphasise the convenience of replacing the three step (Marc) Julia reaction by one-step procedures using reagents 147 in which Ar is a heterocycle. Thus (Sylvestre) Julia used a benzothiazole 157. The adduct 158 with an aldehyde decomposes by fragmentation of the spiro compound 159 without the need for any further reagents to give SO2, the alkene 150 and the benzothiazole 169. The benzene rings are omitted from 158–160 for clarity.31 O O

O O R1

S

S

S

S

R1

O

N

2. R 2CHO

157

O

O O R1

O

R2

R2 O

R2

N

158

S

S

S

1. base

N

R1

O

SO2

159 R1

R2

E-150

S

S N

O N 161

160

This method was greatly improved by Charette’s observation that the hexamethyldisilazide bases (Li, Na or K) and careful choice of solvent made either E- or Z-selective alkene synthesis possible without the side reactions found in the original method.32 Starting from the benzothiazole 164, either E- or Z-162 could be made in high yield and good selectivity by choice of solvent alone. Both partners in this modified Julia reaction are enantiomerically pure cyclopropanes and this method led to no destruction of either stereochemistry or the three-membered rings. This method was used to make a single enantiomer of a natural product containing six cyclopropane rings: details are given in the workbook.

O O 1. (Me 3Si)2NNa DMF

OR 162; 100% yield, 3.4:1 E:Z, R = Sii-Pr3

2. OHC

S

S N

163

1. (Me 3Si)2NNa toluene or CH2Cl2

2. OHC

OR

164

OR 163

162; 100% yield, OR 10:1 Z:E, R = Sii-Pr3

15 Stereoselective Methods for E-Alkenes: The Julia Reaction

241

The Kocienski modification of the Julia reaction The best version of the Julia olefin synthesis (so far) is probably that introduced by Kocienski.33 It uses N-phenyl tetrazolyl sulfones 167 easily prepared from the available thiol 165 by a Mitsunobu reaction with a simple alcohol followed by oxidation.

SH

N N N

HO

R1

S

N N

DIAD, Ph 3P THF

N

N

165

O O

Ph

Ph

Ph

R1

N

NaHCO3 CH2Cl2

N

N

R1

S

N

mCPBA

N

166

167

The reaction uses hexamethyldisilazide bases and is again sensitive to solvent. Polar solvents (dimethoxyethane, DME, is the best) and (Me3Si)2NK give high E-selectivity (usually ⬎95:5 E:Z) while non-polar solvents (toluene is the best) with (Me3Si)2NLi give moderate Z-selectivity. Ph R2

1. (Me 3Si)2NLi toluene

R1

2. R 2CHO

N N

Z-150

O O R1

S

1. (Me 3Si)2NK DME

R1

R2

2. R 2CHO

N N 167

E-150

A recent application is the synthesis of the anti-parasitic agent nafuredin 168 by Omura’s group.34 Nafuredin has four alkenes in the side chain but disconnection near the middle of the molecule is strategically best. As with the Wittig, there is a choice where to put the sulfone and the aldehyde. The sulfone 169 can be easily made from an allylic alcohol. No doubt some of the functional groups in 169 will have to be protected. N N H

O

HO O

O

Julia

HO

Ph O

N O

N

H

S

O

O O 169

168; nafuredin

OHC 170

The asymmetric aspects of the syntheses of 169 and 170 appear in the workbook. The other three alkenes were made by E-selective Wittigs with stabilised ylids or HWE reactions. The sulfone 172 for the Julia reaction was prepared from 171, in turn prepared from glucose and gave the corresponding E-alkene (a modified version of 168) in excellent yield and perfect E-selectivity.

242

15 Synthesis of Double Bonds of Defined Stereochemistry

N N

MsO

OH

H

HO

O N

1. 165 DEAD Bu3P

MsO

2. H 2O2 Mo(VI)

O

BnO

Ph

N

S

H

HO

O 1. (Me 3Si)2NK THF 2. 170

O

BnO

OTIPS 171

E-alkene 79% yield E only

OTIPS 172

Direct Coupling of Carbonyl Compounds and Alkenes Carbonyl compounds: the McMurry reaction The disconnections for the Wittig, HWE, and Julia reactions are at the double bond with the starting materials being a carbonyl compound and a P or S functionalised alkyl group. There is a choice about which component has which functional group.

R2

R1

Wittig, HWE or Julia

R1

R2

CHO +

Z or

+ R1

Z

OHC

R2

Z = PPh3 or P(O)(OR)2 or SO2Ar

We now turn to two reactions, the McMurry and metathesis, that have the same disconnection but both reagents have identical functional groups: carbonyls for the McMurry and alkenes for metathesis. These methods have the advantage of simplicity but there are obvious problems of selectivity.

R2

+

R1

metathesis

R1

R2

McMurry

R1

CHO +

OHC

R2

The McMurry35 uses low-valent titanium to couple two carbonyl groups together and works very well when the two compounds are the same: the symmetrical stilbene E-174 can be made in 95% yield and in almost perfect (99.92:0.04 E:Z) E-selectivity from 3-fluorobenzaldehyde 173. The low-valent titanium is produced by reduction of Ti(III) with lithium metal.36

F

CHO

TiCl3, Li

F

F

DME, reflux 18 hours

173

E-174; 95% yield

If the reaction is intramolecular two different carbonyl groups can be coupled unambiguously and these two examples show that esters37 175 and amides38 177 can be used to make benzofurans 176 and indoles 178.

15 Direct Coupling of Carbonyl Compounds and Alkenes

243

Ph Ph

O

MeO

O

MeO

TiCl3, Zn

NH

Ph

naphthalene THF, 25 ˚C

O Ph

TiCl3, Li

DME

O

O

O

N

O

N H O

N

F3C

CF3 175

176

177

178

Alkenes: olefin metathesis Metathesis is now one of the most important ways to make alkenes. And yet practical catalysts for metathesis were reported only in the 1990s and the reaction39 was largely limited to cyclic alkenes such as 180. Though there are many metathesis catalysts, by far the most important are the two Grubbs ruthenium carbene complexes 181 and 182. They are not the most active catalysts but they are stable and easy to use and compatible with a wide range of functional groups. The simpler 181 is obviously a carbene complex and is usually called ‘Grubbs’ catalyst’. The more active complex 182 (‘second generation Grubbs’) replaces one of the tricyclohexyl phosphines with a heterocycle. In these diagrams the dotted line shows that there may or may not be a double bond on the other side of the ring. Otherwise the diagrams 182a-c show three different ways of drawing the same compound. The fi rst 182a has the advantage of simplicity but wrongly suggests that there is a hydrogen atom on the ring where the Ru atom bonds. The other two 182b,c show more clearly that the heterocyclic ring is also a carbene-like ligand.

PCy3 N R N R N R R N R N R N Cl Ru Cl Ru Cl Ru Cl Ru Cl Ph PCy3 Cl Ph Cl Ph Cl Ph PCy3 PCy3 PCy3 181 182b 182c 182a Grubbs' catalyst

metathesis catalyst

+ CH2=CH2

179

180

The mechanism of the reaction involves a series of [2 ⫹ 2] cycloadditions 183 and 185 and cycloreversions 184 and 186. The first cycle (183 to 185) is different from the main reaction in that the other product is styrene and the catalyst then changes to the methylene complex 187. The reaction goes from left to right as the other product is gaseous ethylene.

Ph

Ph Ru 147

183

Ph

Ru Ru Ru

184

185

186

PCy3 Cl Ru CH 2 180 + Cl PCy3 187; catalyst for rest of reaction

244

15 Synthesis of Double Bonds of Defined Stereochemistry

A simple example is the synthesis of analogues of the antibiotic and antifungal streptazolin by Cossy.40 Enantiomerically pure diene carbamate 188, prepared from the chiral pool (see chapter 23 and the workbook for this chapter) was treated with the Grubbs’ catalyst to form the six-membered ring 189 required for (⫺)-4,5-dihydrostreptazolin 190. This metathesis product inevitably contains a Z-alkene. H

OHC O

O

3 mol% 181

N

protected glyceraldehyde

O

OH

H

N

60 ˚C

O O

188

H

OH

N

O O

189; 90% yield

H

O

190

Metathesis is excellent for the formation of medium- and large rings. In the synthesis of epothilones K. C. Nicolaou41 used the Grubbs catalyst to close the 16-membered ring lactone 192 from the open chain ester 191 in good yield under mild conditions. The most obvious way to prepare 192 is to make an open chain hydroxy-acid and cyclise by lactone formation. The metathesis cyclisation generally works better.

HO 0.1 equivs 181

HO

OTIPS

O

OTIPS

25 ˚C, CH 2Cl2

O O

O

O

O 192; 80% yield

191

The very complex natural product ircinal A Z-196 has difficult Z-alkenes in the eight-membered ring and in the ‘frill’ round the molecule’s waist. Both have been inserted by metathesis with good Z-selectivity.42 The starting material 193 has three alkenes but only two metathesise to give the 13-membered cyclic frill 194. The final metathesis to close the eight-membered ring is obviously difficult and the yield is poor. H

CH(OMe)2

H

N

181

CH(OMe)2

N

1. KOH

O N

O

N O

2. COCl

O 194; 67% yield 8:1 Z:E

193 H

CH(OMe)2 OH

N N

O

195; 75% yield

H

1. 181 2. HCl water

CHO

OH

N H

N

O

Z-196; 26% yield

15 Stereoselective Methods for E-Alkenes

245

A much simpler heterocyclic eight-membered ring43 198 was successfully closed in 95% yield with the second generation Grubbs catalyst 182; R ⫽ mesityl [this is the catalyst with the double bond]. This example shows that trisubstituted alkenes can be made by metathesis too. H

H

N Me

182; R = mesityl

O

O

N

O

Me

the N-mesityl group

N

O 198; 95% yield

197

Me

Stereoselective Methods for E-Alkenes [3,3]-Sigmatropic rearrangements We said a few pages back that there was a strong reason for making a trisubstituted E-γ,δ-unsaturated carbonyl compound 111 by a [3,3]-sigmatropic rearrangement and we have come now to the section of the book where we look at this reaction more closely. The disconnection is of the α,β bond in a γ,δ-unsaturated carbonyl compound 199. O

O

enolate allylation

X

R

+

X

Br

R

199

We call this disconnection “enolate allylation” because we want to alkylate an enolate ion with an allyl halide. We have already discussed (chapter 19) the regioselectivity problems inherent in reactions of allyl derivatives and you should recall that we want to react at the correct (less substituted) end of the allyl system and we want to make an E-alkene. The solution is to use a [3,3]sigmatropic rearrangement. This is the reaction sequence: X

X R

MeO

OMe

+

O

O

[3,3]

X

200

X

R

R

201

OMe

Hg(II)

[3,3]

202

OH 200

RCO2H

202

203

199

Instead of an allylic halide, which can equilibrate all by itself, we use a regiochemically stable allylic alcohol 200. This is combined with an acetal 201 under proton catalysis or with a vinyl (enol) ether 203 under Hg(II) catalysis to give the key vinyl allyl ether 202. These compounds rearrange on heating as the mechanism shows 202 to give the E-γ,δ-unsaturated carbonyl compound 199. A C–C single bond replaces a C–O single bond (a bad bargain this) and the reaction is driven by the replacement of a C⫽C double bond by the more stable carbonyl group. These reactions can be carried out at the aldehyde (X⫽H), ketone (X⫽R), ester (X⫽OR) or amide (X⫽NR2) oxidation level just by choosing the correct starting material. X

X

X

R

O

O

X

OH

R

R 200

O

O R

202

R 204

E-199

E-alkene in 204

246

15 Synthesis of Double Bonds of Defined Stereochemistry

Because the [3,3] sigmatropic rearrangement turns the allylic system inside out we must use the ‘wrong’ allyl alcohol 200 to make 199. The E-selectivity is a result of a chair-like transition state 204 in which the substituent R prefers an equatorial position. The second diagram of the transition state marks the trans arrangement of the three bonds involved. These rearrangements are among the most E-selective reactions known.44 The selectivity is still good even when a tri-substituted alkene 208 is being made, providing that a large substituent is in the equatorial position 207. X

X R

X O

O

R

O

OH R 205

R 207; R > Me

206

E-208

An excellent illustration is Johnson’s synthesis of squalene45 by a series of double [3,3] rearrangements. Reaction of a dialdehyde 209 with propenyl lithium 210 gives a double allylic alcohol 211 that undergoes a double rearrangement at the ester oxidation level with an orthoacetate [MeC(OEt)3]. OH Li

OHC

CHO

MeC(OEt)3 210

EtO2C

CO2Et

EtCO2H

209

211

OH

E,E-212

Reduction of the esters in the product 212 to aldehydes allows the whole process to be repeated twice more. The complete skeleton 214 has four new alkenes and yet contains over 90% of the E,E,E,E isomer. OH 1. LiAlH 4, 2. CrO 3

E,E-212

3. 210

OH

E,E-213 MeC(OEt)3

EtO2C

CO2Et

EtCO2H

E,E,E,E-214

This method too gives Z-alkenes if they are in a ring. The β-lactam 215 can be converted into the E-alkene 216 by a Wittig reaction and hence by a [3,3] sigmatropic rearrangement into the eightmembered heterocycle 217 containing two new Z-alkenes.46 The discerning reader will notice that this is quite a different reaction from the last. That was an oxy-Cope (or Claisen-Cope) driven by the formation of a carbonyl group. This is the Cope itself with all carbon atoms in the (necessarily boat shaped) cyclic transition state. The driving force is the loss of the strained four-membered ring. CO2Me

CO2Me CHO N O

CO2Me toluene

N R

215

Ph3P

O R E-216; 82% yield

reflux

N O R Z,Z-217; 86% yield

15 Stereoselective Methods for E-Alkenes

247

[2,3]-Sigmatropic rearrangements: The Wittig rearrangement Derivatives of allylic alcohols such as 219 may rearrange in base to the homoallylic alcohols 222 with the creation of a new alkene. The reaction is a [2,3]-sigmatropic rearrangement of the anion 220 and works best when the group Z is anion-stabilising otherwise there may be competition with the original Wittig rearrangement to give the alcohol derived from 218. Both these products are more stable than the carbanion 220 because they are oxyanions. The reaction of 220 to give 221 is best called the [2,3]-Wittig rearrangement and is E-selective as the group R prefers a (pseudo-) equatorial position in the half-chair transition state.47 R

[1,2]

O

R

R

base

O

Z

O

Z 219

218

R

[2,3]

R

O

Z

HO

Z

Z 220

E-222

221

The problem of what to do with the usually unwanted anion-stabilising group Z is solved in the Wittig-Still rearrangement by using the R3Sn group that sacrifices itself to form the anion. Alkylation of the allylic alcohol 223 and treatment with BuLi initiate the rearrangement 225 to give the unsubstituted E-allylic alcohol 226. There are many examples.48 R

R

1. KH 2. I

OH

SnBu3

BuLi

O

R

SnBu3

O

224

223

[2,3]

R

R

O 226

225

HO E-227

Most of the examples in this chapter have been disubstituted alkenes, a few have been trisubstituted, but none so far has been tetrasubstituted (except cyclic alkenes) as this is the most difficult case of all. One solution49 starts with ethyl lactate 228 and uses a HWE reaction on the enantiomerically pure phosphonate 229 to make the E-enone 230 with very high selectivity. Chelation-controlled addition of a Grignard reagent gives the stereochemically pure allylic alcohol 231. Chelation control is explained in chapter 21. O

EtO2C

O

O R1CHO

(EtO)2P OH

LiOH,H2O

OR 229; R = BnOCH 2

228; ethyl lactate

HO

R2

R2MgBr

R1

R1 OR

OR E-230; >95:5 E:Z

E-231

Alkylation of the free OH provides the substrate 232 for the Wittig-Still rearrangement 233. The new tetrasubstituted alkene 234 is almost entirely E (⬎95:5 in most cases) and the migrating CH2O group is transferred across the top face of the allylic system as drawn. The [2,3]-sigmatropic rearrangement is suprafacial.

Bu3Sn

O

1. KH

E-231

2. I

SnBu3

R2

1

O BuLi

HO R2

1

R

R1

R OR

E-232

[2,3]

OR

OR

E-233

E-anti-234

248

15 Synthesis of Double Bonds of Defined Stereochemistry

There are many sigmatropic rearrangements and all are E-selective in open chain compounds but can give Z-alkenes if the structure of the compound demands it. In this way they resemble the McMurry reaction and olefin metathesis. Many such reactions are used to transmit threedimensional stereochemistry rather than for E/Z control in alkenes.

Reduction of Alkynes You will already be aware that cis-alkenes can be formed by catalytic hydrogenation of alkynes. This example illustrates how acetylene, a double-ended nucleophile, can be used to build the skeleton of a Z-allyl silane 235 of the type used in chapter 12. The disconnection 236 is next to the ‘alkene’ (now an alkyne) rather than across it. SiMe3

OEt

SiMe3

FGI

EtO reduction

EtO

OEt

235

Cl

EtO OEt 237

236

SiMe3

Cl

238

The synthesis is straightforward except that the normal Lindlar catalyst is not used, Raney nickel being preferred in this case.50 H 1. BuLi

EtO

1. NaNH 2

2. 237

2. 238

H2, Raney Ni

236 H2N

OEt

235

NH2

Reduction of alkynes with sodium in liquid ammonia produces E-alkenes as you will be aware though the reduction of functionalised alkynes with LiAlH4 is perhaps more common nowadays. Here are simple examples of both methods: the LiAlH4 reduction will be discussed in the next chapter. 1. Na, NH 3(l) 2. NH 3, H 2O

240; 80-90% yield, all E-dec-5-ene

239 OH

LiAlH4

OH 241

E-242

In an important paper making enantiomerically enriched amines by 1,3-dipolar cycloadditions, Denmark51 made both isomers of the vinyl ether 243 from the alkyne 244. Lindlar reduction gave pure Z-243 while LiAlH4 reduction gave pure E-243. Ph Ph

O

Ph

THF

E-243

O

LiAlH4

Ph H2, Pd/BaSO 4

Ph O

quinoline

Ph 244

Z-243

15 Stereospecific Methods for Z-Alkenes

249

Stereospecific Methods for Z-Alkenes Using cyclic compounds In a smaller ring (⬍8 membered) an alkene must have the Z-configuration. Cleavage of another bond in such a ring must leave the Z-alkene intact. We shall look at oxidative cleavage of Birch reduction products, elimination in a nitrogen heterocycle and hydrolysis of lactones. Birch reduction of aryl ethers 245 leaves two non-conjugated alkenes 246: the enol ether is much more nucleophilic and so is more easily cleaved by ozone. Corey52 used this strategy to make trisubstituted alkene 247. OMe

1. ozone MeOH, Me 2S

OMe

Li, THF

t-BuOH, NH 3(l)

CO2Me

HO

2. NaBH 4

245

Z-247

246

Simple pyridines, such as 2-methylpyridine 248, can be alkylated and reduced to leave just one necessarily Z-alkene 251 not conjugated with nitrogen. 1. BuLi 2. RCl

N

NaBH4

MeI

R

N

248 2-methylpyridine

MeOH

R

N Me

249

R

N Me Z-251

250

Further alkylation and elimination 253 gives an open chain diene 254. The second alkene is formed stereoselectively with E geometry.53 H MeI

Z-251

OH

R

KOH

MeOH

N Me

R

Me Z-252

R

N Me

Me2N

Me Z-253

Z,E-254; 57% yield

White needed the diacid ester Z,E,Z-255 for his synthesis of the anti-leukaemia compound verrucarin.54 It must clearly be made from the components Z-256 and E,Z-257 though there are obvious selectivity problems in getting the right ester from compounds with three different CO2H groups. O

CO2H

CO2H

O CO2H

OH

+

HO2C

CO2H

Z,E,Z-255

Z-256

E,Z-257

The Z-acid 256 was made from a lactone. Dehydration of the tertiary alcohol in the open chain compound 258 would give the E-alkene but after cyclisation to the lactone 259 dehydration with polyphosphoric acid (PPA) must give Z-260. Careful hydrolysis and methylation gives Z-261, the methyl ester of Z-261.

250

15 Synthesis of Double Bonds of Defined Stereochemistry

HO HO HCl

EtO2C

OTHP

1. NaOH H2O

PPA

OH

EtOH

O

O 259

258

O

O 260

2. CH 2N2

CO2Me Z-261

The Z-alkene in the other half 257 was made from a furan 262. Cycloaddition with singlet oxygen (rose Bengal is a dye that sensitises the conversion of normal triplet oxygen to the singlet) leaves one necessarily Z-alkene in the five-membered ring 264. One possible mechanism for the decarboxylation of the cycloadduct 263 is given 265. O

O2, hν

CO2H

O

O O O 263

rose Bengal

262

CO2H

HO

O 264

O O O

O

O

H

265

The two halves are coupled by an HWE reaction to put in the easiest double bond - the conjugated E-alkene. The alcohol Z-261 is converted into an ester Z-266 that reacts directly with the hemiacetal 264 using two equivalents of base to give the ester of Z,E,Z-255 in 84% yield. Note that this method also keeps the acids distinct. O

CO2Me Z-261

1. NaOMe 2. 264

CO2Me

CO2H

O

P(OEt)2

O

OH

O

Z-266

O CO2Me Z,E,Z-255 (mono methyl ester)

Interconversion of E and Z Alkenes Photochemical isomerisation to the Z-isomer A simple approach to a difficult alkene might be to make whatever mixture results from the easiest synthesis and convert it stereoselectively into one isomer or, with more precise control, to make the wrong isomer, if that is easier to make, and convert that stereospecifically into the other. A simple example is the enone E-5 made by an aldol condensation. It is easy to see why E-5 is the thermodynamic product: there is a steric clash in the Z-isomer between the carbonyl oxygen and one of the ortho-Hs on the benzene ring. O

O

O

Ph O

hν

PhCHO

H

Ph

NaOH H2O

E-5 only

Z-5 85% yield

Z-5; steric clash

Shining light on the pure E-5 isomerises it completely55 into Z-5. The excited π,π* state of the enone 5 has one electron in the π* orbital allowing free rotation about the alkene. But why should it prefer the less stable Z-isomer? Photochemical reactions are determined not by thermodynamic stability but by light absorption. More stable E-5 is planar and conjugated and absorbs light more efficiently and at higher wavelength but Z-5 cannot be planar and absorbs light less efficiently and at a shorter wavelength than does E-5. Irradiation at a higher wavelength converts all E-5 into Z-5.

15 Interconversion of E and Z Alkenes

251

A synthesis of methoxatin Methoxatin 267 is a bacterial co-enzyme that allows the oxidation of methane to more useful one-carbon compounds. The 1,2-dione in the middle ring is the powerhouse of the molecule and can be made from the simpler aromatic compound 268 by oxidation. Oxidative cyclisation of the Z-alkene 269 might give 268. HO2C

HO2C

HO2C

CO2H

NH

O

N

CO2H

NH

C–C

oxidation

oxidation

CO2H

CO2H

N

O 267; methoxatin

CO2H

NH

FGI

CO2H

N

Z-269

268

Making 272 from the aldehyde 270 and the Wittig reagent 271 proved easy once all the carboxylic acids were protected as methyl esters but gave mainly E-272 as either of the possible ylids is stabilised. Clearly this isomer cannot cyclise so isomerisation to Z-272 was necessary.56 CO2Me

CO2Me

MeO2C NH

NaH

+

MeO2C

DMF

N

CHO

N

CO2Me

N H E-272; 84% yield, >95% E

PPh3 270

CO2Me

271

Brief irradiation with UV light through a pyrex filter (to remove short wavelength light) in the presence of diphenyldiselenide (PhSe – SePh) gave the trimethyl ester of 268 by isomerisation to Z-272 and oxidative cyclisation. MeO2C

MeO2C NH

E-272

CO2Me

CO2Me

NH

hν

hν

pyrex filter

PhSe

SePh

N CO2Me Z-272; 100% yield, 100% Z

N

CO2Me

273

Radical isomerisation to the E-isomer Relatively stable radicals such as I •, PhS• or Bu3Sn• add to alkenes 274 and then drop out again. PhS• radicals are easily made by AIBN treatment of thiophenol (PhSH) or simply by refluxing the alkene with PhSH without any precautions to remove oxygen. If the intermediate 275 has a long enough life it can rotate and equilibrate thermally to the E-alkene 277.

R1

SPh R2 274

SPh

R1 R2 275

R1

R2

R2

R1 SPh 276

E-277

252

15 Synthesis of Double Bonds of Defined Stereochemistry

Stilbene formation by the Wittig reaction usually gives a mixture (here 7:3 Z:E) of E- and Z-isomers 280. Irradiation of a heptane solution containing a small amount of iodine in sunlight isomerises the mixture to the E-stilbene 280 which crystallises from solution in 68% yield.57

Br

PPh3

Br

I2, hν

1. base

E- and Z-280

2. Br

heptane

Br

OHC 279

278

E-280

The mechanism is the same: iodine absorbs the sunlight (the solution is purple) to give I• radicals that add to the alkene 281 to give a long lived radical 282 that rotates before ejecting the I• radical. There is an example of the use of the Bu3Sn• radical in the next chapter.

Ar1

I

I

Ar1

Ar2 Z-281

Ar2

Ar1

Ar2

I 283

282

Ar2

Ar1

E-281

Stereospecific Interconversion of E and Z-isomers Some isomers of alkenes are particularly difficult to make and a stereospecific method is required. Fortunately there is a good one. Cis-cyclo-octene Z-284 is easy to make by elimination and is available commercially. The epoxide 285 is made without change of stereochemistry and opening this epoxide with the very nucleophilic Ph2PLi ensures one inversion in the formation of the phosphine 286. H

H Ph2PLi

mCPBA

OH

O H

Z-284

H 286

285

PPh2

The phosphine 286 is easily alkylated to provide the Wittig intermediate 287 that eliminates in the usual stereospecific manner via the oxaphosphetane 288 and must therefore give the strained trans-cyclo-octene E-284, the smallest trans cycloalkene that is stable. It is also chiral.58 H

H

OH

H 286

O

H

H 287

PPh2 Me

H 288

PPh2 Me

H O

base

MeI

PPh2

H

OH

PPh2 H Me 289

H

E-284

15 References

253

References General references are given on page 893 1. Vogel, page 1033. 2. E. J. Corey, N. H. Andersen, R. M. Carlson, J. Paust, E. Vedejs, I. Vlattas and R. E. K. Winter, J. Am. Chem. Soc., 1968, 90, 3245. 3. E. J. Corey, P. B. Hopkins, S. Kim, S. Yoo, K. P. Nambiar and J. R. Falck, J. Am. Chem. Soc., 1979, 101, 7131. 4. D. E. Vogel and G. H. Büchi, Org. Synth., 1988, 66, 29. 5. J. F. Lane, J. D. Roberts and W. G. Young, J. Am. Chem. Soc., 1944, 66, 543, J. F. Lane, J. Fentress and L. T. Sherwood, J. Am. Chem. Soc., 1944, 66, 545. 6. F. A. Abd Elhafez and D. J. Cram, J. Am. Chem. Soc., 1952, 74, 5846; D. J. Cram, F. D. Greene and C. H. Depuy, J. Am. Chem. Soc., 1956, 78, 790. 7. A. Ault and R. Kopet, J. Chem. Ed., 1969, 46, 612; D. B. Ledlie, T. J. Wenzel and S. M. Hendrickson, J. Chem. Ed., 1989, 66, 781; C. B. Fryhle, C. M. Ryback and K. Pulley, J. Chem. Ed., 1991, 68, 1050. 8. S. Ma and X. Lu, Org. Synth., 1995, 72, 112. 9. R. Huber and G. B. Jones, Tetrahedron Lett., 1994, 35, 2655; K. Bowden and M. J. Price, J. Chem. Soc. (B), 1970, 1466, 1472. 10. D. E. Vogel and G. H. Büchi, Org. Synth., 1988, 66, 29. 11. C. A. Henrick, Tetrahedron, 1977, 33, 1845; B. A. Bierl, M. Beroza and C. W. Collier, Science, 1970, 170, 87; for a leading reference to possible detailed mechanisms, see R. Robiette, J. Richardson, V. K. Aggarwal and J. N. Harvey, J. Am. Chem. Soc., 2006, 128, 2394. 12. M. Schlosser, Topics in Stereochem., 1970, 5, 1. 13. E. J. Corey, N. M. Weinshenker, T. K. Schaaf and W. Huber J. Am. Chem. Soc., 1969, 91, 5675. 14. S. Trippett and M. Walker, J. Chem. Soc., 1961, 1266. 15. H.-J. Bestmann, O. Vostrowsky, H. Paulus, W. Billmann and W. Stransky, Tetrahedron Lett., 1977, 121. 16. W. S. Wadsworth and W. D. Emmons, J. Am. Chem. Soc., 1961, 83, 1733; W. S. Wadsworth, Org. React., 1977, 25, 73. 17. F. E. Ziegler, G. R. Reid, W. L. Studt and P. A. Wender, J. Org. Chem. 1977, 42, 1991. 18. T. Sugahara, Y. Komatsu, and S. Takano, J. Chem. Soc., Chem. Commun., 1984, 214. 19. M. Schlosser and K. F. Christmann, Angew. Chem., Int. Ed., 1966, 5, 126. 20. M. B. Gravestock, W. S. Johnson, B. E. McCarry, R. J. Parry and B. E. Ratcliffe, J. Am. Chem. Soc., 1987, 100, 4274. 21. W. C. Still and C. Gennari, Tetrahedron Lett., 1983, 24, 4405. 22. J. Clayden and S. Warren, Angew. Chem., Int. Ed., 1996, 35, 241. 23. A. D. Buss and S. Warren, J. Chem. Soc., Perkin Trans 1, 1985, 2307. 24. K. C. Nicolaou, J. L. Gunzer, G. Shi, K. A. Agrios, P. Gärtner and Z. Yang, Chem Eur. J., 1999, 5, 646. 25. T. K. Jones, S. G. Mills, R. A. Reamer, D. Askin, R. Desmond, R. P. Volante and I. Shinkai, J. Amer. Chem. Soc., 1989, 111, 1157. 26. D. Levin and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1988, 1799. 27. L. F. van Standen, D. Gravestock and D. J. Ager, Chem. Soc. Rev., 2002, 31, 195. 28. A. G. M. Barrett, J. A. Flygare, J. M. Hill and E. M. Wallace, Org. Synth., 1996, 73, 50. 29. M. Julia and J.-M. Paris, Tetrahedron Lett., 1973, 4833. 30. G. Kim, M. Y. Chu-Moyer, S. J. Danishefsky and G. K. Schulte, J. Am. Chem. Soc., 1993, 115, 30. 31. J. B. Baudin, G. Hareau, S. A. Julia and O. Ruel, Tetrahedron Lett., 1991, 32, 1175. 32. A. B. Charette and H. Lebel, J. Am. Chem. Soc., 1996, 118, 10327. 33. P. R. Blakemore, W. J. Cole, P. J. Kocienski and A. Morley, Synlett, 1998, 26. 34. D. Takano, T. Nagamitsu, H. Ui, K. Shiomi, Y. Yamaguchi, R. Masuma, I. Kuwajima and S. Omura, Org. Lett., 2001, 3, 2289. 35. J. E. McMurry, Chem. Rev., 1989, 89, 1513. 36. P. Wyatt, S. Warren, M. McPartlin and T. Woodroffe, J. Chem. Soc., Perkin Trans 1, 2001, 279. 37. S. Talukdar, S. K. Nayak and A. Banerji, J. Org. Chem., 1998, 63, 4925; S. Rele, S. Talukdar, A. Banerji and S. Chattopadhyay, J. Org. Chem., 2001, 66, 2990.

254

15 Synthesis of Double Bonds of Defined Stereochemistry

38. A. Fürstner, A. Hupperts and G. Seidel, Org. Synth., 1999, 76, 142. 39. A Fürstner, Angew. Chem., Int. Ed., 2000, 39, 3012. 40. J. Cossy, I. Pévet and C. Meyer, Synlett., 2000, 122. 41. K. C. Nicolaou, Y. He, D. Vorloumis, H. Vallberg, F. Roschangar, F. Sarabia, S. Ninkovic, Z. Yang and J. I. Trujillo, J. Am. Chem. Soc., 1997, 119, 7960. 42. S. F. Martin, J. M. Humphrey, A. Ali and M. C. Hillier, J. Am. Chem. Soc., 1999, 121, 866. 43. A. Fürstner, O. R. Thiel, L. Ackermann, H.-J. Schanz and S. P. Nolan, J. Org. Chem., 2000, 65, 2204. 44. S. J. Rhoads and N. R. Raulins, Org. React., 1975, 22, 1. 45. W. S. Johnson, L. Werthemann, W. R. Bartlett, T. J. Brocksom, T. Li, D. J. Faulkner and M. R. Petersen, J. Am. Chem. Soc., 1970, 92, 741. 46. B. Alcaide, C. Rodríguez-Ranera and A. Rodríguez-Vicente, Tetrahedron Lett., 2001, 42, 3081. 47. J. A. Marshall, Comp. Org. Synth., 3, chapter 3.11, page 975; T. Nakai and K. Mikami, Chem. Rev., 1986, 86, 885. 48. W. C. Still, J. Am. Chem. Soc., 1978, 100, 1481; W. C. Still and A. Mitra, J. Am. Chem. Soc., 1978, 100, 1927; M. Mikami and T. Nakai, Synthesis, 1991, 594. 49. J. Mulzer and B. List, Tetrahedron Lett., 1994, 35, 9021. 50. W. S. Johnson, Y-Q. Chen and M. S. Kellogg, J. Am. Chem. Soc., 1983, 105, 6653. 51. S. E. Denmark and L. R. Marcin, J. Org. Chem., 1995, 60 3221. 52. E. J. Corey and H. Yamamoto, J. Am. Chem. Soc., 1970, 92, 226, 3523, 6636 and 6637; E. J. Corey, H. Yamamoto, D. K. Herron and K. Achiwa J. Am. Chem. Soc., 1970, 92, 6635. 53. G. Dressaire and Y. Langlois, Tetrahedron Lett., 1980, 21, 67. 54. J. D. White, J. P. Carter and H. S. Kezar, J. Org. Chem., 1982, 47, 929. 55. F. Bourielle-Wargnier, A. Feigenbaum and J. Mozart, J. Chem. Ed., 1978, 55, 339. 56. J. B. Hendrickson and J. G. de Vries, J. Org. Chem., 1982, 47, 1148. 57. A. Terfort and H. Brunner, J. Chem. Soc., Perkin Trans. 1, 1996, 1467. 58. A. J. Bridges and G. H. Whitham, J. Chem. Soc., Chem. Commun., 1974, 142; K. T. Burgoyne, S. G. Davies, M. J. Peagram and G. H. Whitham, J. Chem. Soc., Perkin Trans. 1, 1974, 2629.

16 Stereo-Controlled Vinyl Anion Equivalents Introduction: Reagents for the Vinyl Anion Synthon Vinyl-Lithiums Vinyl-Lithiums from Ketones: The Shapiro Reaction The Aliphatic Friedel-Crafts Reaction Hydrometallation of Alkynes Hydrostannylation of terminal alkynes Hydroboration and Hydroalumination of Alkynes Hydroboration of alkynes Hydroalumination of alkynes Hydroalumination of propargyl alcohols Hydrosilylation Hydrosilylation of alkynes catalysed by Lewis acids Hydrosilylation of alkynes catalysed by transition metals Hydrozirconation Carbo-Metallation Carboalumination Carbocupration Reactions of Vinyl Sn, B, Al, Si, and Zr Reagents with Electrophiles Reactivity of vinyl alanes Preparation of vinyl silanes Reactions of vinyl silanes ‘Ate’ complexes from vinyl silanes

Introduction: Reagents for the Vinyl Anion Synthon Fresh from the discussion in chapter 15 on stereo-controlled alkene synthesis, did you notice that we avoided one disconnection entirely? We never considered combining a vinyl anion 2 with a carbon electrophile. This chapter concerns reagents for the vinyl anion synthon and particularly those that allow control over the stereochemistry of the double bond in the product. Please notice from the start that we mean vinyl anions where the nucleophilic group is in the plane of the alkene 3. This means that we shall chiefly be discussing vinyl metals such as vinyl-lithiums. These are metal σ-complexes 4 with the metal atom also in the plane of the alkene.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

256

16 Stereo-Controlled Vinyl Anion Equivalents

R1

R1

R1

3

R

R2

+

R2

H

R3

X

R2

Li

R2

H

1

R1

H

2

H

3

4

We shall not be using metal π-complexes 5 where the metal sits at right angles to the alkene plane 6. These are very important compounds and have their role to play in synthesis but they do not have the right properties to act as reagents for the vinyl anion synthon. Nor shall we be using the familiar organic reagents such as enamines 8 that act as π-donors at the same atom. The problem here is the same - though electrophilic attack occurs at the correct atom, it occurs at right angles to the alkene plane and stereochemistry is difficult to control. Instead we shall be describing vinyl metal σ-complexes 7. R1

R1

Li

R1

R2 H 5; M = metal

R1

M

M

R2 H

6; metal π-complex

H

R2N

R2

H 8; organic π-donor

7; metal σ-complex

Stereochemical information is built into reagents 9 that have a nucleophilic vinyl-metal bond in the plane of the alkene. This bond already has a stereochemistry - it is cis to R1 and trans to R2 and this information can be passed to the products providing that the reaction with electrophiles is stereospecific. In most of the methods in this chapter vinyl metals 9; M ⫽ Li, Mg, Cu, Sn, Al, Zr, or compounds closely related to them such as vinyl boranes 10 or vinyl silanes 11, react with electrophiles E⫹ with retention of configuration 12. R1

R1 M

R2

BR2

R2

H

R1

H 10

9

R1 SiR3

R2

9, 10, 11

E

E

R2

H 11

H 12

This reaction will be useful only if we are able to prepare these compounds with a fixed (E or Z) configuration. The first part of the chapter concerns the preparation and reactions of simple vinyl metal derivatives and we shall then progress to a study of stereochemistry.

Vinyl-Lithiums Unsubstituted vinyl-lithium and vinyl Grignard 14 reagents can be made directly from the halide by oxidative insertion of Li(0) and Mg(0). Vinyl-lithium is available as a 2M solution in THF from Alfa and vinyl magnesium bromide, which must be prepared in THF, is available in THF from Aldrich. These are quite stable σ-complexes because alkenyl anions are more stable than saturated alkyl anions. They add as nucleophiles to carbonyl groups, e.g. cyclobutanone to give 15 and prefer direct to conjugate addition with enones to give e.g. 13. We have already used them in enone synthesis (chapter 5). OH

O

O

OH

MgBr 13

14

15

16 Vinyl-Lithiums

257

Both the Grignard reagent 14 and vinyl-lithium can be converted into copper derivatives by exchange with Cu(I) halides. Two Cu(I)-catalysed conjugate additions of vinyl Grignard to enones start one of Corey’s gibberellin syntheses.1 In the absence of Cu(I) the vinyl Grignard reagent 14 adds directly to the carbonyl groups of 17 and 20.

O

O

O

(EtO)2P

BnO

14

O

BnO

KOH, EtOH/H 2O

Cu(I)

16

17

pyridine t-BuOH/H2O

18

CHO

OsO4 NaIO4

O

NaOH EtOH

O

BnO

O

BnO

O

14 Cu(I)

BnO

BnO

19

20

21

The extra stability of vinyl Grignard reagents is clearly shown by Knochel’s preparation of 22 by exchange with i-Pr2Mg. Notice that the alkenyl group is transferred rather than the alkyl group to benzaldehyde and that both metallation and reaction with the aldehyde occur with retention of configuration.2 OH Mgi-Pr

i-Pr2Mg

I

–20 ˚C, THF

Hex

PhCHO

Hex Ph 23; 60% yield

22

A few substituted vinyl halides are available, such as E- and Z-1-bromopropene, and a few can be made with stereochemistry. Both E- and Z-1-bromostyrene have been used in stereospecific additions to benzaldehyde with Cr(II) and catalytic Ni(II) as the metals.3 The synthesis of these isomers from cinnamic acid is described in the workbook. OH Br

Ph

Ph

Ph

CrCl2, cat NiCl 2 PhCHO, DMF

E-24

Ph

Br

Ph

OH

CrCl2, cat NiCl 2

Ph

PhCHO, DMF

E-25

Z-24

Z-25

A few more vinyl halides can be made stereospecifically by halogenation and base-catalysed elimination. One example is the vinyl bromide E-28 available by stereospecific trans bromination of crotyl alcohol 26 followed by stereospecific elimination.4 Various regioselectivities are available in the elimination reaction so the formation of that particular alkene is in a way more surprising than the stereopecificity of the reaction. Presumably the bromine atoms increase the acidity of nearby Hs (H-2 and H-3 in anti-27) so that one or other of the vinyl bromides will be formed. One explanation is an intramolecular elimination through an anti-peri-planar transition state in a chair like conformation using OLi as an internal base 26. It can reach H-3 in a five-membered cyclic array. H Br

Br Br2

OH 26; crotyl alcohol

LDA

OH

CCl4

Br anti-27

THF HMPA

Br

OH

E-28

Br

H O Me 29

Li

258

16 Stereo-Controlled Vinyl Anion Equivalents

This type of approach is inevitably limited and we need to look at more general methods of making alkenyl-lithiums since these can be converted into all the other reagents we might need.

Vinyl-Lithiums from Ketones: The Shapiro Reaction This extraordinary reaction converts a ketone into a vinyl-lithium 32. First an aryl sulfonyl hydrazone 31 is made by the obvious condensation with the aryl sulfonyl hydrazine 30. Treatment with two equivalents of butyl-lithium gives the vinyl-lithium.5

O + H2N

N

H N

S

H N

Ar

S

O

Ar O

Li 2 x BuLi

+ N2 + ArSO2

O O 30

31; arylsulfonylhydrazone

32; vinyl lithium

The first equivalent of butyl-lithium removes the rather acidic NH proton to give 33 and the second forms an aza-enolate 34. It may sound difficult to get two lithium atoms into the same functional group but the necessary sulfonyl group is very electron-withdrawing. Two β-eliminations now occur. The more reactive aza-enolate N⫺Li bond expels arylsulfinate anion to create an N⫽N double bond 35. Note that this is not the usual arylsulfonate (ArSO3⫺) containing S(VI) but the lower oxidation state aryl sulfinate (ArSO2⫺) containing S(IV). Though not such a good leaving group as sulfonate, it is still good. The remaining N⫺Li bond now lacks any stabilisation from sulfur and so can initiate the second β-elimination. The main driving force for this reaction is the formation of a molecule of nitrogen. The lithium atom is forced to migrate from nitrogen to carbon to form a vinyl-lithium 32. O O S Ar N

BuLi

31

N

Li

O O S Ar Li

N

N

BuLi

Li ArSO2

Li

N

N 32 + N2

33

34

35

Practically, it is important that the aryl group has no ortho hydrogen atoms as competing ortholithiation (chapter 7) may interfere with the reaction. The favoured aryl group is 2,4,6-tri-isopropylphenyl or ‘trisyl’ shown in the hydrazone 36. Hydrocarbon solvents such as hexane and addition of ligands for lithium such as TMEDA allow the reaction to go with maximum efficiency.

O + H2N

H N O

S

Tr O

N

H N

S

2 x BuLi TMEDA

O O

hexane

36;'trisyl' hydrazone

32 + N2 + TrSO 2

16 Vinyl-Lithiums from Ketones: The Shapiro Reaction

259

The vinyl-lithium may be combined directly with the carbon electrophile such as an alkyl halide or a carbonyl compound, or converted into another vinyl derivative such as the silane for later development. Many common examples use cyclic ketones. Cyclohexanone gives the Z-vinyllithium6 38 (Li has a lower priority than C) and hence the E-vinyl silane 39 (Si has a higher priority than C) in good yield on trapping with Me3SiCl. O

NHTr

N H2N

NHTr

Li

SiMe3 Me3SiCl

2 x BuLi TMEDA hexane

Z-38

37

E-39; 83% yield

This example has stereochemistry of a sort, but the other isomer is impossible. If the ketone is unsymmetrical, the double bond prefers to go into a methyl group 49, as one would expect for kinetically controlled aza-enolate formation (chapter 3), so no stereochemistry can be developed there. With symmetrical open-chain ketones, the lithium atom prefers to sit trans to the chain giving the Z-vinyl-lithium6 41. Reaction with an electrophile, such as DMF, occurs with retention of configuration giving the trans product E-42 (this is usually the E isomer as the electrophile usually has higher priority than saturated C!).

O

N H2N

NHTr

NHTr

Li

CHO Me2NCHO

2 x BuLi TMEDA hexane

40

Z-41

E-42; 70% yield

This process is obviously of limited value as the two side chains in the product will always differ by only one carbon atom. There is a way round this difficulty. If acetone trisylhydrazone 43 is converted to the dilithio derivative 44 and alkylated at low temperature, the monolithio-product 45; R ⫽ Li is stable. You may have noticed that we have always drawn the hydrazone with the NTr group on the same side as the double bond. This is no coincidence. When alkylation occurs this configuration of the C⫽N bond is maintained. H O

H2N

NHTr

N

N

Li Tr

Li 2 x BuLi

N

N

Tr

R1CH2I

N

N

Tr

–78 ˚C

TMEDA hexane

43

R

44

R1 45; R=Li

It appears that the lithiation site is determined by the geometry of the C⫽N bond. When a second (third?) lithiation is induced, it occurs on the same side as the first - and this is now the more crowded side 46. The regiochemistry is determined by the C⫽N bond but the stereochemistry remains as before. When the Shapiro reaction is allowed to go to completion, by warming to 0 ⬚C, the unsymmetrical Z-vinyl lithium Z-47 is produced. This is the isomer that cannot be made by the Shapiro reaction from the methyl ketone 48.

260

16 Stereo-Controlled Vinyl Anion Equivalents

Li Li

N

N

Tr

Li

Li

×

0 ˚C

2 x BuLi

45 R=H

O 1. H 2N.NHTr

TMEDA hexane

1. H 2N.NHTr 2. 2xBuLi

2. 2xBuLi

R1

R

1

R1

Z-47

46

R1

48

49

It can be quenched with electrophiles in the usual way. In this example prenyl bromide is the first electrophile giving the stable lithiated hydrazone 50: the Shapiro reaction gives Z-51 trapped as the vinyl stannane7 E-52. The vinyl silane 39 and stannane 52 products from these reactions can be used to make more elaborate alkenes with control over geometry as we shall see later. Li N

1. 2 x BuLi

43

N

Li Tr

SnBu3 2. Bu 3SnCl

1. s-BuLi

2. Br

Z-51

50

E-52; 90% yield

One recent application (that also introduces the next section) is to quench the vinyl-lithium Z54 with iodine and use the vinyl iodide E-55 in a Suzuki coupling8 (chapter 18) with a vinyl borane made by hydroboration of an acetylene.9 Li 1. 2 x BuLi DME

43

N

N

Tr

OMe

1. 2 x BuLi TMEDA

Li

OMe

I2

I

OMe

2. 0 ˚C

2. ArCH 2CH2I

53

Z-54

E-55; 77% yield OMe

1. 9-BBN, THF, 23 ˚C

Hex 56; 1-octene

2. E-55, NaOH PdCl2(dppf), 0 ˚C

E-57; 99% yield

A recent application involving a kinetic resolution (chapter 28) led to an asymmetric synthesis of the cotton boll weevil pheromone grandisol. Photochemical addition of ethylene to the cyclopentenone 58 gives the racemic ketone 59. The Shapiro reaction using simply tosylhydrazine works well and is completely regioselective in favour of the alkene not at the bridge. Trapping with DMF and Luche reduction gives the allylic alcohol 61. Attempts to make 61 by other methods gave much worse yields.10

16 The Aliphatic Friedel-Crafts Reaction

Me

Me H2N

hν CH2

Me 1. BuLi TMEDA

NHTs

MeOH

CH2

H

O

2. DMF 3. CeCl 3 NaBH4

H

O

H

N NHTs (±)-cis-60; 89% yield

(±)-cis-59; 77% yield

58

261

OH (±)-cis-61; 83% yield

Kinetic resolution (chapter 28) using the Sharpless asymmetric epoxidation with L-(⫹)di-isopropyl tartrate (chapter 25) removed the unwanted enantiomer as the epoxide 62 and left the required enantiomer (⫺)-61 for transformation into (⫹)-grandisol 63. Me

Me

Me

Me

H

H

t-BuOOH Ti(OPri)4

OH

+ O

(+)-DIPT

H

H

OH (±)-cis-61

OH

OH (–)-cis-61; 94% ee 31% yield

62

63; (+)-grandisol

The Shapiro reaction is important and useful but it is limited. We need to explore next a method with a wider scope - vinyl metals made from alkynes. The trick here is to add two things on the same side of the triple bond - typically one is a metal and the other a hydrogen atom or an organic fragment such as an alkyl group.

The Aliphatic Friedel-Crafts Reaction The application of the aliphatic Friedel-Crafts reaction (chapter 5) to alkynes is a stereoselective approach to vinyl-lithiums that forms a bridge between the Shapiro reaction and the next sectionhydrometallation. Acid chlorides 64 react with alkynes under Lewis acid catalysis to give E-β-chloroenones E-65 stereoselectively. The chlorine can be replaced by more reactive iodine by conjugate substitution.11 Cl

Cl

I

NaI acetone

AlCl3

O

O

O E-65

64

E-66

Both these reactions are stereoselective. Chloride prefers to add anti to the ketone in the linear vinyl cation intermediate 67. Rotation of the σ-bond in the enolate intermediate 69 in the conjugate substitution allows loss of chloride to give either the E- or the Z-vinyl iodide 66. Cl

R

Cl

R

I

R

E-65

H

O 67

E-66 I

O 68

Cl

O 69

This compound E-66 could be metallated with lithium directly but it would then destroy itself by reaction with its own keto group so reduction and protection first gives the silyl ether E-70. This vinyl iodide was converted into a vinyl-lithium E-71 and then into a vinyl cuprate E-72 for

262

16 Stereo-Controlled Vinyl Anion Equivalents

conjugate addition. We shall have more to say about these reactions below but note at this stage that both occur with retention of configuration at the alkene. I

1. NaBH 4

E-66

Li t-BuLi

OTBDMS E-70

2. t-BuMe 2SiCl

Pr

Cu (Me2N)3P

OTBDMS E-71 Cu

Li

OTBDMS E-72

Pr

Conjugate addition to cyclohexenone occurred very efficiently to give E-73 in 80% yield. It can be argued that both ends of the alkene in this product were linked to the rest of the molecule by vinyl anion additions to electrophiles.12 E-72

E-73

O

O

OTBDMS

Hydrometallation of Alkynes There are broadly three groups of metals that do this reaction. Some such as tin hydrides do so by a radical mechanism, boron and silicon by hydroboration-style reactions, and transition metals such as zirconium by π-complex formation. We shall take them in that order.

Hydrostannylation of terminal alkynes Tin hydrides readily form radicals. Typically Bu3SnH reacts with AIBN to give tin radicals that add to alkynes at the unsubstituted end to give the more substituted vinyl radical. The vinyl radical captures a hydrogen atom 74 from another molecule of Bu3SnH to produce another tin radical and complete the chain. The other product is a vinyl stannane Z-75. Here is the complete scheme: Bu3Sn Bu3Sn

H

AIBN

Bu3Sn

R

R H

H

SnBu3

Bu3Sn

R

H

H

Bu3Sn +

Z-75

74

This reaction shows a remarkable stereoselectivity. The kinetic product is the Z-alkenyl stannane Z-75. This is easily explained by arguments similar to those we used in chapter 15. The vinyl radical is on a linear (sp hybridised) carbon and reaction with a large reagent like Bu3SnH occurs preferentially on the less hindered side 74. This gives the Z-alkene Z-75. If the reaction is carried out at higher temperatures or with an excess of Bu3SnH the E-alkenyl stannane E-75 becomes the product. This is obviously the thermodynamic product and equilibration occurs by repeated addition 76 and loss 77 of Bu3Sn• radicals to the vinyl stannane. Bu3Sn

R

Bu3Sn

Bu3Sn

R

H

R

H

Bu3Sn

H

H H 76

H

Bu3Sn 77

E-75

16 Hydroboration and Hydroalumination of Alkynes

263

This example of the commoner thermodynamic route was used by Corey13 in the synthesis of prostaglandins as described below. Bu3Sn

Bu3Sn

H

AIBN

OSiMe2t-Bu

OSiMe2t-Bu

E-79

78

Hydroboration and Hydroalumination of Alkynes Boron and aluminium are group three (13) elements so their hydrides R2BH and R2AlH are dimers e.g. 81 linked by hydrogen bonds. Two useful compounds are 9-BBN 80 (9-BoraBicyclo[3.3.1]nonane) and DIBAL 82 (Di-IsoButylALuminium hydride, also known as DIBAH). Because of their large alkyl substituents these compounds are reasonably stable and can be bought. In solution they are in equilibrium with their monomeric forms which are active hydro-boration and -alumination reagents for triple bonds. The monomers 80 and 82 are trigonal with an empty p-orbital on the B or Al atom. H B

H

B

80; 9-BBN monomer

Al

B H

H

81; 9-BBN dimer

82; DIBAL monomer

Hydroboration of alkynes They react with terminal alkynes by electrophilic addition of the empty p-orbital to the unsubstituted end of the triple bond 83. The intermediate would then be the more substituted vinyl cation 84. It is easier to draw this mechanism with R2BH than with the full structure for 9-BBN. The intermediate 84 is not fully formed before hydride transfer begins so that the reaction is semiconcerted and the transition state is something like 86. The result is a regioselective and stereospecific cis hydroboration of the triple bond to give the E-vinyl borane 85. The intermediate 84 is quite like the radical intermediate in hydrostannylation but the difference is that hydrogen transfer is intramolecular and stereospecific in hydroboration. R2B

H

H R2B

R2BH

H

H

R

H 83

(–) H R2B (+) R

R

R

R

R2B

84

H

85

86

Here is a simple example in the field of prostaglandin synthesis where 9-BBN was used on a protected optically active propargyl alcohol.12 The starting material is identical to the alkyne 78 that we reacted with Bu3SnH above and the result is the same - cis hydrometallation with the metal atom at the terminus. However that was a thermodynamically controlled stereoselective radical chain reaction while this is a kinetically controlled stereospecific electrophilic addition to give the vinyl borane E-87. 9-BBN

OSiMe2t-Bu 78

R2B OSiMe2t-Bu

E-87

264

16 Stereo-Controlled Vinyl Anion Equivalents

Of recent years, boronic acids RB(OH)2 have become important in Suzuki couplings (chapter 18) and they too can be made by hydroboration of alkynes. Catechol borane14 89 is prepared from catechol 88 and diborane - a typical reaction of boranes with acidic OH groups to give stable B–O bonds - and is available commercially as a solution in THF. Hydroboration of an alkyne with catechol borane shows the usual regioselectivity and stereospecificity to give a terminal E-vinyl boronate ester15 90 and hydrolysis of the ester under very mild conditions14 (room temperature, no acid or base) gives the boronic acid16 91. We shall see later that you can argue whether these compounds can be considered as vinyl anion equivalents, but it seems sensible to discuss their preparation at this point. OH

O

B2H6

B

H

B

O

OH 88; catechol

R

O

no acid or base

E-90

89; catechol borane

OH

H2O 25 ˚C

O

R

HO

B

R

E-91; boronic acid

Treatment with iodine gives the E-vinyl iodides 93 stereospecifically and quantitatively. There is no doubt that these intermediates act as vinyl anion equivalents after exchange of the iodine for a metal. With bromine in methanol, protected 2-bromo aldehydes 92 are formed in good yield.16 Br

OH Br2, NaOMe

MeO

R

MeOH, –78˚C

OMe 92; 90-94% yield

HO

I2, NaOH

B

R

I

H2O, 0˚C

E-91; boronic acid

R

E-93; 100% yield

A dramatic example of this technique is the last step of Heathcock’s synthesis of myxalamide A 96. Hydroboration of the alkyne 94 with catechol borane gives the E-vinyl borane and this is combined with the Z-vinyl iodide 95 in a palladium-catalysed Suzuki coupling to give the natural product 96 with every alkene correct.17 OH 1. 89, PhNEt 2 2. 95, Pd(OAc) 2 i-Pr2NH, MeCN, H 2O

OH 94

96 44% yield

O

N

H

O

I

N

OH

H

OH

95

96

Hydroalumination of alkynes Reactions of terminal alkynes with DIBAL are very similar - syn addition leads to the E-vinyl alane and hence, by exchange with iodine or bromine, to E-vinyl halides.18 I 1. i-Bu2AlH

1. i-Bu2AlH

I

2. I 2, THF

97

E-98; 94% yield

2. I 2, THF

99

E-100; 72% yield

16 Hydrosilylation

265

Hydroalumination of propargyl alcohols Reduction of acetylenes does not usually occur with LiAlH4 but reduction of propargyl alcohols 101 with LiAlH4 occurs through a cyclic aluminium σ-complex 103 and is well controlled stereochemically. If the σ-complex 104 is trapped with an electrophile such as halogens or Bu3SnOTf, a simple Z-vinyl derivative, e.g. 105 results with a useful OH group for further development.19 The kinetic route to Z-vinyl stannanes above is a radical reaction and more difficult to control so it is just as well that there is a good route to a simple Z-vinyl stannane by a non-radical reaction. H2Al OH

H O

LiAlH4

H3Al 101

102

O

HO

H Al

H

H

H

O

H Al TfO

H

Bu3Sn

SnBu3

103

104

105

A similar process occurs with Grignard reagents.20 Though this is strictly carbometallation, the subject of the next section, it is very similar mechanistically to the reduction with LiAlH4. One magnesium atom guides a second Grignard reagent into the alkyne 106 to form the metallocycle 108 and hence, by reaction with, say, halogens, derivatives 109 with R1 and R2 cis (E or Z depending on priorities!). Two molecules of LiAlH4 or R2MgBr are needed and the intermediates 103 and 108 are vinyl anion equivalents. Cl Mg OH

O

O

R2MgBr

HO

Mg

E

R1

1

R

BrMg 106

R2 107

R1

E

R2

R1

108

109

R2

Hydrosilylation The reaction of alkynes with R3SiH can take place by three different mechanisms: radical, ionic catalysed by Lewis acids, and ionic catalysed by transition metals.21 The radical method is best with tris(trimethylsilyl) silane, a reagent introduced originally to take toxic tin out of radical chemistry. Promoted by Et3B in the presence of oxygen the stable radical 110 adds to the alkyne to give the linear radical 111 that collects a hydrogen atom from another molecule of tris(trimethylsilyl) silane to complete the cycle. As in the formation of vinyl stannanes 75 the large reagent delivers hydrogen anti to the large Si(SiMe3)3 substituent.22 Typical Z:E selectivity is ⬎20:1. The products 112 can be converted into Z-vinyl bromides by reaction with bromine in CH2Cl2 at 25 ⬚C. O2

Et3B

Et

H

Si(SiMe3)3

EtH

Si(SiMe3)3

R

Si(SiMe3)3

H

H

H

R R

Si(SiMe3)3

(Me3Si)3Si

110

H 111

Z-112

Hydrosilylation of alkynes catalysed by Lewis acids Lewis acid-catalysed addition of Et3SiH (Me3SiH cannot be used as it is unstable) also gives the – Z-vinyl silane 115. This is trans addition to the alkyne and probably occurs by ‘H ’ transfer to the

266

16 Stereo-Controlled Vinyl Anion Equivalents

π-complex 113 followed by electrophilic substitution of the ‘ate’ complex 114 with retention of configuration. The substituent R can be alkyl or aryl and can have an alkyloxy functionality.23 AlCl3

Et3SiH

R

R

AlCl3

H

H

R

AlCl3 114

113

Et3SiCl

H

H

R

SiEt3 Z-115

+ AlCl3

Hydrosilylation of alkynes catalysed by transition metals The commonest combination here is trichlorosilane with H2PtCl6 as catalyst. Cis addition of the silane is preferred, probably via a metal complex such as 116, giving the E-vinyl silane 117. The SiCl3 group is not easily replaced but the dianion 118 formed with KF reacts cleanly with NBS to give the E-vinyl bromide 119 with retention of configuration.24 R

R

Cl3SiH

R

cat H2PtCl6

H

Pt SiCl3 116

H

H

SiCl3

KF

R

H

H

E-117

2K

NBS

SiF5

R

H

H

E-118

Br

E-119

R can be alkyl or aryl and can carry various functional groups such as the ester in 120. The vinyl bromide E-121 is formed in 68% yield and is ⬎95:5 E:Z. However other transition metals may give the Z-vinyl silane as is the case with Et3SiH and an Ir(I) catalyst.25 Yet other transition metals give the opposite regioselectivity as with Et3SiH and a Ru(I) catalyst.26 The best advice is to use conditions already determined by reliable workers. 1. Cl 3SiH, H 2PtCl6

MeO2C

MeO2C

Br

2. KF, 3. NBS

E-121

120

Hydrozirconation Among transition metals, zirconium is the most important for hydrometallation of alkynes. The available 16 electron complex Cp2ZrHCl (Cp ⫽ cyclopentadienyl) 122 (zirconocene hydrochloride) adds stereoselectively syn to alkynes to give the E-vinyl zirconium species 126. The initial interaction is the donation of two electrons by the triple bond 123 to make the reasonably stable 18-electron π-complex (or η2-complex) 124. Typically for a transition metal π-complex, a ligand is now transferred from the metal to one end of the π-bond 125 while the metal itself forms a stable 16-electron σ-complex at the other 126. The H transfer is intramolecular so the metal and H atoms must add to the same side of the triple bond. The zirconium atom transfers the least stable – – anion among its ligands. This is obviously H as Cl is much more stable. The metal prefers to take the terminal position because the resulting σ-complex is more stable as it is a σ-complex of the less-substituted carbanion.27

H Zr Cl

Cp2Zr

Cl

Cp2Zr

H R

122; Cp 2ZrHCl 16 electrons

123

Cl

Cp2Zr

H R

124; π-complex 18 electrons

Cl

Cl

H R

125; hydrozirconation

Cp2Zr

H R

126; σ-complex 16 electrons

16 Carbo-Metallation

267

Here is an example from the McMurry flexibilene synthesis quoted in chapter 1. An alkyne 127 with a protected aldehyde group reacts with Cp2ZrHCl to give a vinyl zirconium complex 128 which is coupled to a palladium-allyl complex in the next step. The E double bond so produced is present with the same E configuration in the final product. It is marked 129 with an arrow in the diagram.28

OMe

OMe Cp2ZrHCl

MeO

ZrCp2Cl

MeO

127

128

129

Exchange of the zirconium for zinc allows addition to carbonyl compounds. A free alcohol group, as in 130, must be protected and the zirconium in the initial product 132 replaced by zinc before the aldehyde is added to give the allylic alcohol 134. Both reactions occur with retention of configuration.29 HO

t-BuPh2SiCl imidazole DMAP, CH 2Cl2

130 Me2Zn

RO

Cp2ZrHCl

ZnMe

–60 ˚C

ZrCp2Cl 132

131; 96% yield RCHO

RO

RO

t-Bu

0 ˚C, CH 2Cl2

Si

Ph

O Ph

133

OH E-134; 66% yield

Carbo-Metallation Some of the metals we have been discussing can be used for the simultaneous cis-addition of an alkyl or aryl group and a metal to an alkyne.30 This process is called carbometallation and involves particularly Li, Mg, Al, Cu, and Zn.

Carboalumination Trialkyl aluminiums are not reactive enough to add to alkynes but do with catalysis particularly from Cp2ZrCl2. In reaction with Me3Al, the methyl and the AlMe2 groups are added to the same side of the triple bond to give an unstable alkenyl aluminium E-136 that is reacted immediately with an electrophile to give the product E-137 with retention of configuration.31 The mechanism is uncertain but almost certainly involves transfer of a methyl group from zirconium in a π-complex such as 135 with a number of chloride ligands on one or both metals, or even between them.32 Me3Al

R 0.4 equivs Cp2ZrCl2

R

carboalumination

Zr

Me 135

AlMe2

R

AlMe2 Me E-136

E

R

E Me

E-137

Palladium-catalysed coupling of such an intermediate E-139 with benzyl chloride gives the trisubstituted alkene E-140 as the only product in high yield.33

268

16 Stereo-Controlled Vinyl Anion Equivalents

R

Me3Al 0.4 equivs Cp2ZrCl2

138

Ph

5 mol% Pd(0)

AlMe2 Me E-139

Ph

Me E-140; 92% yield

Cl

A recent application to the synthesis of the macrolide antibiotic concanamycin used carbometallation of the alkyne 141 to give the vinyl iodide E-142 followed by a palladium-catalysed coupling with a vinyl stannane, also created from an alkyne, to give the diene34 E,E-143. OMe

I 1. Me 3Al Cp2ZrCl2

OR1

OR1

OR1

O

2. I 2

R1O

OR2

R1O

OR2

R1O

E-142; 88% yield

141

OR2

O

E,E-143; 72% yield

Carbocupration Various organo-copper compounds carry out carbocupration of alkynes without any extra catalyst. The reagents may be RCu (usually with some ligand such as Me2S), Grignard reagents with added Cu(I), or any of the more complex cuprates also used in conjugate addition (chapter 9). Cis addition is the rule with the metal ending up on the terminal carbon atom35 E-145. carbometallation

R1

R2Cu 1

R

Cu

L = ligand

R2 144

R1

Cu

E

R2

L

R1

E R2

E-146

145

The electrophile (E in E-146) can be an alkyl halide, acid chloride, conjugated carbonyl compound and so on. A simple example illustrating the use of a functionalised alkyl halide, produces the Z-alkene 148 as the larger group is added by carbocupration.36 Addition in the reverse order would have given E-148. 1. Cu 2Br2, LiBr Et2O, –15 ˚C

EtMgBr

I

Me

2. Me

CO2Me

Cu.MgBr2 Et

Me

CO2Me Et

Z-148; 56% yield

147

We expect vinyl copper derivatives to be best at conjugate addition but we are disappointed at the few enones that react satisfactorily. Carbocupration of propyne with n-hexylcopper gives the intermediate 149 that adds to cyclohexenone with retention of the Z-configuration to give the conjugate addition product37 150. With open chain enones results are not so good. n-Hex 1. Cu 2Br2.SMe2

n-HexMgBr

Cu.MgBr2

2. Me

cyclohexenone

O 149

Z-150; 73% yield

16 Reactions of Vinyl Sn, B, Al, Si, and Zr Reagents with Electrophiles

269

Reactions of Vinyl Sn, B, Al, Si, and Zr Reagents with Electrophiles So far we have given an assortment of ways these vinyl metals might be used in synthesis as vinyl anion equivalents. Now it is time to consider their reactions in more detail. One of the commonest ways, and certainly the most obvious way in which they act as vinyl anions, is to transform them into vinyl-lithium reagents. This can be done directly or via halogenation. After carbocupration, exchange with iodine and then lithium gives a vinyl-lithium that reacts cleanly with enolisable aldehydes and ketones to give allylic alcohols, e.g. E-152, all with retention of configuration.38 1. Cu 2Br2.SMe2

[Cu]

EtMgBr

OH

1. I 2, Et 2O, 0 ˚C 2. BuLi, Et 2O –70 ˚C 3. Me 2CO

2. Bu

151

E-152; 87% yield

Treatment of vinyl Sn, B, or Al compounds with BuLi results in effective addition of Bu⫺ to the metal to form a hypervalent anion such as 154. These are often referred to as “ate” complexes. The analogy is with the names of anions such as sulfate or carbonate. You are already familiar with the copper analogues, usually called cuprates. Lithium now replaces tin at the vinyl group 155 to form a vinyl-lithium derivative E-156. The reaction is an electrophilic substitution at carbon - the lithium atom attacks the C-Sn bond and does so with retention of configuration. Li BuLi

Bu3Sn

R

Bu4Sn

E-153

Bu4Sn

R

E-154

Li

R

R

+ Bu4Sn

E-156

155

Vinyl copper derivatives such as 157 do not react with epoxides but transformation of the vinyl copper into a cuprate by the addition of pentynyl lithium gives a cuprate that preferentially transfers the vinyl group (the less stable anion is transferred from copper) to ethylene oxide to give the homoallylic alcohol39 E-159. Note that 157 has the opposite stereochemistry to 149. 1. Cu 2Br2.SMe2 2.

Pr

[Cu]

MeMgBr R

Li

157 O

Li

Cu 158

OH

E-159; 75% yield

Pr

A lithium atom may also be replaced by copper to make a cuprate 161. As in the previous example, using pentynyl copper ensures that no vinyl group is wasted. Such cuprates are superior to simple copper derivatives in conjugate addition to give 162 with enones. All these reactions happen with retention at the vinylic carbon. O Pr

Li

R 160

Cu

Li

Cu

O

R +

Pr 161

162

R

270

16 Stereo-Controlled Vinyl Anion Equivalents

An excellent illustration of both these types of vinyl anion equivalent appears in Corey and Wollenberg’s synthesis40 of the antibiotic brefeldin A 163. This interesting molecule has two E-alkenes, one attached to a hydroxyl group as an allylic alcohol, and so accessible by direct addition of a vinyl lithium 165 to an aldehyde, and one in a 1,3-relationship with an alcohol and so accessible in theory by conjugate addition of a vinyl copper (or cuprate) 166 to the simple double electrophile 164. Each OH group must be protected. H

OH

O

O

HO

165: vinyl-lithium OR for direct addition

Li

O

OR

O

H

Cu

H 163; brefeldin A

164

166; vinyl-copper for conjugate addition

Each vinyl metal reagent was prepared by thermodynamic hydrostannylation of the corresponding alkyne.41,42 The simpler lithium derivative 165; R ⫽ CH2SMe came from the protected propargyl alcohol 167. OH

O

Bu3SnH

SMe 101

Bu3Sn

O

SMe

AIBN

167

168 BuLi

Li

THF –78 ˚C

O

SMe

165; R = CH 2SMe

The vinyl stannane 170 for the copper derivative came from another protected alcohol 169 by a similar route but the protecting groups were made quite different so that they could be removed selectively. OSiMe2t-Bu

Bu3Sn

OSiMe2t-Bu

Bu3SnH

166

Bu3Sn

AIBN

169

170; 94% yield

The vinyl stannane 170 was converted into the vinyl-lithium and then the vinyl cuprate as described for 161 and added to the enolate anion of 171. Notice the protection of the malonate ester group in 162 by deprotonation with sodium hydride. The product 172 is all anti across the five-membered ring and all E, but there is of course no control at the remote stereogenic centre on the side chain. CO2Et

CO2Et 1. BuLi

170 2. Pr

Cu

pentynyl cuprate + O of 170

CO2Et 171

O

CO2Et

OSiMe2t-Bu

172; 84% yield from 170

Functional group manipulation now gives the aldehyde 173 ready for addition to the vinyllithium reagent 165. The product was trapped as the MEM (methoxy-ethoxy-methyl-) derivative 174 in a very high yield but without any stereoselectivity. Fortunately, Corey already knew how to control the stereochemistry at the two OH groups by oxidation to ketones and stereoselective reduction so the synthesis could be completed.

16 Reactions of Vinyl Sn, B, Al, Si, and Zr Reagents with Electrophiles

CHO

OSiMe2t-Bu

O

H

1. 165; R = CH2SMe

271

OMEM OR

O

2. MEMCl

OSiMe2t-Bu

H 174; 82% yield

173

Reactivity of vinyl alanes Vinyl aluminium compounds also carry out conjugate additions and illustrate well the difference in reactivity between vinyl alanes and the related “ate” complexes. If the enone can adopt the s-cis conformation, as in 175, a cyclic mechanism is possible in which aluminium both acts as a Lewis acid and delivers the vinyl nucleophile. Notice that the E-vinyl alane leads to the E-γ,δ-unsaturated ketone 178 as this is again an SE2 reaction at the vinyl carbon.32 i-Bu

i-Bu2Al

R

i-Bu

R

Al

O

i-Bu2Al

R

O

O

O

H H2O

175

176

R

177

178

If the enone cannot adopt the s-cis conformation, it is necessary to convert the vinyl alane into an “ate” complex 179 before conjugate addition. Conjugate addition now occurs to s-trans enones, such as cyclic enones, to give the E-γ,δ-unsaturated ketone 180 after aqueous work-up. Notice that this time the most stable anion is transferred - the choice is between Me, i-Bu, or vinyl. Aluminium is not a transition metal and simply releases the most stable anion as there must be some negative charge on the group being transferred. Me i-Bu2Al

MeLi

R

+

i-Bu2Al

R

O

179

O

180

R

We have already seen examples of cuprates being used in stereochemically controlled conjugate additions to cyclopentenones and similar results can be achieved with ate complexes of vinyl alanes prepared by hydroalumination of alkynes to give e.g. 179; R ⫽ hexyl and hence the antiE-product43 182. O

O R

Me 1. i-Bu2AlH

Hex

i-Bu2Al

2. MeLi

CO2Et

+ Hex

179; R = n-hexyl

181

182; 76% yield

Preparation of vinyl silanes We have seen that vinyl silanes can be prepared by hydrosilylation of alkynes by three different mechanisms giving good control over geometry of these inevitably terminal vinyl silanes. Vinyl silanes are stable compounds and can be isolated, unlike most of the vinyl metals we have seen so far, and other ways of making vinyl silanes allow the more-or-less controlled synthesis of monoor trisubstituted compounds with reasonable control over selectivity. These include the Peterson reaction with two SiMe3 groups on the same carbon atom 183 and, more relevant to this chapter, reactions of vinyl lithiums with silyl chlorides.44

272

16 Stereo-Controlled Vinyl Anion Equivalents

R 1. base

Me3Si

SiMe3

Me3Si

Me3Si

O

2. RCHO

R

SiMe3 184

183

185

Hydrometallation or carbometallation of alkynyl silanes 186 is easier because of the presence of the silicon atom as the metal prefers to be at the silylated end of the resulting alkene 187. Trapping these intermediates with electrophiles allows the stereochemically controlled synthesis of almost any vinyl silane 188. The metal can be Li, Mg, Al, or Cu and both 187 and 188 are single geometrical isomers (whether E or Z depends on priorities of M and the various R groups).30 R2

R2[M]

R1

SiMe3

R2 [M]

R1

R3X

SiMe3 187

186

R3

R1

SiMe3 188

Reactions of vinyl silanes Vinyl silanes resemble alkenes in reactivity: they combine with reactive electrophiles such as bromine without catalysis but need Lewis acid catalysis for reaction with carbon electrophiles. Reaction usually occurs 189 at the silyl end of the alkene so that the intermediate 190 enjoys the β-silyl stabilisation of the carbocation. The silyl group is removed by a nucleophile, usually a halide ion.45

Me3Si

R

X

E 189

Me3Si

E

R

R

E 190

191

If the electrophile also benefits from Lewis acid catalysis, as in the aliphatic Friedel-Crafts reaction, good yields of products are generally found. The anion of the allyl phosphonate 192 can be silylated to give the vinyl silane 193 and this reacts in turn with the classical combination of acetyl chloride and AlCl3 to give the enone E-194 that can be used in HWE reactions to make dienones46 (chapter 15). O (EtO)2P 192

O 1. 2 x LDA 2. Me 3SiCl

(EtO)2P

O SiMe3

E-193; 96% yield

MeCOCl AlCl3, 0 ˚C CH2Cl2

O

(EtO)2P

E-194; 89% yield

The conversion of 193 to 194 occurs with retention of configuration. Many such reactions give only the more stable of the two possible products, depending on the lifetime of the β-silyl cation 190. In favourable circumstances the reaction is stereospecific with retention. The initial attack of the electrophile occurs on either face of the alkene by interaction with the π-bond. As the new C⫺E bond is formed the Me3Si group starts to rotate away from the electrophile as the C-atom becomes tetrahedral. This rotation continues until the C⫺Si bond is parallel with the empty porbital on the other carbon atom and then stops. Loss of the Me3Si group gives the product with retention of alkene geometry.

16 Reactions of Vinyl Sn, B, Al, Si, and Zr Reagents with Electrophiles

E

E

Me3Si

R

E

R

Me3Si H

electrophilic attack

Me3Si starts to rotate downwards

E

R SiMe3

273

R

E

R

SiMe3

Cl

Me3Si rotation complete

product formed with retention

Me3Si lost

Of course, if the intermediate cation, stabilised by the β-silyl cation effect, has a lifetime long enough to allow rotation about the C⫺C bond, the stereochemical information is lost. Clear-cut examples with retention are often intramolecular and the most valuable are those that create an exocyclic alkene. A famous example is the cyclisation of the two acetals E- and Z-195 with SnCl4 as the Lewis acid to give reasonable yields of the oxepanes 196 with complete retention of configuration at the exocyclic alkene.47 t-Bu

SnCl4

CH2Cl2

O SiMe3

O SiMe3 t-Bu

O

–15 °C

OMe

O

E-195

t-Bu

SnCl4

E-196; 57% yield, >98% E Z-195

CH2Cl2

t-Bu

O

–15 °C

OMe

O

Z-196; 71% yield, >98% Z

The lifetime of the key intermediate can sometimes be controlled by choice of Lewis acid. Copper-catalysed conjugate addition of the vinyl Grignard reagent 198 to the enone 197 gives the anti adduct 199 as expected. Intramolecular FriedelCrafts reaction of the corresponding acid chloride with AgBF4 as this Lewis acid gave only the less stable Z-enone 200 but other Lewis acids such as TiCl4 gave E/Z mixtures.48 O 1. Cu(I)

O

SiMe3

+ BrMg CO2t-Bu 197

SiMe3

2. NaI Me3SiCl

CO2H

1. (COCl)2

O

2. AgBF4 MeNO2

H

H 199

198

O

Z-200; 71% yield

Even replacement of silicon by a humble halide can be important in complex syntheses. The epothilones, e.g. 201, are anti-cancer compounds of some promise. Chemoselective epoxidation of intermediate 202 is possible and Shibasaki decided to disconnect at the obvious ester and the far less obvious diene requiring a C⫺C bond to be made stereoselectively between two alkenes49 203 and 204. S

S

S

N

N

N

H

O

H

O

H

O

ester linkage H

O

OH

OH

O

O

OAc

PhO2C

OH

O H

I OH

201; epothilone A

OH 202

O diene linkage 203

OTBDMS 204

274

16 Stereo-Controlled Vinyl Anion Equivalents

Each half of the molecule was made as a single enantiomer by catalytic asymmetric synthesis. We are concerned with the vinyl iodide 203. The obvious reduction of 205 with DIBAL failed and instead hydrotitanation gave the Z-vinyl silane 206. Replacement of Me3Si with iodine was regioand stereospecific giving only the Z-vinyl iodide in 59% yield. S

S Ti(OPri)4

N H

1. I2, CH2Cl2

N

i-PrMgBr

OTBDMS

S

H

Et2O –78 °C

Me3Si

OH

N H

2. Ac2O Et3N DMAP CH2Cl2

I 203

SiMe3 206

205

OAc

‘Ate’ complexes from vinyl silanes As we have seen with E-118, ‘ate’ complexes with two negative charges on silicon are possible and so it should come as no surprise that one fluoride ion can be added to most silyl groups to give an ‘ate’ complex with just one negative charge. These complexes now have a C⫺Si bond of greater nucleophilicity than the π-bond and so behave like the other vinyl metals in this chapter. Silylation of the alcohol 207 with a silyl halide having a Si⫺H bond opens the way for Pt-catalysed intramolecular hydrosilylation (cf. above) to give only the E-vinyl silane50 209. H

i-Pr2Si

OH

i-Pr

Cl

O

Et3N, DMAP

207

i-Pr Si

H

208

O

H2PtCl6.6H2O CH2Cl2

Si

i-Pr 209

i-Pr

Addition of fluoride (as TBAF, the salt Bu4NF) gives the ‘ate’ complex 210 that couples with aryl and alkenyl halides to give a single isomer of a trisubstituted unsaturated alcohol, e.g. 202, with retention at the alkene. OH O

2 x TBAF

209

i-Pr

Si F i-Pr

210

PhI, 5 mol% Pd(dba) 2 THF, –45 ˚C

E-211; 88% yield

The scope of vinyl metals as sources of nucleophilic vinyl groups is very great. As well as the expected electrophiles such as halogens, alkyl and acyl halides, aldehydes and ketones, unsaturated carbonyl compounds and epoxides, they also combine with aryl and alkenyl halides with palladium catalysis. The usual stereochemical course is retention at the vinyl group. It is necessary to decide whether the vinyl metal is reactive enough or whether it must first be transformed into an ‘ate’ complex. Since most of these vinyl metals can be converted into each other with retention, this is an unusually versatile group of reagents.

16 References

275

References General references are given on page 893 1. E. J. Corey and J. G. Smith, J. Am. Chem. Soc., 1979, 101, 1038; Vogel p 539. 2. M. Rottländer, L. Boymond, G. Cahiez and P. Knochel, J. Org. Chem., 1999, 64, 1080; A. Boudier, L. O. Bromm, M. Lotz and P. Knochel, Angew. Chem., Int. Ed., 2000, 39, 4415. 3. K. Takai, K. Sakogawa, Y. Kataoka, K. Oshima and K. Utimoto, Org. Synth., 1995, 72, 180. 4. E. J. Corey, H. F. Wetter, A. P. Kozikowski and A. V. Rama Rao, Tetrahedron Lett., 1977, 777; M. G. Bock, A. P. Kozikowski, A. V. Rama Rao, D. Floyd and B. Lipshutz, Tetrahedron Lett., 1978, 1051. 5. A. R. Chamberlin and S. H. Bloom, Org. React., 1990, 39, 1. 6. A. R. Chamberlin, J. E. Stemke and F. T. Bond, J. Org. Chem., 1978, 43, 147. 7. E. J. Corey and J. P. Dittami, J. Am. Chem. Soc., 1985, 107, 256. 8. N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2457. 9. E. J. Corey and B. E. Roberts, Tetrahedron Lett., 1997, 38, 8919. 10. D. P. G. Hamon and K. L. Tuck, J. Org. Chem., 2000, 65, 7839. 11. C. C. Price and J. A. Pappalardo, Org. Syn. Coll., 1963, IV, 186. 12. E. J. Corey and D. J. Beames, J. Am. Chem. Soc., 1972, 94, 7210. 13. E. J. Corey and N. Raju, Tetrahedron Lett., 1983, 24, 5571; E. J. Corey, L. K. Niimura, Y. Konishi, S. Hashimoto and Y. Hamada, Tetrahedron Lett., 1986, 27, 2199, and see correction on page 3556. 14. H. C. Brown and S. K. Gupta, J. Am. Chem. Soc., 1972, 94, 4370; R. C. Larock, S. K. Gupta and H. C. Brown, J. Am. Chem. Soc., 1972, 94, 4371. 15. N. Miyaura and A. Suzuki, Org. Synth., 1989, 68, 130; T. Ishiyama, M. Murata and N. Miyaura, J. Org. Chem., 1995, 60, 7508. 16. T. Hamaoka and H. C. Brown, J. Org. Chem., 1975, 40, 1189. 17. A. K. Mapp and C. H. Heathcock, J. Org. Chem., 1999, 64, 23. 18. G. Zweifel and J. A. Miller, Org. React., 1984, 32, 375. 19. E. J. Corey and T. M. Eckrich, Tetrahedron Lett., 1984, 25, 2415, 2419. 20. J. G. Duboudin and B. Jusseaume, J. Organomet. Chem., 1979, 168, 1. 21. The Science of Synthesis (Houben-Weyl) Vol 4 page 740, Thieme, 2000, ed Ian Fleming. 22. K. Miura, K. Oshima and K. Utimoto, Bull. Chem. Soc. Japan, 1993, 66, 2356. 23. N. Asao, T. Sudo and Y. Yamamoto, J. Org. Chem., 1996, 61, 7654. 24. K. Tamao, J. Yoshida, H. Yamamoto, T. Kakui, H. Matsumoto, M. Takahashi, A. Kurita, M. Murata and M. Kumada, Organometallics, 1982, 1, 355. 25. R. S. Tanke and R. H. Crabtree, J. Am. Chem. Soc., 1990, 112, 7984. 26. B. M. Trost and Z. T. Ball, J. Am. Chem. Soc., 2001, 123, 12726. 27. J. A. Labinger, Comp. Org. Synth., 8, 667. 28. J. McMurry, Acc. Chem. Res., 1983, 16, 405. 29. P. Wipf and W. Xu, Org. Synth., 1997, 74, 205. 30. J. F. Normant and A. Alexakis, Synthesis, 1981, 841. 31. D. E. Van Horn and E. Negishi, J. Am. Chem. Soc., 1978, 100, 2252. 32. E. Negishi and D. Y. Kondakov, Chem. Soc. Rev., 1996, 25, 417. 33. E. Negishi, H. Matsushita and N. Okukado, Tetrahedron Lett., 1981, 22, 2715. 34. T. Jyojima, N. Miyamoto, M. Katohno, M. Nakata, S. Matsumura and K. Toshima, Tetrahedron Lett., 1998, 39, 6003; 6007. 35. B. H. Lipshutz and S. Sengupta, Org. React., 1992, 41, 135. 36. J. F. Normant, G. Cahiez, C. Chuit and J. Villeras, J. Organomet. Chem., 1974, 77, 269; J. F. Normant, C. Chuit, G. Cahiez and J. Villeras, Synthesis, 1974, 803. 37. A. Marfat, P. R. McGuirk and P. Helquist, J. Org. Chem., 1979, 44, 3888. 38. G. Cahiez, D. Bernard and J. F. Normant, Synthesis, 1976, 245. 39. Y. Horiguchi, S. Matsuzawa, E. Nakamura and I. Kuwajima, Tetrahedron Lett., 1986, 27, 4025. 40. E. J. Corey and R. H. Wollenberg, Tetrahedron Lett., 1976, 4705. 41. E. J. Corey and R. H. Wollenberg, J. Org. Chem., 1975, 40, 2265.

276

16 Stereo-Controlled Vinyl Anion Equivalents

42. E. J. Corey and M. G. Bock, Tetrahedron Lett., 1975, 3269. 43. K. F. Bernady and M. J. Weiss, Tetrahedron Lett., 1972, 4083. 44. P. F. Hudrlik, A. K. Kulkarni, S. Jain and A. M. Hudrlik, Tetrahedron, 1983, 39, 877; L. E. Overman, M. J. Brown and S. F. McCann, Org. Synth., 1989, 68, 182. 45. I. Fleming, J. Dunoguès and R. Smithers, Org. React., 1989, 37, 57; T. A. Blumenkopf and L. E. Overman, Chem. Rev., 1989, 89, 857. 46. B. S. Lee, S. Y. Lee and D. Y. Oh, J. Org. Chem., 2000, 65, 4175. 47. L. E. Overman, A. Castañeda and T. A. Blumenkopf, J. Am. Chem. Soc., 1986, 108, 1303. 48. K. Fukuzawa, E. Nakamura and I. Kuwajima, Tetrahedron Lett., 1984, 25, 3591. 49. D. Sawada, M. Kanai and M. Shibasaki, J. Am. Chem. Soc., 2000, 122, 10521. 50. S. E. Denmark and W. Pan, Org. Lett., 2001, 3, 61.

17 Electrophilic Attack on Alkenes Introduction: Chemo-, Regio-, and Stereoselectivity Chemoselectivity Controlling Chemoselectivity Regioselectivity ‘Markovnikov’ Hydration Mercuration-reduction Wacker oxidation Hydroboration: ‘Anti-Markovnikov’ Hydration of Alkenes Mechanism, regio- and stereoselectivity of hydroboration Alternative Approaches to the Synthesis of Alcohols from Alkenes Regioselective reduction of epoxides Intramolecular hydrosilylation Stereoselectivity Selectivity by Intramolecular Interactions Halolactonisation The synthesis of vernolepin Halolactonisation in the synthesis of erythronolides Sulfenyl- and Selenenyl-Lactonisation The Prins Reaction The original Prins reaction The formation of tetrahydropyrans by the Prins reaction The oxo-ene mechanism Stereoselectivity in the Prins reaction Double Prins reactions: use of Sn and Si to control cation formation Hydroboration as a Way to Make Carbon-Carbon Bonds Carbonylation of Alkyl Boranes Carbon monoxide Ketone synthesis using cyanide ion Reactions with α-halo-carbonyl compounds Polyene Cyclisations Looking Forwards

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

278

17 Electrophilic Attack on Alkenes

Introduction: Chemo-, Regio-, and Stereoselectivity The last two chapters have been concerned with making alkenes of known stereochemistry. An important reason for wanting to do this is that alkenes have unusual potential as starting materials for making polyfunctional compounds with control over all aspects of selectivity. O

O CO2Me

1

2

3

4

5; Cecropia juvenile hormone

Chemoselectivity is often easy to control in alkenes because the alkene is a “weak” functional group. It does have inherent reactivity with electrophiles such as bromine but when another functional group, whether electron-donating or electron-withdrawing, is conjugated to the alkene, the reactivity is normally dominated by the other functional group. Compounds 2–4 would normally not be described as alkenes but as benzene, an enone, and an enol ether. But they are all alkenes. You will already know many reactions that would happen with one of these compounds and not at all with another. The triene 5 has three alkenes in the same molecule and we shall want to react just one and not the others. Chemoselectivity. OH

O

R

R

6

R X

7

R

Br

R

H2O

8

OH

Br

Br2

mCPBA

X

9

10

Regioselectivity usually arises when an unsymmetrical electrophile attacks an unsymmetrical alkene. Familiar examples would be the action of bromine water or peroxyacids on a terminal alkene 8. The intermediate bromonium ion would be attacked by water at the more substituted end 9 and the epoxide by a nucleophile at the less substituted end 7. OH

OH Br2

mCPBA

X

O

Br

H2O

X

Br

11

12

13

1

14

Stereoselectivity comes from a stereospecific syn or anti addition to an alkene of fixed and known geometry. These last reactions, applied to cyclohexene, lead to anti bromohydrin 14 while epoxidation occurs stereospecifically and syn. It doesn’t matter which end of the epoxide 12 or bromonium ion 13 is attacked by the nucleophile: anti addition occurs in both cases since inversion is demanded by the mechanism of the SN2 reaction. Cyclohexene must be Z but in open chain compounds syn addition to the E-isomer would lead to the same diastereoisomer of the product as anti addition to the Z-isomer. In this chapter we explore more advanced versions of these reactions in which usually several types of selectivity will be combined and show how they are used in synthesis. Br

OH R

R

OsO4

R

R

H2O

OH syn-15

E-16

OH

Br2

R

R

Br

R R

R

OsO4

R

Br2 H2O

OH

OH

anti-17

anti-15

R

R OH

Z-16

syn-17

17 Chemoselectivity

279

Chemoselectivity The commonest form of chemoselectivity involves the preferential attack of electrophiles on electron-rich alkenes as in the ozonation of a Birch reduction product 19 used in chapter 15 to make a Z-alkene Z-20; R⫽Me and hence Cecropia juvenile hormone 5. OMe

OMe

Birch reduction

2. NaBH 4

R

R

CO2Me

1. O 3

5 R

OH

19

18

20

The alkene that is attacked is an enol ether and is much more nucleophilic than the other simple alkene. Similarly nucleophilic reagents attack only alkenes that are conjugated with an electronwithdrawing group. Chemoselective epoxidation is usually straightforward. The peracid epoxidation is stereospecific: the alkene 22 is E and so the epoxide 23 is trans (or anti). The other epoxidation is stereoselective as it is a two stage process and gives the more stable trans (or anti) epoxide 21 by choice. There is an example later in the synthesis of vernolepin. O

O

H2O2

RCO3H

NaOH

O

O

21

O

E,E-22

23

There is often no difficulty either if the two alkenes are the same, whether conjugated or not, as some reactions will stop cleanly after one alkene has reacted. A conjugated diene is more nucleophilic (higher energy HOMO) than a simple alkene so the first product, e.g. 25 is a less nucleophilic alkene than the starting material. The monoepoxide of cyclopentadiene 25 can be made by direct epoxidation of cyclopentadiene itself 24 provided that the solution is buffered with sodium carbonate to prevent the acid by-product (MeCO2H) from decomposing 25. Cycloocta-1,5-diene 26 can also give the mono-epoxide 27 cleanly. The difference in reactivity is less here - probably only about a factor of two - there are two equal double bonds in the diene and only one in the epoxide - and this approach is less reliable.1 MeCO3H

MeCO3H

O

Na2CO3

24

O

Na2CO3

25

26

27

Even if the only difference between the two alkenes is the number of substituents, that can be enough for some reactions. If the substituents are simple alkyl or aryl groups, then the more highly substituted alkene will be the more nucleophilic. This is enough to allow the epoxidation of the trisubstituted alkene in citronellene2 28 while leaving the monosubstituted alkene intact and provide a source of the optically active acid 31 for Nicolaou’s synthesis of rapamycin.3 mCPBA

HClO4 H2O

O 28; (+)-β-citronellene

29

HO2C 31; 75% yield from 28

1. NaIO 4

OH

OH 30

2. CrO 3 H2SO4

280

17 Electrophilic Attack on Alkenes

The synthesis of the aglycone (the bit without the sugar) fortamine 44 of the antibiotic fortimicin A illustrates this sort of chemistry perfectly. Selective epoxidation of cyclohexadiene 32 gives a good yield of monoepoxide 33 if the reagent is buffered. Nucleophiles, such as MeNH2 attack at the allylic centre as the SN2 reaction is accelerated by the alkene to give the anti-amino alcohol 34. The amino group is acylated and the OH group methylated ready for the next step.

MeCO3H

MeNH2

Na2CO3

MeOH 70 ˚C

O 33; 84% yield

32

1. ClCO 2Me MeOH, Na 2CO3

MeO2C

MeHN

2. NaH 3. MeI

OH 34; 99% yield

N

Me OMe 35; 88% yield

We shall discuss the conversion of 35 into 36 later in the chapter. Elimination and epoxidation stereoselectively produces the epoxide 37 on the outside (exo-face) of the folded alkene. Nucleophilic opening with PhSe⫺ gives the trans-diaxial product (chapter 21). Br

H

H

O

1. DBU, 85 ˚C toluene

O

35

H

Me

O

OMe 36; 86% yield

SePh

O

EtOH

N

N

H

Me

OH

O PhSeNa

O

2. (CF 3CO)2O H2O2, 0 ˚C CH2Cl2

N

H

O

Me

OMe 37; 89% yield

H

OMe 38

Oxidation to the selenoxide 39 and thermal elimination (chapter 32) produces a new alkene 40 that forms an epoxide 41 on the outside face again. At this point every carbon atom in the sixmembered ring is functionalised, five oxygen-based and one nitrogen, and this all started with symmetrical cyclohexadiene. Each functional group was selectively introduced from an alkene. H 38

mCPBA i-Pr2NH

OH

O

H

SePh

O O

OH

H (CF3CO)2O H2O2, 0 ˚C

O O

N

CH2Cl2

N H

Me

Me

OMe 39

OH

O O

O N

H

Me

OMe 40; 94% from 37

H

OMe 41; 62% yield

Now the epoxide 41 was opened regioselectively and stereospecifically with azide ion to give the trans di-axial product 42 that was reduced to the amine 43. The synthesis of fortamine 44 was completed by the removal of the protecting carbamate.4

H NaN3, MeOH

41

NH4Cl, 65 ˚C

OH

O

H2

O

Pd/C

N Me

H

N3

H 42

OH OMe

OH

OH

O

HO

NH2

NH2

HCl

O N Me

H

OMe 43; 92% yield

OH

H2O

MeHN

OH

OMe 44; 99% yield

17 Controlling Chemoselectivity

281

Controlling Chemoselectivity In view of the last example, you should not be surprised that the non-conjugated diene 45 reacts with mCPBA at the more highly (tri-) substituted alkene to give the epoxide 33 in good yield.5

mCPBA CH2Cl2

HO

O HO

45

46; 75% yield

Chemoselective reaction at the cis-disubstituted alkene is more difficult. One solution is to deliver the reagent intramolecularly through a favourable 5/6 membered ring. The alcohol 45 is converted into the aldehyde 47. Oxidation of the derived imine with the commercial reagent Oxone® (KHSO5.KHSO4.K2SO4) gives the oxaziridine 48 in yet another chemoselective oxidation. Methylation at nitrogen makes the oxaziridine electrophilic at oxygen and intramolecular epoxidation takes place.

OHC

1. BnNH 2 mol sieves

BnN

2. Oxone® NaHCO3

H

1. MeOTf CH2Cl2

O

47

2. H 2O NaHCO3

48; 47% yield

OHC O 49; 48% yield

It is not necessary for the reagent to be covalently bound to the substrate for intramolecular epoxidation. Hydrogen bonding of peroxy acids or hydroperoxides, often with a co-ordinating metal such as VO(acac)2, delivers the reagent to well situated alkenes.6 This can achieve chemo- and stereoselectivity. Epoxidation of the dienol 50 with mCPBA gives a good yield of the epoxide 51 with only 15% of the bis-epoxide. They tried VO(acac)2 /t-BuOOH first but it gave mainly enone.7 Ar HO

O

HO

1. MsCl, DMAP Et3N, CH 2Cl2

mCPBA

O

H 50

–20 ˚C CH2Cl2

O

HO

H

2. CsOAc, DMAP

O

H 51; 73% yield

3. K 2CO3 NaHCO3, MeOH

O

H 52; 83% yield

O

O O

H

O

O

R

53

The intramolecular delivery of the reagent by some intramolecular mechanism such as the one in the margin is confirmed by the exclusive syn relationship between the OH group and the epoxide in 52. The hydrogen bond between the OH group and the peracid makes the electrophilic oxygen atom of mCPBA more reactive and guides it in to the bottom face of the molecule as

282

17 Electrophilic Attack on Alkenes

drawn. Epoxidation of the other alkene would involve an unfavourable medium-ring transition state. In fact the trans relationship 53 was required so the OH group was inverted. Regiospecific opening of the epoxide was also achieved by intramolecular delivery of the nucleophile. The alcohol 52 was converted into the reactive intermediate 53 with methyl isocyanate. Cyclisation can occur only to the nearer end of the epoxide (5-exo-tet) to give 54. The stereochemistry is already correct for the SN2 inversion on the epoxide 53. The free OH group is turned into a good leaving group by mesylation without isolation of 54. Me O

52

1. NaH, THF MeNCO

O

N

N

O

O

O

Me

Me N

O

O

OH

OMs

2. MsCl Et3N

O

O

H

53

O

H 54

H

55; 67% yield

Finally the remaining alkene is dihydroxylated with catalytic OsO4 and stoichiometric N-methylmorpholine (NMO) as oxidant to give the diol 56 that cyclises to the THF 57 stereospecifically. The aldehyde 58 was used to make (⫺)-dysiherbane. O

O

Me

55

N

OsO4 O NMO H2O acetone

OMs OH O

pyridine

O

Me

N O

Me N

H OH

O

DessMartin

O

H O CHO

OH O

H 56

O

H 57

H

58; 78% yield

Regioselectivity Epoxidation and dihydroxylation have become even more important in recent times as asymmetric versions have been developed and widely used. These are the subject of a special chapter (chapter 27) and do not have intrinsic regioselectivity. In this section we shall discuss the hydration of alkenes and how to reverse the normal regioselectivity of the reaction by hydroboration. These reactions do have regioselectivity and are used in organic synthesis as is a related electrophilic addition, iodolactonisation, that we shall deal with later in this chapter. The hydration of an alkene 8 - literally the addition of water - might occur with either regioselectivity to give the primary 59 or secondary 62 alcohol. If the reaction were a simple acid-catalysed addition, the secondary alcohol 62 would be expected as the secondary cation 60 would be an intermediate and the alcohol 62 is sometimes called the Markovnikov product. In fact the reaction does not usually occur just by treating alkenes with acidic water and special methods must be developed for both regioselectivities.

OH

R 59

?

H

R

R 8

OH2

H2O

60

R

61

OH R

62

17 ‘Markovnikov’ Hydration

283

‘Markovnikov’ Hydration Mercuration-reduction The proton is not a good electrophile for an alkene - it is too hard. Some metal ions are much better soft electrophiles and mercury, as Hg(II), is outstandingly good. Addition of mercury (II) salts, usually the acetate but the oxide and sulfate are also used, to alkenes 8 gives regioselective addition via the cation 63 with the mercury adding to the less substituted end of the alkene. This stage, the formation of 64, is oxymercuration. Reduction in alkaline solution removes the mercury, hydrolyses the ester and releases the secondary alcohol 62. The complete process is Brown’s mercuration-reduction method (Vogel, page 545, calls it oxymercuration-demercuration, take your choice) for the hydration of alkenes.8 The radical mechanism for the removal of mercury is discussed in the workbook. OAc

OAc Hg

R

HgOAc

R OAc

8

AcO

HgOAc

R

63

OH NaBH4 NaOH, H 2O

R

64

62

The reaction is surprisingly tolerant of other functional groups. The tertiary alcohol 66 is formed in 100% yield on a 5 gram scale by this method9 and the electron-withdrawing triazine ring in 67 makes the secondary alkyl cation more stable than the alternative.10 In spite of these virtues, this method has fallen out of favour as it requires stoichiometric toxic mercury. N O

1. Hg(OAc) 2 THF, H 2O

Ph2P

2. NaBH 4 NaOH, H 2O

65

O Ph2P

OH

N

N 1. Hg(OAc) 2 THF, H 2O

N

N

N

OH

2. NaBH 4 NaOH, H 2O

66; 100% yield

67

68; 78% yield

Wacker oxidation Wacker oxidation11 provides a way to add water to an alkene 8 and oxidise the product to a ketone 72 all in the one step using oxygen under palladium (II) catalysis. The key to the difference between these two superficially rather similar sequences lies in the great tendency for palladium to undergo β-elimination 70. Oxypalladation 69 gives an unstable alkyl-palladium σ-complex which decomposes at once to regenerate the double bond.

Cl H2O

Pd Cl

HO

PdCl2

R 8

R 69; oxypalladation

R

OH

HPdCl

H PdCl

70; βelimination

R 71; enol

O R 72; ketone

The palladium is released as HPdCl - still Pd(II)- but this decomposes rapidly to Pd(0) to end the cycle. In practice Pd(0) is reoxidised to Pd(II) with catalytic Cu(II) and the stoichiometric oxidant, O2, oxygen itself, regenerates the Cu(II). The Wacker process is run on a large scale in industry to make simple carbonyl compounds but does find some use in the laboratory.

284

17 Electrophilic Attack on Alkenes

It obviously has no stereoselectivity but the regioselectivity is as expected for oxymetallation: water adds to the more substituted centre. This reaction is part of a useful cyclopentannelation sequence. Allylation of ketones is among the easiest alkylations: subsequent Wacker oxidation and cyclisation creates a new five-membered ring 76. This is the α-acyl cation strategy discussed in chapter 6, though this version of it was not mentioned. O

O

X

O

O2

allylation of enol(ate)

base

73

O

aldol

cat. PdCl 2 cat. CuCl 2

O

74

75

76

An illustration that demonstrates stereoselectivity as well comes in Ikegama’s synthesis of coriolin.12 Allylation of the sodium enolate of cyclopentanone 77 gives one diastereoisomer of the precursor 78 for a Wacker oxidation and cyclisation to give the tricyclic intermediate 79. THPO

THPO

H O

H

1. NaH, DME

1. O 2 PdCl2 CuCl2

2.

2. base

Br

H 78; 94% yield

H 77

O

THPO

H O H 79; 77% yield

Hydroboration: ‘Anti-Markovnikov’ Hydration of Alkenes Mechanism, regio- and stereoselectivity of hydroboration Reversing the Markovnikov regioselectivity calls for hydroboration.13 We discussed hydroboration of alkynes in the last chapter and many of the same principles apply here. The reaction is a syn addition of R2B⫺H to an alkene in which the boron bonds to the less substituted end of the alkene. The same sort of hydroborating agents are used – such as 9-BBN –and the mechanism is similar. The most important interaction is between the full π orbital of the alkene (HOMO) and the empty p-orbital on boron (LUMO) 80, but the reverse interaction between σ(B⫺H) and π*(alkene) also comes into play as the reaction proceeds. The result is syn addition via a transition state 81 with some positive charge on carbon and some negative charge on boron. H

R2BH

R

H (+)

BR2

R 8

BR2 (–)

R 80

BR2

R

H2O2 NaOH

81

OH

R

82

59

These alkyl boranes are not used as precursors for alkyl-lithiums but more usually oxidised to alcohols 59 with alkaline H2O2. This reaction involves nucleophilic attack by HO⫺O⫺ anion on boron 83 followed by alkyl migration from boron to oxygen 85.

R

R

BR2 O OH 83

R B

R 84

O

OH

O R 85

OH

HO

BR2

H2O

R 59

17 Hydroboration: ‘Anti-Markovnikov’ Hydration of Alkenes

285

The stereochemistry of all these steps is important. The hydroboration is a cis addition of H and BR2 displayed clearly by the product 87 from the cyclic alkene 86. The addition of HOO⫺ to boron obviously involves no change in stereochemistry at carbon. The migration step 88 goes with retention of configuration as the filled s orbital of the C⫺B bond is used and the last step involves nucleophilic attack by HO⫺ on boron so there is again no change in stereochemistry at carbon. The overall result is retention at the stereogenic centre during the oxidation to give 89. H H

R'

R2BH

R'

R'

NaOH

B

BR2 86

H

H2O2

O

OH

OH

R R 88

87

R'

89

A simple example is the oxidation of an alkene to a carboxylic acid 93 in Evans’ synthesis of cytovaricin. The alkene had been put in by allylation and the optically active unsaturated alcohol 90 was first protected 91 and then subjected to hydroboration and oxidation. Protection before the second alcohol appeared prevented impossible chemoselectivity problems.14 1. 9-BBN

RO

RO

OH

2. H 2O2 NaOH

RuCl3.3H2O

CO2H

RO

K2S2O8.H2O

90; R = H 91; R = t-BuMe2Si

92; R = t-BuMe2Si

93; R = t-BuMe2Si

If the alkene has diastereotopic faces, the less hindered is normally attacked by the borane. A simple example occurs in Grieco’s recent synthesis of the alkaloid ibogamine. The alkene 94 is a trisubstituted cyclohexene with two stereogenic centres in the ring. Hydroboration occurred with the expected regioselectivity on the top face of the alkene. This is the same face as the CO2Me group, but the opposite face to the nearer and larger indole substituent. The alcohol 95 was the only product formed and was isolated in 68% yield. Simple reactions led to ibogamine.15 NHCbz

NHCbz

HO H H

1. B 2H6, THF

H

2. H 2O2, NaOH

N H 94

N H

CO2Me

CO2Me

95

A more complex example with stereochemistry comes from Schreiber’s asteltoxin synthesis.16 Hydroboration of the bicyclic acetal 96 with borane itself occurs on the underside of the folded bicyclic molecule to give 97. But this is not the end of the story. The product of the hydroboration is the THF 98. H

H

R2B Ph BH3.THF

Me

O

O Me 96

hydroboration

H

H Me

H

O

O Me 97

R2B Ph

H

Ph

H Me

O O Me 98

H

BR2

286

17 Electrophilic Attack on Alkenes

Borane was used to achieve a cunning reduction of the acetal in the same reaction mixture by acetal opening 99 followed by intramolecular hydride transfer 100 so that a single diastereoisomer of the borane/borate 98 was formed. Only now is the borane oxidised with alkaline hydrogen peroxide and the new alcohol 101 is formed as expected with retention of configuration. Three new stereogenic centres (the old acetal centre could not be counted as a permanent stereogenic centre) are formed in this sequence. We shall return to hydroboration later as a way of making carbon-carbon bonds.

R2B 97

BH3.THF borane as Lewis acid

H

H

H Me

R2B Ph

Me 99

Ph

H Me

O

O

H

O O

BH2

HO

H Me 100

98

H2O2 NaOH

B

H

Ph

H Me

H

OH O

H

Me 101

H

Alternative Approaches to the Synthesis of Alcohols from Alkenes Regioselective reduction of epoxides In view of the difficulties inherent in all these methods, others have been developed from epoxides and by intramolecular delivery of reagents. These methods are not so general as those we have described so far and we shall quote just two approaches. The unsymmetrical diaryl epoxide 103, easily made by the sulfonium ylid method [details in workbook] reacts with nucleophiles as the benzylic centre rather than the centre next to the pyridine ring. Reduction with LiAlH4 gives just the alcohol 104 while reaction with MgBr2 gives just one bromohydrin 102 both in quantitative yield. It is possible that metals such as Mg, Al, or Li coordinate the pyridine nitrogen and epoxide oxygen atoms.17 Br

MgBr2.2Et2O

O

Et2O, 0 ˚C

N

LiAlH4

N

N

OH

OH 103

102; 100% yield

104; 100% yield

Intramolecular hydrosilylation A hydrosilylation approach with the silicon being delivered intramolecularly from an OH group in 106 is also regioselective.18 Catalysis by H2PtCl6 efficiently gave the heterocycle 107 and an oxidation of the C⫺Si bond (cf. the similar reaction on boranes above) gave one regioisomer of the diol 108. Me OH

H

Me Si

(HMe2Si)2NH

O

0.1 mol % H2PtCl6.6H2O

Me Me Si O

106

OH

H2O2

60 ˚C. 10 min

105

OH

Na2CO3

107

108; 69% from 105

17 Stereoselectivity

287

There is good stereoselectivity with cyclic chiral alcohols such as 109 (⬎100:1) but the more flexible achiral compounds such as 111 are not so good. In both cases the OH group delivers the Si⫺H from the same face but 111 can adopt a different conformation.19 OH

OH 1. (HMe 2Si)2NH

1. (HMe 2Si)2NH

R

2. 0.1 mol % H2PtCl6.6H2O 3. H 2O2, NaHCO 3

109

R

2. 0.1 mol % H2PtCl6.6H2O

OH OH 110; 73% yield 111; R = n-C 5H11

3. H 2O2, NaHCO 3

OH

OH

112; 72% yield

Stereoselectivity Electrophilic attack on alkenes is normally stereospecifically syn (as in hydroboration) or anti (as in bromination). In this section we consider stereoselective aspects of these reactions. In the simplest cases, the two faces of the alkene are diastereotopic because of some stereogenic centre elsewhere in the molecule. Reaction will then occur on the less hindered face opposite the substituent already present. In favourable cases, where the substituent is large and close to the alkene, the selectivity may be high.

mCPBA

OsO4 etc

OH

O 113

OH 115

114

Corey took advantage of the availability of specific syn and anti additions to make the two possible diastereoisomers of a bromohydrin by different methods. In each case the initial attack occurs anti to the t-butyl group and the second reagent, the nucleophile, adds from the other face. The order of events decides which product is formed.20 O

t-Bu

mCPBA

NBS

t-Bu

Br

t-Bu

H2O

117

118

119

HBr

H2O

OH t-Bu

OH t-Bu

Br 116

Br 120

For a more complex example, we turn to Whitesell’s synthesis of iridomyrmecin21 121, a compound from the Argentine ant Iridomyrmex humilis. This compound is a bicyclic lactone with four stereogenic centres. Opening the lactone reveals the basic skeleton 122. This skeleton has a 1,5-diO relationship and one might consider Michael additions at this point, but Whitesell noticed something more fundamental. If the two functionalised atoms were reconnected, the skeleton 123 has some symmetry. The substitution pattern of the two methyl groups in the two five-membered rings is the same, though the stereochemistry is different and an oxidisable double bond is required in one ring but not the other. It might be possible to make the compound from a symmetrical cyclo-octadiene 124.

288

17 Electrophilic Attack on Alkenes

H

H

O

C–O

O

H

HO2C

lactone

?

reconnect

HO H

H 121; iridomyrmecin

H 123

122

124

The synthesis of such eight-membered rings by the nickel-catalysed dimerisation of butadienes, is used here with 2-methylbutadiene. The regio-selectivity and stereo-specificity of the hydroboration of symmetrical 124 are as expected for a trisubstituted alkene. Conversion of the alcohol 125 into a good leaving group 126 made it reactive enough to carry out electrophilic addition on the other alkene. In aqueous solution, the final nucleophile was water and the product a mixture of diastereoisomers of the tertiary alcohol 127.

HO Ni(0)

1. 9-BBN

MsCl

Na2CO3

2. H 2O2 NaOH

Et3N

H2O dioxane

124

H HO 125

H MsO 126

H

H 127

For the transannular reaction to occur, the molecule must fold inwards. In fact medium rings are normally folded like this anyway. The regioselectivity of the electrophilic cyclisation of the alkene gives as normal the more substituted cation. Inversion of course occurs at the mesylate centre and the ring junction has to be cis as 5/5 ring fusion would be very strained it were trans. The third stereogenic centre does not matter. Elimination of the tertiary alcohol 127 (TsCl, pyridine) preferentially gives the alkene 128 away from the ring junction. This is mainly to avoid flattening out the stable folded fused five-membered rings, but the tri-substituted alkene 128 may also be preferred to the alternative tetra-substituted alkene 130. Now a second hydroboration introduces the required oxygen functionality in the left hand ring 129. The regiospecificity is as expected and the borane (9-BBN again) adds stereoselectively to the less hindered lower face of the molecule. This is because of the folded rings. They are folded upwards (the ring junction hydrogens are down) and so the lower surface is outside the fold. It is very difficult to get inside these folded molecules, particularly with a large reagent like 9-BBN. HO

H

H TsCl

1. 9-BBN

pyridine

2. H 2O2 NaOH

H

H

127

128

H

H HO

H 129

H 130: not formed in elimination

The left hand ring must now be cleaved and so the alcohol is converted to a ketone 131. The obvious Baeyer-Villiger reaction will give the wrong cleavage product 133 as the more substituted group migrates from carbon to oxygen in these rearrangements. Stereospecificity is fine - the rearrangement goes with retention- but the regioselectivity is wrong. The problem is solved by ozonolysis of an enol derivative. Kinetic lithium enolate formation (chapter 3) and trapping with

17 Halolactonisation

289

Me3SiCl gives a stable silyl enol ether 132. Silyl enol ethers are of course alkenes too and can be oxidised by ozone. Reductive work-up is necessary not only to prevent the H2O2 by-product from oxidising the aldehyde but also to reduce it to the alcohol in situ. Acid-catalysed lactonisation of 122 finally gives iridomyrmecin 121. H

H

CrO3 H2SO4

H

O

129

H

H 1. O 3

1. LDA

acetone (Jones)

2. Me 3SiCl

Me3SiO

H

122

121

2. NaBH 4

H

O

H

131

O

H 133; Baeyer-Villiger product from 131

132

This synthesis involves four electrophilic attacks on alkenes - two hydroborations, one alkylation by a mesylate, and one ozonolysis. This is possible - even though the starting material had only two alkenes - because one of the alkenes is recreated three times. You will have noticed that each time it reappears, it moves round the structure so that electrophilic attack followed by double bond (re-)creation enables the formation of several new bonds as well as the introduction of new functionality. Now that we have introduced all three kinds of selectivity in the context of electrophilic attack on alkenes, we shall look at some ways to control selectivity by special devices. In the examples which follow it is more convenient to treat the three selectivities together as they are often interdependent.

Selectivity by Intramolecular Interactions Halolactonisation We have seen several instances of regio- and stereochemical control by intramolecular delivery of a reagent. This approach reaches its apogee when the reagent is tethered with a covalent bond to the alkene. The most important of these reactions is iodolactonisation or, more generally, halolactonisation.22 Reaction of an unsaturated acid 134 with iodine in aqueous NaHCO3, to ensure that the acid is present as its anion 135, gives an iodolactone 137 with virtually complete control. Iodine attacks the double bond and the regioselectivity is normal - the nucleophile, the carboxylate anion, attacks the more highly substituted end of the iodonium ion 136. O I2

CO2H 134

O

I2

CO2

NaHCO3 H2O/MeCN

O

O

135

I

I 136

137

However, if the alkene is symmetrically substituted, the nucleophile attacks the nearer end of the iodonium ion as it is easier for it to achieve the 180⬚ angle necessary for the SN2 reaction. This automatically leads to stereoselectivity as well. Perhaps the most famous examples come from Diels-Alder adducts such as 139. The carboxylate anion can reach only the same face of the six-membered ring to which it is already attached (a trans ring junction is impossible with this bridged ring). This means that only the anti iodonium ion 140 can cyclise and the syn iodonium ion 138 must revert to starting materials. The iodolactone 141 has a 1,3-diaxial bridge 141a.

290

17 Electrophilic Attack on Alkenes

CO2H

CO2 I

I

O

O

O O

I

O I

138

139

140

141

141a

O

This result has further implications. An elimination using the bicyclic amidine DBU can occur only on one side of the iodine atom 142 as this is the only way to get H and I anti-peri-planar for the E2 reaction. The alternative H atom is equatorial. Electrophilic attack on this new alkene 143 must occur from the side opposite to the lactone bridge which, because it is diaxial, effectively blocks the alkene on one side 145. Epoxidation 144 again provides a good illustration. I

O

H

H

easy attack from above

O RCO3H

H

×

E2

O

B

O

O

DBU

O

O

O

143

142

no attack from below 144

O

145

If the lactone produced might be either four- or five-membered, then the five-membered ring is usually preferred as this reaction is under thermodynamic control.23 A series of aryl lactones was prepared in this way using KI and an oxidising agent as the source of iodine. Attack on either enantiotopic face of the alkene E-146 gives an iodonium ion which cyclises by the rather awkward mechanism 147 needed to make the five membered ring 148. Note that the stereochemistry of the alkene survives the one inversion. Preparation of the starting materials involves chemistry we met in the last chapter so it is reviewed in the workbook.

Ar

CO2H

KI, NaHCO 3, H 2O

Ar

I

I

I =

Na2S2O8, H 2O

Ar O 147

E-146

O

O H 148a

O

Ar

O 148b

O

If the choice is between a seven- and an eight-membered lactone, the seven-membered ring 151 is preferred to the medium sized ring as in a recent series of experiments24 using the collidine-Br⫹ reagent 150. O CO2H

O + N

149

Br

Br 150

PF6

151; 95% yield

We can now return to the conversion of 35 to 36 mentioned earlier in the chapter. This is a bromolactonisation using a carbamate ester as the free acid would be unstable. The reagent is nearly the same as 150; only the perchlorate salt is used instead in aqueous base. The ester group attacks the bromonium ion 152 to give the oxonium ion 153 that is hydrolysed under the conditions.4

17 Halolactonisation

291

H2O

Br

N Me

OMe

Br

H O

MeO

ClO4 H2O Na2CO3

MeO

N

N Me

Me

OMe

152

35

Br

O

O

O 150

MeO2C

H

N

etc.

H

H

Me

OMe

153

OMe

36

The synthesis of vernolepin In his synthesis of the anti-tumour compound vernolepin, Danishefsky25 used this very style of chemistry to great effect. Vernolepin 154 has just one carbocyclic ring and two lactone rings, but the synthesis was planned around an intermediate with two carbocyclic rings 155 so that all the stereochemistry could be controlled. You can see that the three stereogenic centres in the intermediate are correct, that there is a double bond in ring A 156 that might be cleaved oxidatively and another in ring B where the third ring is to be added. O

OH

O

H

cleave ring A here

O

O

O

H A

O

H

O

O

O 154; vernolepin

O

H 155; bicyclic intermediate

B H

add ring C here

156

We shall discuss the synthesis of the intermediate 155. The combination of a double bond and a lactone in ring B looks as though it could come from an iodolactonisation and elimination. We can draw the disconnections for these steps to reveal a possible starting material 158. O

O

O

O

O

FGI

O

H 155

OH

iodolactonisation

O

I

H 157

O

H 158

This unsaturated acid was made by a Diels-Alder reaction, but not the one you might have expected. Danishefsky invented a special diene ‘Danishefsky’s diene’ 159 for the introduction of enone functionality in the Diels-Alder reaction and here you see an application. The Diels-Alder adduct 161 is a silyl enol ether with a leaving group in the β-position and it hydrolyses easily to the enone 162. OMe

OMe CO2Me

CO2Me

Me3SiO

Me3SiO 159

160

H 161

CO2Me H

HO

H2O

H2O

O

H 162; about 60% yield

158 100% yield

292

17 Electrophilic Attack on Alkenes

Note the chemoselectivity between the two alkenes in the cyclohexadiene 160. This is the expected selectivity as Danishefsky’s diene, with two electron-donating substituents is obviously electron rich and prefers to react with the electron-deficient alkene. The next two steps are the planned iodolactonisation and elimination with the usual reagents. There is chemoselectivity here too as iodine will attack the more nucleophilic double bond in 158, regioselectivity in that the carboxylate anion attacks the nearer end of the alkene 163, and stereoselectivity in that the anion can reach only the bottom face of the alkene. O

158

O

I2, KI

157 88% yield

I

NaHCO3 H2O

O

O

O DBU 24 hours room temperature

H 163

O

O

O

O

O

H 164

H 155; 87% yield

Now the time has come to react one of the alkenes and not the other. The reaction chosen was epoxidation as we could expect only the more nucleophilic non-conjugated alkene to be attacked by mCPBA. Direct epoxidation of the lactone 155 gave only the epoxide 164 with the correct chemoselectivity but the wrong stereoselectivity. The lactone bridge directs the peracid to the top face of the alkene. But if the lactone is first hydrolysed, there is an OH (and CO2H) group on the lower face of the molecule to deliver the peracid to the same face. This is a clear demonstration of the positive nature of this type of stereoselectivity as the lower face of the molecule is the more crowded. The last step closes the lactone again. O

155

OH

O

OH

OH

1. NaOH H2O, THF 2. HCl, H 2O

O

O

OH mCPBA

Ac2O

dioxane

NaOAc

O

H 165; 99% yield

H 166

O

O

O

O H 167; 76% from 155

Now that the more nucleophilic alkene is no more, the less nucleophilic one can be oxidised and ring A converted into a lactone 169. The rest of this synthesis is equally interesting but is outside the scope of this chapter. It has to start with some chemoselective reaction on the two lactones in 169 as described in the workbook. O OHC 167

MeO2C O H 168; 86% yield

O

O

1. LiAlH(OBu t)3

1. OsO 4 2. Pb(OAc) 4 MeOH benzene

O

O

2. H + resin

154 vernolepin

O

O H 169; 100% yield

Halolactonisation in the synthesis of erythronolides A remarkably clever combination of hydroboration and successive halolactonisations gave a vital intermediate in Corey’s synthesis26 of erythronolide B. Compound 170 was required with its five

17 Halolactonisation

293

stereogenic centres around a six-membered ring. Corey realised that both sides of the ring could be developed by halolactonisations with the same carboxylic acid, used twice, once on each alkene. O

O

O

O I

O

FGA

O

O

FGI

OH

iodolactonisation

O

OH

iodo lactonisation

O

O

170

O

C–I

O

172

171

173

The starting material 173 is drawn without stereochemistry because it is not chiral! The symmetry in this dienone means that it does not matter which alkene reacts first as they are the same. You may also have noticed that the stereochemistry of the methyl groups on the double bonds comes out wrong in both reactions, but, as we shall see, this does not matter. The starting material can be made from a simple phenol 174 by ipso-allylation and then hydroboration. The hydroboration is of course regioselective and attacks the electron-rich alkene rather than the enone system. The alcohol produced was further oxidised by Jones reagent without isolation. Now we are ready for the first bromolactonisation. OH

O NaOMe Br

O

O

1. BH 3.THF

Br2 KBr

2. H 2O2, NaOH

H2O

Br O

3. CrO 3, H 2SO4 (Jones)

174

HO2C

175

O 176; 96% yield

173

This reaction goes in a wonderful yield in spite of the rather unreactive enone. Maybe this is why the more aggressive bromine was preferred to iodine. The acid can be delivered only to the bottom face of the ring as it is tethered to that face and the bromine must therefore add to the top face. The next stage was to hydrolyse the lactone in base. The oxyanion released closed spontaneously onto the neighbouring bromide 177 to form an epoxide 178 with inversion. With that inversion, the stereochemistry is correct. This sequence of halolactonisation, hydrolysis and closure to an epoxide has been used in many syntheses to get one particular diastereoisomer of an epoxide. Now that the molecule is unsymmetrical and while it has a free carboxylic acid, it was resolved so all future diagrams show a single enantiomer. The next step was bromolactonisation on the other side to give 170. O

O

176

KOH H2O THF

O Br

Br 171 98% yield

O O2C 177

Br2 KBr H2O

Bu3SnH

O O O 178; 91% yield

AIBN

O O O 170; 93% yield

This also worked very well with all the usual selectivities and the bromine atom was removed with Bu3SnH to give mostly the correct diastereoisomer as the hydride approaches the enolate

294

17 Electrophilic Attack on Alkenes

radical intermediate to give mostly an equatorial methyl group. We shall stop our analysis here but the full story is told in Classics in Total Synthesis, page 173 ff.

Sulfenyl- and Selenenyl-Lactonisation Halolactonisation works well because bromine and iodine form three-membered ring intermediates when they attack an alkene. Sulfur (II) and selenium (II) electrophiles form even better defined intermediates of this kind and similar cyclisation reactions occur with impressive control over selectivity.27 A simple example would be the formation of the bicyclic lactone 181. As the carboxylic acid is tethered to the five-membered ring, it can cyclise only to the intermediate with S(e) on the other face 180. The mechanism is similar to that of iodolactonisation and the stereochemical points are the same. The products are useful for the generation of radicals as Bu3SnH removes a PhSe group and leaves a radical behind while oxidation and elimination leaves a new alkene (chapter 33). Ph

PhS(e)

H

S(e)

O O

PhS(e)Cl

CO2H

O OH

179

H 181

180

We can use sulfenyl-lactonisation to illustrate the stereospecificity of the process when the alkene differs in geometry. Rokach28 has shown that the geometrical isomers of the unsaturated acid Eand Z-182 give the correct diastereoisomers of the lactones from anti addition of Cl and R’S. R'S

R'S R

R'SCl

R

CO2H

Et3N

R

R O

R'SCl

CO2H

Et3N

O

E-182

Z-182 O anti-183

O syn-183

Returning to a reaction we met right at the start of this chapter will illustrate that the regioand stereochemistry of many different electrophilic reactions with alkenes can be controlled by intramolecular nucleophiles. The mercuration of the cis alkene Z-184 leads to a 6:1 ratio of diastereoisomers of a cyclic ether 185 by a related trapping of the intermediate by the internal OH group. CONR2

1. Hg(OAc) 2 MeCN

OR OH Z-184

2. NaCl H2O

OR R2NOC H

O

H

HgCl

185

This is impressive because attack on the cis alkene 186 has to occur on the bottom face to lead to the product 185 and this line is hindered by the allylic OR group. Notice that the regiochemistry is

17 The Prins Reaction

295

determined by the proximity of the OH group to one end of the alkene and that the stereochemistry of the alkene survives (with an inversion) in the product. The HgCl group is removed by reduction and the product converted into a building block 187 for the pammamycins, naturally occurring macrolides.29 CONR2 OH OR

185

MeO2C

HO

H

O

H

Hg AcO

186

187

We shall end this chapter with three ways to make carbon–carbon bonds by electrophilic attack on alkenes. Though useful, they are not as important as the functionalisation reactions we have discussed so far, but they are each special in their own way. The first two, the Prins reaction and hydroboration revisited, use one-carbon electrophiles.

The Prins Reaction The original Prins reaction This is a “forgotten” reaction of remarkable scope which can sometimes give useful molecules very difficult to make by any other means.30 The electrophile is formaldehyde (or occasionally another reactive aldehyde) in acid solution and the first step is straightforward - normal “Markovnikov” addition 188 gives the more stable cation 189, the key intermediate in a Prins reaction. H

H

H

HO

O

R

HO

H 188

H

products

R 189

A number of things can happen now depending on the structure of the alkene and counterion of the acid. Direct cyclisation is not usually one of them as that would give an oxetane, a fourmembered cyclic ether. Formaldehyde is often used as the 37% aqueous solution “formalin” (used to preserve biological specimens) and then there is then plenty of water around. One good example is the reaction of styrene with two molecules of formaldehyde to give an acetal 190 in 100% yield with a solid acidic resin as catalyst.31 Resuming the mechanism where we left off, presumably water adds to the cation to give a diol 191 which cyclises with another molecule of formaldehyde. There are other possibilities as this reaction gives a stable six-membered ring with an equatorial phenyl group and is probably under thermodynamic control. O

CH2O

Ph styrene

H2O

O 189; R = Ph

Ph

190

OH

H2O

Ph

OH

CH2O

190

191

The true product is the 1,3-diol 191, of which the product actually formed 190 is the formaldehyde acetal, and the disconnection should be contrasted with the aldol route 192 to the same product as a different carbon-carbon bond is formed.

296

17 Electrophilic Attack on Alkenes

O Ph

OH

O

OH

Aldol

+ OMe

OH Prins

FGI

CO2Me

Ph

Ph

192

Ph

+

CH2O H2O

191a

The formation of tetrahydropyrans by the Prins reaction Aliphatic alkenes also undergo the Prins reaction and a well-conducted example is with propene, formaldehyde and HCl. The product is 4-chloro-tetrahydropyran in good yield.32 This compound presumably arises from the cyclisation of another diol 194 and that clearly comes from one molecule of propene and two of formaldehyde. What is different from the last example is that the two molecules of formaldehyde must have been added to opposite ends of the original propene. Cl

Cl

CH2O

Prins

FGI

193 HCl

O 193

HO

CH2O

OH

CH2O

194

We can now reconstruct the reaction mechanism. The first cation 189; R⫽Me is formed as before, but in the absence of water elimination must reform an alkene 195 on the other side before a second molecule of formaldehyde is added. The HCl used provides a chloride ion to intercept the second cation 196.

CH2O

CH2O

–H

Cl

HCl

HO 189; R = Me

HO

HO

195

194

193

OH

196

This product (4-chlorotetrahydropyran, 4-Cl-THP) 193 may not look very useful but the chlorine atom can be converted into a nucleophilic Grignard reagent or used in a Friedel-Crafts reaction to give 197 while the ether ring may be cleaved to give a doubly electrophilic unit 198. One useful application is the synthesis of 4-phenyl piperidine 199, a structural unit found in medicinal compounds. Cl

Ph benzene

Ph

Ph NH3

48% HBr

AlCl3

i-PrOH

O

O

193

197; 91% yield

Br

Br

198; 64% yield

N H 199; 87% yield

The disconnections corresponding to this chemistry are remarkable as the molecule is built up quickly from very simple starting materials. In this application, 193 acts as a pentyl 1,3,5-trication synthon! A more modern version33 of these reactions uses a Lewis acid (TiX4) and a pre-formed acetal 200 of a homoallylic alcohol to make the same products. Asymmetric reduction (chapter 26) of ethyl acetoacetate gives the starting hydroxyester as either enantiomer. Yields are much higher and control is exerted in making the protected homoallylic alcohol 200.

17 The Prins Reaction

297

O OH

O

O

OMe

TiBr4

OTBDMS

CH2Cl2

CO2Et

OTBDMS

from ethyl acetoacetate by asymmetric reduction

Br 87% yield

200

The oxo-ene mechanism Returning to the mechanism of 4-Cl-THP 193 formation. When the cation 189 loses a proton to form an alkene, why should it lose that proton to make a terminal alkene? Surely the proton from the other side would be lost preferentially to give the more stable alkene 201? One solution to this dilemma is to make the loss of the proton and the addition of formaldehyde concerted 202. ?

? CH2O

–H

CH2O

H HO 201

H

HO 189; R = Me

O 202

propene

CH2

HO 195

The first step 202 is now a pericyclic reaction. It looks like a cycloaddition, though it involves a hydrogen transfer as well. It is in fact a carbonyl (or ‘oxo-’) “ene” reaction. It is like a Diels-Alder cycloaddition in which a C–H bond has replaced one of the double bonds in the diene and a C⫽O group is the dienophile. Many Prins reactions are probably carbonyl ene reactions. In his excellent review30 in Comprehensive Organic Synthesis, B. J. Snider says “The (carbonyl) ene reaction and the Prins reaction are not mechanistically distinct”. Though this step is pericyclic, it is very polar and the transition state 203 no doubt contains partial charges. It is therefore stabilised and the reaction accelerated by protic acids 205 and Lewis acids 207.

H O

CH2

H

202

195 O (–) 203

(+)

(+)

(+) H O

H

H (+) O

CH2

H 204

O

CH2

H (+) O

AlR2 206

H 205

AlR2 207

Snider has found dialkyl aluminium halides to be excellent catalysts for this reaction and they can be used to stop the reaction after the first step. This is a real advance in the Prins reaction that otherwise tends to give a mixture of products by multiple addition of the aldehyde and intervention by various nucleophiles. A simple example is the addition of formaldehyde to the terpene limonene 208 catalysed by BF3. A single monoadduct 210 is formed in good yield.34 This must be a Lewis acid catalysed carbonyl ene reaction on the external double bond 209. Notice the excellent chemoselectivity in that the internal alkene is not attacked and the excellent regioselectivity in that hydrogen atoms at four other sites might have taken part in the reaction, but do not. H

H

O

BF3

OH

CH2

CH2O, BF 3 Ac2O, CH 2Cl2

208

209

210

298

17 Electrophilic Attack on Alkenes

Stereoselectivity in the Prins reaction Snider’s synthesis of the aryl lignan skeleton is an example with stereochemistry.35 Either geometrical isomer (E or Z) of the starting material (1,4-diphenylbut-2-ene) was combined with formaldehyde and a mixed catalyst of MeAlCl2 and Me2AlCl. The all anti cyclised product 214 was formed in 40-50% yield. The first step must be a carbonyl ene reaction 212. It doesn’t matter which way we write this as the molecule is symmetrical and we know that the stereochemistry is irrelevant. CH2O Me2AlCl MeAlCl2

Ph

Ph

CH2

AlR3

Ph

OH

OH

OH

H

CH2Cl2

211

O

Ph

Ph 212

213

Ph 214

Ph

The next step is a simple electrophilic attack by another molecule of formaldehyde on the alkene - in other words a simple Prins reaction 215 - showing the regioselectivity we expect to produce the secondary benzylic cation 216. The second molecule of formaldehyde has added onto the opposite side from the first. The resulting cation is perfectly placed for an intramolecular Friedel-Crafts alkylation 216 of the benzene ring. This is again a stereoselective reaction giving the more stable anti diastereoisomer 214. This sequence involves three successive C⫺C bondforming reactions and the stereochemistry is simply controlled by the preference for the more stable anti product. Ph

OH

Ph

CH2

O

AlR3

OH

OH

OH

OH H

Ph 216

215

214

Ph 217

Double Prins reactions: use of Sn and Si to control cation formation One problem with the original Prins reaction was uncertainty in the fate of the cation 189. Recent advances have used the chemistry of allyl tins and silanes explored in chapter 12. For example addition of aldehydes to reagent 218 with a chiral catalyst formed from BINOL (chapter 26) and Ti(Oi-Pr) 4 gave the allyl silane 219. This in turn reacted with a second aldehyde using an achiral Lewis acid to give the THPs 220. Yields are very high, enantiomeric excess almost perfect (90-96%) and various aromatic, enolisable aliphatic and functionalised aldehydes can be used in both steps. Control expressed in the isolation of the stable intermediate 219 means that the two aldehydes need not be the same.36

OH Bu3Sn

SiMe3 218

R1CHO BINOL-Ti catalyst

R1

SiMe3

R2CHO Me3SiOTf

219; 74-96% yield

R2 O R1 220; 95-98% yield

The first step 221, whether you call it a Prins reaction or not, gives a cation 222 stabilised by both silicon and tin (chapter 12). Nucleophilic attack on tin is preferred (tin is lower down the

17 Carbonylation of Alkyl Boranes

299

periodic table) and that gives the product. The presence of both silicon and tin stabilise the cation while the tin decides its fate. There is no need to invoke an oxo-ene mechanism. The absolute stereochemistry is decided by the chiral catalyst: you are not expected to see how. L4Ti

R1CHO

SnBu3

O

218 BINOL-Ti catalyst

L4Ti

SiMe3

R1

SnBu3

O

X 219

SiMe3

R1

221

222

The second Prins reaction goes through the oxonium ion 223 to give the final product. Again the nucleophile is an allyl silane 223 and the second intermediate is a cation stabilised by the silicon β-effect. The relative stereochemistry is decided in the second reaction and 220 has the favoured diequatorial conformation. Me3Si

Me3Si X

R2CHO

220

219 Me3SiOTf R1

O

R2

R1

223

O

R2

224

Hydroboration as a Way to Make Carbon-Carbon Bonds Carbonylation of Alkyl Boranes Earlier in this chapter we discussed the oxidation of boranes with hydrogen peroxide. When borane itself was used, the reaction was quite unambiguous. The trialkylborane 225 is the first product and when it is oxidised any of the three identical alkyl groups may migrate first to oxygen. BH3

R

B

R

8

3 225

H2O2

OH

R

NaOH

+ B(OH)3 59

Eventually all three will migrate and three molecules of alcohol will be formed. Borane is supplied as its THF, R3N, or Me2S complex or generated in situ from BF3 and NaBH4. But substituted boranes must be prepared. We met 9-BBN, the most popular dialkyl borane, in chapter 16 and used it earlier in this chapter. The most popular monoalkyl borane is “thexyl” borane whose pet name simply means t-hexyl borane and whose t-hexyl group is often represented as a large letter H. These are obviously bulky boranes and give sterically enhanced regioselectivity in hydroboration of mono substituted alkenes 8. They are both used in the synthesis of alcohols 59. These reactions will be successful providing that the newly introduced alkyl group migrates to oxygen in preference to the t-alkyl group in 228 here or the cage structure in 9-BBN adducts. R B2H6

226

BH2 227

BH2 227

8

H2O2

B

NaOH

228

R

59

300

17 Electrophilic Attack on Alkenes

The rule most people know is that “the group best able to support a positive charge” migrates best. This rule applies to cationic rearrangements like the Baeyer-Villiger where the transition state for the migration has a positive charge. In these boron “ate” complex rearrangements, the transition state has a negative charge and the rule is reversed.37 It might be an exaggeration to say that the group “best able to support a negative charge” migrates as alkyl groups destabilise negative charges, but the migration order is primary⬎secondary⬎tertiary. The group which destabilises the negative charge least migrates best. In the case of 9-BBN, other secondary alkyl groups migrate better than the bicyclononane group because bridgehead atoms migrate worse than ordinary groups. If a bridgehead atom migrates, the whole cage must distort.

Carbon monoxide Armed with this essential information, we can look at C⫺C bond-forming processes. A famous example is the formation of ketones with carbon monoxide.38 The first step is addition of CO 229 and then boron to carbon migration 230 of the primary alkyl group. For the transfer of the second alkyl group an “ate” complex 231 must be formed again, and a nucleophile, usually based on oxygen, is added to the very unstable ketoborane 230 and a second alkyl group is driven across 231 to give a more stable hydroxyborane 232, the product at this stage. R

R t-Hex

t-HexB O

t-Hex

B

O

C

R

O R

R

O

230 t-Hex

B

R

B 2

232

B

R

R

229 t-Hex

R

231 R

O

O

B

Hex-t

t-Hex 2

X

O

R

232a

R

B

231a

The final step is the usual oxidation with alkaline hydrogen peroxide. Both groups on the boron 232 are now tertiary so an excess of oxidant must be used to drive both across to oxygen. The products are t-hexanol and the hydrate of the ketone 234. This doesn’t affect the ketone synthesis, only the by-product.

t-Hex

B

NaOH

R

R 233

OH

H2O2

O

R

OH

O R

R

R 234

If all three groups on borane can migrate, as in normal trialkyl borane, tertiary alcohols are the products. Three migrations from boron to carbon go via the unstable ketone 236 and epoxide 238 to the relatively stable t-alkyl boron derivative 239 that is oxidised to the t-alcohol 240. A t-alkyl group must migrate here as there is no alternative.

17 Carbonylation of Alkyl Boranes

CO

225

B

R

O

3

first

B 2

R migration

R

R

HO

NaOH

R

R

R

R H2O2

migration

238

O

237

OB

third

B

R

R

O second

B 2

O 236

235

migration R

301

R R

239

R

240

A dramatic example is the synthesis of the alcohol 244 by the literal insertion of a CO unit into the triene 241. Heating 242 at 200 ⬚C for six hours after hydroboration equilibrates positional and stereochemical isomers and shows how stable the trialkyl borane is. Glycol is added as the nucleophile so that the stable heterocycle 243 equivalent of 239 can be purified as a crystalline compound. All three B to C migrations go with retention of configuration.38

O BH3.NEt3

H

H

H

CO

O B

H

B

OH

H

H

NaOH HO

OH

H 242

241

H2O2

H 243

H 244; 72% yield

Ketone synthesis using cyanide ion Not everyone wants to use carbon monoxide and a more convenient, though equally deadly, onecarbon reagent is cyanide ion. This does not require the high pressures often needed to make CO react.39 Cyanide forms stable “ate” complexes 247 with boranes and we shall use two different alkenes to illustrate this possibility. Thexyl borane 227 is essential for reaction with two different alkenes and it is better to use the more hindered alkene first. These precautions lead to a more selective reaction with the first alkene. t-Hex B 227

t-Hex H

B R

t-Hex R

B NaCN

CN

R

R = n-hexyl

245

246

247

The “ate” complex 247 has an obvious resemblance to the CO complex 230 but the nitrogen atom is less electron-withdrawing than the oxygen atom as it lacks the positive charge so the cyanide complex is less prone to rearrangement. Acylation with trifluoroacetic anhydride leads to B to C migration of the primary alkyl group 248. This new borane 249 looks stable but the amide group can cyclise onto the boron to regenerate an “ate” complex 250 and initiate migration of the secondary alkyl group. Oxidation of this quite stable heterocycle 251 causes the usual migrations and replaces all B⫺C bonds by B⫺O bonds. The final organic product is a ketone 252. Notice that the first formed borane had primary, secondary and tertiary alkyl groups on it so this sequence illustrates well the order of rearrangements.

302

17 Electrophilic Attack on Alkenes

t-Hex

O t-Hex

B

R

TFAA

247

–78 ˚C

N

B

first

CF3

t-Hex

O

CF3

B

N

N

migration

CF3

O 248

R

R

249 O

t-Hex B O

second

250

H2O2

R

NaOH

migration

N

R

CF3

252; R = n-hexyl 83% yield

251

Reactions with α-halo-carbonyl compounds Any nucleophile that contains a leaving group on its α-atom could in principle be used to initiate rearrangement of alkyl boranes. A simple way to make C⫺C bonds is to use the enolate of an α-halo-carbonyl compound.40 The favoured reagent is 9-BBN (R2BH) and primary or secondary groups can be migrated in preference to that cage structure. The best base for creating the enolate is the very hindered 2,6-di-t-butylphenoxide and B to C migration occurs on the ‘ate’ complex with displacement of bromide 254. The 9-BBN is removed from the product simply with ethanol. R BR2

O Br

R2BH 9-BBN

R B

EtOH

ArO

253

BR2 Br

254

O 256; 73% yield

O

O

255

Polyene Cyclisations In these remarkable reactions, a series of electrophilic attacks on a bank of alkenes e.g. 257 is initiated by an electrophile. Each alkene adds to its neighbour until the bank is exhausted and essentially complete chemo-, regio- and stereo-selectivity is achieved. These reactions were discovered by W. S. Johnson41 and the idea came from nature where steroid synthesis is accomplished by very similar reactions. The main problem is terminating the sequence regioselectively and here allyl silanes (chapter 12) are most effective. SiEt3

CO2R 8

CO2R

4

CF3CO2H

3

CH2Cl2 –20 ˚C 3.5 hours

OH

257

1

H

H 5

7 6 H

2 258; 68-75% yield

The mechanism involves the formation of an allylic cation 259 from the tertiary alcohol 257 and the electrophilic attack on each alkene in turn until the allyl silane is reached when loss of the Et3Si group to the β-silyl cation terminates the reaction. Chemoselectivity is determined simply

17 Polyene Cyclisations

303

by the proximity of the right alkene. Regioselectivity comes from the way the molecule folds up 260 so that each new ring is already in a chair as it is formed. Stereospecificity - the relationship between centres 4 and 5 and between centres 6 and 7 - comes from the two central alkenes reacting in a concerted fashion so that their trans geometry becomes trans 3D stereochemistry in the product. Stereoselectivity - the relationship between centres 3 and 4 and between centres 5 and 6 - is again determined by the way the molecule folds.

X

257

SiEt3

Et3Si CO2R

CO2R

H

258

259

260

This may look like a rather specialised method but the biomimetic inspiration applies to much simpler compounds. Taxodione 261 is an antitumour compound from the swamp cypress Taxodium distichum. It has an extended quinone-like enone system and two trans fused six-membered rings. Livinghouse42 realised that the extended quinone might be derived by oxidation of a simpler aromatic compound 262 where “X” is an oxidisable group. He then saw a possible polyolefin cyclisation in which a tertiary cation derived from a simple alkene 263 starts the reaction and the aromatic ring terminates the sequence. OH

OH

O HO

HO

HO polyene

FGI

cyclisation

oxidation

H

H

O 261; taxodione

X 263

X 262

The unknown “X” now assumes a different role. If it is both oxidisable and anion stabilising, it can be used to build the molecule by a simple alkylation reaction. This disconnection is more or less in the middle of the molecule, both starting materials 264 and 265 will have to be prepared, and so the synthesis will be very convergent. The choice for “X” was cyanide and the phenolic OH groups were protected as OMe groups. The cyclisation was accomplished with BF3 in MeNO2 and gave a very good yield indeed of 267 with complete control over stereochemistry - even the cyanide centre was controlled as the CN group goes equatorial. OMe

OMe

MeO

OMe

MeO

MeO

265 BF3 MeNO2

Cl CN 264

LDA

CN 266; 88% yield

H

CN 267; 83% yield

304

17 Electrophilic Attack on Alkenes

The remaining steps are trivial except for the conversion of the cyanide to a ketone. This was achieved by oxidising the lithium “enolate” of the cyanide. The methyl groups were removed from the phenols with BBr3 and the final oxidation to the extended quinone was done with oxygen adsorbed onto silica. The whole synthesis took only seven steps and gave a remarkable overall 21% yield. Recent developments include asymmetric polyene cyclisations but also include new ways to initiate cyclisations. Iron tricarbonyl stabilises pentadienyl cations and its complexes with dienes are chiral. Addition of the achiral lithium derivative 268 to the chiral complex 269 gives the polyene precursor 270 that cyclises with Lewis acid to give a single diastereoisomer of the bicyclic compound43 271. OMe

OMe

OMe O

Fe(CO)3

Li

BF3.Et2O

OH

CH2Cl2

H Fe(CO)3

Fe(CO)3 268

269

270

271; 79% yield

The Lewis acid removes the OH group from the alcohol 270 to give the Fe(CO)3-stabilised cation 272 that initiates the cyclisation. The product is formed with the large iron diene complex in the equatorial position. OMe H 270

BF3.Et2O

Ar (OC)3Fe

CH2Cl2

Fe(CO)3

272

271

Looking Forwards We have spent some time in this chapter looking at the diastereoselectivity of some very important reactions. We shall be returning to the most important of these in chapter 25 where we discuss two of the most important reactions of the century - asymmetric versions of epoxidation and dihydroxylation by osmium. These developments were possible only because of the selectivities already established by the work described in this chapter and the selectivities we have discussed will be important in many syntheses.

References General references are given on page 893 1. J. K. Crandall, D. B. Banks, R. A. Colyer, R. J. Watkins and J. P. Arrington., J. Org. Chem., 1968, 33, 423. 2. D. R. Williams, B. A. Barner, K. Nishitani and J. G. Phillips, J. Am. Chem. Soc., 1982, 104, 4708. 3. Nicolaou and Sorensen, chapter 31. 4. S. Knapp, M. J. Sebastian and H. Ramanathan, J. Org. Chem., 1983, 48, 4786; S. Knapp, M. J. Sebastian, H. Ramanathan, P. Bharadwaj and J. A. Potenza, Tetrahedron, 1986, 42, 3405; S. Knapp, Chem. Soc. Rev., 1999, 28, 61.

17 References

305

5. A. Armstrong and A. G. Draffan, J. Chem. Soc., Perkin Trans. 1, 2001, 2861. 6. H. B. Henbest and R. A. L. Wilson, J. Chem. Soc., 1957, 1958, A. H. Hoveyda. D. A. Evans and G. C. Fu, Chem. Rev., 1993, 93, 1307. 7. B. B. Snider and N. A. Hawryluk, Org. Lett., 2000, 2, 635. 8. H. C. Brown and G. J. Lynch, J. Org. Chem., 1981, 46, 531, 930. 9. P. F. Cann, D. Howells and S. Warren, J. Chem. Soc., Perkin Trans. 2, 1972, 304. 10. H. L. Nyquist, E. A. Beeloo, L. S. Hurlbut, R. Watson-Clark and D. E. Harwell, J. Org. Chem., 2002, 67, 3202. 11. J. Tsuji, Synthesis, 1984, 369; Comp. Org. Synth., 7, 449. 12. M. Shibasaki, K. Iseki and S. Ikegami, Tetrahedron Lett., 1980, 21, 3587; K. Iseki, M. Yamazaki, M. Shibasaki and S. Ikegami, Tetrahedron, 1981, 37, 4411. 13. G. Zweifel and H. C. Brown, Org React., 1963, 13, 1; K. Smith and A. Pelter, Comp. Org. Synth., 8, chapter 3.10, page 703. 14. Nicolaou and Sorensen, page 491. 15. K. J. Henry, P. A. Grieco and W. J. DuBay, Tetrahedron Lett., 1996, 37, 8289. 16. Nicolaou and Sorensen, page 317; G. Zweifel and J. Plamoden, J. Org. Chem., 1970, 35, 898; S. L. Schreiber and K. Satake, J. Am. Chem. Soc., 1983, 105, 6723; 1984, 106, 4186. 17. A. Solladié-Cavallo, P. Lupatelli, C. Marsol, T. Isarno, C. Bonini, L. Caruso and A. Maiorella, Eur. J. Org. Chem., 2002, 1439. 18. K. Tamao, T. Tanaka, T. Nakajima, R. Sumiya, H. Arai and Y. Ito, Tetrahedron Lett., 1986, 27, 3377. 19. K. Tamao, T. Nakajima, R. Sumiya, H. Arai, N. Higuchi and Y. Ito, J. Am. Chem. Soc., 1986, 108, 6090. 20. E. J. Corey and J. Das, Tetrahedron Lett., 1982, 23, 4217. 21. R. S. Matthews and J. K. Whitesell, J. Org. Chem., 1975, 40, 3312. 22. M. D. Dowle and D. Davies, Chem. Soc. Rev., 1979, 8, 171. 23. J. Thibonnet, M. Abarbri, J-L. Parrain and A. Duchêne, Tetrahedron Lett., 1996, 37, 7507. 24. M.-C. Roux, R. Paugam and G. Rousseau, J. Org. Chem., 2001, 66, 4304. 25. S. Danishefsky, P. F. Schuda, T. Kitahara and S. J. Etheredge, J. Am., Chem. Soc., 1977, 99, 6066. 26. Nicolaou and Sorensen, page 173. 27. S. D. Burke, W. F. Fobare and D. M. Armistead, J. Org. Chem., 1982, 47, 3348. 28. R. N. Young, W. Coombs, Y. Guindon and J. Rokach, Tetrahedron Lett., 1981, 22, 4933. 29. I. Mavropoulos and P. Perlmutter, Tetrahedron Lett., 1996, 37, 3751. 30. B. J. Snider, Comp. Org. Synth., vol 2, p. 527. 31. M. Delmas and A. Gaset, Synthesis, 1980, 871; R. El Gharbi, M. Delmas and A. Gaset, Synthesis, 1981, 361. 32. P. R. Stapp and C. A. Drake, J. Org. Chem., 1971, 36, 522; P. R. Stapp, J. Org. Chem., 1971, 36, 2505. 33. K. Al Mutairi, S. R. Crosby, J. Darzi, J. R. Harding, R. A. Hughes, C. D. King, T. J. Simpson, R. W. Smith and C. L. Willis, Chem. Commun., 2001, 835; S. R. Crosby, J. R. Harding, C. D. King, G. D. Parker and C. L. Willis, Org. Lett., 2002, 4, 577. 34. A. T. Blomquist and R. J. Himics, J. Org. Chem., 1968, 33, 1156. 35. B. B. Snider and A. C. Jackson, J. Org. Chem, 1983, 48, 1471. 36. G. E. Keck, J. A. Covel, T. Schiff and T. Yu, Org. Lett., 2002, 4, 1189. 37. A. Pelter, K. Smith and H. C. Brown, Borane Reagents, Academic Press, London, 1988. 38. H. C. Brown, Acc. Chem. Res., 1969, 2, 65; H. C. Brown and M. W. Rathke, J. Am. Chem. Soc., 1967, 89, 4528; E. Negishi and H. C. Brown, Org. Synth. Coll, 1990, VII, 427. 39. A. Pelter, K. Smith, M. G. Hutchings and K. Rowe, J. Chem. Soc., Perkin Trans. 1, 1975, 129, 138; A. Pelter, M. G. Hutchings and K. Smith, J. Chem. Soc., Perkin Trans. 1, 1975, 142; A. Pelter, M. G. Hutchings, K. Smith and D. J. Williams, J. Chem. Soc., Perkin Trans. 1, 1975, 145. 40. H. C. Brown, H. Nambu and M. R. Rogic J. Am. Chem. Soc., 1969, 91, 6852, 6854, 6855. 41. W. S. Johnson, Y-Q. Chen and M. S. Kellogg, J. Am. Chem. Soc., 1983, 105, 6653; R. Schmid, P. L. Huesmann and W. S. Johnson, J. Am. Chem. Soc., 1980, 102, 5122, Tetrahedron Lett., 1986, 27, 6027. 42. S. R. Harring and T. Livinghouse, J. Chem. Soc., Chem. Commun., 1992, 502. 43. M. Franck-Neumann, P. Geoffroy and D. Hanss, Tetrahedron Lett., 2002, 43, 2277.

18

Vinyl Cations: Palladium-Catalysed C⫺C Coupling Introduction: Nucleophilic Substitution at sp2 Carbon does NOT Occur Stereospecificity in the addition-elimination mechanism Towards Carbon Nucleophiles and Vinyl Cation Equivalents Vinyl cation equivalents Conjugate Substitution Unsymmetrical 1,3-dicarbonyl compound derivatives Unsymmetrical enaminones Conjugate Addition to Alkynes Sulfur-based leaving groups Sulfoxides and sulfones The Diels-Alder Reaction on β-Bromo and β-Sulfonyl Alkynes Choice of leaving group Modified Conjugate Addition The Heck Reaction General description Scope and limitations of the Heck reaction: synthesis of dienes The Heck reaction with electron-rich alkenes A synthesis of strychnine Recent developments in the Heck reaction Sp2 –sp2 Cross-Coupling Reactions by Transmetallation Stille coupling Recent developments in Stille coupling Variations in Stille coupling Suzuki coupling Recent developments in Suzuki coupling Summary

Introduction: Nucleophilic Substitution at sp2 Carbon does NOT Occur If you want to make a diene one obvious place to disconnect is between the two alkenes 1. Though this has a pleasing symmetry, one of the synthons must be a vinyl cation 2 and the other a vinyl anion 3. We have already met stereochemically controlled reagents for the vinyl anion synthon such as vinyl metals 4 (chapter 16) and all we now need is a good reagent for the vinyl cation. That is the subject of this chapter. First we need to examine why you can’t just use a vinyl halide and expect to get substitution of the halide ion.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

308

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

R2

R1 1

R2

+

R1 2; vinyl cation

=

3; vinyl anion

R2 Metal 4; vinyl metal

Nucleophilic substitution of unactivated vinyl halides is a rare reaction.1 The SN2 reaction at a trigonal sp2 carbon atom 5 is generally thought to be unknown. It is puzzling that this innocent looking reaction should be impossible: could it be that the transition state, which would have to contain square co-planar carbon 6, is too high in energy? Recent work 2 suggests the PhI may be displaced from vinyl iodonium salts 8 by halide nucleophiles with inversion of configuration (E-8 gives Z-9 with inversion) but it remains to be seen if this is a useful reaction in synthesis. X

X

R1

R1

H

Nu

Nu

H

R1 Nu

6; square planar transition state

5

X

IPh

1

R

7

R1 X

Bu4N

8

9

Nucleophilic substitution at trigonal carbon is known by the addition-elimination mechanism 10 where an enolate intermediate 11 is formed and the negative charge is deposited on an electronegative atom such as oxygen. This is conjugate substitution and you will see the analogy with nucleophilic aromatic substitution. There is no need for stereospecificity in the two-step process and we might expect Z-alkenyl halides 10 to give E-enone products 12. O

conjugate addition

Br Nu

R

O

conjugate elimination

Br

R

Nu

10

O

R

Nu

11

12

The third possibility is an SN1 reaction with a vinyl cation as intermediate. This is unlikely when we are dealing with vinyl bromide itself 13 as the cation 14 would be very unstable. Notice that the vinyl cation would be a linear species with an empty p-orbital. Such cations are known when they are stabilised by conjugation 17 and the leaving group is excellent, such as triflate (OTf ⫽ CF3SO2O⫺ in 16) but these SN1 reactions are rarely used in synthesis. H

Br 13

×

Ar

Ar

H

Ar

Nu Nu 14; vinyl cation

15

OTf 16

Nu 17

18

Stereospecificity in the addition-elimination mechanism By far the most important of all these mechanisms is addition-elimination. The carbanion intermediate must be stabilised by a group Z such as Ar, COR, CN, RSO, or RSO2. The reaction then becomes a conjugate substitution as 10 to 12. It appears there that, if the intermediate 11 has a lifetime longer than bond rotation, only the more stable of the two products (E in this case) will be formed. However, there are quite a few cases where the first step is rate-determining and the intermediate has a very short lifetime indeed. These reactions go with retention of configuration. An important example is the replacement of halides by thiolate nucleophiles3 in 2-bromo styrenes 19.

18 Towards Carbon Nucleophiles and Vinyl Cation Equivalents

309

Br AlkS

Ar

Ar

Br

SAlk

E-19

E-20

SAlk AlkS

Ar

Ar

Z-19

Z-20

The explanation resembles that for the retention of stereochemistry in the reaction of vinyl silanes with electrophiles (chapter 16). To make this clearer, we have drawn the diagrams as nearly the same as those in that chapter as possible. It does not matter which surface of the alkene is attacked by the nucleophile as long the bromine atom continues to rotate in the same direction. Once it reaches the vertical it can be eliminated by the carbanion and the result is retention (⬍5% of the other isomer). Nu

Nu Nu

Br

Ar

Nu

Ar

Br

Ar

Nu Ar

H

nucleophilic attack

Ar

Br

Br starts to rotate downwards

Br

Br rotation complete

product formed with retention

Br lost

The reaction even goes twice if there are two leaving groups (chloride in this case 21) and the products can be oxidised to the more stable sulfones 24. It is more impressive with the Z-isomer.

Cl

SAlk AlkS

Cl Z-21

SAlk AlkS

Cl

O O Alk O O S [O]

AlkS

Z-22

S

Alk Z-23

Z-24

Other systems that allow stereospecific substitution with retention include ester-activated chloride displacement by an amine (25 to 26) and sulfone-activated chloride displacement by azide (27 to 28). The choice of E- or Z-isomers is arbitrary as each gives ⬎98% retention.1 The reactions shown so far do not include carbon nucleophiles. NH

CO2Et

Cl

N

E-25

Cl

CO2Et

N3

N3

SO2Ar E-26

SO2Ar

Z-27

Z-28

Towards Carbon Nucleophiles and Vinyl Cation Equivalents The subject of this chapter is how we can achieve reaction of nucleophiles with vinyl electrophiles such as vinyl halides. We cannot easily make SN1 or SN2 reactions happen at sp2 carbon atoms but we can make the products of those unfavourable reactions by other reactions in which the same bond is formed. We want to be able to use carbon nucleophiles. We also want to control the stereochemistry of the double bond in the product. This is the disconnection we want to achieve 29: Nu

Z 29

Nu

Z 30

Nu

Z 31

Nu

Z 32

You may think that we already know how to do this by conjugate or Michael addition where Z is an anion-stabilising group. These reactions do add nucleophiles to double bonds but the

310

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

intermediate is an enolate and the double bond disappears in the product so that the disconnection is 31 and the alkene 32 is a reagent and not a synthon. We would be happy with conjugate addition as a solution to the problem if the double bond could be restored in the product. We shall first consider some conventional solutions to the problem such as conjugate substitution and conjugate addition with the restoration of the double bond and then move on to more important modern solutions mostly based on palladium chemistry.

Vinyl cation equivalents We shall start with a pharmaceutical example, the Wyeth analgesic Meptazinol4 33. This compound has a seven-membered ring amine attached through a quaternary carbon atom to a meta-hydroxy benzene ring. The greatest simplification would undoubtedly arise by disconnection of the strategic bond joining the two rings. But the functionality is all wrong for this. Hydroxy groups direct ortho and para and the nitrogen cannot help to activate the required carbon on the seven-membered ring. Changing the functionality of the amine to an amide makes a lot of sense as a correctly placed carbonyl group activates the right carbon atom and amides can be converted into amines by reduction. However this will require the use of a nucleophilic enolate 34 at what is to become the quaternary carbon atom and so forces us to use an equivalent of an aryl cation 35: this route was beautifully realised by the use of benzyne chemistry and this is described in the workbook. OH

OH OH ?

?

?

O

+ N

N

Me 33; meptazinol

Me 34; amide enolate

Me

+

N

FGI

35; benzyne chemistry?

Conjugate Substitution More relevant to this chapter is the realisation of a vinyl cation strategy which was in fact chosen as the development route for the product. Adjusting the oxidation state of the benzene ring reveals a promising enone 37 and requires a reagent for the vinyl cation synthon 38. OH

O

FGI

FGI

reduction

oxidation

?

33 O NMe 36

O

O NMe 37

34; amide enolate 38; vinyl cation synthon?

The best way to derive a reagent for the synthon 38 is to insert a leaving group at the vinylic position. The enol ether 40, easily prepared from the 1,3-diketone 39 turns out to be ideal as we can envisage conjugate substitution by the amide enolate 34.

18 Conjugate Substitution

O

311

O

O

i-PrOH

O

RLi?

toluene

O

i-PrO 40; enol ether

39

i-PrO

R

R 41

42

Addition of the magnesium enolate of the amide followed by acidic work-up gave the amidoenone in one step with the double bond still present in the ring. Notice that direct addition of the enolate to the carbonyl group 41, rather than the conjugate addition envisaged, actually occurred. This doesn’t matter as the enol ether 40 has a symmetrical carbon skeleton. The synthesis was completed by an oxidation of the enone ring with bromine to give the benzene ring in 36 and a reduction of the amide to give meptazinol itself. i-PrO

O

1. i-Pr2NMgBr

H

OH O

2. 40

NMe

H2O

Br2

37

36 80% yield

CHCl3

LiAlH4

33 90% yield

NMe 44

43

The nucleophile does not have to be an enolate: ortho-lithiated aromatics (chapter 7) work well as in this sequence used in the synthesis of sesquiterpene lactones. 2-Methyl furan 45 is lithiated in the one remaining α-position 46 and it adds to the ethyl equivalent of 40 to give the enone 48 in quantitative yield.5

O

BuLi

O

Et2O

O

45

Li

OEt

O

+

O

46

48; 100% yield

47; enol ether

The further development of 48 is interesting. Conjugate addition of Me2CuLi creates a quaternary centre 49 and acid hydrolysis of the furan releases the triketone 50 used to make sesquiterpene lactones. The yields in these three steps are amazingly good. O 48

O

Me2CuLi Et2O, 0 °C

O 49; 100% yield

O

HCl, MeOH room temperature

O 50; 100% yield

Substituted versions of these reagents retain the symmetry of 40 and 47. One interesting type is prepared by a Prins reaction of the diketone 51 with trioxan 52, the trimer of formaldehyde.6 The six- 54 and seven-membered 55 compounds are easily made in the same way.

312

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

O

O O

O

+ O

BF3.Et2O

51a

O

O

O 53; 72% yield

O 54; 84% yield

O

CH2Cl2

O

OH

51

O

O

O

52

O 55; 72% yield

Reaction with organo-lithium compounds works well with all three examples: one molecule of formaldehyde is lost but a CH2OH group is retained in the products 56, 57, and 58 from addition of vinyl, phenyl, and butyl lithium. This chemistry was used to make bertyodionol.7

Li

53

O

O

O OH

THF, –78 °C

54

OH

PhLi

55

BuLi

OH

Ph 57; 93% yield

56; 94% yield

Bu 58; 89% yield

Unsymmetrical 1,3-dicarbonyl compound derivatives For unsymmetrical versions do not start from the diketone. One way is to start from a monoketone and react with an amide acetal (like those used in Claisen rearrangements, chapter 15). Thus cyclopentanone gave the enaminoketone 59 that reacts with BuLi to give the enone 60. There is no doubt here that the nucleophile has added in a conjugate fashion.8 O

O

OMe MeO

O

NMe2

NMe2

110 °C, 12 hours

BuLi

Bu

THF, –30 ˚C

59; 68% yield

60; 75% yield

The same intermediates 62 can be formed from diketones and amines, particularly pyrrolidine, and react well with organolithium compounds. The newly added RLi replaces pyrrolidine in good yield providing that a hydrocarbon solvent is used. The yields of dienyl ketones such as 63 are very good.9

O

O

O

N H

t-Bu

N

t-Bu 61

O Li

62

hexane

t-Bu 63; 84% yield

Unsymmetrical enaminones It is obviously easier to prepare symmetrical enaminones such as 64 by this method but these can be ‘desymmetrised’ by enolate alkylation to give, say, 65 before conjugate substitution leads to the final product 66. The yields are reasonable for a three-step process and the middle step gives the product 65 of γ-alkylation of an extended enolate (see chapter 11).

18 Conjugate Addition to Alkynes

O

313

O

O

O N H

1. LDA

N

BuLi

N

2. EtI

O

Et 64

51

Bu

petrol ether

Et

65

66; 56% yield

The reverse regioselectivity is realised by cerium-catalysed addition of Grignard reagents to enamino ketones 68 formed with primary amines.10 The structure of the product shows that direct addition to the carbonyl group has occurred: the Ce atom presumably delivers the butyl group through chelation to nitrogen. There is reasonable stereochemical control: the newly introduced group goes trans to the carbonyl group in the product 70. O

O

O

MeNH2

Ph

NHMe

CeCl3

Ph 67

Ph

BuMgCl

NMe

Ph

O

HOAc H2O

Bu

68

Bu 70; 65% yield

69

Conjugate Addition to Alkynes Conjugate addition alone gives a vinyl derivative if the electrophile is an alkyne. We have seen examples of this in chapter 15 where, for example, halides and thiolate nucleophiles gave substituted alkenes such as 71 and 73 from 72 as the E-isomers under equilibrating conditions. The bromoester 73 is easily made from the acetylenic acid 72b and esterified afterwards.11 These compounds will now prove useful as vinyl cation equivalents as conjugate substitution on 71 and 73 is generally more successful than conjugate addition to 72, at least with carbon nucleophiles. We are not concerned here with the related carbometallation of alkynes described in chapter 16 as the products are formally vinyl anion equivalents. However, as we shall see later in this chapter, the distinction is blurred in Pd-catalysed couplings. CO2Me

RS

RSH

CO2Me

base

71

MeOH

CO2H

H

72a

1. HBr 2. MeOH,H

72b

CO2Me

Br 73

So lithium enolates react with the halide 73 by conjugate substitution. The lithium enolate of the enantiomerically pure heterocycle 74 prepared from natural alanine gives one diastereoisomer of E-75 in reasonable yield. After further manipulation the heterocycle was hydrolysed to give cyclopropanes that are glutamate antagonists.12 O

CO2H

O

Ph H2N L-alanine

O 1. LiN(SiMe 3)2

N Bz 74

Me

2. 73

O

Ph N Me Bz 75; 65% yield

CO2Me

The corresponding ketones 76 can be made by aliphatic Friedel-Crafts reactions between an acid chloride and acetylene using AlCl3 as catalyst. Under these conditions chloride adds to the acetylenic ketone to give the β-chloroenone 77 directly. These compounds are converted into the sulfides 78 by conjugate substitution13 and are used as vinyl cation equivalents in the next section.

314

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

O

O

H

O

HCI

RCOCI

PhSH

H R

AICL3

Cl

PhS

R

CH2Cl2

77

76

Et3N

R

78

Sulfur-based leaving groups The sulfur compounds like 71 can be made by conjugate addition not only to alkynes but also to enones, e.g. 79, by using the Pummerer oxidation of the products 80 with N-chlorosuccinimide (NCS) 82 to give a β-thio-enone 81 not accessible by conjugate addition to an alkyne. O

O

O ArSH

Me3Si

O

NCS

Me3Si

SAr

79

CCl4

80

N Me3Si SAr 81; Ar = p-Tolyl

O

Cl 82; NCS N-ChloroSuccinimide

While organo-lithium compounds add directly to the carbonyl group of 81, cuprates carry out a conjugate substitution to give the enone 83. Now a conjugate addition gives the ketone 85. It is necessary to use Cu(I) catalysis and silyl trapping (see chapter 9) to make this conjugate addition work but it gives good yields. The group added second (here p-tolyl) ends up trans to the Me3Si group on the other side of the ring.14 And what is the Me3Si group there for? That is explained in the workbook.

O 81

OSiMe3 ArMgBr

Me2CuLi

KF

cat. Cu(I)

Me3Si

Me

Me3SiCl

Ar Me3Si

Me 84; Ar = p-Tolyl

83

O

MeOH

Ar Me3Si

Me 85; 90% yield

Sulfoxides and sulfones One advantage of sulfur chemistry is that reagents can be used at three oxidations levels: sulfide 80, sulfoxide and sulfone. The enones with the more highly oxidised sulfur atoms are more electrophilic and need less reactive nucleophiles. Thus the sulfoxide 87 reacts with unactivated pyrroles and furans, e.g. 86 with or without acid catalysis.15 O 25 ˚C, 6 days, 91% yield

+ Ph S

O 86; 10-fold excess

87

O

TsOH catalyst 25 ˚C, 1 hour, 86% yield

O 88

O

The nucleophiles 89; Z ⫽ O or NR react by normal electrophilic aromatic substitution: both the addition 89 and the expulsion of the leaving group 91 are faster with PhSO instead of PhS. In the acid catalysed reactions the intermediates 90 and 91 would be enols rather than enolates.

18 The Diels-Alder Reaction on β -Bromo and β -Sulfonyl Alkynes

315

O O

PhS

Z

Z

O

H

S

O

89

O

product, e.g. 88 + PhSO

Z Ph

O

90

S

Ph

91

vThe more reactive pyrrole 92 needs a shorter time while imidazole and pyrazole 94 react much faster but through nitrogen. All these reactions go in remarkably high yield. 87

87

N

25 ˚C,14 hours

N

N

H 92; 10-fold excess

N

H O 93; 98% yield

N

25 ˚C, 8 days MeCN

N

H 94; 1.2 equivs

O 95; 92% yield

The sulfone 98 allows a tandem (chapter 36) sequence of remarkable selectivity. The lithium derivative of the allylic phosphine oxide 96 (see chapter 12) adds to the enone 97 to give the lithium enolate 99 trapped by the sulfone 98 to give the complex product 100 in 96% yield: the only imperfection in stereochemical control is 3% of the Z-vinyl phosphine oxide.16 O O

O

PPh2

O

96

S O

97

98

Ph

O O

O PPh2 93

1. BuLi

O

H

2. 97

PPh2

O

3. 98

H

O O

LiO

O 100; 76% yield

99

The Diels-Alder Reaction on β-Bromo and β-Sulfonyl Alkynes An alternative to conjugate addition to alkynes is the Diels-Alder reaction that gives the required β-substituted alkenes without conjugate addition. Thus the deactivated pyrrole 101 reacts quite well with the activated haloalkyne ester 102 to give the bicyclic compound 103 having the right arrangement of Br and CO2Me to allow conjugate substitution. Alkyl copper reagents are best and give good yields of compounds such as 104 used in a synthesis of epibatidine.17 MeO2C

Br

MeO2C

N N CO2Me 101

90-95 ˚C

+

N Br

BuCu

Bu

30 minutes no solvent

CO2Me 102

CO2Me 103; 56% yield

CO2Me 104; 78% yield

The more elaborate diene 105, prepared by enantioselective reduction of the corresponding ketone, (chapter 26) reacts with the β-tosyl alkyne acid 106 to give a good yield of one regio- and

316

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

stereo-isomer of the lactone 107 having the leaving group in the right position for replacement with a nucleophile. A methyl group was needed for the synthesis of forskolin and Me2CuLi was the best reagent.18 O OH

HO2C

O

O

O

23 ˚C 30 hours

S

O 105

Tol

Ts

Me

Me2CuLi BF3.OEt2

CHCl3

O

106

107; 72% yield

108; >95% yield

Choice of leaving group You will notice that simple organo-lithium and copper compounds as well as lithium enolates are used in these conjugate substitution reactions. If the leaving group is a halide, there is a danger of transmetallation with RLi (though not with lithium enolates) and organo-copper compounds are safer. With leaving groups such as NR2, OR, or sulfur in various oxidation states, there is essentially no danger of transmetallation as oxidative insertion into the strong C-N, C-O, or C-S bonds is rare, and simple RLi can be used.

Modified Conjugate Addition The double bond can be restored after a conjugate addition to an electrophilic alkene if the enolate intermediate 110 is trapped as a silyl enol ether 111 and then combined with a sulfur or selenium electrophile which is later eliminated. Organocuprates are ideal nucleophiles for this process as the intermediate enolate can be trapped as a silyl enol ether and reacted directly with PhSCl or PhSeCl. O

R2CuLi

R

109

OLi

Me3SiCl

R

OSiMe3 111

110

This method of making specific enolate equivalents we met in chapter 9. Reactive electrophiles like PhSCl and PhSeCl react rapidly with the silyl enol ether 111 to give the 2-PhS(e)-ketone 114. Addition is to the face of the ketone not blocked by the alkyl group on C-2. Cl S(e)Ph

S(e)Ph

S(e)Ph

PhS(e)Cl

111

R

OSiMe3 112

R

O 113

Cl SiMe3

R

O 114

Oxidation of the sulfur or selenium atom to the sulfoxide or selenoxide stage allows thermal elimination to take place by a cyclic mechanism 115. Because the PhS(e) group is trans to the alkyl group it must be cis to the hydrogen atom at C-2 and so cis elimination is possible. The selenoxides are unstable and decompose spontaneously but the sulfoxides usually need heating. The double bond is put back in its original position 116 so the strategy of the sequence is addition

18 The Heck Reaction

317

of an alkyl anion to a β-enone cation synthon 117. This chemistry is discussed in more detail in chapter 33 under ‘oxidation of enolates’. O S(e)Ph

O

H

[O]

R

Ph S(e)

O

114

PhS(e)OH +

O

R

R

115

O 117 vinyl cation synthon

116

A more general solution to the problem is to change the nucleophile from an organo-lithium or copper species to an organo-palladium compound and this leads on to the Heck reaction.

The Heck Reaction General description We have discussed palladium chemistry in chapter 8 where there was a brief description of palladium σ-complexes. In this chapter we shall see the results of adding palladium σ-complexes to unsaturated carbonyl compounds. Michael addition occurs but the palladium leaves the intermediate in such a way that the original alkene is restored.19 We shall take a simple example first – the addition of iodobenzene to methyl acrylate. PhI +

CO2Me 118

cat. Pd(PPh 3)4

Ph

CO2Me 119

The first step is oxidative addition of the Pd(0) species to iodobenzene. This is called oxidative addition because the Pd atom adds between the benzene ring and the iodine atom and is oxidised from Pd(0) to Pd(II) 120. The mechanism is described in chapter 8. Next the alkene forms a π-complex 121 with the palladium atom. The π-bond is a two-electron ligand and it makes sense for it to displace one of the two-electron phosphine ligands. However, we should emphasise that it is not always possible to be precise about which other ligands are present on the palladium atom in these reactions. Now the nucleophilic attack occurs. The palladium atom transfers its phenyl ligand to one end of the double bond and attaches itself to the other as a σ-complex 122. This is like a phenyl anion doing a Michael addition to the alkene which is made even more electron deficient by complexation to Pd(II). This view explains the regioselectivity of the addition as “Ph⫺” would certainly attack the more electron-deficient end of the alkene and the Pd would prefer to form a σ-complex to the more carbanion-like enolate position. However, you can view it as a coupling reaction on the surface of the metal if you like. Pd(0)

PhI

oxidative insertion

PPh3

I

CO2Me

I

Ph

PPh3 120

Ph

Ph CO2Me

121

PPh3

I

PPh3 Pd

Pd

carbopalladation

Pd PPh3 CO2Me 122

Palladium alkyl complexes are inherently unstable because of β-elimination. We saw this in action in the discussion on hydrogenation in chapter 8. The palladium atom leaves with a hydrogen

318

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

atom from C-3 of the ester 123 and the σ-complex 122 reverts to a π-complex 124 with the product. The product is freed if the palladium (II) atom drops out of the π-complex as 125: it can lose HI and revert to Pd(0) ready for the next cycle. During the reactions Pd is present in all the complexes as Pd(II) but it enters and leaves the cycle as Pd(0). I Ph H

PPh3

I

Pd PPh3 CO2Me 123

reductive elimination

I

PPh3 Pd

CO2Me

H Ph

PPh3

Ph3P

Pd

115 +

PPh3

H

124

PPh3 Pd

125; Pd(II)

Ph3P

PPh3

126; Pd(0)

The stereochemistry of the double bond in the product is a result of the stereoselective β-elimination of palladium - this reaction could give either the E- or the Z-cinnamate but the reaction is reversible and the E-cinnamate is more stable. This sort of Heck reaction is normally done in a more convenient way. Though Pd(0) really starts off the cycle, Pd(II) is involved and Pd(OAc)2 [or (Ph3P)2Pd(OAc)2] is a more convenient reagent than Pd(PPh3) 4. This same product can be made in excellent yield under very mild conditions without any phosphine at all.20 An explanation of the necessary reduction of Pd(II) to Pd(0) is in the workbook. PhI +

CO2Me 118

cat. Pd(OAc) 2 Bu4NCl

Ph

CO2Me 119; 91% yield

K2CO3, DMF, 27 ˚C

Scope and limitations of the Heck reaction: synthesis of dienes Because of β-elimination of palladium, this reaction cannot be used with most alkyl halides. However, vinyl halides are fine as their palladium σ-complexes do not undergo β-elimination to give alkynes. Here is a simple example. I + 127

CO2Me 118

cat. Pd(OAc) 2

CO2Me E-128

The mechanism is the same as before with vinylPdI replacing PhPdI. The E-alkene is again produced under thermodynamic control. But what happens if the vinyl iodide has some stereochemistry? Then there is stereospecificity in the formation of this alkene: Z-vinyl iodides 129 give Z-alkenes and E-vinyl iodides give E-alkenes stereospecifically with retention. In chapter 16 we discussed this sort of reaction - replacement of halide by metal or of metal by metal at a trigonal carbon atom normally occurs with retention. This reaction can give E,E- or Z,E- dienes 130 but not E, Z- or Z,Z-dienes.20 I R Z-129

cat. Pd(OAc) 2 118

CO2Me R R = Bu; 90% yield, 95:5 Z,E:E,E-130

cat. Pd(OAc) 2

R

I E-129

118

R

CO2Me R = Bu; 96% yield, 99:1 E,E:Z,E-130

As we have drawn the products, the right hand double bond is formed stereoselectively under thermodynamic control, and can be E only, while the left hand double bond is formed stereospecifically with retention from the original vinyl halide and can be E or Z.

18 The Heck Reaction

319

The Heck reaction with electron-rich alkenes The obvious limitation of this reaction so far is in the electrophile as we have suggested naturally electrophilic enones as ideal partners for aryl or vinyl palladium σ-complexes. In fact, complexation with palladium makes all alkenes electrophilic and nucleophilic addition can in principle occur to a simple alkene while it and the nucleophile are both bound to the palladium atom. Such reactions are known for aryl halides. Even ethylene itself does satisfactory Heck reactions and its reaction with the bromopyridine 132 is the basis for a large scale process leading to a drug.21 Br

H2 N

Br2

50 psi ethylene

Ac2O

Pd(OAc) 2, (o-Tol) 3P

N

AcHN

131

N 132

AcHN N 133; 72% yield 10 kg scale

MeCN, Et 3N, 90 ˚C

In practice, with vinyl halides, three regioselective problems arise - which alkene is to act as the electrophile, which end of the alkene is attacked by the nucleophile, and which way does the β-elimination occur? I +

Pd(0)

R

?

R

These problems may be avoided by making an organometallic compound [often of Cu(I) or Mg(II)] from the halide intended to become the palladium σ-complex and to use another vinyl halide as the electrophile. Palladium prefers to take out a halide rather than a hydrogen atom in the β-elimination step and all three problems disappear. In the synthesis of this E,Z-diene 134, we prefer to use a Pd σ-complex for the Z-alkene part 136 and a vinyl iodide 135 for the E-alkene. Z I

+ Br

E 135; electrophilic vinyl group

134

Z-136

to be used as an organo -Mg or organoCu reagent

The Grignard reagent from Z-1-bromopropene combines with E-1-iodo-octene with catalytic Pd(PPh3) 4 to give the pure E,Z-diene in 87% yield. The Mg and Pd exchange by transmetallation as Mg forms stronger bonds to oxygen.22 Br

1. Mg

BrMg

Et2O

Z-136

2.

I Hex

cat. Pd(PPh 3)4

Z-137

E,Z-134; 87% yield

This solution takes the method outside the scope of the Heck reaction which is better realised with vinyl triflates 138 [Triflate ⫽ OTf ⫽ OSO2CF3]. These can be made into palladium σcomplexes but do not function as electrophiles since the β-elimination of Pd and OTf is a poor reaction. The small amount of Pd σ-complex 139 formed in solution adds both to the alkene and the remaining vinyl triflate but β-elimination leads rapidly to product, releasing Pd(0) for the next cycle, from only intermediate 141.

320

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

PdL2OTf

OTf

Pd

Pd(0)

OTf 138

β-elimination of Pd and OTf

OTf 140

OTf

L L 139

H

intermediate σ-complex

no product

PdL2OTf FAST

R

R

SLOW

141

β-elimination of Pd and H

R 142; product

Vinyl triflates, which are of course enol triflates, such as 144 and 147, can be made directly from ketones 143 and 146, and react with electron-deficient, conjugated, and electron-rich alkenes. Here are just two examples giving dienes 145 and 148 in good yield.

Ph

1. base 2. Tf 2NPh

O

(Ph3P)2Pd(OAc) 2

TfO

143; steroid O

145; 77% yield OTf

CHO CHO

1. LDA 2. Tf 2NPh

t-Bu

Ph

144

(Ph3P)2Pd(OAc) 2

t-Bu

146

t-Bu 148; 86% yield

147

A synthesis of strychnine Rapid development in the Heck reaction has made it one of the most important of all modern methods. A dramatic example is a recent synthesis of strychnine by M. Nakanishi and M. Mori.23 This synthesis illustrates the intramolecular Heck reaction particularly well, showing examples of regioselectivity in attack on the alkene and in formation of the new alkene. The synthesis starts with a Heck reaction on enantiomerically pure 149. Attack occurs on the nearer end of the alkene to give 150 that cannot eliminate to reform the same alkene as there are no H atoms left and the alkene 151 must be formed. The stereochemistry is controlled by the short tether linking the aryl bromide and the alkene.

Br

NC

CN

PdL2

CN

2 mol% Pd(OAc) 2

N Ts 149

PMe2Ph, Ag 2CO3

H

DMSO, 90 ˚C

N Ts 150

H

N Ts

H

151; 87% yield

The second ring is closed by an amido-palladation reaction. The nitrile in 151 is reduced to a primary amine and protected with a Boc group. Reaction with Pd(II) allows nucleophilic attack by the amide on the nearer end of the alkene and β-elimination again pushes the alkene round the ring 153, this time because there is no syn H atom in the intermediate 154 at the site of addition. Palladium is released as Pd(0) but this reaction needs Pd(II) so the quinone and MnO2 are there to carry out the required oxidation.

18 The Heck Reaction

321

NBoc

NHBoc 151

NBoc

1. LiAlH 4

Pd(OAc) 2

2. (Boc) 2O

quinone MnO2

N

PdL2

H H

N

H

Ts 152; 74% yield

H

Ts 153; 87% yield

154

Now the alkene must be moved yet one more time around the ring to prepare the way for another intramolecular Heck reaction. Hydroboration (chapter 17) of 153 is regioselective because of the large N-Boc group and Swern oxidation completes the insertion of the ketone 155. Reduction and elimination use another palladium-catalysed reaction. Conversion to the triflate 157 is followed by Pd-catalysed transfer hydrogenation, the H atom coming from formic acid HCO2H. NBoc

NBoc

NBoc

H

H

1. 9-BBN

153

2. H 2O2 NaOH

1. PhNTf 2 (Me3Si)2NK

H

3. (COCl) 2 DMSO, Et 3N

N

O

H

Ts 155; 70% yield

2. Pd(OAc) 2 HCO2H, Ph 3P i-Pr2NEt

N

H

OTf

H

Ts 156; 64% yield

157

The molecule is now prepared for the next Heck reaction by reductive removal of the Ts group from nitrogen and acylation with Z-3-bromoacryloyl chloride to give 158. Intramolecular Heck reaction closes the ring 159 in reasonable yield and with the correct and expected regioselectivity. The stereochemistry of the new centre is irrelevant as it will disappear. The double bond is moved back round the ring one place! In this reaction we see a β-bromo unsaturated acid derivative providing the nucleophile for the Heck reaction rather then the electrophile for conjugate substitution. NBoc

NBoc 1. Na naphthalene

156

Pd(OAc) 2, Ph 3P

H

2. RCOCl K2CO3

N

H

i-Pr2NEt, DMSO

H

O 158; 56% yield

N

Br

H

O 159; 46% yield

The double bond is moved into conjugation 160 with base and the side chain for the last Heck reaction is added by deprotection at nitrogen and alkylation with an allylic bromide to give 161. Notice that the allylic bromide reacts rather than the vinylic iodide as this reaction is not catalysed by palladium. N

NBoc

159

I

1. CF 3CO2H

i-PrO

H

H

i-PrOH

N

2. Li 2CO3, DMF

H

I

O 160

N

Br OTBDMS

H

O 161; 53% yield

OTBDMS

322

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

Now the Heck reaction can be induced to close the ring 162. In this case the vinyl iodide forms the palladium derivative and the regioselectivity of the attack is at the δ-position of the dienone mainly because of the tether. The closure of the last ring was already known and the synthesis of strychnine 164 complete. This synthesis uses three Heck reactions and two other palladiumcatalysed reactions. N

161

Pd(OAc) 2 Bu4NCl

N

H

H

1. LiAlH 4

N

K2CO3

N

2. HCl

H

DMF

RO

O

KOH

H

H

HO

O

162; R = TBDMS, 48% yield

EtOH

163; 44% yield N

H N

H

H

H

O O H 164; strychnine

Recent developments in the Heck reaction The big difference between the Heck reaction and those couplings to follow in the next section is that only one of the alkenes is marked with functionality (halide, triflate): the other is a simple alkene and loses a hydrogen atom during the Heck process. This sounds like a disadvantage where regioselectivity is concerned but some recent examples should convince you that the Heck reaction is very important indeed. So the Heck reaction can be combined with a Pd-catalysed ortho alkylation of an aromatic halide 165 to give benzoxepines 166 in excellent yield even with an unprotected amide on the ring.24 CO2Et I

Br O

CO2Et

O

Pd(OAc) 2, TFP, Cs 2CO3 MeCN, reflux

AcNH

AcNH

165

166

A bonanza of Heck reactions was used by Tietze25 in the synthesis of cephalostatin analogues. Corey-Fuchs reaction on the aldehyde 167 gives the alkene 168 from which the trans Br atom is stereoselectively removed by Pd-catalysed tin hydride reduction to give the Z-vinyl bromide 169. This reaction is discussed below under Stille coupling. O

O

O Br

O

Ph3P, CBr 4

Br

O

Bu3SnH

Br

CH2Cl2, 0 ˚C

Br

CHO 167

O

Br

Pd(PPh3)4

Br

Br 168; 80% yield

Br 169; 98% yield

Br

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

323

Heck reaction with the enantiomerically pure bicyclic alkene 170 gives the new C–C bond 171 with retention at the vinyl bromide, stereoselective introduction of two new stereogenic centres, and relocation of the alkene since cis elimination of palladium is required. Repetition of the first two reactions after deprotection of the other aldehyde gives a new Z-vinyl bromide 172 ready for the next Heck reaction. Ot-Bu Ot-Bu

O

169 Pd(OAc) 2 PPh3 Bu4NOAc

1. HOAc 2. Ph 3P CBr4

H

Br

Ot-Bu

O

3. Bu 3SnH Pd(PPh3)4

H Br

170

Br

H H

Br Br

171; 41% yield

172; 59% yield

Now a second Heck reaction between the newly formed vinyl bromide 172 and the same alkene 169 gives the C2-symmetric compound 173 with all the same selectivity. Notice that vinyl bromides are more active in the Heck reactions than the aryl bromides present 169 and 172. However, when there are no more vinyl bromides left, the two aryl bromides do take part in a double intramolecular Heck reaction. Two more stereogenic centres are introduced and the alkene moves one place further round the ring. A special catalyst 175 is needed for this reaction that goes in excellent yield. Ot-Bu

Ot-Bu

170

169 Pd(OAc) 2 PPh3 Bu4NOAc K2CO3

175

H

H H

t-BuO

H

H

Br

Bu4NOAc

H

Br

H

H

H

H

173; 47% yield

174; 80% yield

t-BuO

This synthesis involves two Pd-catalysed reductions and four Heck reactions all using Pd(OAc)2 except the last. The yields are not always wonderful, but the synthesis is made very short and selective by the repeated Heck requirements.

O

O

P

Ar P

Pd

Pd Ar

Ar

O

O

Ar 175; Ar = o-Tolyl

Sp2 –sp2 Cross-Coupling Reactions by Transmetallation Halide and triflate leaving groups are also involved with palladium catalysis in the most general of all these reactions with vinyl electrophiles. These use a main group rather than transition metal, chiefly boron or tin, in stoichiometric amounts to mark one molecule as nucleophilic and then

324

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

transfer that group to the palladium atom by a transmetallation. Combination with an aryl or vinyl halide or triflate completes the synthesis. We can write a general scheme for this class of reaction in four stages: 1. Preparing the organometallic (usually boron 176 or tin 177) compound by various methods from the intended nucleophile. reagents

usually

X

or

Metal

B(OH)2 176

SnBu3 177

2. Oxidative addition of catalytic palladium (0) to the aryl or vinyl halide or triflate. L Br

L Pd

Pd(0)

Br

Br

L

Pd(0)

L Pd

178; Pd(II)

Br

179; Pd(II)

3. Transmetallation to give a Pd-C bond 180 and a B-(or Sn-) halide (shown with an aryl ‘nucleophile’ and vinyl ‘electrophile’). L L Metal

Br

L

L Pd

transmetallation

Pd 180; Pd(II)

The arrows on the transmetallation step are intended to indicate which group joins to which rather than a mechanism. We have seen other transmetallation reactions of this sort - Li exchanged for Cu or boron or tin for example in chapter 16 in particular. They may well go through a fourcentred transition state in which the aryl group (in this case) is bonded to both metals and the halide also. Organometallic mechanisms are not always known in detail. 4. Coupling on Pd to make the new C⫺C bond and to release the Pd(0) for the next cycle. L

L

L Pd

reductive coupling

L Pd L +

181; Pd(II)

182

183; product

L Pd

Pd(0) for next cycle

We have described the two components in the coupling as “nucleophilic’’ and “electrophilic’’. These words are intended to help you make sense of the reactions and should not be taken literally. Of course, when the two groups, vinyl, aryl or whatever, are both bound to the palladium atom, neither is nucleophilic or electrophilic (unless one of them is an unsaturated carbonyl compound) and their combination is a coupling reaction. The coupling occurs through a three-membered transition state 182 and there may be some involvement of the p-orbitals on the carbon atoms joined to palladium.

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

325

The designation “intended as the nucleophile” is helpful when considering the origin of the fragments. When making a disconnection with an sp2-sp2 cross-coupling in mind, one has to decide which molecule is to form the palladium σ-complex and that one is “intended as the electrophile”. If these labels do not help, discard them.

Stille coupling When an organotin compound is the nucleophile, the coupling is called Stille coupling after J. K. Stille of Colorado State University, the inventor.26 In chapter 15 we saw how organo-tin compounds are sources of radicals, but with palladium catalysis the reaction changes. We shall start with an example which may look trivial but is very important for this chapter. Reaction of readily prepared 1,1-dibromoalkenes 184 with tributyltin hydride under Pd(0) catalysis gives Z-1-bromoalkenes 185 with high stereoselectivity.27 RCHO

Ph3P, CBr 4 CH2Cl2

Br

R

Bu3SnH, 4 mol% Pd(PPh 3)4

H

R

room temperature 15 minutes

Br 184

+ Bu3SnBr

Br Z-185

The reaction starts with stage 2 (stage 1 is already done as we can buy the ‘organometallic’ Bu3SnH): oxidative insertion of Pd(0) into the less hindered of the two C-Br bonds of the dibromoalkene, that is the one trans to R to give the Pd(II) σ-complex 186. Notice that palladium is again strongly influenced by steric hindrance. Stage 3 is transmetallation – the palladium gets the H atom and the tin gets the Br atom 187. Most reactions of Bu3SnH end up with tin exchanging the weak Sn⫺H bond (∼310 kJ/mol) for the much stronger Sn⫺Br bond (∼380 kJ/mol). Stage 4 is the coupling of the vinyl group and the H atom - stereospecifically with retention. L 184

oxidative insertion

L Pd

Br

R

L Bu3SnH transmetallation

Br 186

L Pd

Bu3SnBr +

R

H

coupling

Z-185

Br 187

Palladium is released as Pd(0), ready for the next cycle. The relevance of this reaction to the subject of this chapter is that we have already used, and will use again, Z-bromoalkenes for the stereospecific synthesis of dienes. This is one way to make them. Carbon-carbon bonds can be made from palladium-catalysed reactions of vinyl bromides, iodides 189 or triflates with any organotin compounds which cannot undergo β-elimination after transfer to palladium. These include vinyl, methyl, allyl, benzyl, and aryl groups. This raises the question of which group is transferred from a tin atom in such reactions. With tetramethyl tin there is no choice, the methyl group being transferred to give 190 but aryl groups are transferred in preference to alkyl, giving 188.

Bu3SnI + R

Ar 188

Ar

SnBu3

cat. Pd(0)

R

I

Me4Sn cat. Pd(0)

189

R

Me

+ Me3SnI

190

When we want to transfer a group from tin we do not want to waste three of them. The rule is that tin transfers the best anion to palladium (strengthening our view that the organotin compound is the nucleophile). Alkynyl groups are transferred best and the order for other groups is roughly as below.

326

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

alkynyl

alkyl vinyl (alkenyl), aryl

R

allyl, benzyl

R

Me, Bu etc

R Ar

Ar

most easily transferred

least easily transferred

The exact order doesn’t matter. What is important is that aryl, vinyl, allyl and benzyl are all transferred more easily (and that means much more easily) than butyl or methyl. If we use a Bu3Sn– or an Me3Sn– group, we can be sure that no butyl or methyl groups will be transferred. A neat example is the preparation of aryl tin compounds 192 from aryl halides by Stille coupling. How is that possible? If one tin is to be transferred to carbon, another must pick up the halide so we need two tin atoms joined together – hexamethylditin 191 is the reagent. Ar

Br + Me3Sn SnMe3 191

Ar cat. Pd(0)

SnMe3 + Me3Sn 192

Br

If the electrophile is a vinyl triflate, it is essential to add LiCl to the reaction so that the chloride may displace triflate from the palladium σ-complex. Transmetallation takes place with chloride on palladium but not with triflate. This famous example illustrates the similar regioselectivity of enol triflate formation from ketones to that of silyl enol ether formation discussed in chapter 3. Kinetic conditions give the less 198 and thermodynamic conditions the more highly substituted 195 triflate. thermodynamic control O

OTf

1. i-Pr 2NMgBr 2. Tf 2NPh

193

SiMe3

SiMe3

Me3Sn

Pd(PPh3)4, LiCl

195; 90% yield; 98:2 regio

194

kinetic control O

OTf

1. i-Pr 2NLi (LDA) 2. Tf 2NPh

196

Me3Sn

SiMe3

SiMe3

Pd(PPh3)4, LiCl

197

198; 100% yield; 95:5 regio

Now let us look at a complete synthesis.28 The natural product pleraplysillin-1 199 is found in a marine sponge and in the nudibranch (a kind of apparently defenceless shell-less mollusc) that eats it. It has a defensive role - the nudibranch is “rapidly rejected as a food source by carnivorous fish”.29 Pleraplysilin-1 has a diene joined to a furan at the difficult 3-position. Disconnection between the two double bonds suggests a vinyl triflate 201 from a cyclic alkene, as we can make that regioselectively from the corresponding ketone, and a vinyl stannane from the furan half 200. TfO Stille

O

SnBu3

+

O 199; pleraplysillin-1

200

201

The enol triflate comes from regiospecific reduction of the enone (chapter 3) which can be made by a Robinson annelation.

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

O

O

O

FGA

327

CHO

aldol

1,5-di-CO

etc

201 203

202

204

3-Substituted furans are awkward to make but 3-hydroxymethylfuran 207 is available commercially and is converted into the corresponding bromide 205 with PBr3 in pyridine. Alkylation of a vinyl cuprate with this reactive alkyl bromide was used here.

SnBu3

Br

O

OH

+ Br

SnBu3

O 200a

O 205

206

207

The synthesis is summarised below. Note the use of a hexynyl group on the cuprate 208 - the rule for cuprates is the opposite of the rule for stannanes - the less stable anion is transferred first. The coupling requires LiCl and is then very efficient. LiO Cu

SnBu3

205

200

203

Tf2NPh

L-selectride

208

201

209 200 + 201

Pd(PPh3)4, LiCl

199, peraplysillin, 75% yield

The synthesis is convergent, very short, and gives just the one regiochemical and geometrical isomer of the diene. Another peraplysillin-1 synthesis is described in the workbook.

Recent developments in Stille coupling Because Stille coupling uses stoichiometric tin to control the regioselectivity, one might expect it to become rather less popular as concern over the environmental effects of toxic tin grows. In fact it is as popular as ever, particularly in laboratory syntheses. Two recent syntheses of crocacins both use the reaction to make the diene system. Crocacin C is a primary amide 210 while crocacin D has the more complicated side chain 211. The rest is the same. Me

O

Me CONHR

OMe OMe Me 210; Crocacin C: R = H

NH

N H

CO2Me

211; NHR group for Crocacin D

Both syntheses made the bond between the two alkenes by a Stille coupling. One put the tin on the amide part by a Cu(I) catalysed conjugate addition of Bu3SnLi to the acetylenic ester 212 and Weinreb amide formation. Coupling this vinyl stannane with a single enantiomer of the iodide derived from the rest of the molecule gave crocacin C in good yield. The synthesis of the iodide uses an asymmetric aldol reaction and is described in the workbook.30

328

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

Me

CO2Et 212

Bu3Sn

1. Bu 3SnLi CuBr.SMe 2 2. 1.7 equiv. MeOH

Bu3Sn

CO2Et

CONH2

Me3Al

Me 213; 70% yield, 87:13 E:Z

NH4Cl

Me 214; 72% yield, E only

The other synthesis31 had the tin on the larger part of the molecule 215. This starting material was also made by an asymmetric aldol reaction and the vinyl stannane was introduced by a special reaction described in the workbook. This is perhaps the more obvious way to do the coupling with the electrophilic halide combining with the vinyl stannane but these syntheses show that the Stille coupling is versatile. Me

I

Me

CO2Me Me

SnBu3

cat. Pd 2(dba)2

OMe OMe

Ph3As

215 Me

Me CO2Me

OMe OMe

Me

216; 80% yield

Variations in Stille coupling Other reasons for the continuing popularity of the Stille are the range of functional groups compatible with the reaction and the variety of structures that can be incorporated into either component. The allyl isoxazole 218 can be made either from vinyl or allyl-tributyltin using either 217 or 219 It turns out that vinylation of 217 gives a better yield.32

Br Bu3Sn

N

2.5 mol% Pd2(dba)3

Ph

O 217

Ph

(Furyl)3P

Ph

98% yield

Br

Bu3Sn

N Ph

O 218

N

2.5 mol% Pd2(dba)3 (Furyl)3P 49% yield

Ph

O 219

Ph

The ‘electrophile’ can include reagents such as acid chlorides, leading to acylation 221 and cyclisation to pyrones33 222 or bromo-quinones with similar reagents 220 leading eventually to chromenes.34 O Bu3Sn

O

O

RCOCl 5 mol% Pd(PPh3)4

OSnBu3 220; 85:15 E:Z

dioxane, 50 ˚C

OSnBu3

R 221

R

O 222

O

Double Stille couplings are valuable for symmetrical compounds like β-carotene 226. The symmetrical bis-tributylstannyl-pentaene 224 is coupled at both ends to the iodo-triene 223 in one step with catalyst 225 to give an excellent yield of β-carotene 226 with full control over E/Z stereochemistry.35

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

I

329

SnBu3 +

+ Bu Sn 3 223

I

224

223

(PhCN)2PdCl2 225

i-Pr2NEt, 25 ˚C

226; β-carotene, 73% yield

The synthesis of the complex anti-leukemia compound asperazine 231 uses a series of palladium-catalysed reactions including a Stille coupling and a Heck reaction as well as a palladium-catalysed hydrostannylation. This is an asymmetric synthesis as the starting materials are made from serine and tryptophan. We summarise only the key steps but the full description is worth reading.36

O NR O

H

NR I

R N

+ RN

H

SnBu3 227

H

R N

Stille coupling

CO2Me

O

CO2Me

H 229; >84% yield

O

H

O

O Heck Reaction O R N

I

R N

N R

R N

Pd2(dba)3.CHCl3

RN

O

(Furyl)3P, CuI

228

NR HN

RN

Pd2(dba)3.CHCl3

RN

(Furyl)3P, CuI

O

O H

230

H

231; 66% yield

Suzuki coupling In the Suzuki coupling37 the trialkyltin functional group on the “nucleophilic” partner in the coupling reaction is replaced by a boronic acid 234. These stable compounds are easily made from simple organometallic compounds and boronic esters 233 followed by hydrolysis. Oi-Pr Br

B 1. Li or Mg

R

2. (i-PrO)3B

232

R

233

Oi-Pr

OH B HCl

R

234

OH

330

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

A standard way to make vinyl boronic acids 237 is to hydroborate an alkyne with catechol borane38 235 and again hydrolyse the ester products 236. The empty orbital of the boron atom attacks the electron-rich terminus of the alkyne where is the larger HOMO of the π-bond. One detailed example39 includes Suzuki coupling to Z-styryl bromide. Another diester group commonly used in the Suzuki coupling is derived from pinacol - examples appear later. R

syn hydroboration

O H

OH

O B

B

O

R

O 235; catechol borane

B R

236; vinyl boronic ester

OH

237; vinyl boronic acid

There are many other ways of making boronic acids and you are referred to Suzuki’s review37 for a full discussion. A Suzuki coupling would occur if one of these boron compounds 236 was combined with an aryl or vinyl halide or triflate in the presence of Pd(0) and an oxy-anion. O R O 236; vinyl boronic ester

O

catalytic Pd(PPh3)4

+ I

B

RO

238

+ RO R

B O 240

239

The first step in a Suzuki coupling is the formation of the vinyl or aryl boron derivative and this we have already discussed. The second step is the oxidative insertion of Pd(0) into the other partner in the coupling reaction, here iodobenzene 238. This σ-complex 241 forms without the oxyanion. It is in the third step, the transmetallation reaction 243, that the oxyanion is necessary. Boron is barely a metal and it forms strong B⫺O bonds. It will not exchange its organic group with the palladium atom unless it gets an oxygen atom in return. R

O Pd(PPh3)4

238

oxidative insertion

I

RO

Pd

Ph3P

PPh3 241

transmetallation

B RO

Pd

Ph3P

PPh3 242

O

RO

Pd

Ph3P 243

PPh3

239

Pd(PPh3)2

The dotted arrows on the transmetallation step 243 show only what joins to what and are not intended as a serious mechanism. Indeed a better mechanism might involve addition of RO⫺ to the boron atom before transmetallation. This process can be used to couple aryl to aryl, vinyl to vinyl, and aryl to vinyl (either way round!). As boron compounds are much less toxic than tin compounds, the Suzuki coupling is often preferred industrially. Because each partner in the coupling reaction is marked in a different way – one with a boron atom and one with a halide – we can be sure that we shall get cross coupling reactions only. When the bond between the two sp2 centres is in the middle of the molecule, the disconnection becomes very attractive strategically. Markós milbemycin synthesis40 illustrates this point well. Milbemycins belong to a family of anti-parasitic agents and milbemycin β3 244 is one of the simplest. Disconnection of the macrocyclic lactone a reveals a continuous carbon skeleton with an aromatic ring conjugated to a diene and one isolated alkene. Markó next decided to disconnect the isolated alkene b, having in mind a Wittig reaction with 246 on the ketone 245 and it is this ketone we are going to analyse in detail.

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

b

331

O O

O

PPh3 O

O

a C–O

O a

b Wittig

O CO2H

lactone

+ OH

OH 244; Milbemycin β3

OH 245

246

Removing the aromatic ring from the diene makes good sense once we know about Suzuki coupling as it allows us to build the two halves of the molecule separately. There is a free choice as to whether we put the boron atom on the diene or on the aromatic ring, but Markó put it on the diene because he planned to make the vinyl boronate 248 by hydroboration of an alkyne 249. Note the protection of the phenol and the carboxylic acid as methyl ether and ester respectively.

O

O Suzuki coupling

CO2H

CO2H

H

+

X

B(OR)2 248

OH

O

hydroboration

249

OH 247

245a

No doubt joining the alkyne to the alkene could also have been done by a coupling reaction in the coordination sphere of a metal but an alternative is to imagine the alkene as coming from the dehydration of an alcohol 251. This allows disconnection to the known lactone 250. The synthesis of the alkyne uses DIBAL for partial reduction and the differential protection of the two OH groups by more or less hindered silyl groups.

1. DIBAL

O

OH

2. BrMg

OTBDMS

2. TBDMSiCl Et3N, DMAP

O 250 known lactone

1. TBDPSiCl Et3N, DMAP

HO 251; 85% yield

TBDPSO 252; 70% yield

Hydroboration of the alkyne 252 with catechol borane 235 is regioselective for steric reasons and the resulting E-vinyl boronate 253 couples with the iodo-benzene 254 under palladium catalysis.

332

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

The coupling was followed, without isolation, by a Jones oxidation to give the required protected hydroxyketone. Half of milbemycin β3 had been assembled with good stereochemical control thanks to Suzuki coupling.

OTBDMS

O

CO2Me

H

235 TBDPSO 252

+

I

1. Pd(OAc) 2 Ph3P, Tl 2CO3 2. KF, H 2O CrO3, H 2SO4

CO2Me

TBDPSO

B(OR)2 OMe 253; 100% yield

OMe 255; 68% yield from 253

254

All the examples we have chosen so far have been of aryl and vinyl nucleophiles with no possibility of β-elimination when they are transferred to palladium before coupling. With Suzuki coupling it is possible to use saturated alkyl boranes in combination with vinyl halides and get coupling products in good yield. An example is the preparation41 of exo-brevicomin 256, the aggregation pheromone of the Western pine beetle Dendroctonus brevicomis. Almost all syntheses, including this one start with the unravelling of the acetal to reveal a ketodiol 257. The 1,2-diol is syn as drawn and can be made diastereoselectively by cis dihydroxylation of an alkene of the same configuration, that is E-258. The reagent normally used in this reaction is OsO4 with various catalysts. In chapter 25 you will see that this reaction can be made to produce just one enantiomer of the syn diol. H O

H 1,1-diO acetal

O

HO

256; (+)-brevicomin

H

OH

O

1,2-diO

O

functionalisation

H

H 257

E-258

Using a Suzuki disconnection, we intend to add an alkyl boronic acid 260 to an E-vinyl bromide 259 (from but-1-yne) and the boronic acid will come from hydroboration of a simple alkene 261 rather than an alkyne.

H

H E-258a

O

H O

Suzuki

O

+ Br H E-259 from alkyne

B(OH)2 260

261

Each stage of this synthesis uses interesting reactions. The unsaturated ketone was made by allylation of a d1 equivalent – a methoxyvinylcuprate from 263 and it was necessary to protect the ketone as the acetal 265 during the rest of the reactions.

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

MeO

OMe

MeO

1. BuLi

Li

262

333

2. Cu(I)

HO

Br

3.

263

O

OH H

264

O

265

The E-butenyl bromide 259 was made by hydroalumination and bromination of but-1-yne 266. Hydroalumination occurs stereospecifically cis (chapter 16) 267 and electrophilic replacement of aluminium by bromine occurs with retention at the sp2 centre.42 H

H Br2

i-Bu2AlH

H 266; but-1-yne

[DIBAL]

H

Ali-Bu2 E-267

Br H E-259

Hydroboration of the unsaturated ketone was carried out with the very hindered borane 9BBN and the alkyl borane 268 was coupled with the vinyl bromide 259 without isolation using Pd(0) as catalyst and NaOH to force the transfer of the alkyl group from boron to palladium. The pure trans (E-) coupled product 269 is formed in 85% yield so coupling must be faster than βelimination in this case. H O

O

9-BBN

O

R2B

O

E-259

NaOH

265

O

Pd(PPh3)4

O

H E-269; 85% yield

268

Asymmetric dihydroxylation with OsO4 gave the syn diol 270 which was deprotected and cyclised with toluene-p-sulfonic acid (TsOH) catalysis in over 90% yield to complete this short and interesting stereochemically controlled synthesis of exo-brevicomin 256 in 65% yield from allyl bromide. H

H O H

O

OsO4

H OH O

HO

O

TsOH

O

H

E-269

O

syn-270

256; (+)-brevicomin

Recent developments in Suzuki coupling You will have realised that the Suzuki coupling is the best of those we have mentioned. It is regiospecific as the positions for coupling are marked, one with a boronic acid and one with a halide. It does not use toxic tin. It can be used for all the aryl, heteroaryl, vinyl, benzyl and other groups that Heck and Stille use and in addition it can be used for saturated alkyl boronic acids with β-hydrogen atoms without β-elimination. It is not surprising that it is very widely used and we give a few representative examples. Direct coupling of aryl with heteroaryl compounds works well with o-Cl and o-MeO groups on the phenylboronic acid in coupling with isoquinolines or coumarins. In the first case 271 it was

334

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

necessary to have a bromine (rather than a chlorine) atom on the isoquinoline to ensure chemoselectivity as there is a chlorine atom on the boronic acid.43 In the second 272 a chlorine is all right but this is an easier reaction as it is like conjugate substitution.44 CO2Me

CO2Me

N Br

MeO

N

Pd(PPh3)4

B(OH)2

K3PO4

Cl

B(OH)2

Pd(PPh3)4

MeO

K2CO3

CHO

Cl

Cl

CHO 271; 90% yield

O 272; 98% yield

O

Heterocycles are also coupled to vinyl boronic acids and these can be made by hydroboration of alkynes. This example45 shows the compatibility with functionality such as the NHBoc group in 273.

1.

Ipc2BH

O B

NHBoc

2. 30 x MeCHO 3. pinacol

N Tr

273

N

NHBoc

Pd(OAc) 2

+

O

I

N

NHBoc

N

(o-Tol) 3P K2CO3

Tr

274

275; 51% yield

Direct coupling of two vinyl groups remains an important job for the Suzuki reaction. In their synthesis of the antibiotic (⫹)-fostriecin, Reddy and Falck46 coupled the Z-vinyl bromide 276 with the Z-boronic ester 277 to make the triene 278 with Z,Z,E-alkenes. Free OH groups obviously do not interfere.

OH

Br

OR

+ O O

Pd(PPh3)4

B O

O

Ag2O

OR OH 276

277 OH

O

OR

O OH

OR

278; 74% yield

The most distinctive contribution of the Suzuki to modern methodology is the coupling of alkyl boronic acids, formed by hydroboration of alkenes, to vinyl and aryl halides.47 The simple mono-substituted alkene 279 reacts with 9-BBN followed by Suzuki coupling with aryl iodides to give products 281 in good yield regardless of the nature or the position of the substituents.48

18 Sp2–sp2 Cross-Coupling Reactions by Transmetallation

335

BR2

Ar = phenyl substituted:

ArI, 5 mol%

2 equivs 9-BBN-H

BocN

Ar

PdCl2(dppf).CHCl3

BocN

THF

O

O

279

BocN

3M K3PO4 THF-DMF

p-OMe, 71% yield m-NO2, 69% yield o-F, 77% yield

O

280

281

Intramolecular coupling of a Z-vinyl iodide and a saturated alkyl chain in 282 creates a new 11-membered ring 283 providing the ‘ansa’ bridge in Danishefsky’s synthesis49 of xestocyclamine A. The yield in the formation of this difficult medium ring and the toleration of functionality are both impressive.

HO

HO

OR

OR Ts

1. 9-BBN 2. H 2O

N

N O

Ts

N O

3. Pd(dppf)Cl 2 TlCO3, Ph 3As

282

I

N

283; 60% yield

The coupling of two enantiomerically pure fragments to form an alkene of fixed geometry in the centre of ebelactone allows a convergent synthesis.50 One fragment, the vinyl iodide 285 is ultimately derived from a coupling between a chiral aldehyde and a chiral allyl boronate but more immediately by silyl-cupration of an alkyne 284 and iodination.

1. PhMe 2SiLi, CuCN 2. N-I-succinimide

OR1

I

OR2

OR1

284

OR2

285; 90% yield

The other fragment, the alkene 286 originates in an Evans’ aldol reaction (chapter 27) and is coupled to 285 by hydroboration and Suzuki reaction. The yield is very impressive, as is the stereocontrol over the new centre next to the acetal. The full synthesis appears in the workbook together with some stereochemical explanations. BnO

BnO 1. 9-BBN, THF room temperature

O

O 286

2. 285 Pd(dppf)Cl2 NaOH, H 2O room temperature

O

O

OR1 287; 70% yield

OR2

336

18 Vinyl Cations: Palladium-Catalysed C–C Coupling

Summary The palladium-catalysed reactions with alkenyl, aryl, and heteroaryl halides or triflates as one partner (‘electrophilic’) and alkenes (Heck), aryl or vinyl stannanes (Stille) or aryl, vinyl and even alkyl boronic acids (Suzuki) as the other (‘nucleophilic’) partner provide a synthetic method of astonishing power and versatility. These reactions are only just starting to be explored and great things are expected of them.

References General references are given on page 893 1. Z. Rappoport, Acc. Chem. Res., 1981, 14, 7. 2. M. Ochiai, S. Yamamoto and K. Sato, Chem. Commun., 1999, 1363. 3. M. Tiecco, L. Testaferri, M. Tingoli, D. Chiantelli and M. Montanucci, J. Org. Chem., 1983, 48, 4785. 4. G. Bradley, J. F. Cavalla, T. Edington, R. G. Shepherd, A. C. White, B. J. Bushell, J. R. Johnson and G. O. Weston, Eur. J. Med. Chem., 1980, 15, 375. 5. F. J. Moreno-Dorado, F. M. Guerra, F. J. Aladro, J. M. Bustamante, Z. D. Jorge and G. M. Massanet, Tetrahedron, 1999, 55, 6997. 6. K. Chen, C. S. Brook and A. B. Smith, Org. Synth., 1998, 75, 189; A. B. Smith, B. D. Dorsey, M. Ohba, A. T. Lupo and M. S. Malamas, J. Org. Chem., 1988, 53, 4314. 7. A. B. Smith, B. D. Dorsey, M. Visnick, T. Maeda and M. S. Malamas, J. Am. Chem. Soc., 1986, 108, 3110. 8. R. F. Abdulla and K. H. Fuhr, J. Org. Chem., 1978, 43, 4248. 9. T. Mukaiyama and T. Ohsumi, Chem. Lett., 1983, 875. 10. G. Bartoli, C. Cimarelli, E. Marcantoni, G. Palmieri and M. Petrini, J. Chem. Soc., Chem. Commun., 1994, 715. 11. J. R. Weir, B. A. Patel and R. F. Heck, J. Org. Chem., 1980, 45, 4926. 12. D. Ma, Z. Ma, J. Jiang and C. Zheng, Tetrahedron: Asymmetry, 1997, 8, 889. 13. R. K. Haynes, S. C. Vonwiller, J. P. Stokes and L. M. Merlino, Austr. J. Chem., 1988, 41, 881. 14. M. Asaoka, K. Takenouchi and H. Takei, Tetrahedron Lett., 1988, 29, 325. 15. K. Hayakawa, M. Yodo, S. Ohsuki and K. Kanematsu, J. Am. Chem. Soc., 1984, 106, 6735. 16. R. K. Haynes, S. C. Vonwiller and T. W. Hambley, J. Org. Chem., 1989, 54, 5162. 17. C. Zhang and M. L. Trudell, Tetrahedron, 1998, 54, 8349. 18. E. J. Corey, P. D. S. Jardine and J. C. Rohloff, J. Am. Chem. Soc., 1988, 110, 3672; E. J. Corey and P. D. S. Jardine, Tetrahedron Lett., 1989, 30, 7297. 19. R. F. Heck, Org. React., 1982, 27, 345; R. F. Heck, Palladium Reagents in Organic Syntheses, Academic Press, London, 1985; J. T. Link, Org. React., 2002, 60, 157. 20. T. Jeffery, Tetrahedron Lett., 1985, 26, 2667. 21. J. W. Raggon and W. M. Snyder, Org. Process Res. Dev., 2002, 6, 67. 22. H. P. Dang and G. Linstrumelle, Tetrahedron Lett., 1978, 191; D. Michelot, Synthesis, 1983, 130. 23. M. Nakanishi and M. Mori, Angew. Chem., Int. Ed., 2002, 41, 1934. 24. M. Lautens, J.-F. Paquin and S. Piguel, J. Org. Chem., 2002, 67, 3972. 25. L. F. Tietze and W.-R. Krahnert, Chem. Eur. J., 2002, 8, 2116. 26. J. K. Stille, Angew. Chem., Int. Ed., 1986, 25, 508; V. Farina, V. Krishnamurthy and W. J. Scott, Org. React., 1997, 50, 1. 27. J. Uenishi, R. Kawahama, Y. Shiga and O. Yonemitsu, Tetrahedron Lett., 1996, 37, 6759. 28. W. J. Scott, G. T. Crisp and J. K. Stille, J. Am. Chem. Soc., 1984, 106, 4630. 29. J. E. Thompson, R. P. Walker, S. J. Wratten and D. J. Faulkner, Tetrahedron, 1982, 38, 1865. 30. L. C. Dias and L. G. de Oliveira, Org. Lett., 2001, 3, 3951. 31. J. T. Feutrill, M. J. Lilly and M. A. Rizzacasa, Org. Lett., 2002, 4, 525. 32. B. Clapham and A. J. Sutherland, J. Org. Chem., 2001, 66, 9033. 33. J. Thibonnet, M. Abarbri, J.-L. Parrain and A. Duchêne, J. Org. Chem., 2002, 67, 3941.

18 References

337

34. K. A. Parker and T. L. Mindt, Org. Lett., 2001, 3, 3875. 35. B. Vaz, R. Alvarez and A. R. de Lera, J. Org. Chem., 2002, 67, 5040. 36. S. P. Govek and L. E. Overman, J. Am. Chem. Soc., 2001, 123, 9468. 37. N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2457. 38. H. C. Brown and S. K. Gupta, J. Am. Chem. Soc., 1972, 94, 4370. 39. N. Miyaura and A. Suzuki, Org. Synth., 1989, 68, 130. 40. I. E. Markó, F. Murphy and S. Dolan, Tetrahedron Lett., 1996, 37, 2507. 41. J. A. Soderquist and A. M. Rane, Tetrahedron Lett., 1993, 34, 5031. 42. G. Zweifel and W. Lewis, J. Org. Chem., 1978, 43, 2739. 43. Y. L. Janin, E. Roulland, A. Beurdeley-Thomas, D. Decaudin, C. Monneret and M.-F. Poupon, J. Chem. Soc., Perkin Trans. 1, 2002, 529. 44. S. Hesse and G. Kirsch, Tetrahedron Lett., 2002, 43, 1213. 45. F. Berrée, P. Girard-LeBleis and B. Carboni, Tetrahedron Lett., 2002, 43, 4935. 46. Y. K. Reddy and J. R. Falck, Org. Lett., 2002, 4, 969. 47. S. R. Chemler, D. Trauner and S. J. Danishefsky, Angew. Chem., Int. Ed., 2001, 40, 4544. 48. P. N. Collier, A. D. Campbell, I. Patel, T. M. Raynham and R. J. K. Taylor, J. Org. Chem., 2002, 67, 1802. 49. A. Gagnon and S. J. Danishefsky, Angew. Chem., Int. Ed., 2002, 41, 1581. 50. A. K. Mandal, Org. Lett., 2002, 4, 2043.

19

Allyl Alcohols: Allyl Cation Equivalents (and More) This chapter analyses the problems associated with allylic electrophiles and the unique usefulness of allylic alcohols in their solution. The preparation and other chemistry of allylic alcohols is here too.

PART I – INTRODUCTION The Problem of the [1,3]-Shift Nucleophilic attack on unsymmetrical allylic halides Rearrangement and stability of allylic alcohols PART II – PREPARATION OF ALLYLIC ALCOHOLS Traditional Methods Reduction of enones prepared by the aldol or Wittig reactions Elimination on or reduction of iodolactonisation products Addition of vinyl metals to aldehydes and ketones Reduction of acetylenic alcohols made by addition of alkynes to aldehydes or ketones Wittig examples Allylic Alcohols by [2,3] Sigmatropic Shifts Sulfoxides and sulfonium ylids: Regio- and stereochemical control [2,3] Sigmatropic shifts in the allylation of ketones Recent applications PART III – REACTIONS OF ALLYLIC ALCOHOLS [2,3] Sigmatropic Rearrangements Reminder of the conversion to allylic sulfoxides The [2,3] Wittig rearrangement Reactions with Electrophiles The Simmons-Smith cyclopropanation reaction Stereochemically controlled epoxidations Regio- and Stereocontrolled Reactions with Nucleophiles Claisen-Cope rearrangements Stereochemistry in the Claisen-Cope rearrangement The Claisen-Ireland rearrangement Pd-catalysed reactions of allylic alcohols Pd-allyl acetate complexes Stereochemistry of Pd-allyl cation complexes Pd and monoepoxides of dienes The control of remote chirality Recent developments Summary Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

340

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

PART I – INTRODUCTION The Problem of the [1,3]-Shift Nucleophilic attack on unsymmetrical allylic halides At the start of chapter 12 (Allyl Anions) we discussed briefly the general problem that faces any use of allylic derivatives in synthesis. Many of them are not stable with respect to the [1,3]-shift or allylic rearrangement. The bromides 1 and 3 are in equilibrium under most conditions: in polar solvents they equilibrate via the allylic cation 2 to give mostly the compound 3 with the more highly substituted alkene.1 Br Br 1; 13-20%

3; 80-87%

2

This does not prevent the use of allylic bromides in synthesis. As nucleophiles may attack an allylic derivative from either end of the allylic system, there is a further opportunity to get the product with the more highly substituted double bond. You will probably know that prenylation of nucleophiles, particularly enolates, gives overwhelmingly the prenyl derivative 5. The primary allylic halide 6 (prenyl bromide) reacts by the SN2 reaction while the tertiary halide 4 reacts by the SN2’ reaction. Br

Nu

SN2'

Nu

SN2

Nu 4

Br

5

6

This chapter is about ways to get greater control over processes like these. It concerns ways to make allylic derivatives react at the more substituted end to give the less stable product, ways to control allylic substitution when both possible products have equal numbers of substituents on the alkene and ways to control the stereochemistry. The key reagents in this chapter are allylic alcohols. Though these do equilibrate in acid solution, under other conditions they are stable and they are easily prepared regio- and stereospecifically by many ways. Strictly ‘allyl’ refers to the prop-2-enyl group CH2 =CH-CH2- though nowadays it usually means any group next to an alkene but not directly attached to it. We shall use ‘allylic’ rather than ‘allyl’ for that meaning though the distinction is on the way out.

Rearrangement and stability of allylic alcohols The mixture of 1 and 3 can be made by treatment of either alcohol 7 or 8 with HBr. Under these strongly acidic conditions equilibration between the alcohols as well as between the allylic bromides occurs through the allylic cation 2.

1; 13-20% +3; 80-87%

OH HBr

H

H

HBr

OH 7

2

8

1; 13-20% +3; 80-87%

19 Traditional Methods

341

This can be useful when oxidation of an equilibrating alcohol is carried out in acidic conditions, say with a Cr(VI) reagent. The first formed alcohol 10 cannot be oxidised but the rearranged alcohol 11 can. This rearranged alcohol might be formed via an allylic cation or it might be by [3,3]-sigmatropic rearrangement of the chromate esters 13. O

R

R

OH

RMgX or RLi

R

OH O Cr O

Cr(VI)

H

O

OH 9

R

10

O

11

12

13

Under other conditions such as base or ionic solvents or radical formation, 1,3-shifts do not occur either with the alcohols themselves nor with simple derivatives like carboxylate esters. They are therefore ideal for use as specific allylic electrophiles except for the fact that OH – is a terrible leaving group and simple nucleophilic displacements on allylic alcohols are unknown. Special chemistry will be needed.

PART II – PREPARATION OF ALLYLIC ALCOHOLS Traditional Methods Reduction of enones prepared by the aldol or Wittig reactions Types of reactions we meet elsewhere in the book include the reduction of α,β-unsaturated carbonyl compounds prepared by the aldol or Wittig reactions. The cyclic enone 14, prepared by a Robinson annelation is reduced regio- and stereo-selectively by LiAlH4 to the allylic alcohol2 15. The reduction of acetylenic alcohols 16, prepared by addition of metallo-alkynes to aldehydes, also with LiAlH4 is E-selective (chapter 15) giving the allylic alcohol3 E-17. R1

LiAlH4

O

–20 ˚C

LiAlH4

R2

OH

R2

R1 OH

OH 14

15

E-17

16

Elimination on or reduction of iodolactonisation product Slightly more unusual is the elimination and hydrolysis or reduction of iodolactonisation products 20 (chapter 17). Elimination replaces the alkene but in a different position 20 and cleavage of the lactone bridge4 by methanolysis gives 21 or by reduction gives 22. As we get on to the reactions of allyl alcohols you will see how these and other methods are used to make particular compounds. O

O

O CO2H

O

I2

OH HO

DBU

CO2Me HO

NaHCO3

I 18

19

20

21

22

342

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

Addition of vinyl metals to aldehydes and ketones Addition of vinyl metals to aldehydes and ketones (chapter 16) also allows control over stereochemistry. The stereospecific vinyl anion equivalent 24 is one we made in chapter 16 by the Shapiro reaction. We repeat this scheme because it uses prenylation of a hydrazone aza-enolate to make the starting material for the vinyl-tin compound 25. Li

N

SnBu3

NTr

N

NHTr

1. s-BuLi

1. 2 x BuLi

2. Bu 3SnCl

2. Br

23

25; 90% yield

24

Conversion of the vinyl tin 25 to the vinyl lithium 26 and coupling with the keto-epoxide 27 gives one diastereoisomer of the allylic alcohol 28 (17:1), used in a synthesis of ovalicin.5 Notice that the vinyl-lithium 26 is one geometrical isomer and that it attacks the carbonyl group of 27 on the same face as the oxygen atom of the epoxide. This looks like chelation control. O

Li BuLi

O

O

OH

+

25

OMe 27

26

OMe 28

Reduction of acetylenic alcohols made by addition of alkynes to aldehydes or ketones Conjugate addition of the complete allylic alcohol fragment is possible with the mixed cuprate reagents 33 prepared by asymmetric reduction (chapter 26) of acetylenic ketones 29 to give the alcohols 30, protection as a silyl ether 31 and hydroboration-iodination. Lithiation and reaction with hexynyl copper (I) gives the mixed cuprate 33 from which the less stable anion is transferred selectively to an enone.3 This approach has been widely used in the synthesis of prostaglandins. asymmetric reduction

O

R

1. 2 x t-BuLi, Et 2O 2. Bu

OTBDMS 32

TBDMSCl DMAP, DMF

OH 30

29 I

R

R OTBDMS 31 O

Cu

OTBDMS 33

2. I 2, Me 3NO

R

Cu

Bu

1. 9-BBN

cyclohexenone

R 34

OTBDMS

Wittig examples Two simple examples of an E-selective Wittig reaction followed by reduction of the carbonyl group giving allylic alcohols with full control over double bond geometry are given in chapter 15 and the following section in that chapter on Z-selective Wittig variants shows how Wittig-style reactions followed by reduction can be used to give Z-allylic alcohols. The methods used to make enones

19 Allylic Alcohols by [2,3] Sigmatropic Shifts

343

in chapters 2-4 and summarised in chapters 5 and 6 are also methods to make allylic alcohols as reduction with anionic reducing agents (e.g. NaBH4 or LiAlH4) is usually regioselective. One special case with a non-stabilised Wittig reaction is the remarkably Z-selective reaction6 with THP-protected α-hydroxyketones 37. These are among the most Z-selective Wittig reactions known and it is the influence of the α-OTHP group that is decisive. The product 38 is of course a protected allylic alcohol. In chapter 33 we shall see how to make these α-hydroxyketones stereo- and regiospecifically. O (Me3Si)2NK

Ph3P

Ph3P

Me

O

OTHP

+

O

Me

35

36

38; 95% yield, >200:1 Z:E

37

All these methods give allylic alcohols regiospecifically, that is with the OH group and the alkene in predictable positions. They are often stereoselective for the alkene and may also be enantioselective as we shall see later. Provided the allylic alcohol is kept away from strong acid, its structure is more secure than that of most allylic compounds.

Allylic Alcohols by [2,3] Sigmatropic Shifts Sulfoxide and sulfonium ylids: Regio- and stereochemical control We shall be using [2,3] sigmatropic shifts in the reactions of allylic alcohols so this section is an introduction to that as well as describing perhaps the most tightly controlled method for making allylic alcohols. Allylic sulfides 40 are easily oxidised chemoselectively with reagents such as sodium periodate to the corresponding sulfoxides 41. More powerful oxidants tend to form the sulfone 42. O Br

PhSH

R

PhS

R

NaIO4

O O

PhS

R

Ph

NaOH

39

40

41

S

R 42

These allylic sulfoxides 41 are in equilibrium with a sulfenate ester 43 by a [2,3]-sigmatropic rearrangement 41a. It is not usually possible to detect the sulfenate ester by NMR so there must be less than about 3% of it, but it can be trapped by various nucleophiles that like to attack sulfur. These ‘thiophiles’ include secondary amines, thiolate anions and, most important, phosphite esters. The reaction is carried out in a protic solvent (usually the alcohol already present in the phosphite ester) and a rearranged allylic alcohol 45 is formed. R

O Ph

S

[2,3]

Ph

S

O

R

R

O

R

HO MeOH

(MeO)3P 41a

43

44

45

Since the original preparation of the allylic sulfide involved nucleophilic attack on an allylic halide 39, the more stable allylic sulfide - the one with the more highly substituted alkene 40 - is formed and this is inevitably and usefully transformed into the less stable allylic alcohol 45. The reaction is more useful still because the intermediate allylic sulfoxide 41 may be alkylated before rearrangement and this gives the unsymmetrically substituted E-allylic alcohols 47. These reactions

344

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

are exceptionally E-selective and we shall delay an explanation of this stereoselectivity until we have discussed the use of such allylic alcohols in [3,3] sigmatropic rearrangements later in the chapter. O R1

PhS

O

1. BuLi 2. R 2Br

(MeO)3P

R1

PhS

R1

R2

MeOH

OH

R2 46

41

47

An example is the synthesis of nuciferal by David Evans and group.7 Symmetrical allylic halides like ‘methallyl’ bromide 48 can give only one product on reaction with nucleophiles. Alkylation of the sulfoxide with a large alkyl iodide gives the allylic sulfoxide 49. The [2,3] sigmatropic rearrangement gives a single geometrical isomer of the primary allylic alcohol 50: the corresponding aldehyde, formed in 99% yield on oxidation of 50 with MnO2 is the natural product nuciferal. Both regio- and stereochemical control are complete.

1. PhSH, NaOH

Br

Ph

2. NaIO 4

S

3. BuLi

(MeO)3P

O

MeOH

4. RI

OH

48

49

50

These [2,3] sigmatropic rearrangements on sulfur compounds are not restricted to sulfoxides as they occur equally with sulfonium ylids. A good example showing both is the synthesis of yomogi alcohol 51, an irregular monoterpene. Reversing the [2,3] sulfoxide rearrangement gives an allylic sulfide 52 that we should like to make by alkylation of the simple prenyl sulfide 54 with the allylic halide 53. However, we know that won’t work: the alkylation will occur at the less hindered end of the allylic halide. OH

[2,3]

?

+ Br PhS 52

51; yomogi alcohol

PhS

53

54

The solution is surprising. Evans deliberately made the wrong doubly allylic sulfide 55 and rearranged it twice to give yomogi alcohol. At first the sulfur atom must be alkylated (therefore adding a phenyl group is not suitable and a methyl group is used instead). Then the first [2,3] rearrangement (on the sulfonium ylid) 56 gives the right skeleton 57 and the second (on the sulfoxide 58) gives the target molecule. In this reaction, the older thiophile EtNH2 was used.8

Na2S

S

6; prenyl bromide

2. BuLi

55

S Me 56; sulfonium ylid

EtNH2

NaIO4

SMe 57

[2,3]

1. MeI

Br

Me

S 58

OH

MeOH

O 51; 69% yield

19 Allylic Alcohols by [2,3] Sigmatropic Shifts

345

[2,3] Sigmatropic shifts in the allylation of ketones The allylation of ketones is a simple reaction. Most of the specific enol equivalents described in chapters 2-6 react well with allylic halides. Disconnection of the γ,δ-unsaturated ketone 59 to the enolate 61 and the allylic halide 60 is trivial and the reaction of, say, an enamine will occur at the less hindered primary centre to give 59. However, supposing the target molecule is the isomeric ketone 63. The same disconnection gives the isomeric allylic bromide 62 which will almost certainly give the alternative product 59 by reaction at the less hindered end of the alkene. O

Br

O

R2

R1

R2

59

+

+

R1

60

R1

Br 61

R2

O

R2

62

63

One solution is a [2,3] sigmatropic rearrangement of a sulfonium ylid. A stable allylic thiol reacts cleanly with an α-haloketone 64 to give a sulfide and hence the sulfonium salt after ethylation. Only a weak base is need to make the ylid 65 as this is also an enolate. Notice the regioselectivity here: the alternative ylid would also be stabilised but only by an alkene. The [2,3] shift joins the allylic fragment to the enolate of the ketone and at the same time turns it inside out. This automatically produces the more difficult product 63 after desulfurisation with Raney nickel. We shall see alternative solutions to this problem later in the chapter.9 O

SEt

O 1. HS

Ph Br 64

Ph

Ph

Ph

2. Et 3O+ BF4–

[2,3]

Et

3. K 2CO3, EtOH

Ph

S

Ph

EtOH

O

65

Raney Ni

63

66

Recent applications The preparation of the allylic alcohol 68 by Horner-Wadsworth-Emmons reaction followed by reduction illustrate how this method is compatible with varied functionality and stereochemistry. BnO

Me

CHO N O 67

O

base

(EtO)2P

CO2Me

1.

O

BnO

Me 2. DIBAL 3. RCl, NaH

N O 68

O OR

The diene alcohol derivative 68 is used to prepare the starting material 69 for an intramolecular hetero-Diels-Alder reaction to give a new heterocyclic ring 70 that can be cleaved with phenyl Grignard to release a sulfoxide 71 for the preparation of a new allylic alcohol 72. Notice that the stereochemistry of the sulfoxide is shown and that there is complete control over 2D and 3D stereochemistry. The allylic alcohol 72 was used in Weinreb’s synthesis of toxins10 produced by fresh water blue-green algae.

346

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

BnO

BnO

H

H

N

N

68 N

Me

N O 69

BnO H

BnO

Me

S OR

O

BnO

H

S

O 70

BnO

OR

O

H

H

N

NH

OH

(MeO)3P

PhMgBr

N

Me

NH PhS O 71

BnO

MeOH

Me

OR

O

OR

O 72

BnO

PART III – REACTIONS OF ALLYLIC ALCOHOLS [2,3] Sigmatropic Rearrangements Reminder of the conversion to allylic sulfoxides You saw in the last section that the preparation of allylic sulfides by displacement of, say, a halide from an allylic system is likely to give the more stable allylic sulfide whichever allylic halide is used and that the corresponding allylic sulfoxides are made by oxidation of the sulfide. The same is true of phosphorus compounds 73 used in chapter 15 to make dienes of fixed configuration. If R = OR, 73 is an allylic phosphonate for the Horner-Wadsworth-Emmons reaction but if R = Ph, 73 is a phosphine oxide for the Wittig-Horner reaction. Br

R2P

OMe

R1

R2P

R1

R2P

O

39

R1

OMe

Br

73

74

The specific preparation of either phosphine oxide 73 or 80 can easily be carried out from the appropriate allylic alcohol. The reaction is like the one we saw for the preparation of allylic alcohols from allylic sulfoxides but the [2,3]-sigmatropic rearrangements 76 and 78 run only in this direction with phosphorus.11 It is also the preferred direction with sulfur. Since allylic alcohols 75 and 78 are stable to the allylic rearrangement, each alcohol gives specifically one allylic phosphine oxide 73 or 80 or allylic sulfoxide 41 or 77. This is the theme for most of the rest of the chapter. O

SPh

1. PhSCl pyridine 2. heat

R1

OH

Ph2PCl pyridine

R1

41

O

1. PhSCl pyridine

SPh 77 O

Ph2PCl

2. heat

R1 78

R1 76

OH

PPh2

[2,3]

R2

75

R1

O

PPh2

73 R1

PPh2 O

[2,3]

pyridine

R1 79

PPh2 80 O

19 [2,3] Sigmatropic Rearrangements

347

The [2,3] Wittig rearrangement Ethers of allylic alcohols also undergo a [2,3] sigmatropic rearrangement. The driving force is the conversion of an unstable C-Li compound into a stable O-Li compound. These intermediates are often described as anions and in some cases may be so - it is certainly easier to draw the mechanism of the [2,3] shift that way. The commonest ethers used are benzylic 81 and 85 so that the proton to be removed has some acidity. Again a specific product 84 or 88 is formed from each isomer of the allylic alcohol 75 or 78. The original allylic alcohols are transformed into homoallylic alcohols. This rearrangement 82 and 86 is often called the [2,3]-Wittig rearrangement.12 Ph

Ph OH BnCl

R

NaH

R

83

84

Ph

O

NaH

O

O

BuLi

HO

Ph

Ph

[2,3]

R

R

78

Ph

R

82

Ph

R

HO

R

81

BnCl

Ph

[2,3]

BuLi

R

75

OH

O

O

O

85

R

86

R

87

88

There is good, if somewhat substrate-dependent, stereoselectivity when there is a substituent at the end of the alkene in the starting material. The E-ether 90 gives almost exclusively (98%) the stereochemistry shown in the product.13 The transition state has a ’half-chair’ conformation.14 Ph

Ph

O O

O

BuLi

R

R

R 89

Ph

HO

Ph

[2,3]

90

R 91

92

In this form the reaction is limited as an anion-directing substituent such as phenyl is needed and this group is incorporated into the structure. A more general approach is to provide an R3Sn group 93 that can be replaced by lithium even if there is no other substituent at all on the carbon atom. This variation is often known as the Wittig-Still rearrangement.15

I

O

SnBu3

SnBu3

75 R

O

BuLi

R 93

Li

LiO [2,3]

94

OH

R

R 95

96

This approach has been applied to the synthesis of peptide isosteres, compounds that have the same shape as peptides but lack the amide link. In this case 99 it is replaced by an E-alkene. The starting material 97 is made from the natural amino acid leucine and the Wittig-Still [2,3] shift on 98 rearranges this into an E-alkene flanked by two chiral centres 99.

348

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

Me

OH

Me KH

BocN

O

SnBu3

Me BuLi

BocN

BocN

OH

Bu3SnCH2I 18-crown-6

97

98

99

The migrating group passes suprafacially across the Z-alkene but appears to invert because the skeleton must be redrawn to show the E-alkene. This is an elegant way of transforming relatively easy to control 1,2 stereochemistry into more remote 1,4 stereochemistry.16 There is another example of the conversion of 1,2- into 1,4-stereochemistry at the end of this chapter. Not everyone is happy with stoichiometric toxic tin, but an alternative is available in the silicon compounds such as 101

1. BuLi

[2,3]

BuLi

2. Me3Si

OTf

O

Me3Si

HO 100; natural (R)-(+)-perilla alcohol

Li

101

OH

O 102

103

Alkylation of the allylic alcohol 100 with the silyl alkyl triflate gives an ether 101 from which the Me3Si group can be removed with BuLi. The new substituent (CH2OH) in 103 must be axial as it has to overlap with the p-orbitals of the alkene as it moves.17

Reactions with Electrophiles This section is mainly about the role of the OH group in directing reagents (mostly electrophiles) to one face or other of the alkene. This does not strictly fit with the title of the chapter but these reactions are important and this seems to be the logical place for them.

The Simmons-Smith cyclopropanation reaction In its most general form the Simmons Smith reaction18 is the addition of a carbenoid to an alkene to form a cyclopropane. The reagent is an organometallic compound made from a gem-dihalide and zinc usually with catalysis from a more active metal such as copper or silver, and probably has a structure such as ICH2ZnI. It is best known with allylic alcohols as the OH group directs both the regio- and stereochemical outcome. In the simple diene Z,Z-104, only the allylic alcohol forms a three-membered ring even with an excess of the reagent and does so stereospecifically.19 H

H

CH2I2, Zn

HO

cat Cu

Z,Z-104

HO 105

When the OH group is on a chiral centre the faces of the alkene become diastereotopic and the face syn to the OH group is preferred. These two diastereoisomers anti-106 and syn-108 show clearly that this is chelation control and not mere steric hindrance as the cyclopropane ring is formed on the same side as the OH group regardless of the position of the benzene ring.20

19 Reactions with Electrophiles

Ph

349

Ph

Ph

Ph

Ph

Ph

Ph

Ph

CH2I2, Zn

CH2I2, Zn

cat Cu

cat Cu

H OH anti-106

H OH

OH syn-108

107

OH 109

The zinc atom probably coordinates to the OH group and delivers the CH2 group in a mechanism such as 111. It appears that the iodine atom is simultaneously transferred to the metal so that the transition state is something like 112. At all events, the new ring is formed on the syn face to the OH group 113. These diagrams have been stripped of the benzene rings for clarity. I I

CH2I2, Zn

Zn

I I

OH

H

Zn OH CH2

cat Cu

H

OH

OH

110

111

112

113

The many recent publications affirm the importance of this reaction in modern synthesis. Tanaka and Takemoto’s synthesis of the marine metabolite halicholactone uses a sulfoxide route to the allylic alcohol 115. Notice the suprafacial [2,3]-sigmatropic rearrangement and that the OH group ends up in the middle of the diene. Attempted cyclopropanation of 115 gave a poor yield of a mixture of products.21 Ph

S

O

OH (MeO)3P

R

PivO O

O

R

PivO

MeOH

OTBDMS

O

CHCl2 114

O

OTBDMS

CHCl2 115; 88% yield

It was necessary to exchange protecting groups so that there was only one allylic alcohol 116 and to use a modification of the Simmons-Smith reaction with Et2Zn and CH2I2 to get a good yield of one diastereoisomer of the cyclopropane 117. OSEM

OSEM Et2Zn, CH 2I2

R

PivO

–20 ˚C, CH 2Cl2

OH

R

PivO

OTBDMS

OH

116

OTBDMS

117; 69% yield

Further transformations gave a precursor 118 for a ring-closing olefin metathesis using the Grubbs catalyst described in chapter 15. This gave the unsaturated nine-membered lactone required and deprotection revealed halicholactone 119. The solution to the regioselectivity problem of cyclo-propanation with three OH groups and three alkenes embedded in the skeleton of 119 is elegant. OSEM

OH R

O O

OTBDMS 118

1. Grubbs catalyst 72% yield 2. K 2CO3 MeOH 61% yield

R O O 119; R = n-C 5H11

OH

350

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

Stereochemically controlled epoxidations Many epoxidising agents, notably mCPBA and t-BuOOH/VO(acac)2, are capable of bonding to the OH group of an allylic alcohol. Epoxidation then occurs more readily on allylic alcohols and on the same side of the alkene as the OH groups, rather in the style of the Simmons-Smith reaction. The stereochemical outcome is easiest to see in cyclic allylic alcohols. Thus the cyclohexenol 120 gives the syn epoxide 121 cleanly with mCPBA via the conformation 122 in which the axial OH delivers the reagent to the same face of the alkene. Ar

OH

O

OH mCPBA

H 120

O

H

H

H

O

O

O

122; possible arrangement of peracid and allylic alcohol

121

A good illustration of many aspects of this chemistry comes from Koreeda’s synthesis of shikimic acid.22 The silyl diene 123 adds cleanly to methyl acrylate to give essentially one isomer of the adduct 124 (9:1). Dihydroxylation with catalytic OsO4 and the stoichiometric oxidant NMO (N-methylmorpholine-N-oxide) occurs on the opposite face to all three substituents. The alkene 124 is an allylic acetate but this group has no attractive interaction with the reagent. OAc

OAc

OAc CO2Me

CO2Me

HO

cat OsO4

+

O

CO2Me

NMMO

N

HO SiMe3

SiMe3

SiMe3

124; 72% yield

123

Me

125; 96% yield

O

NMO

A Peterson elimination, necessarily in acid solution as the OH and SiMe3 groups are anti (see chapter 15) gives a new alkene 126 that is allylic to the remaining OH group. Now epoxidation with mCPBA occurs on the face of the alkene syn to the OH group to give 127. OAc HO 125

OAc CO2Me

TsOH

HO

mCPBA

HO

CO2H

CO2Me HO

CH2Cl2

OH

O 127; 91% yield

126; 98% yield

shikimic acid

An example showing both regio- and stereoselectivity occurs in Ogawa’s synthesis of pipoxide.23 Epoxidation of the diene 128, with no free OH groups, gave a mixture of epoxides in low yield with very poor regio- and stereoselectivity. OBz OAc

OBz O mCPBA

OBz

OAc

OBz

OAc +

OAc +

pH 8 buffer

OAc 128

OAc 129; 19% yield

OAc O 130; 24% yield

OAc O 131; 14% yield

19 Reactions with Electrophiles

351

By contrast, the diene 132 with one free OH group gave a single epoxide in excellent yield. Epoxidation has occurred only at the allylic alcohol and only on the same side as the OH group despite the crowded nature of the alkene. This product 133 is the plant metabolite pipoxide. OBz

OBz O

OH

132

OH

mCPBA pH 8 buffer

133 87% yield (+)-pipoxide

OBz

OBz

The stereoselectivity with non-cyclic allylic alcohols is more complicated. The molecules tend to adopt a Houk conformation (chapter 21) with a hydrogen atom in the alkene plane and the OH group above or below. If the reagent is guided in by bonding to the OH group, syn epoxidation results. This is the case with allylic alcohols with a trisubstituted or a cis-disubstituted alkene 134 since there is then a group in the plane of the alkene favouring the Houk conformation 134b since whichever group on the chiral centre sits in the alkene plane must eclipse the cis substituent.24

= Me

Me

OH 134a

O

OH Me

Me

OH Me

mCPBA

Me H 135; >95:5 syn:anti

H 134b

However, trans-disubstituted alkenes are in general not epoxidised so selectively. In the simplest case with two methyl groups 136, the selectivity is about 2:1 syn:anti as the three conformations 136a, 136b, and 136c are about equally favourable since whichever group on the chiral centre sits in the alkene plane needs eclipse only the cis hydrogen atom. Me

Me

mCPBA

Me

O

OH Me

+ H

H

137 OH Me

H

OH Me

H

OH 136 Me

O

OH H

Me

H 136a

64:36

H 138 H Me

Me H

H Me 136b

OH 136c

Even mono-substituted allylic alcohols 139 that generally give very poor diastereoselectivity can give good results if the steric interactions are boosted by a Me3Si group 143. The Me3Si group pushes the Ph and OH groups out of the plane and favours the Houk conformation so that stereoselectivity is complete25 142. The Me3Si group can be removed with retention of configuration by fluoride ion. These more complex cases really fit into chapter 21 where you will see other examples of the Houk conformation in action.26

Ph OH 139

O

Ph

mCPBA

OH 140

O

Ph

+

35:65

OH 141

O Bu4NF

SiMe3

SiMe3 Ph

Ph

mCPBA

OH 142; 100% yield

OH 143

352

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

Fortunately the synthesis of complex molecules is likely to involve the epoxidation of more rather than less crowded allylic alcohols. An outstanding recent example is Isobe’s synthesis27 of 11-deoxytetrodotoxin 146, one of the metabolites of the notoriously poisonous Japanese fugu (puffer fish). Epoxidation of the allylic alcohol 144 gave an almost quantitative yield of one regio- and stereo-isomer of the epoxide 145. The mCPBA reagent was buffered with phosphate but that explains only the survival of the acetal: the lack of epoxidation at the other alkene is remarkable as it has an allylic relationship with the rather acidic NH group.

O

H

O

O O HN

HO

O COCCl3

mCPBA

H

HN

COCCl3

H2N

Na2HPO4 CH2Cl2

OH

HN OH O 145; 96% yield

144

H N

O OH

O

HO OH

146; 11-deoxytetrodotoxin

Regio- and Stereocontrolled Reactions with Nucleophiles We now return (at last!) to the original theme of the chapter: how can we add allylic electrophiles to other molecules with control over all the aspects of allylic systems we have been considering? We need to control: • Which end of the allylic system is attacked (regioselectivity) • Which face of the allylic system is attacked (stereoselectivity) • The geometry of the alkene in the product (stereoselectivity) Among the methods available for these reactions, two have gained prominence: the use of Claisen-style [3,3]-sigmatropic rearrangements for the selective allylation of enolates and the more general reactions of palladium allyl cation complexes.

Claisen-Cope rearrangements The Cope rearrangement is a [3,3]-sigmatropic reaction of a 1,5-diene to give another 1,5-diene. The reaction may be degenerate or there may be some reason for the reaction to go one way or the other such as relief of strain or development of conjugation. The Claisen version of the Cope rearrangement has a built-in direction finder. An oxygen atom in the carbon chain becomes a carbonyl group in the product and a C⫽O bond is significantly more stable than a C⫽C bond.28

The Cope Rearrangement:

[3,3]

The Claisen-Cope O Rearrangement:

[3,3]

O

Corey29 used a Cope rearrangement 147 in one of his syntheses of gibberellic acid. This reaction is driven both by the release of strain (the fused bicyclic product 148 is less strained than the bridged bicyclic starting material) and by development of conjugation. But the Cope is difficult to use and there is no general synthesis of appropriate starting materials.

19 Regio- and Stereocontrolled Reactions with Nucleophiles

H

4

5

CO2Me

6 1 147

CO2Me

CO2Me

[3,3]

3

R

353

= R

2 OSiMe3

OSiMe3

OSiMe3

R

148

148

The Claisen-Cope rearrangement is much easier to use. The starting material is an allyl vinyl ether 149 that can be made from an allylic alcohol. The product looks as though it might have been made by the allylation of an enol(ate) and the same disconnection 150 does for both. In this simple case, the product could equally well be made from an enamine of acetaldehyde and allyl bromide but you should by now realise that more complicated examples need a lot of control. OR

OH

O

H

+

[3,3]

O

149

use enamine

O

Br

150

The starting materials are made from the allylic alcohol 75 by exchange with either an acetal 151 or a vinyl ether. Though acid catalysis is needed for this reaction, it must not be strong enough to cause isomerisation of the allylic alcohol so carboxylic acids such as propionic (EtCO2H) are strong enough. Derivatives of aldehydes 152, ketones, acids, and amides 155 can all be used. Here are examples of an aldehyde 153 and an amide 156. OH

MeO +

R

OMe 151

75 OH

MeO +

R 75

Me2N

OHC

O

EtCO2H

R

heat

O

EtCO2H

OMe

153 NMe2 O

heat

R

154

R

152 NMe2

R 155

156

Regiocontrol is easy to explain: the allyl part of the molecule is turned inside out (or back to front if you prefer) with the new C–C bond being formed at the other end of the alkene from the OH group 158. Since the easier allylic alcohol to make is the one with the more highly substituted alkene (e.g. geraniol 157) it is actually easier to make the more ‘difficult’ allylated product.30

O

MeO

OH 157; geraniol

dry HCl

188 ˚C

CHO

90 minutes

R

158

159; 70% yield

The stereochemistry of the alkene needs a little more explanation. The transition state for the reaction is a six-membered ring and it prefers a chair-like conformation with the substituent R in an equatorial position 161. This gives an E alkene in the product and these reactions, like the [2,3]

354

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

shifts we have already seen, are very E-selective. The heavy bonds in the transition state 161 show the trans arrangement of the alkene that is being formed. OHC

O

R

R

O

O

R

R 152

160

161

153

An example of the formation of an ester of an E-4,5-alkenoic acid 164 comes in a synthesis of chrysanthemic acid by Ficini and her group.31 Reduction gives the allylic alcohol 163 and [3,3] Claisen rearrangement with triethyl ortho-acetate gives the product 164 in one step. NaBH4

MeC(OEt)3

CO2Et O 162

OH 163; 100% yield

H

164; 70% yield

Stereochemistry in the Claisen-Cope rearrangement Control over the geometry of the allylic ether portion of the starting material is easy - just use the appropriate allylic alcohol. Control over the vinyl ether is more difficult as that alkene is created in the exchange reaction. If the vinyl ether is cyclic, the problem disappears. The allyl vinyl ether 165, made from cyclopentanone and the allylic alcohol but-2-en-1-ol 8 rearranges with catalysis by 2,6-dimethylphenol into one diastereoisomer of the allylated product32 168. The stereoselectivity is easily predicted by the chair transition state 167. H

O O

120 ˚C ArOH

Me

165

O H

O Me

166

Me

H

H 167

168

If a Pd(II) catalyst is used instead the other diastereoisomer 170 is formed. Presumably Pd coordinates to both alkenes and holds the molecule in a boat conformation 169 during the reaction. In both cases the more difficult regiochemistry is achieved with the bonus of stereochemical control. O

room temperature

Pd O

Me

H Me

PdCl2(MeCN)2

H 165

O [3,3]

H 169

H 170

The easiest way to find the Claisen-Cope disconnection is to reverse the reaction. First you must recognise that this is worth doing and the simplest clue is to look for a piece of skeleton with two terminal π-bonds (alkene or carbonyl), the ends being 1,6-related. Here is a difficult example that makes an eight-membered ring 171 and also illustrates a different way to make the enol ether.33 Reversing the [3,3] looks awkward 173a, but redrawing the structure 173 makes it look much better.

19 Regio- and Stereocontrolled Reactions with Nucleophiles

5

1 O

3 2

4

recognise 1,6-link

6

355

reverse [3,3]

O

= O OR

O 171

O

OR

O

OR OR

O

172

O

173a

173

This particular enol ether is also an acetal (a ketene acetal) so we should be able to make it from the diol 174. This has 1,2- and 1,3-diO relationships and we much prefer to disconnect the 1,3 as we can then use some sort of aldol reaction.

O

1,1-diX

HO

FGI

OR O

2

aldol

O

OR

OR

HO2C 1

HO 173

3

HO

OR

HO2C

174

There are also stereochemical considerations here and Holmes used the Evans asymmetric aldol reaction (chapter 27) to make the starting material 174; R⫽Bn. The formation of any allyl vinyl ether reagent involves no change in the stereochemistry of the allyl alcohol - this is acetal exchange at the vinyl ether or acetal centre. The enol ether was added in masked form as a selenium compound 175 (chapter 32) as selenoxides eliminate at room temperature. The stereochemistry is developed directly from that in 177 as it transforms during the [3,3] shift.

PhSe

OEt

HO

cat. TsOH

PhSe

O

NaIO4

OBn OEt 175

OBn

HO 174; R=Bn

O 176; 70% yield

NaHCO3 MeOH, H 2O

O PhSe

O

[3,3]

O OBn

O 177

OBn O 178

O

O

OBn

179; 73% yield

The Claisen-Ireland rearrangement One special application comes when silyl enol ethers of allylic esters provide the starting material: this is known as the Claisen-Ireland reaction. Esters normally form E-enolates 181 with LDA at low temperature. Because the E/Z nomenclature depends on the hierarchy of the substituents it is particularly ridiculous when applied to enol derivatives of esters. A lithium enolate (Li ⬍ C) would have the opposite stereochemical label from a silyl enol ether (Si ⬎ C) so a uniform scheme is adopted whereby the metal – O bond always has priority over the other. The mechanism 180 may remind you of the reasons for this - they are discussed in more detail in chapter 4.

356

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

O

Li

R2 N LDA

R

O

Li

O

H

O

– 78 ˚C THF

R

O

R

R

O

R 181; ' E'-Li-enolate

R 180

This is true also of allylic esters 182 so that the geometry of both alkenes involved in the Claisen-Cope rearrangement can be controlled if the lithium enolate is trapped by silylation to give 183. Drawing the rearrangement in the chair conformation 183a → 184 gives a predictable diastereoisomer of the product 185. O

OSiMe3

OSiMe3 R2

O

O

1. LDA, THF – 78 ˚C

R1

R1

182 O R2

R2

183

H

1

R

183a

R2

OSiMe3

H [3,3]

R2

O

2. Me 3SiCl

H2O

CO2H

H

R1 184

R1 185

Ireland also discovered a remarkable way to form the Z-enolate.34 All that was necessary was to add HMPA (HexaMethylPhosphorAmide 189) to the solvent in making the lithium enolate. This excellent (though also dangerously carcinogenic) ligand for Li reverses the stereochemistry to give the more stable Z-silyl enol ether 186 and hence the opposite stereochemistry of the product 188. 1. LDA, HMPA, THF, –78 ˚C

182

OSiMe3

O

O

2. Me 3SiCl

OSiMe3

H

[3,3]

R1

R2 R2

1

R2

Me2N

H2O

CO2H Me2N P Me2N

R H

186

R1

187

O

189 HMPA

188

Most of the complex polyether antibiotics made using this reaction have actually used cyclic allylic alcohols and with them a boat-like conformation is preferred in the Claisen-Ireland rearrangement. A simple example is the dihydrofuran 190, prepared as a single enantiomer from a sugar (in chapter 23 we shall call this a ‘chiral pool’ strategy). Acylation and E-selective silyl enol ether formation 191 followed by Claisen-Ireland rearrangement gives one diastereoisomer of the product35 192. OSiMe3

OH OMOM

1. EtCOCl 2. LDA, THF –78 ˚C

1. room temperature

O OMOM

O

3. Me 3SiCl

2. CH 2N2

OMOM MeO2C H

O

O 190

191

192

The observed stereochemistry requires the enol ether to be transferred suprafacially across the top face of the five-membered ring 193 (this fixes the stereochemistry at the new centre on

19 Regio- and Stereocontrolled Reactions with Nucleophiles

357

the ring) with the bulk of the side chain away from the ring in a boat-like arrangement. The stereochemistry of the centre on the side chain is determined by the geometry of the enolate: this E-enol ether has the methyl group over the heterocyclic ring 194. Me3SiO

Me3SiO O

O

H

191

OMOM

OMOM

H H

192

Me O

O 193

194

This result in particular should warn you that the choice between a chair- and boat-like transition state is narrow. In general open-chain compounds prefer the chair and cyclic compounds the boat but do not rely on it! In spite of this apparent disadvantage the Ireland version of the Claisen rearrangement is one of the most widely used ways to set up complicated molecules. Ireland himself has used it to make a catalogue of polyether antibiotics with the monensin synthesis being perhaps the most remarkable. These syntheses are beyond the scope of this book but are worth reading.36 When both alkenes (allyl and vinyl ether) are inside a ring the stereochemical problems are much simplified. The reaction inevitably leads to ring contraction and a simple example is Funk’s synthesis37 of chrysanthemic acid 197 from the seven-membered lactone 195. The silyl enol ether 196 rearranges to cis-chrysanthemic acid in very high yield. A chair transition state is impossible in this restricted molecule.

1. LDA

O

O

2. t-BuMe 2SiCl

195

1. 65 ˚C

OTBDMS

O

2. HF, MeCN

CO2H

196; 97% yield

197; 94% yield

Knight’s synthesis of (⫺)-α-kainic acid provides a heterocyclic example.38 The starting material for the Claisen rearrangement is made from natural aspartic acid coupled to an allylic chloride 198 with a protected allylic alcohol 200 also present. Deprotection and lactonisation gives the nine-membered nitrogen-containing lactone 201. OTHP Cl

CO2H +

OTIPS

HN CO2Et

OTHP

CO2H

O 1. H , MeOH

O

2.

N EtO2C

OTIPS

N Me

198

199

200

Cl

N EtO2C 201

OTIPS

Formation and trapping of the lithium enolate gave only the E silyl enol ether 202 (that is with a cis alkene in the ring). As this is a cyclic enol ether, a boat-like transition state is expected and the formation of a single diastereoisomer of 203 confi rms that this is true. Diagrams 204 and 205 should clarify the conformation of the molecule as it reacts, the mechanism, and the stereochemistry of the product. The product was converted into kainic acid 206 in a few steps.

358

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

OTBDMS CO2H 1. room temperature

O

LDA

201

OTIPS

OTIPS

N

t-BuMe2SiCl –100 ˚C

N

2. K 2CO3 H2O, MeOH

CO2Et 203

CO2Et 202 TIPSO

TIPSO

TBDMSO Me

TBDMSO Me

NCO2Et

O

O

H

CO2H

NCO2Et

CO2H N H 206; (–)-a-kainic acid

H

204

205

Recent developments in this important reaction have helped to reveal the relationship between 2D and 3D stereochemistry. The synthesis of the phospholipase inhibitor cinatrin B 207 illustrates a serious stereochemical problem.39 Cinatrins have spiro-fused lactones. Each ring contains chiral centres and the spiro C-atom is also chiral. It is very difficult to relate stereochemistry between such orthogonal rings. The second lactone might be added by cyclisation onto an alkene such as 208 and that could arise from allylation of an enolate 209. Clearly some protection will be needed. HO

HO

CO2H

HO

HO

CO2H

enolate allylation

R

O

O

O

O

O

CO2H X

O

CO2H

O 207; R = n-C 9H19 (–)-cinatrin B

HO

210

+

R O

HO

208

R

CO2H HO

209

211 R

The starting material 212 was chosen as it was available from natural arabinose and contains enough functionality for the purpose. Esterification with racemic alcohol 211 gave a mixture of esters 213. This does not seem to matter as that centre is destroyed in the Claisen-Ireland rearrangement giving 214 but this compound was formed as a 43:57 mixture of diastereoisomers. BnO

MeO

OBn

CO2H

O

BnO

MeO

212

OBn

1. LDA, Me 3SiCl

O O 213

O

HO

OBn

THF, HMPA –100 ˚C 2. NaOH 3. CH 2N2

R

HO

R O

CO2Me 214

Use of the enantiomerically pure alcohol 211 revealed a surprising stereospecificity. One enantiomer 215 gave (mostly) the required diastereoisomer 216 and hence the iodolactone 217 having the right stereochemistry at the spiro centre and at the lactone centre in the new ring. Removal of iodine, selective oxidation and addition of a d1 reagent gave (⫺)-cinatrin B.

MeO

BnO

OBn

BnO

1. LDA, Me 3SiCl

O 215

R

2. NaOH 3. CH 2N2

1. H 2O NaOH

MeO

OBn

I

OBn

THF, HMPA –100 ˚C

O O

BnO

R O

CO2Me 216

2. I 2, NaHCO3

MeO

R O

O O 217

19 Regio- and Stereocontrolled Reactions with Nucleophiles

359

The scope of the Claisen-Cope rearrangement is very great but you have been given enough mechanistic and stereochemical detail to unravel all but the most difficult. The disconnection is always simple - the reaction corresponds to an allylation of an enolate and, providing you remember to turn the allylic system inside out, you will find the starting materials. All that remains is to identify which method to use and how to control the stereochemistry. Oh, and you also have to make the starting materials!

Pd-catalysed reactions of allylic alcohols In the last chapter we introduced the Heck reaction as a way to add nucleophiles to electrophilic alkenes. The details of the mechanism are discussed there. The reaction also occurs with allylic alcohols 218 and is a good way to make carbonyl compounds, particularly aldehydes, 219 as the alternative conjugate addition of an organometallic compound to an enal often shows poor regioselectivity (chapter 9). ArI

OH

?

Ar

cat Pd(0)

218

ArMet +

CHO 219

CHO 220

Formation of the aryl palladium (II) complex ArPdI by oxidative insertion is followed by π-complex formation and transfer of the aryl group from palladium to the alkene 221. Then β-elimination 222 gives the enol 223 of the product. The other ligands on palladium are omitted. I Pd(0)

ArI

218

Ar Pd

ArPdI

OH

Ar

221

PdI H

OH

Ar

222

OH

220

223

The regioselectivity is not as good with allylic alcohols as it is with unsaturated carbonyl compounds as there is only a steric preference for transfer of the aryl group to the end of the alkene. Examples 224 and 226 show how this can be improved with extra substitution.40 PhI

OH

Ph

cat Pd(0)

218

PhI

OH

CHO 225

224; 71% yield 84:16 regioselectivity

cat Pd(0)

Ph

CHO 226; 96% yield 95:5 regioselectivity

Pd-allyl acetate complexes Allylic alcohols 218 can be converted into their acetates 227 by standard methods, e.g. Ac2O/pyridine or DMAP, without any allylic rearrangement. Reaction with Pd(0) gives initially a π-complex 228 but this loses acetate as the Pd atom donates a pair of electrons to form an η3 allyl cation complex 229 of Pd(II). OH

Ac2O

OAc

pyridine

218 allyl alcohol

227 allyl acetate

Pd(0)

OAc Pd 228 π-complex

Pd 229; η3-allyl cation Pd(II) complex

360

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

These allyl cation complexes 229 are electrophilic and react with a variety of nucleophiles, most notably with the stabilised enolates of β-dicarbonyl compounds such as malonates. The immediate product is again a π-complex of Pd(0) 230 but there is now no leaving group so the Pd(0) drops off and is available for a second cycle of reactions. Though the reaction strictly requires Pd(0), the more convenient Pd(II) compounds are often used with phosphine ligands. Reduction to Pd(0) occurs either because the phosphine is a reducing agent or by oxypalladation and β-elimination. CO2Me

CO2Me Pd 229; η3-allyl cation

CO2Me

CO2Me Pd 230; π-complex of Pd(0)

CO2Me

Pd(II) complex

CO2Me 231 + Pd(0)

The allyl cation complex has the Pd(II) atom at right angles to the allyl plane and more or less in the middle of the triangle. It has therefore lost regiochemistry as the same complex would be formed by either regioisomer. Simple versions of the reaction are still useful - geranyl acetate 232 reacts only at the alkene that is part of the allylic acetate and gives mostly the product of attack at the less hindered end of the allyl cation.41 When malonate is the nucleophile, the regiochemistry is 9:1 in favour of attack at the terminal carbon. With the sulfone 233, the reaction is totally regioselective. CO2Me + OAc 232; geranyl acetate

(Ph3P)4Pd

CO2Me NaH

SO2Ph

SO2Ph 234; 84% yield (>97:3 regio)

233

You will appreciate that CO2R groups may be removed from these various products by hydrolysis and decarboxylation. The sulfones can be removed by reduction, for example with sodium amalgam in ethanol - this is ‘dissolving metal’ reduction and works by electron transfer in the style of a Birch reduction.

Stereochemistry of Pd-allyl cation complexes The most important aspect of this chemistry is its stereoselectivity. The stereochemical integrity of the alkene in the starting material may be preserved. This is not surprising with the geranyl acetate just described but is also the case with the Z-isomer neryl acetate 235. The regioselectivity is less (90:10, attack at the terminus being favoured) but the major product contains only the Z-alkene present in the starting material. There is no rotation at any stage in the reaction.

CO2Me + OAc

SO2Ph

(Ph3P)4Pd NaH

CO2Me SO2Ph

235; neryl acetate

233

236; 74% yield (90:10 regio)

19 Regio- and Stereocontrolled Reactions with Nucleophiles

361

In three dimensions the reaction is stereospecific with retention by double inversion. The palladium approaches the alkene from the opposite face to the leaving group and the nucleophile approaches from the opposite face to the palladium. This is clearly shown by reaction of the two diastereoisomers 237 and 240 with malonate. [Note: These compounds are of course racemic and attack at either end of the achiral η3 allyl cation complexes 238 or 241 would give racemic material anyway.] CO2Me

CO2Me

CO2Me (Ph3P)4Pd

NaH

CO2Me

+

CO2Me

CO2Me

OAc 238 Pd

237 CO2Me

239 CO2Me CO2Me

CO2Me CO2Me

(Ph3P)4Pd

NaH

+ CO2Me

CO2Me

OAc 241 Pd

240

242 CO2Me

The combination of stereo- and regiochemistry can be very powerful. The lactone 243 reacts with malonate esters to give one regio- and stereoisomer of the triester product41 244. 1. (Ph 3P)4Pd

MeO2C

CO2Me

NaH, CH 2(CO2Me)2

O O

MeO2C

2. CH 2N2

244; 84% yield

243

The palladium forms a π-complex 245 from the top face of the alkene, opposite the carboxylate leaving group. The nucleophile then adds from the opposite face to the palladium. The η3 allyl cation complex is unsymmetrical and the nucleophile adds at the less hindered end. The selectivity is very marked in both steps because the starting material is a folded molecule and the palladium much prefers to add to the exo face of the alkene. The nucleophile has to add from the same face as the CH2CO2⫺ side chain so it prefers addition to the end of the complex as far from this side chain as possible 246. MeO2C

CO2

(Ph3P)4Pd

O

243 O

Pd 245

MeO2C

etc

CO2Me

244

Pd 246

CO2Me

When the reaction is intramolecular, the question of ring size may affect both the regiochemistry and the stereochemistry of the alkene. In general, we might expect smaller or larger rings to be preferred to medium rings (8-13 membered) and it is obviously impossible to have an E-alkene in, say, a seven-membered ring. So it is not surprising that the 16-membered ‘large’ ring 248 is formed from 247 by attack at the end of the allylic system and with complete E-selectivity. This is a good way to make macrolides (large ring lactones).42

362

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

O

O (Ph3P)4Pd

O

O

NaH

PhO2S OAc

CO2Me

PhO2S 248; 69% yield

247

CO2Me

However, the formation of the eight-membered ring 253 with a Z-alkene from the E-allylic acetate 249 with only a trace (6%) of the six-membered ring 250 was described by Trost42 as a ‘shocking preference for cyclisation to the eight-membered ring’. Clearly the geometry of the η3 allyl cation complex can change during the reaction and this seems to be particularly the case for intramolecular reactions. The initially formed η3 allyl cation complex 251 can cyclise either to the six-membered ring or to the very unstable E-isomer of the eight-membered ring. Rotation to the other η3 allyl cation complex 252 must be faster than this cyclisation and the new complex can cyclise to either product. O NaH, (Ph 3P)4Pd

PhO2S

O

Ph2P

PPh2

250 + 253

OAc 249

PhO2S

PhO2S

O

O

O

O

O Pd

250

O O

PhO2S

Pd 252

251

O

PhO2S

253

The regioselectivity of the addition is affected by functionality. In another synthesis of chrysanthemic acid, Ficini43 combined the monoacetate of the cis (Z)-alkene diol 254 with malonate to give only the trans (E)-adduct 255 in 80% yield.

Pd(PPh3)4

HO

CO2Me HO

CH2(CO2Me)2

OAc 254

CO2Me 255; 80% yield

Evidently cis/trans isomerisation of the η3 allyl cation complex is faster than addition. The regioselectivity is remarkable. Either the steric hindrance of Me2COH next to the alkene is worse than that of the tertiary centre that is actually attacked or the OH group electronically discourages nucleophilic attack as it does in simpler SN2 reactions. We shall see a related regioselectivity in the reactions of diene monoxides in the next section.

254

CO2Me

Pd

Pd(PPh3)4

HO

HO 256

255 Pd 257

CO2Me

19 Regio- and Stereocontrolled Reactions with Nucleophiles

363

Conjugating and electron-withdrawing groups such as CN or CO2R make the remote end of the allyl cation complex more electrophilic. The CN substituted compounds 259 are easily made44 by acetylating the cyanohydrins of enals 258. Superficially this might look rather like a Heck reaction but the nucleophile is added in a different position. These complexes are reagents for the a4 synthon 262 having umpolung while the Heck reaction gives products from natural conjugate addition (a3 263). The product 261 is formed as an E/Z mixture: if an ester group (CO2Me) is the activating group, only the E-isomer is formed. OAc CHO

R

R

Pd(PPh3)4

CN

258

262

5 mol%

259

R

CN

a4

O O

Pd R

CO2Me

CO2Me

263 R

CN NaH

CN

a3 synthon for Heck reaction

R CN 261; 97% yield

260

synthon for

this reaction

Pd and monoepoxides of dienes A clever variation on the reactions of allylic acetates is the palladium-catalysed reaction of monoepoxides of dienes with nucleophiles. These monoepoxides can be made in two chief ways. If the diene is symmetrical 264, epoxidation of one of the alkenes occurs more rapidly than the epoxidation of the monoepoxide because the HOMO (ψ2) of the diene is higher in energy and therefore more nucleophilic than the HOMO (π) of an alkene. It is simply necessary to buffer the epoxidation reagent as these monoepoxides 265 are sensitive to acid (chapter 17).

mCPBA

mCPBA

O

buffer

264

slow

265

O

O 266

If the diene is unsymmetrical, particularly if the monoepoxide of the less substituted and therefore less nucleophilic alkene is wanted, alternative methods are required. The enone 269 provides both epoxides 270 and 267. The sulfur ylid route converts the enone directly to one epoxide while base-catalysed epoxidation followed by a Wittig reaction provides the other. O Ph3P

H2O2

CH2

O

O 267

O

268

O Me2S

CH2

NaOH

269

270

Reaction of these epoxides with Pd(0) follows the pattern of the allylic acetate reactions. The epoxide oxygen atom is the leaving group but the alkoxide intermediate is basic enough to deprotonate nucleophiles such as malonates without added base. With cyclopentadiene monoxide 265, one regio- and stereoisomer 275 is formed.45 The palladium adds to the opposite side of the

364

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

ring to the epoxide and the nucleophile adds to the opposite side to the palladium and at the end of the allyl cation away from the OH substituent 274, as you should by now expect. O

(Ph3P)4Pd Pd

O

CH2(CO2Me)2

O Pd

265

271

272 MeO2C

MeO2C

OH OH

MeO2C MeO2C

Pd

273

274

275; 57% yield

In general, substitution occurs at the less hindered end of the allyl cation and, unlike the cations formed from allylic acetate, these cations are never symmetrical - one end always has the nearby OH group that was the oxygen atom of the epoxide. The two isomeric epoxides 267 and 270 also give isomeric products though there is no stereochemistry here. Nu (Ph3P)4Pd

Pd

OH (Ph3P)4Pd

Nu

267

Pd

270 OH

OH Nu

OH

276

Nu

277

278

279

In the simplest cases of open chain epoxides with terminal alkenes,46 reaction is predominantly at the end of the chain and the product is overwhelmingly the E-isomer.

R

CH2(CO2Me)2

O

(Ph3P)4Pd

R

R Pd

–O

280

CO2Me

OH CO2Me 282; 84% yield, 98:2 E:Z

281

The control of remote chirality Allylic carboxylates and epoxides give the opportunity to control stereogenic centres that are not next to one another and may be as far apart as a 1,4-relationship across a trans alkene. There is an example of a similar problem solved by a [2,3] sigmatropic rearrangement earlier in this chapter. In particular vinyl lactones such as 283 have special uses. Trost47 first checked the regiochemistry of the reaction with lactone 283. This forms the usual η3 allyl cation complex 284 that reacts with the enolate at the end further from the carboxylate group. The alkene moves along the structure and emerges in the E configuration 285 corresponding to the conformation of the lactone. O O

(Ph3P)4Pd NaH CH2(CO2Me)2

283

MeO2C Nu

CO2H

CO2 Pd 284

CO2Me 285; >95% yield, >98% regioselectivity

19 Regio- and Stereocontrolled Reactions with Nucleophiles

365

Now for the stereochemistry. If a single diastereoisomer of the methylated lactone 286 is reacted under the same conditions with the same nucleophile, a single diastereoisomer of the product 288 is formed in excellent yield. Because the stereogenic centres are so far apart, it was difficult to tell one diastereoisomer from the other and Trost explains carefully how this was done. O

(Ph3P)4Pd

MeO2C

O

CO2H

CO2

NaH CH2(CO2Me)2

Pd

286

CO2Me 288; 90% yield >98% regioselectivity

287

These reactions are also used industrially. The Japanese company Teijin Ltd. uses the reaction of a stable carbanion with an allylic carbonate to make a stable analogue 290 of prostacyclin (PGI2) for use in the treatment of diseases of the circulatory system. Prostacyclin itself 289 is an enol ether and is too labile for medicinal use.48 CO2Me

CO2Me

O

HO

HO

OH 289; PGI 2 or prostacyclin

OH 290; Teijin's isocarbocyclin

The reaction starts with an allylic carbonate 291 that gives the usual η3 allyl cation complex 294 with Pd(0) though the ligand is the less usual chelating diphos 292 (Ph2PCH2CH2PPh2). The leaving group is a carbonate anion 293. MeO

O

MeO Ph2P

O

+ Ph2P

Pd

PPh2 O PPh2

R RO 291

+ R

293 methyl carbonate anion

292 Pd(dppe)2

OR

Pd

O

RO 294

OR

The carbonate anion loses CO2 to give methoxide ion which, like the anion released from a diene monoepoxide, is basic enough to deprotonate the nucleophile, a bis-sulfone 296. The anion from this 295 adds to the less hindered end of the allyl cation complex 294. The product 297 has the skeleton of the prostacyclin analogue and the sulfones simply need to be removed by reduction. PhO2S

CO2Me SO2Ph

PhO2S

SO2Ph CO2Me

294

295 293

CO2 + MeO

290

PhO2S H

CO2Me SO2Ph 296

R RO

OR 297

366

19 Allyl Alcohols: Allyl Cation Equivalents (and More)

Recent developments The palladium-catalysed coupling of allylic esters with carbanions has now been extended to include other reactions. The allylic carbonate 299 couples with the nitro compound 300 to give the expected product 301 in good yield. CO2Me RO

NO2

NO2 O

(Ph3P)4Pd

CO2Me

O

+ O

O 298; R = Ac 299; R = CO 2i-Pr

301; 83% yield

300

But if the acetate 298 was used the additional base required (here Cs2CO3) catalysed a further intramolecular conjugate addition of the nitro-stabilised anion to the unsaturated ester to give a new six-membered ring 302 with good stereoselectivity. This compound was converted into the Erythrina alkaloid49 303. O

O O

O (Ph3P)4Pd

298

301 Cs2CO3

N

O CO2Me

Cs2CO3

H MeO

NO2 302; 80% yield

303; Erythrina alkaloid

Since vinyl stannanes (chapter 18) couple with electrophiles under catalysis by Pd(0), it makes sense to create the electrophile by the palladium-catalysed reactions of allylic acetates in the same pot.50 The vinyl stannane 304 reacts cleanly with the more complex allylic acetate 305 with 0.2 equivalents of Pd(0) to give the triene 306 in good yield.51 PMP

O

O

O O

SnMe3 + AcO

304

Pd2(dba)3

PMP

O O

LiCl

305

306; 87% yield

Summary The chemistry of allylic alcohols and their derivatives is remarkably rich. They can be used to add allyl groups to other molecules with good regio- and stereochemical control. The methods in this chapter, particularly the palladium-catalysed reactions and the sigmatropic rearrangements, should allow you to make many otherwise difficult target molecules.

References General references are given on page 893 1. J. F. Lane, J. D. Roberts, and W. G. Young, J. Am. Chem. Soc., 1944, 66, 543, J. F. Lane, J. Fentress and L. T. Sherwood, J. Am. Chem. Soc., 1944, 66, 545. 2. E. J. Corey, M. C. Desai and T. A. Engler, J. Am. Chem. Soc., 1985, 107, 4339. 3. E. J. Corey and D. J. Beames, J. Am. Chem. Soc., 1972, 94, 7210.

19 References

367

4. B. M. Trost and T. R. Verhoeven, J. Org. Chem., 1976, 41, 3215. 5. E. J. Corey and J. P. Dittami, J. Am. Chem. Soc., 1985, 107, 256. 6. C. Sreekumar, K. P. Darst, and W. C. Still, J. Org. Chem., 1980, 45, 4260. 7. D. A. Evans, G. C. Andrews, T. T. Fujimoto, and D. Wells, Tetrahedron Lett., 1973, 1389. 8. D. A. Evans and G. C. Andrews, Acc. Chem. Res., 1974, 7, 147. 9. R. W. C. Cose, A. M. Davies, W. D. Ollis, C. Smith and I. O. Sutherland, Chem. Comm., 1969, 293. 10. G. R. Heintzelman, W.-K. Fang, S. P. Keen, G. A. Wallace and S. M. Weinreb, J. Am. Chem. Soc., 2002, 124, 3939. 11. M. P. Savage and S. Trippett, J. Chem. Soc., 1966, 1841; S. Demay, K. Harms and P. Knochel, Tetrahedron Lett., 1999, 40, 4981; F. Liron and P. Knochel, Chem. Commun., 2004, 304. 12. K. Mikami and T. Nakai, Synthesis, 1991, 594. 13. K. Mikami, Y. Kimura, N. Kishi and T. Nakai, J. Org. Chem., 1983, 48, 279. 14. Clayden, Organic Chemistry, page 943. 15. W. C. Still and A. Mitra, J. Am. Chem. Soc., 1978, 100, 1927. 16. A. C. Bohnstedt, J. V. N. Prasad and D. H. Rich, Tetrahedron Lett., 1993, 34, 5217. 17. J. Mulzer and B. List, Tetrahedron Lett., 1996, 37, 2403. 18. H. E. Simmons, T. L. Cairns, S. A. Vladuchick and C. M. Hoiness, Org. React., 1973, 20, 1. 19. E. J. Corey, H. Yamamoto, D. K. Herron and K. Achiwa, J. Am. Chem. Soc., 1970, 92, 6635. 20. J. S. Swenton, A. R. Crumrine and T. J. Walker, J. Am. Chem. Soc., 1970, 92, 1406. 21. Y. Baba, G. Saha, S. Nakao, C. Iwata, T. Tanaka, T. Ibuka, H. Ohishi and Y. Takemoto, J. Org. Chem., 2001, 66, 81. 22. M. Koreeda and M. A. Ciufolini, J. Am. Chem. Soc., 1982, 104, 2308. 23. S. Ogawa and T. Takagaki, J. Org. Chem., 1985, 50, 2356. 24. A. S. Narula, Tetrahedron Lett., 1981, 22, 2017. 25. P. Chamberlain, M. L. Roberts and G. H. Whitham, J. Chem. Soc. (B), 1970, 1374. 26. B. E. Rossiter, T. R. Verhoeven and K. B. Sharpless, Tetrahedron Lett., 1979, 4733. 27. T. Nishikawa, M. Asai and M. Isobe, J. Am. Chem. Soc., 2002, 124, 7847. 28. S. J. Rhoads and R. N. Rawlins, Org. React., 1975, 22, 1. 29. E. J. Corey and J. E. Munroe, J. Am. Chem. Soc., 1982, 104, 6129. 30. S. Julia, M. Julia, H. Linarès and J.-C. Blondel, Bull. Soc. Chim. Fr., 1962, 1947. 31. J. Ficini and J. d’Angelo, Tetrahedron Lett., 1976, 2441. 32. K. Takai, I. Mori, K. Oshima and H. Nozaki, Bull. Chem. Soc. Jap., 1984, 57, 446; J. L. van der Baan and F. Bickelhaupt, Tetrahedron Lett., 1986, 27, 6267; K. Mikami, K. Takahashi and T. Nakai, Tetrahedron Lett., 1987, 28, 5879. 33. M. A. M. Fuhry, A. B. Holmes and D. R. Marshall, J. Chem. Soc., Perkin Trans. 1, 1993, 2743. 34. R. E. Ireland, R. H. Mueller and A. K. Willard, J. Am. Chem. Soc., 1976, 98, 2868. 35. R. E. Ireland and J.-P. Vervet, J. Org. Chem., 1980, 45, 4259. 36. R. E. Ireland, R. S. Meissner and M. A. Rizzacasa, J. Am. Chem. Soc., 1993, 115, 7166. 37. R. L. Funk and J. D. Munger, J. Org. Chem., 1985, 50, 707. 38. J. Cooper, D. W. Knight and P. T. Gallagher, J. Chem. Soc., Perkin Trans. 1, 1992, 553. 39. A. N. Cuzzupe, R. Di Florio and M. A. Rizzacasa, J. Org. Chem., 2002, 67, 4392. 40. J. B. Melpolder and R. F. Heck, J. Org. Chem., 1976, 41, 265; A. J. Chalk and A. Magennis, J. Org. Chem., 1976, 41, 273; Review: B. M. Trost, J. Org. Chem., 2004, 69, 5813. 41. B. M. Trost and T. R. Verhoeven, J. Am. Chem. Soc., 1980, 102, 4730. 42. B. M. Trost and T. R. Verhoeven, J. Am. Chem. Soc., 1980, 102, 4743. 43. J. P. Genêt, F. Piau and J. Ficini, Tetrahedron Lett., 1980, 21, 3183. 44. J. Tsuji, H. Ueno, Y. Kobayashi and H. Okumoto, Tetrahedron Lett., 1981, 22, 2573. 45. B. M. Trost and G. A. Molander, J. Am. Chem. Soc., 1981, 103, 5969. 46. J. C. Fiaud and J. L. Malleron, Tetrahedron Lett., 1980, 21, 4437. 47. B. M. Trost and T. P. Klun, J. Am. Chem. Soc., 1979, 101, 6756. 48. P. W. Collins and S. W. Djuric, Chem. Rev., 1993, 93, 1533; K. Bannai, T. Tanaka, N. Okamura, A. Hazato, S. Sugiura, K. Manabe, K. Tomimoro and S. Kurozumi, Tetrahedron Lett., 1986, 27, 6353. 49. C. Jousse-Karinthi, C. Riche, A. Chiarioni and D. Desmaële, Eur. J. Org. Chem., 2001, 3631. 50. V. Farina, V. Krishnamurthi and W. J. Scott, Org. React., 1997, 50, 1. 51. W. D. Shipe and E. J. Sorensen, Org. Lett., 2002, 4, 2063.

Section D: Stereochemistry 20. Control of Stereochemistry – Introduction ...................................................................... 371 21. Controlling Relative Stereochemistry .............................................................................. 399 22. Resolution ........................................................................................................................... 435 23. The Chiral Pool — Asymmetric Synthesis with Natural Products as Starting Materials —....................................................................... 465 24. Asymmetric Induction I Reagent-Based Strategy .......................................................... 505 25. Asymmetric Induction II Asymmetric Catalysis: Formation of C–O and C–N Bonds ................................................................................... 527 26. Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds ................................................................................... 567 27. Asymmetric Induction IV Substrate-Based Strategy..................................................... 599 28. Kinetic Resolution ............................................................................................................... 627 29. Enzymes: Biological Methods in Asymmetric Synthesis ................................................ 651 30. New Chiral Centres from Old — Enantiomerically Pure Compounds & Sophisticated Syntheses — .............................................................. 681 31. Strategy of Asymmetric Synthesis......................................................................................717

20 Control of Stereochemistry – Introduction Introduction Stereochemistry denial The Words We Use Four ketones Diastereomeric and diastereotopic Enantiomers and enantiotopic Homotopic How to decide if two groups are homotopic, enantiotopic or diastereotopic Stereoselectivity and stereospecificity The Structures We Draw Racemates with one chiral centre Racemates with two chiral centres Optically pure compounds with one chiral centre Optically pure compounds with two chiral centres Sophisticated diagrams Wiggly lines and two chiral centres Racemates with one chiral centre which react to give a racemate with two Achiral diastereomers The Fundamentals of Stereochemical Drawings Drawing tetrahedra Drawings to avoid Mental manipulation of tetrahedra Stereochemical Descriptors Including: d, l; d, l; R*, S*; r, s; Re, Si; (
), (); and lk, ul The Next Ten Chapters Stereochemical Analysis Introduction, steps 1 to 9, summary Some Principles Chiral auxiliaries The Horeau principle Popular Misconceptions Misuse of words ‘Chiral’, ‘Chiral synthesis’, ‘Racemic synthesis’, ‘Homochiral’ Chiral centres and stereogenic centres Pseudo-asymmetric centres Unhelpful drawings Bond rotation Stereospecificity revisited Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

372

20 Control of Stereochemistry – Introduction

Introduction Stereochemistry denial

O O

O O

This chapter is – and the next ten chapters are – about stereochemistry. Nobody these days would question the importance of stereochemistry. But inexperienced undergraduates – and, it has to be said, some practising organic chemists – who are faced with a synthesis which involves a little stereochemistry often ignore the stereochemical aspect, and focus entirely on the bond-forming side of things. It is not hard to see why somebody might suffer from this ‘Stereochemistry Denial’. Stereochemistry isn’t easy; it complicates matters and is probably the most difficult aspect of organic chemistry for most chemists. One need only look in the literature from the early part of the 20th Century to see syntheses where compounds were produced as diastereomeric mixtures, where the fact that they were mixtures was of no concern to the authors, and where the final relative stereochemistry of the product was not known, and, at the time, could not have been. That was then. Not any more. In this chapter we introduce the language of stereochemistry and some of the concepts that are used in later chapters within this section and also in the rest of the book. If you are unfamiliar with some of the finer stereochemical points, then this chapter could well be a ‘baptism by fire’. Don’t be put off – come back to any difficult parts of this chapter after some of the others in this Section (chapters 21–31). Before coming to the first section, it is worth pointing out that we assume you know: what an enantiomer is, what a racemate is, what is meant by relative stereochemistry and what is meant by absolute stereochemistry.1 We also assume you know the term enantiomeric excess.2

The Words We Use Four ketones Consider the four ketones 1 – 4. The first three are achiral but the last is chiral and racemic.

20 The Words We Use

O

373

O

O

O R

1 achiral

2 achiral R

3 achiral

R

R

4 chiral R

If we just imagine the six membered rings to be flat for a moment (they are not of course but it doesn’t matter for our purposes) then the first molecule has two mirror planes running through it – one is in the plane of the paper and the other is a right angles to the paper 1a. The next two molecules, 2 and 3, each have one mirror plane. The last ketone 4 has no mirror planes and has a single stereogenic centre.

1a Two mirror planes

O

R R

Now let us consider a reaction of each of these ketones with a nucleophile, Nu. In every case, an sp2 centre is replaced by an sp3 centre to give alcohols 5 to 10.

O

O

O

O R

1 achiral

2 achiral R

R

HO Nu 5 achiral

R

R

HO Nu

HO Nu (±)-8 chiral

6&7 achiral R

4 chiral

3 achiral

R

R

HO Nu (±)-9 & (±)-10 chiral

R

R

Diastereomeric and diastereotopic We will come back to ketone 1. With ketone 2, there are two possible products 6 and 7. The nucleophile could attack the ketone from one face or from the other 2a. Only the top face is on the same side as the R group. The two possible products from the reaction are diastereomers and so the two faces of the ketone are diastereotopic.

374

20 Control of Stereochemistry – Introduction

O O

2 achiral

DIASTEREOTOPIC FACES

2a

R R

HO Nu OH

6&7 achiral

Nu Nu

R

OH

R

6

R

DIASTEREOMERS

7

A similar situation arises with ketone 4. Attacking the carbonyl on one face leads to one diastereomer while attack on the other face leads to another. Hence the faces are diastereotopic. The ketone 4 is chiral and racemic so the two diastereomers 9 and 10 will also be racemic.

Enantiomers and enantiotopic Like ketone 2, ketone 3 can give two possible products and the nucleophile could attack the ketone from one face or from the other. However, this time, the two lines of attack lie on either side of the mirror plane that runs through ketone 3a. The two products are enantiomers and the two faces of the molecule are thus enantiotopic. This reaction was the only reaction that started with an achiral material but gave a chiral product. Because a reaction of ketone 3 leads to chiral products we can also describe ketone 3 as prochiral. O O

3 achiral

3a ENANTIOTOPIC FACES R R Nu HO

Nu

HO

Nu

HO Nu (±)-8 chiral

ENANTIOMERS

R

R (R)-8

R (S)-8

Homotopic With ketone 1, it doesn’t matter at all which side of the molecule is attacked. The exact same compound 5 results and there is no possibility of different enantiomers or diastereomers. The compounds are superimposable. The two faces of the molecule are homotopic.

20 The Words We Use

375

O O

1 achiral

HOMOTOPIC FACES R

R

HO Nu Nu

5 achiral

OH R

OH

R

IDENTICAL COMPOUNDS – SUPERIMPOSABLE –

Nu 5

5

How to decide if two groups are homotopic, enantiotopic or diastereotopic There is a test you can do to decide. It always works. First, substitute one of the groups you are interested in with something else – which we will call X – to give a new molecule A. Now, instead, substitute the other group you are interested in with X to give you a second molecule B. Now decide upon the relationship between your two new molecules A and B. If they are enantiomers of one another then the two groups must have been enantiotopic. If they are diastereomers of one another then the two groups must have been diastereotopic. And if the two molecules you end up with are exactly the same – by which we mean that one can be superimposed upon the other – then the two groups must have been homotopic. The Stereochemical Test

O

O

O

replace HA with X

O

NH

NH X 12A

O replace HB with X

HB HA

O

NH

H

H

11 A and B are DIASTEREOMERS

X

12B

In the example above, the two protons HA and HB in oxazolidinone 11 are replaced in turn to give two new molecules 12A and 12B. This replacement gives us diastereomeric products — the protons themselves in the oxazalidinone (HA and HB) are therefore diastereotopic. In an NMR spectrum we would expect protons HA and HB to have different chemical shifts. We could have looked at other aspects of oxazolidinone 11 – the stereochemical relationship between the two methyl groups for example.

Stereoselectivity and stereospecificity Stereospecificity. Let us state at the outset that stereospecificity is not just perfect stereoselectivity. This is a very common misconception. There is a difference between 100% stereoselectivity and stereospecificity. The reaction from 13 to 14 proceeds by an SN2 mechanism. The nucleophile, PhS, attacks from behind the leaving group and so, of course, the reaction proceeds with inversion. This happens whether the molecule likes it or not. It is forced by the mechanism to give a specific product from a specific starting material. A stereospecific reaction conveys the idea that the other diastereomer will yield the opposite result: 15 must give 16.

376

20 Control of Stereochemistry – Introduction

SPh R

SPh

SN2

Br

R

13

SPh R 14

SN2

Br

R

SPh

15

16

Stereoselectivity. To continue with similar examples – in a stereoselective reaction the molecule selects a particular reaction pathway for some reason. That reason might include steric hindrance, kinetics or thermodynamics but the molecule isn’t forced by mechanism to react one way only. In the reaction of 17, which proceeds by SN1 rather than SN2, the intermediate cation 18 may be attacked from either side by a nucleophile (there is no longer the necessary inversion we see with SN2). Attack occurs selectively from the less hindered face of the cation 18. OH2

SN1

R

R

Br

R

17, anti

OH 19, anti

18

With stereoselectivity, there is the distinct possibility that different isomers will give the same product. The intermediate cation 18 is the same whether that starting material is syn 17 or anti 17 and must give the same product 19. OH2

SN1

R

R

Br

17, syn

R

OH 19, anti

18

There are plenty of reactions that contain both stereoselective and stereospecific components. One reaction might also be regioselective and even chemoselective and it is important to be able to dissect a reaction into these different selectivity components. An example concerns the two epoxidation reactions of 20 and 22. OH R

OH

ArCO3H

R

R 20

OH

O

21

OH

ArCO3H

OH O

23

O

R

R

R 22

OH

O

24; anti

25; anti

These reactions contain stereoselective and stereospecific components. The starting materials 20 and 22 are chiral and racemic and the faces of the double bonds (in both cases) are diastereotopic. Epoxidation is stereospecific since there is only one extra atom in the epoxide it must react with the two carbon atoms on the same side if the reaction is concerted. But, with allylic alcohols the reactions are also stereoselective since the OH will form a hydrogen bond with the peracid and hence deliver the oxygen syn to the alcohol. It could in principal react to give the anti product but it is not possible to make (as the major product anyway) the anti isomers 24 and 25 by this method. Another example of a reaction that has both stereoselective and stereospecific components appears in the pumiliotoxin synthesis in Chapter 21.

The Structures We Draw Racemates with one chiral centre If the compound 26 is a racemate with only the one chiral centre then the bonds are best drawn as ordinary straight lines. This implies a racemate – that there are equal amounts of the two enantiomers (S)-26 and (R)-26. There is no need to draw a wiggly line, for instance.

20 The Structures We Draw

What we draw OH

Correct but unnecessary

What it means OH

(S)-26

(±)-26

377

OH

50:50

OH

(R)-26

unnecessary wiggly line

(±)-26

Racemates with two chiral centres Now things are more complicated. This is because, whether we have racemates or not, we need to specify relative stereochemistry. The syn diol 27 contains two chiral centres. We have to draw in stereochemistry because we need to define which diastereomer we are concerned with. It is, however, racemic. Similarly, what we draw with the anti diol is anti-27. What we draw OH

OH

OH

What it means OH

OH

OH

OH

OH

OH

OH syn-27 What we draw OH

OH

50 : 50 What it means OH

OH

27a

anti-27

50 : 50

Interestingly, this is the point at which many chemists bail out of stereochemistry. Somehow, even the thought of relative stereochemistry becomes all too much (a bit like the character in the cartoon). Such a person would draw this diol as 27a which might mean anything. It really is worth taking the time to think about what the stereochemistry is and drawing it in.

Optically pure compounds with one chiral centre We can now draw the compound 26 as we mean it. There is no ambiguity. OH

OH (S)-26; We draw this.

(R)-26; We draw this.

We mean this enantiomer and only this enantiomer

We mean this enantiomer and only this enantiomer

Optically pure compounds with two chiral centres Now if we draw syn-27, we would like it to mean only that enantiomer. But now comes the difficulty. How do we distinguish this diagram from the one we used for the racemic compound? They both look the same. Sadly there is no easy answer to this and with two chiral centres it is ambiguous but for a start we can label the drawing (±)-syn-27 if racemic. Sound advice is to spell it out on your

378

20 Control of Stereochemistry – Introduction

diagrams. Write (±) under the racemic compound and ‘optically pure’ under the optically pure one, or else label it (S,S)-27 or (
) or whatever is right. OH

OH

OH

syn-27

OH

OH

(±)-syn-27

OH

optically pure syn-27

Sophisticated diagrams An attempt has been made to fix this ambiguity and to use ‘blocks’ for relative stereochemistry (±)-syn-27 and ‘wedges’ for absolute (S,S)-syn-27. And this would be quite a good system if every body did it, but they don’t. Sadly, for the moment at least, this extra level of complexity is not widely accepted and cannot be relied upon. OH

OH

OH

(±)-syn-27

OH

optically pure syn-27

Wiggly lines and two chiral centres A wiggly line can mean several things. It can simply mean that both configurations are present, though as we saw above, it is not necessary with a racemic mixture. Suppose that we have only the S configuration of the amine 28 at the benzylic site. NMe2

NMe2

S

S

NMe2

R

O 28

S

S

O (S,R)-28

O (S,S)-28

This diagram 28 could mean several things — i) Both diastereomers (S,R)-28 and (S,S)-28 are present (though not necessarily in equal amounts). ii) There is only one diastereomer but we don’t know which one it is. and it might even mean — iii) We don’t care. (Because, say, of reactions to follow that make it unimportant). Racemates with one chiral centre which react to give a racemate with two Consider a racemate with one chiral centre such as 29. We do not define the stereochemistry since there is only one centre. Note that the bond is not bold or hashed or even wiggly. If we reduce this in a stereoselective reaction so that anti-27 results then we draw in the relative stereochemistry. Even though the product is a racemate, and we’ve added the ‘±’ symbol to avoid any confusion, we must define the relative stereochemistry. Of course, we could equally have drawn the other enantiomer 27a.

20 The Fundamentals of Stereochemical Drawings

OH

O

OH

379

OH

OH

OH

reduce

(±)-anti-27

(±)-29

(±)-anti-27a

Achiral diastereomers You do not need chirality to make definition of the stereochemistry essential. If there are two possible diastereomers stereochemistry needs to be defined. This is the case where the achiral ketone 30 is reduced to the achiral alcohol 31. Even though both starting material and product are achiral – there is stereochemistry to be addressed, say anti-31 is formed and not syn-31. O

OH

reduce

t-Bu

t-Bu

OH

t-Bu anti-31

30

OH

OH

OH

and not syn-31

32

We’ve gone on about how important it is to always draw in the stereochemistry but we will end this section with one last point—beware of adding too much stereochemical information. Later in this chapter is a subsection on Unhelpful Drawings under Popular Misconceptions. Triol 32 is an example of a compound drawn with too much stereochemical information.

The Fundamentals of Stereochemical Drawings Drawing tetrahedra We’ve all got into trouble from time to time when drawing stereochemistry. Most often it happens when we are standing in front of a group of people. A few simple, and probably entirely obvious, guidelines might help. As far as possible, try to keep two bonds in the plane of the paper. You cannot get more than two tetrahedral bonds in the plane at a time but it is certainly possible to draw structures with fewer and such drawings rapidly become very messy. If you need to draw only three bonds then keep them at 120 to one another. Good: 120˚ angles, two bonds in plane of paper

Bad: poor angles

If you need to draw all four bonds then draw the third and fourth bond with a shallow angle between them. If you need to indicate stereochemistry, keep the shallow angle between the thick and hashed bonds. The other two are in the plane of the paper. shallow angle

Keep the shallow angle between stereochemical bonds shallow angle

Approx. 120˚

shallow angle

380

20 Control of Stereochemistry – Introduction

Drawings to avoid There are many. One of the most confusing is to use a stereochemical bond (ie a hashed or a bold bond) between two stereocentres. For example, amine 33 and alcohol 34 can be drawn without confusion in several ways two of which are shown below. NH2

NH2

redraw

33

redraw

OH 34

33a

OH 34a

The potential confusion arises when we have two stereogenic centres next to one another. The first depiction of aminoalcohol 35 makes perfect sense but the redrawn form is starting to get confusing. Which centre does the bold bond refer to? NH2

NH2 redraw

OH syn-35

NH2

NH2

syn-35a confusing stereochemistry

redraw

OH ? anti-35a

OH anti-35

OH syn-35a

In fact, we can work out the stereochemistry in the second drawing of 35. If we look at each centre in isolation and consider the bold bond to both then it works. But this is rare and works here because we happen to have syn-35. How we would redraw anti-35 in a similar way? Would we use a bold or a hashed bond to do it? There is no good answer. The best thing to do is follow these two guidelines— i) Keep the main chain of the molecule in the plane of the paper. This means that stereochemical bonds will not appear in the main chain of the molecule. ii) Do not use stereochemical bonds between two stereocentres. This will ensure the same.

Mental manipulation of tetrahedra Over the next few chapters it will often be necessary to redraw tetrahedra in different ways. You need to become proficient at this. We introduce two manipulations that will take you a long way. We will start by looking back at the trivial manipulation we just did with amine 33. There is nothing new here but we’ve invented the terms CBP and GSR to help you. Imagine a single enantiomer of the amine 33. We would not normally draw in the hydrogen but we’ve done it in the box just to remind you where it is. We know with tetrahedra that only two bonds can be in the plane of the paper at any one time. The first manipulation we encounter simply changes which bonds are in the plane. This is something you very often have to do in reaction schemes. We shall simply call this manipulation a “Change of Bonds in the Plane” or a CBP. H NH2

33

NH2

33

CBP

NH2

33a

CBP

NH2

33b

Notice that if we start with one of the bonds coming out of the paper then one of the bonds is always coming out of the paper (but remember, you are not allowed to swap groups). If you are

20 Stereochemical Descriptors

381

not convinced by this, make a tetrahedral model and leave the final hydrogen off the back. The result leans like a (static) spinning top on its side and if you hold two bonds in the plane then the other one pops up. The reverse is true too of course—if one of the groups is pointing away from us then one must always be pointing away. See what happens when we do a CBP as with alcohol 36 below. OH

OH CBP

Ph

OH CBP

Ph 36

Ph 36a

36b

No groups swap location when we perform a CBP. Now we meet the second type of manipulation we shall call a “Group Swap by Rotation” or a GSR. If we do swap the location of two groups then we must do so by rotating about one of the bonds 33c and, as a consequence, we move from one thick bond to one hashed bond 33d (or from one hashed bond to one thick bond). rotate about this bond

NH2

one thick bond

one hashed bond

Me

Me

NH2

GSR

33c

33d

amino and methyl groups swap positions

You shouldn’t use these devices without understanding how they work. If necessary use models to satisfy yourself that you understand how it works. Amaze yourself at how easy it is. If up until now you have been a stereochemical denialist and work among other denialists then learning these basic skills will earn you instant awe and respect at the flipchart! Imagine the ring opening of cis-stilbene oxide 37 by an amine in an SN2 reaction. The product is the amino alcohol 38. Redrawing this 38a so that we have the longest chain in a zig-zag is easy by application of a GSR to the right hand portion of the molecule. Note that we have not drawn any structure with a hashed or thick line between two chiral centres. In this sequence the question of absolute stereochemistry does not arise as cis-stilbene oxide 37 has a plane of symmetry and is therefore achiral.

O

OH RNH2

Ph

Ph Ph 37

rotate about this bond

OH NHR

Ph 38

GSR

Ph

Ph NHR 38a

Stereochemical Descriptors This is a short section to describe what the various stereochemical descriptors mean. There are several in common usage and we assume you know what R and S mean. The more antiquated terms include d and l and D and L. Antiquated they may be but their use is still common and you need to, or at least ought to, know what they mean. d, l These were once used to indicate the sign of optical rotation. So, d for dextrorotatory and l for levorotatory. These have been replaced by (
) and () respectively.

382

20 Control of Stereochemistry – Introduction

d, l These are used mostly for carbohydrates and amino acids. The terms are obsolete and come from a time when the configurations of lots of chiral compounds relative to each other were known but the absolute configuration was not. D-(
)-Glyceraldehyde is taken as the standard. Enantiomerically pure compounds that could be related to D-(
)-glyceraldehyde by a series of reactions or degradations were labeled D [or L if they related to L-(
)-glyceraldehyde]. Without going into details, an amino acid is L if the amino group is on the left in a Fischer projection which has the carboxylate at the top. They are smaller than normal capitals. R*, S* These are a sensible way to describe relative stereochemistry of racemic compounds. Suppose we have a compound 39 with three asymmetric centres. We know that we have only one diastereomer. The enantiomer drawn could be called (1R, 2S, 3S)-1,3-dihydroxy-2-aminobutylbenzene. But the enantiomer, (1S, 2R, 3R)-1,3-dihydroxy-2-aminobutylbenzene will also be present. To indicate that both are there we could either call it (1RS, 2SR, 3SR)-1,3-dihydroxy-2-aminobutylbenzene or (1R*, 2S*, 3S*)-1,3-dihydroxy-2-aminobutylbenzene. OH 1

OH 3 2

Br MeS

NH2 39

(S)-40

r, s These are a bit more complicated but important terms and are discussed in ‘Pseudoasymmetric Centres’ below. Re, Si These two terms are in common usage and refer to which face of a flat functional group such as a carbonyl. The faces will either be enantiotopic or diastereotopic. They are very commonly used terms and easy to understand if you already understand R and S. If the decending priority of the groups define a clockwise motion then the face being viewed is the Re face. Similarly, an anticlockwise motion indicates that the Si face is being viewed. One thing to remember here is that whether a face is Re or Si has nothing to do with whether the resulting compound is R or S. If it were, then Re and Si would have to also depend upon the priority of the incoming group. re, si These are often, mistakenly, used for Re and Si but, in a way, they are to Re and Si as r and s are to R and S. In other words they concern the formation of pseudoasymmetric centres from sp2 centres. And we shall just leave it at that as they are not encountered very frequently. If you want to know more, have a look in Eliel’s book on Stereochemistry. (
), () These simply refer to the direction of rotation of polarised light. Unlike R and S which are determined by configuration, their determination is entirely empirical. In other words, although we know that hypothetical compound 40 has the S configuation, we have no idea whether it is (
) or () and will not know until we put it into a polarimeter. lk, ul These might be called ‘derivatives’ since they are further removed from the stereochemical information than R or S. As such they are less useful since they need more ‘decoding’ to get back to the actual structure. They are used for compounds with adjacent chiral centres. They are easy enough to understand – a compound with two R chiral centres (or one with two S chiral centres) is a lk (lk for ‘like’) compound whereas a compound with one R and one S centre would be ul (ul for ‘unlike’). Although this has the advantage of being unambigous, it does not convey a pictorial

20 The Next Ten Chapters

383

image at all. Since organic chemists think with structures, ul and lk is considered by many not to be helpful. For a more transparent device, see syn and anti below. l, u These are related to lk and ul but, instead of describing the relationship between two chiral centres, they describe the relationship between a prochiral centre and the resulting chiral centre. A situation where a Re face gives after reaction, an R centre is l (and Si face giving S centre) but Re face giving S centre is u (and Si giving R).3 They are not used very commonly. syn, anti We use these routinely in this book. Once a molecule has been drawn in the standard way with its most extended chain 41 then if the two groups concerned are on the same side of the molecule it is syn and if they are on the opposite side then they are anti. This is illustrated with the aldol products syn- and anti-42. O

longest chain fully extended

OH

O

syn-42

41

OH

anti-42 OH

redraw

O

OH

longest chain extended but not in plane

O

syn-42a

longest chain not extended

syn-42b

The advantage (or disadvantage depending on your point of view!) with syn and anti is that they always need a picture to go with them. Compound 41 shows the skeleton of 42 with the longest chain in its most extended form (zig-zag). Aldol syn-42a shows syn-42 redrawn but with the zigzag chain not being kept in the plane and syn-42b with the longest chain not extended. Neither of these is as good as syn-42 but it might well be necessary to draw some compounds like this (if there is a macrocyclic ring for example we might have no choice). Something drawn in the way of 42b would then look like anti rather than syn. There are no rules according to atoms priorites etc – the picture is king. It is also fast to go from name to picture and no ‘decoding’ is needed. If you take one thing away from this section it should be that any name or stereochemical description should be tied firmly to a clear stereochemical diagram drawn on the lines we have suggested. Everybody will then know what you are talking about.

The Next Ten Chapters The next ten chapters are all to do with stereochemistry. The next chapter and Chapter 30 concern relative stereochemistry and its control. All the other chapters are concerned with asymmetric synthesis and the different strategies that can be employed. The different strategies can be broadly divided into three groups. These are resolution, chiral pool and asymmetric induction. Resolution

Chiral Pool

Asymmetric Induction

These main groups can be subdivided as shown below. This diagram shows not only the structure of the field asymmetric synthesis but how they are arranged in this book. Resolution, for instance divides into ‘Classical’, ‘Kinetic’ and what we call ‘Fancy Kinetic’ which means dynamic kinetic

384

20 Control of Stereochemistry – Introduction

resolutions and all the complicated stuff. All the resolution subdivisions involve the separation of enantiomers. Classical resolution includes methods like the formation of diastereomeric salts whereas in a kinetic resolution enantiomers may be separated because one reacts faster than the other. Some of the subgroups can be divided further so reagent-based asymmetric induction divides into catalytic and stoichiometric methods and even these may subdivide. Some of the subdivisions are so important that they have entire chapters dedicated to them and others even have several chapters devoted to them. For instance, we devote two chapters to catalytic asymmetric induction (or three if you count enzymes) – one for C–O and C–N bond forming reactions and one for C–H and C–C bond forming reactions. Although it looks as if there are clear divisions between the topics it is, of course, not that simple and one chapter will refer to another. Finally, Section D rounds up with a chapter on Strategy of Asymmetric Synthesis. The other chapters are self explanatory. Resolution

Asymmetric Induction

Chiral Pool

CHAPTER 23

Substrate Based

Fancy Kinetic

Kinetic

Classical

Reagent Based

CHAPTER 27 Catalytic CHAPTER 22

Stoichiometric

CHAPTER 28

CHAPTER 28 Enzymatic

CHAPTERS 26 & 27

'Strategy'

Chemical

CHAPTER 31

CHAPTER 24 CHAPTER 29

Division of Topic and Chapters in Section D. Many enantiomerically pure compounds can be made by a variety of strategies. Several conceivable strategies for the synthesis of the quinolone antibiotic ofloxacin 45 are shown below. The molecule contains only one chiral centre and this may be introduced as a fragment from the chiral pool 47 or by the resolution of a key intermediate 44 or indeed the final product.4 A synthesis developed5 for the preparation of racemic ofloxacin which incorporates racemic aminoalcohol (±)-47, could in principle F

F

Chiral Reagent Catalytic or Stoichiometric

F

Resolution

N

F

O F

43

O F

X*

F

F

N

F 46

OH

CO2H

N

O Me

N

NH O

O

OAc 44

N O

45; ofloxacin

Substrate Based X* = chiral auxiliary

Chiral Pool

NH2

HO 47

20 Stereochemical Analysis

385

be turned into an asymmetric synthesis.6 Alternatively an achiral intermediate 43 could be asymmetrically reduced or a chiral auxiliary could be attached 46 to direct an asymmetric reduction.7 The application of different strategies means that several compounds show up in more than one chapter.

Stereochemical Analysis Stereochemical analysis is a formal descripton of the stereochemical aspects of molecules and their reaction. In many ways it is simply a formalisation of what chemists do anyway and is not dissimilar to retrosynthetic analysis. But clarity is vital and clarity can be achieved only by using the stereochemical words we have just met. Stereochemical analysis uses stereochemical words to describe the properties of molecules and of reactions. It uses words of symmetry and it avoids ambiguous words. The word ‘same’ means different things to different people and depends upon the context. One chemist might describe two enantiomers as being the ‘same’ because they have the same structure whereas another chemist concerned with asymmetric synthesis might well describe them as ‘different’. ‘Different’ is not much better but by using the word ‘enantiomer’ we know where we are. We have divided the analysis into nine steps but, as we shall see, you will often feel you don’t need all of them. Stereochemical words

Properties of molecules • Diastereotopic • Enantiotopic • Homotopic • Chiral • Racemic

Properties of reactions

Diastereomer Enantiomer Homomer Achiral

• Stereoselective • Enantioselective • Stereospecific

Step No. 1 Look for symmetry in the starting material. Identify axes, planes or (less commonly) centres of symmetry. For example, diol 48 has a axis of symmetry whereas diol 49 has a plane. Diol 48 is C2 symmetric and chiral; 49 is achiral. OH

OH Ph

Ph

Ph

OH

HO

OH

Ph

Ph

Ph

Ph

Ph

OH 48a; axis of symmetry coming towards us

OH 48

HO

49a; plane of symmetry

49

A word of warning straight away – be careful about how the molecule has been drawn (especially if you have previously suffered from stereochemistry denial!). Redrawing the apparent syn diol 49 reveals the anti relationship between the hydroxyl groups 49b. The plane of symmetry is no longer evident. In fact, the molecule has a centre of symmetry in this conformation. Diol 49 is achiral however we draw it. Normally, our priority should be for planes and axes before we start thinking about centres. HO

OH

OH redraw

Ph Ph

Ph

'syn' diol 49

Ph

OH anti-diol 49b

OH

centre of symmetry Ph

C2 axis of symmetry

Ph

OH syn-diol 48

Step No. 2 Identify relationships within the molecule. For example, once you have identified a plane of symmetry in a molecule, you might decide that two ester groups are enantiotopic. ‘Enantiotopic’ is

386

20 Control of Stereochemistry – Introduction

the relationship. Make sure, by looking back at Enantiomers and Enantiotopic if necessary, why the ester groups in 50 are enantiotopic while those in 51 are homotopic. Imagine being at a party where no relationships were revealed – you would not know who is whose brother, sister, husband or wife. It would be unfortunate to be without information of that kind (especially) if you were a bit of a gossip. Relationships count and, just because a girl has a sister, it doesn’t stop her from having a brother too. So don’t stop your analysis once you have identified one relationship—be exhaustive. We have identified that the ester groups in 50 are enantiotopic, but we might also notice that the two protons at the α-position are diastereotopic 50a. One is on the same side as the phenyl group and one is not. O

Ph

O

EtO

O OEt

O

O

EtO

OEt

50

Ph

O

EtO

OEt

H H 50a: diastereotopic protons

51

We shall now examine the epoxidation of the doubly allylic alcohol 52 to show how steps 1 and 2 give insight into the possible stereochemical outcome of this reaction. We shall assume that mono epoxidation is possible. The five-membered ring is flat and the two bonds to the epoxide oxygen must of course be on the same side of the ring. Step 1, looking for symmetry in the starting material, reveals one symmetry element - a mirror plane through the OH group at right angles to the ring 52a. The molecule is achiral. O RCO3H

OH

OH

52

mirror plane

OH

53

52a

Step 2, identifying relationships within the starting material, is more revealing. The two alkenes are reflected in the mirror plane so they are enantiotopic. But the top and bottom faces of the ring are diastereotopic - one is on the side of the OH group and one is not. So it is possible for any epoxidising agent to choose either the top or bottom faces by steric hindrance or by bonding to the OH group. Step No. 2

Step No. 1

Step No. 2

diastereotopic faces

OH

mirror plane

OH

OH

mirror plane

mirror plane

diastereotopic faces

But the choice between the two enantiotopic top faces or the two enantiotopic bottom faces would produce a single enantiomer, and that would require the introduction of asymmetry, probably from the reagent. Step No. 2

Step No. 1 OH

mirror plane

enantiotopic faces OH

mirror plane

mirror plane

Step No. 2

OH

enantiotopic faces

20 Stereochemical Analysis

387

Step No. 3 Decide whether or not the molecule is chiral (not to be confused with step number 4). The amino acid 54 is clearly chiral (whether or not it is racemic of course). A cursory glance at compound 55 may fail to expose that it is not chiral. Compound 55 in fact contains a centre of symmetry (rather more difficult to spot) but redrawing 55 as 55a exposes a mirror plane within the molecule. Either way, it cannot be chiral. O OH

Ph

Ph

mirror plane

redraw

NH2

Ph

Ph

54

55

55a

Step No. 4 If the molecule is chiral, decide whether or not it is enantiomerically pure. You will often have to know more about the reaction than is evident in the scheme. For instance, trans-stilbene oxide 56 is always chiral but may or not be racemic. It is not evident in the scheme below and it could easily be either. It all depends how it was made. If only achiral starting materials and reagents were used, the epoxide 56 must be racemic. If asymmetry was present somewhere, it may be a single enantiomer. But had we put optically pure stilbene oxide 56 into the reaction then we could expect to get an optically pure product 57 at the end. Amino acid 54 is likely to be enantiomerically pure anyway (as are other compounds from the chiral pool) and if we were in any doubt it has been drawn as a single enantiomer. In other situations we need to know the context to be sure.

asymmetric starting material or reagent?

O Ph

OH N3

Ph

Ph

Ph N3 57

56

These first four steps are the most important. They are not the whole story and it may be necessary to go through the next five steps as well. Experience will show you if they are needed and enough experience may allow you dispense with this formal analysis. Step No. 5 Identify the source of chirality and indicate the chiral centres or axes. You may feel this is unnecessary but it doesn’t hurt to draw a ring around a chiral centre or draw in the symmetry plane with a coloured pen. Step No. 6 Look for more subtle symmetry relationships. These might be where molecules are only just short of symmetry or where a small change in the molecule could increase its symmetry. For example, we could raise the level of symmetry of the ester 58 by either hydrolysing it to give the diol 59 or by esterification of the remaining hydroxyl group to give the diester 60. It can often be instructive to notice this sort of thing if only as a warning. mirror plane O HO

O 58

mirror plane O

O HO

OH 59

O

O 60

388

20 Control of Stereochemistry – Introduction

A completely different issue, but another where we are just short of some symmetry, is that of the pseudo-C2 axis. The diol 62 has a pseudo-C2 axis (Ψ-C2) – removal of the phenyl group would give us the genuine article 61 and 61a. Note that this term is applied to compounds where the disruptive element is on the axis itself (or rather where this axis could be) because it is only here that the substituent will be non-stereogenic. Hence we describe 62 as pseudo-C2 but not 63 or 64. The pseudo-C2 phenomenon is very real but there are those who do not like the name. There is a movement to rename it pro-C2. We hope this movement gathers momentum. It is a better description (akin to prochiral) and is without the negative connotation of ‘pseudo’. OH

OH

OH

Ph

O

OH HO 63

Ψ-C2 axis

C2 axis OH

O

62

61

OH

OH

Ph

OH

OH

OH Ph

62a

61a

64

Step No. 7 Decide how the reaction influences symmetry. Is symmetry created, destroyed or even left unchanged? If a molecule contains a mirror plane, for instance, and the new bonds are made within the mirror plane we can expect the symmetry element to be retained. However, attack at one side or other of the mirror plane will naturally destroy it. Ketone 65 has mirror plane that slices through the ring 65a. The mirror plane contains the carbonyl and the central carbon of the tertbutyl group. If we attack the carbonyl with a nucleophile, the nucleophile will have a line of attack that is within the mirror plane and the resulting alcohol 67 will retain the mirror plane. This is not the case if we make an enolate 66. The proton must be removed from one side or the other and this will destroy the mirror plane. Notice that when this happens a chiral centre is formed where the tert-butyl group connects to the ring.

O

O

mirror destroyed

mirror plane O Nu–

Base

t-Bu 65

mirror Nu OHintact

Chiral centre 66

65a

67

Step No. 8 Is there an asymmetric aspect to the reaction? Are we, for instance, attacking one of two enantiotopic groups selectively? Step No. 9 How does the product differ to the starting material? If necessary, apply step numbers 1 to 6 to the product too. It might be worth explicitly stating the ‘obvious’ that, in the very simple conjugate

20 Some Principles

389

addition of a nucleophile to ethyl cinnamate 68, we started with an achiral planar material but the product 69 is chiral and racemic. O

Nu

O

Nu

OEt

OEt

68

69

Summary So then, to summarise, here are the nine steps of stereochemical analysis — 1. 2. 3. 4. 5. 6. 7. 8. 9.

Look for symmetry in the starting material Identify relationships within the molecule Decide whether or not the molecule is chiral Decide whether or not it is enantiomerically pure Identify the source of chirality and indicate the chiral centres or axes Look for more subtle symmetry relationships Decide how the reaction influences symmetry Identify any asymmetric aspect to the reaction Consider how the product differs to the starting material. Reapply steps 1 to 6.

Applying these nine steps to the simple nucleophilic conjugate addition to a cinnamate ester 68, we can summarise our findings.

1

4-6

mirror plane O

2

enantiotopic faces

OEt 68a

3

Do not apply

7

Mirror plane is destroyed

8

Reaction not enantioselective

9

Product has a chiral centre

racemic product 69

achiral molecule

Some Principles In this section we look at the influence of some numbers on the outcomes of reactions.

Chiral auxiliaries In Chapter 27 (Substrate Based Strategy) we shall meet the use of chiral auxiliaries. If you ask someone what properties a good chiral auxiliary should have you will get all the usual suspects – cheap, both enantiomers available, easy to put on, easy to take off, effective at induction etc. However, there is one property that gets very little attention. After the reaction the diastereomers should be separable. Suppose we have a chiral auxiliary 70 with an enantiomeric excess of 96% which reacts

390

20 Control of Stereochemistry – Introduction

to produce the product 71 in 74% diastereomeric excess. What is the ee of the product 71? If the product were hydrolysed without separating the two diastereomers of 71 what would be the enantiomeric excess of the liberated acid 72? If the two diastereomers were separated before hydrolysis what would be the ee of the liberated acid 72? Try to work the answers out for yourself before you look at the answers. O

O N

O

O O

reaction

N

O O

reaction

OH

Ph 70; 96% ee

Ph 71; 74% de

72

The answer to the first question is easy. The ee of 71 is 96%. The ee of the ‘wrong’ diasteromer (which we’ll call 73) is also 96%. The answer to the second question is a bit surprising. It is 71% ee. The best way to explain this is to draw out all four possibilities. A 96% ee means we will have 98% of one enantiomer 71a and 2% of the other 71b. And a 74% de means that we will have 87% of one diastereomer 71 and 13% of the other 73. The ratio of 71a to 71b will be 98 to 2 as will the ratio of 73a to 73b (keep your attention on the stereochemistry in the ring here – this will not change regardless of the reactions going on nearby). O

O N

O

O O

N

O

98 Ph

Ph

0.853

71a enantiomers

O

O N

2

0.127

73a O

O O

N

Ph

O

Ph

0.017

71b

87

0.003

73b

diastereomers

13

The ratio of 71a to 73a is 87 to 13 but the same is true with the other enantiomer in the reaction so the ratio of 71b to 73b is also 87 to 13. The relative amount of each of the four structures can then be determined. For example, the proportion of 71a is 0.98  0.87  0.853. If the compounds are not separated before hydrolysis then everything will be included. One enantiomer of the acid 72 will result from 71a and 73b (0.853
0.003  0.856) and the other from 73a and 71b (0.127
0.017  0.144). The ee is 0.856 – 0.144/0.856
0.144  71.2%. This might surprise you as it is lower than either the de or the ee of the reaction! It is a different story if the diastereomers are separated first. In this case both 73a and 73b will have been thrown away. The ee is therefore 0.853 – 0.017/0.853
0.017  96%. This is an important lesson. It shows how important separating the diastereomers is. As long as you can do

20 Some Principles

391

this, your de could be 0% (which would mean your auxiliary was a resolving agent) and you could still get an ee of 96% or whatever the ee of the auxiliary was. This brings us to the next point, your ee will only be as good as the ee of the auxiliary, so get 100% if you can.

The Horeau principle We saw above how two numbers conspired to lower the quality of the product. In this next example the reverse happens. We want to improve the ee of 36. Suppose we have alcohol 36a and its enantiomer 36b in an 85:15 ratio, an ee of 70%. We’ll represent this by the area in the box. OH

36a

36b OH

Ph

70% ee 85:15 ratio of enantiomers

Ph

If the alcohol is reacted with oxalyl chloride then two molecules of alcohol react with each molecule of oxalyl chloride. Two diastereomers result. One has enantiomers 75a and 75b and the other is achiral 74. Assuming the molecules react statistically (and there is no chiral recognition) then the proportion of 75a should be 0.85  0.85  72.25%. Here’s where the method works – most of the unwanted enantiomer gets incorporated into the other diastereomer 74 (2  0.85  0.15  25.5%). Only a small amount of the enantiomer of 75b is produced (0.15  0.15  2.25%). 36a OH

36b OH

Ph

Ph

70% ee 85:15 ratio of enantiomers

Ph OH 36a

Ph

O

O

Ph

O

O O

Ph 74

O

Ph 74

O

Ph

O

O O

Ph

75a OH

36b

O Ph

74

75b

75b

O

The diastereomer 74 is removed to leave 75a and the small amount of 75b in a ratio of 97:3. If these esters are cleaved then we have the alcohol 36 with an ee of 94% which is quite an

392

20 Control of Stereochemistry – Introduction

improvement. This is done at the expense of yield of course as some of the alcohol we wanted was thrown away in diastereomer 74.

76

OH Ph

wrong diastereomer (74) separated and thrown away

O O

O

Ph

O

Ph

O

75a Successful Application 92% ee → 99.6% ee 78% yield

O

O 97 : 3 ratio of enantiomers 75b remains. 94% ee

Ph

O 75b

This method was successfully used8 to improve the enantiomeric excess of alcohol 76 from 92% ee to 99.6% ee. The combination of substrates that are already of quite good enantiomeric excess to give a product of even better ee is not uncommon and is sometimes referred to as the Horeau Principle.9 On the way to the antibiotic FR-900848, an intermediate with four contiguous cyclopropane rings 80 was required.10 Cyclopropane 77 was made in 88–90% ee. Dimerisation of 77 yields a product of 98% ee for the same reasons as outlined above – if the correct enantiomer of one compound combines with the wrong one of the other then a different diastereomer is produced and removed, thus removing the wrong enantiomers. Only on the rare occasions that two wrong enantiomers combine do they form the enantiomer of 78 and slip through the net to the next round. Dimerisation of 79 improves the ee again, this time to 99.9%.

SnBu3 + Bu3Sn

TBDPSO 77

OTBDPS

88-90% ee

i) sec-BuLi ii) [ICuPBu3]4 iii) O2

OTBDPS

TBDPSO

98% ee 78

SnBu3

TBDPSO

Br 79

= TBDPSO

TBDPSO

SnBu3

i) tert-BuLi ii) [ICuPBu3]4 iii) O2

> 99.9% ee OTBDPS

TBDPSO 80

The authors of the above work specifically mention the Horeau principle but its operation is often less explicit. Take for example the production of acid 84 which can be produced in 90–92% ee by the reaction of optically pure diester 83 of hydrobenzoin 81 in a Diels-Alder reaction.11

20 Popular Misconceptions

393

Since two asymmetric reactions are going on the ee of the product will be enhanced (so long as the diastereomer of 83 is separable and removed the auxiliary is forgiven if it only gets it wrong once – it has to get it wrong twice for the wrong enantiomer to find its way through). The acid is cleaved from the hydrobenzoin which is recycled. Acid 84 was needed in part of a synthesis of FK-506. In fact the authors say that the 90–92% ee material is formed from the reduction of the crude cycloadduct 83. In which case the Horeau principle cannot be operating as we intimate above unless the authors somehow inadvertently purified their cycloadduct without realising during, for example, the filtration through silica. O H O

OH Ph

O Ph

Ph Et3N

OH 81

LiOH MeOH H2O

O

Cl

Ph

O

O 82; O >98% yield

TiCl4

H

O OH

Ph

Ph

84; 90-92% ee

O 83 O

H

You might like to work out what the selectivity is in each cycloaddition reaction. To give you a clue, look back at the graphical diagrams used for diester 84/85 and think about using square roots. We shall see this sort of double reaction to enhance selectivity again in Chapter 28. One reaction is a kinetic resolution and the reagent reacts twice. On the second pass it removes the small amount of the wrong compound that was formed the first time round.

Popular Misconceptions Misuse of words ‘Chiral’. For the purist, this is an especially irritating misuse. A worker who has been using racemic stilbene oxide is about to move onto using enantiomerically pure stilbene oxide. He describes the enantiomerically pure stilbene oxide as ‘chiral stilbene oxide’ or more probably, in the abreviated colloquial language of the lab as ‘my chiral compound’. Racemic stilbene oxide is no less chiral than optically pure stilbene oxide and so ‘chiral stilbene oxide’ says no more than ‘stilbene oxide’. ‘Chiral’ is sometimes used in this way even by experienced and distinguished chemists! ‘Chiral Synthesis’. For the same reason as above, a reaction that features racemic materials is no less chiral than one with optically pure materials. It should not, therefore, be refered to as a ‘chiral synthesis’: ‘asymmetric synthesis’ is a better term. ‘Racemic Synthesis’. We’ve heard many people object to this term but we really don’t know why. If you can have an asymmetric synthesis, then why can’t you have a racemic synthesis? The objection runs along these lines - ‘it is the compounds that are racemic and not the synthesis and so it should not be refered to as a racemic synthesis’. We suppose so, but the same is true for asymmetric synthesis isn’t it? Presumably we should not have a ‘comfortable armchair’ since it is the person who sits in the armchair who is comfortable and not the armchair itself but we would be allowed to have a ‘warm house’ since the house is indeed warm. Objections to ‘racemic synthesis’ are dogmatic in that they deny the transferred epithet (as in comfortable armchair) which is a common feature of English. ‘Homochiral’. Is sometimes used to mean ‘optically pure’ and both the authors have used it this way. But this use to mean optically pure should probably be avoided since ‘homochiral’ has another meaning; if two molecules have the same absolute stereochemistry they can be said

394

20 Control of Stereochemistry – Introduction

to be homochiral. Hence if we say ‘those two aminoacids are homochiral’ then we have two different meanings depending on how we are using ‘homochiral’. ‘Enantiomerically pure’ is unambiguous.

Chiral centres and stereogenic centres There has been a move away from describing something as a ‘chiral centre’. The argument is that a centre, a point, cannot be chiral. Instead it is refered to as a stereogenic centre. If you replace your ‘chiral centres’ with ‘stereogenic centres’ you will be on safe ground but you should be aware that they are not necessarily the same thing. In the way that all elephants are mammals but not all mammals are elephants so all chiral centres are stereogenic centres but not all stereogenic centres are chiral ones. The dimethylcyclohexanes 85 and 86 illustrate a simple case. Neither compound is chiral as they both contain a mirror plane. There are no chiral centres. However, changing the configuration of the carbon atom where the methyl joins the ring in 85, gives a different (achiral) diastereomer 86. The centres are thus stereogenic (as they give rise to other stereoisomers) even though they are not chiral centres.

86

85

Pseudo-asymmetric centres We can continue with the above theme but things start to get a bit more complicated. Consider the triol 87. It is not chiral and might be described by many as a meso compound. We’ll come back to this compound but for now look at the central carbon atom. This is not a chiral centre and not an asymmetric centre. You can see this if you try to decide whether it has an R or S configuration. You immediately run into trouble because two of the groups on either side are the same (enantiotopic). This means we cannot give one a greater priority than the other as we usually would.

OH

OH S

R

OH

R

r

87

OH

OH S

R

3

2 OH

OH S

1 OH 87a; (R, r, S)

OH 88; (R, s, S)

It is, in fact, r (note lower case). In this situation the group with the R configuration is given greater priority than the one with S. Those priorites in 87a would mean an R configuration if it were a real chiral centre but since it isn’t we call it r. If we change its configuration (to s) we get a different meso diastereomer 88. In other words we get another achiral compound. Whatever we do with that centre we get achiral diastereomers so it can scarcely be asymmetric—it is not an asymmetric centre. However, when we change the configuration we do get different diastereomers so it is certainly stereogenic (gives rise to stereochemistry). So we see, something can be a stereogenic centre without being a asymmetric, or chiral, centre. Such centres are sometimes referred to as pseudoasymmetric centres. In a detailed and highly cited paper, Mislow points out

20 Popular Misconceptions

395

that chirotopicity (something that is chirotopic is in a chiral environment) and stereogenicity are conceptually distinct. The centres above could be referred to as achirotopic (because the molecule is not chiral they are not in a chiral environment) and stereogenic centres.12

Unhelpful drawings Let’s look at a diastereomer 89 of the triol 87 that we have just met. Again we focus on the central carbon. Imagine trying to make the diastereomer of 89 by changing the configuration at the central atom (89 changes to 89a). We now rotate 89a by 180 and find that 89b is identical to 89. Whatever we do with the central OH (whether we have it coming forwards or going backwards) makes no difference at all. In other words there is no stereochemical information at all. Triol 89 should be drawn 89c with an ordinary bond to the central OH. rotate 180˚

OH

OH

OH

change configuration

OH 89

OH

OH

OH 89a

OH 89b

OH

correct drawing OH OH

OH 89c

This lack of stereochemistry is typical of substituents that lie on a pseudo-C2 axis and compound 60 was drawn correctly without stereochemistry to the central phenyl group. However, the central substuent does disrupt the C2 axis that would otherwise be there. Consider the two hydroxyl groups on the left and right of the molecule. One of them will be on the same side of the molecule as the central hydroxyl and one will not. The hydroxyl groups on the left and right of the molecule are diastereotopic whereas they would be enantiotopic if the central hydroxyl were not there (because they would be related by the C2 axis)

Bond rotation The misconception is that if a bond rotates fast enough, protons interchanging position by that rotation will be the same. For example, we have already discussed diastereotopic protons in molecules like 90. The diastereotopic nature of the protons HA and HB is quite obvious here because it is clear that one will be on the same side of the five-membered ring as the phenyl group and the other will not be. O 90

HN Ph

O

H2N

OH

Ph

HB HA

91 HB

HA

The protons in 91 are diastereotopic too. We can see this easily with the diastereotopicity test we have discussed earlier. Nevertheless, there is sometimes a misconception that, because we now have bond rotation, they are equivalent. After all, with 91 (unlike 90) we can swing proton HB into the position occupied by HA and vice versa. But the rotation itself, changes the environment too – putting HB where HA was also puts the OH in a different place, so HB does not experience

396

20 Control of Stereochemistry – Introduction

the environment that HA did. Rotating the bond faster will not help as the environment will be changed at exactly the same rate! As a dog cannot catch its own tail so we cannot get away from the diastereotopicity.

Stereospecificity revisited If you find yourself in a group of organic chemists and the conversation seems to be flagging a bit, try throwing in the word ‘stereospecific’ and before you know where you are you will find yourself in the middle of a lively debate about what the word means. We have already established in the earlier section that stereospecificity is not merely 100% stereoselectivity and we won’t look at that again here, but there are other issues for chemists to argue about. The situation is nicely illustrated by the reactions of the Schwartz reagent [Cp2Zr(H)Cl] 93. This reacts across alkenes by hydrozirconation to give a species 94 with a carbon–zirconium bond. This bond can be cleaved by adding water (which adds a proton to the carbon as we would expect with an organometallic) or electrophiles such as halides. It can be a very useful reaction. H H

Cl

Zr

+

Br

ZrCp2

hydrozirconation

Br2

93

92

Cl

94

95

In one example alkyne 96 is activated by hydrozirconation. The zirconium is in the terminal position 99. The two Cp rings are quite big and so it is usual for the zirconium to add preferentially to the less hindered end of the triple (or double) bond. This is the regioselective aspect of the reaction and not what concerns us here. Transmetallation to the organolithium and again with Cu(I) gives the cuprate and this cuprate does a conjugate addition to the double bond of an enone that looks suspiciously Robinson annelation-derived.13 Before we move on, notice that the double bond has a trans geometry – we shall come back to this.

OSiR3 96

1. Cp 2ZrHCl 2. 3eq MeLi

Me

O ZrClCp2 H

3. CuCN•LiCl

Me

O

4. 97

OSiR3

OSiR3

O

99 O 97

98

Now then, the hydrogen and zirconium are added from the same side of the molecule at the same time. If we have a molecule of styrene that is deuterated we can detect this. If we start with the trans-dideuterated styrene 100 we get one diastereomer 101. But if we start with the cisdideuterated styrene 102 we get the other diastereomer of product 103. It is a similar story with a substrate with a triple bond. With phenylacetylene 104, addition of hydrogen and zirconium occur from the same side. Thus after reaction of the organozirconium species we have a trans double bond as we observed with 98 and 99.

20 Popular Misconception

D

H Cp2Zr(H)Cl

D

93

D

Cl

Cp2Zr(H)Cl

Cl ZrCp2

D Ph

93

HD

D H 101

102

Ph 93

103 H

Cp2Zr(H)Cl

104

H D

ZrCp2

D Ph

100

397

Ph

Cl ZrCp2

105

Some chemists would argue that the reactions of the double bonds are stereospecific, but the reaction of the triple bond is not. Their argument is that, for a reaction to be stereospecfic, a specfic diastereomer of starting material must lead to a specfic diastereomer of product and this implies that there is an alternative diastereomer of starting material which would give the alternative diastereomer of product. BUT, since the actylene has no diastereomer, this cannot be stereospecific. Other chemists would argue that this is a load of rubbish and that both are examples of stereospecificity. We should be concerned with the mechanistic element of the reaction. We want to know how the zirconium complex reacts, and this is what the stereospecificity refers to; delivery of hydrogen and zirconium to the same side of a multiple bond. The first group of chemists would probably come back with the argument that in order to probe the mechanism in the first place you need two diastereomers of starting material and how can we be sure that the reaction with the triple bond goes by the same mechanism – and so on. Perhaps we should mention that Schwartz himself describes the addition across a triple bond as stereospecific.14 The definition used by Eliel15 says that something is stereospecific when ‘starting materials differing only in their configuration are converted to stereoisomerically distinct products’. He goes on to talk about changing the catalyst instead of the substrate and to discourage the meaning ‘highly stereoselective.’ This definition embraces the meaning of stereospecificity in the double bond scenario but not its extension to the triple bond. This denies the mechanistic spirit of the word stereospecific. To return to an ordinary SN2 reaction of, say, an alkyl halide with an unspecified nucleophile. We know SN2 reactions always go with inversion and that an SN2 reaction is stereospecific. The reaction goes with inversion because the antibonding orbital is populated as the reaction proceeds and there is no other option. Is this stereospecific because it is defined by the population of the antibonding orbital AND because there is another enantiomer of alkyl halide for us to consider or is it just because the antibonding orbital defines the process exactly? This is turning into a bit of a soap box but we would suggest that it is more instructive to use the latter definition. Thus the different stereospecific outcomes from different starting material isomers is a consequence of stereospecificity of the reaction and not the other way round. Perhaps the madness of needing to have another diastereomer of a starting material in order for a reaction to be stereospecific is more obvious in the next example. Heptene 106 may be dihydroxylated stereospecifically to give 107. But the reaction of cycloheptene 108 would not be a stereospecific reaction because there is no such thing as a trans cycloheptene. Clearly this is crazy – both reactions are stereospecific. But see what your friends think.

HO

106

HO

OH

OsO4

OsO4

stereospecific

stereospecific?

107

108

OH

109

398

20 Control of Stereochemistry – Introduction

References General references are given on page 893 1. Clayden, Chapter 16. 2. Clayden, page 1230. 3. Eliel, page 119. 4. S. Atarashi, S. Yokohama, K. I. Yamazaki, K. I. Sakano, M. Imamura and I. Hayakawa, Chem. Pharm. Bull., 1987, 35, 1896. 5. H. Egawa, T. Miyamoto and J. Matsumoto, Chem. Pharm. Bull., 1986, 34, 4098. 6. L. A. Mitscher, P. N. Sharma, D. T. W. Chu, L. L. Shen and A. G. Pernet, J. Med. Chem., 1987, 30, 2283. 7. S. Atarashi, H. Tsurumi, T. Fujiwara and I. Hayakawa, J. Heterocycl. Chem., 1991, 28, 329. 8. I. Fleming and S. K. Ghosh, Chem. Commun., 1994, 99. 9. J. P. Vigneron, M. Dhaenens and A. Horeau, Tetrahedron, 1973, 29, 1055. 10. J. R. Falck, B. Mekonnen, J. R. Yu and J. Y. Lai, J. Am. Chem. Soc., 1996, 118, 6096. 11. J. A. Marshall and S. Xie, J. Org. Chem., 1995, 60, 7230. 12. Eliel p. 53 and 123; K. Mislow and J. Siegel, J. Am. Chem. Soc., 1984, 106, 3319. 13. B. H. Lipshutz and E. L. Ellsworth, J. Am. Chem. Soc., 1990, 112, 7440. 14. D. W. Hart, T. F. Blackburn and J. Schwartz, J. Am. Chem. Soc., 1975, 97, 679. 15. Eliel, page 1008.

21

Controlling Relative Stereochemistry Note Before We Start Introduction PART I – NO CHIRALITY IN PLACE AT THE START Formation of Cyclic Compounds via Cyclic Transition States Cycloadditions Formation of Acyclic Compounds via Cyclic Transition States Stereoselective aldol reactions Nomenclature Early work The transition state model Getting control – ketones Getting control – esters Reactions of Cyclic Compounds Conformational Control with Six-Membered Rings Addition of nucleophiles to cyclic ketones Reactions of exocyclic enolates PART II – CHIRALITY IN PLACE FROM THE START Using one chiral centre to control another Reactions of Cyclic Compounds: Conformational Control The half-chair Reactions of endocyclic enolates Opening epoxides Atoms move towards each other when a bond is made Microscopic reversibility Formation of a Cyclic Intermediate Iodolactonisation Predicting stereochemistry Application of an iodolactonisation reaction Variations on this theme Formation of an Acyclic Compound via a Cyclic Transition State Claisen rearrangement The Ireland-Claisen rearrangement Stereoselective Reduction of β-Hydroxy Ketones Syn selective reductions Anti selective reductions The Evans-Tishchenko reduction Kinetic diastereoselection of 1,3-diols Aldol revisited Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

400

21 Controlling Relative Stereochemistry

Open Chain Chemistry - with Chelation Control Grignards Zinc borohydride Open Chain Chemistry - in its Most Genuine Form Felkin control over reactions of nucleophiles with carbonyl compounds Role of electronegative substituents The Houk conformation for reactions of alkenes with electrophiles Allyl silanes

Note Before We Start: We assume you already know about conformational control in making new chiral centres and about folded molecules preferring to react on the convex or exo face. You are referred to Clayden chapters 33 and 34 for a basic treatment of this material.

Introduction Most of the compounds we shall be looking at in this chapter will be in racemic form. We are concerned only with the control of relative stereochemistry and not with the control of absolute stereochemistry. However, many of the reactions have been developed into asymmetric versions. It is certainly true that many of the reactions have been employed within asymmetric synthesis – that is, where the asymmetric part has come from elsewhere and this idea will be revisited in Chapter 30. If we are to concern ourselves simply with relative stereochemistry then, for there to be any stereochemical relationship, we must have at least two chiral centres. If there is no chirality in the starting materials this means that two chiral centres must form in one reaction and if there is only one new chiral centre that forms in the reaction then there must have been a chiral centre already in one of the starting materials. Some sort of ‘cyclic aspect’ is usually a feature of good stereocontrol. This aspect might be in the form of a cyclic transition state or perhaps in the generation of a genuinely cyclic compound but in any case a cyclic aspect is associated with fewer degrees of freedom – or less floppiness – in whatever system we are considering which usually means better control. Without trying to be exact, the cyclic aspect of reactions broadly decreases down the series – • • • • •

Formation of cyclic compounds via cyclic transition states Reactions of cyclic compounds Cyclic intermediates Formation of acyclic compounds via cyclic transition states Open chain stereochemistry

Each of the above topics can be looked at either with no pre-existing chirality, or with a chiral element already in one of the molecules.

PART I – NO CHIRALITY IN PLACE AT THE START Formation of Cyclic Compounds via Cyclic Transition States Cycloadditions Cycloadditions are supremely powerful when it comes to the controlled formation of many stereogenic centres. The stereochemistry of the products can be the result of many different stereochemical aspects. Though strictly this example does not have a place in this chapter (because relative stereochemistry is not generated), it is instructive.

21 Formation of Cyclic Compounds via Cyclic Transition States

CO2H

401

CO2H

+ 1

2

3

This is a straightforward Diels-Alder reaction. The product is the racemic acid 3. The faces of the dienophile 2 are enantiotopic – if the diene attacks from one face 4a we get one enantiomer (R)-3 and the other (S)-3 if it attacks from the other 4b. However, it doesn’t matter at all which face of the diene 1 reacts. Its faces are homotopic. CO2H

CO2H

CO2H

(R)-3

4a

CO2H

(S)-3

4b

The Diels-Alder reaction is, of course, an excellent way of introducing several new chiral centres at once. The E,E-diene 5 and unsaturated acid 6 can be expected to give the adduct 7. Note that we move from no chiral centres in the starting material to four chiral centres. The product is racemic (±)-7 since we have made no effort to make a single enantiomer.

R

No Chiral Centres

R

Four Chiral Centres

+ HO2C 5

CO2H (±)-7

6

The selectivity in the above reaction comes from two sources. Firstly there is the stereospecific aspect. This demands that stereochemistry in the starting materials shows up in the products. That is, the trans relationship between R and CO2H in the unsaturated acid 6 shows up in the product 7 as a trans relationship. The same stereospecific dictat ensures that the two methyl groups of the diene 5 end up on the same side in the product. This is less easy to see with a diene, but we would expect the methyl groups to be trans to one another in the product only if we instead started with the E,Z-diene. Secondly, there is the stereoselective aspect to the reaction - often referred to in this context as endo selectivity. We will not go into why this happens here1 but with normal endo selectivity, the electron-withdrawing group of the dienophile will prefer to be close to the electronrich diene. When 5 reacts with 8, we know that the two methyl groups must be on the same side of the molecule and that the two CO2Me groups must be on the same side of the molecule. But two structures 9 and 10 answer this specification. Which is formed?

CO2Me

CO2Me

?

CO2Me 5

CO2Me or

+

8

CO2Me 9

CO2Me 10

402

21 Controlling Relative Stereochemistry

Predicting the stereochemical outcome of a Diels-Alder reaction is easy once you know the likely transition state. The trick is to identify all the groups that are on the same side of the sixmembered ring. The figure below shows one way to do this with the bonds on the right hand side exaggerated and placed in a box. Once the new six-membered ring has been made then these groups stay on the same side. In this case all the hydrogen atoms are on the same side: the product is 10.

H MeO2C endo selectivity operating here MeO2C

H H

H

H H

H H

groups on the same side of the forming six-membered ring

H

H

H

H

CO2Me CO2Me

10

groups on the same side of the newly-formed six-membered ring

Many Diels-Alder reactions have both stereospecific and stereoselective aspects. It is as important as ever at this point not to confuse perfect stereoselectivity with stereospecificity. So, let’s look at a real example. Pumiliotoxin C 11 (a toxin from a tropical frog) contains four chiral centres and in the following synthesis, three of them were controlled by a Diels-Alder reaction.2 The retrosynthetic analysis is quite instructive. The first disconnection 11b (reductive amination) takes no consideration of the chirality at all – we will worry about this later. We already have a six membered ring and it now contains all the chirality. The introduction of double bonds by FGA is used to considerable effect in this retro-synthesis. The first new alkene allows us to do an aldol disconnection 13 that not only removes a carbon chain but also reveals an electron-withdrawing group on the six-membered ring 14. This is, of course, one vital component of the Diels-Alder reaction. The next FGA puts in the double bond that allows us to do the retrosynthetic Diels-Alder step 15. Me

H H H

Me H

N H

FGA

amination

11a; pumiliotoxin C

N H H H 11b; pumiliotoxin C

Me

Me

O

H

Bu

H

H

reductive

NH2 13

H

aldol

12 Me

CHO

Bu

O NH2

H

FGA

CHO

CHO

DielsAlder

17 H 14

NH2

H 15

NH2

NH2 16

The trans relationship between the methyl group and the aldehyde 15 must be preserved in the unsaturated dienophile 17. The other component (the primary enamine 16) will clearly not be a stable compound and something will have to be done about that in the forwards route. The reagent used for enamine 16 is carbamate 22 which is made3 from unsaturated acid 18.

21 Formation of Cyclic Compounds via Cyclic Transition States

1. ClCO 2Et

CO2H

N

N

403

110 ˚C

N

N

2. NaN 3

O 19; acid azide

18

Curtius

N

rearrangement

C

O

O 20; acyl nitrene O

Ph

OH

N H 22

21; isocyanate

O

Ph

We are now ready for the key step in the sequence which is the reaction of the carbamate 22 with the unsaturated aldehyde 17 to give the Diels-Alder adduct 23. H O N H

H CHO

CHO

O

Ph

17

NHCO2CH2Ph H 23, 67% yield

22

This adduct 23 contains three of the four stereocentres and we should look a little more carefully at the reaction that has just occurred. The two molecules 22 and 17 could have approached each other in two ways which correspond to endo and exo selectivity. In the endo case, on the left, giving 23, the anti relationship between the methyl group and the formyl group stems from the trans geometry of crotonaldehyde and is stereospecific. The syn relationship between the formyl group and the carbamate comes from the endo selectivity in the transition state and is stereoselective. Had the two components of the cycloaddition chosen to come together in the alternative exo way (on the right) then there would have been an anti relationship here and 24 would have been formed. exo

endo syn

H H CHO

HH

NH

H

CO2CH2Ph

CHO NH CO2CH2Ph syn stereospecific aspect

anti syn H

anti anti

stereoselective aspect

H

H CHO NHCO2CH2Ph

H endo-adduct 23

exoadduct 24 (not formed)

CHO H NHCO2CH2Ph H

404

21 Controlling Relative Stereochemistry

We have now discussed the main point of interest but we may as well finish off the synthesis. Reaction of the aldehyde 23 with a phosphonate salt gives the unsaturated compound 25. This is the Horner-Wadsworth-Emmons reaction described in chapter 15. Notice that the compound has two double bonds that we do not actually want – they were there to aid the synthesis and have now done their job. O

H

H

O

CHO NHCO2CH2Ph

O

H

(MeO)2P NaH

NHCO2CH2Ph H 25, 80% yield

H 23

The final step is hydrogenation which removes both of these unwanted double bonds. But the hydrogenation does much more than this! For a start the nitrogen is deprotected 27 and exercises its nucleophilic muscles once the α,β-unsaturation has been removed (it cannot reach the ketone until this happens). The result is the imine 28 which can also be hydrogenated. Now the final stereocentre is introduced. The imine 28 is hydrogenated from its less hindered face. This is an example of a folded molecule reacting on its convex face. There are several examples of diastereoselective reactions on folded molecules in chapter 17.

Me

O

H

H

H2, Pd/C HCl, EtOH

N H H H 26; 95% yield

NHCO2CH2Ph H

Me

25

Hydrogenate from this face

H

H imine

H

N

Bu

formation

O H NH2 27

28

Formation of Acyclic Compounds via Cyclic Transition States Stereoselective aldol reactions From a stereochemical point of view, aldol reactions can provide us with another one of those fiat lux moments – stereochemistry appears to spring from nothing. As in the Diels-Alder reaction, the stereochemistry comes from two places. Firstly, there is the geometry of the starting materials and secondly, there are choices that are made in the transition state (remarkably like the Diels-Alder reaction in fact). Let us focus on the geometry of the starting materials—whether enolates have a cis or a trans geometry.

21 Formation of Acyclic Compounds via Cyclic Transition States

405

Nomenclature Just before we go any further we need a quick note on nomenclature. We see below a lithium enolate 29 and a titanium enolate 30 of an ester. By definition, the lithium enolate 29 contains a Z double bond (lithium is lighter than carbon) whereas the titanium enolate 30 contains an E double bond (titanium is heavier than carbon). But to call one E and the other Z is plainly crazy. The important thing, as far as we are concerned, is that the substituent R is trans to the reactive oxygen. In order to avoid confusion we shall refer to both of these as trans enolates regardless of the metal 31. This is not universally done (some chemists will be outraged by such simplicity) but is what we shall be doing in this book.

O 29 Z 'trans'

Li

OMe R

TiCl3

O 30 E 'trans'

OMe

O 31 E or Z 'trans'

Metal

OMe

R

R

Early work The pioneering work of Dubois4 showed that trans enolates 32 of cyclopentanone react to give selectively anti aldol products 33. The enolates have to be trans because the substituent is locked into a ring. Heathcock showed that ketone 34 (note in particular the tert-butyl group on the left) gave the cis enolate 35 and mostly syn products5 36.

O

OLi

O

OH

i-PrCHO

H >95% anti-33

trans-32 O

OLi

O

OH

PhCHO

Ph 34

cis-enolate 35

>98:2 syn : anti 36

The transition state model So, trans enolates give anti aldol products whereas cis enolates give syn aldol products. Why? The enolate and the electrophile fit together in the transition state into a six-membered ring. Chemists often refer to the model as the ‘Zimmerman-Traxler’ transition state.6 The six-membered ring is in a chair conformation. Consider the reaction of some aldehyde RCHO with a trans lithium enolate. The lithium, the oxygen atom, adjacent carbon atom and nucleophilic carbon atom of the enolate react via the six-membered ring transition state 37. The methyl group must go into an equatorial position as it is trans to the oxygen atom of the enolate and is forced into this position. Finally, there is one sp2 centre (ArOC) that stays as an sp2 centre from start to finish.

406

21 Controlling Relative Stereochemistry

O

Li

ArO

Li

O

O ArO

R

O

ArO

R

Me trans-enolate

H

sp2 centre

O

Me

Li

Me

equatorial position – choice –

O R

H

equatorial position – no choice – 37; conformation

37

The R group of the aldehyde, however, can choose to take up an equatorial position 37 or it could fold round the other way and go into the axial position 40. The favoured conformation 37 gives the anti-aldol 39 while the disfavoured conformation 40 would give the syn-aldol 42. H O Li ArO Me

O R

equatorial position – choice –

Li

O

FAVOURED

H

O

ArO

equatorial position – no choice – 37; conformation H O Li O BUT... H NOT FAVOURED axial ArO Me R position – choice – equatorial position – no choice –

O

aqueous work-up

ArO

R

Li

O

O would give

ArO

OH

ArO

R

R

syn-aldol 42

41

40

R

anti-aldol 39

38

O

OH

What about cis enolates? Consider this time the reaction of a boron enolate. As the lithium was in the six-membered ring last time so the boron atom is this time. Now the methyl group of the enolate is forced to adopt the axial position 43 but the aldehyde still chooses to have its R group in the equatorial position. The result is a syn aldol product 44. O

BR2 Me

PhS

Axial position - no choice -

O

Me O

R PhS

cis boron enolate

H 43

B O

equatorial position – choice –

R

O

OH

PhS

H

R

syn-aldol 44

If it is not obvious to you how the six-membered transition states gives rise to the stereochemistry in the aldol products then read on. The trick is to identify the carbon zig-zag in the six-membered transition state that corresponds to the zig-zag in the open chain compound. The way we drew the chairs allows for easy mental transfer of the stereochemistry. Firstly the conformation 37 giving rise to compound 38 is redrawn 37a but with the hydrogen atoms on the stereocentres emphasised. Now notice the front part of the six-membered ring – the zig-zag is highlighted in bold. The hydrogen atoms are quite clearly anti to one another and will be anti to one another on the zig-zag in the final molecule 39a. In this molecule forward bonds are shown by wedges. The absolute stereochemistry indicated in the transition state is the same as that in the final product (satisfy yourself that this is the case) but note that the zig-zags kink in opposite directions – we still have a little bit of work to do in making the mental transfer. H O Li O ArO Me

Li O

O H

R ArO

H

37 conformation

R H 38a

H O Li ArO Me

O

O

Li

O H

R ArO H

R H

37a showing anti H atoms 39a showing anti H atoms

21 Formation of Acyclic Compounds via Cyclic Transition States

407

Getting control – ketones It is all very well knowing what enolates give rise to what aldol products, but this is not much use unless we can choose to have cis or trans enolates. Clearly with cyclopentanone there is no option but to form a trans enolate. But with an acyclic ketone we will need a little more care. Boron enol ethers provide an answer. Two common boron reagents used are 9-BBN chloride 46 (derived from 9-BBN 45) and another is dicyclohexylboron chloride 47. Other reagents include the corresponding triflates. There are also the chiral boron reagents that we shall meet later chapters. H

Cl

B

B B

B Cl 46; 9-BBN-Cl

46; 9-BBN-Cl

45; 9-BBN

Cl 47; c-Hex2B-Cl

Amazingly, different combinations of base, alkyl groups, and leaving groups on the boron give different enolate geometries. Broadly speaking, cis-enolates 49 are obtained by having a small alkyl groups on boron (like butyl), a good leaving group (triflate) and a more bulky neutral base (like i-Pr2NEt). In contrast, trans-enolates 48 are obtained with sterically demanding alkyl groups on boron (like cyclohexyl), a less good leaving group (chloride) used in combination with a small neutral base (Et3N).7 The methodology to obtain trans-enolates selectively8 was later in coming. We will avoid explanations about why these combinations work but suggestions are available.9,10 O

B(RBig)2

O (RSmall)2BOTf

i-Pr2NEt

Et3N

trans

B(RSmall)2

O

(RBig)2BCl

pentan-2-one

48

49

cis

Interestingly, although the 9-BBN group looks like a large group, it (like dibutyl) is often complementary to a dicyclohexyl boronyl system. Since the groups on the boron are ‘tied back’ it probably has fewer steric demands than one might initially imagine but also 9-BBN enolates readily undergo equilibration.11 Although the leaving group is important,12 even keeping boron’s leaving group the same and varying the nature of the alkyl groups can have dramatic effects.8 The reaction of propiophenone can thus be varied to give either anti 54 or syn 52 aldol products.

O 9-BBN-Cl

i-Pr2NEt

O

Me

2. H 2O2

Ph

OH Ph

Me 52; 98% syn

Ph cis enolate 51

Me Ph 50; propio c-Hex BCl 2 -phenone Et3N

O

1. PhCHO

B

c-Hex O

B

c-Hex

Ph Me trans enolate 53

O

OH

1. PhCHO 2. H 2O2

Ph

Ph

Me 54; 95% anti

408

21 Controlling Relative Stereochemistry

But is isn’t as simple as this. The stereochemistry of the enolisation depends also on the group on the other side of the carbonyl (phenyl in the above case). So, for example, while pentan-2-one gives mostly cis boron enolate 55, branched ketone 56 give mostly trans.13 The aldol reaction is immensely complicated as there are so many variables. However, all the fundamentals from cis and trans enolates to double stereodifferentiation can be found in a review by Heathcock.14

O

O

(c-C5H9)2BOTf

B(c-C5H9)2

O

B(c-C5H9)2 82 : 18

pentan-2-one

trans-55

cis-55

O

O

B(c-C5H9)2

B(c-C5H9)2

O

(c-C5H9)2BOTf

19 : 81 56

trans-57

cis-57

Ester 58 is a key intermediate in a synthesis concerned with swinholide A.15 We can see that two of the stereocentres could be controlled through the use of a syn-selective aldol.

MeO

MeO O

TBSO

O

O

syn selective

TBSO

HO

58

O

59

aldol

O

O

O

Indeed, the forwards reaction uses a boron triflate and a bulky base of the type we have seen in order to make the cis boron enolate and achieve exactly this control. There are, of course, two syn-aldol products possible here, 58 and 60, by virtue of the chiral centres present in the aldehyde fragment, and both do indeed form (in a 16:84 ratio). Trying to achieve selective formation of one of these syn diastereomers rather than the other syn diastereomer is beyond the scope of this chapter, even though that too is relative stereocontrol. It is complicated because it involves enantiomerically pure reagents in combination with the enantiomerically pure aldehyde and a match/mismatch issue. These issues are explored more fully in Chapter 30. Examples include combinations of chiral or achiral aldehydes with both achiral and chiral boron reagents. R

R

TBSO Bu2BOTf

59

O HO

TBSO +

O HO

i-Pr2NEt

O pentan-2-one

O

BBu2 cis boron enolate O syn-58

16 : 84

O syn-60

21 Formation of Acyclic Compounds via Cyclic Transition States

409

Getting control – esters With esters there is yet another variable that may be used – the nature of the esterifying alcohol. In fact, we saw the enolates resulting from esters 61 and 64 earlier in the chapter when we were discussing the transition states. The ester may have a hindered aromatic component16 61 or be a thioester 64 instead of an ester. O

Li

O

LDA

O

O ArO

O

61

R2B i-Pr2NEt

S

R2BOTf

Ph 64

R

anti product 63

trans enolate 62 O

OH

RCHO

O

O

OH

RCHO

S

PhS

R

Ph

syn product 66

cis enolate 65

Corey used an ester and a thioester in combination with an optically pure boron reagent in order to control both relative and absolute aldol stereochemistry.10 The optically pure boron reagent was busy with the absolute control (the details belong in chapter 27 but 69 was 94% ee and 71 97% ee by virtue of the chiral boron reagent). OBR*2

O

O

R*2BBr

t-BuO

Et3N

t-BuO 67

i-Pr2NEt

64

anti aldol 69 O

OBR*2

R*2BBr

PhS

93% yield Ph 98:2 anti:syn

t-BuO

trans enolate 68

O

OH

PhCHO

OH

PhCHO

PhS

PhS

Ph

93% yield 1:99 anti:syn

syn aldol 71

cis enolate 70

The advantage with esters is that once the stereocontrol has been achieved, the alcohol or thiol component of the ester may be replaced with something else entirely. That something else may not contain a heteroatom at all as we shall see in the next illustration. Serricornine 72 is a sex pheromone for the cigarette beetle and synthetic serricornine is used in the control of the beetle. We are interested in this instance in the stereochemistry between the γ- and δ-positions. To use aldol chemistry here would require a carbonyl at the β-position and it is conspicuous by its absence. We shall have to put it in. Disconnection 72 of serriconin and FGI leads to an acid 74 which is looking a little more like the kind of substrate we need for an aldol reaction.

O

β

OAc γ

enolate alkylation

O

OAc

O

OH

FGI

δ

72; serricornin

I +

pentan-2-one

HO 73

74; syn aldol required

410

21 Controlling Relative Stereochemistry

The synthesis starts with the thioester 64 to get the cis boron enolate 75 we need for the syn product. Once the adduct 76 has been made, the sulfur has done its work and can be removed. Transesterification with MeOH and HgCl2 and straightforward steps lead to iodide 78. Once this iodide has reacted with the enolate of pentanone, there will be no trace of the carbonyl that originally allowed the aldol chemistry to be performed. R2B

O

i-Pr2NEt

PhS

O

O

OH

EtCHO

PhS

PhS

9-BBN-OTf

64

etc

76; syn only

cis enolate 75 O Si

O 1. HgCl 2 / MeOH

O Si

I 1. DIBAL

MeO

2. TBDMSCl Im / DMF

2. TsCl, Et 3N 3. NaI

77

78

One noteworthy step is the control of the final chiral centre. You will have noticed that we did not worry about the stereochemistry at the α-position at all. Natural serricornine readily epimerises at that position and so the workers knew they could leave nature to sort that out later. After deprotection of 79 the mixture epimerised to serricornine 72. This synthesis by Baker and Devlin was used as the basis for their enantioselective synthesis of serricornine. To intrigue you, something other than PhS was used and this something could also be removed and was a chiral auxiliary.17 O Si

O

O 1. LDA

O

OH

Bu4NF

stand in CHCl3

2. 78

72 serricornine

72; mixture of epimers

79

Reactions of Cyclic Compounds Conformational Control with Six-Membered Rings

Addition of nucleophiles to cyclic ketones Cyclic compounds have fewer degrees of freedom than their acyclic counterparts and this means that not only can we expect a higher level of control, but the resulting stereochemistry is easier to visualise. So reduction of the achiral ketone 80 could lead to the syn alcohol 81 or the anti alcohol 82. The question is which do we get and why? OH

O "H

OH

"

or

80

81; syn

?

82; anti

In fact, either can be achieved by using different reducing agents18 but let us fi rst look at the issues concerned in conformational terms. Attack on the carbonyl group 80a will either be

21 Reactions of Cyclic Compounds

411

axial or equatorial leading to the two alcohols 81 and 82 Note that both 81 and 82 are achiral molecules—achiral molecules with stereochemistry. axial attack O

80a

thermodynamic product H

OH "H

"

equatorial attack

H

OH

axial alcohol 81; from equatorial attack

equatorial alcohol 82; from axial attack

Reducing Agent Li, NH 3, ROH

1

:

99

NaBH4

7 96.5

: :

93 3.5

L-Selectride

The product with the equatorial hydroxyl group 82 will certainly be the thermodynamic product because both of its substituents are in the equatorial position. Let’s look at what reagents yield what products. The combination of Li in NH3 will generate solvated electrons. The reducing agent is an electron. We will come back to this in a minute but for now notice that it is highly selective at producing the diequatorial compound. Conversely L-Selectride®, LiBH(s-Bu)3, is highly selective at producing the other diastereomer.19 At lower temperatures or with an even bulkier reducing agent ⬎99.5% 81 is formed. There is stereochemical control in this reaction but we must be clear about one thing – the tertiary butyl group is NOT responsible simply by steric hindrance. It is too far away to be of any direct consequence. Its function is to lock the conformation of the ring. It is the ring itself that provides the steric hindrance. Structure 80b shows that the there will be steric hindrance from the axial protons on the ring. Another way to look at this is to consider the plane of the ketone 80c/d. Some of you will find the ‘right and left’ distinction of 80c easier to grasp and some the ‘top and bottom’ distinction of 80d. Either way you should see that on one side is the rest of the ring while on the other side there are merely two protons.

more hindered axial approach H O H

O more hindered approach H H

more hindered approach H less hindered H approach t-Bu H H

less hindered 80b equatorial approach

t-Bu 80c

O

H H less hindered approach 80d

We can explain the result with the L-Selectride quite easily. It is a very large reagent and will thus approach from the less hindered side to give the compound with the axial OH. The result with the Li in NH3 takes a little more explaining. Small things are less affected by steric issues than large things. Some of us know from experience that while closing doors and windows is a very effective way of keeping cats out of your house, mice will get in through the tiniest of holes. Trying to block the holes up is simply a waste of time – they will find a way in! Similarly, while large reducing agents can be expected to respond to steric hindrance, small ones, if they are influenced at all, will be influenced by other things. With this in mind we should look back to the result with the electron (Li in NH3). This is the smallest reagent that can be! It will take no notice of steric constraints whatsoever but may well

412

21 Controlling Relative Stereochemistry

be influenced by other things. If the extra stability of the thermodynamic product shows up in the transition state of the reaction then we would expect this reaction to go faster. This is clearly what is happening with the electron. Although that appears to wrap it up, it may be that this argument is simplest, largely because the transition state is probably more like the starting material than the product and there are other factors at work.20 Summary (whatever the explanation): large reagents add equatorially and small ones add axially providing that the thing being added is smaller than the thing already there.

Reactions of exocyclic enolate The above concerns an electrophilic centre in a six membered ring. The reverse would involve a nucleophilic centre in a six membered ring. If anything the chemistry is now more straightforward and is not so influenced by the size of the electrophile. With the exocyclic enolate 83, equatorial attack is preferred.21 But if the enolate is too reactive then there is little selectivity. axial attack

O

R OLi

R

E

equatorial attack

product from E equatorial attack

preferred

83

84

In this section we are not concerned with chirality being in place before the reaction but there is an additional warning here. Even if a molecule does have a chiral centre in it, it may be that this centre has nothing whatever to do with the stereocontrol observed. Consider the two unsaturated ketones below. One of them 85 is chiral, one of them 86 is not. O

O 85 chiral

Nu

Nu

86 achiral

In both cases, a nucleophile will come in across the unsaturated side of the molecule. The two chiral centres at the bridgeheads of the chiral molecule 85 have nothing to do with this. The only consequence of the chiral centres will be that the product will be a chiral diastereomer whereas (with the achiral ketone 86) the product will be an achiral diastereomer.

PART II – CHIRALITY IN PLACE FROM THE START Using one chiral centre to control another

Reactions of Cyclic Compounds: Conformational Control The half-chair The first concept that needs to be introduced here is the half-chair. Assuming you know how to draw the chair form of a cyclohexane, then the half-chair is a way to describe cyclohexene. One half of the normal chair has been flattened – hence the term ‘half-chair’.

21 Reactions of Cyclic Compounds: Conformational Control

chair cyclohexane

cyclohexane

cyclohexene

413

half-chair cyclohexene

The usual way to draw a half-chair is with the double bond at the front of the six-membered ring 87. The two carbons that are furthest away from the double bond are less perturbed from their chair-like arrangement and the substituents are axial and equatorial. Those next to the double bond are not genuine axial or equatorial substituents and occupy what are called the ‘pseudo-axial’ and ‘pseudo-equatorial’ positions. Pseudo is indicated by a ‘Ψ’ symbol. We shall see in the next section how a half-chair can develop into a chair. Ψ-axial 87 substituents in a half-chair cyclohexene

Ψ-equatorial equatorial

axial

Reactions of endocyclic enolates The ketone 88 reacts with Me3SiCl and base to give the more substituted silyl enol ether 89. This has only one chiral centre and that is at the point where the tert-butyl group joins the ring. The silyl enol ether 89 is turned into a reactive lithium enolate by the application of MeLi. This enolate then reacts with the methyl acrylate to give 90 with the stereochemistry as shown.22 If the tertiary butyl group is in the equatorial position, which is to be expected, then the new group has been installed in an axial position 90a. It is this that needs to be explained.

Me3SiCl

O 88

CO2Me

1. MeLi 2.

Et3N

OSiMe3

CO2Me

89

O 90

CO2Me

O 90a

We need to consider how the conformation of the half-chair changes as it reacts. Enolate 91 is drawn as a half-chair with the tertiary butyl group at the back of the molecule in an equatorial position. If the electrophile attacks from the top face then, as a bond forms to the electrophile and the sp2 hybridisation changes to sp3, a chair 92 starts to take shape (for clarity, we’ve omitted the carbonyl double bond in the conformational diagram). We see that the electrophile occupies an axial position. However, if the electrophile attacks from below, then a twist boat 93 starts to develop instead. If this boat were flipped into a chair, the new electrophile would be in the equatorial position.22

414

21 Controlling Relative Stereochemistry

E

t-Bu

E

O 92 as chair

t-Bu

88

O 91

O 92 flip to chair

t-Bu O

E

E

=

t-Bu

t-Bu

E

E O 93 as chair

93 as twist-boat

So attack on one side leads to an axial substituent whereas attack on the other leads to an equatorial substituent (but only after the ring flips). Boats are, of course, much less favourable than chairs and so axial addition is the usual outcome as the chair is formed during the reaction and not afterwards. We shall refer back to this as the normal development of a half-chair. Summary: If the starting material is not a chair (it might be a half chair), the most important thing is to get a chair. The question of whether groups go equatorial or axial is less important. On the other hand, if the starting material is already a chair (it might be ketone 80), the most important question is whether groups or reagents are in equatorial or axial positions.

Opening epoxides We can look at how epoxides are opened in a similar way. The flattening of the six membered ring by the strain of the three membered ring 95a is not dissimilar to the flattening caused by a double bond 94. We can attack the end of the epoxide 95 (with generic nucleophile, Nu⫺) that leads to a chair 98 rather than the end that leads to a twist-boat 97. Compound 96 is the trans-di-axial product as both Nu and OH occupy axial positions. t-Bu

t-Bu

mCPBA

t-Bu

Nu

94 O 95 Nu

Nu

t-Bu

HO chair

boat

O

Nu

Nu

t-Bu

t-Bu 97

Nu 96

O 95a

O 98

Atoms move towards each other when a bond is made When you think about the development of a half-chair into a chair it is important to remember that atoms move towards each other as they form a bond. This sounds like rather an obvious statement but there is subtle psychology at work – although chemists rarely make the mistake when the half chair is the nucleophilic component, it seems all too easy to think absent-mindedly of an attacked electrophilic carbon atom in a half-chair retreating away from the incoming nucleophile – and this is not the case. The atoms move towards each other.

21 Reactions of Cyclic Compounds: Conformational Control

Nu

E

Cl

Nu

415

atoms move towards each other

HH R atoms move towards each other

atoms move towards each other

t-Bu

Nu O

Nu

t-Bu

E HH

95b

R

98

O

Microscopic reversibility Another way to look at the opening of epoxides uses the principal of microscopic reversibility. Microscopic reversibility is all about considering a reaction in reverse like watching a piece of film run backwards. You would expect the film running backwards to be the exact reverse of the film running forwards. A clown receiving a custard pie from the left of his face, will have it removed to the left of his face. We would never expect that, in reverse, it would disappear off to the right! The same is true of reactions – as then run forwards, so shall they run backwards. At every step of the way (frame by frame) they will be identical. The justification for this is that a reaction running forwards will run along the lowest energy pathway. Hence it will run back that way too. If there were a lower energy pathway on the way back then the forwards reaction would have found it and used it too! The lowest pass between two valleys is the same whichever way one is travelling. Why should we bother with this? For the simple reason that it is sometimes easier to consider a reaction in the reverse sense – easier to see the orbitals concerned, the stereochemistry involved, the conformation necessary or whatever. We will stick with the epoxide 95 we used in the half-chair analysis above and react it with RLi. Since the nucleophile is organometallic the reaction is not reversible. We should be clear about this point — the reaction does not have to be reversible to use a microscopic reversibility argument. The two conceivable products 101 and 102 are drawn (assuming SN2 reaction). R

t-Bu

O 99 or

RLi

O

R

t-Bu

aqueous work-up

OH 101 or

O

t-Bu

OH

95 t-Bu

R

t-Bu

100

R 102

We now draw the conformational diagrams 99a and 100a of the intermediates and consider the reverse reaction that would lead back to the same epoxide 95. In the reverse reaction, the alkoxide would have to come in behind the R group and displace it. R t -Bu

R

line t -Bu of attack O 99a

R

t -Bu O 99b

line of attack 100a

O

RO

t -Bu

X impossible 100b

416

21 Controlling Relative Stereochemistry

In one case this is clearly possible 99 and in the other it is not 100. And if R⫺ cannot be displaced by O⫺ then the forwards reaction cannot happen either. So, once again, the diaxial product is the product we would expect. It is not really necessary to get too involved with orbitals here as the alignment (or lack of it) is very evident. But if we wanted to we could think of this from an orbital point of view. The σ* of the C–O bond is populated when the epoxide is opened and goes into forming the lone pair on oxygen and the C–R σ-bond. They would have to realign to turn back into their former selves in the reverse reaction too. Summary: Opening cyclohexene oxides with nucleophiles gives the trans-di-axial product.

Formation of a Cyclic Intermediate Iodolactonisation A stable cyclic intermediate can be a good way to control stereochemistry. The idea is often to create stereochemistry in forming the ring and cleave the ring later with the stereochemistry already in place. Treatment of acid 103 with iodine yields the iodolactone 104. This must take place by an intramolecular version 105 of the reactions we have just been discussing.23 This reaction was also discussed in chapter 17. O

O

O

CO2H

I2, NaHCO 3

103

O

I

104

KI

I 103

105

104

This reaction is worth looking at in a little more detail. It depends upon formation of an intermediate species, the iodonium ion 105. There is nothing to stop the diastereomeric iodonium ion 106 from forming but cyclisation with the carboxylate ion cannot occur and, since the reaction is reversible, the reaction will run back to the unsaturated acid 103 and try again. The two alternative iodonium ion conformations with equatorial CO2⫺ groups cannot cyclise until the chain flexes to put the CO2⫺ axial. attack here O leads to chair

O

I 105a

attack here leads to boat

O

O

O I

carboxylate cannot attack here 106

O equatorial proton

H

axial proton

H I

H

equatorial proton 104a

The iodonium ion that does react cyclises 105 to give a chair 104a and not a boat. The 1,3- lactone bridge must be diaxial - another reason why the CO2⫺ group must be axial for cyclisation to occur. We can also see that the product 104 has the trans-diaxial relationship between the iodine and the CO2⫺ group we are familiar with from the opening of epoxides. The iodolactone 104 can be elaborated. Once again we should look at the compound from a conformational point of view 104a. There is only one proton in an antiperiplanar relationship to the axial iodine atom and so elimination will occur with this proton 107. When it comes to reaction of the new double bond in 108, stereoselectivity will be dominated by the lactone bridge which

21 Formation of a Cyclic Intermediate

417

will provide a barrier to incoming reagents. So, epoxidation leads to the oxygen being introduced on the underside of the ring to give epoxide 109. On the other hand, if the lactone bridge is first opened a different epoxide 110 is formed. B

O

O

O

O

O H

O ArCO3H

DBU

I

CO2Me

base

104

H

MeOH

O

H 107

O

108

110

109

Just before we continue let’s have a look at the reagents used in the iodolactone formation. The iodine is obviously needed but notice the weak base, NaHCO3. This will deprotonate the acid (making it more nucleophilic) and make the reaction faster. Only one product is possible from 103 but in some cases the presence or absence of base leads to different products. The example 103 is probably the most straightforward – we start with a ring to aid the control of another one. Things become more complicated when we form iodolactones from open chain compounds The hexenoic acid 112 can give two different iodolactones24 111 and 113. I

I I2

Ph

NaHCO3

Ph

I2

Ph

O CHCl3

111

O

O

MeCN

CO2H 112

113

O

In one reaction base is used and in the other it is not. When base is not used the product is the thermodynamic trans lactone 113 whereas with base then the kinetic cis lactone 111 is formed. The decision about stereochemistry is made when the iodine first attacks the double bond. But this is not the end of the story. Once again we might expect addition of I2 to be reversible (which means that the decision is not yet irrevocable) and so the selectivity will be down to which of the iodonium ions 114 or 115 cyclises. The influence of the base might be to change the rate determining step of the reaction. With a more reactive nucleophile to hand (carboxylate instead of carboxylic acid) we could imagine that the iodonium ion 114 would not have any time to equilibrate before nucleophilic attack occurs.

111

I

Ph

I2 NaHCO3

I2

Ph

CHCl3

CO2 114

Ph

I 113

MeCN

CO2H 112

CO2H 115

Predicting stereochemistry It tends to be easier to predict the stereochemistry of thermodynamic products and in general, thermo-dynamic control give good diastereoselection. With six-membered rings the bulkier substituents will end up trans if they are substituted in a 1,2 pattern and cis if they are 1,3-related (this is presuming there is some element of choice of course). With five membered rings then the substituents need to be 1,2-related if there is to be diastereoselection and then they will end up trans.25

418

21 Controlling Relative Stereochemistry

Application of an iodolactonisation reaction (±)-Methyl shikimate 116 has been made using a strategy with iodolactonisation at its heart.26 Note that shikimate has three chiral centres. Interestingly, these three centres are controlled in the synthesis by a centre that is finally destroyed. The first steps are iodolactonisation of 103 and subsequent elimination and epoxidation to give 109 as described above. O

O

O CO2H

O

1. NaHCO 3, I 2, KI

ArCO3H

2. DBU, THF

Ar = 3,5-dinitrophenyl

O 109; 81% yield (+6% endo)

108; 84% yield

103

Epoxide 109 is opened with Me3SiBr (Ph3P-catalysed) and another elimination with DBU gives the allylic alcohol 117. The epoxidation that follows is controlled in a couple of ways. Firstly, the axial alcohol can direct the reagent (the same peroxyacid used to make 109) to the bottom face of the six-membered ring which is also the less hindered side of the molecule (opposite the lactone bridge) 118. The synthesis is completed with basic methanolysis of the lactone 118 to give racemic methyl shikimate 116.

O

O

1. Me 3SiBr Ph3P

O

O

O

HO

ArCO3H

109 2. DBU

HO

MeOH

HO

HO

O 118; 85% yield

117

OMe

K2CO3

OH 116; 98% yield

The last step in this reaction (the methanolysis) warrants a little more discussion as there is quite a lot going on—the epoxide has gone and a new double bond has appeared. The first step is merely the methanolysis to give ester 119 but, under the basic reaction conditions, the ester can be deprotonated to form enolate 120 which spontaneously opens the epoxide to give our target material. Note that the stereochemical information at the α-position – the only chiral centre in the starting material 103 – is lost with the deprotonation and never comes back. O K2CO3

HO

CO2Me

K2CO3

HO

OMe

118 MeOH

HO

O 119

HO

116

O 120

Variations on this theme There are several reactions that have a similar flavour to iodolactonisation.27 That is, where an intramolecular nucleophile attacks an electrophilic site that is derived from a double bond. The most

21 Formation of an Acyclic Compound via a Cyclic Transition State

419

obvious cousin must be the bromolactonisation. Acid 121 was subjected to bromolactonisation in the first stage in the synthesis of a pheromone component.28 But for a more complicated substrate, bromolactonisation was used in Corey’s synthesis29 of erythronolide B. Dienone 122 was reacted with bromine and KBr to give bromolactone 123 in 96% yield. In fact, a few steps later, another bromolactonisation was done which featured the other double bond. O

O

Me

CO2H

Me

Me Br

Me Br2 / KBr

O

CO2H

122

121

O

Me

Me

123; 96% yield

Other varieties of this sort of reaction include bromoetherification (where the nucleophilic component is an alcohol rather than a carboxylic acid). Kishi30 used this reaction to form one of the tetrahydrofuran rings 124 to 125 in his synthesis of Monensin but he used NBS instead of bromine as the source of electrophilic bromine.31 Me

Me

Me

Ar

Me

Ar H

O

Me

OH

Et

H

O

Et H

124

O

H

Br Me

125

Formation of an Acyclic Compound via a Cyclic Transition State Claisen rearrangement The Claisen rearrangement32 converts an allyl vinyl ether 126 into a γ,δ-unsaturated carbonyl compound 128. This is already useful simply because it is a good way to make γ,δ-unsaturated carbonyl compounds but we are interested in the additional stereoselective aspects of the reaction. This is a [3,3]-sigmatropic reaction 127. R'

R' O

[3,3]

R

R'

R

126

127

R

R

γ

trans geometry of 128 controlled by what R' chooses to do

128

R' O

126a

O

δ

heat

O

R' O

129

R

R' O

128a

R

R' O

129a

R

At first sight it may appear that this reaction is not going to be terribly useful stereo-chemically. For instance, in the above scheme we start with one chiral centre and end with none at all. However, it should be noted, before we start to make the chemistry any more complicated, that the trans geometry of the double bond is controlled by what R’ chooses to do in the transition state 129.

420

21 Controlling Relative Stereochemistry

The choice is reminiscent of what we have seen with the aldol reaction – where the electrophilic aldehyde chooses to place its R group equatorial. Here too, R’ chooses the equatorial position. The equatorial positions in cyclohexane are always parallel to one of the existing bonds within the cyclohexane ring 129a. The relationship between that bond and the substituent R’ is already trans before 126a and during the reaction 129. So, a trans double bond results. This reaction is discussed in chapter 15. We can make the situation much more complicated because substituents can be introduced to both of the double bonds 130. Of all the substituents R’ and a-d, R’ is the only one that has any choice—the others just have to do what they are told. Allylic alcohol 131 reacts with enol ether 132 to give an allyl vinyl ether 133 which is ripe for rearrangement. As it happens, 133 is optically pure and contains a trans double bond. R'

OH O

i-Pr

c a

Me

i-Pr

Me

131

R

b

[ 3,3 ]

i-Pr

132

O

Me

O

/ H+

EtO

133

H 134

d 130

The two options for rearrangement are that the isopropyl group can choose whether it adopts an axial 135b or an equatorial 135a position in the transition state.33 The methyl group has no choice at all – it has to be in an equatorial position because it is attached to the trans double bond involved in the reaction. The product from 135a contains a trans double bond and an S configured carbon atom while the product from 135b has a cis double bond and the R configuration. equatorial i-Pr

135a

H O

i-Pr

axial

i-Pr

=

i-Pr =

i-Pr

H

E-(S)-134; Favoured

i-Pr O

O Me

Me

S

=

i-Pr

Me

H

135b

O O

=

O

i-Pr

O Me

NO choice here

H

H

O

R H Z-(R)-134; Not favoured

The product E-134 resulting from the equatorial isopropyl group is the one that is favoured. Note that the chiral centre we started with has been destroyed and a new chiral centre has been introduced. This process is sometimes known as ‘chirality transfer’. Interestingly, if we start with the other absolute stereochemistry and the other bond geometry 136, the outcome is the same.34 So, cis compound 137 also gives aldehyde 134. equatorial O

i-Pr

i-Pr 136

H

O

OH

137

i-Pr H 138

i-Pr =

O

E-134 Favoured

H

To the shrewd operator, this can be quite a handy trick since it provides a way of using both enantiomers, from a resolution say, to give the same enantiomer of product.

21 Formation of an Acyclic Compound via a Cyclic Transition State

OH OH

OH

i-Pr

i-Pr

134

i-Pr

resolution

(R)-140

Me

421

Me

136

OH

(±)-139

OH i-Pr (S)-140

Me

134

i-Pr 131

The Ireland-Claisen rearrangement There are many variations on this Claisen theme. Take, for instance the Ireland-Claisen reaction35 where an enol ether of an ester (a silyl ketene acetal) 143 is used instead of the simple vinyl group in the standard Claisen reaction. The double bond geometry can be controlled by the enolisation process. In this example 143 we have one chiral centre and two double bonds before the reaction. Once again, the R group has choice but the substituents on the double bonds do not.

OSiMe3

O OH

1. LDA

O R 141

1 × chiral centre 2 × double bonds

O

2. Me 3SiCl

R

R 143

142

Following the reaction we have two chiral centres and one stereochemically controlled double bond 145. Note that the chiral centre that was in place at the start has been destroyed but its chiral essence has not — it is reborn in the two new chiral centres in 145.

H Me

OSiMe3 R O

H Me

144

H

[ 3,3 ]

H OSiMe3 Me R = Me H O H Me Me 145

O OSiMe3 2 × chiral centres 1 × double bond R 145

An Ireland-Claisen reaction was employed by Hulme and Paterson in their synthesis of the ebelactones.36 Interestingly, the synthesis featured an unprotected ketone 146. This nice piece of chemoselectivity probably relied upon the conditions of the enolisation process as well as the steric congestion around the ketone. The enolisation reaction used an in situ trapping by having a mixture of Me3SiCl and Et 3N together with the ester before the LDA was added. The cis silyl enol ether (or ‘E silyl ketene acetal’ if you prefer) 147 resulted and following rearrangement and hydrolysis the E-ester 149 was isolated in 83% yield and 96:4 diastereomeric ratio.

422

21 Controlling Relative Stereochemistry

–78 ˚C TMSCl / Et3N

O

O

OTBDMS

O

LDA

O

OTBDMS 147

146

O

warm [3,3]

OSiMe3

Me MeO

O

H O

O

Me

H OSiMe3

OTBDMS

149; 83% yield; 96:4 diastereomers

O

OTBDMS

148

With all this talk of chair transition states it is easy to think that the outcomes of Claisen reactions and their friends are easy to predict. Beware however, because they can occasionally proceed through a boat transition state if the substituents make the chair too unfavourable. A related reaction is the [2,3] Wittig rearrangement.33, 37 This goes via a five-membered transition state – we shall not go into any more detail about that – but it too is a useful reaction both for making homoallylic alcohols and because of the stereocontrol that can be achieved in the process. Allylic ether 150 gives38 only the diastereoisomer shown of alcohol 152. The [2,3]sigmatropic rearrangement 151 creates an E-alkene at the expense of a Z-alkene and two new chiral centres at the expense of one. The immediate product of the [2,3]-shift is an oxyanion instead of a carbanion. BuLi

O

O

THF –78 ˚C

Ph

HO

Ph

Ph

150

151

152

Stereoselective Reduction of β-Hydroxy Ketones We shall look in this section at two methods to reduce β-hydroxy ketones 154. One method gives the anti product 153 whereas the other gives the syn 155. In both cases, the existing chirality at the carbinol carbon is dictating the reagent’s performance. OH

OH R2

R1

OH

OH-directed reduction

O R2

R1

153 (anti)

OH

OH-directed reduction

OH R2

R1 155 (syn)

154

Syn selective reductions The overall process involves a stoichiometric reducing agent which is plain old NaBH4 and another boron species Et2BOMe. We shall see in a minute that it is the job of this second boron species to hold the β-hydroxy ketone 156 in a ring as the reduction proceeds.39 OH Bu

O

Et2BOMe, NaBH 4

Bu 156

MeOH / THF –70 ˚C

OH

OH

Bu Bu 157; 99% yield

diastereoisomer ratio: 99:1 syn:anti

21 Stereoselective Reduction of β-Hydroxy Ketones

423

But first consider the yield of the reaction. It is 99%. What is more, the diastereoselectivity is 99:1. If you work in a lab yourself, you might reflect on how many of your reactions go in 99% yield with such high selectivity. It really is a very good reaction indeed and this is how it works. Under the reaction conditions the cyclic intermediate 158 is formed. Methoxide has been displaced from Et2BOMe and one of the carbonyl lone pairs also gets in on the act and binds the boron in place. Reduction of 158 gives another boron chelate 159. Et OH

O

Bu

O

Et2BOMe

Et B

Et NaBH4

O

O

Et B

O

MeOH

Bu

Bu

Bu

Bu

Bu

158

156

159

We have two sp2 centres in the ring (the carbonyl) and so have a half-chair 158a. We already know from this chapter that half-chairs, by axial attack, develop into chairs. Axial delivery of hydride from NaBH4 gives 159a. Note the two hydrogen atoms in the resulting chair 159a. In your mind, highlight the zig-zag of the carbon chain from one butyl group to the other to reveal that the two hydrogen atoms are on one side of the chain and the two oxygen atoms on the other side. The stereochemistry is in place. All that remains is to remove the boron to liberate the diol syn-157. This hydrolysis is done with acetic acid. syn protons H Et H B Bu O Et O Bu 159a

axial attack Et Bu Bu

NaBH4

O B O Et 158a

OH

AcOH

OH

Bu

Bu

syn diol 157

Anti selective reductions Boron reagents are used once again in the anti selective reduction but this time one reagent does all the work. Again, a look at the overall reaction reveals a reaction of the same calibre.40 OH

O

Bu

Me4N+ (AcO)3BH–

Bu

AcOH / MeOH, –40 ˚C

156

OH

OH

Bu Bu 160; >90% yield

diastereoisomer ratio: 95:5 anti:syn

Once again it’s the hydroxy group that starts things off by binding to the boron 161. But this time the hydride is delivered internally to the carbonyl group through a chair transition state 162. The hydrogen atom forms part of the ring with the carbonyl oxygen axial and the butyl group equatorial. Again, highlight the carbon chain in your mind and notice the substituents on it 163. This time the two hydrogen atoms (and the two oxygen atoms) are on opposite sides of the chain. anti protons (AcO)2B

H OAc

H O

O

O

Bu Bu

Bu

161

O H

B OAc

H Bu

OAc OH O B OAc H

OH Bu

OH Bu

Bu

Bu 162

163

anti-diol 160

424

21 Controlling Relative Stereochemistry

The Evans-Tishchenko reduction Another way to produce 1,3-diols in an anti-selective reaction also starts with a β-hydroxy ketone. The reaction is catalysed by samarium iodide but the stoichiometric reducing agent is acetaldehyde which ends up as the acetate ester in the product 165. OH

O

OAc

MeCHO

OH

15% SmI2

165; 96% yield > 99:1 anti : syn

164

The overall reaction gives the ester 165 and not, as we might expect, a diol or an acetal. The alcohol 164 reacts with acetaldehyde to form a hemiacetal 166. It is the hydrogen on this hemiacetal that is delivered as a hydride to reduce the neighbouring carbonyl but only after a Sm chelate 167 is formed. Compare this transition state 167 with 162 where trisacetoxyborohydride donated the hydride ion. They are very similar: both transfer a hydride via a six-membered cyclic transition state. In 162 the boron holds the two oxygens in a ring while in 167 it is samarium.41 H MeCHO

164

HO Me

Sm O

SmI2

O Hex

O

O

Me

O H

becoming equatorial

165

Hex 167

166

The selectivity (⬎99:1) is astounding but made all the more impressive when we see that substrates with other chiral centres react very selectively too. So, both syn and anti β-hydroxyketones 168 and 170 give excellent anti diastereoselection. But this reaction can be as fickle as frustrated chemists rather expect – compound 172 does not react at all under these conditions while compound 173 reacts as expected. O OH

i-Pr

O

O

OH

O

BnO

15% SmI2

anti 169; 95% yield, > 99:1 anti : syn

168, syn

OH

OH

i-PrCHO

O

OAc

OH

172; no reaction BnO

OH

O

MeCHO 15% SmI2

170, anti

anti 171; 86% yield, > 99:1 anti : syn

173; reacts as expected

Kinetic diastereoselection of 1,3-diols While we are on the subject of anti and syn-1,3-diols, there is a very useful reaction that allows their separation by means of reaction rather than, say, chromatography. The reaction is a kinetic diastereoselection. With a kinetic resolution, one enantiomer of substrate reacts faster than the

21 Stereoselective Reduction of -Hydroxy Ketones

425

other. With a kinetic diastereoselection, one diastereomer of substrate reacts faster than the other. As with a kinetic resolution, the quality of the material that is left behind improves as the reaction is taken to greater levels of completion. Acetals 174 were made from a mixture of the corresponding anti and syn diols. The acetals are hydrolysed with dilute acid and then the reaction is stopped prematurely by the addition of base. The anti diastereomer hydrolyses more quickly to produce the anti diol 175 while the syn acetal 176 is left behind.42

O

OH

2M HCl (5 mol%) 20 ˚C / CH 2Cl2 / 1 hr

O

O

OH

O +

63% conversion

175; syn : anti, 30:1

174; syn : anti, 1.2:1

176; syn : anti, 1:2

Aldol revisited Finally we’ll have a quick look at how combinations of these methods have been applied. In the aldol reactions we have looked at so far there has been no chirality at the start. Both the aldehyde and the enolate have been achiral species that have reacted in a stereoselective way to give a particular diastereomer. With the aldol reaction there is a lot of opportunity to introduce aspects of chirality. The enolate could be chiral as could the aldehyde. In addition to this, the whole reaction could be mediated by a chiral catalyst. Although chiral enolates are most commonly associated with asymmetric methods (most famously the method of Evans in Chapter 27) it is important to remember that the components could just as easily be chiral and racemic. The diastereoselectivity that allows the Evans’s chemistry to work with optically pure materials will operate whether the auxiliary is optically pure or not. In the next example, the chiral ketone happens to be optically pure but it is still an example of relative stereocontrol. We shall see more of relative stereocontrol with optically pure materials in Chapter 30. We saw in the Diels-Alder reaction that different features of reactivity are responsible for different aspects of resulting stereochemistry — geometry of starting materials, stereospecificity of reaction and endo selectivity all have their part to play. Ketone 177 is reacted with a boron chloride to give boron enolate 178. Significantly, this is a trans enolate which means we can expect an anti relationship from the aldol reaction which we do indeed see 179.

anti O

O

B(cHex)2

O

syn

OH

O

(cHex)2Cl

RO

Et3N, –78 ˚C

OBn 177

OH

RO

RO

OBn

OBn

trans 178

RO

anti 179

180

OBn anti

However, another anti relationship also appears in the product. The trans enol was not responsible for this but the existing chiral centre was. Assuming that the trans-boron enolate will give its anti relationship come what may, the two possible diastereo-mers were the anti, anti 179 and the syn, anti 180. One chair transition state leads to one and another chair to the other.

426

21 Controlling Relative Stereochemistry

If you would like to know more about how the existing chiral centre makes its influence felt then you are directed to the paper43 and a suggestion for transition states.44 For now we continue with the manipulations. The new hydroxyl group that has been formed in the aldol reaction can be used to direct a stereoselective reduction of the ketone in the way that we have already seen. So, β-hydroxyketone 179 is reduced to diol 181. O

OH

OH

OH

Me4N•(AcO)3BH

RO

MeCN, –30 ˚C to –20 ˚C

RO

OBn

OBn

179

181

We now have four chiral centres. Three of them are new and put in place relative to the starting chiral centre. Since that was optically pure we have an optically pure product. Diol 181 is an intermediate on the way to a fragment of Spongistatin 1 – a chain extension, asymmetric dihydroxylation and stereoselective intramolecular Michael addition all feature on the way there.43

Open Chain Chemistry - with Chelation Control Grignards When students first encounter stereochemistry, there is a tendency for them to justify stereochemistry based upon the way the compounds happen to have been drawn. So, when asked to explain why ketone 93 reacts with BuMgBr to give mostly diastereomer 94 the answer is easy. Why yes, obviously the butyl group comes in on the less hindered side of the carbonyl 182a which is the one opposite the R group. O Ph

HO BuMgBr

O

Ph

O

attack from behind

Bu

BuMgBr

Ph

Me

Me

O

R 183

R 182

O Me

R 182a

But of course, the exact same question could be asked with the structures merely drawn in an alternative conformation 182b and 183a. Sticking with the same enantiomer we now find that attack still occurs from behind even though that is now from the same side as the R group. O R Ph

HO Me

O 182b

R

BuMgBr

Ph

Me O 183a

attack from behind?

O

Bu R

BuMgBr

Ph

Me O 182c

Something, somewhere must be controlling the conformation to give control of stereochemistry and just drawing the molecules in a convenient way is going to give the right answer only by chance. Chelation is the answer. In the example we’ve just seen, it is the magnesium of the Grignard that is doing the work. Both the oxygen atoms bind to it to lock the conformation of the

21 Open Chain Chemistry - with Chelation Control

427

molecule 184. The butyl group does indeed attack the ketone from the less hindered side (opposite the R group) and now we have a reason why. Mg Ph

182

attack from behind

O

MgX2

O

Mg Ph

Me

O

Bu

O

R 184

183

Me

R 185

Zinc borohydride Chelation control of this kind is commonly exploited in reduction and a favourite reagent is zinc borohydride45 [Zn(BH4)2]. As with the Grignard, we need a second oxygen atom to do the binding. It works well with α-hydroxyketones 186. Generally speaking, the anti diastereomer will be favoured. We can draw a conformation where the oxygen atoms are on the same side of the molecule both bound to the zinc 187. The hydride will then be delivered from the less hindered side of the molecule. If R1=Ph and R2 =Me, the ratio anti:syn-188 is 98:2. Zn

OH Zn(BH4)2

R1

R1

O

R2 H3B

O 186

H

hydride comes in opposite R2

OH

O

R2

R2

R1 187

OH +

R1

OH

R2 OH

syn-188 : anti-188 anti:syn = 98:2, R 1=Ph and R2=Me

A similar sort of thing can be achieved with β-ketoesters where the major product is the syn diastereomer as we shall see in the example given below. Interestingly, Zn(BH4)2 does not generally work very well with β-hydroxyketones. But there are other very powerful methods available for that sort of transformation which we have see already. Both the reduction of a β-keto ester and an α-hydroxy ketone are put to use in the synthesis of racemic blastmycinone46 190 – a key fragment of blastmycin 189.

H N

HN CHO OH

O

O

O O

O

O 2 x C–O ester

O

189; (±)-blastmycin

O

O

O

O

190; (±)-blastmycinone

The synthesis of blastmycinone begins with a Reformatsky reaction. Although the resulting allylic alcohol 192 is a required intermediate, it is formed in this reaction without any stereocontrol. The alcohol 192 is oxidised in a Swern reaction to the enone 193 only to be resurrected in the next reaction but this time with stereocontrol.

428

21 Controlling Relative Stereochemistry

OH

O

Bu

O

Ph

O

(COCl)2 Me2SO

1. Zn 2.

Swern

CHO

Br

O

191

OBn

O

O

192

Ph

193

The ketone 193 is reduced with Zn(BH4)2 via the chelate 194. In the enantiomer drawn, the hydride attacks the molecule on the opposite side from the butyl group. After hydrolysis of the zinc chelate of the product 195 the molecule is drawn as we have drawn it before 196 but in the reacting conformation 194 both the oxygen atoms are on the same side of the molecule because they are chelating the zinc atom. O

Zn

O

H

Zn(BH4)2

193

O

Zn

OH

O

OBn

OBn O

attack from the front 194

O

195

Ph

196

The first new stereocentre is thus taken care of. Next another carbonyl is revealed 197 by ozonolysis of the double bond. We are now in a position to introduce with control the second stereo-centre using the one we have just made! OH

OH

OH Zn(BH4)2

1. O 3 2. Me 2S

O

O

Ph

O

O

196

O

HO

Ph

O

O

197

Ph

198

This time chelation is with the ketone and the hydroxyl group 199. The chelate 199 is drawn without the ester group for clarity and note (once again) that the oxygen atoms are on the same side because they chelate the zinc atom and that attack occurs opposite the side chain. Zn

OH Zn(BH4)2

O

CO2Bn

=R

O

O

OH HO

R

R = 198

attack from the front

197a

199

198a

The benzyl group is removed by hydrogenation which leads, by spontaneous ring closure, to the lactone 200. The synthesis is completed by esterifying the remaining hydroxyl group. O O

OH

OH H2 / Pd / C

O

Cl pyridine

HO

O

O 198

Ph

O

O O 200

O 190

21 Open Chain Chemistry - in its Most Genuine Form

429

Open Chain Chemistry - in its Most Genuine Form The most genuine form of open chain chemistry has no chelation to hold things in rings. But there must be something controlling the conformation or there would be no control. We look at two examples – the addition of nucleophiles to acyclic ketones and the addition of electrophiles to alkenes.

Felkin control over reactions of nucleophiles with carbonyl compounds We need to consider firstly the line of attack nucleophiles use in approaching carbonyl groups. The nucleophiles come in to the C=O bond at an angle close to the tetrahedral angle of 109⬚ and not just at 90⬚. We shall see that this is significant. This is often referred to as the Bürgi-Dunitz trajectory from the work by Bürgi and Dunitz who found an angle of 107⬚ from crystal structures where a nitrogen atom approached a carbonyl group.47 Theoretical values are not wildly different.48 Imagine nucleophilic attack 201 on a carbonyl compound with a stereogenic centre next door. We will call the groups L (for large) and M (for medium). The hydrogen can be S (for small). What influence will the old chiral centre have on the newly forming chiral centre? Nu

Nu O

107˚ O

H

O Nu

L

R

H

M 201

?

Nu

L

R M 202

We view the molecule from the carbonyl carbon down towards the chiral centre which gives us a useful (Newman) projection of the molecule. The most comfortable conformation for the molecule is with the larger group being at 90⬚ to the carbonyl functional group. The two rotamers with this conformation are 203 and 204. Step 1: Place the large group L orthogonal to the carbonyl group

O S

M O L

L

R M

S R 203

204

The nucleophile will come in opposite the large group and it would rather come in alongside the small group 203a (which is often a hydrogen atom) than alongside the medium group 204a. We see now why 107⬚ is significant. The model indicates that, as far as the nucleophile is concerned, there would be no difference between the rotamers if the nucleophile came in at 90⬚ as it would come in between (rather than alongside) the M and S group. Step 2: Choose the more reactive of the two conformations

M O

√

O S L

Nu

S R 203a

L R M 204a

× Nu

Consider then, the reduction of ketone 205 with LiAlH4. Unlike some of the reductions we have seen so far in this chapter, there are no coordinating sites in the rest of the molecule for chelation control of any kind. The result of the reduction is alcohol 206 complete with stereocontrol.

430

21 Controlling Relative Stereochemistry

O

OH LiAlH4

Me

Ph

Me

Me

Ph 205

Me 206

And now the explanation. The phenyl group is the largest group and sits at 90⬚ to the carbonyl. Attack is alongside H 207. When working through the stereochemistry it is helpful to draw the result with a Newman projection before drawing the molecule in the standard way. In redrawing we have then performed a GSR (Chapter 20) to reveal that we have anti 206. Me O

Me Ph

O

OH Ph

H

GSR

Me

Me

usual way

H3Al

H

Me H 208; product in Newman Projection

H Me

207

OH

Redraw in

LiAlH4

Ph

Me

Ph

Me

206a

anti-206

The selectivity improves if the group adjacent to the carbonyl is increased in size.49 Diastereoselectivity can be further improved by lowering the temperature. The 98% (i.e. 49:1) selectivity that is observed when R=tert-Bu is increased to 99.8% when the temperature is lowered from 35 ⬚C to ⫺78 ⬚C. O Me

OH

LiAlH4

Ph

R Ph 209

R

R Me anti-210

anti : syn ratio

Me

Et

i-Pr

t-Bu

2.8:1

3.2:1

5:1

49:1

Role of electronegative substituents Another thing to consider is the role of electronegative substituents. These can interact with the carbonyl in an electronic way (as opposed to steric). They too have a tendency to sit at 90⬚ to the carbonyl. So the α-chloroketone 211 reacts with nucleophiles 212 to give the adduct 213 and hence the alcohol 214 with the nucleophile anti to the OH group. 90˚ to carbonyl

O

Me O

Me Cl

Cl 211

Nu

H 212

HO

Nu

O

Nu

Cl

Nu H

Me 213

Cl 214

The explanation for this depends upon the interaction of orbitals and there are a couple of ways of looking at this. One way is to think of the π*CO orbital interacting with the σ*CX orbital to give an even lower antibonding orbital50 216. This overlap can occur only if the orbitals are correctly aligned 215 and this will occur when the substituent is at 90⬚ 211a. There are no electrons in this orbital of course, so the benefit from the overlap will only occur once the nucleophile starts to feed in electrons.

21 Open Chain Chemistry - in its Most Genuine Form

431

π*CO O

π*CO

=

O σ*CX

Cl

σ*CX X

new LUMO 215

211a

216

The other argument (or the other part of the same argument) applies when the nucleophile has already started to attack 212a. The σ-bond between the nucleophile and the carbonyl carbon is partly formed (σC–Nu). Once again, this can overlap with the σ*CX orbital 217 to lower the energy a bit further.51 The angles of the orbitals are slightly different because the attack has started. σ C–Nu {from π*CO and nNu}

Nu O

O

σ*CX Cl

X 212a

217

These explanations talk about the position of the electronegative atom just before the reaction happens or as the reaction is happening (and are thus concerned with the ground state or the transition state of the reaction). The electronegative atom may or may not be at 90⬚ in the ground state but who cares – it is where it is when the reaction is happening that counts!

The Houk conformation for reactions of alkenes with electrophiles The Houk conformation depends upon allylic 1,3 strain so we will look at this first.52 The strain we are concerned with is between substituents on the double bond with substituents at the allylic position. They will be in the same plane for the strain to be present 218. When the substituent on the double bond is merely a proton, the allylic 1,3 strain is not too bad between the proton and a larger group. About 25% of the conformations will have the proton and the substituent in the same plane 218b but the rest will have two hydrogens in the plane 218a. R

R R

X

Y allylic 1,3-strain 218

R H

H

218a; 75%

R

allylic 1,3-strain not too bad with H on double bond

H H

R

218b: 25%

This changes significantly when the substituent on the double bond is something larger. Even with a methyl group then the only significant conformation 219b will have the proton eclipsing the methyl group. And, just in case you are wondering, the methyl group will not prefer to stagger itself between the two larger groups 219a. This is because it would result in allylic-1,2 strain on the other side of the molecule. Allylic-1,2 strain is very bad news, worse than allylic 1,3-strain, even between two protons. Conformations 219b and 219c have no 1,2-strain but 219c has worse allylic 1,3-strain than 219b. Since we have only one significant conformation, this is the only one we need to consider for reactions. It is known as the Houk conformation.

432

21 Controlling Relative Stereochemistry

allylic 1,2 strain

H

least strain R R

H Me

Me

RR 219a

worse allylic 1,3-strain R H

H

Me

219b

R 219c

If R is larger than X in 220 we can expect electrophilic attack on the double bond to occur opposite R 221. If the electrophile were a peroxyacid, the product would be the epoxide 222. R O

X R Me

Me

H E

220

X

H

R Me

X 221 (R > X)

H 222

Epoxidation of the cis allyl sulfone Z-224 gives the stereochemical outcome we would expect from a Houk conformation.53 The selectivity with the trans compound E-224 is less good – also as we would expect. Things are rarely quite so simple as this of course and dihydroxylation of the same substrates gives rise not only to reduced selectivity but a reversal in stereoselectivity 226. OH 80 : 20

OsO4 pyridine

HO

O m-CPBA

95 : 5 223

SO2Ph

SO2Ph

SO2Ph

Z-224

225

OH OsO4 pyridine

m-CPBA

O

78 : 22

75 : 25 OH

SO2Ph

SO2Ph 226

SO2Ph 227

E-224

Allyl silanes The steric interactions that go into controlling the Houk conformation can be reinforced by electronics in the case of allyl silanes (or other σ-donors instead of silicon).54 The silicon substituent is large and also electron donating. We have already learned about β-cations being stabilised by silyl groups. From a Houk point of view we can expect electrophiles to react as shown in 228. The small group (S) is in the plane of the double bond and the large group (L) is usually the silyl group. Hence reaction of 229 leads to 231 by attack of the electrophile (t-Bu⫹) on the opposite side to the Me3Si group in the Houk conformation 230 (an example of chirality transfer) and the usual loss of silicon from the β-cation. Me3Si Ph

R3Si

Me3Si t-BuCl

S M E 228

H H 229

Me

TiCl4

Ph

t-Bu 230

Me

t-Bu H

Ph H

Me 231

21 References

433

Reaction with electrophiles can be by SE2’ reaction 232 as in the last example but the other end of the double bond can react if it is part of, for example, an enolate 233. Reaction of 234 is an example of the latter and highly stereoselective. Nu

PhMe2Si

O

Li MeI

PhMe2Si

R3Si Ph R

Ph

OMe

E

O

PhMe2Si +

OMe

O

Ph

OMe

E

232; S E2' reaction

233

97:3

234

235

If the group on the double bond is merely a proton 236 then it is possible for the medium group to move into the plane 237 so long as the medium group is itself quite small (typically a methyl group). Although this 237 would not be the major conformation, it allows the electrophile to react alongside the small group. E

L 236 M

S

S

H

M

L

E

H

237

For large electrophiles, which clearly have difficulty coming in alongside the medium group, this selectivity is observed and illustrated by the reaction of 238 with OsO4 (large) to give mostly 240. However, once the medium group is made a bit larger (Ph instead of Me) there is a return53 to the selectivity expected by 236. The reactivity depicted in 237 is not expected if the double bond has anything larger than a proton cis to the chiral centre – the allylic strain would be too great. PhMe2Si

OsO4

PhMe2Si

+

Me

Me

OH 239

238 PhMe2Si

OsO4

PhMe2Si

241

OH 242

OH

Me OH 240

34:66

PhMe2Si

OH +

Ph

Ph

PhMe2Si

OH

92:8

OH

Ph OH 243

References General references are given on page 893 1. See Clayden, Organic Chemistry chapter 35 or Fleming, Orbitals for an explanation. 2. L. E. Overman and P. J. Jessup, Tetrahedron Lett, 1977, 1253. 3. L. E. Overman, G. F. Taylor, C. B. Petty and P. J. Jessup, J. Org. Chem, 1978, 43, 2164. 4. J. E. Dubois and M. Dubois, Chem. Commun., 1968, 1567. 5. C. H. Heathcock, C. T. Buse, W. A. Kleschick, M. C. Pirrung, J. E. Sohn and J. Lampe, J. Org. Chem, 1980, 45, 1066. 6. H. E. Zimmerman and M. D. Traxler, J. Am. Chem. Soc., 1957, 79, 1920. 7. C. J. Cowden and I. Paterson, Org. React., 1997, 51, 1. 8. H. C. Brown, R. K. Dhar, R. K. Bakshi, P. K. Pandiarajan and B. Singaram, J. Am. Chem. Soc., 1989, 111, 3441. 9. J. M. Goodman and I. Paterson, Tetrahedron Lett., 1992, 33, 7223.

434

21 Controlling Relative Stereochemistry

10. E. J. Corey and S. S. Kim, J. Am. Chem. Soc., 1990, 112, 4976. 11. T. Inoue and T. Mukaiyama, Bull. Chem. Soc. Jpn., 1980, 53, 174; This reference details equilibration leading to regioisomers rather than stereoisomers. 12. H. C. Brown, K. Ganesan and R. K. Dhar, J. Org. Chem., 1993, 58, 147. 13. D. A. Evans, J. V. Nelson, E. Vogel and T. R. Taber, J. Am. Chem. Soc., 1981, 103, 3099. 14. C. H. Heathcock, in Asymmetric Synthesis, vol. 3, pp. 111–212. 15. I. Paterson, J. G. Cumming, J. D. Smith and R. A. Ward, Tetrahedron Lett., 1994, 35, 441. 16. C. H. Heathcock, M. C. Pirrung, S. H. Montgomery and J. Lampe, Tetrahedron, 1981, 37, 4087. 17. R. Baker and J. A. Devlin, J. Chem. Soc., Chem. Commun., 1983, 147. 18. N. Greeves in Comp. Org. Synth., vol. 8, page 14. 19. H. C. Brown and S. Krishnamurthy, J. Am. Chem. Soc., 1972, 94, 7159. 20. Eliel, pages 736 and 881. 21. D. A. Evans in Morrison, vol3, p 29–37; L. Mander in Eliel, pages 901–2; R. E. Ireland and L. N. Mander, J. Org. Chem., 1967, 32, 689 (the year and volume number of this reference are wrong in Eliel); H. O. House and T. M. Bare, J. Org. Chem., 1968, 33, 943. 22. Clayden, Organic Chemistry, pages 858–859. 23. M. D. Dowle and D. I. Davies, Chem. Soc. Rev., 1979, 8, 171; B. N. French, S. Bissmire and T. Wirth, Chem. Soc. Rev., 2004, 33, 354. 24. F. B. Gonzalez and P. A. Bartlett, Org. Syn. Coll., 1990, VII, 164. 25. Eliel p. 909. 26. P. A. Bartlett and L. A. McQuaid, J. Am. Chem. Soc., 1984, 106, 7854. 27. P. A. Bartlett, in Asymmetric Synthesis, vol. 3, pp. 411. 28. S. Kuwahara, D. Itoh, W. S. Leal and O. Kodama, Tetrahedron Lett., 1998, 39, 1183. 29. For a full discussion of Corey’s synthesis of Erythronolide B, see Nicolaou and Sorenson p. 167. 30. T. Fukuyama, C.-L. J. Wang and Y. Kishi, J. Am. Chem. Soc., 1979, 101, 260. 31. For a full discussion of Kishi’s synthesis of Monensin see Nicolaou and Sorenson p. 185. 32. P. Wipf in Comp. Org. Synth., vol 5, chapter 7.2, pages 827–873; A. M. M. Castro, Chem. Rev., 2004, 104, 2939. 33. R. K. Hill in Morrison, vol 3, pages 503–572. 34. K.-K. Chan, N. Cohen, J. P. De Noble, A. C. Specian and G. Saucy, J. Org. Chem., 1976, 41, 3497. 35. R. E. Ireland and D. W. Norbeck, J. Am. Chem. Soc., 1985, 107, 3279; R. E. Ireland, D. W. Norbeck, G. S. Mandel and N. S. Mandel, J. Am. Chem. Soc., 1985, 107, 3285. 36. I. Paterson and A. N. Hulme, J. Org. Chem., 1995, 60, 3288. 37. T. Nakai and K. Mikami, Org. React., 1994, 46, 105. 38. N. Sayo, E. Kitahara and T. Nakai, Chem. Lett., 1984, 259; D. J.-S. Tsai and M. M. Midland, J. Org. Chem., 1984, 49, 1842. 39. K.-M. Chen, G. E. Hardtmann, K. Prasad, O. Repic and M. J. Shapiro, Tetrahedron Lett., 1987, 28, 155. 40. D. A. Evans, K. T. Chapman and E. M. Carreira, J. Am. Chem. Soc., 1988, 110, 3560. 41. D. A. Evans and A. H. Hoveyda, J. Am. Chem. Soc., 1990, 112, 6447. 42. S. E. Bode, M. Müller and M. Wolberg, Org. Lett., 2002, 4, 619. 43. I. Paterson and L. E. Keown, Tetrahedron Lett., 1997, 38, 5727. 44. I. Paterson and R. D. Tillyer, J. Org. Chem., 1993, 58, 4182. 45. L. A. Paquette, Ed., Encyclopedia of Reagents for Organic Synthesis, vol. 8 (John Wiley & Sons, Chichester, 1995). 46. T. Nakata, M. Fukui and T. Oishi, Tetrahedron Lett., 1983, 24, 2657. 47. H. B. Bürgi, J. D. Dunitz and E. Shefter, J. Am. Chem. Soc., 1973, 95, 5065. 48. H. B. Bürgi, J. M. Lehn and G. Wipff, J. Am. Chem. Soc., 1974, 96, 1956. 49. M. Chérest, H. Felkin and N. Prudent, Tetrahedron Lett., 1968, 2199. 50. Eliel p. 877. 51. A. J. Kirby, Stereoelectronic Effects (Oxford University Press, Oxford, 1996) p. 51. 52. I. Fleming and J. J. Lewis, J. Chem. Soc., Perkin Trans. 1, 1992, 3257; L. Mander in Eliel, p 894–897; K. N. Houk, N. G. Rondan, Y.-D. Wu, J. T. Metz and M. N. Paddon-Row, Tetrahedron, 1984, 40, 2257. 53. E. Vedejs and C. K. McClure, J. Am. Chem. Soc., 1986, 108, 1094. 54. I. Fleming, J. Chem. Soc., Perkin Trans. 1, 1992, 3363.

22 Resolution Resolution Introduction and an example (1-phenylethylamine) Choice and Preparation of a Resolving Agent Resolving a hydroxy-acid on a large scale Resolving a hydroxy-amine on a large scale Resolving an amino-acid on a large scale Resolution via covalent compounds Advantages and Disadvantages of the Resolution Strategy When to Resolve General rule: resolve as early as possible Resolution of Diastereoisomers Resolution of compounds made as diastereoisomeric mixtures The synthesis of Jacobsen’s Mn(III) epoxidation catalyst by resolution Resolution with half an equivalent of resolving agent Physical Separation of Enantiomers Chromatography on chiral columns Resolution of triazole fungicides by HPLC A commercial drug separation by chiral HPLC Differential Crystallisation or Entrainment of Racemates Conglomerates and racemic compounds Typical procedure for differential crystallisation (entrainment) Conventional resolution of L-methyl DOPA Resolution of L-methyl DOPA by differential crystallisation Finding a differential crystallisation approach to fenfluramine Resolution with Racemisation Resolution of amino acids by differential crystallisation with racemisation Differential crystallisation and racemisation when enolisation is impossible Kinetic resolution with racemisation Other methods of racemisation during resolution: the Mannich reaction Resolution with racemisation in the manufacture of a drug Resolution with Enzymes Enzymes as resolving agents Resolution by ester hydrolysis with enzymes Resolutions of secondary alcohols by lipases Kinetic resolution with proteolytic enzymes Kinetic resolution with racemisation using proteolytic enzymes

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

436

22 Resolution

Kinetic resolution on diastereoisomeric mixtures Comparison between enzymatic and classical resolution Asymmetric Synthesis of a Prostaglandin with many Chiral Centres

Resolution Introduction and an example (1-phenylethylamine) Resolution is the separation of a racemic compound into its right and left handed forms. In the world at large it is an operation we carry out whenever we sort chiral objects such as gloves or shoes. Simply inserting your right foot into any shoe tells you at once whether it is a left or right shoe: the combination of right foot and right shoe has different physical properties (it fits) to the combination of right foot and left shoe (it hurts). Resolution also needs a “right foot”, a single enantiomer of a resolving agent which we combine with the racemic compound to form a 1:1 mixture of diastereoisomers. These will probably have different physical properties so that any normal method of separation (usually crystallisation or chromatography) separates them; removal of the resolving agent then leaves the optically active target molecule. We shall begin with a classical resolution - almost the classical resolution.1 O A chemical reaction

NH2

NH4OAc NaB(CN)H3

Ph 1; prochiral

Ph 2; 25 g

HO HO2C

3; chiral enantiomerically pure one diastereoisomer one enantiomer [α]D +12.4 one peak by chiral HPLC

OH CO2H

3; (+)-(R,R)-tartaric acid cool slowly in MeOH

HO O2C

OH CO2H NH3

Ph 4a; in mother liquor NaOH, H 2O extract with ether distil

NH2 Ph (+)-(R)-amine 2 [α]D +24.5 recrystallise sulfate

4.4 g (+)-(R)-amine 2 [α]D +38.3

HO O2C

2; chiral but racemic one compound one NMR spectrum one peak by HPLC

OH CO2H NH3

4a and 4b; two salts diasterosiomers different properties different NMR spectra two peaks by HPLC

Ph 4b; crystallises out NaOH, H 2O evaporate to dryness

NH2 Ph 6.9 g (–)-(S)-amine 2 [α]D –38.2

(R)-2 and (S)-2; two separate enantiomers same NMR spectrum different by chiral HPLC equal and opposite rotations

22 Choice and Preparation of a Resolving Agen

437

The amine 2 is made by a chemical reaction - the reductive amination of ketone 1. The starting material 1 and the reagents are all achiral so the product 2, though chiral, must be racemic. Reaction with one enantiomer of tartaric acid 3 forms the amine salt 4, or rather the amine salts 4a and 4b. Examine these structures carefully. The stereochemistry of tartaric acid 3 is the same for both salts but the stereochemistry of the amine 2 is different so these salts 4a and 4b are diastereoisomers. They have different physical properties: the useful distinction, discovered by trial and error, is that 4b crystallises preferentially from a solution in methanol leaving 4a behind in solution. Neutralisation of 4b with NaOH gives the free amine (S)-2, insoluble in water and essentially optically pure. Crystallisation doesn’t remove all of 4b from solution so the mother liquor contains mostly 4a with some 4b. This is clear when the solution is neutralised and the free amine 2 isolated from it by distillation. The rotation has the opposite sign to that of (S)-2, but is smaller. Recrystallisation of the sulfate salt brings the rotation to the same value as that of (S)-2, but with the opposite sign and we have a sample of pure (R)-2. You may feel that we have laboured this very simple resolution but it is important that you understand this process before continuing not only this chapter but all this section - chapters 22–31. The amine 2 is itself an important compound as you will see in the next section.

Choice and Preparation of a Resolving Agent Resolving a hydroxy-acid on a large scale Chemists at Parke-Davis have been making hydroxy acids of the general structure 5 in their development of an HIV protease inhibitor and they sought a method of resolution that would give them both enantiomers.2 The obvious resolving agent would be a single enantiomer of some kind of amine so that a salt would be formed between the resolving agent and the carboxylic acid 5. This would be the reverse of the resolution we have just seen. They tried many amines including 2, but the best by far was 6. The salts between 5 and 6 were easily crystallised, the separation of the diastereoisomers was straightforward, and the yield and % ee of the recovered 5 was excellent.

R1

CO2H 2

R

resolution?

R1

CO2H 2

OH

R

OH

and

R1

CO2H R2 OH

5a

5

Ph

5b

N H (R)-6

Ph

Now a difficulty emerged. They wanted to carry out the resolution on a large scale but enantiomerically pure 6 is expensive. The solution was to make it themselves from previously resolved cheap 2. The obvious route is reductive amination using benzaldehyde and the only danger is racemisation of the intermediate imine 7. They found that the imine 7 did not racemise as it was prepared in toluene but that some racemisation took place when NaB(CN)H3 was used for the reduction. The solution was to use catalytic hydrogenation and they prepared 53 kg batches of optically pure 7 in 98% yield by this method and used that to resolve the hydroxy acid 5.

H2/Pd/C

PhCHO

Ph

NH2

(R)-(+)-2

toluene reflux

Ph

N imine 7

Ph

toluene water

Ph

N H (R)-6

Ph

438

22 Resolution

The preparation of 6 is not a resolution but the starting material 2 was prepared by a resolution and enantiomerically pure 6 was used in a resolution. This sequence of identifying the best resolving agent and then preparing it from a resolved starting material is standard practice. You will meet the lithium derivative of compound 6 in chapter 26 as a chiral reagent. In the past many racemic acids were resolved using toxic alkaloids such as strychnine. Nowadays simple amines such as 2 or 6 are preferred. Top Tip: If you need to resolve an acid, try first amine 2 or some derivative of it such as 6.

Resolving a hydroxy-amine on a large scale The Bristol-Meyers Squibb company wanted the simple heterocycle 8 for the preparation of a tryptase inhibitor. As 8 is an amine, tartaric acid was the first choice for a resolving agent. It again turned out that a modified version of the first choice was the best. Tartaric acid is so good at resolutions that simple variations, such as the dibenzoate ester 9, often work well. CO2H

HO2C

OH +

O

N H (±)-8

O

O O 9; (–)-di-benzoyl tartaric acid

1. crystallise from EtOH at 74 ˚C 2. NaOH, pH 10.5 3. Boc 2O, MeOBu- t

OH N O

Ot-Bu (S)-(+)-10

In this instance, the exact proportions of the resolving agent and 8 and the purity of the crystallisation solvent were important in getting good results.3 After one crystallisation, the ee of the salt was about 50% but this improved by 10-15% with each recrystallisation and reached ⬎99% after five recrystallisations. By then the yield had dropped to 30% from a theoretical maximum of 50%. For the next stage in their synthesis, they really needed the Boc derivative (S)-(⫹)-10 so the salt was directly converted to 10 in ⬎99% ee on a 50 g scale. You will see later in this chapter that an enzyme can be used to do the same resolution. These two examples, 5 and 8, show that with two functional groups in a molecule it is better to choose the one that can form a salt (here CO2H and R2NH) rather than the OH group as it would be necessary to make a covalent compound to use that group.

Resolving an amino-acid on a large scale Other companies (Cilag AG and R. W. Johnson) required the pyridine-containing β-amino acid 11 or, to be more accurate, the ester dihydrochloride4 12. This combination of acidic and basic functional groups offers a wide choice of resolving agents.

NH2

NH3 CO2H

N 11

CO2Me N H

.2Cl 12

The synthesis of the racemic compound is interesting and relevant. The simple aldehyde 13 could be combined with ammonia and malonic acid all in the same operation to give racemic 11.

22 Choice and Preparation of a Resolving Agen

439

One of the functional groups now should be protected so that the other can be used for the resolution and the amine was blocked with a Boc group to give 14. NH2

NHBoc CO2H

CHO CH2(CO2H)2 NH4OAc, EtOH

N

NaOH H2O/THF

N 13

CO2H

Boc2O

N

(±)-11

14

The best resolving agent was also a bifunctional compound, the natural amino alcohol ephedrine 15. Mixing 14 with ephedrine in warm ethyl acetate gave an immediate precipitation of the salt 16. The crude salt already had an ee of around 90% but one recrystallisation again from ethyl acetate gave pure salt 16 in 42% yield and ⬎98% ee. Conversion to 12 required merely neutralisation (NaOH) and reaction with HCl in MeOH to remove the Boc group and make the methyl ester. The product 12 was isolated on a large scale in 82% yield with ⬎98% ee. This is a spectacularly successful resolution. NHBoc CO2

14

OH

MeHN

OH

MeH2N

EtOAc

Ph

Ph N 16; salt of 14 and (1 R,2S)-(–)-ephedrine

15; (1 R,2S)-(–)-ephedrine

Resolution via covalent compounds The calcium channel blocking dihydropyridine drugs 17, used in the important field of heart disease and easily prepared by the Hantszch pyridine synthesis, are chiral but ‘only just.’ The molecule does not quite have a plane of symmetry, because there is a methyl ester on one side and an ethyl on the other and because R may not be Me. An important example is amlodipine 18, a best seller from Pfizer, and this is more asymmetrical than some. Nevertheless resolving these compounds is difficult.

X

X

CO2Et

MeO2C N H 17

R

Hantszch pyridine synthesis

Cl MeO2C

CHO

CO2Et

NH2 O

R

MeO2C

CO2Et N H

O

NH2

18; Amlodipine

The method published by Pfizer5 relies on the formation of an ester 21 of an intermediate carboxylic acid 19 with the alcohol 20 derived from available mandelic acid and the separation of the diastereoisomers by chromatography rather than crystallisation. We can assume that the classical crystallisation of diastereoisomeric salts was not successful. Removal of the ester was simplified as a transesterification. CDI is carbonyl-di-imidazole.

440

22 Resolution

Ph

Cl

OMe

O

MeO2C

Cl MeO2C

HO

OH

20

O

N H

O

Ph

OMe 1. separate (chromatography)

O O

CDI

N H

N3

19

N3

18

2. EtOH 3. H 2, Pd/CaCO 3

21

The second example of resolution via a covalent compound also involves a decision about when to resolve. Ketone 22 is the pheromone of the southern corn rootworm. It has the one functional group and one stereogenic centre in a 1,9 relationship. Disconnection was guided by the long distance between the ketone and the stereogenic centre and by the availability of undecenoic acid6 25. The ketone is changed to an alkene and the 10-methyl group to CO2H to allow disconnection to a readily available starting material 25. O

H

H FGA

FGI

1

2

10 22

23 H

CO2H

CO2H

enolate alkylation

+ X

25

24

We need to add a propyl group to the di-lithium derivative 26 (chapter 2), reduce the CO2H group to CH3, and convert the alkene into a ketone by the mercuration-reduction sequence described in chapter 17. LiO

OLi

2 x LDA

25

Pr–I

R

R

OH

CO2H

H

LiAlH4

R

H

(±)-27

26 1. Ph 3PBr2

(±)-28

1. Hg(OAc) 2

23

22 2. NaBH 4 3. Cr(VI)

2. LiEt 3BH

The CO2H group also helps resolution. Amide formation with the amine (S)-(⫺)-2 gave the amide 30 - a likely crystalline derivative. It is of course impossible to predict with certainty which compounds will crystallise, and particularly which diastereoisomer will crystallise. It turns out that (R,S)-30 crystallises out, leaving (S,S)-30 in solution. Recrystallisation purifies this diastereoisomer until it is free from the other. H

H

R

H

CO2H

SOCl2

R

H

COCl

H2N

(S)-(–)-2)

(±)-27

(±)-29

Ph

Ph

R

O

NH S

H

R

(R,S)-30 this diastereoisomer crystallises out

22 When to Resolve

441

The resolving agent must now be removed by hydrolysis of the amide. This is a risky business as enolisation would destroy the newly formed stereogenic centre, and a cunning method was devised to rearrange the amide 30 into a more easily hydrolysed ester by acyl transfer from N to O. The rest of the synthesis is as before. By this means the alcohol 28 was obtained almost optically pure, ⬍0.4% of the other enantiomer being present. No further reactions occur at the newly formed stereogenic centre, so the absolute chirality of 22 is as shown. H

Ph O

1. LDA

O

(R,S)-30 2. O

R

N

H

H

OH

H2O

R

OH

H

+ HO

(R)-(–)-27

31

H N H

Ph

32

Advantages and Disadvantages of the Resolution Strategy These examples expose the main weakness of the resolution strategy: the maximum yield is 50% as half the chiral molecule 2, 6, 8, 11, 19, or 24 must be the wrong enantiomer. In addition, extra steps are needed to add and remove the resolving agent and, in the removal of the resolving agent, racemisation is a danger. There are advantages too: in principle you get both enantiomers of the target molecule so if you are making a chiral auxiliary, or don’t know the structure of a natural product, or want to investigate the relationship between biological activity and stereochemistry, all situations where having both enantiomers is a distinct advantage, resolution may be the best strategy. You can minimise the disadvantages by resolving as early as possible: that way there is least waste of time and materials. In favourable cases you can neutralise either or both disadvantages, as we shall see soon. The maximum yield may be made 100% if the wrong enantiomer can be recycled. Some extra steps may be avoided if no covalent compound is formed at all. However, the fact remains that, even in the 21st century, most drugs that are sold as single enantiomers are manufactured by resolution. When you see a paper about the preparation of a single enantiomer that has in its title words like ‘practical’ ‘expedient’ or ‘efficient’ you may guess that resolution is going to be used. This situation will change. Asymmetric methods, the subjects of chapters 26–28, particularly the catalytic methods, gain in efficiency and ease of operation every year and are likely to become steadily more important.

When to Resolve General rule: resolve as early as possible Verapamil 33 is used in the treatment of cardiovascular disease. An asymmetric synthesis by the resolution strategy would normally be planned around a synthesis of the racemic compound and the important decision would be: when do you resolve?

CN MeO

MeO

Me N

OMe

OMe

33 verapamil

442

22 Resolution

The most satisfactory answer is ‘as early as possible’. If the starting material can be resolved then nothing is wasted. If the final product is resolved then half of the starting material, the reagents, energy, time and so on is wasted. And probably more than half; for few resolutions produce even close to 50% yield of the wanted enantiomer. Here is the outline racemic synthesis of verapamil without distracting details - where would you resolve?

CN

CN MeO

CN

H

CN

MeO

CN

hydrolysis MeO

MeO

MeO

MeO

CO2H

34

35

MeHN

35

37

Me

CN

OMe

OMe

36

MeO

N

OMe

O

MeO

reduce amide

(±)-33

OMe

38

These are the questions you should ask, and the answers in this case: 1. What is the first chiral intermediate? Answer: the starting material 34. 2. Is it a suitable compound for resolution? Answer: No doubt it could be resolved, though a nitrile is not particularly convenient, but the chiral centre is immediately destroyed in the next reaction. No. 3. Which is the first intermediate that can be conveniently and safely resolved? Answer: The carboxylic acid 36. It has a very helpful functional group and the chiral centre, being quaternary, is secure from racemisation. 4. Do any reactions occur later in the synthesis that might racemise the molecule? Answer: No. The one chiral centre is unchanged in the rest of the synthesis. We already have a good idea how to resolve a carboxylic acid by making a salt with an enantiomerically pure amine. In this case the first amine you think of, phenylethylamine 2, works very well. Here is the asymmetric synthesis, carried out on a 50–100 g scale at Celltech.7 The hydrolysis of the dinitrile 35 is chemoselective because the intermediate 39 is formed. The salt with 2 crystallises in good yield (39% out of a possible 50%) and in excellent ee. H N

HN CN

34

35; not isolated

0.5 mol% t-BuOK t-BuOH

O

remove t-BuOH

36 NaOH, H 2O, EtOH reflux, 18 hours

MeO

i-Pr

91% yield

39

MeO

CN H2N

2

Ph

36 EtOAc seed with product

MeO

Me2S.BF3

H

CO2

H3N

MeO 40: salt of 39 and 2, 39% yield, >95% ee

Ph

Verapamil

38 THF, 20 ˚C

22 Resolution of Diastereoisomers

443

Resolution of Diastereoisomers When the compound itself contains more than one chiral centre the question of diastereoisomers takes precedence over that of enantiomers. Resolution is normally performed on the wanted diastereoisomer rather than on the mixture. In the case of sertraline 45, an anti-depressant that affects serotonin levels in the brain, the active isomer was not known when both diastereoisomers were prepared by a unselective route.8 The starting material 41 was made by a Friedel-Crafts reaction between 1,2-dichlorobenzene and succinic anhydride. OH

O NaBH4

Cl

Cl

acidic

CO2

CO2H NaOH H2O, 80 ˚C

Cl

41

work-up

Cl

42

O

O

Cl

O

benzene

1. MeNH 2 TiCl4

Cl conc. H 2SO4

Cl 43

2. H 2, Pd/C

44

Cl

NHMe

NHMe OH

Cl

Cl Ph

Cl

Cl 45

CO2H

46; (+)-(R) mandelic acid

(+)-syn-45; sertraline

Separation of the syn and anti diastereoisomers by crystallisation of the HCl salt revealed that it was the syn diastereoisomer that was active and the reductive amination of 44 could be controlled to give 70% syn-45. The diastereoisomers of 45 were separated before the resolution. There is no point in resolving any earlier compound in the synthesis as even more material would be wasted in the reductive amination step. Natural (⫺)-(R)-mandelic acid 46 was a good resolving agent for 45 and 50% of the material derived from 44 could be isolated as the active (⫹)-syn-(1S,4S)-45.

Resolution of compounds made as diastereoisomeric mixtures It may be possible to prepare the correct diastereoisomer, assuming that this is known, by stereoselective synthesis and avoid the problem. The anti isomer of the amino alcohol 48 can be prepared from cyclohexene oxide 47 in high yield and with minimal contamination (⬍3%) of the syn-diastereoisomer.9

OH O

HO2C

CO2H

MeNH2 EtOH reflux

47

Me

O

O

NHMe (±)-anti-48 90% yield, >95% anti

O O 49; (+)-di-toluoyl tartaric acid

Me

444

22 Resolution

Resolution with tartaric acid 3 required up to seven recrystallisations to get pure material and by that time the yield was only 8%. Di-p-toluoyl tartaric acid 49 (cf 9 used earlier) was spectacularly better when used in the right proportions (4:1 48:49). The solubility of the required diastereoisomer as the salt of one molecule of (⫹)-49 with two molecules of (⫹)-48 was very much less than that of (⫹)-49 with two molecules of (⫺)-48 so that merely mixing 48 and (⫹)-49 in the right proportions in ethanol at 60 ⬚C for twenty minutes, cooling, and filtering off the crystals gave a 45% yield in ⬎99% ee. Neutralisation with NaOH and extraction with t-BuOMe gave pure (⫹)-48. OH

49

OH

OH NaOH

(±)-anti-48 EtOH, 60 ˚C

NH2Me.(+)-49

NHMe (–)-anti-48 in solution

t-BuOMe

salt of (+)-anti-48 and (+)-49 crystallises out in 45% yield

NHMe (+)-anti-48 98% yield, >99% ee

The synthesis of Jacobsen’s Mn(III) epoxidation catalyst by resolution Possibly the easiest resolution known is of the related trans diaminocyclohexane 50, used to make the catalyst for Jacobsen’s asymmetric epoxidation (chapter 25). It is not even necessary to separate the diastereoisomers first and this is a big advantage as the commercial mixture of about 40:60 cis and trans-50 costs about one tenth of the pure racemic trans and about one hundredth of the resolved trans isomer. You can usually tell if a commercial product is made by resolution as the two enantiomers cost about the same.

H2N

NH2

H2N

NH2

enantiomers of trans-50

H2N

NH2

achiral cis-50

The resolving agent is tartaric acid: 150 g are dissolved in water in a litre beaker. Then 240 ml of the mixture of isomers of 50 is added at 70 ⬚C followed by 100 mls acetic acid at 90 ⬚C and the solution cooled to 5 ⬚C. The pure salt 51 separates out in 99% yield - that is 99% of all that enantiomer originally present - and with 99% ee. This is almost incredibly good. Though the free trans-diamine 50 can be isolated from this salt, it is air-sensitive and it is better to make the chiral catalyst 52 directly from the salt as shown. The yield is better than 95% and the catalyst 52 can be made in multi-kilogram quantities by this resolution.10

HO2C HO

CO2H

1. add mixture of cis and trans 50 at 70 ˚C

OH

2. 100ml HOAc at 90 ˚C 3. cool to 5 ˚C, filter

(+)-tartaric acid 150 g in 400ml water

H3N O2C

NH3 CO2

1. ArCHO K2CO3 2. Mn(OAc) 2 air t-Bu 3. NaCl

HO OH 51; 160 g 99% yield, >99% ee

N

N Mn

O Cl O

t-Bu

t-Bu 52

t-Bu

22 Resolution of Diastereoisomers

445

Resolution with half an equivalent of resolving agent Since only half the compound (the maximum amount of either enantiomer) crystallises out, it may seem extravagant to use a whole equivalent of resolving agent and in some cases the resolution is much better with only just enough to crystallise one enantiomer, as with 48. A case in point is methylphenidate11 53. The HCl salt of racemic syn 53 is marketed as ‘Ritalin’ for treatment of children with ADHD (attention deficit hyperactivity disorder). The resolving agent is unlike anything we have seen so far: an axially chiral BINOL-derived cyclic phosphate 55. If the right amount is added to a solution of 54 to crystallise just one enantiomer a very good yield (38% out of 50%) of the salt can be isolated on a 50 g scale. Separation is now easier as one enantiomer is a salt and the other a neutral compound: simple solvent/solvent extraction is used. The active enantiomer, (R,R)-53, can be isolated in an overall yield of 32%.

53; racemic syn ('threo-') diastereoisomer H2 N

Ph

H2 N

CO2Me

53; active enantiomer

Ph resolution

CO2Me

Cl

H2 N

Ph CO2Me

Cl

Cl 32% yield, >99.8% ee

1. i-PrOAc 2. NaOH, NaCl, H 2O

H N

3. separate organic layer

Ph

H N

CO2Me

1. conc. HCl 2. crystallise

Ph CO2Me H N

54; racemic syn ('threo-') diastereoisomer 0.5 equivs resolving agent 55 i-PrOAc, MeOH 60-65 ˚C seed with salt of product then cool to 0-5 ˚C

O

Ph CO2Me

(R,R)-54 in organic layer

O P OH

O

Resolving Agent: binaphthyl hydrogen phosphate 55

O

O P

O

O

H2 N

Ph CO2Me

NaOH

organic layer i-PrOAc

H2O

aqueous layer salt of 55 and (R,R)-54; 36% yield on 50 g scale

It is not possible to give a comprehensive guide to the essentially practical skill of classical resolution in this book. You are referred to the papers we quote and also to Eliel and Wilen, chapter 7, pages 297 to 441 for a fuller account. We must move on to other styles of resolutions particularly those that do not involve the separation of specially prepared diastereoisomers.

446

22 Resolution

Physical Separation of Enantiomers Chromatography on chiral columns One good way to separate enantiomers physically is separation on a chiral chromatography column. There are now many of these available12 usually consisting of silica functionalised with a linker such as a 3-sulfanyl- or 3-amino-propylsilyl group 56 to which are attached enantiomerically pure groups such as the covalently bound anthryl alcohol 57. Separation occurs when suitable racemic compounds are passed down the column, usually in hexane containing about 10% i-PrOH. This method is one of the best ways to assess the enantiomeric purity of a compound and ees are routinely measured using chiral columns. OH OH silica surface

(EtO)3Si

NH2

Si

NH2

O

OH

CF3

OMe

O

OEt

O

H

Si

S

O

56

57

A column loaded with the amide 58 that is held in place merely by hydrogen bonds and electrostatic forces is used preparatively to resolve the important axially chiral binaphthol13 59. You will meet compounds of this type as reagents and ligands in chiral catalysts in chapters 24–26.

H HO2C

O

Ph

NO2

N H

NEt2

HN

OH OH

NO2

N 60; chloroquine

Cl 59; binaphthol

58

The anti-malarial compound chloroquine 60 is a salutary case. For many years it was thought that both enantiomers were equally active against the parasites that transmit malaria. This was because the only optically active samples available (rotations ⫹12.3 and ⫺13.2) were obtained by conventional resolution with bromocamphor sulfonic acid 61 and were of low purity. H HN

Cl

N (+)-(S)-chloroquine 60 [α]D +12.6

H

NEt2

HN

Cl

N (–)-(R)-chloroquine 60 [α]D –13.2

NEt2

Br

O SO2OH 61; bromo-camphor sulfonic acid

When the racemic compound was resolved on a poly-N-alanylacrylamide column, samples of rotation ⫹86.9 and ⫺86.9 showed not only that the (⫹) isomer of 60 was more active against

22 Physical Separation of Enantiomers

447

malaria, but that it also had fewer toxic side-effects. The active enantiomer is now made from the dichloroquinoline 63 and the enantiomerically pure diamine 62 prepared by conventional resolution. H Ph

CO2Et

HN

chromatography of racemic 60

O

(+)-(S)-60 chloroquine [α]D +86.9

H

N

Cl

NEt2

H2N

n

Cl

nucleophilic aromatic substitution

62; prepared by conventional resolution

poly-N-alanyl acrylamide

63

Resolution of triazole fungicides by HPLC The triazole fungicides of general structure 64 such as hexaconazole 65 and flutriafol 66 are a rare case of human and plant medicine using similar compounds. They were initially used as racemates but it was soon essential to discover the active enantiomers. Conventional resolution by crystallisation of diastereomeric derivatives proved difficult. F HO

Cl HO

R

X

N

N

Cl

N N 64; a triazole fungicide

F HO

N

N

N

N 65; hexaconazole

N 66; flutriafol

The solution was to make the usual sort of diastereoisomers by acylation with camphanic acid chloride and to separate them by standard (not chiral) HPLC on Dupont ‘Zorbax’ columns. Only 60 mg was separated at a time but that was enough to accumulate grams of material. The craft needed in such separations is best illustrated by the solvent composition needed for good separation of 65. You must use 918:80:2 of F3C-CCl3, MeCN, and Et3N. The (⫺) isomer was biologically active.14 O 1. NaH DMF

2.

COCl O

O O Ar

64 hexaconazole

O (–)-camphanic acid chloride

O

1. HPLC

N N 67

N

2. NaOMe THF

separated (+)- and (–)-64 hexaconazole

If you need any more convincing, applying the same method to flutriafol 66 gave the camphanic esters as before but now no separation could be achieved even with HPLC. Esters 69 of a different acid 68 could be separated on the same column but using 1:1 CH2Cl2 /EtOAc as eluent. You will not be surprised to know that fluconazole 70 is now a leading fungicide in this area. It is not chiral.

448

22 Resolution

F

66 flutriafol

OPh

1. NaH, DMF

F 2.

1. HPLC

O

OPh

ClOC (+)-68

2. NaOMe THF

F HO

separated (+)- and (–)-66 flutriafol

N N

F

O

N

N

N

N

N

N N

70; fluconazole

69

A commercial drug separation by chiral HPLC Cetirizine 71·2HCl is an antihistamine marketed as Zyrtec. It has a ‘low grade’ chiral centre (arrowed) - the molecule is chiral only because one of the benzene rings has a para-chloro substituent. It is very difficult to resolve cetirizine or to synthesise it asymmetrically. One company, Sepracor Inc., whose business is making single enantiomers, found that the related amide 72 could be separated by chiral HPLC on Chiral Technologies ‘Chiralpak AD’ columns: the two enantiomers having very different retention times (4.8 and 8.8 minutes). They could separate nearly 40 g of racemic material with one injection and, by repeated injections, could easily separate 1.6 kg of (⫺)-(R)-72 with 99.8% ee. The separate enantiomers were converted to cetirizine in two steps and the (⫹)-(R)-enantiomer 73 found to be biologically active.15 O

CO2H

O

N

N

N

N

Cl

Cl 71; cetirizine (as .2HCl)

CONH2

O

CO2H

N 1. HCl, MeOH 2. HCl, H 2O

N

.2HCl

Cl (–)-(R)-amide 72

(+)-(R)-cetrizine 73

Chiral HPLC is the method of choice for analysing enantiomers and determining % ee. It can be used preparatively. In either application it is best to consult an expert when choosing columns and solvents.

Differential Crystallisation or Entrainment of Racemates Conglomerates and racemic compounds Most chiral compounds crystallise as racemic crystals, each crystal containing equal numbers of right and left handed molecules, and cannot be separated by crystallisation. Some racemates, unfortunately only about 15% of those known, crystallise as “conglomerates” or “racemic compounds”; that is each crystal consists only of one enantiomer though the crystalline mass contains equal numbers of right and left handed crystals.16 You may recall that Pasteur17 did the very first resolution by picking out right and left handed tartrate crystals by eye. If a saturated solution of such a compound is seeded with crystals of one enantiomer (usually quite a lot is needed: 5–25% by weight), that enantiomer may crystallise out first in the process known as differential crystallisation (or sometimes as entrainment).

22 Differential Crystallisation or Entrainment of Racemates

449

Racemates (Racemic Mixtures) 1:1 mixtures of enantiomers of any kind Conglomerates Each crystal contains a single enantiomer

Racemic Compounds Each crystal contains a 1:1 mixture of enantiomers

10-15% of compounds

*crystallisation of racemic compounds of >85% ee usually enhances optical purity

85-90% of compounds

Conglomerates can be recognised by a number of features: 1. The m.p. of the racemate is the same as that of the single enantiomer 2. The IR spectrum of the solid racemate is the same as that of the single enantiomer 3. The racemate is more soluble than either enantiomer in a chosen solvent. There is no guarantee that a given group of molecules nor any derivatives of them will provide conglomerates but there are some well known cases, thus with α-amino acids, certain known derivatives, such as the N-acetyl amides, generally crystallise as conglomerates. We shall give some examples to show how the method works.

Typical procedure for differential crystallisation (entrainment) cis-Stilbene diols form conglomerates. A solution of 11 g racemic cis stilbene diol 74 and a small amount (usually about 5–10%, here 0.37 g) of (⫺)-74 in hot ethanol is cooled and seeded with 10 mg (⫺)-74. After 20 minutes 0.87 g pure (⫺)-74 has crystallised out. The excess of the one enantiomer is less soluble than the racemate but the yield is only about twice as much as the amount of pure enantiomer added in the first place. OH

OH Ph +

Ph

OH Ph +

Ph

OH

11 g syn racemic 74

Ph

Ph

OH

1. dissolve in 85 g 95% EtOH by heating

OH 0.37 g (S,S)-(–) 74

2. Cool to 15 ˚C seed with 10 mg (–) 74 stir 20 min

OH Ph

Ph OH

0.87 g (S,S)-(–) 74 optically pure

The solution is now enriched in the other enantiomer so we replace the lost racemate by adding another 0.87 g heating and cooling as before but seeding with the other enantiomer (⫹)-74. We get about 0.87 g of (⫹)-74 crystallising out. And so on. After fifteen cycles 6.5 g (⫺)-74 and 5.7 g (⫹)-74, each 97% ee, had been separated. This compound is more efficiently prepared by asymmetric dihydroxylation (chapter 25).

mother liquor

OH

1. add 0.87 g racemate heat to dissolve 2. Cool to –15 ˚C seed with 10 mg (+) 74 stir 20 min

Ph

Ph OH

0.87 g (R,R)-(+) 74 optically pure

450

22 Resolution

Conventional resolution of L-methyl DOPA The aromatic amino acid L-methyl DOPA (Di Hydroxy PhenylAlanine) 80 is used to make an antihypertensive compound. Synthesis by the Strecker method clearly requires the aromatic ketone 77, and the synthesis follows the pattern below.18 The intermediates and final product have been resolved in various ways. CN

CN EtOAc NaOAc

O

O

EtOH

O

O

H2SO4

O

O

EtOH, H 2O

O

O

76; 98% yield

75

(NH4)2CO3, KCN

H2O

77; 57% yield

Me O O NH

HN

O

Ba(OH)2 H2O

48% HBr

NH2

O

O

78; 88% yield

CO2H

O

HO

CO2H NH2

HO

(±)-79, 100% yield

(±)-80, 84% yield

The synthesis of L-methyl DOPA 80 by the Strecker reaction was straightforward and of course produced racemic material. A conventional resolution by crystallising the menthyl ester 83 from hexane and hydrolysis of acetal, ester and amide in 48% HBr (note that no racemisation by enolisation can occur) gives good yields.19

(±)-79

O

CO2H

O Ac2O

Ac2O

HN

O

heat

O

O

81

N

O

O

82

O menthol base

O

1. crystallise from hexane

O NHAc

O

2. 48% HBr

(S)-80 L-methyl DOPA

83

Resolution of L-methyl DOPA by differential crystallisation Careful study of m.p./composition diagrams and infra red spectra revealed that aryl sulfonate salts of aromatic amino-acids may form conglomerates, and 79 has indeed been purified by differential crystallisation. Batches of racemic salt are seeded with about 20% by weight of pure L-79. This gives a yield of about 40% of pure L-79. This is again about twice the original excess of the single enantiomer. The mother liquor is rich in D-79, so seeding with that enantiomer gives a similar yield of pure D-79, and the process can be repeated.20 CO2H

O

NH2

O (±)-79

O O

OH

CO2H NH3

O3S

sulfonate salt of 79

22 Resolution with Racemisation

451

Finding a differential crystallisation approach to fenfluramine The anorectic drug fenfluramine 85 is made by the alkylation of the simpler amine 84. Either compound could be resolved and, though a classical resolution of the camphoric acid salt of fenfluramine was known, chemists at Rouen were determined to achieve better results by differential crystallisation.21 CF3

CF3 NH2

HN

84

85; Fenfluramine

Just how determined you will see. They looked at salts of both amines with ‘about fifty’ achiral acids of which eight proved to be conglomerates, three for 84 and five for 85 (this is about what would be expected - about 10%). Of these eight, two could not be separated by crystallisation because one salt 86 had crystal facets that acted as seeds for the other enantiomer while the conglomerate of another 87 was unstable and easily reverted to a racemic compound. CF3

CF3 NH3

H2N

CHCl2CO2

86; salt of 84 with chloracetic acid

PhO

CO2

87; salt of 85 with phenoxyacetic acid

That left six candidates, two for 84 and four for 85. All six salts could be separated by crystallisation as we have described giving alternately one enantiomer and then the other but the best were the salts of 85 with the two arylacetic acids. You may feel that this heroic effort, though successful, is rather discouraging. CF3

CF3 NH3

H2N

salts of 84 with acids: Ph

CO2

Me

salts of 85 with acids: SO2O

Ph

CO2

Ph3C

CO2 CO2

CO2 Cl

O2N

Resolution with Racemisation Resolution of amino acids by differential crystallisation with racemisation The separation of the enantiomers of most amino acids can be achieved by differential crystallisation of their N-acetyl derivatives, such as that of leucine. Racemic N-acetyl leucine 88 is dissolved in the right solvent mix cooled and seeded with 4% by weight of natural (S)-88. Pure (S)-88 crystallises out in good yield.22

452

22 Resolution

solubility in 10:1 AcOH:Ac2O is 88 g CO2H

1. cat Ac 2O (10 ml) 90 ml AcOH, reflux

CO2H

2. cool to 100 ˚C and seed with 6 g enantiomer

HN

HN 3. cool at 10 ˚C/hour with stirring

O 150 g racemic 88

O 112.6 g (S)-88, 98.8% ee

4. at 75 ˚C, add 10 ml Ac 2O 5. stop at 40 ˚C and isolate

In fact your suspicions may have been aroused by the quantity of material put in. We started with 150 g racemic 88, that is 75 g of each enantiomer and we seeded with 6 g of (S)-88 making 81 g of (S)-88 altogether. But the yield of pure crystalline (S)-88 was 112.6 g - too much! Clearly the other enantiomer is somehow being converted into the enantiomer that crystallises. The clue is the addition of that extra Ac2O at step 4. This forms a mixed anhydride 89 that racemises by enolisation 90 and crystallisation can continue. O CO2H

Ac2O

O

OH

O

HN

O

HN O

HN

O 89

88

O

O 90

Sometimes these differential crystallisations with racemisations are very easy to do. Racemic N-butyroyl proline 91 gives a good yield of one enantiomer in moderate ee just by melting the racemic compound with catalytic acetic anhydride and seeding with one enantiomer. Further crystallisations improve the ee. The yield is 6.3 from 10.5 g or 60% of the total material. This process is clearly a great improvement on simple differential crystallisation both in simplicity of operation and because it is no longer necessary to alternate the isolation of enantiomers. 100 ˚C (melt), cat Ac 2O

CO2H

N

N

seed at 100 ˚C with 0.5 g enantiomer then add toluene

O 10 g racemic 91

CO2H

O 6.3 g (S)-91, 62% ee

Differential crystallisation and racemisation when enolisation is impossible We return to the asymmetric synthesis of L-DOPA noting that it is an amino acid that cannot racemise by enolisation as it has no proton between the NH2 and CO2H groups. We again use the Strecker reaction - the only change is a minor alteration of phenolic protecting groups. The Strecker synthesis gave a good yield of the amino-nitrile 93 that could be converted into L-DOPA 80 with conc. HCl. Unfortunately no derivatives of 93 could be separated by differential crystallisation.

MeO

MeO

O

CN

NH3, HCN i-PrOH

HO 92

MeO

CN

Ac2O

NH2 HO 93; 93.5% yield on 400 g scale

97% yield

NHAc HO 94; 25 g racemic

22 Resolution with Racemisation

453

When the N-acetyl derivative of the intermediate 94 was crystallised from isopropanol with seeding a good yield of enantiomerically pure L-DOPA could be crystallised out. Treatment of the residue from the mother liquor with NaCN in DMSO led to complete racemisation and the differential crystallisation could be repeated.23 MeO

crystallise from i-PrOH

HO

CN

CO2H

conc. HCl

(±)-94

NH2

NHAc seed with 5% enantiomer

HO

HO

heat

(S)-94

(S)-80; 100% yield

This racemisation is a separate chemical reaction from the crystallisation but one cycle gave 63% yield of L-DOPA of 100% ee. Evidently the cyanide is lost from 94 by elimination and readdition 95 gives racemic 94. MeO

MeO

CN N

HO

H

N

CN HO

O

(±)-94

CN

O

(R)-94

95

Kinetic resolution with racemisation Amine (S)-(⫹)-97 is needed for the synthesis of a gastrointestinal hormone antagonist, Merck’s cholecystokinin antagonist 96, by acylation with indole-2-carboxylic acid. Me

Me

O C–N

H N

N

N H

O

N NH2

O

N H

amide

+ HO2C N H

N H Ph

Ph 96; L-364,718

97

The racemic compound is available by methods used (Disconnection Textbook, page 250) for the closely related benzodiazepines like diazepam, used in the treatment of depression. The one stereogenic centre is next to a primary amine, and the compound forms a crystalline salt with camphor sulfonic acid 98 (cf. 61). The maximum yield is 50%, but the amine (S)-97 can be released from the salt simply by neutralisation.24 This is a classical resolution by crystallisation of diastereoisomers. Me

Me

O

O

N

N

NH3

NH2 + N Ph racemic 97

O SO2OH 98; (+)-camphor sulfonic acid

N H

O SO2O

Ph (+)-97 salt crystallises in 40-42% yield

A remarkable improvement happens if the salt is crystallised in the presence of 3,5-dichlorobenzaldehyde 99: 90% optically active salt crystallises out. Clearly the non-crystalline enantiomer is racemising via the still chiral imine 100 by imine exchange with achiral imine 101.

454

22 Resolution

Me (+)-camphor sulfonic acid 3 mol % ArCHO, 20-15 °C

6 kg racemic 97

O

N NH2

seed with 10 g (S)-(+)-97 crystallise from MeCN/i-PrOAc then neutralise

Cl

N H Ph

99 3,5-dichloro benzaldehyde OHC

90% yield (S)-(+)-97

Cl

Cl Me

Cl

Me

O

N

O

N N

Cl

N

N Ph

Cl

N Ph

100

101

Other methods of racemisation during resolution: the Mannich reaction Even more complicated reactions can be used to racemise during a resolution. The amino ketone 102 is needed for the synthesis of the analgesic and useful asymmetric reagent (see chapter 24) DARVON. Classical resolution with dibenzoyl tartaric acid 9 succeeds in crystallising the (⫹) enantiomer and racemising the mother liquors by reverse Mannich reaction.25 O

PhCOO

Ph

H

OCOPh

NMe2

HO2C

reversible Mannich

Ph NMe2

CO2H

O

O

OH Ph

OCOPh

O2C

CO2H

9; (–)-dibenzoyl tartaric acid

racemic ketone (±)-102

CH2

PhCOO crystallise

+

Ph H

H

NMe2

(–)-(S)-102; this enantiomer racemises by reverse Mannich

NMe2 H

salt of (+)-(R)-102 and 9 crystallises out

The enantiomerically pure ketone (⫹)-(R)-102 can be converted to DARVON 104 by a chelation controlled diastereoselective addition of benzyl Grignard and acylation. Using the other enantiomer of 9 gives the other enantiomer (⫺)-(S)-104 which is called NOVRAD. O Ph

O Ph

Ph

H

NMe2

(+)-(R)-102

MgCl

OH EtCOCl

Ph

H 103

Ph

O

NMe2 Ph

H

NMe2

104

A case where two stereogenic centres are equilibrated in one step, and a third in another occurred in Oppolzer’s synthesis26 of (⫹)-vincamine 105, a natural alkaloid used for the treatment of cerebral insufficiency in man. It has three stereogenic centres, but one of them epimerises at

22 Resolution with Racemisation

455

room temperature by equilibration with the ketone 106 so need not be controlled - it equilibrates in nature too, so it will automatically be correct. H 105; (+)-vincamine

N

H

N

N H

N

106

HO CO2Me MeO2C

O

Earlier in the synthesis the heterocycle 107 was combined with 108 to give an advanced intermediate 109 as a mixture of diastereoisomers. As both starting materials are achiral 109 is also racemic. OSiMe3 N

N H

H

+

N H OHC

Br

racemic (±)-109 mixture of diastereoisomers

108

107

N

Heating the mixture of isomers of 109 with TsOH equilibrates the diastereoisomers so that the required more stable syn (H and Et syn) diastereoisomer of 109 crystallises out. Treating this with (⫹) malic acid leads to crystals of the natural enantiomer (⫹)-syn-109. Both the unwanted antidiastereoisomer and the unwanted enantiomer can be equilibrated again with TsOH. This cannot be enolisation as there are no α-hydrogens and is presumably equilibration by reversible Mannich reaction as 110 lacks either stereogenic centre and so epimerises both!

racemic (±)-109 as mixture of diastereoisomers

TsOH

H

TsOH

N

N H

N H OHC

HO

racemic syn-109

110 crystallise (+)-malate

TsOH

HO

CO2H CO2H

(R)-(+)-malic acid

H N H

N

N

H N H

N

OHC

OHC

(–)-syn-109 in mother liquor

crystals of (+)-syn-109

Resolution with racemisation in the manufacture of a drug Resolution with racemisation is used in the manufacture of the cardioprotective drug CP-060S 111 by the Chugai Pharmaceutical company.27 The drug itself can be resolved with some difficulty but as it is made from the simpler carboxylic acid 112 this looks a better bet.

456

22 Resolution

S t-Bu

CO2H

HO2C

O

S t-Bu

N N

HO

O

CO2H

O HO

Me

t-Bu

O N

O

t-Bu

(S)-111; CP-060S as hydrogen fumarate

112

The resolving agent for the acid 112 is the simple amine 6 that we discussed at the start of this chapter. A 27% yield of the salt of (S)-(⫺)-6 with the required (S)-112 with ee ⬎99% was crystallised from i-PrOH/i-Pr2O. This is not a good recovery but they preferred to sacrifice yield to near perfect ee as the mother liquor was racemised by treatment with NaOH in water at room temperature. The resulting solution had ⬍1% ee – the one time when low ee is a good thing – and was used in the next resolution. S

Me (±)-112 +

t-Bu Ph

N Ph H (S)-(–)-6

Me

O

dil HCl

N

CO2 Ph

HO

N H2

Ph

(S)-112 23% overall yield 99.8% ee

t-Bu salt of (S)-112 and (S)-(–)-6; 27% yield, >99% ee

You are fortunate if you can find a way to resolve and racemise in the same solution but, in theory, any reversible reaction such as the important Diels-Alder or aldol reactions could do the job so this method has wider application than might appear at first sight.

Kinetic Resolution with Enzymes Enzymes as resolving agents There is a chapter soon (29) about enzymes as reagents in organic synthesis, but they are also widely used in industry as resolving agents. The basis of this application is the enantioselectivity of enzymes. They react with only one enantiomer of a racemic mixture and can be used to create the wanted enantiomer or destroy the unwanted. A process in which one enantiomer reacts and the other does not is called a kinetic resolution (chapter 28) since the resolution depends on the rates of two competing reactions. The result is that instead of separating enantiomers or even diastereoisomers, one is separating two compounds of different structure that also happen to belong to the two opposite stereochemical series. Zeneca market a herbicide for broad leaved crops, Fusilade28 (fluazifop butyl) 113. This is a carboxylic ester with two ether linkages, one between two aromatic rings. O CF3

O H N

O

(R)-113 Fluazifop Butyl "Fusilade"

O

Disconnection of the ethers 113a is guided by our mechanistic knowledge that nucleophilic substitution is possible on alkyl halides, and on 2-halo-pyridines such as 114, but not easy on halobenzenes.

22 Kinetic Resolution with Enzymes

457

O CF3

O H N

OBu

3 x C–O one ester two ethers

O 113a

CF3

O

OH + Cl

+ N

Cl

OH

H

HO

114

115

+ BuOH

(S)-116

The only chiral intermediate is the chloroacid 116 which Zeneca manufacture as a racemic compound. They use the enzyme chloropropionic acid dehalogenase to destroy the unwanted isomer by conversion to lactic acid. It is easy to separate these two compounds as they are not even isomers.

O

O chloropropionic acid dehalogenase

Cl H

OH

Cl H

racemic (±)-116

O OH +

HO H

OH

(+)-(R)-lactic acid (natural)

(S)-116

It is important to follow the fate of stereogenic centres made early in a synthesis and they have established that the nucleophilic substitution of 115 with 116 to give 117 goes with inversion and not, as happened to amino acid derivatives in chapter 23, with unexpected retention. The active product is therefore the (R)-113 enantiomer shown.

O (S)-116

114

O

CF3

H

115 NaOH

CO2H

H 2 x NaOH

HO

CO2H

N

117

BuOH

(R)-113 H

O 118

Kinetic resolution by ester hydrolysis with enzymes By far the commonest reaction used in kinetic resolution by enzymes is ester formation or hydrolysis. Normally one enantiomer of the ester is formed or hydrolysed leaving the other untouched so one has the easy job of separating an ester from either an acid or an alcohol. There are broadly two kinds of enzymes that do this job. Lipases hydrolyse esters of chiral alcohols with achiral acids such as 119 while esterases hydrolyse esters of chiral acids and achiral alcohols such as 122. Be warned: this definition is by no mans hard and fast! If the unreacted component (120 or 123) is wanted, the reaction is run to just over 50% completion, to ensure complete destruction of the unwanted enantiomer, while if the reacted component (121 or 124) is wanted it is best to stop short of 50% completion so that little of the unwanted enantiomer reacts.

458

22 Resolution

Me

O

Me

lipase

Me

O +

R

R

O (±)-119

+ AcOH R

O (S)-120

OH

(R)-121 O

O

O esterase

R

MeO

R

MeO

+

+ MeOH

Me (R)-124

Me (S)-123

Me (±)-122

R

HO

There is an inherent problem with either type of enzyme as the reactions are reversible. One way to make the reaction run in the direction of ester formation is to use a non-aqueous solvent (you may be surprised that enzymes function in, say, heptane, in which they are insoluble, but lipases do). One way to make the reaction run in the other direction is to make the alcohol component an enol so that, on hydrolysis, it gives the aldehyde or ketone and does not reverse.

Resolutions of secondary alcohols by lipases Simple secondary alcohols 121 (R ⫽ alkyl or aryl) are enantioselectively esterified by the reactive trichloroethyl ester 125 using porcine pancreatic lipase in anhydrous ether. The products, one enantiomer of 121 and the other enantiomer of the ester 126, are both formed in ⬎90% ee, and are easily separated from each other and from the insoluble enzyme.29 O

Me

Me

Me

porcine pancreatic lipase

+ O 125

R OH (±)-121

CCl3

R OH (S)-121

anhydrous ether

O

+ R

O (R)-126

The reaction occurs in the reverse direction in aqueous buffer (pH 7, 20 ⬚C) using a lipase from Pseudomonas spp. The reactive chloroacetates, e.g. 127, give the best results with nearly quantitative yields of (R) alcohol 128 and (S) ester 127 separated by flash chromatography on a 250 g scale. The ester (S)-127 was easily hydrolysed to the (S) alcohol 128 without racemisation.30

Me Ph

O

O (±)-127

Pseudomonas lipase

Cl

Me

Me

O

+ water, pH 7, 20 °C

Ph OH Ph (R)-(+)-128 48% yield, 99% ee

O (S)-(–)-127 47% yield

K2CO3

Cl MeOH

Me Ph OH (S)-(–)-128 97% yield, 99.5% ee

The same enzyme catalyses the esterification of racemic 129 in non-aqueous solution with vinyl acetate. The released alcohol is CH2⫽CHOH, the enol of acetaldehyde: it immediately forms acetaldehyde which self condenses and is removed from the equilibrium. The enzyme is filtered off, the enantiomerically pure alcohol (S)-129 and acetate (R)-130 separated by flash chromatography, and the ester hydrolysed to the alcohol without racemisation. Either method (esterification or hydrolysis) gives both enantiomers of a range of secondary alcohols.31

22 Kinetic Resolution with Enzymes

OH

459

OH

OH

OAc

OAc Pseudomonas lipase

K2CO3

+

MeOH

t-BuOMe, 20 ˚C

(+)-(S)-129 48%yield, 95% ee

(±)-129

(+)-(R)-130 46%yield, >99% ee

(–)-(R)-129 48%yield, 95% ee

Kinetic resolution with proteolytic enzymes The simple furan alcohol 131 is successfully resolved32 with a lipase from Candida cyclindracea and you should note that the same enzyme is used to form the octanoate 132 and, under different conditions, to hydrolyse it to the pure alcohol (⫹)-(R)-131. C7H15CO2H Candida lipase

O OH (±)-131

+

O

hexane room temperature

OH (–)-(S)-131

Candida lipase

O OCOR (+)-(R)-132

O

water, pH 7.5 room temperature

OH (+)-(R)-131

However, the closely related amino acid 133 was not a substrate for either lipase (from pigs or Candida) but could be resolved with the proteolytic enzyme papain. This acted as an esterase, hydrolysing the methyl ester rather than the amide. Note that this kinetic resolution produces a single enantiomer of the carboxylic acid rather than the alcohol and that separation of 134 from 133 is very easy as the free acid can be extracted from organic solvents by aqueous base in which it is soluble as the anion.

CO2Me

papain

O

CO2Me O

NHCO2Et (±)-133

DMF/H2O pH 7

+

CH2N2

CO2

NHCO2Et (+)-(S)-133

CO2Me O

O NHCO2Et (–)-(R)-134

pH 3

NHCO2Et (–)-(R)-133 42% yield, 97% ee

Perhaps the ideal enzymatic resolution is that of phenylalanine 137 by the proteolytic enzyme subilopeptidase, marketed as Alcalase® acting as a lipase on the ester amides 135. The reaction stops after one enantiomer is consumed: the yields and ees of the two products are close to 100%. The amide (S)-136 can be hydrolysed directly to natural phenylalanine (S)-137. The unnatural (R)-135 can of course be hydrolysed to the arguably more valuable (R)-137 but it can also be racemised by NaOMe in MeOH for the next cycle of reactions.33 CO2Me

CO2Me

Alcalase®

CO2H

5M HCl

+ NHAc (±)-135

water, pH 7.5 45 minutes room temperature

NHAc (–)-(R)-135, 98% yield CO2H NH2 (+)-(S)-137 78% yield, 98% ee

NHAc (+)-(S)-136 96% yield, 98% ee

reflux

460

22 Resolution

More recently acids have been resolved with enzymes cloned and over-expressed in their own organisms, such as an esterase from Bacillus subtilis that resolves ibuprofen methyl ester 138 to give ibuprofen 139 of 99% ee. A range of anti-inflammatory arylpropionic esters, including 138, could also be resolved with a cell-free extract from Pseudomonas fluorescens showing that purified enzymes are not essential.34

CO2Me

CO2H

esterases

(S)-139; Ibuprofen

(±)-138

Kinetic resolution with racemisation using proteolytic enzymes In all these enzymatic resolutions so far the maximum yield of one enantiomer is of course 50% and only in the last example have we seen the recycling of the other enantiomer. However in the resolution of esters 140 of the anti-inflammatory drug ketorolac 141, the easily enolised ester 140 racemises in moderately basic conditions and, after a survey of four lipases and four proteases, one was found35 that hydrolyses the ester 140 enantioselectively at pH 9.7. The product of the reaction is (⫺)-(S)-ketorolac 141 in 92% yield but only 85% ee. One recrystallisation improves this to 94% ee.

CO2Me

Streptomyces griseus protease

N

CO2H

(+)-(R)-140 +

N

water, pH 9.7

O

O (–) (S)-141; 92% yield, 85% ee

(±)-140

Kinetic resolution on diastereoisomeric mixtures Derivatives of chrysanthemic acid such as (1R,3R)-permethrinic acid 143 are in demand for the manufacture of highly specific insecticides that do not persist in the environment. Mixtures of the esters 142 are easy to make and contain various proportions of the cis and trans diastereoisomers. Pig liver esterase accepts only the trans esters as substrates so complete hydrolysis gives the unchanged cis esters and hydrolysed but poorly resolved trans acids. At 50% conversion, kinetic resolution of the trans esters occurs.36 Cl porcine liver esterase

Cl Cl

Cl

CO2Me

Cl

CO2Me (1R,3S)-cis 142

Cl

CO2Me (1S,3R)-cis 142

Cl

Cl

(±)-cis/trans-142 Cl

CO2Me (–)-(1S,3S)-trans 142

Cl

CO2H

(+)-(1R,3R)-trans 143

22 Asymmetric Synthesis of Prostaglandins with many Chiral Centres

461

At 50% conversion, the product contains three esters, the two cis-142s and the (1S,3S)-trans ester together with a little (1S,3S) -trans acid and all of the (1R,3R)-permethrinic acid 143. The ratio of these last two is 10:90 but one recrystallisation from petrol gives (1R,3R)-permethrinic acid 143 in 98% ee. This approach works for several different cyclopropane-based carboxylic acids in much the same way.

Comparison between enzymatic and classical resolution Earlier in this chapter we described the resolution of the piperidine 8 by classical resolution. The Bristol-Meyers Squibb company also investigated an enzymatic resolution.3 Various esterases and lipases had been used before on various derivatives of 8 but none was efficient. Two methods were successful. Accurel PP, a lipase immobilised on polypropylene, resolved the N-Cbz derivative 10 to give 46% yield of (R)-(⫺)-10 with 99.4% ee. Unfortunately this is not the wanted enantiomer and the ester (S)-(⫹)-144 could be isolated only by chromatography.

OH

OAc

OH

OH

OAc +

Accurel PP

N

N H O (±)-8

Ot-Bu 10

N

N Ot-Bu

O

Ot-Bu

O

(R)-(–)-10

(S)-(+)-144

The separation problem was solved by esterifying racemic 10 with succinic anhydride. The (R)-ester 145 now has a free carboxylic acid group and can be separated simply by extraction with weak base (NaHCO3) without any chromatography. The ester is easily hydrolysed to (S)(⫹)-10 which was obtained in 32% overall yield (maximum 50%) and 98.9% ee. They do not say whether they prefer this resolution to the classical version with dibenzoyl tartaric acid but both are good. O

10 +

O

O

CO2H

+ N

N

Accurel PP

succinic anhydride

O

OH O

O

Ot-Bu (R)-(–)-10

O

Ot-Bu (S)-(+)-145

Asymmetric Synthesis of Prostaglandins with many Chiral Centres We end with a stereochemically involved synthesis of a prostaglandin 155. The racemic synthesis is summarised below - each compound is a single diastereoisomer and all from 146 to 155 are chiral but all are racemic as the synthesis starts with achiral materials. There are 14 steps in the synthesis and the final product 155 contains four chiral centres.37 If we want a single enantiomer, where should we resolve?

462

22 Resolution

O

Cl COCl

Cl

O

O

Cl

Cl

O

O

O RCO3H

Zn

Prins

protect oxidise

Et3N

HO cis-146

cis-147

cis-148

O

OH O

O

O

protect deprotect

149

O

O

O

OH

Wittig

reduce

oxidise

O

C5H11

O

O OSiPh2t-Bu

C5H11

O O

O

O CHO

O

150

151

OH

152

OH

153

OH CO2H

Wittig

CO2H

deprotect

C5H11

O

C5H11 O

O OH

OH

154

155

The obvious answer is as early as possible. The first chiral intermediate is 146 and that already has two chiral centres. The first intermediate with a useful functional group is 149 with its two alcohols. The chemists at what was then Glaxo (now GlaxoSmithKline) chose an ingenious resolution of the stable ketone 147. Because this is a strained ketone it forms a stable adduct with bisulfite and that could be resolved as a salt with our old friend 1-phenylethylamine 2. The bisulfite compound reverts to the ketone 147 on treatment with base and the resolution was complete.

O O O

HO

O

S O 1. crystallise

SO2, H 2O, CH 2Cl2

2. Na 2CO3, H 2O H2N

racemic 147

Ph 2

H2N

Ph

salt of bisulfite adduct and 2

enantiomerically pure 147

More recently enzymes have been used to resolve related intermediates also used in prostaglandin synthesis.38 Acylation with vinyl acetate catalysed by twelve lipases was tried and the best was ‘Amano PS’ from Pseudomonas cepacia. At 55% conversion the alcohol (⫹)156 was obtained in 40% yield and 91% ee and the acetate (⫹)-157 in 78% ee. This low enantiomeric purity could be enhanced to 95% by hydrolysis to (⫺)-156 and reacetylation with the enzyme. Note that the alcohol and acetate of the same series have opposite signs of rotations.

22 References

O H

463

O

O

H

O

O

H

O

O

H

OAc

K2CO3

+ Amano PS

H

H

OH

MeOH

H

OH

H

OAc

(+)-156

(±)-156

O

OH (–)-156

(+)-157

It should not be assumed that all these adventures are successful. The same workers attempted to resolve the first chiral intermediate 159 in the synthesis of 156 by acylation of the related alcohol 160 with the same range of lipases but with only very limited success.39 The ee of 160 was 34–66% with the various enzymes. Neither could 160 be resolved with chiral HPLC. You may see that it is asymmetric in the ring remote from the alcohol. H

H

BaeyerVilliger

O

O

(±)-145

O

H 158

H

O O

HO

O O

H 160

H 159

Returning to the resolution of amines, an enzymatic acylation of racemic amines in aqueous solution by a penicillin acylase (enzymes used in industry in the synthesis of penicillins) from Alicaligenes faecalis has recently been reported.40 The best acylating agent is the amide 161 of phenylacetic acid. There is a big advantage here. Unlike the ester formations and hydrolysis we discussed earlier, no amide exchange occurs with simple amides like 161. The amide (R)-162 was formed in 45% yield and 98.5% ee. Once it has been separated from free (S)-2, it can be hydrolysed to free (R)-2 with the same enzyme! This automatically perfects the ee. penicillin acylase

NH2 + H2N O 161

H N

+ H2N

water, pH 10

Me (±)-2

O Me (R)-162; amide of (R)-2

Me (S)-2

References General references are given on page 893 1. A. Ault, J. Chem. Ed., 1965, 42, 269. 2. B. L. Huckabee, S. Lim, T. L. Smith and T. L. Stuk, Org. Process Res. Dev., 2000, 4, 594. 3. A. Goswami, J. M. Howell, E. Y. Hua, K. D. Mirfakhrae, M. C. Soumeillant, S. Swaminathan, X. Qian, F. A. Quiroz, T. C. Vu, X. Wang, B. Zheng, D. R. Kronenthal and R. N. Patel, Org. Process Res. Dev., 2001, 5, 415. 4. H. Boesch, S. Cesco-Cancian, L. R. Hecker, W. J. Hoekstra, M. Justus, C. A. Maryanoff, L. Scott, R. D. Shah, G. Solms, K. L. Sorgi, S. M. Stefanick, U. Thurnheer, F. J. Villani and D. G. Walker, Org. Process Res. Dev., 2001, 5, 23. 5. J. E. Arrowsmith, S. F. Campbell, P. E. Cross, J. K. Stubbs, R. A. Burges, D. G. Gardiner and K. J. Blackburn, J. Med. Chem., 1986, 29, 1696; S. Goldmann, J. Stoltefuss, and L. Born, J. Med. Chem., 1992, 35, 3341.

464

22 Resolution

6. P. E. Sonnet, P. L. Guss, J. H. Tumlinson, T. P. McGovern, and R. T. Cunningham, ACS Symposium 355 Synthesis and Chemistry of Agrochemicals, eds D. R. Baker, J. G. Fenyes, W. K. Moberg, and B. Cross, ACS, New York, 1987, 388. 7. R. M. Bannister, M. H. Brookes, G. R. Evans, R. B. Katz and N. D. Tyrrell, Org. Process Res. Dev., 2000, 4, 467. 8. M. Williams and G. Quallich, Chem. Ind. (London), 1990, 315; M. Lautens and T. Rovis, J. Org. Chem., 1997, 62, 5246; E. J. Corey and T. G. Gant, Tetrahedron Lett., 1994, 35, 5373. 9. X. Lu, Z. Lu and G. Tang, Org. Process Res. Dev., 2001, 5, 184. 10. J. F. Larrow, E. N. Jacobsen, Y. Gao, Y. Hong, X. Nie and C. M. Zepp, J. Org. Chem., 1994, 59, 1939. 11. M. Prashad, B. Hu, O. Repic, T. J. Blacklock and P. Giannousis, Org. Process Res. Dev., 2000, 4, 55. 12. W. H. Pirkle and J. M. Finn, J. Org. Chem., 1981, 46, 2935; W. H. Pirkle and J. L. Schreiner, J. Org. Chem., 1981, 46, 4988. 13. G. Blaschke, Angew. Chem., Int. Ed., 1980, 19, 13. 14. P. A. Worthington, Pestic. Sci., 1991, 31, 457. 15. D. A. Pflum, H. S. Wilkinson, G. J. Tanoury, D. W. Kessler, H. B. Kraus, C. H. Senanayake and S. A. Wald, Org. Process Res. Dev., 2001, 5, 110. 16. A. Collet, M. Brienne and J. Jacques, Chem. Rev., 1980, 80, 215. 17. L. Pasteur, Ann. Chim. Phys., 1850, 28, 79; G. B. Kaufman and R. D. Myers, J. Chem. Ed., 1975, 52, 777. 18. G. A. Stein, H. A. Bronner, and K. Pfister, J. Am. Chem. Soc., 1955, 77, 700. 19. S. Terashima, K. Achiwa, and S. Yamada, Chem. Pharm. Bull., 1965, 13, 1399. 20. S. Yamada, M. Yamamoto, and I. Chibata, J. Org. Chem., 1973, 38, 4409. 21. G. Coquerel, R. Bouaziz and M. Brienne, Chem. Lett., 1988, 1081. 22. S. Yamada, C. Hongo and I. Chibata, Chem. Ind. (London), 1980, 539; C. Hongo, S. Yamada and `I. Chibata, Bull. Chem. Soc. Jpn., 1981, 54, 3286, 3291. 23. D. F. Reinhold, R. A. Firestone, W. A. Gaines, J. M. Chemerda and M. Sletzinger, J. Org. Chem., 1968, 33, 1209. 24. P. J. Reider, P. Davis, D. L. Hughes and E. J. J. Grabowski, J. Org. Chem., 1987, 52, 955. 25. E. A. Pohland, L. R. Peters, and H. R. Sullivan, J. Org. Chem., 1963, 28, 2483. 26. W. Oppolzer, H. Hauth, P. Pfäffli, and R. Wenger, Helv. Chim. Acta, 1977, 60, 1801. 27. T. Kato, T. Ozaki, K. Tsuzuki and N. Ohi, Org. Process Res. Dev., 2001, 5, 122. 28. “Fluazifop-butyl” Technical Data Sheet, Zeneca Plant Protection, Fernhurst, 1981; D. W. Bewick, Eur. Pat., 133,033 (Chem. Abstr., 1985, 102, 165,249). 29. A. M. Klibanov, Acc. Chem. Res., 1990, 23, 114. 30. K. Laumen and M. P. Schneider, J. Chem. Soc., Chem. Commun., 1988, 598. 31. K. Laumen and M. P. Schneider, J. Chem. Soc., Chem. Commun., 1988, 1459. 32. D. G. Drueckhammer, C. F. Barbas, K. Nozaki, C.-H. Wong, C. Y. Wood and M. A. Ciufolini, J. Org. Chem., 1988, 53, 1607. 33. J. M. Roper and D. P. Bauer, Synthesis, 1983, 1041. 34. I. Kumar, K. Manju and R. S. Jolly, Tetrahedron: Asymmetry, 2001, 12, 1431. 35. G. Fülling and C. J. Sih, J. Am. Chem. Soc., 1987, 109, 2845. 36. M. Schneider, N. Engel and H. Boensmann, Angew. Chem., Int. Ed., 1984, 23, 64. 37. E. W. Collington, C. J. Wallis, and I. Waterhouse, Tetrahedron Lett., 1983, 24, 3125. 38. G. Zanoni, F. Agnelli, A. Meriggi and G. Vidari, Tetrahedron: Asymmetry, 2001, 12, 1779. 39. G. Zanoni and G. Vidari, J. Org. Chem., 1995, 60, 5319. 40. D. T. Guranda, L. M. van Langen, F. van Rantwijk, R. A. Sheldon and V. K. Svedas, Tetrahedron: Asymmetry, 2001, 12, 1645.

23

The Chiral Pool — Asymmetric Synthesis with Natural Products as Starting Materials — Introduction: The Chiral Pool PART I – A SURVEY OF THE CHIRAL POOL The Amino Acids Important reactions of amino acids Reduction of amino acids Hydroxy Acids Amino Alcohols Terpenes Carbohydrates - the Sugars The Alkaloids PART II – ASYMMETRIC SYNTHESES FROM THE CHIRAL POOL Amino Acids The synthesis of captopril The synthesis of ramipril HIV-Protease inhibitors Hydroxy-Acids The synthesis of bestatin The synthesis of (⫹)-laurencin Streptazolin from tartaric acid Amino Alcohols Synthesis of quinolone antibiotics The oxazolidinone antibiotics The synthesis of anti-viral nucleoside analogues The New Chiral Pool Glycidol as a chiral pool member Phenylglycine as a chiral pool member C2 symmetric diamines as chiral pool members Amino-indanols The synthesis of crixivan (Indinavir) Conclusion: Syntheses from the Chiral Pool The synthesis of agelastatin The synthesis of LAF389, a Novartis anti-cancer agent PART III – THE CHIRAL POOL A listing of potential starting materials from the old and new chiral pool

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

466

23 The Chiral Pool

Introduction: The Chiral Pool The chiral pool is that collection of available natural products considered cheap enough to use as starting materials for organic synthesis. The chiral pool strategy is the incorporation of part or all of one of these compounds into the target molecule.1 Chiral pool compounds were used as resolving agents (chapter 22) and derivatives of them will be used as reagents (24), catalysts (25 and 26) and chiral auxiliaries (27) and in this chapter we give the synthesis of many such compounds. There is of course no hard and fast definition of which natural products qualify for the chiral pool but our list appears at the end of the chapter. Most amino acids are cheap enough for anyone, and sucrose is one of the cheapest pure organic compounds anywhere, but if you plan the synthesis of a very active compound, say a prostaglandin, when only a few hundred grams per annum is needed, then a quite expensive compound becomes a suitable starting material. The Syntex headquarters is in Mexico City because R. E. Marker found that the Mexican yam Dioscorea could replace animals as a source of steroids since the vegetable “steroid” diosgenin 1 could be converted into optically active birth control ingredients2 like pregnenolone acetate 2.

H O

O

H O

H

H

H H

H

HO 1; diosgenin

H

H

AcO 2; pregnenolone acetate

We shall introduce the chiral pool in the fi rst half of this chapter, taking each type of natural product, such as the amino acids, in turn and give the key reactions that allow their use in synthesis. We shall illustrate this section with the synthesis of important resolving agents, reagents and catalysts for asymmetric synthesis, and chiral auxiliaries. We shall reveal that there is now what we shall call the ‘new chiral pool’ - compounds so easily made from natural compounds that they are readily (and often commercially) available. Then we shall look at syntheses of more complex targets, showing how the decision was made to use the chiral pool strategy.

PART I – A SURVEY OF THE CHIRAL POOL The Amino Acids The α-amino acids found in proteins are widely available and reasonably priced - many indeed are cheap.3 They have the general structure 3, where R can be alkyl 4, 5, cycloalkyl 6 and functionalised alkyl 7 or aryl. They are all L, most are also (S) (all except cysteine and cystine), and some are (⫹) and some (⫺) as you would expect. Some are also available as the D enantiomer, usually more expensive, but the synthesis4 of D amino acids (e.g. by enzymatic resolution, chapter 29) is making them cheaper. These examples 4–7 are very important.

23 The Amino Acids

O

O

R

OH

Ph

O

O OH

OH

NH2

NH2 3; (L)-amino acids

467

NH2

4; (S)-(–)-Phe phenylalanine

5; (S)-(+)-Val valine

CO2H

N H

HO2C

OH NH2

6; (S)-(–)-Pro proline

7; (S)-(+)-Glu glutamic acid

If you are manufacturing aspartame 8, it is pretty obvious that you will use the natural amino acids Asp 9 and Phe 4, modified only by turning Phe into its methyl ester 10. Standard peptide coupling gives the artificial sweetener. However, to make more interesting compounds, the natural amino acids must be transformed by stereospecific reactions before they are incorporated. Ph

O HO2C

O

Ph

C–N

N H

OH

HO2C

CO2Me

amide

NH2 9; Asp (S)-(+)-aspartic acid

NH2 8; Aspartame (Nutrasweet)

+ H2N

CO2Me

10; (S)-(+)-Phe-OMe phenylalanine methyl ester

Important reactions of amino acids Diazotisation of amino acids gives diazonium salts such as 11 derived from valine 5. This reaction obviously goes with retention as the chiral centre is not affected. The internal CO2H group is the best nucleophile and displaces N2⫹ with inversion. Any nucleophile now opens the α-lactone 12 by SN2 displacement at the alkyl centre, again with inversion to give 13 with overall retention from the original amino acid.5 Most lactones react with nucleophiles at the carbonyl group but α-lactones react at saturated carbon as that relieves the strain of the 3-membered ring. O

O

O SN2

1. AcOH

OH

OH 2. NaNO 2

NH2 5; (S)-(+)-Val

N2 11

intramolecular

O

AcO

SN2

OH O 12; α-lactone

AcO

OAc 13; 98% yield

The same reaction works well with chloride as nucleophile to give 14 and, after reduction of the acid group and cyclisation of the alcohol 15, the epoxide 16. Such epoxides, made from many of the amino acids, now count as members of the new chiral pool. The diazonium salt 11 and the chloroacid 14 are susceptible to racemisation by enolisation and it is only after two distillations that the epoxide 16 has 97% ee. There are very detailed Organic Syntheses procedures for both these steps.6

O

O OH NH2 5; (S)-(+)-Val

NaNO2 HCl

OH Cl 14; 58-65% yield

LiAlH4

OH Cl 15; 70% yield

KOH

O 16; 87% yield 97% ee

468

23 The Chiral Pool

Glutamic acid 7 has its own internal nucleophile: the diazonium salt initially cyclises to the α-lactone 17 which is then opened by the other CO2H group to give the γ-lactone 18. Selective reduction of either the lactone or the acid gives two useful intermediates7 19 and 21.

CO2H

HO2C

O

O 17

HO

H

19

(S)-(+)-18

CO2H

O

CO2H H (S)-(+)-18 O

B2H6

DIBAL

CO2H H HO

O

HO

NH2 7; ( S)-L-(+)-Glu

O

O

HONO

OH

O

O

H

20

H 21

This type of reaction can be used to invert the natural series to give a rare and expensive (R)-amino acid. Phenylalanine 4 is converted into the hydroxyacid 22 with retention. Displacement of triflate from 24 with inversion and removal of the benzyl ester gives Boc-protected (R)-Phe 26. In fact this method was used to make 15N-(R)-Phe so the nitrogen nucleophile was labelled but it is the inversion that matters to us.8

O

O

Ph

OH

Ph 2. NaNO 2

O Ph

Boc

Boc

OTf

BuLi

24; 98% yield

–78 ˚C

lutidine

OH

22; 70% yield

23; 98% yield

O

O

H N

OBn

OBn

DMF

OH

4; (S)-(–)-Phe

Tf2O

Ph

OH

NH2

O

BnBr C2CO3

1. H 2SO4

Ph

OBn NBoc2 25; 98% yield

1. TFA

Ph

OH NHBoc

2. Pd/H 2

(R)-26; 88% yield

Reduction of amino acids Reduction of the carboxyl group of the amino acids gives a range of amino alcohols, usually named after the parent acid such as valinol 27 and prolinol 28. It is important that the reduction does not threaten racemisation. Various methods have been used such as NaBH4 in acid solution9 (no doubt producing borane in situ) and LiBH4 with Me3SiCl.10

O [H]

OH NH2 5; (S)-(+)-Val

[H]

OH NH2 27; (S)-valinol

CO2H

N H 6; (S)-(–)-Pro

OH N H 28; (S)-prolinol

23 Hydroxy Acids

469

Valinol 27 and phenylalaninol 29 are used to make the Evans chiral auxiliaries used in asymmetric aldol reactions (chapter 27) and Evans prefers reduction with borane itself as its complex with Me2S. The phenylalanine based auxiliary 30 is generally preferred as the compounds are more likely to be crystalline and can easily be made11 on a 150 g scale. O 1. BF 3.Et2O

H2N

OH

H2N

(EtO)2CO

2. BH 3.Me2S

CO2H Ph

3. NaOH, H 2O

4; (S)-(–)-Phe

HN

O

K2CO3

Ph

Ph 30; 78-79% yield

29; 73-75% yield

Prolinol methyl ether 31 is used in Enders’s SAMP and RAMP chiral auxiliaries (chapter 27) and (S)-(⫺)-SAMP 32 can be made in 50–58% overall yield from (S)-L-(⫺)-proline on a 75 g scale.12

OMe CO2H

N

OMe

N

N

H

H 6; ( S)-(–)-Pro

NH2 32; ( S)-(–)-SAMP

31

Hydroxy Acids A few hydroxy acids are very cheap. (S)-(⫹)-Lactic acid 33 occurs in milk, (R)-(⫺)-malic acid in apples, and both enantiomers of the very important tartaric acid 35 and 36 are reasonably cheap. (⫹)-Tartaric acid 35 is usually called “natural” as it occurs in grapes, but (⫺)-tartaric acid 36 is natural too as it occurs in the West African tree Bankinia reticulata. The other enantiomers of malic and mandelic acids are commercially available and are not that much more expensive. H

OH

HO

H

HO

CO2H

CO2H

OH

HO2C CO2H 34; (–)-(R) -malic acid

33; (+)-(S) -lactic acid

HO

CO2H

HO2C

(R,R)-(+)-35 tartaric acid "natural"

OH

OH

CO2H

Ph

CO2H

(S,S)-(–)-36 tartaric acid "unnatural"

37; (R)-(–)mandelic acid

One of the most important specific applications of tartaric acid is the preparation of Seebach’s TADDOLs, e.g. 40. The dimethyl ester 38 is protected as the acetal 39 and reacted with four molecules of an aryl Grignard to give the TADDOL13 40. All these compounds are C2 symmetric and various TADDOLs have found applications as resolving agents, NMR additives, asymmetric catalysts and so on.14 Some of these will feature later in the book.

O

O

OH

O

O

OMe

MeO

BF3•OEt2

OH

O

38; dimethyl tartrate

MeO2C

CO2Me

39; 77% yield

O PhMgBr

O

Ph Ph

Ph Ph OH HO 40; TADDOL

470

23 The Chiral Pool

Both enantiomers of a member of the new chiral pool, propylene oxide 44, can be made from lactic acid 33. The idea is to reduce the acid and cyclise in two different ways. One is simple enough. Ethyl lactate 41 is mesylated, to turn the secondary alcohol into a leaving group 42, and then the ester is reduced. Cyclisation uses the primary alcohol of 43 as the nucleophile in an internal SN2 reaction so that inversion gives (R)-propylene oxide15 44. H

OH

MsCl

OMs

H

Et3N PhMe

CO2Et 41; (+)-(S)ethyl lactate

AlH3, THF [LiAlH4 + H2SO4]

CO2Et

OMs

H

H

KOH

OH

(S)-43 not isolated

(–)-(S)-42 98% yield

O

H2O

(+)-(R)-44 72% from 42ß

The second sequence starts with reduction of the ester to give the diol 45 with no change in stereochemistry. Reaction with HBr and HOAc gives an 89% yield of a mixture (94:6) of 46 and 47 which sounds like disaster. Not so as 46 and 47 react in different ways. H

LiAlH4

OH

HBr

H

OH

41

OAc Br +

H

Br

H

OAc

ROH

HOAc

THF

O

RO K

(S)-46 94 parts

(+)-(S)-45 81% yield

(–)-(S)-44 54% from 41

(R)-47 6 parts

The HBr/HOAc reaction goes via the cation 48 that reacts either with or without inversion to give 46 or 47. Treatment of the mixture with RO⫺ cleaves the ester to release the oxyanions 48 and 49 which cyclise to 44 without inversion 48 or with inversion 49. The molecule either undergoes no inversion (via 46) or two inversions (via 47) and the result is retention.16 Me H

O 47 inversion

Br

Br

O

O

46 no inversion

Br Br

O

49 no inversion

48

50 inversion

Malic acid 34 is useful for the simple functionalised skeleton it contains and so it is more likely to be incorporated into the target molecule than form part of a reagent. It will feature more in the second half of the chapter.

Amino Alcohols The amino alcohols 51 derived from amino acids have been mentioned already. Amino alcohols available directly from nature are the ephedrine family: ephedrine 52, its anti-diastereoisomer pseudoephedrine 53 and norephedrine 54 lacking the N-methyl group. The enantiomers of these compounds are also available: (⫺)-52 is about twice as expensive, (⫺)-53 slightly (about 25%) more expensive, and (⫺)-54 about the same price. OH R

S

S OH NH2

51; amino-alcohols from amino-acids

OH S

R

Ph

OH

NHMe 52; (1 S, 2 R)(+)-ephedrine

S

S

Ph

R

Ph NHMe

53; (1 S, 2S)-(+)pseudoephedrine

NH2 54; (1 S, 2 R)-(+)norephedrine

23 Amino Alcohols

471

Derivatives of these aminoalcohols are used as bases for asymmetric deprotonation. There is more about this in the next chapter (asymmetric reagents) but we should note here that amino alcohols such as 55 and 58 or diamines such as 56 and 57 have been used successfully. Some, such as 55, are derived from amino acids, others, such as 58, are derived from the ephedrine family.17 H N

H Ph N H

R

LiN

MeN

Ph R

OLi 55

Ph

N

N H 56

OLi Ph NR2

57

58

One useful reaction is the desymmetrisation of 4-substituted cyclohexanones 59 to give the silyl enol ethers 60 using the chiral base 57. The chirality can be made more permanent by oxidative cleavage of the alkene to give the dicarboxylic acid 61, or by Pd(II) oxidation (chapter 33) to the enone 62. R

R

R

1. 57

MoO2 (acac)2

2, Me 3SiCl

t-BuOOH

HO2C

CO2H

OSiMe3 60

O 59

R

O 62

61

A dramatic application was the asymmetric synthesis of epibatidine 70 by Simpkins.18 DielsAlder reaction of the deactivated pyrrole 63 with the alkynyl sulfone 64 gave the bicyclic core 65 of epibatidine. Selective reduction gave the compound 66 needed for epibatidine, but in racemic form. A directed lithiation (chapter 7) and sulfonation led to achiral bis sulfone 67. Boc

H

N NBoc +

63

H2, Pd/C

1. BuLi

MeCN

2. ArSO 2F

SO2Ar 65

64

N

N

85–90 ˚C

SO2Ar

Boc

Boc

SO2Ar SO2Ar 67

SO2Ar (±)-66

Catalytic reduction from the exo-face gave achiral 68 ready for desymmetrisation. Elimination with the chiral base (simply the sodium salt 69 of pseudoephedrine 53) gave a single enantiomer of 66 ready for conjugate addition of the pyridine and conversion into epibatidine 70. Boc H2 Pd(OH)2/C 800psi

67 CH2Cl2 EtOAc

Boc Ph

N H H SO2Ar SO2Ar 68

Me

NaO

N

Cl

H N N

NHMe 69

SO2Ar 66

70; epibatidine

The same type of amino alcohol is used as ligand for various organo-lithium additions. N-Dialkylated norephedrines 58, made by dialkylation of ephedrines,19 are effective in catalysing the addition of lithium acetylides to carbonyl compounds.20 This was particularly useful in Merck’s

472

23 The Chiral Pool

preparation of the anti-HIV drug efavirenz.21 The ketone 71 was combined with the lithium derivative 74 and the lithium acetylide 72 (both made with BuLi) to give the adduct 73 in 96–98% ee improved to ⬎99.5% by crystallisation. OLi

CF3

CF3

74

Cl

Cl

O

OH

N

Li

N Ar H 71; Ar = p-methoxy-C 6H4

72

N Ar H 73: 80% yield, 96-98% e.e.

74

Terpenes The molecules we have seen so far have usually been incorporated into the target molecule complete. There are two further and most important groups of larger molecules, the sugars and the terpenes, from which pieces are usually snipped out for incorporation. The simple monoterpenes (C10 compounds) citronellol 75, citronellal 76, and citronellic acid 77 are good examples. They are not cheap but both enantiomers are available, not always in excellent ee. CHO

OH 75; (R)-(+)-citronellol

CO2H

76; (R)-(+)-citronellal

77; (R)-(+)-citronellic acid

These compounds are often used by oxidative cleavage of the alkene to give a C7 unit with functionality at both ends, e.g. 78 from citronellic acid. Notice how the functional groups are carefully made different so that one end can be reacted selectively. This compound 78 was used by Mori in natural product synthesis.22 H CO2H

1. O 3, MeOH 2. Me 2S 3. TsOH, MeOH

77; ( R)-(+)-citronellic acid

H MeO

CO2H OMe

78

The cyclic monoterpenes are also very useful. Menthol 79 is very cheap and the ketones pulegone 81 and carvone 82 are moderately cheap. All are available as the other enantiomer, e.g. 80. An important application is as a chiral auxiliary, the favourite being 8-phenylmenthol 83 made from pulegone,23 see chapter 30. O 5

2

1 OH

79; (1R, 2S, 5R) -(–)-menthol

OH 80; (1S, 2R, 5S) -(+)-menthol

O 81; (R)-(+)pulegone

H 82; (R)-(–)carvone

Ph

OH

83; (–)-8phenylmenthol

Pulegone has also been used to make the chiral auxiliary 86 that acts as an asymmetric d1 reagent. Conjugate addition of the thiol PhCH2SH gives a diastereomeric mixture of ketones 84 from which the single compound 85, having all substituents equatorial, can be made by reduction.

23 Carbohydrates - the Sugars

473

Formaldehyde introduces one carbon that is the d1 centre in bicyclic 86. A carbonyl group 87 is added by lithiation and reaction with an aldehyde followed by Swern oxidation. Nucleophiles add to this ketone from the front, i.e. opposite the sulfur rather than the oxygen atom, to give, after removal of the auxiliary, single enantiomers of the alcohols24 88.

Ph

O 81; ( R)-(+)-pulegone

SH

S

NaOH THF reflux

Na, NH 3(l)

Ph

MeOH –78 ˚C

O 84; 90% yield, mixture 1. BuLi

S

(CH2O)n

O

cat. TsOH benzene reflux

R1

S

2. R 1CHO

O

3. DMSO (CF3CO)2O

OH 85; >80% one isomer O

1. R 2MgBr 2. NCS AgNO3

O

86; 86-89% yield

SH

R1

H

R2

HO

87

88

There are bicyclic monoterpenes too - α-pinene 89 and β-pinene 91 share a common skeleton with four- and six-membered rings but have the alkene in different places. There is a discussion in chapter 24 on the variable ee of α-pinene and it is better to make the (⫺)-enantiomer from the more reliable β-pinene 91 (99% ee) that can be isomerised with strong base (‘KAPA’) 92 in 93% yield to (⫺)-90 without loss of ee. Many asymmetric reagents for reduction (chapter 24) and chiral auxiliaries for asymmetric aldol reactions (chapters 27 and 30) are based on α-pinene.25 KAPA

NH K

93% yield

NH2 89; (1R)-(+)-α-pinene

(–)-90

91; (–)-β-pinene

92; KAPA

Camphor 93, a bridged monoterpene with one six- and two five-membered rings is available in both enantiomeric forms and was used by Woodward in the synthesis of vitamin B12. Sulfonation on camphor occurs on the bridgehead methyl group by a series of rearrangements described in the workbook. You will see in chapter 27 how Oppolzer’s sultam 95 is used as a chiral auxiliary and in chapter 33 how oxaziridines such as 96 are used in asymmetric oxidation. Both are made from camphor-10-sulfonic acid 94.

R

R

H2SO4

NH R O

HO3S

S

O

S O

93; (+)-camphor

94; (+)-camphor10-sulfonic acid

O

95; Oppolzer's chiral sultam

N S

O

O

O 96; Davis's asymmetric oxidant

Carbohydrates - the Sugars The sugars are generally both cheap and optically pure, and more compounds have been made from glucose 97 than any other member of the chiral pool. It is rare that a target molecule looks

474

23 The Chiral Pool

much like glucose, but by keeping in mind that glucose can exist in three forms, the normal pyranose 97, the open chain 99, and the furanose 101, resemblances are easier to find. Each form may be trapped by suitable acetal formation: benzaldehyde prefers the pyranose form with all substituents (including the Ph group) equatorial except the anomeric OMe group, which prefers to be axial 98. A thiol traps the free aldehyde 99 as the bis-alkylthioacetal 100. Acetone prefers five-membered acetals from cis substituents, and so traps the furanose form 102. HO HO HO

97; α-D-glucose pyranose form

O

OH

OH

OH

HO

OH

100

SR

O O

HO

SR

HO

CHO

OMe

OH

RSH, H

OH

101; α-D-glucose furanose form

98 HO

OH

99; α-D-glucose open chain form HO

O

O HO

2. PhCHO, H

OH OH OH

O

Ph

1. MeOH, H

acetone, H

OH

O

O

H

HO

HO

OH

102

O

H

H

O

H

Mannose 103 resembles glucose closely except for the axial OH at C-2. This is a significant difference as the open chain form 104 makes clear: it is nearly C2 symmetric and, on reduction, becomes so in the form of mannitol 105. HO HO HO

OH O

OH HO

2

OH 103; mannose

OH

OH CHO

NaBH4

OH OH

HO OH OH 105; mannitol

OH OH 104; open-chain mannose

If mannitol is protected as the bis-acetal 106 (acetone prefers to form two acetals without stereochemistry rather than a single acetal in the middle - acetal formation is under thermodynamic control) and cleaved oxidatively, both halves give the same product - a protected form of the unstable D-glyceraldehyde26 107. O OH 105 mannitol

ZnCl2 acetone 20 ˚C

O O

O O

OH

106; bis-acetal of mannitol

Pb(OAc) 4

H

O O

or NaIO4

107; acetal of glyceraldehyde

An alternative sequence that gives 100% yield uses silyl and benzyl protecting groups 109. This form of protected glyceraldehyde reacts with allyl silane to give 110 with good stereoselectivity that is dependent on the Lewis acid used as catalyst.27 The products have been used in the synthesis of the immunosuppressant FK506 useful in transplant surgery.28

23 The Alkaloids

OH

475

O

OBn Pb(OAc)4

OTBDMS

TBDMSO

TBDMSO

H

OBn 109; 100% yield

OBn OH 108; protected mannitol

OBn

OBn BF3•Et2O

TiCl4

TBDMSO OH 110; syn:anti >92:2 >98% yeild

TBDMSO

109 SiMe3

SiMe3

OH 110; syn:anti 19:81 90% yield

Glyceraldehyde, of which 107 and 109 are protected forms, is the only three-carbon sugar. There are two four-carbon sugars: erythrose 111 and threose 113. Both exist in hemiacetal (furanose) and open chain forms and both are chiral but their symmetry properties differ. The tetrols formed by reduction of the aldehydes are meso erythritol 112 and C2 symmetric threitol 114 having the same stereochemistry as tartaric acid 35. Threose or threitol can be oxidised to tartaric acid. These sugars are the origin of the terms ‘erythro’ and ‘threo’ sometimes used to describe such diastereomeric relationships. O

OH

OH

OH H

HO HO OH 111; erythrose O

NaBH4

OH

HO

OH O 111; erythrose

OH 112; erythritol

OH

OH HO

HO OH 113; threose

OH H

NaBH4

OH

HO OH O 113; threose

OH 114; threitol

Threitol 114 has been used in many asymmetric syntheses, for example, the preparation of the C2 symmetric bis-epoxide 117 by protection of the secondary and activation of the primary alcohols. Cyclisation of 116 could in principle give two fused four-membered rings, but threemembered rings are greatly preferred kinetically. The bis-epoxide 117 reacts with nucleophiles at the terminal carbons.29 OH 114 threitol

O

1. MsCl pyridine

O

HO

OH 115

2. MsOH EtOH, H 2O

H OMs

MsO OH 116

O

KOH EtOH H2O

O

H 117

The Alkaloids These nitrogen-containing natural products, often with powerful biological properties, are not usually incorporated into target molecules. However, they are important in asymmetric syntheses as the foundations of many reagents and catalysts. Quinine 119 is familiar as an anti-malarial and an ingredient in tonic water. Quinine and its twin cinchona alkaloid quinidine 118 are referred to as pseudo enantiomers. Each occurs naturally as one enantiomer only but the two structures are nearly enantiomeric: only the vinyl side-chains disturb the symmetry and they act as enantiomers. The vinyl side-chains are reduced and two molecules of, say, dihydroquinine (DHQ) are joined

476

23 The Chiral Pool

by a spacer such as phthalazine (PHAL) to form the catalyst (DHQ)2-PHAL 120 that accelerates dihydroxylation of alkenes by OsO4.

N

N OH

N

N

HO

N

N

O

O

OMe MeO

MeO

OMe

N 119; quinine

N 118; quinidine

N

N 120; (DHQ) 2-PHAL

Other alkaloids such as brucine were used as resolving agents but their expense and extreme toxicity has made them unpopular and the availability of a range of amines (chapter 22), often made from the chiral pool, has made them unnecessary.

PART II – ASYMMETRIC SYNTHESES FROM THE CHIRAL POOL Here we discuss syntheses of target molecules based on the chiral pool strategy using the same sequence of natural products as part 1. This time we discuss why the chiral pool strategy was chosen and try to demonstrate the particular value of each type of compound.

Amino Acids With some targets the structure of an amino acid is immediately obvious. The thyroid hormone thyroxine 121 present in the thyroid gland contains the skeleton of the proteinaceous amino acid tyrosine 122 and can indeed be made from it. I HO

I

CO2H

CO2H HN

I

NH2

O

NH2

HO

I 121; thyroxine

CO2H 122; tyrosine

6; proline

We have already met proline 6 and its structure is clearly contained in the Bristol-Meyers Squibb ACE inhibitor captopril 123. In other cases the relationship is not so obvious. Other ACE inhibitors, also used to treat high blood pressure, such as ramipril 124 and fosinopril 125, also contain the elements of proline but it is not obvious how we should use that observation.

RO2C

N SH

O

N H

CO2H Ph

123; captopril

H

H N

O

124; ramipril

OR

N

P CO2H

Ph

O

O

125; fosinopril

CO2H

23 Amino Acids

477

The synthesis of captopril The first disconnection is easy 123a: it is in the middle of the molecule, at the amide bond and reveals the molecule of proline 6. The thiol acid 126 has a 1,3-relationship between SH and C⫽O and can be made by conjugate addition.

C–N

N

HS

HN

+ HS

amide

O

1

3

CO2H

HS

+ O 127

126

6; proline

OH

1,3-diX

O

CO2H

123a; captopril

C–S

OEt

2

Thiolacetic acid 126 is used as a reagent for nucleophilic SH (it is a carbonyl compound and the S atom is the nucleophile) and does the conjugate addition well. The racemic product 127 is coupled to protected proline 128 and the t-butyl ester hydrolysed to give a mixture of diastereoisomers of 129. The salts of these can be separated and, after cleavage of the thiolester, the right isomer gives captopril. Proline acts as a resolving agent as well as providing half of the target molecule. O

127

SH

AcS

CO2H +

132

AcS

N

1. 134 2. separate

2. TFA

O 131

t-BuO2C 130

129

128

1. 133 (DCC)

HN

CO2H

123

3. NH 3 MeOH

Three reagents deserve some comment. Quinol 132 is easily oxidised to quinone and protects the thiolacid 128 against disulfide formation. DCC 133 is a standard peptide coupling agent uniting free NH2 and CO2H groups by removing water to give dicyclohexylurea. The simple amine 134 is sufficient for separation by crystallisation during the resolution as the resolving agent (proline) is already present in the molecule and a crystalline salt is all that is needed. Making both diastereoisomers separately (each as a single enantiomer of course) established that 123 was several orders of magnitude more active than the other isomer.30 OH

O

HO

N

O 132

quinone

C

N H 134

N

133: DiCyclohexylCarbodiimide

The synthesis of ramipril Ramipril 124, the Hoechst ACE inhibitor, is obviously more complicated. The first amide disconnection is the same and we can see the structure of alanine in the acid 135. The amine 136 looks like proline but no reactions spring to mind to make the bonds corresponding to the disconnections. proline

alanine RO2C

H N H

Ph

RO2C

H C–N

N

N H

amide

O

124a; ramipril

H

CO2H

Ph 135

OH O

+

H ?

HN

HN CO2H 136

CO2H 6

478

23 The Chiral Pool

In practice 136 was not made from proline. Breaking open the other, more functionalised, ring 136a is possible with the idea of a radical cyclisation onto an alkene 137. FGI of hydroxyl for iodine reveals the elements of another amino acid, serine 140 and an allylic bromide 139.

H

radical cyclisation

H

H

H

FGI

HN

HN

I

OH

Br

allylation

H2N

OH CO2H

CO2H 138

CO2H 137

CO2H 136a

C–N

HN

139

140; serine

The key step in the synthesis was the radical cyclisation of 141 that gave an excellent yield of two diastereoisomers: both had cis ring junctions and the compounds could be separated after conversion to the bis benzyl esters 143 and 144. Hydrogenation of 143 released 136. Once again the chiral pool material (serine) acts both as starting material and resolving agent though here there was some selectivity (1.25:1) in favour of the right diastereoisomer31 143.

Bu3SnH

H Z

AIBN

N

Z

I

H

H

PhCH2OH

N

78% yield

CO2Me

H

H

(i-PrO)4Ti

N

Z

CO2Me

141; Z = CO 2Bn

H

H

+

Z

N

CO2Bn

142; 88% yield

143

CO2Bn 144

1.25:1

An ester of 136 is now used as resolving agent in coupling to the rest of the ramipril molecule. Conjugate addition of alanine to the ketoester 145 gives a 2:1 mixture of 146 and its diastereoisomer. Coupling with 136 allows separation of the right compound and hence completes the synthesis of ramipril. EtO

O

EtO2C

1. EtOH, Et 3N

H2N

Ph

O 145

N H

CO2Bn

2. Pd/C, H 2, AcOH

Ph

OH 1. 136 benzyl ester K2CO3, (EtMePO) 2

O

124 2. H 2, Pd/C, MeOH

146; 2:1 diastereomers

The losses in yield when these semi-resolutions occur are not ideal and some ACE inhibitors are made with much better selectivity. Enalapril 147, Merck’s Vasotec, is easily disconnected to three simple pieces, proline 6, alanine 149, also part of ramipril, see 135, and the keto-acid 148, assuming we can make the amine by reductive amination. EtO2C Ph

EtO2C N H

2 x C–N

N

O 147; enalapril

CO2H

amide amine

Ph

O 148

+

H2N

OH + HN

O 149; alanine

CO2H 6; proline

23 Amino Acids

479

Normal peptide coupling (N-Boc-Ala with proline benzyl ester coupled with DCC and removal of the protecting groups) gives the dipeptide 150. Formation of the imine 151 and reductive amination gives a 62:38 selectivity in favour of 147 providing that the reducing agent is hydrogen with Pd/C. Sodium cyanoborohydride gives almost 50:50. The maleate salt of 147 was separated by crystallisation.32 EtO2C N

155

CO2H O 150; L-ala-L-pro

EtOH

H2N

H2, Pd/C

N

N

Ph

147 enalapril

EtOH

CO2H O 151; E/Z-mixture of imines

62:38 diastereomers

HIV-protease inhibitors The HIV-protease inhibitors, probably the most successful drugs used to prevent HIV turning into AIDS, are all protein mimics with an unreactive C⫺C bond replacing an amide link. That bond is marked with a broad arrow in norvir 152, Bristol-Myers Squibb’s entry in this class. Amide disconnection reveals that the key central part of the molecule 154 looks like a phenylalanine dipeptide except that it has the C⫺C bond. So can we make 154 from phenylalanine 4? N

H N

O

HO

Ph

O N H

S O

amides

Ph O

H2N

X

O

NH2

S 153

O

152 norvir HIV-protease inhibitor

N N

OH

N

S

O

Ph 2 x C–N

Me

H N

Me

H N

X

S

N N O

Ph

154

155

Phenylalanine 4 is protected with benzyl groups 156 and combined with the anion of acetonitrile to give 157. Benzyl Grignard attacks the nitrile to give the enaminone 158 which is immediately reduced to the boron complex 159 that shows remarkably high 1,4-stereoselectivity (95:5). Still without isolation, reductive cleavage of 159 gives 160, a protected version of 154, again with excellent stereoselectivity.33 The deprotected intermediate 154 is isolated in 60% yield from the enaminone 158 after debenzylation by Pd-catalysed transfer hydrogenation with ammonium formate. O H2N

O OH

BnCl

Bn2N

O

K2CO3

Ph 4; phenylalanine

Bn2N

MeCN

OBn

CN

Ph

MgCl THF

NaNH2

Ph 156; 94% yield

Ph 157; 78% yield OR

O

NH2

O 2.5 NaBH4

Bn2N Ph 158

Ph

MsOH THF, i-PrOH

Bn2N

B

NH

OH NaBH3(OR)

1 4

NH2

Bn2N

R = CF3CO

Ph

Ph 159: 95:5 diastereoselectivity

Ph Ph 160; 93% of mixture

480

23 The Chiral Pool

Hydroxy-Acids The synthesis of bestatin The potent enzyme (aminopeptidase) inhibitor bestatin 161 provides another perfect bridge between the amino- and the hydroxy-acids. At first sight bestatin appears to be a dipeptide but only one of the components, leucine 163 is a normal α-amino acid. The other half 162 is a β-amino acid and also an α-hydroxyacid and that might give us the idea that malic acid 34, with the same absolute configuration at the secondary alcohol, could provide a chiral pool starting material. Ph

Ph

O

O

H2N

N H

H2N

CO2H

O OH

H2N

OH

OH 161; bestatin

HO

+

162

OH

CO2H

O

OH

34; malic acid

163; leucine

First the benzyl group must be added. The di-lithium derivative of diethyl malate 164 is alkylated on the opposite side to the OLi group (chapter 30). Now the two ester groups must be distinguished. Reaction of the free diacid with trifluoroacetic anhydride and then with ethanol gives 166. Presumably the two acids form a cyclic anhydride and the free OH forms an ester. Reaction with ethanol occurs at the more reactive (next to OCOCF3) and less hindered (not next to benzyl) carbonyl group of the anhydride and the trifluoroacetyl groups fall off during work-up.

Ph

O EtO

OEt

O OH 164; diethyl malate

1. 2 x LiHMDS 2. PhCH 2Br

Ph

O

EtO

OEt O

OH 165

O

HO

1. NaOH 2. (CF 3CO)2O

OEt O

3. EtOH

OH 166

The free acid group is converted into an amine with diphenylphosphoryl azide 167. A Curtius rearrangement with retention at the migrating group gives the isocyanate 168 which is captured by the OH group to give the cyclic carbamate 169. Peptide coupling with leucine methyl ester using a water soluble carbodiimide (EDC), N-methyl morpholine (NMM) as base, and hydroxybenzotriazole (HOBt) as catalyst gives a slightly disguised form of bestatin 161 with all the stereochemistry correct.34

Ph

Ph

Ph

O

O (PhO)2P

N3

167 diphenyl phosphoryl azide

CO2Et

166

N C O

CO2Et HN O

OH 168

O

169

1. LiOH 2. 170-OMe EDC, NMM HOBt

HN O O

N H 170

CO2Me

23 Hydroxy-Acids

481

The synthesis of (⫹)-laurencin Some molecules that can be made from the chiral pool give little hint of that when you first inspect their structures. It is only after the disconnection process had started that smaller fragments reveal a possible natural starting material. Laurencin 171 reveals its marine origin by the bromine atom and the eight-membered ring that is the heart of the molecule. Trimming off the side chains and writing a lactone 172 having the right sort of functional groups in the right places makes the problem much simpler and opening the lactone reveals a relatively simple open chain intermediate 173. HO

FGI ?

H

H HO

Br

C–O lactone

O

O

HO

OH

O

OAc 171; (+)-laurencin

H

HO2C

OH

172

173

Since there is a cis alkene in the middle of 173 a Wittig disconnection 173a makes sense and the starting materials 174 and 175 turn out to be four-carbon units with functionality, mostly in the form of oxygen atoms, at C-1, C-2 and C-4. This is job description for malic acid and both compounds can indeed be made from (R)-(⫹)-malic acid35 34. OH CHO

HO2C

OH

Wittig

OH

PPh3 OH

Wittig

OH

HO2C

174

OH 175

173a

A protected version of the aldehyde 174 comes from acetal formation and reduction of the remaining CO2H group. Acid catalysed hydrolysis of the acetal gives the lactone 177 that is protected as 178. Hydrolysis and oxidation give 179 ready for the Wittig reaction. MeO

OH

OMe

O

CO2H

HO2C

1. BH3

CO2H

O

O 176; 75-85% yield O

O

O

2. H+

TsOH

34; (R)-(+)-malic acid

3. Ph

OH

O

177

OBn

Cl

O

1. EtOH, K2CO3

O

2. Swern

i-Pr2NEt

EtO2C

O 178; 75% yield

OBn CHO

179; 58% yield

The phosphonium salt 175 requires similar operations but malic acid 34 is reduced to the triol 180 before selective protection 181 leaves just one alcohol for conversion to tosylate, bromide and finally protected phosphonium salt 182.

34

OH

BH3.SMe2 (MeO)3B

OH

HO 180; 98% yield

1. TsCl, pyr Me2C=O

O

O

OH

2. NaBr 3. Ph 3P, LiI, melt

TsOH

181; 82% yield

O

O PPh3 182; 100% yield

482

23 The Chiral Pool

The Wittig reaction between the ylid formed from 182 with BuLi and the protected aldehyde 179 does indeed give the Z-alkene 183, a protected version of 175, ready for cyclisation and completion of the synthesis of laurencin. Notice how far apart (1,6-related) are the chiral centres in 183. Making both independently from malic acid ensures that the right diastereoisomer, as well as the right enantiomer, results.

O

2. 185

Ph3P

O

1. BuLi, THF

O

HO2C

182

OBn

O

O

Z-183; 73% yield

Streptazolin from tartaric acid Streptazolin 184, an antibiotic from a Streptomyces species, has a cluster of three rings, an awkward exocyclic alkene with geometry and three neighbouring chiral centres. There is no obvious chiral pool starting material. Disconnecting the lactone is obvious and we might be able to control the alkene geometry by the kind of Pd-catalysed coupling we met in chapter 18. This gives 185 from which we can remove the carbamate. The aldehyde can be reconnected to the nitrogen atom to give 186 - atoms 1 and 5 have been joined as an amide. Now the possibility emerges of making streptazolin from tartaric acid36 35. Me 2

OH N

H

O

O

Pd-diene synthesis

N EtO2C

H

184

3

4

OBn

CHO 5

H

1

2

1 C–O lactone add protection

H

OBn

reconnect 1-5

N

FGI reduce

?

3

O

185

3

OH

HO2C

4

OH

OBn

2

5 4

HO2C

5

OBn

186

(R,R)-(+)-35 tartaric acid

To get back to tartaric acid from 186 we must disconnect as in 186a and the obvious synthon is 187 with the proviso that the nucleophilic end must control the geometry of the alkene. A vinyl silane is a good answer (chapter 18) with some leaving group at the other end 188 together with the imide 189 obviously derived from tartaric acid 35. O

Me3Si H N

+

OBn

OBn

HN O

O 186a

OBn

X synthon 187

HO reagent 188

OBn 189

HO2C

OH

HO2C

OH

(R,R)-(+)-35 tartaric acid

The imide 189 is easily made from protected tartaric acid 190 and the vinyl silane 188 coupled in a Mitsunobu reaction. Reduction of one of the carbonyl groups of the imide 191 may seem tricky but the molecule is C2 symmetric so the two carbonyl groups are the same and once one is reduced the molecule is a much less reactive amide. The stereochemistry of the acetate in 192 does not matter as it disappears in the cyclisation to 186. Notice that the alkene geometry is retained.

23 Hydroxy-Acids

483

Me3Si

Me3Si EtO2C

OBn NH3

189 EtO2C

OBn

188

1. NaBH 4

O

DEAD Ph3P

OBn

N

BF3•Et2O

OAc

2. Ac 2O DMAP

186 99% yield

OBn

N

190 O

O

OBn 192; 93% yield

OBn 191; 77% yield

The next few steps including the vital palladium-catalysed cyclisation revealed a problem in this synthesis and it is continued in the workbook. An improved synthesis by Trost uses two alkynes in the cyclisation and the problems disappear. This starts from protected mannitol 106 and the oxidative cleavage is followed at once by imine formation 193. Addition of the lithium derivative 194 goes with good selectivity (13:1 in favour of 195). Adjustment of protecting groups and hydrolysis of the acetal leads to the key intermediate37 197.

PMBO

OH

O

1. NaIO 4, H 2O, 0 ˚C

O O

2.

O

OH

N

O

NH2

Me3Al

O

106; bis-acetal of mannitol

Li 194; PMB = p-MeO-benzyl

193; 95% yield

PMBO

PMBO 1. Pd(dba) 2 dppb, 2-thiosalicylic acid 2. PhOCOCl NaHCO3, H 2O

HN

O O

195; 69% yield

from alkyne and BuLi

+

PMBO 1. TsOH MeOH

O

2. K 2CO3

PhO

N H

HN

O O

OH

O O 197; 88% yield

196; 99% yield

The second alkyne is introduced after oxidation to an aldehyde again with good selectivity (6:1 in favour of 198). Protecting groups are added or adjusted before the Pd(0)-catalysed cyclisation gives both alkenes in 199 in the right place with the right geometry. The mechanism of the cyclisation was already established by Trost38 and is discussed in the workbook. Cyclisation and deprotection gave streptazolin 184 in only 11 steps from the chiral pool.

197

1. DMSO (COCl)2 Et3N CH2Cl2 2. ZnCl 2 RMgBr R = propynyl

Me

1. TBDMSCl imidazole 75% yield 2. DDQ 93% yield

PMBO

HN

OH O

O 198; 50% yield

3. TsCl, pyr 75% yield 4. Pd 2(dba)3 HCO2H, Et 3SiH 64% yield

OR TsO

HN

H

O 199

O

H

1. NaH, THF 84% yield 2. TBAF THF 99% yield

184

484

23 The Chiral Pool

Amino Alcohols Synthesis of quinolone antibiotics Quinolone antibiotics such as ofloxacin 200 are totally different from β-lactams, tetracyclines and so on and work in a different way. They offer some hope that resistance may be slow to appear. They all contain the ‘quinolone’ core: a benzene ring fused to a γ-pyridone. Most have various amine substituents and ofloxacin has a fluorine atom. Disconnection of the enamine reveals a benzene ring with a series of heteroatom substituents (N, N, O, F) and the one fluorine suggests that a series of nucleophilic substitutions might provide a synthesis.

O

O

O CO2H

F

C–N

N

N O

MeN

N

enamine

F

CO2H

F

NH

CHO

200; ofloxacin

SNAr

F

2 x C–N C–O

O

MeN

CO2H F

CHO

F

201

202

The acid chloride of tetra-fluoro benzoic acid 203 acylated the magnesium enolate 204 of diethyl malonate to give 205. The chelated magnesium enolate avoids O-acylation. Condensation with ethyl orthoformate puts in the masked aldehyde group and 206 is ready for a succession of aromatic nucleophilic substitutions.39 O O

Mg

F

OEt

1. 204

F

2. hydrolyse 3. heat (–CO 2)

F

Cl

O

EtO

O

F

F

CO2Et

CH(OEt)3 Ac2O

F

F 203

204

O

F

F

CO2Et

F

F F

205

OEt

206

Reaction with the amino alcohol alaninol 207 from the chiral pool forms the enamine 208 and introduces the only chirality. The first two substitutions are controlled by the tether. The amine can reach only the ortho position to give the quinolone 209 and now the hydroxyl group can reach only the meta position to give 210. The geometry of the enamine 208 means nothing as it is a conjugated enaminone and such double bonds rotate easily. The displacement of F by N to give 209 is activated by the carbonyl group but that of F by O to give 210 is not. The three fluorine atoms in the ring help and it is an intramolecular reaction. O

NH2 HO

O CO2Et

F 207

F

F F

NH

DMF

F

N

HO 208

F

CO2Et

KF

KF

206

O CO2Et

F

F 209

DMF

F

N O

OH

210

23 Amino Alcohols

485

That leaves just one nucleophile to introduce, the piperazine ring. After hydrolysis of the ester, N-methylpiperazine 212 displaces the fluorine para to the ketone and ofloxacin is formed. Amino alcohols are closely related to amino acids so this one example will be enough. The next group of antibiotics also look as though they might come from amino alcohols, but wait… O

O

F

CO2H

CO2H

F

NaOH

210

H2O

F

N

N

NH

O

MeN

Me

212

N O

N

200

211

The oxazolidinone antibiotics The oxazolidinones, general structure 213, such as the Upjohn compound U100766, 213; R ⫽ morpholine, are antibiotics that act by a different mechanism to all others. Amide disconnection takes us back through amine 214 and azide 215 to alcohol 216. R

O

O

O

F

N

O

H 213

H N

O

FGI

C–N

ArN

amide

FGI

O

ArN

O

NH2 H 214

O

ArN

O

N3

OH H 216

H 215

Opening the oxazolidinone ring reveals a three carbon chain with one functional group on each carbon atom 217. Glycidol, available as either enantiomer, is an ideal starting material. O C–O

ArN

EtO2C

ester

216a

1,2-diO

O

ArN

O

H

OH

H

OH

C–N

= H

OH

217

OH

218

O

H

O

O

Pr

O 222 (R)-(+)-glycidol (R)-Glycidyl butyrate OH

One of these antibiotics 224 was made by Upjohn from the Cbz-protected aromatic amine 219 and the ester 222 in place of glycidol. The lithium derivative of 219 attacks the epoxide releasing an oxyanion that attacks the Cbz group 220 releasing PhCH2O⫺ that in turn cleaves the butyrate ester and gives 221 in one pot and 85% yield.40 O O

Ar

1. BuLi –78 ˚C

N

Pr

N

F

NHCbz 219

OBn

O

O

O

2. 222

F

N

O

O 220

N

O OH

221; 85% yield

486

23 The Chiral Pool

The synthesis is completed with mesylation and azide displacement to give 223 then reduction of N3 to NH2 and acetylation. This 224 is U100766, an antibiotic with a potency similar to that of vancomycin that might not stimulate resistance as quickly as β-lactams.41

O

O N

221

F

2. NaN 3 DMF

N

O

1. MsCl Et3N

N

O

1. H 2 Pd/C

O N3

F

2. Ac 2O pyridine

N

O NHAc

223

224; 85% yield from 226

The synthesis of anti-viral nucleoside analogues Many anti-viral compounds are analogues of the natural nucleic acid nucleosides such as adenosine 225. This example 226, known as (S,S)-iso-ddA (dda stands for di-deoxy-adenosine), has a severely modified sugar 227 but the purine, adenine 228, is unaltered. The purine is attached to C-2´ instead of C-1´ making the molecule more stable and two of the OH groups have been removed. Coupling adenine to the modified sugar 227 needs a Mitsunobu process.42

H2N N HO

NH2 N

N

N

RO

N

O

N

NH2 OR

N

N

N

+

Mitsunobu

N H

O

N

OH

2' 226; adenosine analogue

OH OH 225; adenosine

N

O

227 sugar analogue

228; adenine a purine

Since 227 has a 1,2-diX relationship, one way to make it would be to change the terminal OR into I 229 and use an iodine cyclisation of the alkene 230. That allows us to recognise a glycidol unit 231 by disconnection of the vinyl group. O FGI

I

Iodoetherification

OH

227 229

OH

Protect?

OH 230

OH C–C

O

= O

Protect?

H OH

231 (R)-(+)-glycidol

It turns out that we need to protect only one of the hydroxyl groups: the primary OH of glycidol as a t-butyl diphenylsilyl ether 232. The epoxide is opened by a vinyl cuprate and the iodoetherification occurs regio- and stereoselectively (no four-membered rings are formed and the trans stereochemistry of 230 is preferred). Iodide is displaced by the p-nitrobenzyloxide 233 and finally the Mitsunobu reaction with adenine goes on the right nitrogen atom to give 234.

23 The New Chiral Pool

OH

1. MgCl 2 CuCN

OTBDPS

487

O I2, NaHCO 3

231 O

I

CH3CN

2. TBAF

OH 229; 71% yield

OH 230; 98% yield

232

O2N

O2N O 234

O

K 233

N

adenine PPh3

RO

DMSO 18-crown-6

OH

NH2

O

N

N

N

O

DEAD DMF

227; 73%

234; 62% yield

The New Chiral Pool Glycidol as a chiral pool member You may have felt that all this was out of place as glycidol is not available from natural sources. But both enantiomers are available quite cheaply by several processes such as Sharpless asymmetric epoxidation (chapter 25) and enzymatic hydrolysis of glycidol esters (chapter 29). We propose to elect glycidol and a number of other small molecules to the ‘New Chiral Pool’ as they now fill the same role as, say, the amino acids – cheap sources of enantiomerically pure small fragments of molecules that can easily be added to other fragments with predictable stereochemistry. Whenever you require an electrophilic three-carbon fragment with functionality at all three carbons, glycidol 231 or 235 is a natural choice. But there’s another, arguably even more useful reagent that does much the same thing – epichlorhydrin 236 and 237.

O

H OH

231 (R)-(+)-glycidol

O

H

O

OH

235 (S)-(–)-glycidol

H

H

O

Cl

236 (R)-(–)-epichlorhydrin

Cl

237 (S)-(+)-epichlorhydrin

There is a complication: it is obvious that the terminal epoxide atom of glycidol is attacked by nucleophiles, but epichlorhydrin could be attacked at the terminal epoxide atom or at the CH2Cl group. This matters a great deal: direct displacement of chloride 240b gives one enantiomer of 241 but attack at the epoxide 240a followed by cyclisation 239 gives the other 238.

O

Nu

(R)-(–)-epichlorhydrin O

O Nu

Nu Cl 238

239

Cl 240a

Nu 240b

O

Nu

241

This sequence is particularly important in the synthesis of β-blockers such as propranolol 242. Disconnecting α-naphthol 243 and i-PrNH2 reveals a C3-synthon 244, electrophilic at both ends, that must be chiral. Epichlorhydrin is ideal.

488

23 The Chiral Pool

OH O

OH

H N

2 x 1,2-diX

OH

H2N

C–O, C–N

synthon 244

243; α-naphthol

242; propranolol

i-PrNH2

If we react epichlorhydrin with α-naphthol 243 and an amine, the amine (pKaH about 11) will remove the proton from the phenol and that will attack epichlorhydrin 245 to give the alkoxide 246 that will remove a proton from the amine salt to give the intermediate 247.

OH

O

O

ArO

R2NH

ArO

R2NH2

ArOH pKaH ~11

Cl

ArO

pKa ~10

Cl

pKa ~11

246

245

Cl 247; pKa ~15

If we want to do the reaction in one step we must use a stronger base, such as NaOH, so that the intermediate 246 has a long enough lifetime to cyclise to the new epoxide 248 that will react with i-PrNH2 to give propranolol.43

O

Cl

NaOH

O

O

OH

i-PrNH2

O

H N

243

236; R-(–)epichlorhydrin

242; propranolol

248

To make enantiomerically pure propranolol, it doesn’t matter which way round the reaction goes as long as you know which and as long as it goes one way or the other. Glycidol 235, easily made from mannitol, was made by Sharpless using asymmetric epoxidation (chapter 25) and converted into the tosylate 249. This reacts with the anion of α-naphthol 243 at the CH2OTs end and not at the epoxide end to give 250 and hence propranolol.44 But we must use the other enantiomer, (S)-glycidol 235, whereas we used (R)-epichlorhydrin 236.

O

H

TsCl

OH

235 (S)-(–)-glycidol

O

H

243

OTs

O

H

i-PrNH2

OAr

i-PrNH

OAr

NaH DMF

Et3N

249

250

242; propranolol

Phenylglycine as a chiral pool member Phenylglycine doesn’t occur in proteins but is available as either enantiomer. When ScheringPlough made the racemic drug 251 as an NK1 antagonist, the task was easy. Because a trusty hydantoin synthesis already existed, they mixed the ketone 253 with potassium cyanide and ammonium carbonate to get 252 and reduced out the spare amide carbonyl with LiAlH4.

23 The New Chiral Pool

CF3

O

O

NH

+ KCN, (NH 4)2CO3 hydantoin synthesis

NH FGA

HN O

Ph

489

HN

O O

Ph

CF3 251

Ar

2 x C–N

O O

Ph

252

Ar

253

When they wanted a single enantiomer, this synthesis was useless. Where could chirality be introduced? Instead they disconnected the other two C⫺N bonds in the ring 252a and worked their way back to the enolate of phenylglycine 255 and some alkylating agent 256. O NH HN

2 x C–N

O O

Ph

Ar

amides

H2 N

C–C

CO2H O

Ph

252a

H2N

Ar

CO2H + X

Ph 255; enolate

alkylation

254

O

Ar

256

They did not know, of course, which enantiomer would be more active but it turned out that the methyl ester 257 was the right choice. Conversion to the amide 258 and aminal formation with pivalaldehyde 259 allowed control of enolate alkylation by Seebach’s relay method (chapter 30). Hydrolysis of 260 and reaction with ClSO2NCO gave one enantiomer45 of 251. t-Bu

t-Bu H3N Ph

CO2Me

MeNH2 H2O

H

H2N Ph

CONHMe H

HN Ph

257; >99%ee

258; >97%ee

1. LDA

NMe

t-BuCHO

O H

259; 49% yield

2. 256 (X = Br)

NMe HN

O O

Ph

Ar

260; 70-75% yield

C2 symmetric diamines as chiral pool members One mono-amine 261 and one diamine 262 were resolved in chapter 22 by the simplest of processes. They are joined in the new chiral pool by two closely related C2 symmetric diamines, 263 and 264, made by rather different procedures. Though we show only one enantiomer of these amines, both are available at about the same price as they are all made by resolution.

NH2

NHMe

NH2

NH2 NH2 261

262

NHMe 263

NH2 264

Diamine 264 is made from the α-diketone benzil by a simple but multiple cyclisation to give the heterocycle 265. Stereoselective dissolving metal reduction gives anti-266 and hence racemic 264 on aminal hydrolysis. Resolution with tartaric acid brings down crystals of the tartrate of one enantiomer. With L-(⫹)-tartaric acid, the salt with (S,S)-264 is less soluble while (R,R)-264 requires treatment with D-(⫺)-tartaric acid. One important application is the asymmetric Lewis acid46 267.

490

23 The Chiral Pool

O

NH2 Ph

Ph

O

N

1. Li, THF 2. EtOH

N

HN

NH

NH4OAc, AcOH

O benzil

Ph

Ph

Ph

NH2

1. L-(+)-tartaric acid

Ph +

2. filter 3. recrystallise 4. NaOH

Ph

Ph

NH2

Ph

N

Ph

N

Al Me

NH2

(S, S)-264 57-66% yield, >98%ee

NH2 (±)-264; ~ 90% yield syn but racemic SO2CF3

NH2

Ph

3. HCl 4. NaOH

anti-266

265; ~ 95% yield RESOLUTION

Ph

Ph

Ph

SO2CF3 267

(R, R)-264 in solution

Diamine 263 is made by the radical (pinacol style) dimerisation of the benzaldehyde imine 268. This gives a diastereomeric mixture equilibrated in favour of the syn isomer with lithium in isoprene and separated (51% yield) from the meso isomer by crystallisation with racemic tartaric acid.47 Finally, (±)-263 is resolved with a single enantiomer of tartaric acid giving 90% yield of either (S,S)-263 or (R,R)-263, depending on which enantiomer of tartaric acid is used, in 99% ee. The isomer remaining in solution can be isolated with only slightly worse ee: 96%.

O Ph

N

MeNH2

H

H2O

Ph

Me

H

NHMe 1. Zn, Me 3SiCl MeCN 2. Hydrolysis

268; 95% yield

Ph

NHMe

1. Li,

Ph

NHMe 269; syn + anti use crude

Ph

Ph 2. racemic tartaric acid 3. NaOH, H 2O

NHMe (±)-263 syn but racemic

This diamine has had many uses. The bis m-trifluoromethyl compound 270 made in the same way from m-trifluoromethylbenzaldehyde, is a chiral NMR shift reagent that gives better spectra than the more traditional Eu-based shift reagents and allows the reliable determination of ee. Terpenes are notorious for low ees and the ee of myrtenal 271 was determined as 96% by formation of the aminal 272 and the measurement of its 19F NMR spectrum.48 The signals for the two diastereoisomers were easily distinguished and integrated even though the compounds were difficult to separate by chiral HPLC. Ar NHMe F3C

CHO CF3

Me

N N

NHMe 270; chiral NMR shift reagent

Ar

+ Me 271; myrtenal

272

Compound 263 itself forms C2-symmetric aminals and the one from glyoxal 273 is valuable in controlling nucleophilic attack on the remaining aldehyde group 274 or on electrophilic alkenes such as 275 formed in a Wadsworth-Horner-Emmons reaction (chapter 15). The aminal can be hydrolysed with acid in conditions that do not cause any racemisation even next to an aldehyde.49

23 The New Chiral Pool

491

Ph

Ph Me

O

N

Ph

263

H

H

H

O 273; glyoxal

Me

O (EtO)2P

N

Me

N

Bu2CuLi

Me E-275; 77% from 267

H

Ph

H2O

N

EtO2C

Ph N

EtO2C

NaOEt

Me 274 Ph

O

N

CO2Et

Me 276; >99:1 diastereoselectivity

CHO

EtO2C

(S)-277; >99% ee

The aminal 279 with pyridine-3-aldehyde 278 adds cuprates in the 4-position with excellent stereoselectivity to give partly reduced pyridines 280 that are immediately acylated 281. This example can be reduced and cyclised to the alkaloid 282 with two chiral centres under control.50

N

N 263

MeLi 2CuBr.SMe 2

Me N

CHO 278

H N

Ph

MeN

Ph

MeN

279 Ph MeO

ArCH2COCl

Me N

4 LiBr

280 Ph

O MeO

MeO

N

1. LiAlH 4

Me N Ph

MeN 281

2. CF 3CO2H

N

MeO

3. (CF 3CO)2O 4. MeOH, NaOMe

H CHO

Ph

282

Amino-indanols These may seem strange molecules to have a place in the new chiral pool but they were made on a vast scale by Merck in the synthesis of their HIV protease inhibitor crixivan (Indinavir). They both come from Jacobsen epoxidation of indene 283 (chapter 25): the anti-compound 285 by opening the epoxide 284 at the benzylic centre with azide ion.51 NH2

NH2 O

Jacobsen epoxidation

OH

OH

chapter 25

283; indene

284

285

286

The syn compound 286 is more interesting. It was absolutely necessary for the synthesis of crixivan and is clearly visible at the right hand end of 287. It is not a natural product.

492

23 The Chiral Pool

N

OH

Ph

N 287 Merck's HIV-protease inhibitor crixivan (Indinavir)

t-BuNH

OH

H N

N O

O

Merck devised a synthesis that was so efficient that 286 is now available commercially (it was only a rumour that Merck would give it away free as they had so much) and has been used in asymmetric synthesis in many ways. The synthesis is described52 in full by the Merck team in Organic Syntheses. The epoxide 284 is protonated by acid to give the cation 288 that captures acetonitrile in a Ritter reaction giving the heterocycle 290 that is easily hydrolysed to 286.

284

O

MeCN

N

H MeCN

O

OH

H2SO4 (fuming)

288

289

H2O

286

290

The stereochemistry of this reaction looks all wrong. In fact the cation 289 prefers to capture acetonitrile 292 from the top side, opposite the OH group, to give anti-291. But this intermediate cannot cyclise and is in equilibrium with 292. Only when syn-291 is formed can cyclisation occur and the cations are removed from the equilibrium. Even the cyclisation that does occur looks bad enough but 5-endo-dig reactions are all right. Me

Me Me

N

N

N OH

OH

anti-291

290

OH

292

syn-291

The amino-indanol 286 is the basis for many important asymmetric reagents such as the enolates derived from 293 and the metal complexes such as 295 derived from the dimer 294. These are effective as catalysts for Diels-Alder reactions.53 O Ph

N

O

O

O

O N

N

Cu(OTf) 2

O N

Cu

TfO

293

294

N

OTf

295

23 Conclusion: Syntheses from the Chiral Pool

493

The synthesis of crixivan (Indinavir) Further steps in this synthesis are interesting as the alkylation of the enolate of 293 with glycidol tosylate 249 puts in the next two chiral centres. Attack occurs at the epoxide end of the electrophile 296 and the stereochemistry at C-2 is controlled by the amino-indanol. The stereochemistry at C-4 is controlled by the electrophile and is again unchanged when the epoxide 297 is attacked by the piperazine nucleophile.54 The remaining centre in crixivan is in the piperazine ring and this is secured by asymmetric hydrogenation (chapter 26). We have now explored both new and old chiral pool enough to look at a few syntheses without prejudging the starting materials. O 293

(Me3Si)2NLi

Ph

O

Li

N

O O

Ph

O

N

Ph

N

O

O

TsO

O

2

O

4

O

249 OTs TsO 296

297

298; 72% yield

Conclusion: Syntheses from the Chiral Pool The synthesis of agelastatin Agelastatin 299 is an interesting molecule partly because it is an anti-neoplastic and partly because of the massing of functionality around such a small frame. It contains a cluster of heteroatoms all having 1,1-diX relationships and removal of these reveals the basic carbon skeleton having a carbonyl group in a 1,3-relationship to the pyrrole nitrogen. Disconnection 300 with conjugate addition in mind reveals a much simpler compound 301 with just two neighbouring chiral centres. Me HO

N

Br

O

O NH

O

Br

Br

1,1-diX

N

NH2

N 3 x C–N

NH

NH2

NH C–N

NH

O 299; agelastatin

1,3-diX

NH

O 300

O 301

Disconnecting the amide 301a isolates the five-membered ring containing both chiral centres 303. It is not obvious whether this compound could be made from the chiral pool but literature search revealed a glucosamine derivative 304 that looked promising. Neither of your authors has detailed knowledge of the sugar literature, and neither author would have seen this possibility. O Br

Br

O

O Ph

NH2

NH NH

amide

?

NH

302

OMe

NH2

C–N

CO2H O 301a

+

O O

NH2 303

N H 304

494

23 The Chiral Pool

Glucosamine can be N-benzoylated 305 and protected as a methyl glucoside and a benzylidene acetal leaving just one OH free 306. Tosylation of this free OH and treatment with base gives the three-membered ring55 304. HO HO HO

1. MeOH, H +

O OH

O O HO

Ph

2. PhCHO ZnCl2

NHBz

O O TsO

Ph

TsCl

O

pyridine

304 i-PrOH

BzNH OMe 307

BzNH OMe 306

305

Na

O

The synthesis of agelastatin from 304 is too long to describe in detail so we shall just pick out the most important stages. The trans relationship between the two nitrogen atoms in 303 was established by opening the N-acylated aziridine 308 with azide ion. The trans-diaxial product 309 is formed (chapter 21). Further functional group and protecting group manipulations (five steps) give 310. The protecting group SES is Me3SiCH2CH2SO2⫺, designed to be removed by fluoride ion. O

Ph ClCO2Me

304

O

NaN3

N OMe CO2Me 308; 86% yield

pyridine

NH4Cl DMF

N3 O

O

Ph

O

O

MeO2CNH

NHR O

I Et3SiO MeO2CNH

OMe

309; 88% yield

OMe

310; R = SES

Now two critical steps. The iodide 310 is fragmented by zinc in aqueous THF 311 to give the unsaturated aldehyde 312 in astonishing yield. The mechanism 311 has the NHR group omitted for clarity: the loss of the MeO group may not be concerted with the rest. Though Wittig and Peterson failed, a Kocienski-Julia reaction (chapter 15) gave the diene 313 used immediately in the next step. Zn

N

I

O

Et3SiO MeO2CNH 311

Et3SiO OMe

CHO

MeO2CNH NHSES 312; 97% yield

N

N N Ph

SO2Me

(Me3Si)2NK

Et3SiO MeO2CNH 313

NHSES

We hope you guessed that metathesis was used to close the five-membered ring. The modified Hoveyda-Grubbs catalyst 314 gave the best results and the product 315 cyclised to 316 after removal of the Et3Si and CO2Me protecting groups. This compound contains all the functionality and stereochemistry of 303. So was it all worth it, just to use glucosamine from the chiral pool? Well, this was the first asymmetric synthesis.56 R N Cl Ru Cl O i-Pr 314

N R

314

O

Et3SiO

K2CO3

313

MeOH

MeO2CNH

NHSES 315

O

N H

NHSES

316; 36% yield from 312

The final stages57 are better seen if we turn 316 round to make it look more like 303. Protection of the ring nitrogen 317 allowed acylation of the other nitrogen by the pyrrole 318 having Me3Si in place of Br.

23 Conclusion: Syntheses from the Chiral Pool

O

O

O

O

1. BuLi

NH

N Bn

Et3N, DMAP

316

THF, RT four hours

Cl

Bn

Me NHSES

NHSES

319 89% yield

NH

+

N

N ClOC

Me3Si

O

2. Me

495

O 318

317; 88-94% yield

While 319 could be converted into the required enone 320, no amount of trying allowed further progress with the synthesis. Conjugate addition occurred but other reactions followed. The answer was bromination. This gave a mixture of mono-, di-, and tri-brominated pyrroles but they all cyclised with Hünig’s base to a mixture of protected 300 brominated in various places from which agelastatin was finally made. O O

O

O N

Me3Si

N

Bn

Me NSES

H N

Me

H N

N

O

H N

Me3Si

Bn

NH

O

O

319

320

The synthesis of LAF389, a Novartis anti-cancer agent The natural product bengamide B shows promising anti-cancer activity and Novartis have initiated a programme to synthesise the analogue LAF389 321 on a large scale.58 It contains a side chain with four contiguous chiral centres joined by an amide to a seven-membered ring heterocycle with two chiral centres. OH

OMe

t-Bu

H N

O 4

1

OH

OH

321 Novartis's LAF389

NH O

O 1

O

There is a 1,4-relationship both between the chiral centres on the ring and between those on the side chain and the nearer on the ring. This makes the ester and amide disconnections shown on 321 attractive giving three starting materials. OH C–O and C–N

321

O

OMe OH +

t-Bu

H2N

CO2H NH

+

ester and amide

OH 322

OH

O

OH 323

324

The side chain fragment 322 has the look of a sugar and, if the alkene is removed with a Julia reaction in mind (as it is an E-alkene, chapter 15), it has six carbon atoms with oxygen on each one 325. Unfortunately two of the centres are not those in glucose 99. However a C7 sugar, the

496

23 The Chiral Pool

gulo-heptonic acid 326 is available. It has all the right stereochemistry at C-2 to C-5, only C-6 is unwanted but as it also has one too many carbon, that centre can be removed in making the aldehyde 325. Again, a search of the sugar literature would be helpful. OH

Julia

322

OH

OH

OHC

OH

CO2H HO OH

OH CHO

OH OH 99; glucose

OH 325

HO

OH 5

OH 1

3 4

6

2

CO2H

OH OH 326; gulo-heptonic acid

The acid 326 exists as the lactone 327 and can be selectively protected as 328 for oxidative cleavage by periodate to give 329. The development synthesis was similar to the laboratory synthesis but in this step as in many others careful improvements were needed for it to be safe and effective on a large scale. O

O OH

OH

O

OH

HO

O

O

O OMe

HO O

OH OH 327; gulo-heptonic acid γ-lactone

NaIO4

OMe

OHC

O

O

328

O 329

The other chiral component, the lactam 323, is clearly a cyclised version of 5-hydroxy-lysine 330. This and hydroxyproline 331 occur in the connective protein collagen. The hydroxy groups are added after the collagen is assembled but as they can be obtained by hydrolysis of collagen, they are members of the chiral pool. O

HO NH2

H 2N

NH

C–N

H2N

CO2H

amide

OH

OH 323a

330; (2S,5R)-5-hydroxy-lysine

CO2H N H 331; (2S,4R)hydroxy-proline

Hydroxyproline is abundant but hydroxylysine is not. Fortunately it can be made59 from an abundant member of the chiral pool - malic acid 34. Borane reduction gives the triol 332 and anisaldehyde gives the diequatorial acetal 333 under thermodynamic control. HO2C

CO2H OH

34; (R)-malic acid

Me2S.BH3 B(OMe)3

OMe ArCHO

HO OH 332

OH

O

pyr.HOTs

O HO

333; 89% yield from 34

The free OH group can be converted into the azide 334 by the usual sequence. Now comes a critical step. The acetal is opened by reduction with remarkable (in their words ‘unexpected’) selectivity to give 335. The key is to have a Lewis acid (a silyl chloride) present. If a proton is used, the opposite selectivity is achieved.

23 Conclusion: Syntheses from the Chiral Pool

333

O

1. TsCl, Et 3N, DMAP (98%)

O

Ar

N3

2. NaN 3, DMF, 55 ˚C (88%)

497

NaB(CN)H3

N3

t-BuMe2SiCl

TBDMSO

334

OCH2Ar

335; 89% yield

Now the terminal OH group can be deprotected, turned into a leaving group 336, and used to alkylate an asymmetric glycine anion equivalent, the heterocycle 337. The azide is then reduced and protected to give 338, a protected version of 330. The yield in the alkylation was 94% and only one diastereoisomer was detected. Ph Cbz

Ph Ph

N

N3

O

TBDMSO

I O 337

336

Cbz

1. (Me 3Si)2NNa 15-crown-5 2. 336 3. Ph 3P, H 2O 4. (Boc) 2O

Ph

N O

BocNH TBDMSO

O 338

The synthesis of 323 needed activation as the methyl ester for cyclisation to occur spontaneously. Protection of the primary amine was needed to get selective acylation of the alcohol to give 340. O 330

Me3SiCl MeOH

methyl ester of 330

1. Et 3N

BocNH

O NH

NH2

1. RCOCl

2. (Boc) 2O

NH O

2. 2M HCl EtOAc

OH

O 340

339

The final stages of the synthesis of LAF389 require only the combination of 340 with the Julia reagent 341 and deprotection and the synthesis of the drug molecule 321 from two different members of the chiral pool is complete. O O O t-Bu

S

N S

1. BuLi 2. Me 3SiCl

O OMe

t-Bu

3. 329

341

O

O

1. 340, base

321

2. 1M HCl 3. H 2O EtOH

342

PART III – THE CHIRAL POOL In the following tables the main members of the traditional chiral pool with our selection from the ‘new chiral pool’ are listed with the compound numbers given them in this chapter. The most accessible full listing is in Morrison60 and the most comprehensive in Houben-Weyl61 (this volume is in English). Valuable (and free) source books are the catalogues of chemical suppliers. These are the only way to find out prices as these change year-by-year.

498

23 The Chiral Pool

The amino acids The proteinaceous α-amino acids are all available. They have the general structure 3 where R can be an alkyl group or a variety of substituted groups.3,4 There are full listings in many textbooks.62 Important members of this class described in the chapter include phenylalanine 4, valine 5, proline 6, glutamic acid 7, aspartic acid 9, alanine 149 and leucine 163. To these we add two functionalised amino acids, the thiol cysteine 343 and the amine lysine 344 with two heterocyclic amino acids, the imidazole histidine 345 and the indole tryptophan 346. O R

OH

3; ( L)-amino acids

NH2 O Ph

O

O OH

NH2 4; (S)-(–)-Phe phenylalanine

OH NH2 5; (S)-(+)-Val valine

O HS

CO2H

N H

HO2C

6; (S)-(–)-Pro proline

OH

343; (R)-(+)-Cys cysteine

HO2C

OH

NH2 9; Asp (S)-(+)-aspartic acid

O OH

NH2

OH

NH2 7; (S)-(+)-Glu glutamic acid

O

NH2

O

NH2

H N

OH

OH

NH2

N

344; (S)-(+)-Lys; lysine

O

NH2

N H 346; (S)-(–)-Trp; tryptophan

345; (S)-(–)-His; histidine

Hydroxyamino acids are useful because the OH group can be converted into a leaving group. We saw serine 138 earlier in the chapter. The other three have a secondary alcohol at a potentially useful chiral centre. Threonine 347 is found in normal proteins. The other two are present in collagen and the extra hydroxyl groups are added by oxidation after the protein is formed (‘post-translational modification’). Hydroxyproline 331 is abundant, hydroxylysine 330 less so. We might also add synthetic phenylglycine as a new member. O HO

O OH

NH2 138; (S)-(+)-Ser serine

HO

HO

NH2 OH

NH2 347; (2S,3R)-(–)-Thr threonine

H2N

CO2H OH

330; (2S,5R)-5-hydroxy-lysine

CO2H N H 331; (2S,4R)hydroxy-proline

Hydroxy-acids A few hydroxyacids are available: lactic 33, malic 34, tartaric in both forms 35 and 36, and mandelic 37. H

OH

HO

H

CO2H

CO2H 33; (+)-(S) -lactic acid

HO HO2C

CO2H 34; (–)-(R) -malic acid

OH CO2H

(R,R)-(+)-35 tartaric acid "natural"

HO HO2C

OH

OH CO2H

(S,S)-(–)-36 tartaric acid "unnatural"

Ph

CO2H

37; (R)-(–)mandelic acid

23 Conclusion: Syntheses from the Chiral Pool

499

Derived from them and from other chiral pool members is a small collection of epoxides that are members of the new chiral pool: both enantiomers of propylene oxide 44, of glycidol 231 and 235, and of epichlorhydrin 236 and 237, and the bis-epoxide 117. H

O

H

O

O

H

H

O

OH

OH

235 231 (+)-(R)-44 (–)-(S)-44 (R)-(+)-glycidol (S)-(–)-glycidol

H

O

Cl

H

H

O

Cl O

237 236 (S)-(+)-epi (R)-(–)-epi chlorhydrin chlorhydrin

O H 117

Amino alcohols Reduction of amino acids (usually with borane) gives amino alcohols such as valinol 27, prolinol 28 or phenylalaninol 30: they have the general structure 51.

OH NH2 27; (S)-valinol

Ph

OH N H

OH

R S

OH

NH2

28; (S)-prolinol

NH2 51; amino-alcohols 30; (S)-phenylalaninol from amino-acids

Other amino alcohols occur naturally, including the ephedrine family 52 – 54. We can add the famous amino-indanol 286 to this collection. OH S Ph

OH

OH S Ph

R NHMe

52; (1S, 2R)(+)-ephedrine

S Ph

S

R OH NH2

NHMe 53; (1S, 2S)-(+)pseudoephedrine

NH2

54; (1S, 2R)-(+)norephedrine

286 the aminoindanol

Diamines These important compounds are all members of the new chiral pool and are all C2 symmetric.

NH2 NH2 262

NHMe

NHMe 263

NH2

NH2 264

Terpenes The variety of terpenes is great but they may not be available in high ee. Here is a selection: these are all monoterpenes (C10) and may be linear, cyclic or bicyclic. We have included the new chiral pool members 8-phenyl-menthol 83 and camphorsulfonic acid 94.

500

23 The Chiral Pool

O 5

2

1 OH

OH

79; (1R, 2S, 5R) -(–)-menthol

80; (1S, 2R, 5S) -(+)-menthol

OH

83; (–)-8phenylmenthol

82; (R)-(–)carvone

81; (R)-(+)pulegone

CHO

OH 75; (R)-(+)-citronellol

CO2H

76; (R)-(+)-citronellal

77; (R)-(+)-citronellic acid

R

R

93; (+)camphor

R

89; (1R)-(+) -α-pinene

Ph

H

O

HO3S

O

91; (–)-β-pinene

S

94; (+)camphor10-sulfonic O acid

Sugars There is again a great abundance of sugars with every stereochemistry at every centre available at a price and their use in asymmetric synthesis is described in recent books.63 The most important C6 compounds are glucose 99 and mannose 104 and its reduction product mannitol 105. The sugars are shown in their open chain forms so that the stereochemistry is clear. You are referred to the earlier parts of the chapter for the pyranose and furanose forms. OH

OH

OH

HO

HO

CHO

OH

OH

OH

HO

CHO

OH OH 105; mannitol

OH OH 104; open-chain mannose

OH OH 99; open chain α-D-glucose

OH

The C5 sugars were not featured in the chapter but include arabinose, ribose and xylose. They each have corresponding alcohols among which beware that achiral xylitol is not much use in asymmetric synthesis.

OH HO

OH

OH CHO

OH OH D-(–)-arabinose open chain form

HO

CHO OH OH D-(–)-ribose open chain form

HO

OH CHO

OH OH D-(+)-xylose open chain form

HO

OH

OH OH xylitol WARNING! achiral

The two C4 sugars and their alcohols belong to the erythro or threo series. One alcohol 114 is C2 symmetric (the bis epoxide 117 is made from it) but the other 112 is achiral with a plane (or a centre) of symmetry.

23 References

OH

OH

OH H

HO

501

OH

HO

OH O 111; erythrose

OH H

HO OH O 113; threose

OH 112; erythritol

OH

HO OH 114; threitol

The only C3 sugar is the rather unstable glyceraldehyde best used in a protected form such as 107 or 109 available from the oxidative cleavage of mannitol derivatives such as 106.

O OH

O

O O

O

O

OH

Pb(OAc) 4

H

O O

or NaIO4

106; bis-acetal of mannitol

107; acetal of glyceraldehyde

TBDMSO

H OBn

109; alternative protected form of glyceraldehyde

Alkaloids Not much used as building blocks for target molecules but invaluable for designing reagents and catalysts. See structures 118 and 119.

References General references are given on page 893 1. J. W. Scott, Morrison, volume 4, pages 1–226. 2. L. F. and M. Fieser, Steroids, Reinhold, New York, 1959, p. 547 ff. 3. G. M. Coppola and H. F. Schuster, Asymmetric Synthesis: Construction of Chiral Molecules from Amino Acids, Wiley, New York, 1987. 4. R. M. Williams, Synthesis of Optically Active Amino Acids, Pergamon, Oxford, 1989. 5. T. Kolasa and M. J. Miller, J. Org. Chem., 1987, 52, 4978. 6. B. Koppenhoefer and V. Schurig, Org. Synth., 1988, 66, 151; 160. 7. L. R. Smith and H. J. Williams, J. Chem. Ed., 1979, 56, 696; J. P. Vigneron, R. Méric, M. Larchevêque, A. Debal, G. Kunesch, P. Zagatti and M. Gallois, Tetrahedron Lett., 1982, 23, 5051. 8. F. Degerbeck, B. Fransson, L. Grehn and U. Ragnarsson, J. Chem. Soc., Perkin Trans. 1, 1993, 11. 9. A. Abiko and S. Masamune, Tetrahedron Lett., 1992, 33, 5517. 10. D. A. Quagliato, P. M. Andrae and E. M. Matelan, J. Org. Chem., 2000, 65, 5037. 11. J. R. Gage and D. A. Evans, Org. Synth., 1990, 68, 77, 83. 12. D. Enders, P. Fey and H. Kipphardt, Org. Synth., 1987, 65, 173, 183. 13. A. K. Beck, P. Gysi, L. La Vecchia and D. Seebach, Organic Syntheses, 1999, 76, 12. 14. E. Weber, N. Doerpinghaus, C. Wimmer, Z. Stein, H. Krupitsky and I. Goldberg, J. Org. Chem., 1992, 57, 6825; C. van dem Bussche-Hünnefeld, A. K. Beck, U. Lengweiler and D. Seebach, Helv. Chim. Acta, 1992, 75, 438; D. Seebach, G. Jaenschke and Y. M. Wang, Angew. Chem., Int. Ed. Engl., 1995, 34, 2395; G. Jaeschke and D. Seebach, J. Org. Chem., 1998, 63, 1190; D. Seebach, A. K. Beck, B. Schmidt and Y. M. Wang, Tetrahedron, 1994, 50, 4643; D. Seebach, R. Dahinden, R. E. Marti, A. K. Beck, D. A. Plattner and F. N. M. Kühnle, J. Org. Chem., 1995, 60, 1788.

502

23 The Chiral Pool

15. B. T. Golding, D. R. Hall and S. Sakrikar J. Chem. Soc., Perkin Trans. 1, 1973, 1214. 16. L. R. Hills and R. C. Ronald, J. Org. Chem., 1981, 46, 3348. 17. R. Shirai, M. Tanaka and K. Koga, J. Org. Chem., 1986, 108, 543. 18. C. D. Jones, N. S. Simpkins and G. M. P. Giblin, Synlett, 1997, 589; Tetrahedron Lett., 1998, 39, 1021; 1023. 19. D. Zhao, C. Chen, F. Xu, L. Tan, R. Tillyer, M. E. Pierce and J. R. Moore, Org. Synth., 2000, 77, 12. 20. S. Niwa and K. Soai, J. Chem. Soc., Perkin Trans. 1, 1990, 937. 21. A. S. Thompson, E. G. Corley, M. F. Huntington and E. J. J. Grabowski, Tetrahedron Lett., 1995, 36, 8937. 22. K. Mori, S. Kuwahira, and H. Ueda, Tetrahedron, 1983, 39, 2439. 23. O. Ort, Org. Synth., 1987, 65, 203. 24. E. L. Eliel and J. E. Lynch, Tetrahedron Lett., 1981, 22, 2855; E. L. Eliel, J. E. Lynch, F. Kume and S. V. Frye, Org. Synth., 1987, 65, 215. 25. See discussion in Org. Synth., 1987, 65, 224. 26. Vogel, pages 654, 592. 27. M. T. Reetz and K. Kesseler, J. Org. Chem., 1985, 50, 5434. 28. Y. Morimoto, A. Mikami, S. Kuwabe and H. Shirahama, Tetrahedron:Asymmetry, 1996, 7, 3371. 29. M. A. Robbins, P. N. Devine and T. Oh, Org. Synth., 1999, 76, 101. 30. M. A. Ondetti, B. Rubin and D. W. Cushman, Science, 1977, 196, 441; Saunders, Top Drugs, chapter 1. 31. H. Urbach and R. Henning, Heterocycles, 1989, 28, 957. 32. M. J. Wyvratt, E. W. Tristram, T. J. Ikeler, N. S. Lohr, H. Joshua, J. P. Springer, B. H. Arison and A. A. Patchett, J. Org. Chem., 1984, 49, 2816. 33. T. L. Stuk, A. R. Haight, D. Scarpetti, M. S. Allen, J. A. Menxia, T. A. Robbins, S. I. Parekh, D. C. Langridge, J. J. Tien, R. J. Pariza and F. A. J. Kerdesky, J. Org. Chem., 1994, 59, 4040. 34. B. H. Norman and M. L. Morris, Tetrahedron Lett., 1992, 33, 6803. 35. J. W. Burton, J. S. Clark, S. Derrer, T. C. Stork, J. G. Bendall and A. B. Holmes, J. Am. Chem. Soc., 1997, 119, 7483. 36. H. Yamada, S. Aoyagi and C. Kibayashi, J. Am. Chem. Soc., 1996, 118, 1054. 37. B. M. Trost, C. H. Chung and A. B. Pinkerton, Angew. Chem. Int. Ed., 2004, 43, 4327. 38. B. M. Trost and D. C. Lee, J. Am. Chem. Soc., 1988, 110, 7255; B. M. Trost, F. J. Fleitz and W. J. Watkins, J. Am. Chem. Soc., 1996, 118, 5146. 39. H. Egawa, T. Miyamoto and J. Matsumoto, Chem. Pharm. Bull., 1986, 34, 4098. 40. S. J. Brickner, D. K. Hutchinson, M. R. Barbachyn, P. R. Manninen, D. A. Ulanowicz, S. A. Garmon, K. C. Grega, S. K. Hendges, D. S. Toops, C. W. Ford and G. E. Zurenko, J. Med. Chem., 1996, 39, 673. 41. see Chemistry in Britain, March 2001, page 22. 42. Y. Díaz, F. Bravo and S. Castillón, J. Org. Chem., 1999, 64, 6508. 43. A. F. Crowther and L. H. Smith, J. Med. Chem., 1968, 11,1009. 44. J. M. Klunder, S. Y. Ko and K. B. Sharpless, J. Org. Chem., 1986, 57, 3710. 45. G. A. Reichard, C. Stengone, S. Paliwal, I. Mergelsberg, S. Majmundar, C. Wang, R. Tiberi, A. T. McPhail, J. J. Piwinski and N.-Y. Shih, Org. Lett., 2003, 5, 4249. 46. S. Pikul and E. J. Corey, Org. Synth., 1993, 71, 22, 30. 47. A. Alexakis, I. Aujard, T. Kanger and P. Manganey, Organic Syntheses, 1999, 76, 23. 48. D. Cuvinot, P. Mangeney, A. Alexakis, J. F. Normant and J. P. Lellouche, J. Org. Chem., 1989, 54, 2420. 49. A. Alexakis, J.-P. Tranchier, N. Lensen and P. Mangeney, J. Am. Chem. Soc., 1995, 117, 10767. 50. P. Mangeney, R. Gosmini, S. Raussou, M. Commerçon and A. Alexakis, J. Org. Chem., 1994, 59, 1877. 51. M. Takahashi and K. Ogasawara, Synthesis, 1996, 954. 52. J. F. Larrow, E. Roberts, T. R. Verhoeven, K. M. Ryan, C. H. Senanayake, P. J. Reider and E. N. Jacobsen, Org. Synth., 1999, 76, 46. 53. I. W. Davies, C. H. Senanayake, L. Castonguay, R. D. Larsen, T. R. Verhoeven and P. J. Reider, Tetrahedron Lett., 1995, 36, 7619; I. W. Davies, L. Gerena, L. Castonguay, C. H. Senanayake, R. D. Larsen, T. R. Verhoeven and P. J. Reider, Chem. Commun., 1996, 1753; I. W. Davies, C. H. Senanayake, R. D. Larsen, T. R. Verhoeven and P. J. Reider, Tetrahedron Lett., 1996, 37, 1725; A. K. Ghosh, P. Mathivanan and

23 References

503

J. Capiello, Tetrahedron Lett., 1996, 37, 3815; A. K. Ghosh and P. Mathivanan, Tetrahedron: Asymmetry, 1996, 7, 375. 54. D. Askin, K. K. Eng, K. Rossen, R. M. Purick, K. M. Wells, R. P. Volante and P. J. Reider, Tetrahedron Lett., 1994, 35, 673; P. E. Maligres, V. Upadhyay, K. Rossen, S. J. Cianciosi, R. M. Purick, K. K. Eng, R. A. Reamer, D. Askin, R. P. Volante and P. J. Reider, Tetrahedron Lett., 1995, 36, 2195; K. Rossen, R. A. Reamer, R. P. Volante and P. J. Reider, Tetrahedron Lett., 1996, 37, 6843; P. E. Maligres, V. Upadhyay, K. Rossen, S. J. Cianciosi, R. M. Purick, K. K. Eng, R. A. Reamer, D. Askin, R. P. Volante and P. J. Reider., Tetrahedron, 1996, 52, 3327; W. J. Thompson, P. M. D. Fitzgerald, M. K. Holloway, E. A. Emini, P. L. Darke, B. M. McKeever, W. A. Schleif, J. C. Quintero, J. A. Zugay, T. J. Tucker, J. E. Schwering, C. F. Homnick, J. Nunberg, J. P. Springer and J. R. Huff, J. Med. Chem., 1992, 35, 1685. 55. Y. Ali, A. C. Richardson, C. F. Gibbs and L. Hough, Carbohydrate Res., 1968, 7, 255, 56. K. J. Hale, M. M. Domostoj, D. A. Tocher, E. Irving and F. Scheinmann, Org. Lett., 2003, 5, 2927. 57. M. M. Domostoj, E. Irving, F. Scheinmann, and K. J. Hale, Org. Lett., 2004, 6, 2613. 58. D. D. Xu, L. Waykole, J. V. Calienni, L. Ciszewski, G. T. Lee, W. Liu, J. Szewczyk, K. Vargas, K. Prasad, O. Repic and T. J. Blacklock, Org. Process Res. Dev., 2003, 7, 856. 59. T. Gustafsson, M. Schou, F. Almqvist and J. Kihlberg, J. Org. Chem., 2004, 69, 8694. 60. J. W. Scott, Morrison, volume 4, pages 1–226. 61. R. Herrmann in Houben-Weyl volume E21f, eds G. Helmchen, R. W. Hoffmann, J. Mulzer and E. Schaumann, Thieme, Stuttgart, 1996, pages 5760–6001. 62. Clayden, Organic Chemistry, pages 1353–1356. 63. G.-J. Booms and K. J. Hale, Organic Synthesis with Carbohydrates, Sheffield Academic Press, 2000; S. Hanessian, Total Synthesis of Natural Products: The ‘Chiron’ Approach, Wiley, 2004.

24

Asymmetric Induction I Reagent-Based Strategy Introduction to Reagent-Based Strategy New stereogenic centres from prochiral units in planar molecules Enantiotopic and diastereotopic groups Asymmetric Reduction of Unsymmetrical Ketones Asymmetric boron or aluminium hydrides Asymmetric reduction by boranes Reduction of ketones with Ipc2BCl (DIP-Chloride™) Asymmetric Electrophiles Asymmetric hydroboration of alkenes Asymmetric Nucleophilic Attack Transfer of Allylic Groups from Boron to Carbon Reactions of chiral allylic boranes with carbonyl compounds Reactions of chiral allyl boranes with imines Asymmetric Addition of Carbon Nucleophiles to Ketones Addition of alkyl lithiums to ketones Asymmetric epoxidation with chiral sulfur ylids Asymmetric Nucleophilic Attack by Chiral Alcohols Deracemisation of arylpropionic acids Deracemisation of α–halo acids Asymmetric Conjugate Addition of Nitrogen Nucleophiles An asymmetric synthesis of thienamycin Asymmetric Protonation Enantioselective lactonisation of achiral hydroxy diacids Asymmetric Deprotonation with Chiral Bases Asymmetric deprotonation with sparteine Asymmetric Oxidation

Introduction to Reagent-Based Strategy So far we have used racemic compounds and resolved them (chapter 22) or started with enantiomerically pure materials (chapter 23). In the next four chapters we deal with the more demanding but ultimately more versatile and important creation of new stereogenic centres. Naturally the stereogenic centres are not really created: the chirality already exists and is transferred to the growing molecule from what may be called a chiral auxiliary: this process is called asymmetric (sometimes chiral) induction. In this chapter (reagent-based strategy) the chiral auxiliary is either part of the reagent or a chiral ligand on a metal. In the next two chapters it will be part of a chiral catalyst. In the fourth chapter, the chiral auxiliary is covalently bound to the growing molecule Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

506

24 Asymmetric Induction I Reagent-Based Strategy

(substrate-based strategy) and so remains attached at the end of the reaction. Later in chapter 29 we shall discuss another kind of reagent - the enzyme. A new stereogenic centre normally appears when a trigonal carbon atom becomes tetrahedral during a reaction (chapter 21). The unsymmetrical ketone 1 could be reduced to the chiral alcohol 2 or enolised and combined with an electrophile to give 3. The starting material 1 is prochiral, and both the CH2 and the C⫽O groups are prochiral centres. If one hydrogen atom, say HA rather than HB, is replaced by the electrophile E, a single enantiomer of 3 is formed. Similarly, if one face of the C⫽O group, say the one on our side of the paper, is attacked by the reducing agent, a single enantiomer of the alcohol 2 is formed. The two hydrogen atoms and the two faces of the carbonyl group are enantiotopic: you need to know your right from your left to tell them apart. An NMR machine, or an achiral reagent such as NaBH4, cannot tell its right from its left, so HA and HB have the same chemical shift, and NaBH4 gives an exactly 50:50 mixture of the enantiomers of 2. O

OH

?

HA 2

HB

reducing agent

Ph

Ph HA 1

HB

prochiral starting materials

new stereogenic centre

O ? OH E

electrophile

new stereogenic centre

Ph

Ph

E 3

1a; enol form

If we want to produce high yields of one enantiomer of 2 or 3, we must use a reagent which can tell its right from its left, that is an enzyme or an asymmetric reagent with some built in mechanism for distinguishing two enantiotopic features. This is the subject of this chapter. Chapter 27 deals with the alternative strategy, making the Hs or faces of the C⫽O group diastereotopic so that achiral reagents can tell them apart. Other trigonal carbons which can give rise to new stereogenic centres are simple alkenes undergoing electrophilic attack, e.g. epoxidation of 4, and unsaturated carbonyl compounds 6 undergoing Diels-Alder reactions. These reactions may produce several stereogenic centres at once whose relative stereochemistry is controlled by the stereochemistry (E or Z) of the double bond, and by the endo selectivity of the Diels-Alder reaction. epoxidation

Ph

O Ph

4

O

O

and/or Ph

5a

5b

Diels-Alder reaction

R and/or

+

R

R O 6

7a

O 7b

We are concerned with the absolute stereochemistry of the product: does epoxidation give 5a or 5b? Does the Diels-Alder reaction give 7a or 7b? There is also a small group of prochiral tetrahedral carbon atoms with enantiotopic functional groups such as the diester 8 or the diene 9. We shall meet examples of all these (and more!) in this chapter.

24 Asymmetric Reduction of Unsymmetrical Ketones

507

R 8

9 MeO2C

CO2Me

This chapter and the next two deal with two approaches. Asymmetric reagents are enantiomerically pure compounds used in stoichiometric amounts to make single enantiomers of the products. Asymmetric catalysts are enantiomerically pure compounds used in sub-stoichiometric amounts to catalyse the reaction of stoichiometric but achiral reagents to achieve the same result. You might think it would be easy to distinguish these approaches and often it is. However it can be difficult and broadly we shall describe stoichiometric compounds that transfer the odd atom to the final product as ‘reagents’ and compounds that are used in substoichiometric amounts and usually transfer no atoms to the product as ‘catalysts’. In outline this chapter will deal with asymmetric reduction, asymmetric acids and bases, and asymmetric nucleophiles and electrophiles. Asymmetric oxidation will mostly be dealt with in chapter 25

Asymmetric Reduction of Unsymmetrical Ketones The reduction of unsymmetrical ketones such as 1 with achiral reagents NaBH4 or LiAlH4 gives racemic alcohols. The obvious way to make such a reduction asymmetric is to replace some of the H atoms around B or Al with ligands that form good bonds to boron or aluminium. This is usually done by reacting the achiral reagents with chiral diols or amino-alcohols.

Asymmetric boron or aluminium hydrides One of the earliest was the darvon complex of LiAlH4. We saw in chapter 22 that both enantiomers of the amino-alcohol 11 (DARVON and NOVRAD) are available by resolution of the precursor 10. DARVON 11 is also available as Chirald®. Reaction of DARVON 11 with LiAlH4 replaces two hydrides with the OH and NMe2 groups to give the reagent DARVON-H that presumably has the structure 12. H Ph

O Ph

Ph

H

OH

MgCl

NMe2

Ph

(+)-(R)-10

O

LiAlH4

H

Al

H NMe2

Ph

NMe2

Ph

(2S,3R)-(+)-11; DARVON

H 12

This complex chiral hydride 12 reduces acetylenic ketones such as 13 with reasonable selectivity.1 The other enantiomer of 14 comes from reduction of 13 with the enantiomeric reagent derived from NOVRAD. Note that the absolute sense of the induction in the reduction of 13 is the same with 12 and with the Alpine borane from (⫹)-α-pinene, below. OH H R1

NOVRAD

LiAlH4

R1

DARVON

R2 14a 72-82% ee

OH H

O LiAlH4

R1 R2

R2 13

14b 72-82% ee

508

24 Asymmetric Induction I Reagent-Based Strategy

It is necessary for the two groups on the ketone to be sterically different and one of the best results is achieved with a linear acetylenic group on one side and a branched alkyl chain on the other. Even so the ee of this example is only 82% and newer reagents can improve on this. O

OH H

LiAlH4 darvon

13; R 1= i-Bu, R 2= Me

14; R 1= i-Bu, R 2= Me

A more successful development2 was the reagent derived from LiAlH4 and BINOL. Enantiomerically pure BINOL 15, available as either enantiomer, is combined with LiAlH4 to give 16 and then with a simple achiral alcohol such as ethanol. The reagent, known as BINAL-H, probably has structure 17.

OH OH

O

LiAlH4

O

EtOH

AlH2

THF

O

15

O

OEt Al H

probable structure of 17; BINAL-H

16

This is a better reagent in two ways. Acetylenic alcohols give higher yields and higher ees. The scope is also wider: though some contrast between the two sides of the ketone is still needed, simple ketones and, most usefully, α,β-unsaturated ketones work well. The reductions of α-tetralone 18 and ionone 20 illustrate the typical high yields and high ees. OH

O

O

OH

BINAL-H

BINAL-H

–100 ˚C to –78 ˚C

–100 ˚C to –78 ˚C

19; 91% yield; 87% ee

18

21; 87% yield; 100% ee

20

Asymmetric reduction by boranes A different style of reduction comes with the use of trialkyl boranes derived from α-pinene. Hydroboration of α-pinene 22 occurs from the less hindered face away from the gem-dimethyl groups to give the tertiary borane 23 with the usual regioselectivity (chapter 17). The marked hydrogen atom is transferred from boron to carbon.3

H

=

a-pinene

BR2

BR2

=

a-pinene

H

22

23

24 Asymmetric Reduction of Unsymmetrical Ketones

509

On reaction with ketones, a bond is first formed between oxygen and boron 24 and then the same marked hydrogen atom is transferred 25 to the ketone through a tight six-membered cyclic transition state 26. These reducing agents deliver a hydrogen atom from carbon rather than from boron. The larger group (R L) on the ketone prefers to occupy an ‘outside’ position in the transition state 26 away from the pinene-derived cage leading to one enantiomer 27 of the alcohol. R

R

R B

B

O

R

R O

RS

RL

RS 24

OH

O RL

H

H

H

R B

25

RL

RS

RS

RL 26

27

An early success4 was Midland’s Alpine-Borane®, derived from 9-BBN 28 and α-pinene 22. Hydroboration takes place from the less hindered side of the double bond, away from the gem dimethyl groups, to give alpine borane 29. The reagent works well for acetylenic alcohols and the transition state 30 puts the acetylene in the ‘outside’ position. H B

H

reflux

+

B

28; 9-BBN 22; (+)-α-pinene 11 mmol 0.5 molar THF solution

O

R1

THF 25 hours

B

H 29; (+)-Alpine-Borane® 0.5 molar THF solution

30

R2

The general scope of the ketones 31 is large and the results, especially for ee in the alcohol products 32, are good. An example is the simple alcohol 34. This enantiomer is the same as the one obtained from DARVON-H above. OH H

O (+)-Alpine-Borane®

O

OH H

(+)-AlpineBorane®

R1

R1 R2

two days room temperature 5 mmol

31

R2 32; 59-98% yield 77-100% ee

34; 78% yield 99% ee

33

One advantage of this reagent is that both enantiomers of α-pinene 22, the chiral auxiliary, are available from nature. Both enantiomers of Alpine-Borane® are available as THF solutions. However, the quality of natural α-pinene is erratic, and the table shows that you get what you pay for here as elsewhere. The prices are approximate relative prices to that of benzoic acid. Table: Commercially Available α-pinene (1S)-(−)-α-pinene rotation about −51.7 from European conifer oils

(1R)-(+)-α-pinene 22 rotation about +51.7 from North American conifer oils [a]21D

ee

price/100 g

[a]21D

ee

price/100 g

+50.7

97%

240

−50.7

97%

235

91%

22

87%

36

0%

10

81%

8.4

510

24 Asymmetric Induction I Reagent-Based Strategy

Reduction of ketones with Ipc2BCl (DIP-Chloride™) There are many other reducing agents derived from α-pinene and various boranes. It is worth referring to Brown’s articles for details.3 We shall describe just one of these5 as it is rather different both sterically and electronically and seems to be the reagent of choice at the moment. Hydroboration of α-pinene with ClBH2 occurs twice to give Ipc2BCl 35 (strictly B-chloroisopinocampheylborane and available as DIP-Chloride™).

Cl

BCl

BH2

=

=

H 2

a-pinene

a-pinene

22

35

This borane has been used in the preparation of enantiomerically pure prozac,6 (S)Fluoxetine 38. The simple ketone 36 is reduced to give the alcohol 37, easily transformed into prozac 38. F3C OH

O Ipc2BCl

O

Cl

Cl

NHMe

36

38; ( S)-Fluoxetine (Prozac)

37

It is a mistake to dismiss these reagents as expensive academic toys not fit for serious chemistry as Ipc2BCl has been used by Alcon Laboratories7 in the large scale production of their glaucoma treatment Azopt 39. Disconnection of most of the structural C–X bonds reveals the simple heterocycle 40 (note that inversion will occur when the EtNH group in 39 replaces the OH group in 40). The saturated ring could come from cyclisation of 41 and either 40 or 41 could be made by enantioselective reduction of the corresponding ketones.

HN O H2N

Et

O S S

S O

N O

39; AL-4862 (Azopt)

OMe

C–N

Cl note inversion

Br

HO

OH 2 x C–N 1 x C–S

S

S O 40

NH

Cl

S

S

O

NH2

O O 41

In fact they chose to make 41 by asymmetric reduction of the ketone 45 easily made from a thiophene available from Lancaster Synthesis.

24 Asymmetric Reduction of Unsymmetrical Ketones

O

O

O 1. Cl 2

NaSBn

Cl

511

Cl

S

Cl

EtOH

Lancaster Synthesis

SBn

S

Cl

toluene 98% ee OH O

1. BuLi

2. TsCl, Et 3N

S

H2N

2. SO 2

1. MeC(OMe) 3

O S

3. NH 2OSO2OH

S

N

OMe 3. EtNH 2

O O

39; AL-4862 (Azopt) 61% yield, >99% ee 500 g scale

4. 12M HCl

47; 69% yield on 1 kg scale

The functional groups in 45 clearly do not interfere with the reaction but some functional groups have a large effect. Reduction of β-keto-esters 48 gives poor yields and ees. By contrast the free acids 49 give alcohols 50 with high yields (87–92%) and very high ees (91 – ⬎98%). Clearly the CO2H group binds to the boron atom and cooperates with the Ipc group in delivering the H atom to one face of the ketone.8 We shall return to these reagents under ‘asymmetric electrophiles’ when we discuss hydroboration of alkenes.

O

O CO2R2

R1 48

OH

Et3N, (–)-Ipc 2BCl

CO2H

R1 49

CH2Cl2, 0 ˚C, 2 hours

CO2H

R1 50

512

24 Asymmetric Induction I Reagent-Based Strategy

Asymmetric Electrophiles Asymmetric hydroboration of alkenes A hydroborating agent must have a free B–H bond and the obvious choice for asymmetric hydroboration is something derived from borane (BH3 or B2H6) and α-pinene. The Me2S-BH3 complex reacts rapidly once with α-pinene, nearly as rapidly a second time to give good yields of Ipc2BH 51, but fails to react a third time. This is good news in a way as the reagent 51 is easy to prepare but is also a warning that it is unlikely to react with any hindered alkenes.9

Me2S

=

B

BH3

=

H

H

2

α-pinene

α-pinene

51; Ipc 2BH

22

In fact, Ipc2BH 51 reacts well with cis-disubstituted alkenes. Using crystalline Ipc2BH 51 of 99.1% ee, cis butene gave (R)-(⫺)-2-butanol 53 of 98.1% ee in 73% yield and norbornene 54 gave exo-alcohol 55 of 83% ee in 63% yield. In this second case there is diastereoselectivity too. There are other examples in Brown’s reviews.3 Ipc2BH

BIpc2

–25 °C

1. Ipc2BH –25 °C, THF

OH

H2O2

OH

2. H2O2, NaOH

NaOH

THF

(R)-(–)-53

52

54; norbornene

(–)-55

For more hindered alkenes the mono-substituted borane IpcBH2 is better.10 Direct hydroboration of α-pinene cannot be stopped after one hydroboration and gives Ipc2BH 51 in good yield but the hydroboration is reversible. Adding TMEDA to a solution of Ipc2BH 51 in ether precipitates the crystalline complex 56 of two molecules of IpcBH2 and one of TMEDA. The yield is only 80% but if α-pinene of 94% ee is used, 56 crystallises with 100% ee.

Me2S

BH

BH3

Me2N Me2N

H TMEDA

Et2O

NMe2

NMe2

BH2 H

H2B

2

22

51; Ipc 2BH

56; (IpcBH 2)2.TMEDA, m.p. 140-141 ˚C

The reagent itself, IpcBH2 57 is released from the crystalline complex 56 by treatment with BF3 in THF. TMEDA prefers to form a complex with the more electron-deficient BF3: this crystallises leaving pure IpcBH2 57 in solution ready for hydroboration of alkenes. Reactions with simple trans-disubstituted alkenes 58 are reasonable: trans butene gives (⫹)-(R)-2-butanol of 73% ee in 73% yield.11

24 Transfer of Allylic Groups from Boron to Carbon

Et2O

OH

BH2

BF3

H

56

513

R

R

THF

57; IpcBH2

1. 57, THF, –25 °C

R

R

2. H2O2, NaOH

trans (E-)-58

(R)-59

Trisubstituted alkenes - the type that do not react at all with Ipc2BH - perform particularly well. Regioselectivity is usually excellent and the ees are generally better than those with simple trans-disubstituted alkenes. 2-Methylpent-2-ene 60 gives only 58% ee but the champion is 1-phenylcyclopentene 62 that gives 92% yield of the anti alcohol 63 in 100% ee. Brown12 has improved on most of these ees by using modified ‘pinenes’ with larger bridgehead substituents but these ‘pinenes’ have to be prepared and the Ipc reagents retain their popularity. Ph

Ph

1. 57, THF, –25 °C

1. 57, THF, –25 °C

2. H2O2, NaOH

2. H2O2, NaOH

HO 60 2-methylpent-2-ene

OH (S)-(–)-61

(1S,2R)-(+)-63

62

More recently, the effects of functional groups on the alkene have been explored. Enol ethers are more nucleophilic than ordinary alkenes and the two enol ethers E- and Z-64, give good yields of different diastereoisomers of the half-protected diol 65. Though 64 is a trisubstituted alkene, it reacts rapidly with Ipc2BH with good enantioselectivity.13 1. Ipc 2BH –25 ˚C, THF

OMe

E-64

2. H 2O2, NaOH

1. Ipc 2BH –25 ˚C, THF

OH

OMe (2R,3R)-(–)-65 72% yield, >97% ee

MeO

2. H 2O2, NaOH

Z-64

OH

OMe (2R,3S)-(+)-65 77% yield, 90% e

Asymmetric Nucleophilic Attack This commoner type of reaction involves the attack of carbon or heteroatom nucleophiles onto carbonyl compounds, by direct or conjugate addition, and onto imines. Because we have just dealt with boranes we shall start with the reaction of allylic boron compounds with such electrophiles. You might strictly not call this nucleophilic attack.

Transfer of Allylic Groups from Boron to Carbon Reactions of chiral allylic boranes with carbonyl compounds Simple alkyl boranes do not react with aldehydes and ketones but allyl boranes react rapidly because they can use a six-membered cyclic transition state 68 not unlike that of the Claisen rearrangement. The evidence for this is that the allyl group is rearranged in the product. In this example a crotyl borane 66 adds to an aldehyde to give anti-70 since both R and Me prefer to be equatorial in the transition state 68.

514

24 Asymmetric Induction I Reagent-Based Strategy

R O BR2

R R

B O

RCHO

66 a crotyl borane

OH

B

R R

BR2

O

R

R R

Me 68

67

69

70

Allyl dialkyl boranes do this reaction, transferring only the allyl group as the mechanism 67 requires, so asymmetry can be introduced by replacing BR2 with Ipc2B to give 71. In this specific instance, the reaction with acetaldehyde gives predominantly just one enantiomer (2R,3S) of the anti-diastereoisomer of 72 accompanied by some of the other anti enantiomer and traces of the syn-diastereoisomer.3 The diastereoselectivity is therefore 99:1 and the ee 90%. OH

OH BIpc2

OH

OH

MeCHO

71 Ipc2 crotyl borane

(2R,3S)-72; 94.5% (2S,3R)-72; 4.5% (2R,3R)-72; 1% (2S,3S)-72; trace

The allylic boranes can be made from allylic Grignard or lithium reagents with Ipc2BCl 35 (used above in reductions). An interesting example is Barrett’s synthesis14 of the C2-symmetric compounds 76. The dilithium derivative 74 reacts with two molecules of Ipc2BCl 35 to give the C2-symmetric allylic double borane 75, both Ipc groups having the same chirality. Reaction with an aldehyde gives the C2-symmetric diols 76. BuLi TMEDA hexane

73

Ipc2BCl

Li

Li

Et2O

OH

RCHO

Ipc2B

OH

BIpc2 R

74

75

R 76

Various aliphatic and aromatic aldehydes can be used and either enantiomer of the metric diastereoisomer 76 can be made depending on the enantiomer of Ipc2BCl used. Two allylic transfers occur, each with same sense of asymmetric induction and this is easier to see if 75 is redrawn and the intermediate 77 also redrawn before reaction. The ratio of the C2 to the meso diastereoisomers is usually greater than 90:10 and the ees uniformly ⬎95%. OH

BIpc2 RCHO

IpC2B

Ipc2B

Ipc2B R

OH

RCHO

=

76

R 75

77

77

Reactions of chiral allyl boranes with imines Boranes other than those based on α-pinene are particularly useful in allylic transfer to imines to make single enantiomers of unsaturated amines. One good combination is an allylboron compound 80 complexed with an N-sulfonyl amino alcohol such as 78, derived from nor-ephedrine (see chapter 23) with an N-silyl imine such as 81. The unsaturated amines 82 are formed in good

24 Asymmetric Addition of Carbon Nucleophiles to Ketones

515

yields15 with ees around 90%. Brown has shown16 that the water content of such reactions is critical - no reaction occurs in the complete absence of water! Ph

Ph

OH +

Me

Et3N

B

THF

3

NHTs 78

O B

Me

79

N Ts 80

SiMe3

N

+ R

H

NH2 HCl workup

R

81

82

Asymmetric Addition of Carbon Nucleophiles to Ketones The chapters on asymmetric catalysis will describe catalytic versions of the reactions we are about to describe. Catalysis is a fundamentally more efficient way of asymmetric induction and we shall restrict our discussion here to cases where catalysis proved ineffective.

Addition of alkyl lithiums to ketones Merck’s HIV-1 reverse transcriptase inhibitor Efavirenz 83 is one of the simplest anti-HIV drugs yet produced. The most straightforward synthesis17 is based on closure of the amino alcohol 84 with some phosgene equivalent and the preparation of 84 by asymmetric addition of cyclopropylethynyl-lithium 86 to the ketone 85.

F3C

C–N C–O

Cl

O N H

CF3

F3C Cl

C–C

OH

O

Cl

O

NH2

83

Li

NH2

84

85

86

It was already known that amino alcohols of the kind we have just used 78 were good at this kind of asymmetric addition but this particular combination of an acetylenic nucleophile and an aryl trifluoromethyl ketone was uncharted territory. After some exploration, stoichiometric pyrrolidine alcohol 88 prepared by alkylation of norephedrine 87 from the chiral pool (chapter 23) proved the best and the ketone 85 had to be used as its N-4-methoxybenzyl derivative18 89. CF3 HO Ph

NH2 Me

Br

Br

NaHCO3, toluene

87 (1R,2S)-norephedrine

HO

N

Ph

Me

88; 97% yield

Cl

O

N Ar H 89; Ar = 4-MeOC6H4

The conversion of ligand 88 into the reagent 91 required the specific conditions set out in the chart. Titration of BuLi against Ph3CH produced the lithium alkoxide 90 with enough BuLi left over to make the lithium derivative 86 and the reagent 91 turned out to be a tetramer.18

516

24 Asymmetric Induction I Reagent-Based Strategy

88 3.31 kg

1. 1.6M BuLi in hexane added until Ph3CH went red 2. 1.6M BuLi in hexane add same amount again

LiO

N

Ph

H

Ph

O

99.5% ee

2. NaOH MeOH NaBH4

93

3. COCl 2 Et3N toluene

F3C Cl

N Ar H H

or 3. ClCO2Me heat

OH

F3C Cl

NH2

O N O H 83; 83-95% yield

94; 94% yield

This one example reveals that it is usually necessary to find the right ligand for any direct addition of RLi or RMgX to a carbonyl group. If you are lucky you may find a catalytic method. If not, an amino-alcohol based stoichiometric method should be possible.

Asymmetric epoxidation with chiral sulfur ylids A very different type of chemistry occurs when sulfur ylids add to carbonyl compounds. Epoxides are formed and recent progress with chiral sulfur ylids allows good asymmetric induction in this reaction. The easily prepared C2 symmetric sulfide 96 reacts with alkyl halides and then with aryl and alkyl aldehydes to give good yields of trans epoxides 97 with reasonable ees.19 OH 1. MsCl, Et 3N 2. Na 2S, EtOH

Br

Me

S

Me

OH 95

96; 95% yield

Ph

PhCHO, NaOH t-BuOH/H2O

Ph O

Ph

97; 92% yield 88% ee

24 Asymmetric Nucleophilic Attack by Chiral Alcohols

517

The alkylation of the sulfide 96, the formation of the ylid 99, and the reaction with the aldehyde are all carried out in one operation. The sulfide is a good nucleophile for alkyl halides and forms the sulfonium salt 98. This gives the ylid 99 with NaOH as a convenient base in aqueous tertbutanol. The ylid selectively attacks the aldehyde to give the betaine 100 that closes to the epoxide and releases the sulfide 96 for the next round.

Br

Ph

Me

96

Me

S

NaOH

Me

Ph

Me

S

Me Ph

Ph Ph

O

96 + 97

Ph O

99 ylid

98 sulfonium salt

Me

S

100

Enolisable aldehydes such as 101 or 103 do not give quite such good yields but the ees are still good and the diastereoselectivity in favour of the trans epoxides 102 and 104 is excellent. The secret of this method is the simple preparation of the reagent 96. In the next chapters you will see that superior catalytic methods are available for asymmetric epoxidation of allylic alcohols and of cis-alkenes but they are less good for the trans disubstituted alkenes that would give 97, 102, or 104. You will also see catalytic versions of sulfur ylid epoxidation.

CHO CHO 101

O

Ph

O 102; 75% trans epoxide 94% ee

Ph

104; 70% trans epoxide 93% ee

103

Asymmetric Nucleophilic Attack by Chiral Alcohols Deracemisation of arylpropionic acids Arylpropionic acids such as ibuprofen 105 are important NSAIs (non-steroidal anti-inflammatories). Only one enantiomer is active and some are administered as enantiomerically pure compounds through there is a problem with racemisation in the body by enolisation. This can be turned to advantage in ‘deracemisation’. Weak bases are enough to convert the acid chloride 106 into an enolate that eliminates 107 to form the achiral ketene 108. Addition of, say, ethanol then gives racemic esters 109; R ⫽ Et of ibuprofen. H

H

CO2H

SOCl2 DMF

COCl

heptane

105; racemic Ibuprofen

O

EtNMe2 heptane

106

ROH

C

O

H CO2R

Cl 107; achiral enolate

108; achiral ketene

109; racemic ester of ibuprofen

518

24 Asymmetric Induction I Reagent-Based Strategy

Asymmetric induction would in theory be possible if a chiral alcohol were used in the last step20 and, amazingly, providing α-hydroxy-esters or lactones are used, can actually be realised.21 The elimination is followed by infra red until the ketene is completely formed (2,100 cm⫺1) and then the asymmetric oxygen nucleophile is added in heptane at ⫺78 ⬚C. The best reagents are (S)(⫺)-ethyl lactate and (R)-(⫹)-pantolactone 112 from the chiral pool - they give opposite enantiomers with excellent ees. In this case R/S does give a good indication of chirality as the functional groups are virtually the same. The induction takes place on the protonation of the enolate rather than the addition of the alcohol. Ethyl lactate gives the ester 111 in a 97:3 ratio of diastereoisomers that becomes 89% ee on hydrolysis to 105. Even under the very mild hydrolysis conditions shown, some racemisation occurs during hydrolysis. H

OH

O EtNMe2

106 heptane

CO2Et

CO2Et

108 detected by IR

O

heptane –78 ˚C

H

O

proton transfer

110 CO2Et

O

H

HOAc, 2M HCl H2O, 85 ˚C

O

or LiOH, H 2O MeCN, 5 ˚C

111; 92% yield; 97:3 diastereoisomers

OH

(R)-105; 89% yield; 89% ee

The other enantiomer (S)-105 comes from reaction with (R)-(⫹)-pantolactone 112 in rather better ee - evidently the hydrolysis of the ester 113 is faster and occurs without racemisation. In both cases, rather precise conditions are necessary. This should be seen as both a warning and an encouragement. Conditions for asymmetric induction by attack of homochiral alcohols on ketenes Same solvent and catalyst used for synthesis and reaction of ketene: Solvent: Heptane best, toluene all right, polar solvents no good. Catalyst: Me3N and N-Me-pyrollidine best, Et3N and i-Pr2NEt all right. Temperature unimportant: only 7% decrease in ee at room temperature cf. ⫺78 ⬚C. Dilution critical: 90% ee. at 1M; 97.2 ee at 0.02M. Work-up: treat with aqueous AcOH to cleave any anhydride formed. H

OH EtNMe2

106 heptane

108 detected + by IR

O O

heptane

O

O O O

(S)-105 89% yield 99% ee

–78 ˚C

112; pantolactone

113

Deracemisation of α–halo acids The same reagent (R)-(⫹)-pantolactone 112 gives good results with α-halo acids that are good precursors for many other compounds such as α-amino acids by nucleophilic displacement. The α-halo acid chlorides 115 are prepared directly from the simple acids 114 and treated with the same tertiary amine EtNMe2 used above to give the unstable ketenes 116 and hence the esters 117 by reaction with (R)-(⫹)-pantolactone22 112.

24 Asymmetric Conjugate Addition of Nitrogen Nucleophiles

519

X X

PCl5

R

X

EtNMe2

C

CO2H

X2

R

114

COCl

O O

112

R

O

116

115

O

O

R

117

It is best not to isolate the ketene but to add a solution of the acid chloride 115 in THF to a mixture of (R)-(⫹)-pantolactone 112 and EtNMe2 also in THF at ⫺78 ⬚C. R is always alkyl and X can be Br or I. The esters 117 are isolated as mixtures (7:1 to 19:1) of diastereoisomers: two examples follow. I

X

OH R

O

COCl

O O

O

O

EtNMe2

O

Br

O O

O

O

THF, –78 °C

117a; 19:1, 73% yield

112

117b; 8:1, 78% yield

It is essential that the ketene is formed: reaction of the acid chlorides 115 with (R)-(⫹)pantolactone 112 also gives the esters 117 but as a 1:1 mixture of diastereoisomers. A useful application is the preparation of unusual amino acids such as 120 by SN2 displacement of halide with azide and reduction of the azido group. It is best to hydrolyse the ester at the azide stage 118 to reduce racemisation but even so some does occur both in the azide displacement and in the hydrolysis. Br O

112

COCl

EtNMe2

115c

O O

NaN3

O

O

THF, –78 ˚C

N3

O

Br

DMF 60 ˚C

117c; 14:1, 79% yield

O

O 118; 7:1, 70% yield

N3

NH2

LiOH

OH

OH

THF/H2O

O

O

119; 70% yield, 72% ee

120

Asymmetric Conjugate Addition of Nitrogen Nucleophiles The work of Steven Davies23 at Oxford has made conjugate addition of nitrogen nucleophiles a reliable and useful reaction. The essential reagent is the amine 121, prepared on a large scale by resolution (chapter 22). This amine does not react with unsaturated esters like 123 but the lithium derivative 122 adds in a conjugate fashion to give the ester 125 via the enolate 124. Ph

Ph

Ph

N H

THF –78 ˚C

Ph

N Li

121

122

Ph

CO2t-Bu

Ph

BuLi

123

Ph H

Ph

OLi

N

Ot-Bu

Ph 124

H2O

Ph

N

CO2t-Bu Ph 125; 92% yield; 95% de

520

24 Asymmetric Induction I Reagent-Based Strategy

Here at last some explanation can be offered. The lithium derivative 122 forms a Li–O bond to the carbonyl group of the ester 123 in its best ‘s-Z’ conformation and the nitrogen atom adds to the conjugate position ‘like a butterfly settling on a leaf’ 126 to give the lithium enolate 124 in the same conformation 124a.

H

Li

N

O

122 + 123

Ph Ph

Ph Ph H H

O H 126

N

O

Me Ph

LiO

H

124

Me Ph

H 124a

The two benzyl groups on the nitrogen atom can be removed by hydrogenation – remarkably the third benzylic bond to nitrogen is not cleaved - and the t-butyl ester hydrolysed in acid solution to give ‘β-phenylalanine’ 128 in good yield and ee. The reagent 122 has supplied just one nitrogen atom to this final product 128 and no recycling of 121 is possible. Fortunately 121 is very easy to make. NH2

H2, Pd/C

O

O

NH2

CF3CO2H

125 Ph

HOAc

O

Ph

127; 84% yield

OH

128; 100% yield; 95% ee

An asymmetric synthesis of thienamycin In this sequence valuable chemical information is discarded. Not only is 124 a specific enolate, it has a specific geometry (‘Z’ lithium enolate). This information can be used in a tandem style of reaction described in chapter 36. We shall continue with an alternative. Replacing the benzyl group in 121 with an allyl group 129 and adding to a dienoate ester (t-butyl sorbate) gives the product 131 from regioselective addition24 at C-3. Ph H

1. BuLi

N

Li

Ph

CO2t-Bu

1.

N

Ph

N

Me

2. NH 4Cl, H 2O

t-BuO 129

O

131; 87% yield, 99:1 diasts

130

This ester 131 forms the usual ‘E’-enolate with LDA that can be trapped as the boron enolate 132. Reaction with acetaldehyde gives, as expected, an anti aldol (chapter 4). The major product is 133 in a 91:9 ratio with the other anti aldol. This diastereoisomer can be isolated in 75% yield. The absolute stereochemistry is decided by the chiral centre already in place in 132 - it remains as C-4 in 133. The next centre along (C-3) is controlled by C-4 and the relative stereochemistry of C-3 and C-4 is controlled by the geometry of the enolate. Ph

OH

1. LDA

Me

131

N

2. B(OMe) 3

(MeO)2B

Me

OR

132; ' E' boron enolate

CHO

Me

Ph 3

2

t-BuO2C

4

N

H

133; 75% yield

Me

24 Asymmetric Deprotonation with Chiral Bases

521

This compound was used in an asymmetric synthesis25 of the antibiotic thienamycin 138. The N-allyl group was removed with Pd(0) catalysis to give 134 and the t-butyl ester hydrolysed to give the redrawn acid 135. Cyclisation by the pyridyldisulfide reagent gave the βlactam core 137 of the antibiotic 138. OH

(Ph3P)4Pd

Ph

OH

CF3CO2H

133 Me

136

H

t-BuO2C

N H

Me

PyrS-SPyr

N H

Me

OH H

Me

H

N

N

H

Me

NH2

H S

N

Ph

N O 136

Me

Me

Ph3P

Me

O

CO2H 135; 98% yield

134 OH

O

Ph

H

O

O

CO2H

Me 137; 98% yield

138; (+)-thienamycin

Asymmetric Protonation Enantioselective lactonisation of achiral hydroxy diacids Reduction of the achiral ketodiester 139 gives the racemic lactone 140. Hydrolysis of both ester groups then gives the again achiral hydroxydicarboxylate 141. This compound is prochiral and the − two CO2 groups are enantiotopic. If one could be protonated selectively by an enantiomerically pure acid, one enantiomer of the monoacid would be formed. This sounds like an improbable event. O O EtO2C

NaBH4

O

OH

NaOH

CO2Et

CO2Et

O2C

CO2

(±)-140

139

2Na

141

As you may guess, it does work. Camphor sulfonic acid 142 (used in chapter 22 as a resolving agent) protonates just one of the carboxylate anions and cyclisation then gives the lactone26 144 in remarkably high ee (94%). The two carbonyl groups are now different enough to be developed separately. O OH

141

O SO2OH 142; (+)-camphor sulfonic acid

HO2C

Na CO2

143

HCl

O

EtOH

CO2H 144; 94% ee

Asymmetric Deprotonation with Chiral Bases A more widely used method is deprotonation to make, say, an enolate, with one enantiomer of a base. The base must form a close association with the substrate and that usually means a lithium

522

24 Asymmetric Induction I Reagent-Based Strategy

amide. You may object that a planar enolate is not intrinsically chiral, but desymmetrisation of a symmetrical ketone is ideal. Simpkins’ synthesis of anatoxin 145 is a good example.27 This highly toxic compound (known as ‘sudden death factor’) is the model for potentially useful pain killers. Simpkins wanted to make it from readily available tropinone derivatives 147. O

O O

HN

HN

HN

?

R

?

N O

145 (+)-anatoxin-a

146

145a

147

A lot of work has to be done. A carbon atom needs to be added and the ketone moved round to allow addition of a d1 reagent. But this is an attractive strategy because tropinone is achiral and enolate formation desymmetrises the molecule. It was not easy to find the right base. In the end the C2-symmetric double base 148 gave the silyl enol ether 149 in good yield, and, as we shall see later, good ee. MeO2C

Ph

Ph

MeO2C

N

Me3SiCl

+ O

N Ph

147; R = CO 2Me

N

N Li Li 148

OSiMe3

Ph

149; 90% yield

Clearly the lithium enolate 150 must be formed asymmetrically and is trapped by Me3SiCl to give the silyl enol ether 149. Simmons-Smith-style cyclopropanation adds the extra carbon and the product 151 is cleaved with Fe(III) to give the enone 153. This is not the same as 146 but has an electrophilic carbon in the right place. The enone 153 is crystalline and can be obtained optically pure for conversion to a single enantiomer of anatoxin 145. Further details are in the workbook. MeO2C

MeO2C N

148

Me3SiCl

147

OLi

149 90% yield

Et2Zn ClCH2I

151; 99% yield N

NaOAc

N

HN DMF MeOH

DMF MeOH

O 152

O

MeO2C

Cl

MeO2C

OSiMe3

ClCH2CH2Cl

150

FeCl3

N

O 153; 71% yield 99% e.e. after recrystallisation

145

Asymmetric deprotonation with sparteine By far the most popular asymmetric deprotonation uses sparteine as a ligand for lithium.28 Sparteine 154 is a bicyclic alkaloid from lupins and gorse and has two nitrogen atoms that can, in one favourable conformation 154a, chelate a lithium atom. In spite of appearances, it is not (quite) C2-symmetric and only one enantiomer is available.

24 Asymmetric Oxidation

523

H H

N N

N

N

N

N

H H 154; (–)-sparteine

154a; chelating

154b; non-chelating

A simple application is the asymmetric alkylation of carbamates 155 of cyclic amines. Removal of one of the diastereotopic protons from 155 gives a more stable sparteine chelate 156. This is not an enolate but has a C–Li bond stabilised by coordination to the carbonyl group. Alkylation occurs with retention of configuration to give, after hydrolysis, one enantiomer of the secondary alcohol28 158.

Cby O N

H

N

H

O

Bu

O

sec–BuLi sparteine

N

155

N Li H

O

MeI

Bu

156 Me

H

hydrolysis

CbyO

Me

H

HO 158: 93% yield, 98% ee

157; 87% yield

Hoppe has extended this work to d3 reagents 159 (homoenolates - see chapter 13 for achiral versions) by the addition of a double bond.29 Lithiation occurs at C-1 by removal of one of the enantiotopic protons at C-3. Aldol reaction with acetone occurs at C-3 of the complex 160 as expected for a homoenolate (chapter 13) giving a single enantiomer of the homoaldol product 161. All these reactions use an excess of sparteine. O

i-Pr2N

O

i-Pr2N O

H

H

O O

N Sp Li N Me2CO

n-BuLi

1

3

i-Pr2N

O

OH H

(–)-sparteine

159; Z:E 92:8

160: Sp = sparteine

161; 69% yield, 97% ee

Asymmetric Oxidation The most important example of stoichiometric asymmetric oxidation is probably the titaniumcatalysed conversion of sulfides into sulfoxides by cumene hydroperoxide in the presence of stoichiometric diethyl tartrate. A simple example is the efficient asymmetric synthesis of methyl p-tolyl sulfoxide 162, an important starting material for much sulfoxide-controlled asymmetric synthesis.30

524

24 Asymmetric Induction I Reagent-Based Strategy

O S

S

Ti(OiPr)4, CH 2Cl2, H 2O

Me

Me

OH OH

CO2Et

EtO2C

OH (S,S)-(–)-DET

O

cumene hydroperoxide

162; 68% yield 99.5% ee

The very important anti-ulcer drug omeprazole is a sulfoxide and is invariably made by oxidation31 of the sulfide 163. AstraZeneca workers have reported32 that the same mixture gives the enantiomerically pure drug 164, to be marketed as esomeprazole, in good ee. An important difference is that the asymmetry comes from only 60 mol% of tartrate and they were able to get 91% ee with only 4% tartrate. We are entering the more efficient realms of catalysis - the subject of the next two chapters. OMe

OMe cumene hydroperoxide

N

S

N

Ti(OiPr)4, H 2O

HN

O

60 mol% (S,S)-(–)-DET

OMe

163

i-Pr2NEt toluene

N

S

N HN

OMe

164; esomeprazole, >99.5% ee

References General references are given on page 893 1. R. S. Brinkmayer and V. M. Kapoor, J. Am. Chem. Soc., 1977, 99, 8339. 2. R. Noyori, I. Tomino, M. Yamad and M. Nishizawa, J. Am. Chem. Soc., 1994, 106, 6709; 6717. 3. H. C. Brown and P. V. Ramachandran, Pure Appl. Chem., 1991, 63, 307; Acc. Chem. Res., 1992, 25, 16. 4. M. M. Midland, D. C. McDowell, R. L. Hatch and A. Tramontano, J. Am. Chem. Soc., 1980, 102, 867; M. M. Midland, Chem. Revs., 1989, 89, 1553. 5. H. C. Brown, J. Chandrasekharan and P. V. Ramachandran, J. Am. Chem. Soc., 1988, 110, 1539; P. V. Ramachandran, A. V. Teodorovic, M. V. Rangaishenvi and H. C. Brown, J. Org. Chem., 1992, 57, 2379. 6. D. W. Robertson, J. H. Krushinski, R. W. Fuller and J. D. Leander, J. Med. Chem., 1988, 31, 1241. 7. R. E. Conrow, W. D. Dean, P. W. Zinke, M. E. Deason, S. J. Sproull, A. P. Dantanarayana and M. T. DuPriest, Org. Process Res. Dev., 1999, 3, 114. 8. Z. Wang, C. Zhao, M. E. Pierce and J. M. Fortunak, Tetrahedron: Asymmetry, 1999, 10, 225. 9. H. C. Brown, M. C. Desai and P. K. Jadhav, J. Org. Chem., 1982, 47, 5065. 10. H. C. Brown, A. K. Mandal, N. M. Yoon, B. Singaram, J. R. Schwier and P. K. Jadhav, J. Org. Chem., 1982, 47, 5069; H. C. Brown and B. Singaram, J. Am. Chem. Soc., 1984, 106, 1797. 11. H. C. Brown, P. K. Jadhav and A. K. Mandal, J. Org. Chem., 1982, 47, 5074. 12. H. C. Brown and U. P. Dhokte, J. Org. Chem., 1994, 59, 2365. 13. D. Murali, B. Singaram and H. C. Brown, Tetrahedron: Asymmetry, 2000, 11, 4831. 14. A. G. M. Barrett, D. C. Braddock and P. D. de Koning, Chem. Commun., 1999, 459. 15. S. Itsuno, K. Watanabe, T. Matsumoto, S. Kuroda, A. Yokoi and A. El-Shehawy, J. Chem. Soc., Perkin Trans. 1, 1999, 2011. 16. G. Chen, P. Ramachandran and H. C. Brown, Angew. Chem., Int. Ed., 1999, 38, 825. 17. M. E. Pierce, R. L. Parsons, L. A. Radesca, Y. S. Lo, S. Silverman, J. R. Moore, Q. Islam, A. Choudhury, J. M. D. Fortunak, D. Nguyen, C. Luo, S. J. Morgan, W. P. Davis, P. N. Confalone, C. Chen, R. D. Tillyer,

24 References

525

L. Frey, L. Tan, F. Xu, D. Zhao, A. S. Thompson, E. G. Corley, E. J. J. Grabowski, R. Reamer and P. J. Reider, J. Org. Chem., 1998, 63, 8536. 18. A. Thompson, E. G. Corley, M. F. Huntington, E. J. J. Grabowski, J. F. Remenar and D. B. Collum, J. Am. Chem. Soc., 1998, 120, 2028. 19. K. Julienne, P. Metzner, V. Henryon and A. Greiner, J. Org. Chem., 1998, 63, 4532; K. Julienne, P. Metzner and V. Henryon, J. Chem. Soc., Perkin Trans. 1, 1999, 731. See also C. L. Winn, B. R. Bellenie and J. M. Goodman, Tetrahedron Lett., 2002, 43, 5427. 20. C. Fehr, Angew. Chem., Int. Ed., 1996, 35, 2566. 21. R. D. Larsen, E. G. Corley, P. Davis, P. J. Reider, and E. J. J. Grabowski, J. Am. Chem. Soc., 1989, 111, 7650. 22. T. Durst and K. Koh, Tetrahedron Lett., 1992, 33, 6799. 23. S. G. Davies, N. M. Garrido, O. Ichihara and I. A. S. Walters, J. Chem. Soc., Chem. Commun., 1993, 1153; S. G. Davies and D. R. Fenwick, J. Chem. Soc., Chem. Commun., 1995, 1109. 24. N. Asao, N. Tsukada and Y. Yamamoto, J. Chem. Soc., Chem. Commun., 1993, 1660; N. Asao, T. Uyechara and Y. Yamamoto, Tetrahedron, 1990, 46, 4563. 25. S. G. Davies, C. J. R. Hedgecock and J. M. McKenna, Tetrahedron: Asymmetry, 1995, 6, 827; 2507; S. G. Davies and D. R. Fenwick, Chem. Commun., 1997, 565. 26. K. Fuji, M. Node, S. Terada, M. Murata, H. Nagasawa, T. Taga and K. Machida, J. Am. Chem. Soc., 1985, 107, 6404. 27. N. J. Newcombe and N. S. Simpkins, J. Chem. Soc., Chem. Commun., 1995, 831. 28. D. Hoppe, F. Hintze, P. Tebber, M. Paetow, H. Ahrens, J. Schwerdtfeger, P. Sommerfeld, J. Haller, W. Guarneri, S. Kolczewski, T. Hense and I. Hoppe, Pure Appl. Chem., 1994, 66, 1479; Clayden, Lithium, pages 226–235 and 258–269. 29. M. Seppi, R. Kalkofen, J. Reupohl, R. Fröhlich and D. Hoppe, Angew. Chem., Int. Ed., 2004, 43, 1423. 30. S. H. Zhao, O. Samuel and H. B Kagan, Org. Synth., 1989, 68, 49. 31. Saunders, Top Drugs, chapter 5, page 37. 32. H. Cotton, T. Elebring, M. Larsson, L. Li, H. Sörensen and S. von Unge, Tetrahedron: Asymmetry, 2000, 11, 3819.

25

Asymmetric Induction II Asymmetric Catalysis: Formation of C–O and C–N Bonds Introduction: Catalytic methods of asymmetric induction PART I – SHARPLESS ASYMMETRIC EPOXIDATION The AE Method The ligands The catalyst Catalyst structure The mnemonic device The synthesis of propranolol Modification after Sharpless Epoxidation Oxidation after Sharpless epoxidation The Payne rearrangement Asymmetric Induction at the Allylic Alcohol Centre: AE is anti-Selective No Asymmetric Induction from Remote Allylic Alcohol Centre: Reagent Control Asymmetric Synthesis of Diltiazem Summary of Sharpless Epoxidation PART II – SHARPLESS ASYMMETRIC DIHYDROXYLATION The AD Method The ligand Solvent dependence The catalytic cycle Methanesulfonamide Substrate Dependence and the Mnemonic Device Recommended ligands for different classes of alkenes Applications of the Sharpless AD Reaction The synthesis of indicine The synthesis of a C2-symmetric piperidine The synthesis of the alkaloid reticuline Dihydroxylating Compounds with More than One Double Bond I—Regioselectivity Dienes Guidelines When guidelines combine or compete A synthesis of aspicillin Designer ligands to improve selectivity Dihydroxylating Compounds with More than One Double Bond II—Diastereoselectivity Advanced dihydroxylation strategy Scaling Up the Asymmetric Dihydroxylation

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

528

25 Asymmetric Induction II Asymmetric Catalysis

PART III – AMINOHYDROXYLATION PART IV – CONVERTING 1,2-DIOLS INTO EPOXIDES Applications of the AD to epoxide transformation: propranolol and diltiazem PART V – JACOBSEN EPOXIDATION PART VI – DESYMMETRISATION REACTIONS Opening anhydrides Opening epoxides Comparison of desymmetrisation and kinetic resolutions Mono esterification of diols PART VII – HETERO DIELS-ALDER REACTIONS

Introduction: Catalytic methods of asymmetric induction In the previous chapter we discussed various reagent-based strategies for the enantioselective formation of chiral centres. The next two chapters are also concerned with reagent-directed strategy but with an important difference. The source of asymmetry is used in a much lower concentration: it is a catalyst. We may put in a few milligrams of an optically active compound and get out grams or even kilograms of enantiomerically pure product. This chapter concerns the formation of C–O and C–N bonds, the next the formation of C–H and C–C bonds. We dedicate a large part of this chapter to two very important, and extraordinarily useful, enantioselective methods – catalytic asymmetric epoxidation (AE) and catalytic asymmetric dihydroxylation (AD). Impressively, both these methods were developed by Professor Barry Sharpless’s research group and are therefore often referred to as the Sharpless epoxidation and the Sharpless dihydroxylation. Both are examples of ligand-accelerated catalysis.

PART I – SHARPLESS ASYMMETRIC EPOXIDATION There are many asymmetric epoxidation reactions but none is as reliable or has been as widely used as the Sharpless Asymmetric Epoxidation1 (AE), the subject of this section, and the Jacobsen epoxidation, the subject of a later section in this chapter.2

The AE method The Sharpless epoxidation converts a prochiral allylic alcohol 1 into a chiral epoxy alcohol 2 and gives that epoxy alcohol with a very high enantiomeric excess.

O R

OH 1

R

OH 2

There is a great deal of flexibility allowed within the framework of this reaction – the substituents on the starting material can vary a great deal and the double bond can be E or Z. However, there is one restriction and that is that the substrate must be an allylic alcohol. If it isn’t the reaction will not work because, as we shall see, the hydroxyl group plays a crucial role in the reaction. The reagents used are an oxidising agent (usually, tert-butyl hydroperoxide t-BuOOH, but other

25 The AE method

529

hydroperoxides can be used), a catalyst, and – for the asymmetric induction – a chiral and enantiomerically pure ligand. There is an Organic Synthesis procedure for a typical case.3

The ligands The ligands used are the dialkyl tartrates. Most commonly, the diethyl or diisopropyl tartrates are used. Diethyl tartrate is often abbreviated to DET, diisopropyl tartrate to DIPT and dimethyl tartrate to DMT. OH

O

OH

HO

OH O

O

EtO

OH

OEt O

3; ( R,R) L-(+)-tartaric acid

OH

OH

OH

O

OEt

OEt O

O

OH

5; ( S,S) D-(–)-DET

4; ( R,R) L-(+)-DET

O

EtO

EtO

OH

6; meso-DET

Importantly, both enantiomers of tartaric acid are available. However (and also importantly) one is very much more expensive than the other. The cheap one is L-tartaric acid 3 and is often referred to as ‘natural’ tartaric acid. It is about 35 times cheaper than the ‘unnatural’ D-tartaric acid though, in fact, there is nothing unnatural about this acid at all – it too comes from natural sources. The ethyl 4, 5 and iso-propyl esters of both enantiomers are also available. These are the enantiomers of C2 symmetric DET, the other diastereoisomer, meso-DET 6 is achiral. The ligand can be used catalytically providing that molecular sieves are added.4

The catalyst The metal at the heart of the active catalyst is titanium. This holds everything together—the tartrate ligand, the substrate (the allylic alcohol) and the oxidant tert-butyl hydroperoxide t-BuOOH. The structure of the catalyst before substrate or ligand binding is thought to be as shown in the figure below. There are two titanium atoms and two unsymmetrically bound tartrate units.

Catalyst structure The most likely structure of the catalyst is based on X-ray structures of related compounds and reinforced by calculations.5 It is dimeric with two Ti atoms, two tartrates and four alkoxides. We show one of two degenerate structures. When the catalyst combines with the allylic alcohol and the hydroperoxide, the same titanium atom takes both and only one diastereotopic face of the alkene is presented to the active oxygen atom. RO2C OR E RO O

Ti

O

RO2C

OR Ti

O

O OR

O O RO Likely catalyst structure E = CO2R

OR

OR E R

OH

t-BuOOH

RO O

Ti

O

O Ti

O O

O

O E O

R

t-Bu RO Proposed 'loaded' catalyst structure E = CO2R

530

25 Asymmetric Induction II Asymmetric Catalysis

The mnemonic device The absolute configuration of the epoxide product depends upon the enantiomer of the tartrate ligand that is used. The empirical observation is very reliable—when the allylic alcohol is drawn with the carbinol carbon in the SE quadrant as in the figure below, the L-tartrate directs attack of the oxygen atom from the oxidising agent t-BuOOH to the bottom face of the olefin while the Dtartrate directs it to the top. D-(–)-dialkyl

tartrate

O R

OH 2

NE

NW

R3

R2

OH R1 O SE

SW L-(+)-dialkyl

R

OH

ent-2

tartrate

The synthesis of propranolol Propranolol 7 is a β-blocker. Let’s consider how it could be disconnected. The first thing to notice is that there are two 1,2-related functional groups. We have an ether and an alcohol on the lefthand side of the molecule and, on the right, a 1,2-aminoalcohol. 2

1

O 7; (–)-propranolol

OH

2 N H

Because we have an alcohol in the middle of the molecule, we can imagine disconnections 7a or 7b that rely upon the attack of a nucleophile on the terminal position of an epoxide. Notice that there is only one chiral centre in this molecule. a OH

N H

O 9

8

a

b

O OH

N H

7a; disconnection of (–)-propranolol

b

O

H2N O

11

10

Either disconnection a or b would be fine. However, the way it was done6 involved disconnection b. The next disconnection is akin to a in that it uses α-naphthol 8 as the nucleophile.

25 Modification after Sharpless Epoxidation

O

OH

O

531

OMs

C–O

Ms = MeSO2

+ ether

O 8

10a

9

Mesylate was chosen as the leaving group and the epoxy mesylate 9 could be made from the corresponding alcohol 12 and that, in turn, from allyl alcohol 13 by Sharpless epoxidation. O O Me

S

C–O

AE

HO

O

O

9a

ester

HO

O 12

13

In fact, allyl alcohol 13 was not used at the start of this synthesis because low ees result if there is no substituent on the alkene trans to the alcohol (R2 in the mnemonic should be something other than H). It is much better here to add a removable group and so the silyl alkene 14 was used instead of 13. The silyl group allowed the isolation of the epoxy alcohol 15 in a much higher yield than was possible using allyl alcohol itself and the ee of 15 was ⬎95%. The alcohol was converted to a mesylate 16. (–)-DET

HO

HO

SiMe3 t-BuOOH

O

SiMe3

MsCl

MsO

15; 60% yield, >95% ee

14

O

Et3N

SiMe3

16

The mesylate was then displaced using sodium naphthoxide made from 8 with base. The silyl group is not wanted in the final product and was removed with TBAF (tetrabutylammonium fluoride, Bu4N⫹ F⫺). Finally the epoxide 10 was opened with isopropylamine 11. This synthesis depends on regioselective opening of an epoxide at the less substituted end.

8

16 base

Bu4N

O

O

11

F

SiMe3

O

17

7 O

10

Modification after Sharpless Epoxidation Oxidation after Sharpless epoxidation Disparlure 18 is the sex attractant of the gypsy moth. The only functional group in disparlure is the epoxide and we need an alcohol somewhere if the stereoselective introduction of this group is to be done by AE. Our disconnection strategy7 is dominated by that thought – an alcohol must be introduced somehow at one of the allylic carbons in 19. OH? AE?

O 18; (+)-disparlure

OH?

19

532

25 Asymmetric Induction II Asymmetric Catalysis

It is best to add the OH group in a reaction that allows a helpful disconnection so an alkene was introduced (Functional Group Addition) to give the vinyl epoxide 20. The double bond can be disconnected by a Wittig reaction. Although we might expect a mixture of double bond geometries using the Wittig reaction it doesn’t matter here—this double bond is going to be removed anyway. Aldehyde 21 could come from alcohol 22 which is the product from an AE reaction on the Z-allylic alcohol 23. The geometry of this alkene must be controlled and fortunately Z-allylic alcohols are easily made by partial reduction of alkynes (chapter 15). OH CHO

Wittig

FGA

18

O

FGI

AE

O R 20; R = n-C 10H19

OH

O R 21

R

R 23

22

The (Z)-allylic alcohol 23 was epoxidised at ⫺40 ⬚C to give 22 in 91% ee and 80% yield. Sharpless7 says ‘The crystallinity of the epoxy-alcohol 22 greatly simplifies its isolation (no chromatography needed.)’. Oxidation, Wittig reaction and hydrogenation of the alkene complete the synthesis of (⫹)-disparlure 18 that did indeed attract the moths. OH

OH

1.

(–)-DET (i-PrO)4Ti

t-BuOOH –40 ˚C

R 23

O

Ph3P

CHO

CrO3. 2pyr

O

CH2Cl2

18; (+)disparlure

2. RhCl, H 2

R 21; 88% yield

R 22; 80% yield, 91% ee

47% yield

The Payne rearrangement The products of a Sharpless epoxidation, such as epoxides 12, 16, or 22, are potentially unstable in base as the anion of the alcohol can attack the epoxide 24 in the Payne rearrangement. This is easily seen with the simplest compound 12. It doesn’t and we have rather given the game away by the compound numbers. The OH groups in the right hand and in the left hand compounds 12 are homotopic. Sharpless made the definitive statement of this in his propranolol synthesis.6 base

HO

O

O

O 24

12

O

O

OH

O

24

12

If the carbon skeleton is unsymmetrical 2 the danger is different: the Payne rearrangement gives a mixture of two products 2 and 25. There are now three places where a nucleophile might attack: C-3 in 2, C-1 in 25, and C-2 in both but with different stereochemical outcomes. We much prefer to make 1 rather than 26 by AE as 26 is already chiral and the AE reaction would be a kinetic resolution. 3

1

Payne

AE

HO

R 1

HO

2 2

O

R base

1

OH

OH AE?

O

R

3 R 2 25

26

Of the two compounds 2 and 25, the more substituted epoxide 2 is the more stable and therefore the more abundant. But the mixture reacts with nucleophiles preferentially at the least substituted

25 Asymmetric Induction at the Allylic Alcohol Centre: AE is anti-Selective

533

carbon, that is C-1 in 25. The minor component of the mixture 25 gives the major product 27. So AE on 1 can be used to make products from 2 or 25 according to the reaction conditions.

R

HO

3

1

AE

HO

2

1

R

O

base

OH

1

Payne

O

OH Nu

3 R 2

2

Nu

R OH 27

25

Our illustration comes from Sharpless’s sugar syntheses.8 The mono benzyl ether 28 responds well to AE and reaction of the epoxide 29 with PhSH in basic solution gives mostly (4.5:1) the product 30 of the Payne rearrangement.

OH

BnO

(+)-DET (i-PrO)4Ti

O BnO

OH OH

t-BuOOH

BnO NaOH

OH 30; 88% yield, 4.5:1

29; 85% yield, 99:1

28

SPh

PhSH

Asymmetric Induction at the Allylic Alcohol Centre: AE is anti-Selective It is possible to make the CH(OH) centre of the allylic alcohol asymmetric by an AE reaction. This is a kind of symmetry breaking, It is well illustrated by Fürstner’s synthesis of the only interesting piece 32 of the natural product (⫺)-balanol 31: a potential lead for development of protein kinase inhibitors.9 The two disconnections are trivial. O

H H O O

OH

H

O

OH NH2

C–O ester

N HO2C

O

C–N amide

O

OH

N H

N H 31; (–)-balanol

32

It does not look initially as if the cyclic amine 32 can easily be made by AE. However, the two substituents on the seven-membered ring are anti and could come from an epoxide opening. The key idea is to make the seven-membered ring by olefin metathesis (chapter 15). An alkene must again be added and Fürstner placed it so as to make the alcohol allylic 33. Reversing the metathesis gives a non-cyclic starting material 34. Now he removed one of the amines and replaced it by a second OH group to give 35. OH

OH NH2

FGA

N H 33

metathesis

OH FGI

C=C

32

OH NH2

N

N

H 34

H 35

534

25 Asymmetric Induction II Asymmetric Catalysis

Disconnection of a 1,2-diX relationship gives allyl amine 36 and an epoxide 37 that might be made by AE. The starting material is indeed an achiral allylic alcohol 38 but it has two alkenes. They are in fact enantiotopic and AE will probably be able to discriminate between them. A more difficult question is what will be the stereochemistry of the alcohol in 37 as it becomes a stereogenic centre during the AE reaction. It turns out (by experiment) that AE has anti-selectivity: the catalyst delivers the epoxide oxygen to the opposite face to the OH group. This is surprising because both oxygen atoms are Ti-bound. But inspection of the loaded catalyst structure above shows that a substituent (here vinyl) on the allylic alcohol would prefer to be on the opposite face of the alkene to that attacked by t-BuOOH when both are bound to Ti. The symmetry aspects of this reaction are discussed in more detail in chapter 28. OH

OH OH

OH 1,2-diX

OH

AE?

O = NH

C–N

O

NH2 37

36

35

38

‘Unnatural’ D-DET was needed to get the right absolute stereochemistry and 37 was not isolated but protected as a benzyl ether 39 before reaction with allylamine gave the carbon chain of 35. Conversion of OH into NH2 now required inversion. OH

(D)-(–)-DET

OBn

NaH

OBn

36

37 t-BuOOH (i-PrO)4Ti

BnBr

N H

O

OH 40; 94% yield

39; 95% yield

38

After N-Boc protection, metathesis worked very well to give doubly protected 42 and inversion with diphenylphosphoryl azide under Mitsunobu conditions gave the azide 43. This order of events allowed a triple reduction - of the alkene, the azide and the benzyl group - in one step and Boc-protected 44 could then be used in the synthesis of balanol itself. OBn

OBn OH

OBn OH

Mo-carbene catalyst

Boc 41

N Boc 42; 94% yield

NH2

N3 H2, Pd/C

(PhO)2PO.N 3

metathesis

N

OH

Ph3P, DEAD

N

acid

Boc 43; 91% yield

N Boc 44

No Asymmetric Induction from Remote Allylic Alcohol Centre: Reagent Control If there is already a chiral centre in the molecule that has already been controlled by asymmetric synthesis, the AE reaction will ignore it if it is far enough away. Yamada’s synthesis10 of the biologically active diterpenoid 45 illustrates this and demonstrates control when nucleophiles attack an unsymmetrical epoxide. Preliminary disconnections with a Diels-Alder in mind took the chemists back to the diol 48.

25 No Asymmetric Induction from Remote Allylic Alcohol Centre: Reagent Control

NHCHO

H

H

H

O

O

H

H

535

OH OH

H

DielsAlder

H

C–C

H O

O

O

Cl 45; kalihinene X

Cl

O

Cl

46

Cl 48

47

The tetrahydropyran ring 48 has a tertiary carbon joined to oxygen 48a and this C–O bond can be made by addition of an alcohol to an alkene. Addition of an anion-stabilising sulfone 50 allows a C–C disconnection back to the epoxide 51 that does look like an AE product except for the allylic chloride. OH OH

OH OH

H

OH OH

H

OH

H

FGA

C–O

O

OH

t-ether

OH

Cl 48a

O

Cl

SO2Ph

+

sulfone + epoxide

OH

Cl 49

C–C

SO2Ph

52

Cl 51

50

The substrate chosen for the AE reaction was the bis allylic alcohol 53 having the same carbon skeleton as 51 but with an inverted alcohol replacing the chloride. This was a strange choice as one of the allylic alcohols must be protected and there remains the question of the influence of the stereogenic centre on the AE reaction. Protection was easy: silylation with t-BuMe2SiCl (TBDMSCl) occurred only on the primary alcohol, silylation with t-BuPh2SiCl (TBDPSCl) could be forced onto the one remaining OH and selective methanolysis of the less hindered silyl group gave 54. The AE reaction was, as expected, regioselective for the allylic alcohol but most significantly was highly reagent controlled (10:1 in favour of 55 over the other diastereomer). Reaction with the lithium derivative of the sulfone 52 gave 56 which could be converted into 51 by acetylation of the two OH groups, desilylation and Mitsunobu-style inversion of the alcohol with Ph3P and CCl4. These reactions are discussed in the Workbook.

1. TBDMS-Cl, DMAP, Et 3N

HO

2. TBDPS-Cl, imidazole

HO

3. MeOH, PyrH + TsO –

OH 53

HO

O

t-BuOOH

1. 52-OBn, BuLi

HO (–)-DIPT (i-PrO)4Ti

OTBDPS 54; 83% yield BnO

2. Na/Hg, Na 2HPO4

OTBDPS 55; 95% yield

OTBDPS

HO 56; 89% yield

536

25 Asymmetric Induction II Asymmetric Catalysis

Should the starting material 53 be enantiomerically pure? Yes, as there is virtually no kinetic resolution. The asymmetric centre in 53 is evidently too far from the alkene that reacts by AE to have more than a minor effect on the stereoselectivity. Enantiomerically pure alcohol 53 was already available from asymmetric reduction (using CBS, see chapter 26) of the corresponding enone. Sharpless epoxidation does have limitations but it is supremely practical.

Asymmetric Synthesis of Diltiazem Diltiazem is a calcium channel blocker used in the treatment of angina and hypertension. The active compound is the (⫹) enantiomer of the cis-diastereoisomer 57. The side chain (R ⫽ Me2NCH2CH2- in 57) can be added by alkylation of 58. The amide is an obvious disconnection. OMe

OMe

S

S

C–N

C–N

O

O alkylation in base

OMe

N

O R 57; (+)- cis diltiazem

amide

OMe

N O

H

58

The intermediate 59 has two 1,2-diX relationships but only one can reasonably be derived from an epoxide and redrawing 59a to show the anti- stereochemistry between S and O that would result from opening an epoxide unfortunately reveals an epoxide 60 from a cis unsaturated ester. NH2 S

2 Ar

Ar

Ar redraw

O 2 1 RO2C

S OMe

NH2

59

C–S

CO2R O

OMe

1,2-diX

CO2R O 60

59a

We need an allylic alcohol if we are to use AE and we much prefer a trans allylic alcohol. So two inversions will be needed. The epoxide 63 was made in excellent yield and ee provided cumyl hydroperoxide 61 was used as the oxidant. Oxidation gave the required ester 64.

O

(i-PrO)4Ti

Ph

OOH 61

Ar

OH 62

(+)-DET, 61

Ar

OH 63

O

1. RuO 2, NaIO 4 2. Me 2SO4

Ar

CO2Me

64

Now for the first inversion. It turns out that nucleophilic attack by chloride ion occurs next to the electron-rich aromatic ring rather than next to the ester. The thiol 66 was used so that the thiolate anion could be made with a weak base, and the required regio- and stereochemistry is in place 67. The rest of the synthesis11 is in the Workbook. We shall be seeing other syntheses of diltiazem later in this chapter that use other asymmetric methods.

25 Summary of Sharpless Epoxidation

Cl

NO2

py•HCl

64

CO2Me +

Ar

HCl

537

Ar Et3N

CO2Me

S SH

OH 65

NO2

OH

66

67

Summary of Sharpless Epoxidation This exceptionally accommodating reaction can be made to work for most kinds of allylic alcohol but there are some general guidelines that we shall summarise here: 1. The substrate must be an allylic alcohol such as 1. The next best is a homoallylic alcohol 68 but then the reaction is slow and the ees generally poor. R

OH

OH

R

1

68

2. The best results come with terminal allylic alcohols that have E-alkenes with some substitution such as 1, 69, and 70.

R

OH

R2 R

OH R1

OH

1

69

70

3. Allylic alcohols with Z-alkenes (that is an alkyl group cis to the OH group 71 – 73) generally give poor results. The mnemonic device above shows that the fit is poor. R2

R

R3 OH

OH

R1 72

71

R2

OH 73

4. Terminal alkenes other than allyl alcohol itself are not much use. The epoxide from one achiral type 75 is very susceptible to nucleophilic attack and the Payne rearrangement, while the other achiral type 76 gives poor ees. The chiral compounds 77 and 78 belong to the realm of kinetic resolution (chapter 28). OH 74

OH

R

R

R OH

75

76

R1

R OH 77

R2

OH 78

PART II – SHARPLESS ASYMMETRIC DIHYDROXYLATION The Sharpless dihydroxylation reaction (AD) is arguably the most important discovery in organic chemistry of the 20th century. There is no question that asymmetric synthesis is at the forefront of modern synthetic organic chemistry. And the most impressive asymmetric methods are those which are catalytic. The Sharpless epoxidation achieved this and more. But those methods that

538

25 Asymmetric Induction II Asymmetric Catalysis

combine high enantiomeric excesses and high catalyst turnover with activity over a broad range of substrates become supremely useful when they are simple to perform and require no fancy solvents, temperatures or other rigorously controlled conditions. The Sharpless Asymmetric Dihydroxylation (AD), an asymmetric version of the familiar dihydroxylation of alkenes with osmium tetroxide, is such a method.12 As testimony to their ease of use, both the Sharpless epoxidation and the Sharpless dihydroxylation have now found their way into undergraduate laboratory practicals.

The AD Method Professor Barry Sharpless13 wrote, ‘It reacts only with olefins and it reacts with all olefins (slight poetic licence here)’ The “It” is, of course, osmium tetroxide which reacts with those olefins to give 1,2-diols.14 It does so stereospecifically, introducing the two hydroxyl groups on the same side of the flat double bond 79 to give the syn diastereoisomer 80. In addition to this stereospecificity, in the presence of an appropriate ligand, osmium tetroxide may react with olefins enantioselectively, attacking one or other of the two enantiotopic faces to give 80 or ent-80. OH R2

R1

OH

OsO4

R2

R1

+

OH 80

79

R2

R1 OH ent-80

The reaction has been developed by Sharpless15–18 and others19 from one which required a stoichiometric amount of prohibitively expensive osmium tetroxide to one where as little as 0.002 mol% of osmium may be used.13 What is more, the former method performs a racemic transformation whereas the latter yields essentially optically pure diols. Look at that quote again. The variety of substrates that can be used is enormous and even for an asymmetric reaction, no other functionality is required. Of the five compounds illustrated below, only the allylic alcohol 83 forms a suitable substrate for the Sharpless epoxidation but all are suitable for the asymmetric dihydroxylation (AD) reaction. Ph

OBn Ph

Ph 81

OH

82

Ph

83

OAc

CO2Me

Ph

84

85

It is certainly appropriate for us to describe here some of the important aspects of Sharpless dihydroxylation methodology. This background is essential to chemists who do, or are thinking about doing, the reactions. However, a detailed discussion of, for instance, the mechanism of the reaction, or the explanations behind the origin of the enantioselectivity, are beyond the scope of this book and so, where this sort of thing is described at all, it is done mostly from an empirical standpoint. Some history. The development of the current state of the art makes an interesting story. The reaction of osmium tetroxide with olefins was discovered14 in the 1930s. The reaction first gives an osmate ester 86. This ester is a stable species but can be hydrolysed to yield the diol 80. O O stoichiometric dihydroxylation

R2

R1

79

Os

OsO4

O

OH hydrolysis

O

R2 R1 86; osmate ester

R2

R1 OH (±)-80

25 The AD Method

539

Osmium tetroxide is very expensive and very toxic which made using it quite unattractive. For a long time, many people who used osmium tetroxide to convert olefins to diols—and this was long before enantioselective dihydroxylations came on the scene—used the Upjohn procedure.20 This process used catalytic amounts of osmium tetroxide, NMO (N-methylmorpholine N-oxide) 87 as the stoichiometric oxidant, and one solvent phase. The solvent was water, acetone and tert-butyl alcohol. The osmate ester 86 was hydrolysed under these conditions and the osmium (VI) species was reoxidised to OsO4 by NMO. catalytic dihydroxylation O

N Me

O

R1

O

OH

Os

hydrolysis

O 79

87; NMO N-methyl morpholine N-oxide

O

“Upjohn procedure”

R2

O

Os (VI)

OsO4

R1 NMM

NMO

86

R2

R2

R1

H2O acetone t-BuOH

OH (±)-80

The ligand The simplest asymmetric ligands that were used, and from which all of the state-of-the-art ligands are derived, were esters of either dihydroquinidine 88 or dihydroquinine 90. Sharpless first developed a stoichiometric process using these ligands but many years passed before the catalytic asymmetric dihydroxylation was achieved when the ligands were added to the standard Upjohn process.15 The two cinchona alkaloids (and their ester ligands 89 and 91) are not, in fact, enantiomers, but diastereomers. The absolute configurations of the centres in one complement those in the other at all but one centre – where the ethyl group joins the quinuclidine unit – and so they function as enantiomers. Cinchona is pronounced sinkowner: “Named by Linnæus after the countess of Chinchon, who, when vice-queen of Peru, was cured of a fever by Peruvian bark, and afterwards brought a supply of it into Spain” according to the Shorter Oxford English Dictionary. Et

Et

N

N R

O

O

H

H

R

MeO

OMe

N N

N 88; dihydroquinidine (R = H) (DHQD) 89; (R = p-chlorobenzoyl)

90; dihydroquinine (R=H) (DHQ) 91; (R = p-chlorobenzoyl)

92; quinuclidine (achiral)

These ligands were superseded with the development of the phthalazine (PHAL) ligands in which two cinchona alkaloid units are connected together. The two most widely used ligands are the phthalazine ligands (DHQD)2-PHAL 93 and (DHQ)2-PHAL and these are used in the two commercially available asymmetric dihydroxylation mixes – AD-mix-α and AD-mix-β. Et

Et

N

N O

H

N

N

N

Alk*O

O

N OAlk*

H

MeO

OMe

N

N 93; (DHQD) 2-PHAL

94; phthalazine ligands (DHQ)2-PHAL or (DHQD)2-PHAL simpler representation

540

25 Asymmetric Induction II Asymmetric Catalysis

The diphenylpyrimidine (PYR) 95 and the indoline (IND) substituted ligands 96 are other types. The indoline ligands contain just the one alkaloid unit. The three different ligand types (PHAL, PYR and IND) are suitable for different types of olefin. Et

Ph Alk*

N

O

O N

Alk*

N

N

O

H

O

MeO

Ph N

95; diphenylpyrimidine ligands (DHQ)2-PYR or (DHQD)2-PYR simple representation

96; (DHQIND)

More recently discovered ligands21 include the diphenylpyrazinopyridazine (DPP) 98 and anthraquinone (AQN) 99 ligands as well as the diphenyl analogue of the phthalazine ligands (DP-PHAL) 97. OAlk*

OAlk* Ph

Ph

N N OAlk*

97; diphenylphthalazine (DP-PHAL) ligands

Ph

N

Ph

N

O

OAlk*

O

OAlk*

N N OAlk*

98; diphenylpyrazinopyridazine (DPP) ligands

99; anthraquinone (AQN) ligands

Solvent dependence There are two solvent systems commonly used for dihydroxylations. The first involves just one phase and consists of either water and acetone22 or water, acetone and tert-butyl alcohol.14 The second, and more modern, solvent system is a two phase mixture and consists of water and tert-butyl alcohol. The impact that the two phase system has on enantioselectivity is profound. This is because it prevents the deleterious ‘second cycle’ which can otherwise occur. In fact, the single phase method is still used for certain applications including multiple dihydroxylations.

The catalytic cycle One of the problems encountered in the development of the AD reaction was a second cycle.23 The first cycle is the same as the cycle we saw in the Upjohn procedure with one modification. It is possible for the osmium in the osmate (VI) ester to become oxidised before hydrolysis to form an osmate (VIII) ester. If this species simply hydrolyses to diol and OsO4 then there is no problem but otherwise this is where the trouble starts. In the second cycle the osmate (VIII) ester reacts with another olefin instead of being hydrolysed. It does so with low enantioselectivity (and sometimes in the opposite sense to the first cycle) which results in the poorer enantiopurity of the product (step D rather than step C below). One way round this is to keep the olefin concentration as low as possible which can be done by adding the olefin slowly.22,23 Although slow olefin addition resulted in a profound improvement in enantiomeric excess, it was inconvenient and no way near as effective as the alternative which was to alter the conditions. The use of two phases makes it impossible for the second cycle to occur.19,24

25 The AD Method

541

OH R

R O O Os O O L

R

R

L

H2O

OH

R

R step C

step D

O First Cycle R high enantioselectivity O Os O O O R L Osmate (VIII) R ester

step A

O O

[O]

O O O Os O O O

step E [O]

R R

R

R

R

Osmate (VI) ester

R R

Second Cycle low enantioselectivity

R

OH

L

R

step F

step B

R

Os O

O

O O Os O O O

R

H2O, L

OH

Before the osmate (VI) ester, resulting from the addition of olefin and osmium tetroxide, can react with a second olefin it has to be reoxidised (step B). With two phases, the osmate ester and olefin are in the organic layer and the oxidant [Fe(CN) 63⫺] is in the aqueous layer. Hence the osmium cannot be oxidised while it is in the form of the osmate (VI) ester to an osmate (VIII) ester. Only after the osmate ester has hydrolysed may the osmium enter the aqueous phase and be reoxidised for another cycle (step Y).12 It is critical that the reaction mixture is biphasic. Just using a water/tert-butyl alcohol solvent system is not enough because these two solvents can be miscible. It was demonstrated24 that stoichiometric osmylation of styrene in 1:1 tert-butyl alcohol/water, a homogeneous reaction, gave a product with 58% ee. That same stoichiometric reaction performed in the presence of either K3Fe(CN) 6 or potassium carbonate led to a 74% ee. These salts stop the water and tert-butyl alcohol mixing. The addition of K2CO3 is essential. Although it may be partially replaced with NaHCO3 for base-sensitive materials the AD reaction fails to turn over with its complete omission.12,24

step W

R

R

R Osmate (VI) ester

O O Os O O L

Organic

+ L OH R

L+

R

R

O O O

Os

O

OH step Z

step X

Aqueous

2–

2 OH–, 2 H2O

O HO HO

Os

2– O

OH OH

HO Os HO

O

O O

O

step Y 2

OH–, 2

Fe(CN)6

3–

2 H2O, 2 Fe(CN) 43–

2 OH–

542

25 Asymmetric Induction II Asymmetric Catalysis

Methanesulfonamide Another key development was the discovery that methane-sulfonamide MeSO2NH2 accelerated the rate of hydrolysis of the intermediate osmate ester (not to be confused with the rate of the addition of osmium tetroxide to the olefin). Reaction times can be reduced by as much as 50 fold. After 3 days at 0 ⬚C in the absence of methanesulfonamide, trans-5-decene had been only 70% converted to the corresponding diol but the diol was isolated in a 97% yield after just 10 h at 0 ⬚C in the presence of methanesulfonamide. This improvement means that reactions can be run at 0 ⬚C instead of room temperature. However methanesulfonamide slows down the reaction of terminal olefins. It is thus omitted from such reactions.16

Substrate Dependence and the Mnemonic Device In order to predict facial selectivity, Sharpless and co-workers invoke a mnemonic device.25 To an approaching olefin, the greatest steric constraints are presented by the NW, and to an even greater extent, the SE quadrants. The SW and NE quadrants are more open and, in addition, the SW quadrant contains what is described as an “attractive area”. The attractive area is particularly well suited to accommodate flat aromatic groups. The olefin positions itself according to the constraints imposed by the ligand and is dihydroxylated from above (β-face), in the case of dihydroquinidine derivative, or from below (α-face) in the case of dihydroquinine derivatives. The commercially available AD-mix-α and AD-mix-β are chosen according to this mnemonic.

The Mnemonic Device

dihydroquinidine derivatives β-face NE

NW RM

RS attractive area SW

RL

H α-face dihydroquinine derivatives

SE

Inspection of this mnemonic illustrates why stilbene is such a good substrate for these ligands: its two aromatic portions can align nicely with two-fold degeneracy into the ligand pocket with only hydrogen atoms in the sterically demanding areas. In contrast, we can see why cis alkenes make such poor candidates – one of the substituents will have to be in a sterically demanding area. Of all the alkene classes,25 the cis-di-substituted alkenes are the least good. They need a unique ligand - the indoline substituted alkaloid18 96 and even then enantiomeric excesses rarely exceed 80%. The figure below outlines the ligands that Sharpless recommends for each of the six olefin classes. Five of the six work very well with either PHAL or PYR. The AQN ligands are particularly suited to allylically substituted terminal olefins21 and the PYR ligands are good for sterically congested olefins.

25 Applications of the Sharpless AD Reaction

543

Recommended ligands for different classes of alkenes gem-di-

mono-

PHAL PYR AQN

30-97% ee

97% ee (R) (DHQD)2PHAL

70-97% ee

PHAL Cl 90% ee (S) (DHQD)2AQN

cis-di-

30% ee (R) (DHQD)2PHAL

trans-di

Ph Ph 99.8% ee (R, R) (DHQD)2PHAL

72% ee (1R, 2 S) DHQD-IND IND

94% ee (R) (DHQD)2PHAL

20-80% ee

PHAL

90-99.8% ee 95%ee (R, R) (DHQD)2PHAL

56% ee (1R, 2 S) DHQD-IND tri-

tetraPh Ph 95% ee (R, R) (DHQD)2PHAL

PHAL

90-99% ee

47% ee (R) (DHQD)2PYR PHAL PYR

98%ee (R) (DHQD)2PHAL

20-97% ee

MeO

64% ee (S) (DHQD)2PHAL

Applications of the Sharpless AD Reaction That was the background. And now we look at the application of the AD reaction. First of all we look at a few retrosyntheses before moving on in the section dihydroxylating with more than one double bond to consider the more complicated issues. It is often easy to spot when a Sharpless AD strategy might be used if there are two hydroxyl groups in the target molecule which are 1,2related.

The synthesis of indicine Indicine 100 is an alkaloid which has antitumour activity but hepatic toxicity.26 The molecule clearly consists of two enantiomerically pure chunks which we would like to separate in our retrosynthesis. What is more, those two portions are conveniently connected by an ester linkage making the first disconnection very easy.

544

25 Asymmetric Induction II Asymmetric Catalysis

OH

OH

O HO

HO

C–O

OH

H +

H ester

OH

O

HO

N

O

OH

N 101

100; indicine

102

We shall not concern ourselves with the whole synthesis of (⫹)-retronecine 101 at this stage but just with the diol 102. Diol 102 comes from the corresponding alkene by Sharpless AD. OH

Sharpless AD

HO

HO

O 103

O OH 102

The substrate for the Sharpless AD was, in fact the ethyl ester 104 of the α,β-unsaturated acid 102. This was dihydroxylated using AD-mix-α and then the ester 105 was hydrolysed using barium hydroxide. Before the resulting acid was coupled with alcohol 101, the diol functionality was protected as an acetal 106. The coupling was performed using DCC and DMAP and the protecting group was removed under acidic conditions to give indicine 100. OH

Sharpless AD

EtO AD-mix-α

O

HO

O OH 105; 81% yield, 90%ee

104

O OH (–)-102; 96% yield

O

MeO OMe 1. DCC, DMAP

O cat. HCl

OH

Ba(OH)2·8H2O

EtO

HO2C

O

cat. H +

O

O HO

H+

H

2. (+)-100

O

100 84% yield

N (+)-106; 76% yield

107

When fitting olefin 104 into the pocket provided by the mnemonic device we notice that the isopropyl group and the ester function in olefin are of similar size. As a result, we might expect the ee to be low and yet 90% is achieved.

The synthesis of a C2-symmetric piperidine When we think about retrosynthesis, two 1,2-related oxygen atoms, each on stereogenic carbon atoms, as in 100, may scream to be ‘retro-dihydroxylated’ but it is harder to spot where the strategy may be useful if there is only one oxygen atom or where heteroatoms are not even 1,2-related. Consider, for instance, the C2-symmetric 2,6-disubstituted piperidine27 108. We notice for a start that the two oxygen atoms are most certainly not 1,2-related. At best they are 1,5. Both the oxygen atoms are 1,2-related to the nitrogen atom however. Two carbon – nitrogen bonds can be disconnected at once to give the tosylated tetraol 109. We are going to have to think about protection

25 Applications of the Sharpless AD Reaction

545

because the secondary alcohols are the ones that need to be tosylated – we will worry about that in the forwards synthesis. As the tetraol 110 is C2 symmetric it could be made by bis dihydroxylation of diene 111 using the same asymmetric ligand. OTs

2 x C–N

HO

OH

N Bn 108

OTs

FGI

HO

OH 109

OH

OH

HO

Sharpless AD

OH 111

110

Diene 111 was dihydroxylated using AD-mix-β to give the C2 symmetric tetraol 110 and its meso diastereomer 114 in a combined yield of 87%. The primary hydroxyl groups are selectively protected using TBDMS chloride and imidazole – otherwise they would react with tosyl chloride. The secondary alcohols are then activated with tosyl chloride to yield 113 and meso isomer in 47% yield from the tetraol. Reaction with benzylamine yields the target material 108 in 47% and the unwanted meso amine 115 in 30% yield. AD mix β

OH

TBDMSCl

111

110, 87% yield imidazole

OH

TBDMSO

OTBDMS 112

OTs

TsCl

OTs

TBDMSO

Et3N

1. Ph

OTBDMS

NH2

TM 108 47% yield

2. deprotect

113; 47% yield

Take particular note that two hydroxylations are done on the same substrate 111 at the same time. We will not dwell too much on this here, but from the point of view of enantiomeric excess this is an aspect of asymmetric methodology that can be very useful. See Double Methods in chapter 28. OH

OH

HO

HO

OH 114

OH

N Bn

115

The synthesis of the alkaloid reticuline If it is difficult to spot where a product results from a dihydroxylation because there is only one oxygen in the product then it must be even more difficult to spot when there are none. Reticuline 117 is a natural product which can be used in the synthesis of morphine 116. OH

HO NMe

OMe

O NMe MeO HO

116; morphine

OH

117; ( R)-reticuline

546

25 Asymmetric Induction II Asymmetric Catalysis

We will come onto the stereochemistry in just a moment but first we can simplify the structure by disconnecting the six-membered amine ring. We should be able to make this C–C bond 117a using a Friedel-Crafts reaction. Right, we notice that there is no oxygen functionality involved in reticuline’s one chiral centre. If we are to use a Sharpless AD reaction then the first thing to do is to put in some oxygen functionality. OH

OH OMe

NMe

1. C–C Friedel -Crafts?

X

C–N

2. FGA MeO

SN2

OH

MeO OH

OMe

NMe

OH

117a

118

The C–N bond of the amine can now be disconnected 118. If we imagine that it was introduced by an SN2 displacement, then we invert that stereocentre and introduce an oxygen-based leaving group to give the diol 119. Now we have the sort of compound that we are looking for! We can instantly see that this comes from the corresponding stilbene 120. Notice that the hydroxyl groups in 120 are homotopic and that the diol 119 is C2 symmetric. OH

OH OMe

OH

OMe Sharpless AD

OH

MeO OH

MeO OH

119

120

The stilbene 121 (which is dibenzylated 119) is dihydroxylated to yield diol 122 in 82% ee and 88% yield. This diol needs to be activated. It is reacted with SOCl2 to yield a cyclic sulphite which is oxidised to cyclic sulphate 123. This is redrawn as 123a. Both cyclic sulphites and cyclic sulphates are very useful synthetic building blocks with, of course, two electrophilic carbon atoms.28 The two benzylic oxygen atoms here are homotopic so attacking either benzylic position will yield the same enantiomer of product. In fact a protected aminoaldehyde was used as nucleophile as this makes the cyclisation easier.29 This and the deprotection steps are given in detail in the Workbook. OBn OMe

OH AD

Ar

Ar

2. RuCl 3 NaOCl

OH MeO

122; 82% ee OBn

O

1. SOCl 2 Et3N

O SO2 Ar =

Ar O 123

O S

O

O

Ar

Ar 123a

121 OBn

MeO

NHMe

(MeO)2CH

NMe

MeO 123 MeO

OH OBn 124; 65% yield from diol

OMe

117 (R)-reticuline

25 Dihydroxylating Compounds with More than One Double Bond I—Regioselectivity

547

Note that to keep a straight carbon chain in our drawing of the cyclic sulfate 123a we need to draw the two S–O bonds in a very curvaceous way 123. This technique is often used to keep emphasis on the syn nature the two hydroxyl groups of diols when the diols are converted to cyclic sulfates or acetals. The disconnection we used to plan the synthesis of reticuline is rather unusual. Reticuline is often drawn in a way 117b that is rotated clockwise by 60⬚ from the way drawn above 117a. The disconnections are more often a C–C disconnection between the aromatic ring and an imine resembling the Mannich reaction as the imine is formed by the attack of an amine 125 on an aldehyde 126. This is an easy reaction to carry out but it forms two bonds and a chiral centre all in one step and is difficult to make asymmetric. Because we wanted to use AD, we needed a different disconnection. MeO

MeO

HO

C–N C–C

NMe H

OH

CHO

NHMe

HO

OH

+

Mannich style

OMe 117b

OMe

125

126

Dihydroxylating Compounds with More than One Double Bond I—Regioselectivity Here we think about which double bond will be dihydroxylated if a molecule contains more than one. We postpone the question of how one already dihydroxylated double bond influences the diastereoselectivity in the dihydroxylation of a second double bond until the next section. If there is more than one double bond in a molecule then, unless we have something like a symmetrical diene, there will be a question of regioselectivity. Predicting that regioselectivity can be a bit of a minefield and you need to know what you’re doing.12 In this section we will deal with some of the simpler guidelines. The Workbook deals with some of the more subtle rules and finer methods. We start with dienes.

Dienes With two double bonds only, we need to worry about two main things – whether or not they are conjugated to each other, and whether or not they are equivalent from a symmetry point of view 127 – 130. A substrate may contain two double bonds which are – because they are related by a symmetry element that makes them so – equivalent. Then they may be conjugated or not. The double bonds may, of course, be different electronically and/or sterically. For the sake of strategy we must have some idea how one double bond reacts in the presence of another. For the moment we will concern ourselves only with monodihydroxylation of dienes and outline a few guidelines. Guidelines are not rules. For a given substrate, the Guidelines may support each other or they may conflict with one another. In the latter case, predicting what will happen can be a bit tricky!

127 conjugated and equivalent

128 non-conjugated and equivalent

129 conjugated and non-equivalent

130 non-conjugated and non-equivalent

548

25 Asymmetric Induction II Asymmetric Catalysis

Guideline One. – Double bonds which are trans-di- 131 or tri- 132 substituted react faster than those which are either cis-di- 133 or mono- 134 substituted. The simple diene 135 reacts regioselectively at its trans double bond when it is dihydroxylated using (DHQD)2PHAL as the ligand.

OH

> 131

133

OH

AD

OH

+ OH 3:1 136; 90% ee

> 132

135

134

137; 72% ee

Guideline Two. – Electron-rich double bonds react in preference to more electron-deficient double bonds. You can see that this is partly responsible for Guideline One but the principle extends beyond this. The benzoate ester 138 contains two trans double bonds but one is evidently much more reactive than the other. The double bond on the left is part of an allylic ester and is deactivated by the electron-withdrawing benzoate group. allylic portion Ph

AD

O O

OH

OH BzO

BzO +

(DHQD)2 -PHAL

OH

OH

138

139

140

14:1

Guideline Three. – In a conjugated system, the double bond that reacts will be the one that leaves the most conjugated system behind. This means that the double bonds at the end of a conjugated system are the more reactive. Conjugated diene 141 has one ‘end’ only because the benzene ring is included in the conjugated system but is not reactive. It is perhaps worth noting that in a simple conjugated system – like, for instance, octatetraene – the highest HOMO coefficient and the highest LUMO coefficient will both be on the terminal carbon. OH

OH AD (DHQD)2 -PHAL

141

+ OH

OH

142

143

4:1

This guideline predicts that dihydroxylation of the diene 141 would give the diol 142 as the main product.30 This is indeed the case but, as we shall see, this selection is easy to overturn. In this case, if the bulk of the aromatic ring is increased then the selectivity is reversed. Guideline Four. – Other Guidelines being equal, the least hindered double bond will react preferentially.30 Thus 130 gives mainly 144. HO OH

AD

+ (DHQD)2–PHAL

130

OH 144

OH 20:1

145

25 Dihydroxylating Compounds with More than One Double Bond I—Regioselectivity

549

When guidelines combine or compete Combinations of these simple systems can be seen. Triene 146 has three different sorts of double bond—a cis, a trans and a terminal double bond. Which of these will react preferentially? Two of the guidelines come into play here. We have a conjugated system and so we would expect one of the double bonds at the end of that system to react by Guideline Three. However, Guideline One tells us that double bonds which are trans-di- substituted are more reactive than those which are either cis-di or mono- substituted. So maybe the triene reacts in the middle! The answer is, in fact, that it reacts at the right hand end. Conjugation is more important than alkyl groups.30 C5H11

AD with (DHQD)2-PHAL

C5H11

OH OH

0 ˚C

146

147

A synthesis of aspicillin In the synthesis of (⫹)-aspicilin 148, a lichen macrolide, by Sinha and Keinan31 all the chiral centres were introduced using the AD reaction. An intermediate in the synthesis of this compound was compound 149.

OH

O OH

O

O O

OH

OH

O

OH

O 148; (+)–aspicilin

149

If we draw the carbon backbone of 149 in a slightly straighter fashion 149a it’s easier to see the retrosynthetic step that follows. This means that we are forced to draw the acetal 150 in the same distorted way that we used with cyclic sulfates. We can install the diol at the end of the molecule using AD-mix-α.

O

OH

O AD-mix-α

O

OH

O

O O

O O

149a

150

We can then remove both of the acetal protecting groups to reveal the tetraol 151. The stereochemistry of both the diols in 151 demands the use of AD-mix-β on 152.

550

25 Asymmetric Induction II Asymmetric Catalysis

OH 1,1-diX

AD-mix-β

150 acetals

OH

OH OH 151

152

It is when we look at the synthesis in the forward direction that we see just how clever it is. We react triene 152 with AD-mix-β and there are three double bonds which could react. There are two terminal double bonds and one trans double bond. Guideline One would suggest that the first double bond to react would be the trans one to give product 153. Sinha and Keinan31 who thoroughly investigated the dihydroxylations of this triene, found that this was indeed the case. The diol forms with a 96% enantiomeric excess. We are then left with two terminal double bonds. Which of these will react? In a conjugated diene, the dihydroxylation of one double bond reduces the reactivity of the second one considerably. Hence the second dihydroxylation yields mostly 151 which is formed with an 86% diastereomeric excess. The tetraol is protected before we react the final double bond with, this time, AD-mix-α. OH AD-mix-β

AD-mix-β

152

OH

151 86% de

AD-mix-α

acetal

150

149

153; 96% ee

This synthesis uses both enantiomeric asymmetric dihydroxylation reactions. They are used successively to exploit the regioselectivity dictated by the inherent relative double reactivities. For certain substrates this makes the asymmetric dihydroxylation a tremendously powerful reaction.

Designer ligands to improve selectivity It is also possible to select our regioselectivity by the synthesis of ‘designer’ ligands. Corey32 has synthesised a ligand which provides superior selectivity for some substrates. One difference is that, rather than having the methyl ether of the quinoline ring we find in quinine, it has a 4-heptyl ether. The idea is that this further sterically encumbers the ligand. One might imagine that regioselectivity in the case of either geranyl geranyl acetate 154 (with four double bonds) or squalene 155 (with six double bonds - notice the symmetry here) would be a tricky business.

OAc

154; geranylgeranyl acetate

155 squalene I

II

III

It is. If squalene 155 is mono-dihydroxylated with OsO4 using (DHQD)2PHAL as the ligand, the three double bonds I, II and III react in a ratio of 2.4:1.8:1.0. If no ligand is present at all they react in a ratio of 1:1:1 but with Corey’s ligand, reaction of double bond I and the sum of the reactions of II and III give a product ratio of 8:1.

25 Dihydroxylating Compounds with More than One Double Bond II—Diastereoselectivity

551

Dihydroxylating Compounds with More than One Double Bond II—Diastereoselectivity There are two main issues here. Firstly, if one of two double bonds has reacted, does the remaining one become more or less reactive than it was in the first place? Secondly, how does the stereochemistry of the first dihydroxylated double bond influence the stereochemistry of subsequent reactions? The factor that has the biggest influence - at least on the first of these questions - is the solvent system that is used. Up until now we have been looking at the regioselective monodihydroxylation of substrates containing more than one double bond. The solvent system that we have used has been the biphasic tert-butanol and water system.

Advanced dihydroxylation strategy Just have a look at cyclododecatriene 156. It has twelve carbon atoms, two trans double bonds and one cis double bond. This compound, with its three double bonds, presents us with all sorts of opportunities.33 The first thing to realise is that the trans double bonds will react with OsO4 more quickly than the cis double bond. In fact, both the trans double bonds will react before the cis double bond. We can dihydroxylate 156 to give us the diol 157. A better ee is obtained (94%) when this reaction is allowed to go to completion because the minor enantiomer is removed by kinetic resolution. The phenomenon of enantiomeric enhancement by kinetic resolution and the idea of match/mismatch in the context of a diol substrate with AD reagents are explored in Chapter 28.

OsO4

HO

(DHQD)2PYR

OH 157; 84% yield, 94% ee

156

There is clearly an option for further dihydroxylation. The diol can be protected and how we protect it depends on the dihydroxylation we want to do next. It turns out that the disilyl ether 158 is matched with (DHQ)2PYR whereas dioxolane 159 is matched with (DHQD)2PYR. Hence we may, by judicious protection, maximise subsequent stereoselectivity. Obviously, the disilyl ether of the enantiomer of 157 would be matched with (DHQD)2PYR and the enantiomer of dioxolane 159 with (DHQ)2PYR. Matched with (DHQ)2PYR

Matched with (DHQD)2PYR OsO4

RO

O OR

158; R = TBDMS

O

159

+

HO O

(DHQD)2PYR

O

HO 160

HO O O

HO

99:1

161

The major product from the dihydroxylation of 159 is the partially protected tetraol 160. If this were deprotected it would form the C2 symmetric tetraol 162. Note that 162 results from two

552

25 Asymmetric Induction II Asymmetric Catalysis

successive treatments of dodecatriene 156 with OsO4 in the presence of (DHQD)2PYR. If we wanted to make tetraol 163 – which is achiral – we would be advised to use diether 158 and treat it with OsO4 and (DHQ)2PYR before deprotection. This chemistry is explored in the Workbook.

HO OH

HO OH OH

HO

OH

HO

162

163

Scaling Up the Asymmetric Dihydroxylation This important reaction has been applied on a large scale by the pharmaceutical industry. The amplification of chirality from small amounts of catalyst and the minute amounts of osmium make it very attractive but the expense of K3Fe(CN) 6 and the difficulty of removing it from the reaction mixture on a large scale suggest using NMO instead. The by-product is then the simple amine N-methylmorpholine that can simply be extracted in the workup as can the ligand. The first of these examples was a 2.5 kg reaction34 while the second gave 42 kg of crystalline diol 166 in ⬎98% ee from methyl trans cinnamate.35 OH

OH OH

K2OsO2(OH)4

CO2Me

82 g (DHQD)2PHAL

O MeO

OH

O

NMO 87

166; 42 kg, 70% yield >98% ee (crystals)

MeO

t-BuOH, H 2O

164

165

PART III – AMINOHYDROXYLATION Not content with revolutionising the world of asymmetric synthesis with the catalytic asymmetric epoxidation reaction and the asymmetric dihydroxylation, Sharpless went on to develop the asymmetric aminohydroxylation 36 (AA). This is very similar to the dihydroxylation except one of the oxygen atoms is replaced with a nitrogen atom in the product 169. This complicates matters because we may have the additional question of regioselection. You will now be familiar with all of the ingredients in the reaction – K 2OsO2 (OH) 4 is the source of OsO 4 although the reactive species in this reaction is not OsO 4 since one of those oxygen atoms has to go. The ligand is the familiar double cinchona ligand (DHQD) 2PHAL 93 as is the mixture of alcohol and water (though isopropanol is used in this case) for the solvent. The one new component is the source of nitrogen. This is chloramine-T 167 (the sodium salt of N-chloro-p-toluenesulfonamide).37 Here is a typical asymmetric aminohydroxylation reaction. O S R

R

+

Ts

O N

Cl Na

167; chloramine-T

NH

NH2

cat. K 2OsO2(OH)4 ligand i-PrOH, H 2O

R

R OH 168

R

R OH 169

25 Scaling Up the Asymmetric Dihydroxylation

553

The reaction in the box below shows a typical AA with methyl cinnamate as the substrate. Methyl cinnamate 170 reacts in moderate yield and good, though not spectacular ee. Well never mind, this is a powerful transformation and not to be sneezed at. And from a practical point of view, it is often the case that the resulting amino alcohols can be recrystallised to very high ee. Note the regioselection in the reaction. With a cinnamate (or other α,β-unsaturated carbonyl compound) we have a nicely differentiated double bond which, in the absence of symmetry, is much better than two ends of a double bond that are only slightly different. We find the nitrogen atom bound to what was the more electrophilic end of the double bond 171. This is a handy mnemonic for the selectivity with a standard ligand – the more nucleophilic nitrogen atom binds to the more electrophilic end of the double bond (though, of course, it is only a mnemonic and not any sort of explanation). Three alternative nitrogen sources are 172 – 174 (but there are plenty more). While chloramine-T 167 and N-chloromethylsulfonylamide 172 are added to the reaction as salts, the carbamate salts (BocNClNa 173 and CbzNClNa 174) are prepared in situ from the carbamates. 168; chloramine-T

O

Ts

NH

O

Ms

4 mol% K2OsO2(OH)4

Ph

OMe

5 mol% ligand 92 room temperature i-PrOH, H 2O

170

Ph

OMe

Boc

OH 171; 60% yield, 82% ee

Cbz

N N N

Cl Cl Cl

172 173

174

Incredibly, changing the ligand to (DHQD)2-AQN 94 gives us the opposite regioselection (regioisomer 175) and yet the ee is similar to that obtained with (DHQD)2-PHAL and the same enantiotopic side of the double bond attacked. OH

168; chloramine-T

O

O

4 mol% K2OsO2(OH)4

Ph

OMe 170

5 mol% ligand 93: (DHQD) 2-AQN room temperature i-PrOH, H 2O

Ph

OMe NHTs 175

The reaction also scales up nicely38 – using N-bromoacetamide this time – and in a reaction that was 630 times larger than Sharpless’ standard mmol scale reaction, 120g of isopropyl cinnamate was reacted to give acetamide 177. The yield is good and the ee excellent. Hydrolysis of both the ester and the amide give (2R, 3S)-3-phenylisoserine 178 which is a precursor to one of the side chains of Taxol.® The β-amino alcohol function is found in many biologically active compounds and this catalytic reaction will be attractive to process research groups. Medium Scale O Ph

Oi-Pr 176; 120 g

NHAc O

AcNHBr, LiOH 1.5 mol% K2OsO2(OH)4

Ph 1 mol% (DHQ)2PHAL, 92 t-BuOH, H 2O, 4 ˚C

NH2 Oi-Pr

OH 177; 71% yield, 99% ee

Ph

O OH

OH 178

Many of the application of AA have used conjugated alkenes but simple, even mono-substituted alkenes such as 179 can be used providing there is some definite difference between the ends - here with this naphthalene compound 180 between no substituent and an aromatic ring. The synthesis of lactacystin using AA is described in chapter 31.

554

25 Asymmetric Induction II Asymmetric Catalysis

BocNH N

Boc Cl 174 4 mol% K2OsO2(OH)4

OH

1 mol% (DHQ)2PHAL, 92 MeCN, H 2O, 4 ˚C

180; 70% yield, 98% ee

179

Aminohydroxylation is the least well developed of the methods in this chapter but in many ways the most promising. Research continues. We shall now return to epoxidation as this is an alternative route to unsymmetrically substituted derivatives.

PART IV – CONVERTING 1,2-DIOLS INTO EPOXIDES We have already seen how cyclic sulfates such as 123 can be converted into amino alcohols that might have been made by amino hydroxylation. The conversion of 1,2-diols, made by the AD reaction, into epoxides has been very widely used. A combination of acetyl bromide and an orthoester gives a bromoacetate via an oxonium ion 182. The ion is formed with retention, bromide opens it with inversion at the benzylic centre, and epoxidation in base inverts it again. The net result is retention.39 OMe OH Ar

O

AcBr

OH

Br

Br

O MeC(OMe)3

181

Ar

K2CO3

OAc

Ar 182

MeOH

O Ar

183

184

It seems at first that regioselectivity is going to be a problem with diols such as 79 but this is not the case. A mixture of bromoesters 185 and 186 is indeed formed but both give the same epoxide 187 in base as each centre undergoes either no change or a double inversion. The result is again retention at both atoms and we have made the epoxide 187 from the alkene 79 asymmetrically.40 OH R2 AD

R1

Br R2

R1

AcBr

R1

MeC(OMe)3

OH ent-80

79

OAc R2

OAc 185

+

R2

R1

O

K2CO3

R1

MeOH

Br 186

R2

187

Applications of the AD to epoxide transformation: propranolol and diltiazem We return to two compounds we made earlier by the AE reaction: propranolol 7 and diltiazem 67. In both cases the synthesis is easier as we do not have to start with an allylic alcohol. The synthesis of propranolol41 uses the allyl ether 188 that gives the diol 189 with good ee in the AD reaction for a monosubstituted alkene. Transformation to the epoxide 190 shows no loss of ee. Reaction with i-PrNH2 was already known to give propranolol 7. This is a very short synthesis from easily made starting materials. O

O AD-mix-β

OH OH

1. AcBr MeC(OMe)3

O

O

2. MeOH K2CO3

188

189; 90% yield, 91% ee

190; 92% yield, 91% ee

25 Scaling Up the Asymmetric Dihydroxylation

555

The synthesis of diltiazem42 starts with the cinnamate that was actually used to make the allylic alcohol 62, the substrate for the AE route described above. No special extra inversion is required here if the intermediate 193 is used as the electrophile rather than the epoxide derived from it. Again, this is shorter and simpler than the previous synthesis. OH CO2Me

Cl CO2Me

CO2Me

AD

OH MeO 192; 80% yield, >97% ee

MeO 191

OAc

MeO 193

NH2 NH2

S

S

193

N

R

CO2Me

S 194

OAc

MeO

O

MeO

O OMe

62; diltiazem

195

However, it would be better if we had an asymmetric epoxidation that, unlike AE, did not require any particular functionality. We do: it is the Jacobsen epoxidation, the third of the great triumvirate of catalytic reactions in this chapter.

PART V – JACOBSEN EPOXIDATION The catalyst in this type of epoxidation is a salen, that is a compound derived from a salicylaldehyde and a diamine.43 One of the best is 198, made from the substituted salicylaldehde 197 and the tartrate salt of the C2 symmetric diamine 196 that results from a very simple resolution (chapter 22). There are others: they are all C2 symmetric salens made from C2 symmetric diamines and have large groups on the two benzene rings.

CHO

t-Bu

H3N

NH3

HO

OH

O2C

CO2 196

H

OH

H N

197

N

t-Bu

EtOH, H 2O, K2CO3, 80 ˚C

t-Bu

OH

t-Bu

HO

t-Bu

t-Bu

198; 95-99% yield

The active catalyst is a Mn(III) complex 199 of the salen used at 1 mol% or less. The stoichiometric oxidant is bleach (NaOCl) and again a two-phase system (CH2Cl2 and water) is best. Various additives such as pyridine-N-oxides improve performance.

556

25 Asymmetric Induction II Asymmetric Catalysis

H

H N

1. Mn(OAc) 2. 4H 2O

N Mn

EtOH

198

t-Bu

2. Air, NaCl, H 2O

t-Bu

Cl O

O

t-Bu

t-Bu

199; 95-99% yield

The Jacobsen method gives very good results with cis-alkenes, including cyclic alkenes, and trisubstituted alkenes. But is not so good with trans alkenes. Typical ees for these three classes of alkenes are shown for 200 to 202.

Two phase system: water and CH 2Cl2 Stoichiometric oxidant: NaOCl

O

Catalytic salen complex (99% ee on 600 Kg scale

208

Another important application is the asymmetric epoxidation of cis cinnamate esters 210. They are poor substrates for AD and the allylic alcohols derived from them by reduction are poor substrates for AE. Jacobsen epoxidation gives epoxides in good yield (⬎90%) but not all is the cis-epoxide 211. There is some leakage to the trans epoxide 212 a typical ratio of 211:212 being 5:1. The cis epoxide is formed in excellent ee but the trans epoxide in only moderate ee. Ph

NaOCl (13 mol%), 203 (1 equivalent)

CO2Et

Ph

CO2Et

O

+ (R,R)-199 (5 mol%), ClCH 2CH2Cl, 4 ˚C

Ph

O 211; 93% ee

210

CO2Et

212; 50-60% ee

Electron-donating groups on the benzene ring improve the cis:trans ratio and iso-propyl esters improve the ee. This make for a simple synthesis of diltiazem 62: a compound we have already made by AE and AD. Epoxidation of 213 gives a reasonable yield of nearly enantiomerically pure cis-epoxide 214 (only 10% of the trans epoxide is formed). MeO MeO NaOCl (R,R)-199 (5.5 mol%)

CO2i-Pr

CO2i-Pr 3 mol% Me

N

O

O 214; 62% yield, 96% ee

213

The epoxide is opened regioselectively with the anion of the nitrobenzene thiol 66 and the rest of the synthesis is essentially as described earlier (reduction, cyclisation, addition of substituents to O and N). OMe

NO2 NO2

S

214

S

NaHCO3

CO2i-Pr

SH OH

MeO 66

O

R 215

OMe

N O

62; diltiazem

558

25 Asymmetric Induction II Asymmetric Catalysis

Other improvements include the development of more sterically hindered ligands46 and changes in oxidant to mCPBA and NMO to give efficient asymmetric synthesis of monosubstituted epoxides.47 The asymmetric epoxidation of dimethylchromenes such as 216, a previously problematic task, is easy with Jacobsen epoxidation and these epoxides such as 217 have been used to make calcium channel antagonists 218. O

O NC

O

NC

Jacobsen epoxidation

N

Me

OH

ArOH

(S,S)-199 (3.7 mol%)

O

NC

N

O

O 217; 96% yield, 97% ee

216

218; 85% yield, 97% ee

PART VI – DESYMMETRISATION REACTIONS So far we have looked largely at substrates which are flat achiral objects that have chiral centres introduced by means of a reagent. Desymmetrisations are slightly different. Molecules which are desymmetrised tend to be of a meso type. That is, they are achiral because they have a mirror plane and the sides of the molecule contain left and right handed portions 219. This is in contrast to the C2 axis present in many catalysts such as the TADDOLate 218. Desymmetrisations are powerful because there may be several chiral centres embedded in an achiral molecule which suddenly become much more useful in the newly formed chiral molecule. There are a large variety of symmetrical substrates that have been enantioselectively desymmetrised48 and we look at a few of the more important classes. Ar Ar H O

O

Ti O

O

Oi-Pr

C2 axis

plane

(R) H

mesoanhydride 219

O

Oi-Pr

H Ar Ar

O

H

(S)

C2symmetric TADDOLate 218

O

Opening anhydrides For example, acid anhydride 219 is achiral and the molecule can be drawn above to highlight the mirror plane running through it. It has an R and an S chiral centre, one on either side of the mirror. The anhydride can be cleaved with Al(Oi-Pr)3 in a reaction catalysed by a titanium TADDOLate 218. The resulting ester49 220 is formed in 88% yield and 88% ee.

O

10 mol% TADDOLate 218 90 mol% (i-PrO)3Al

O O 219

THF, –20 ˚C

CO2H CO2i-Pr 220; 88% yield 88%ee

25 Scaling Up the Asymmetric Dihydroxylation

559

We have seen much of quinine and quinidine in this chapter alone. They are among those ‘privileged structures’ which seem to turn up time and again as being useful for asymmetric reactions. Oda50 had found that 10% of guinidine would catalyse the opening of anhydride 221. The trouble was that the ee of the ester 222 was modest (67%) and the reaction took 4 days. MeOH in Toluene room temperature

221 O

O

HO2C

10% Quinidine / 4 days

O

CO2Me

222 100% yield 67% ee

Bolm51 was able to improve matters by piling in the quinine (so that over an equivalent was used – 110%) which enabled the reaction of anhydride 223 to be done at a lower temperature (giving a better ee) in a mere 1.5 days. The pseudo-enantiomer quinidine could also be used to give ester 224. A range of rings 225 to 229 works. MeOH in toluene/CCl4 –50 ˚C

MeOH in toluene/CCl4 –50 ˚C

219

CO2H

CO2Me

110% quinidine 1.5 days

110% quinine

CO2H

CO2Me 223; 85% yield, 98%ee

224; 83% yield, 98%ee

Ph

O

OH O

O

O O

O

O

O

O

225

O

O

226

O

O

O

O

227

228

229

O

But matters were really significantly improved by Chen et al.52 If one cinchona alkaloid works, what about the double cinchona alkaloid ligands that are used in the asymmetric dihydroxylation reactions? Amazingly, these ligands work really well for desymmetrisations of anhydrides such as 230. Only 5% of catalyst is needed and they can be done at room temperature or ⫺20 ⬚C for better enantiomeric excess. Once again a range of anhydrides can be used 232–234. O 5 mol% (DHQD)2AQN

230

O O H 232

O

231; 85% ee 100% yield (based on SM consumed) or 93%yield, CO2Me 98% ee 231 CO2H

O

O O

H

MeOH / toluene room temperature 2 hours or MeOH / Et2O –20 ˚C

O O

233 O

O

234

O O

Opening epoxides Another privileged structure is the salen complex we have already seen as the catalyst 199 for Jacobsen epoxidations. Here the equivalent cobalt complex 235 is used to catalyse the opening of an achiral epoxide.53

560

25 Asymmetric Induction II Asymmetric Catalysis

H

H N

N Co

t-Bu

O

t-Bu

O

t-Bu t-Bu (S, S)-235

Cyclohexene oxide 236 is opened with benzoic acid catalysed by 2.5% of the cobalt salen complex 235 to give the asymmetric half ester 237. Although the ee is only 75% this is improved by recrystallisation to 98% ee. Epoxides on other ring sizes can also be opened enantioselectively. OH

OH

PhCO2H

oligomeric Co salen

O

2.5% Co catalyst 235

O

i-Pr2NEt 0– 4 ˚C

OH 238; 98% yield 94% ee

236

O

237 75% yield 98%ee

recrystallise

Ph

237; 98% yield, 77%ee

The (very useful) hydrolysis of cyclohexene oxide 236 to give the C2 symmetrical diol 238 is difficult with the standard Co salen 235. However, this reaction can be achieved in high yield and enantiomeric excess with a special oligomeric form of the salen.54 The oligomeric form simply contains several salen complexes connected together in each molecule. This has enhanced reactivity for certain reactions. Reactions which proceed exceptionally well with a dimeric form of the catalyst are believed to make use of two active sites and thus benefit from the entropic advantage of having them both in the same species.55

Comparison of desymmetrisation and kinetic resolutions As we shall see further in Chapter 28, a kinetic resolution is the more rapid reaction of one enantiomer of starting material over the other. In the absence of anything fancy (like a dynamic kinetic resolution) they are limited to 50% yield of product (or starting material). But because a desymmetrisation starts with one achiral molecule (instead of a pair of enantiomers) this limitation is removed as are other complications we face in kinetic resolutions such as the build up of the wrong enantiomer which makes selectivity more difficult. However, it is worth noting that desymmetrisation and kinetic resolutions are brothers in the stereochemical world. OH

OH

OH kinetic resolution

+

R-239

S-239 OH

OAc +

R-239

OH

S-240 OH

OAc

desymmetrisation

241

242

25 Scaling Up the Asymmetric Dihydroxylation

561

Consider, for example, a hypothetical kinetic resolution of enantiomers (R) and (S)-239 by esterification of one of them. Compare this to the esterification and desymmetrisation of diol 241. In many ways, a desymmetrisation is a kinetic resolution but with the two enantiomers joined together in the same molecule. The features that make a reagent work well in a desymmetrisation, may well make it work in a kinetic resolution. Indeed, in Double Methods in Chapter 28 we see that this is the case.

Mono esterification of diols Diol 241 is very enantioselectively acylated with 1% of Fu’s chiral56 version 243 of DMAP 244. The enantiomeric excess is a staggering 99.7% ee. Other, rather less fancy, chiral nucleophilic catalysts57 have been used very effectively including pyrrolidine derivative 247 for the asymmetric benzoylation of 245.

OH

Me

OAc

N

1 mol% 243

N

Me

241 Ac2O, 0 ˚C NEt3, t-AmylOH

N R Fe R

R 242; 91% yield 99.7% ee OH

243

N 244; DMAP

R

R

OH

Me

0.5 mol% 247

N OH

PhCOCl, Et 3N

N

OBz 246; 83% yield 96% ee

245

247

Bn

Me

PART VII – HETERO DIELS-ALDER REACTIONS Hetero-Diels-Alder reactions are not as straightforward as ordinary Diels-Alder reactions. In an ordinary (normal electron demand) Diels-Alder reaction we are accustomed to having an electronrich diene and an electron-deficient dienophile. In asymmetric catalysis, early successes with a hetero-Diels-Alder reaction58 typically needed a very electron rich diene, such as Danishefsky’s diene 248, or very electron deficient dienophiles like a glyoxylate 249. usual hetero-Diels-Alder reactions

very electron-rich dienes e.g. 248

Me3SiO

OEt and/or

OMe

R

O O

very electron-deficient dienophiles e.g. 249

In the example below with a modified Danishefsky’s diene59 250, the asymmetry comes from a menthyl auxiliary while the reaction is catalysed by a europium complex. The ee (determined on a subsequent fragment) is a modest 27% but can be dramatically improved by using an

562

25 Asymmetric Induction II Asymmetric Catalysis

optically pure catalyst, Eu(hfc)3, in the reaction together with the auxiliary. Excellent yield and enantiomeric excess have been achieved using a copper catalyst featuring a bisoxazoline ligand60 but again we see a very electron-rich diene 248 coupled with a very electron-deficient dienophile 252.

Me3SiO

Ph +

Eu(fod)3

Me3SiO

F

Ph

F

t-Bu

CF3

O

O

O OMen 251; 27% ee

OMen 250; Men = menthyl

F

O

F

fod = 6,6,7,7,8,8,8-heptafluoro -2,2-dimethyl-3,5-octanedionate

OMe O

O

1. 10% Cu bis-oxazoline –78 ˚C to –40 ˚C

OMe

COMe Me

O 2. H +

Me3SiO O 252

248

253; 96% yield 99% ee

A further complication is whether the HDA reactions really proceed by a pericyclic mechanism or whether they are stepwise reactions (aldol then cyclisation). Both are conceivable and possible.61 The salen complexes like those of Co 235 or Mn 199, but this time with chromium at their heart, catalyse the reaction of Danishefsky’s diene and benzaldehyde.62 The yield and ee are excellent and the reaction works with a variety of aldehyde dieneophiles and, although the ee’s are not always spectacular, the products can sometimes be recrystallised to high ee. The question of [4⫹2] versus aldol-then-cyclisation is also addressed. The intermediate compound that would be expected from the aldol reaction was found not to cyclise under the reaction conditions which indicates that the [4⫹2] mechanism is active.

Me

Me3SiO Ph

2. 2% Cr Salen –30 ˚C

O

Ph

MeO

H

Bu

+ O

2. H +

O

OMe 248

254; 85% yield 87% ee

?

O Me 255

256

This is all very well but what about using dienes which are more typically electron rich (with one oxygen substituent instead of two) 255 in combination with normal aliphatic aldehydes 256? A catalyst that could achieve this would be very useful. One solution is a modification of the chromium salen complex which replaces half the salen with an adamantyl group and the other half with cis-aminoindanol 209. The synthesis of this complex 258 is straightforward since commercially available compounds 257 and 209 are combined in high-yielding reactions and the complex itself is impressively enantioselective.58 The hexafluoroantimonate catalyst 260 was more enantioselective than the corresponding chloride 259.

25 Scaling Up the Asymmetric Dihydroxylation

563

Me Me

1. SnCl4, (CH2O)n 2,6-lutidine

1. CrCl2 2. air

2.

3. 2,6-lutidine

NH2

OH

N

OH

OH

OH 258; 88% yield from 257

209

257

Me

Me

AgSbF6

N

N

O

O

O Cr

Cr O

Cl

259; 99% yield

SbF6 260

The application of this catalyst is nicely illustrated in a synthesis of ambruticin.63 Ambruticin 261 is an antifungal compound. The tetrahydropyran on the left hand side of the molecule and the dihydropyran on the right both contain two stereogenic centres which could be controlled in a hetero Diels-Alder reaction via 262 and 263.

OH

OH H

O CO2H

261; ambruticin

OH

OH

Me

O

O

CO2H

O

Me

Me

262

Me

Me

O Me

OH

263

Me

The first HDA reaction is between aldehyde 264 and diene 265 at room temperature. The diene is reasonably electron rich and the aldehyde an ordinary aliphatic. Note that there are no stereocentres that arise merely from stereospecific reactions of the substrates. The pyran is formed in 97% ee and the subsequent hydroboration is regio- and diastereoselective to give alcohol 267 as a single diastereomer. The other pyran ring (for the right hand side of ambruticin)

room temperature

1. BH 3•THF

OR

10% ent-259

RO

O

O OBn

264; R=TBDPSi

265

OTBDMS

OTBDMS

OTBDMS

OBn 266; 64% yield, 97% ee

2. H 2O2 NaOH

OH

OR O

OBn 267; 94% yield

564

25 Asymmetric Induction II Asymmetric Catalysis

is made in a very similar manner but using the other enantiomer of catalyst 259. This time the yield was 87% and the ee ⬎99%! Several other interesting reactions are employed in the synthesis including an asymmetric cyclopropanation and an asymmetric hydroformylation but we’ll end by highlighting the KociénskiJulia olefination. Rather than employing the ordinary sulfone of the Julia reaction64 or the one-pot (Sylvestre) Julia method65 which employs benzothiazolyl sulfones, the Kociénski variant66 uses a tetrazole substituted sulfone 269 (chapter 15). The key advantage over the benzothiazolyl sulfone being that the aldehyde and sulfone do not need to be premixed before the addition of base. The E/Z selectivity by using LiHMDS in DMF/HMPA is very good at ⬎30:1 but with NaHMDS in THF the E/Z selectivity is reversed to 1:8. A lithiated tetrazolesulfone sporting a tert-butyl

2. Pd/C, H 2

N

OTBDMS

1. TBSOTf base, –30 ˚C

N

OTBDMS

OR

+

267

N Ph

3. NMO cat. TPAP

N

Me S

O

O

O

O CHO 268; R=TBDPSi

Me 269

Me

Me

LiHMDS DMF/HMPA –35 ˚C

OTBDMS OTBDMS

OR

Me

Me 2. TBAF

O

O Me Me 270; >30:1 E / Z

Me

Me

2. Pt, O 2 H2O, acetone

TM261 ambruticin

(as opposed to the phenyl group seen here) is even more stable.67 Returning to the synthesis of ambruticin – deprotection and oxidation of 270 gave the final product 261. Catalytic reactions look like the best bet for the future of asymmetric synthesis. In the next chapter you will see how C– H and C–C bonds can also be made by catalytic asymmetric reactions. The only limitation at the moment is the relatively small number of reactions that have been developed. Some of those you have met in this chapter – Sharpless AE and AD and Jacobsen epoxidation – are among the best.

References General references are given on page 893 1. R. A. Johnson and K. B. Sharpless, Ojima, Asymmetric, pages 231–285. 2. T. Katsuki, Ojima, Asymmetric, pages 287–325. 3. J. G. Hill and K. B. Sharpless, Org. Synth., 1985, 63, 66. 4. R. M. Hanson and K. B. Sharpless, J. Org. Chem., 1986, 51, 1922. 5. M. G. Finn and K. B. Sharpless, J. Am. Chem. Soc., 1992, 113, 113 and in Morrison, Vol. 5, pp. 247–308; R. A. Johnson and K. B. Sharpless in Ojima, Asymmetric, pp. 271–2. 6. J. M. Klunder, S. Y. Koo and K. B. Sharpless, J. Org. Chem., 1986, 51, 3710. 7. B. E. Rossiter, T. Katsuki and K. B. Sharpless, J. Am. Chem. Soc., 1981, 103, 464. 8. T. Katsuki, A. W. M. Lee, P. Ma, V. S. Martin, S. Masamune, K. B. Sharpless, D. Tuddenham and F. J. Walker, J. Org. Chem., 1982, 47, 1378. 9. A. Fürstner and O. R. Thiel, J. Org. Chem., 2000, 65, 1738.

25 References

565

10. H. Miyaoka, H. Shida, N. Yamada, H. Mitome and Y. Yamada, Tetrahedron Lett., 2002, 43, 2227. 11. Diltiazem by AE: J. Saunders, Top Drugs. 12. H. C. Kolb, M. S. VanNieuwenhze and K. B. Sharpless, Chem. Rev., 1994, 94, 2483. 13. K. B. Sharpless, Tetrahedron, 1994, 50, 4235. 14. R. Criegee, Liebig’s Ann. Chem., 1936, 522, 75; Angew. Chem., 1937, 50, 153; Angew. Chem., 1938, 51, 519; R. Criegee, B. Marchand and H. Wannowius, Liebig’s Ann. Chem., 1942, 550, 99. 15. E. N. Jacobsen, I. Markó, W. S. Mungall, G. Schröder and K. B. Sharpless, J. Am. Chem. Soc., 1988, 110, 1968. 16. K. B. Sharpless, W. Amberg, Y. L. Bennani, G. A. Crispino, J. Hartung, K.-S. Jeong, H.-L. Kwong, K. Morikawa, Z.-M. Wang, D. Xu and X.-L. Zhang, J. Org. Chem., 1992, 57, 2768. 17. G. A. Crispino, K.-S. Jeong, H. C. Kolb, Z.-M. Wang, D. Xu and K. B. Sharpless, J. Org. Chem., 1993, 58, 3785. 18. L. Wang and K. B. Sharpless, J. Am. Chem. Soc., 1992, 114, 7568. 19. M. Minato, K. Yamamoto and J. Tsuji, J. Org. Chem., 1990, 55, 766. 20. V. VanRheenen, R. C. Kelly and D. Y. Cha, Tetrahedron Lett., 1976, 1973. 21. H. Becker and K. B. Sharpless, Angew. Chem., Int. Ed., 1996, 35, 448. 22. B. B. Lohray, T. H. Kalantar, B. M. Kim, C. Y. Park, T. Shibata, J. S. M. Wai and K. B. Sharpless, Tetrahedron Lett., 1989, 30, 2041. 23. J. S. M. Wai, I. Markó, J. S. Svendsen, M. G. Finn, E. N. Jacobsen and K. B. Sharpless, J. Am. Chem. Soc., 1989, 111, 1123. 24. H.-L. Kwong, C. Sorato, Y. Ogino, H. Chen and K. B. Sharpless, Tetrahedron Lett., 1990, 31, 2999; K. P. M. Vanhessche, Z.-M. Wang and K. B. Sharpless, Tetrahedron Lett., 1994, 35, 3469; H. C. Kolb, Y. L. Bennani and K. B. Sharpless, Tetrahedron: Asymmetry, 1993, 4, 133. 25. H. C. Kolb, P. G. Andersson and K. B. Sharpless, J. Am. Chem. Soc., 1994, 116, 1278; K. B. Sharpless, W. Amberg, M. Beller, H. Chen, J. Hartung, Y. Kawanami, D. Lübben, E. Manoury, Y. Ogino, T. Shibata and T. Ukita, J. Org. Chem., 1991, 56, 4585. 26. M. Nambu and J. D. White, Chem. Comm., 1996, 1619. 27. H. Takahata, S. Kouno and T. Momose, Tetrahedron: Asymmetry, 1995, 6, 1085. 28. Y. Gao and K. B. Sharpless, J. Am. Chem. Soc., 1988, 110, 7538; B. B. Lohray, Synthesis, 1992, 1035. 29. R. Hirsenkorn, Tetrahedron Lett., 1990, 31, 7591. 30. H. Becker, M. A. Soler and K. B. Sharpless, Tetrahedron, 1995, 51, 1345. 31. S. C. Sinha and E. Keinan, J. Org. Chem., 1994, 59, 949. 32. E. J. Corey, M. C. Noe and S. Lin, Tetrahedron Lett., 1995, 36, 8741. 33. U. Hugger and K. B. Sharpless, Tetrahedron Lett., 1995, 36, 6603. 34. L. Ahgren and L. Sutin, Org. Process Res. Dev., 1997, 1, 425. 35. Z. Hu and P. W. Erhardt, Org. Process Res. Dev., 1997, 1, 387; X. Lu, Z. Xu and G. Yang, Org. Process Res. Dev., 2000, 4, 575. 36. V. V. Fokin and K. B. Sharpless, Angew. Chem. Int. Ed., 2001, 40, 3455; M. A. Andersson, R. Epple, V. V. Fokin and K. B. Sharpless, Angew. Chem. Int. Ed., 2002, 41, 472; C. Bolm, J. P. Hildebrand and K. Muñiz in Ojima, Asymmetric, page 399. 37. P. O’Brien, Angew. Chem. Int. Ed., 1999, 38, 326. 38. M. Bruncko, G. Schlingloff and K. B. Sharpless, Angew. Chem. Int. Ed., 1997, 36, 1483. 39. W. Harmann, H.-G. Heine and D. Wendisch, Tetrahedron Lett., 1977, 2263. 40. H. Kolb and K. B. Sharpless, Tetrahedron, 1992, 48, 10515. 41. Z.-M. Wang, X.-L. Zhang and K. B. Sharpless, Tetrahedron Lett., 1993, 34, 2267. 42. K. G. Watson, Y. M. Fung, M. Gredley, G. J. Bird, W. R. Jackson, H. Gountzos and B. R. Matthews, J. Chem. Soc., Chem. Commun., 1990, 1018. 43. J. F. Larrow and E. N. Jacobsen., Org. Synth., 1998, 75, 1. 44. E. N. Jacobsen, L. Deng, Y. Furukawa and L. E. Martínez, Tetrahedron, 1994, 50, 4323. 45. J. F. Larrow, E. Roberts, T. R. Verhoeven, K. M. Ryan, C. H. Senanayake, P. J. Reiderand E. N. Jacobsen, Org. Synth., 1999, 76, 46. 46. K.-H. Ahn, S. W. Park, S. Choi, H.-J. Kim and C. J. Moon, Tetrahedron Lett., 2001, 42, 2485. 47. M. Palucki, G. J. McCormick and E. N. Jacobsen, Tetrahedron Lett., 1995, 36, 5457; N. H. Lee, A. R. Muci and E. N. Jacobsen, Tetrahedron Lett., 1991, 32, 5055.

566

25 Asymmetric Induction II Asymmetric Catalysis

48. M. C. Willis, J. Chem. Soc., Perkin Trans. 1, 1999, 1765. 49. G. Jaeschke and D. Seebach, J. Org. Chem., 1998, 63, 1190. 50. J. Hiritaki, M. Inagaki, Y. Yamamoto and J. Oda, J. Chem. Soc., Perkin Trans. 1, 1987, 1053. 51. C. Bolm, A. Gerlach and C. L. Dinter, Synlett, 1999, 195. 52. Y. Chen, S.-K. Tian and L. Deng, J. Am. Chem. Soc., 2000, 122, 9542. 53. E. N. Jacobsen, F. Kakiuchi, R. G. Konsler, J. F. Larrow and M. Tokunaga, Tetrahedron Lett., 1997, 38, 773. 54. J. M. Ready and E. N. Jacobsen, J. Am. Chem. Soc., 2001, 123, 2687. 55. R. G. Konsler, J. Karl and E. N. Jacobsen, J. Am. Chem. Soc., 1998, 120, 10780. 56. J. C. Ruble, J. Tweddell and G. C. Fu, J. Org. Chem., 1998, 63, 2794. 57. T. Oriyama, K. Imai, T. Sano and T. Hosoya, Tetrahedron Lett., 1998, 39, 3529. 58. A. G. Dossetter, T. F. Jamison and E. N. Jacobsen, Angew. Chem. Int. Ed., 1999, 38, 2398. 59. M. Bednarski and S. Danishefsky, J. Am. Chem. Soc., 1983, 105, 6968. 60. S. Yao, M. Johannsen, H. Audrain, R. G. Hazell and K. A. Jørgensen, J. Am. Chem. Soc., 1998, 120, 8599. 61. Fleming, Orbitals p. 142 62. S. E. Schaus, J. Brånalt and E. N. Jacobsen, J. Org. Chem., 1998, 63, 403. 63. P. Liu and E. N. Jacobsen, J. Am. Chem. Soc., 2001, 123, 10772. 64. M. Julia and J.-M. Paris, Tetrahedron Lett., 1973, 4833. 65. J. B. Baudin, G. Hareau, S. A. Julia and O. Ruel, Tetrahedron Lett., 1991, 32, 1175. 66. P. R. Blakemore, W. J. Cole, P. J. Kocienski and A. Morely, Synlett, 1998, 26. 67. P. J. Kocienski, A. Bell and P. R. Blakemore, Synlett, 2000, 365.

26

Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds Introduction: Formation of C–H and C–C Bonds PART I – ASYMMETRIC FORMATION OF C–H BONDS Introduction: Catalytic hydrogenation with soluble catalysts Hydrogenation with C2 -Symmetrical bis-Phosphine Rhodium Complexes C2 symmetric ligands (DIPAMP, DIOP, PNNP) Reduction of 2-acylamino acrylates to give aminoacids Hydrogenation with C2 -Symmetrical BINAP Rh and Ru Complexes Asymmetric Hydrogenation of Carbonyl Groups Regioselective asymmetric hydrogenation of enones Asymmetric reduction of ketones with kinetic resolution A Commercial Synthesis of Menthol Noyori’s synthesis of menthol by Rh⫹-catalysed [1,3]H shifts Corey’s CBS Reduction of Ketones A synthesis of the H1 blocker cetirizine PART II – ASYMMETRIC FORMATION OF C–C BONDS Organic Catalysis The Robinson annelation The proline-catalysed aldol reaction Reactions with hydroxy-acetone Conjugate addition Catalysed asymmetric Baylis-Hillman reactions Catalysed Asymmetric Diels-Alder Reactions Organic catalysis Diels-Alder reactions catalysed by metal complexes C2-Symmetric Lewis acid catalysis Origins of enantioselectivity in catalysed Diels-Alder reactions Box and pybox catalysts in asymmetric Diels-Alder reactions Cyclopropanation Application to the synthesis of medicinal compounds Asymmetric Alkene Metathesis Asymmetric Pericyclic Additions to Carbonyl Groups 2 ⫹ 2 Ketene cycloadditions Synthesis of enalapril Carbonyl ene reactions Alder ene reactions

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

568

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

Nucleophilic Additions to Carbonyl Groups Allyl metals Alkynyl metals Addition of alkyl organo-zinc reagents Palladium Allyl Cation Complexes with Chiral Ligands Summary

Introduction: Formation of C–H and C–C Bonds The last chapter was mostly about the asymmetric introduction of functionality using catalytic reagents. The first part of this chapter concerns the modification of the carbon skeleton, or of existing functional groups, by asymmetric catalytic reduction, usually by hydrogenation with Rh or Ru catalysts. The second part is more varied: metals are again used for some reactions but there are also examples of an exciting modern approach: organic catalysis. New C–C bonds are formed by a variety of reactions ranging from cyclopropanation through Robinson annelation to alkene metathesis. All are asymmetric. A valuable general reference for this chapter and the last is Ojima’s book1 Catalytic Asymmetric Synthesis. It contains a useful list of ligands and their acronyms on pages 801-856 just before the index.

PART I – ASYMMETRIC FORMATION OF C–H BONDS Introduction: Catalytic hydrogenation with soluble catalysts Catalytic hydrogenation using hydrogen gas and insoluble palladium catalysts such as Pd/C will be well known to you already. This story starts with the invention of Wilkinson’s catalyst, the Rh(I) phosphine complex (Ph3P)3RhCl that can be used for catalytic hydrogenation in organic solvents such as benzene. It catalyses the reduction of simple alkenes such as 1 or conjugated alkenes such as 3 to give (racemic) products 2 and 4 having new chiral centres.2 (Ph3P)3RhCl H2

Ph

(Ph3P)3RhCl

H Ph

1

Ph

CO2H

2

H2

H Ph

3

CO2H 4

Hydrogenation using Wilkinson’s catalyst is still used for diastereoselective reduction as in the transformation of the functionalised enone 5 into the ketone 6 with high diastereoselectivity.3 Other (homogeneous or heterogeneous) catalysts were less effective. Me3Si

O

Me3Si

O

(Ph3P)3RhCl

Me

H2

H RO

5

P

benzene, 65 °C

H RO

6

7

26 Hydrogenation with C2-Symmetrical bis-Phosphine Rhodium Complexes

569

Horner4 at Mainz in Germany and Knowles5 at the Monsanto works in the USA realised at more or less the same time (1968) that if Ph3P were replaced by an enantiomerically pure phosphine such as 7, asymmetric induction might accompany the reduction. Unsymmetrical phosphines are enantiomerically stable, unlike the corresponding amines, as pyramidal inversion is much slower with P than with N. The phosphine was not available in high enantiomeric purity but Horner made compound 2 in 8% ee while Knowles made compound 4 in 15% ee. Such results would be scorned today but they were the first.

H2, catalyst 8

Me

P

Rh

Cl

3 MeOH, 25-80 °C

Pr Catalyst 8: 69% 'optical purity'

(S)-(+)15% ee

300-400 lb/in2

Hydrogenation with C2-Symmetrical bis-Phosphine Rhodium Complexes C2 symmetric ligands (DIPAMP, DIOP, PNNP) The real breakthrough came when Kagan6 realised that only one of the three phosphines in Wilkinson’s catalyst was displaced by the alkene substrate - alkenes, like phosphines, are two electron donor ligands. If the remaining two phosphines were linked together to form a more rigid chiral ligand, results should be much better. There is now a large selection of such catalysts, usually having C 2 symmetry, some chiral at both phosphorus atoms such as DIPAMP 9, some chiral in the backbone such as DIOP 10, derived from tartrate and Kagan’s first successful catalyst, or PNNP 11, derived from α-methylbenzylamine and chiral in the side chain.7 H

Ph

Ph P OMe

P MeO

9; ( S,S)-(+)-DIPAMP

O

PPh2

Ph

PPh2

O H 10; DIOP

N Ph2P

N

Ph

PPh2 11; PNNP

Reduction of 2-acylamino acrylates to give aminoacids One of the most famous examples is a general synthesis of aromatic amino acids 14, both natural like phenylalanine 14; Ar ⫽ Ph, and unnatural. The easily prepared N-acyl enamines 12 are hydrogenated in methanol catalysed by the rhodium-DIPAMP complex, to give the saturated amides 13 in almost perfect ee. Hydrolysis releases the amino acid 14. This process was discovered by Knowles at Monsanto. H N O

CO2H

H2, MeOH, 3 atmospheres 50 ˚C, ca 1 hour

Ar 12

H H N CO2H

(COD)2Rh (DIPAMP) ca 0.01 g per g substrate

O

Ar 13; 96-100% ee

amide hydrolysis

H2N

H

CO2H

Ar 14; 96-100% ee

570

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

Though it is often difficult to say exactly how asymmetric induction occurs, the complex probably adopts a C2 symmetric conformation 15 allowing the enamine to bind to the rhodium both at the alkene and at the amide group 16. Transfer of hydrogen from Rh to the bound alkene gives 13. The enantiotopic alkene faces in 12 become diastereotopic by complexation with the chiral catalyst.

MeO Ph P

Rh

P

12

Ph

H N

Me

OMe Ph

A P Rh

O

OMe

P

OMe Ar 15; (DIPAMP)Rh

Ph 16

Phenylalanine is manufactured for incorporation into the sweetener Aspartame™ (nutrasweet) 18 by a process using PNNP 11. This has the advantage that even lower catalyst loadings are possible and a high pressure is not needed. AcNH

CO2H

Ph

1. H 2, (PNNP)Rh (15,000:1), EtOH 2. MeOH, H

12; Ar = Ph

H2N

NH2

CO2Me HO2C

Ph

H N

CO2Me

O

17; Ph-OMe; 83% ee crystallise to 97% ee

Ph 18; aspartame

Hydrogenation with C2-Symmetrical BINAP Rh and Ru Complexes The first place in catalytic hydrogenation nowadays is taken by Rh or Ru complexes of BINAP. This ligand has axial chirality as the naphthalene rings cannot rotate past each other. These compounds were developed by Noyori, who with Knowles and Sharpless received the 2001 Nobel prize for their contributions to asymmetric synthesis. BINAP 20 is usually made from BINOL 19 and either 19 or 20 can be resolved. Rhodium complexes similar to those we have met include a molecule of cyclooctadiene and, as these are Rh(I) compounds, a counterion, often triflate 21. Both enantiomers of BINAP are available commercially.8 Ph OH

PPh2

OH

PPh2

P

Ph

Rh P Ph 19; (R)-(+)-BINOL

20; (R)-(+)-BINAP

Ph

21; [(R)-(+)-BINAP(COD)Rh]OTf

Though BINAP is almost always represented as 20, it is important to realise that 20a is an equally good representation. The two naphthalene rings are not co-planar and can rotate by less than 90⬚ around the bond joining them. The true structure is more like 20b (seen looking along the bond joining the two naphthalenes).

26 Hydrogenation with C2-Symmetrical BINAP Rh and Ru Complexes

PPh2

571

Ph2P

PPh2

Ph2P

PPh2

20; (R)-(+)-BINAP

20b; (R)-(+)BINAP

PPh2

20a; (R)-(+)-BINAP

Rhodium complexes of BINAP can be used in much the same way as those we have just seen. In the synthesis of the Merck HIV protease inhibitor crixivan an enantiomerically pure piperazine 23 essential for activity was prepared by reduction of an N-acyl enamine 22 with 2% of catalyst 21. Though the catalyst loading is relatively high, the essentially perfect yield and enantiomeric purity of the product make this worthwhile.9 O t-BuO

O H2, 2% catalyst 21 70 bar, MeOH

N N

t-BuNH

O

t-BuO

N

OBn

N

O

t-BuNH

22

O

OBn O

23; 96% yield; 99% ee

The achievements of the Noyori group using ruthenium BINAP catalysts are extensive and remarkable. Simple alkenes are reduced whether conjugated or not: the very efficient synthesis of the analgesic and anti-inflammatory naproxen 25 from an unsaturated acid 24 resembling 1 is a good example.

H2

CO2H

CO2H Ru(BINAP)2

MeO

MeO 25; Naproxen: 97% ee

24

The presence of functionality does not disturb these excellent results. The cyclic enol ester 26 is cleanly reduced without reduction of the lactone. O 26

O

O H2

O

27; 94% ee

(S)-BINAP-Ru

Reduction of geraniol 28 shows remarkable chemoselectivity: only the allylic alcohol is reduced to give citronellol 29 of higher enantiomeric excess than that found naturally. The catalyst is used at 7 mol%: that means 3 mg/g of geraniol. The catalyst must have acetate as the counterion or it is not so chemoselective.10 You will see an alternative synthesis of the other enantiomer of citronellal shortly.

572

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

H2 Ru(OAc) 2[(R)-BINAP]

OH

OH MeOH, H 2O 90 atmospheres

28; geraniol

29; (S)-(–)-citronellol

20 ˚C, 8-16 hours

93-97% yield, 98% ee

Asymmetric Hydrogenation of Carbonyl Groups It is more difficult to hydrogenate carbonyl groups than alkenes as the C⫽O group is considerably more stable thermodynamically than the C⫽C bond. Asymmetric hydrogenation of ketones with Ru-BINAP catalysts is possible, though increased pressure and higher temperatures are usually needed. Hydrogenation of β-ketoesters 30 is a general reaction giving good yields and excellent chemo- and enantio-selectivity.11

O

H2

O

R

OH

RuCl2[(R)-BINAP]

OMe 30

O

R OMe 31; >90% yield 95-100% ee

MeOH 4 atmospheres 100 ˚C, 6 hours

R = alkyl, aryl, alkyl with O- or N-substituents

Both carbonyl groups of β-diketones 32 are reduced and the two reductions are independently enantioselective in the same sense so that the chiral (C2 symmetric) product 33 predominates by 99:1 over the meso compound 34. This makes symmetrical 1,3-diols such as 33 readily available.12 Elsewhere you will meet methods that use one centre to induce another in a 1,3-relationship, maybe by diastereoselective reduction - this is a quite different strategy. O

O

H2, 50 kgw/cm 2

OH

OH

OH

RuCl2[(R)-BINAP] MeOH 50 ˚C, 20 hours

32

OH

+ 33; 98% yield, >99% ee

34

99:1 anti:syn 33:34

Regioselective asymmetric hydrogenation of enones The situation is different when both C⫽C and C⫽O groups are present in the same molecule. We have seen the rather special case of the enol ester 26 and the usual result with conjugated enones is also that reduction of the alkene is preferred. High pressures are needed but good enantiomeric excess can be achieved with unsaturated lactones 35 and ketones 37. The lactone needs the RuCl2 catalyst but the ketone gives better results with the corresponding acetate.13 Notice that use of each enantiomer of BINAP gives a different enantiomer of the product 36 or 38.

O

100 atmospheres H2 RuCl2[(S)-BINAP]

O

CH2Cl2

35

50 ˚C, 60 hours

O

O

O 100 atmospheres H2

Bu O

Ru(OAc) 2[(R)-BINAP]

Bu

CH2Cl2

36; 100% yield 95% ee

37

50 ˚C, 60 hours

38; 100% yield 98% ee

26 A Commercial Synthesis of Menthol

573

When asymmetric reduction of the ketone of an enone such as 39 is wanted, the catalyst must be further modified14 by using the very crowded BINAP derivative XYLBINAP 41 and the diamine additive DAIPEN 42. The results are then excellent but this example makes the point that it is essential to follow closely related examples when attempting the asymmetric reduction of any but the simplest alkene.

O

OH

10 atmospheres H2 RuCl2[(S)-XYLBINAP].[(S)-DIAPEN] 10,000:1 substrate:catalyst i-PrOH, K 2CO3, 30 ˚C, 43 hours

39

40; 99% yield, 100% ee OMe

P

2

RuCl2

OMe

H2N

P H2N 2

41; (S)-XYLBINAP

42; (S)-DAIPEN

Asymmetric reduction of ketones with kinetic resolution Since β-keto-esters enolise rapidly and extensively, dynamic kinetic resolution is possible in which one enantiomer is reduced more rapidly than the other.11 By this means two chiral centres may be developed simultaneously especially when there is an N-acetyl substituent on the middle atom as in 43. The synthesis of unusual functionalised amino acids related to threonine becomes straightforward. A fuller discussion of dynamic kinetic resolution appears in chapter 28.

O O

NHAc

O 43

OH

H2

CO2Me

RuCl2[(R)-BINAP]

O

CH2Cl2 100 atmospheres

O

20 ˚C, 6 hours

CO2Me NHAc

44; 99:1 syn:anti, 94% ee

A Commercial Synthesis of Menthol One of Noyori’s most remarkable achievements is a commercial synthesis of (⫺)-menthol 51 used since 1983 by the Takasago International Corporation on a scale of thousands of tonnes a year. This and related processes are discussed in detail by S. Akutagawa and K. Tani in chapter 3 of Ojima’s Catalytic Asymmetric Synthesis. The process is summarised here:

574

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

known reaction

1. Et2N

Rh [(S)-BINAP]2 10,000:1

NEt2

Li

100 ˚C, 15 hours 7 tonne batches

2. H 2O

45; β-pinene

46; myrcene

H

H

47

CHO

NEt2

ZnCl2

H

H2

OH

H2O

50

49; citronellal

48; 100% yield, 98% ee

OH

catalyst

51; (–)-menthol

Noyori’s synthesis of menthol by Rh⫹-catalysed [1,3]H shifts Various natural terpenes, including β-pinene 45, can be used to make myrcene 46. Addition of lithium diethylamide to the diene portion of 46 gives the achiral allylic amine 47. Now comes the asymmetric catalytic process: a [1,3]H shift rather than a hydrogenation giving the enamine 48 easily hydrolysed to citronellal of higher enantiomeric purity than can be found naturally. The [1,3]H shift probably involves complexation of Rh to the nitrogen atom 52, removal of hydrogen to create an imine complex 53, transformation to an allyl cation Rh complex 54, rather like the Pd complexes we met in chapter 19, and replacement of the original H atom to give the enamine 55, more stable because of conjugation than the original allylic amine 47. The decisive moment is the selection of one of the enantiotopic H atoms (marked in 52) so that the Rh atom occupies the lower face of 53–55 and returns the H atom to that same face. H 47

Rh

H R

R

Rh

Rh H 52

48

R

R

NEt2

N Et2

H

Rh

Rh NEt2

NEt2

53

54

55

The rest of the synthesis involves a Zn(II)-catalysed oxo-ene reaction 56 going through a chair transition state 57 to put all the substituents equatorial in the product 50 and thus create the relatively remote two remaining chiral centres. A classical hydrogenation gives the iso-propyl group. Menthol itself is in great demand but Takasago also find a market for the other terpene intermediates such as citronellal.

49

O

ZnCl2 O

H

H 56

OH

57

50

26 Corey’s CBS Reduction of Ketones

575

Corey’s CBS Reduction of Ketones Among the very few worthwhile methods for catalytic reduction that do not depend on hydrogenation is Corey’s CBS15 reagent 60, derived from proline 58 by protection and addition of phenyl Grignard, giving the bulky amino-alcohol 59. This is converted into the stable crystalline boron compound 60. H

Ph H

CO2H

OH NH

NH

58 (S)-L-(–)-proline

59

Ph H

Ph

Ph

MeB(OH)2 toluene Dean Stark

N

O B Me

60

Even this method is only semi-catalytic as usually about 10-15% of the reagent 60 is used with stoichiometric borane, to give 61 which reduces simple ketones with high enantioselectivity to give 62. The advantage of this method is that reduction of carbonyl groups is the favoured reaction and takes place under mild conditions - not the situation with hydrogenation. H

Ph

O

N

H

Ph

B

Ph

BH3.THF

N

Ph

O

O

OH Ph

Ph

B BH3 Me

THF

Me 60; 10 mol %

62; 99% yield 96.5% ee

61

This 61 is a genuine chiral borohydride, but presumably relies on complexation of the ketone oxygen by the other boron atom 63. In general, high enantioselectivity on acyclic substrates requires a rigid complex such as 63 to stop the substrate rotating. It is obviously no use ensuring delivery of H from, say, the top face of the molecule if the molecule is rotating and the “top” face is continually changing. Intramolecular transfer of hydride gives the boron ester 65 of the product 62 and returns the catalyst 60 ready for reaction with another molecule of borane. Ph H

Ph H

Ph O

N B

61 H

RL

Ph H

Ph

O N B

Me

O

B H H

Ph

Rs

H B O H H RL Rs

63

N B

Me

O

Me

64

60

+ H2B H

62; RL = Ph Rs = H

O

RL Rs 65

The general rule is expressed in 65 using RL and RS to show the large and small substituents on the ketone. Proline is naturally and cheaply available only as one enantiomer so it is interesting to note that this most general of asymmetric ketone reductions generally gives the opposite enantiomer to the most general of enzymatic reductions, that with baker’s yeast. (see chapter 29) O

OH

RS

OH

baker's yeast

RL

sucrose

CBS

RS

RL

catalyst from (S)-pro

RS

RL

576

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

A synthesis of the H1 blocker cetirizine The synthesis of the H1 blocker cetirizine 66 is an interesting application of this reagent. Cetirizine can obviously be made by substitution on some derivative of one enantiomer of the alcohol 68 by the piperazine 67. Asymmetric reduction of the ketone 69 might provide 68 but the difference in size between the two substituents is altogether insufficient for asymmetric induction. O

CO2H

N

CO2H

O N

C–N

N

N H

67

O

OH

+

Cl

Cl

Cl 68

66; Cetirizine

69

Corey’s solution15 was to make one of the aromatic rings much larger than the other by complexation with Cr(CO)3 70. This is a very large substituent as the three CO molecules stretch out linearly from the Cr atom. To get full value from this substituent, he used it to direct lithiation and acylation to give the ketone 71 with one ring definitely much larger than the other and reduction with the CBS reagent 60 using catechol borane as the stoichiometric reducing agent gave the alcohol 72 in excellent yield and ee. OH

O 1. n-BuLi TMEDA, THF

15 mol% 60

2. CuBr•Me 2S

Cr(CO)3

Cl

3. ArCOCl

O

Cl B

Cr(CO)3

H

Cr(CO)3

O

70

71; 78% yield

72; 99% yield; 98% ee

You may have noticed that this appears to be the wrong enantiomer for cetirizine. Not so as the Cr(CO)3 has a final job to do. It stabilises cations and this substitution is likely to occur by the SN1 mechanism as the doubly benzylic cation is so stable. Treatment of 72 with HBF4, a strong acid with a non-nucleophilic counterion, creates the cation. The Cr(CO)3 sits on one side of this planar system and preserves its chirality. The nucleophile attacks on the face not occupied by Cr(CO)3 and the substitution occurs with retention. O

CO2t-Bu

CO2t-Bu

N

N 1. HBF 4 CH2Cl2

72

O

pyridine

HCl

66 86% yield

N

N 2. 67 as t-Bu ester

Cl

Cl

Cr(CO)3 73; 86% yield; 98% ee

74; 92% yield; 98% ee

26 Organic Catalysis

577

Nowadays much less catalyst is needed. This is particularly important in large scale syntheses as of the intermediate needed by Merck16 for the drug intermediate 79. The starting material is ketone 75 easily made by a Friedel-Crafts reaction. Reduction using only 0.5% CBS catalyst 60 gave alcohol 76 of 98.9% ee. Even with only 0.1% 60, the ee of the product 76 was still 94.2%. F

F

F t-BuNH2

(S)-60

O F

OH

Et3N-BH3

F

Cl

75

OH

MeOH

F

Cl

76

NHt-Bu

77

The synthesis continued with displacement of Cl by an amine, actually via an epoxide, to give 77, conjugate addition of acrylonitrile 78 and ring closure with inversion and high stereoselectivity at the nitrile centre. F

F

CN

OH

(EtO)2P

77

F

CN

O

CN

Cl

F

LiHMDS

N R

78; R = t-Bu

N R

79; R = t-Bu

PART II – ASYMMETRIC FORMATION OF C–C BONDS Metal complexes remain with us in the formation of C–C bonds, indeed a wider variety of metals as Al, Co, Cu, Mo, Ru, Ti and Zn all have parts to play in a wide variety of reactions. But we must also introduce a new type of catalyst – the simple organic compound – and we must start with organic catalysis as it is involved in some of the same reactions as the metal complexes.

Organic Catalysis The Robinson annelation Pride of place here must go to the asymmetric Robinson annelation, discovered as long ago as 1974, but not appreciated at the time for the landmark that it undoubtedly was.17 The normal Robinson annelation (text chapter 36) is first a Michael addition to produce an achiral triketone 82 which cyclises when the enol of the methyl ketone adds, aldol fashion, to the syn face of either of the other ketones to give 83, and hence the enone 84, the north east corner of a steroid.

H + O 80 achiral

N H

H2O

O 81 achiral

O

O

O

20 °C 5 days

O

82; achiral 88% yield

TsOH

O

O HOAc, pH 7 3 days

O

OH

83; 100% yield all syn but racemic

O 84

578

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

The top and bottom faces of the five-membered ring in 82 are diastereotopic and the intramolecular reaction selects the syn face. But the two ketone groups in 82 are enantiotopic. If the enol attacks the right hand ketone one enantiomer of 83 is formed while if it attacks the left hand ketone, the other enantiomer is formed. O

Me O

O

O

OH

O

O

Me O

HO

O

OH one enantiomer

OH

the other enantiomer

The reaction is normally catalysed by weak acids and weak bases, such as piperidine and acetic acid. Replacing these with a catalytic (50:1) amount of proline 58 led to optically active 83 in an astonishing 93.4% ee and 100% yield. Dehydration gave the enone 84 with some loss of ee (by reverse aldol?), but recrystallisation from ether gave 100% ee. O

O

O

O 35 mg 58 (S)-(+)-proline

O

N H

82; 2 g scale

CO2H H

TsOH

O

O

OH (+)-aldol 83 100% yield, 93.4% ee

(+)-enone 84 99.4% yield, 88% ee

The chiral auxiliary must obviously be involved intimately with the cyclisation, perhaps by forming an imine 85 or an enamine 86, so that only the enantiomer 83 of the cyclisation mechanism applies. In both these proposals, the free CO2H group of proline interacts by hydrogen bonding with a ketone or OH group. Explanations involving two molecules of proline have now been discredited and the most recent explanation18 is that the enamine mechanism 86 is correct with that step as the rate determining step, closely followed by addition of water to the iminium ion 87. The transition state has a chair-like central portion 88. Me

Me N

O

HO O 85

O

O

HO

OH

O

O

86

O 83

O

N

H

O N

87 (–) H O

O

O N (+)

O 88

Proline is now recognised as a first class chiral auxiliary, having a rigid structure and a high grade stereogenic centre carrying contrasting sterically and electronically demanding substituents.

26 Organic Catalysis

579

This was an extraordinary result when it was published in 1974 as it was well ahead of its time. It is still one of the best asymmetric reactions with catalytic amounts of a chiral reagent. An example of an asymmetric Robinson annelation using stoichiometric ‘catalyst’ is in the workbook.

The proline-catalysed aldol reaction The asymmetric Robinson annelation relies on an intramolecular aldol reaction to create the new chiral centre. More recently List19 and MacMillan20 have used proline 58 to catalyse intermolecular aldol reactions with nearly as good results. Acetone and isobutyraldehyde 89 can be condensed to give a single enantiomer of the aldol 90 in excellent yield and ee providing 30% proline is used as catalyst. O

O

O

30 mol % proline

OH

+

N H

4:1 DMSO/acetone

H

room temperature

89

90; 97% yield, 96% ee

CO2H H

58; ( S)-(–)-proline

Acetone, the component that must enolise, is present in large excess but the achievement is considerable. The reaction involves formation of the proline enamine of acetone 91 which then attacks the aldehyde through a chair-like transition state 92 held together by the acidic proton of proline’s carboxylic acid. This gives the imine salt 93 hydrolysed to the product with regeneration of proline. The intermediates are like those in the Robinson annelation: enamines and imines. Organic catalysis with amines relies on equilibria between these intermediates and carbonyl compounds.

acetone + proline

N

CO2H

CO2 OH

RCHO

N

N O

R

H O

H

91; proline-acetone enamine

O

H2O

84 + proline

93

92

At the moment the limit is reached when both components in a crossed aldol reaction are enolisable aldehydes. One is indeed propionaldehyde 94 - a compound notorious for self-condensation - that reacts cleanly with isobutyraldehyde 89. However, it is necessary to add the propionaldehyde slowly by syringe pump. The aldol 95 is produced in very high yield considering the similarity of the two aldehydes but the most amazing aspect is the very high diastereo- and enantioselectivity.20 O

O +

H

O

OH

DMF, 4 ˚C

H

94

10 mol % proline

89

95; 82% yield, >99% ee 24:1 anti:syn

Reactions with hydroxy-acetone It turns out that one of the best ketones for these asymmetric crossed aldol reactions is hydroxyacetone 96. Combination with isobutyraldehyde 89 gives an aldol that is also an anti-diol 97 with almost perfect selectivity.21 The proline enamine of hydroxyacetone is evidently formed preferentially on the hydroxy side. You will recall from chapter 25 that asymmetric synthesis of anti-diols is not as easy as that of syn diols.

580

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

O

O +

96

OH

O

OH

97 82% yield, >99% ee >20:1 anti:syn

20-30 mol % proline

H

DMSO

OH

room temperature

89

Hydroxyacetone 96 is a reagent in an even more remarkable reaction: the asymmetric direct three-component Mannich reaction. It is combined with an aromatic amine 98 and the inevitable isobutyraldehyde 89 with proline catalysis to give a very high yield of a compound 99 that might have been made by an asymmetric amino-hydroxylation. The proline enamine of hydroxyacetone, must react with the imine salt formed from the amine and isobutyraldehyde. This is a formidable organisation in the asymmetric step. OMe

O +

+ H2N

OH 96

O

20 mol% (S)-proline

OMe OHC

DMSO room temperature

OH

89

98

HN

99; 92% yield, >99% ee >97:3 syn:anti

There is something different here. The absolute stereochemistry at the OH group (the one that comes from hydroxyacetone) is the same in 99 as it was in 97 but the relative stereochemistry is different: anti in 99 but syn in 97. The electrophile (ketone in 97 or imine in 99) must approach the proline enamine in different ways. List’s suggestion is that the large N-aryl group prefers to keep away from the rest of the molecule in the transition state 100 leading to 99 but that the side chain on the aldehyde is more important in the transition state 101 leading to 97. The dotted arrows in 100 and 101 show where that clash would come and the black dots mark the atoms that join to form the new bond. There is a review of the catalytic asymmetric aldol reaction that includes material from other chapters.22 O Ar

NHAr

O

N

N H

O

O

O

N H

O

OH

O

R

R H

OH 99

HO

HO R 100

H OH

R 101

97

Conjugate addition α,β-Unsaturated carbonyl compounds also reversibly form iminium salts with secondary amines and that offers opportunities for asymmetric conjugate addition and, as we shall soon see, asymmetric Diels-Alder reactions. The pioneer in the development of modern organic catalysis has been Dave MacMillan23 at Berkeley and he reported the asymmetric conjugate addition of heterocycles such as pyrroles to enals catalysed by the amine 102 prepared from phenylalanine. The amine forms an iminium salt 103 reversibly and this salt is more electrophilic than the enal.

26 Organic Catalysis

O

Me

CO2H Bn

Bn

phenylalanine

Bn

CHO Ph

N

NH2

581

Ph

Me N H

O

N

Me

N

Me

Me

Me

102

103

When N-methyl-pyrrole 104 is the nucleophile, clean conjugate addition occurs at its α-position with good ee in the product. Conjugate addition is normally a problem with aldehydes (chapter 9) so both regio- and enantioselectivity are impressive. The catalyst loading is quite high but clearly the catalyst can detach itself quickly enough from the first formed adduct to maintain a concentration of 103 large enough to swamp the slower reaction of 104 with the enal. 20 mol% catalyst 102

+

N

Ph

CHO CF3CO2H, –30 ˚C

Me

CHO

N Me

Ph

105; 87% yield; 93% ee

104

A new catalyst 106, also derived from phenylalanine but with extra chirality, is needed for conjugate additions to crotonaldehyde. Nucleophilic benzene rings such as 107 react with good regio- and stereoselectivity24 in the same sense as 105. O

OMe

Me

OMe Me CHO

N Bn

CHO H

N H

10 mol% catalyst

t-Bu

106

Me2N

Me2N

CH2Cl2, –40 ˚C

108; 86% yield, 89% ee

107

This method has been applied to the synthesis of a COX-2 inhibitor 110, one of a new generation of analgesics under development by the Merck company. This requires conjugate addition to the 3-position of an indole 109 and the same catalyst is used.25 MeO

CHO

1.

20 mol% catalyst 106

N

MeO

CO2H N

CH2Cl2, i-PrOH –40 ˚C 2. [O]

Br 109

Br 110; 82% yield, 87% ee

Catalysed asymmetric Baylis-Hillman reactions Other reactions are possible with iminium salts derived from α,β-unsaturated carbonyl compounds. The Baylis-Hillman reaction (chapter 18) also uses conjugate addition as a first step and we shall deal with that immediately. Much more work has been done on the Diels-Alder reaction: that will follow. The Baylis-Hillman reaction between an electrophilic α,β-unsaturated carbonyl compound and an

582

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

electrophilic aldehyde is catalysed by tertiary amines and lowish yields are generally acceptable as so much is achieved in a one-pot reaction. The obvious source of asymmetry is a chiral amine catalyst but only recently has the right amine and the right α,β-unsaturated carbonyl compound been identified.26 OH O

O

10 mol% QD4 (113)

CF3

OH

O

CF3

O

+ R

O

CF3

DMF –55 °C

111

R

O

CF3

N N

112; ~50% yield, 91-99% ee

113; catalyst 'QD4'

The unusual catalyst 113 must add to the unusual ester 111 in a reversible conjugate addition 114 to give the enolate that adds to the aldehyde in the asymmetric step 116. The bicyclic amine must be placed close to the carbonyl group of the aldehyde: Hatakeyama suggests an H-bonding interaction with the OH group on the quinoline ring of the catalyst. Finally, elimination of the catalyst launches a second cycle. The next few years are likely to see considerable development here. O

O

CF3 O

O

R

CF3

R*3N

CF3 O

O

CF3

R

R*3N 114

CF3 O

R*3N

115

O

CF3

112

116

Catalysed Asymmetric Diels-Alder Reactions Organic catalysis This is an important area with many available methods. We shall look first at organic catalysis and then change to catalysis by metal complexes. The same type of intermediate 117 used for conjugate addition is clearly also suitable for Diels-Alder reactions with the same proviso: it must be more reactive then the α,β-unsaturated carbonyl compound as that too can do Diels-Alder reactions. And of course the first formed product 118 must hydrolyse rapidly to release the catalyst 102. Bn 102

CHO

R

R

O

N N

Me Me

Me

117 Bn

R

R O

N

CHO + 102

N

Me Me

Me

118

An early example was the reaction between cyclohexadiene and cinnamaldehyde catalysed by the phenylalanine-derived amine 102 to give the endo adduct 119 with good selectivity of all kinds.23

26 Catalysed Asymmetric Diels-Alder Reactions

5 mol% catalyst 102

+

119 82% yield 94% ee 1:14 exo:endo

Ph

CHO

Ph

583

MeOH, H 2O, 21 hours room temperature

CHO

More recent developments include the refined catalyst 123 with extra bulk and an extra chiral centre. This catalyses reactions between functionalised dienes such as 120 and acyclic enones such as 121. The very high yields, ees, and endo:exo selectivity of the product 122 compensate for the 20% catalyst loading. It is very difficult to get high ees in Diels-Alder reactions with acyclic ketones and you will see in the next section why this catalyst is successful.27

O

+

20 mol% HClO4 EtOH, –30 ˚C

NHCbz 120

CbzNH O 122

121

O

122 91% yield 98%ee, >100:1 endo:exo

20 mol% catalyst 123

Me N

Bn

O N Me H catalyst 123

Diels-Alder reactions catalysed by metal complexes C2-Symmetric Lewis acid catalysis The best results before organic catalysis were with amides 125 and Lewis acid catalysts based on Al, Ti, and Cu(II) with C2-symmetric ligands. Corey’s aluminium complex 127 derived from the diamine whose resolution was described in chapter 22 works well with substituted cyclopentadienes 124 and the product 126 was used in prostaglandin synthesis.28 There are three aspects of stereoselectivity in this reaction: which diastereotopic face of 124 is attacked? (that anti to the CH2OBn group), is the product exo or endo? (endo) and which endo product is formed, 126 or its enantiomer? Only for the last question is asymmetric catalysis necessary, though Lewis acid catalysis of any kind enhances endo/exo selectivity. OBn +

N

Ph

BnO

O

O

10 mol% catalyst 127

O

O O

124

126 94% yield 95% ee endo only

N

Ph

TfN

NTf Al

Me catalyst 127

O

125

Seebach’s TADDOL auxiliaries, derived from tartrate (chapter 23), combine well with Ti(IV) to make an effective Lewis acid catalyst 130 for Diels-Alder reactions. The reaction of isoprene with the doubly activated amide dienophile 128 gives one adduct 129 in good yield.29 Polymer supported versions of this catalyst are available. CO2Me +

CO2Me

Ph O

N

Ph Me

O isoprene + amide 128

10 mol% TADDOL Ti catalyst 130

N

O

O

O

Ti O

129; 94% ee

Cl Cl

O Ph

O O

O

Ph

Ph TADDOL-Ti catalyst 130

584

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

Origins of enantioselectivity in catalysed Diels-Alder reactions You may be wondering why these dienophiles 125 and 128 have a heterocyclic ring attached to the carbonyl group. Later in chapter 27 you will see that chiral versions of this oxazolidinone are essential for other Diels-Alder reactions. But this ring is not chiral and its role is to chelate the metal 131. By this means the two lone pairs on the ketone oxygen atom are distinguished and the asymmetry of the catalyst activates one face of the alkene rather than the other. Such a distinction is not possible for a ketone such as 121 and that is why organic catalysts such as 123 are much more effective. The iminium salt 132 shows the same kind of shape as the complex 131. In both these diagrams 131 and 132 the orientation of the alkene cannot be stated for certain. L

lone pair discrimination

O

O

L M

catalyst 123

O

O

Bn N

Me N N

O 121

131

O Me

132

Box and pybox catalysts in asymmetric Diels-Alder reactions Another important group of ligands for metal-catalysed reactions are based on oxazolines derived from amino acids. Evans30 has used the bis-oxazoline or ‘box’ ligands in Cu(II)-catalysed DielsAlder reactions. Ligand 133 is derived from the unnatural amino-acid tert-leucine and, complexed with Cu(OTf)2, catalyses Diels-Alder reactions of the same dienophile 125 with various dienes. The best counterions are triflate (OTf) and SbF6.

N

+

N

Cu(II) 10 mol% box 127

O

O

O

O

N

O

–50 ˚C

O O

N

O

t-Bu

t-Bu

bis-oxazoline ligand 133

134; 97% ee 98:2 endo:exo

125

The more highly chelating pybox ligands such as 135, also prepared from tert-leucine, catalyse Diels-Alder reactions even with aldehydes such as 136. Now the counterion really must be the very non-nucleophilic SbF6. In these examples diastereo- (endo/exo) and enantioselectivity are excellent.31

Cu(SbF6)2 5 mol% pybox 135

O O

O

N N

+

N

H –40 ˚C

t-Bu

t-Bu pybox ligand 135

136

O

H

137; 92% ee 97:3 endo:exo

26 Cyclopropanation

585

Cyclopropanation Historically one of the first asymmetric methods to be explored, cyclopropanation came of age32 with box and salen ligands on Cu(I). Diazo compounds, particularly diazoesters 138, react with Cu(I) to give carbene complexes 140 that add to alkenes, particularly electron-rich alkenes to give cyclopropanes 141. The reaction is stereospecific with respect to the alkene - trans alkenes giving trans cyclopropanes - and reasonably stereoselective as far as the third centre is concerned. Any enantioselectivity comes from the chiral ligand L*. You have already seen the Ru carbene complexes are intermediates in olefin metathesis (chapter 15). N

N

CO2Et

N L*Cu(I)

H

H

N

CO2Et

–N2

L*Cu

138

H

CO2Et

L*Cu

139

CO2Et

R R

R

R

L*Cu(I)

140

141

A good test case is styrene. With the box ligand 133 the best diazoester is the acetate of ‘BHT’, 2,6-di-t-butyl-4-methylphenol. This gives excellent diastereoselectivity and near perfect ee in the trans-cyclopropane30 142. t-Bu

CO2R

Me N2

R=

Ph

O

Ph 142; trans:cis 94:6 trans: 99% ee

CuOTf bis-oxazoline 133

styrene

t-Bu

('BHT')

CO2R

The early salen-type catalysts 143 of Aratani gave moderate results when menthyl diazoacetate was used,32 but more modern versions, such as the salen-Co catalyst 145 of Katsuki33 are much more impressive. You saw such catalysts in the last chapter under Jacobsen epoxidation. CO2M* N O

Cu

N2

R O

Ph

R

143

Aratani catalysts 143

M* = menthyl

N

CO2t-Bu

N

N2

Co O

Ph 144; trans:cis 86:14 trans; 70-80% ee

Ph

Ph

MeO

CO2M*

Br

O

OMe

CO2t-Bu

Ph

Katsuki salen catalyst 145

Ph Co-salen 145

146; trans:cis 96:4 trans; 93% ee

Application to the synthesis of medicinal compounds Cilastatin 147 is used to increase the potency of β-lactam antibiotics. It is manufactured commercially by asymmetric cyclopropanation of isobutylene with ethyl diazoacetate catalysed by the Aretani Cu(I) complex 143 that is more effective with this trisubstituted alkene than with styrene.34

586

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

H N O

CO2

S CO2

N2CHCO2Et

CO2Et

143

NH3

148; 92% ee

147

The antidepressant tranylcypromine 151 was made using the Evans box ligand 133. A different bulky ester 149 gave good ee but only moderate diastereoselectivity.35 The trans diastereoisomer hydrolyses faster than the cis so the free acid was essentially pure trans. Further reactions including a Curtius rearrangement with retention gave tranylcypromine 151.

N2

CO2CHR2

CO2CHR2 CuOTf

Ph

CO2H

NH3Cl

25% NaOH EtOH

bis-oxazoline 133 R = cyclohexyl

Ph 149; trans:cis 88:12 trans: 97% ee

10 hours

Ph

Ph 150; 79% yield trans:cis 130:1

151 tranylcypromine

Asymmetric Alkene Metathesis Metal carbene complexes are also involved in metathesis (described in chapter 15). Exchange of carbene complexes with alkenes via a metallacyclobutane releases volatile alkenes such as ethylene with the formation of new alkenes. Ring closing metathesis is particularly favoured but normally leads to no new chiral centres. The simple Mo and Ru carbene catalysts described in chapter 15 cannot of course be used to induce asymmetry but a new generation of asymmetric Schrock 152 and Grubbs 153 catalysts can create asymmetry if a choice between two enantiotopic alkenes is offered.36 i-Pr Mes N O

N Mo

i-Pr

O

Ph

Cl

N

Mes

Ru

O RO

Mes 152; Asymmetric Schrock

153; Asymmetric Grubbs

As we largely discussed Grubbs Ru catalysts in chapter 15 we shall redress the balance by using examples of the Mo catalyst 152. The necessary parts of the substrate are (a) a monosubstituted alkene so that metathesis starts there and (b) a choice between two at least disubstituted enantiotopic alkenes for the second stage in the metathesis. So the amine 154 forms first the metallacyclobutane 155 and hence the new carbene complex 156 that can choose between the remaining two alkenes. In this case it does so very well indeed to give a high yield of the eight-membered cyclic amine 158 as a single enantiomer. The conditions are mild: 2 mol% of 152 in benzene at 22 ⬚C for 25 minutes. Other than the solvent, all is fine.

26 Asymmetric Pericyclic Additions to Carbonyl Groups

587

R *Mo *Mo

N

R

R

catalyst 152

Ph

*Mo

N

N

Ph

Ph

154

155

+

156

Mo*

H

H

*Mo

+

CH2

CH2

catalyst for next cycle

N

N Ph

Ph

H

158; 93% yield, >98% ee

157

Examples of oxygen heterocycles include the five-membered ring 160 and the six-membered ring 161. It may seem at first sight that the obligatory extra alkene in all the products is a disadvantage but this is a useful functional group that can be used to develop functionality or can simply be oxidised to a carbonyl group. It is generally more reactive than the alkene embedded in the ring. The silicon compound 162 can also be cleaved at both Si – O and, if necessary, Si – C bonds to give an open chain compound.

2 mol% 146

O

benzene 22 ˚C 25 minutes

159

O

O

2 mol% 152

Me

Me Si

benzene 22 ˚C 25 minutes

161

160; 99% yield up to 95% ee

O

Si

Me Me 162; >99% yield >95% ee

Asymmetric Pericyclic Additions to Carbonyl Groups 2 ⫹ 2 Ketene cycloadditions In chapter 25 you met the ‘pseudoenantiomers’ quinine and quinidine when their dihydroderivatives were used in Sharpless asymmetric dihydroxylation. The cinchona alkaloids themselves are used to catalyse a 2 ⫹ 2 cycloaddition of ketene to chloral in an important industrial process.37 Chloral reacts with ketene 163 from a generator in the presence of quinine or quinidine to give the β-lactone 164. Catalysis by quinine gives the other enantiomer. Hydrolysis opens the lactone ring and hydrolyses the CCl3 group to CO2H to give either enantiomer of malic acid 165. O

O + Cl3C chloral

•

0.5-2 mol% quinidine

H 163 ketene

O

–50 ˚C toluene

O

OH NaOH H2O

HO2C

CO2H

Cl3C 164; 95% yield >96% ee

165; 95% yield >96% ee (S)-(–)-malic acid

588

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

This is probably not strictly a 2 ⫹ 2 cycloaddition but an addition of the chiral amine 169 to the ketene 166 to give a chiral enolate that adds to chloral in the step that induces asymmetry 167. Ring closure 168 gives the lactone 164 and regenerates the quinidine for the next cycle.38 OMe O

R*3N

Cl3C

OH

O

O

N Cl3C

H 167

166

O

O

O

NR*3

•

R*3N

O

Cl3C

H 168

N

164

169; (+)-quinidine

The mechanism of the hydrolysis both of the lactone and of the CCl3 group by hydroxide must be understood to interpret the stereochemistry.39 Whereas simple β-lactones tend to react at sp3 carbon by an SN2 reaction (BAlk2) as that releases ring strain in the slow step, the CCl3 group slows down the SN2 reaction and lactone 164, like most lactones, is attacked at the carbonyl group by hydroxide ion 170. The configuration at the chiral centre is retained in this step. The hydrolysis of the CCl3 group in 171 to CO2H however occurs with inversion at the chiral centre - the details of this reaction are in the workbook. The other enantiomer (R)-(⫹)- of malic acid thus joins the new chiral pool (chapter 23). O O NaOH

O

164

O OH

BAC 2

OH

=

HO

H2O

Cl3C CCl3 170

NaOH, H 2O

CO2

(S)-165 INVERSION

CCl3 (R)-171

Reaction with trichloroacetone is similarly high-yielding and enantioselective giving the βlactone 172 and after hydrolysis with inversion, (S)-citramalic acid 173. Again, quinine gives the other enantiomer. O

O + Me

•

CCl3

trichloroacetone

O O

–50 ˚C, toluene

OH

NaOH

INVERSION

0.5 — 2 mol%

HO2C

Cl3C

quinidine 169

Me 172; 95% yield >96% ee

163 ketene

CO2H

173; >96% ee (S)-citramalic acid

Hydrolysis of both lactones 163 and 172 by hydroxide ion occurs with inversion during the conversion of the CCl3 group to CO2H. Weaker nucleophiles clearly attack the carbonyl group 176 to make available a series of derivatives such as the amides 174 and 175.

R2N

O

Me 174

Cl3C

O

O HNR2

HO

HO HO2C

R2N

RETENTION

Me 175

Cl3C

O

NaOH

173 INVERSION

Me 176

26 Asymmetric Pericyclic Additions to Carbonyl Groups

589

Synthesis of enalapril The ACE (Angiotensin Converting Enzyme) inhibitors are a group of important drugs that reduce high blood pressure and help to prevent strokes and heart attacks. One is enalapril 182 that can be made from the β-lactone 164. A Friedel-Crafts acylation (note that attack at the carbonyl group is preferred) gives the ketone 177 reduced to the diol 178 and hence to the lactone 179 both as single enantiomers of mixtures of diastereoisomers.40

164 >96% ee

O

benzene 3 x AlCl3

OH

Ph

OH

LiAlH4 THF

CCl3

Ph

NaOH

CCl3

5 ˚C

177; 90% yield >99% ee

OH

OH

H2O

Ph

178; 100% yield

O

O

179; 75% yield

Catalytic hydrogenation of the benzylic C–O bond in 179 removes the second chiral centre and the hydroxy-ester 180 is prepared for coupling to an enantiomerically pure amine by conversion to the triflate 181. The coupling goes with inversion to give enalapril41 182. 1. H 2, Pd/C

OH

AcOH

179 2. EtOH

Ph

CO2Et

EtO2C

OTf

Tf2O

Ph

180; 67% yield

CO2Et

N

N H

Ph

O

CO2H

182; enalapril

181

Recent developments extend the asymmetric β-lactone formation to dichloroaldehydes 183 with formation of the ketene 163 by elimination on acetyl chloride rather than from a ketene generator. A mixture of the aldehyde 183, Hünig’s base and 2 mol % quinidine 169 is treated with acetyl chloride to give the β-lactones 184 in good yield (40–85%) and excellent ee. Reduction with DIBAL gives the diol 185 with the two chlorine atoms intact.42

O

i-Pr2NEt

R • +

Cl

O

O

O

H

Cl 163 ketene

2 mol% quinidine

Cl

OH

O DIBAL

R

toluene –25 ˚C

25 ˚C CH2Cl2

Cl Cl (R)-184; 93-98%ee

183

R Cl

OH Cl

(R)-185; 95% yield

Among various further developments, the formation of the azido ketone 188 is especially promising. It also illuminates aspects of the mechanism of the hydrolysis of 157 and 165. The closure of the epoxide 187 and its opening by azide ion with inversion and the exposure of the carbonyl group 188 by loss of chloride all also occur in the hydrolysis of the CCl3 group in 157 and 165 (see workbook). OH (R)-185; R = C6H13

i-Pr3SiCl imidazole DMF

OTIPS THF

Cl

O

NaH

R Cl

(R)-186; 97% yield

OTIPS

R Cl (R)-187; 92% yield

N3 NaN3 DME H2O

OTIPS

R O (S)-188; 83% yield

590

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

Carbonyl ene reactions The carbonyl ene reaction is between a similarly reactive aldehyde and an alkene rather than a ketene. Glyoxalate esters and chloral are typical of the carbonyl compounds involved. Asymmetric versions rely on Lewis acid catalysts similar to those used in Diels-Alder reactions based on Ti and Al among other metals.43 A simple example is the formation of 190 from methyl glyoxalate and an alkene 189 catalysed by 0.5 mol% of a BINOL-Ti complex 191. Other alkenes that react well are 192 - 195. H

O

Ti*

OH

0.5 mol% catalyst

O Ph

CO2Me

Ph 189

CO2Me

Ti

190; 84% yield, 95-95% ee

OR OR

O 192

193

194

195 191; ( R)-(+)-BINOL-Ti catalyst

Alder ene reactions The original ene reaction, invented by Alder, was between an alkene such as 192 and a dienophile - that is a conjugated unsaturated carbonyl compound. If the reaction is intramolecular, the alkenes need not be specifically activated by conjugation but they must be clearly distinct in reactivity. Xumu Zhang and group44 have recently reported asymmetric tetrahydrofuran syntheses using the BINAP Rh(I) catalyst 21 so successful in hydrogenations. A Z-allyl propargyl ether 196 reacts with this Rh catalyst in an ene reaction that corresponds to the tortured mechanism 197. The substituents R1 and R2 can be a range of alkyl, aryl, and functionalised groups. No absolute stereochemistry is given but each enantiomer of BINAP produces a different enantiomer of 198. R1

R1 R2

10 mol% Rh(COD)Cl2

H

R2

R2

R1

12 mol% BINAP

O

O O

AgSbF6

196

197

198; 82-96% yields 99.0 to >99.9% ee

Nucleophilic Additions to Carbonyl Groups Allyl metals Closely related to the previous section is the asymmetric addition of allyl stannanes to aldehydes. The catalyst is the same and the main difference is that an R3Sn group is lost instead of a proton. This has the advantage that there is no ambiguity about the position of the alkene in the product. An example is the addition of an allyl group to the functionalised aldehyde 199. Complexation of either or both oxygen atoms to the Ti is indicated.45 O +

BnO

SnBu3

H 199

OH

(S)-BINAP-Ti catalyst 178

BnO –20 ˚C, CH 2Cl2

200

201

26 Nucleophilic Additions to Carbonyl Groups

591

Recent improvements include the replacement of the toxic tin reagent 200 with the allyl silane46 203. Addition to the alkoxyaldehyde 202 gives 204 that can be transformed into the half-protected dialdehyde 205. OR

O H

SiMe3

+

OR

(S)-BINAP-TiF4 catalyst

OH

O

O

204; 97% yield, 91% ee (97% after recrystallisation)

203

O

H

0 ˚C, CH 2Cl2

202

OH

205

Alkynyl metals Ideally it should be possible to add organo-Li and Grignard reagents to prochiral aldehydes and ketones with a chiral catalyst to give alcohols in good enantiomeric excess. This is difficult since the uncatalysed rate of addition is so high. It is possible to add the less reactive lithium acetylides such as 207 to ketones in the presence of stoichiometric lithium derivatives of amino alcohols such as 209, analogues of ephedrine. The product 208 is used in the asymmetric synthesis of the Merck anti-HIV compound Efavirenz.47 OLi

CF3 CF3

Li

Cl

O

Cl

OH

207

N H

N

Ar N Ar H 208; 80% yield, 96-98% e.e.

206; Ar = 4-MeOC 6H5

209

Among the best catalytic versions of such reactions are Carreira’s additions of alkynyl zinc reagents to carbonyl compounds.48 Organo-zinc reagents are very much less nucleophilic than organo-Li or Mg compounds so the rate of the uncatalysed reaction is small. The catalysts are again amino alcohols related to ephedrine (such as N-methylephedrine 212) and a big advantage is that no separate step is needed to create the alkynyl zinc reagent as the acetylene, Zn(OTf)2 and the amine do the trick. With stoichiometric catalyst 212 the reaction is general for propargyl alcohols 210 such as 211 and the toluene used as solvent need not be completely dry. H R2 Zn(OTf) 2, 212

O R1

H

OH

OH

Ph

Me

HO

NMe2

R1 toluene, 23 ˚C

R2 210

Ph 211; 88% yield, 99% ee

212; N-methylephedrine

The catalytic reaction, using admittedly 20 mol% Zn(OTf)2, 22 mol% catalyst 212, and 50 mol% Et3N, is successful with a variety of functionalised aldehydes and acetylenes such as 213 and 214. Though toluene is used as solvent, in fact no solvent is necessary.49 OH

O H H +

TIPSO

NBn2

20 mol%Zn(OTf) 2 22 mol% 212 50 mol% Et3N

TIPSO

toluene, 100 ˚C

213

214

215; 80% yield, 95% ee

NBn2

592

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

Addition of alkyl organo-zinc reagents Though not so general as the reactions we have just seen, the catalysed addition of dialkyl zincs to certain aldehydes sets a new standard for catalysis that needs some explaining. Dialkyl zincs add to the pyrimidine aldehyde 216 under catalysis from amino alcohols, amino acids such as leucine 219, hydroxy acids, and simple secondary alcohols or amines such as 218 to give enantiomerically enriched alcohols 217. Plain sailing so far, except for the extraordinary range of catalysts. O

OH NHMe

N Me

i-Pr2Zn

H

'initiator' ~0.1% ee

N 216

NH2

N Ph Me N 217; 88% yield; 85% ee

CO2H 218

219; ( S)-leucine

The strange thing is that catalysts of low enantiomeric excess, even as low as 0.1% ee, induce asymmetry of a much higher order (here 95%). This ‘asymmetric amplification’ is particularly intriguing since ees of about 0.1% can be produced by irradiating racemic leucine with circularly polarised light. The true catalyst in these reactions is the zinc alkoxide 220 of the product. The ‘initiators’ as they are better known, such as leucine, merely create a small amount of this true catalyst.50 O N Me

OZni-Pr H

i-Pr2Zn catalyst 220

N

1. HCl, H 2O

N Me

N

2. NaHCO 3, H 2O

217 88% yield 85% ee

220

216

This is easily demonstrated in what Soai calls ‘practically perfect asymmetric autocatalysis’ with the special aldehyde 221 using 20 mol% of the previously prepared product 222 as catalyst and enough i-Pr2Zn to allow for the conversion of the catalyst to the zinc alkoxide. There is considerable amplification: if the catalyst 222 has 5.5% ee, the product (also 222) is 70% ee. But if enantiomerically pure catalyst is used, the yield and the ee of the product are practically perfect. Each molecule of catalyst produces five molecules of itself as product. The amplification factor is six and if the whole of the product is used in a second reaction with five times as much aldehyde a second batch of 222, 36 times as much, is the product.51 O N

H N

t-Bu 221

OH 1.7 equivalents i-Pr2Zn 20 mol% catalyst 222 cumene, 0 ˚C, 3 hours

N N

t-Bu 222; >99% yield; 70% ee 5.5% ee catalyst 222; >99% yield; >99.5% ee >99.5% ee catalyst

So how are we to explain such remarkable results? The key point is that all the aldehydes must have a pyrimidine 216, 221, pyridine 223, or quinoline 225 ring joined to the carbonyl group and the heterocyclic nitrogen must be meta to the aldehyde.

26 Palladium Allyl Cation Complexes with Chiral Ligands

NR2

NR2 CHO

O

OH

OH

O

CHO

i-Pr

i-Pr2Zn

N

N

i-Pr

i-Pr2Zn

N

224; 56 → 85% ee

223

593

N 226; 9 → 59% ee

225

Non-linear relationships between ee of catalyst and ee of product require some dimeric species. Suppose the true catalyst, the zinc derivative of the product, forms dimers. There are two diastereoisomers of these, a meso dimer 229 and a homochiral dimer 230. If the meso dimer is more stable than 230 it will absorb all the minor enantiomer 227 leaving only the major enantiomer 228 to catalyse the reaction. Of course, it could be the homochiral dimer 230 that is the active catalyst as the dimers have spare coordination sites on zinc. This phenomenon is described in more detail in an excellent paper.52 H

216

H

OZnLn

i-Pr2Zn

i-Pr OZnLn

H

i-Pr

R

Zn O

H N

N

R

Zn O

N

N

+ N

N

N

N

N Me

N

Me

227 (R)-monomer

228 (S)-monomer

Zn

O

N H

N

Zn

O

H i-Pr R 230; (S,S)-dimer

i-Pr R 229; racemic (R,S)-dimer

Palladium Allyl Cation Complexes with Chiral Ligands In chapter 19 we discussed the uses of palladium allyl cation complexes as electrophiles. We established that Pd(0) adds to the opposite face of the allylic system to the leaving group 232 to form an η3 cation complex 233 and that the nucleophile attacks from the opposite face to the Pd so that the two inversions lead to retention. We established that regioselectivity and diastereoselectivity can be well controlled. If this seems unfamiliar we suggest you read the relevant section of chapter 19 before proceeding. OAc

OAc Pd(0)

R

R 231

Nu R

R

R Pd

232

R Pd 233

Nu

Nu

R

R 234

R

R 235

What is new here is that attack of the nucleophile on the complex 233 can lead to asymmetric induction in a number of ways. For example either enantiomer 234 or 235 could be formed from the symmetrical complex 233. This involves attack at one enantiotopic end of the allylic system rather than one enantiotopic face and in this sense rather resembles the ring closing step (82 → 83) in the asymmetric Robinson annelation. Though essential early work53 was done on the symmetrical complex 233; R ⫽ Ph, we shall concentrate on more recent results. Trost established54 that the ligand 236 was the best for many of these reactions. Thus racemic allylic acetate 237 gives a symmetrical allylic cation complex like 233 and the ligand directs the

594

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

malonate nucleophile to one end only to give 238. High ees are achieved only if the counterion is large - (Hexyl) 4N⫹ being the best.

O

O NH

CO2Me

HN

H

OAc

CO2Me

CO2Me

PPh2 Ph2P

CO2Me

Pd(0). 236

237

ligand 236

238; 98% yield, 98% ee

Desymmetrisation of bis allylic esters such as 239 adds another dimension. Asymmetry is already introduced in the η3 cation complex 240 by selection of one of the enantiotopic benzoates as leaving group. Attack of the nucleophile occurs on the same side as the other benzoate, as that is the side opposite the Pd atom and so there is strong regioselectivity for the site away from the second benzoate. Since the first formed product 241 is still an allylic benzoate, further reaction is possible. Ph Ph

O

O

O

Ph

O

PhO2S

SO2Ph

O

O

Ph

NO2

O

O O N

Pd(0). 236

O

Pd 240

239

241

Cyclisation of the intermediate 241 in the same pot closes the five-membered ring 243 leading to 244, an intermediate in the synthesis of the anti-viral drug carbovir. Both stereo- and regioselectivity are controlled by the short tether.55 PhO2S

PhO2S 241

H

MeO2C

Pd(0). 236 base

O N

O

O

N HO

O

Pd

H 243; 97% yield, 96% ee

242

244

Catalyst 236 is by no means the only one to achieve such results. Helmchen56 uses the manganesecontaining monophosphine 245 for alkylation of malonates with racemic allylic chloride 246 to give the adduct 247 that is used in the synthesis of the lactone 248. O Mn(CO)3

PdCl

N

Cl Ph

2

P CO2Me

H

OAc CO2Me

H

CO2Me

O

AcO

246 245

OH

O CO2Me

247; 96% yield 99.5% ee

H 248

26 Palladium Allyl Cation Complexes with Chiral Ligands

595

Enantioselectivity with open chain compounds is more difficult but Trost uses ligand 236 for the simplest of these: the symmetrical allylic carbonate 249 (note the different leaving group: CO2 is lost, see chapter 19) with malonate anions formed with CsCO3. However, this is the best result from a variety of different open chain allylic carbonates.57 O O

CO2Et OEt

EtO2C

Cs2CO3

+

CO2Et H

CO2Et

236

249

250; 98% yield, 92% ee

Intramolecular reactions generally give excellent results. Allylation of the nitroalkane portion of the indole 251 closes the six-membered ring of the ergoline alkaloids such as chanoclavine 253. Genet’s study of various ligands revealed that BINAP provided the best results. 58 AcO

NO2

NO2

NHMe

Pd(OAc) 2 (S)-BINAP

Ph

K2CO3 THF

NH 251

NH

NH

252; 60% yield, 95% ee

253; (–)-chanoclavine

More recently, Trost has developed 59 a clean way to form the difficult bicyclic core of the cinchona alkaloids with good diastereo- and enantioselectivity using his favourite ligand 236. Chirality is developed at both ends of the new bond: at the allylic carbon (marked with an empty circle) and at the former enol carbon (marked with a black blob). The two diastereoisomers of the product 255 and 256 form the core of the quinine alkaloids. They have the same absolute stereochemistry in the bicyclic core but the stereochemistry of the vinyl group is different. There is a strong match/mismatch effect: (R,R)-236 gives 4.6:1 diastereoselectivity in favour of 255 and excellent ee while (S,S)-236 gives 8:1 diastereoselectivity in favour of 256 and poor ee. H CO2Et

Pd2dba3.CHCl3

HO

H

3 mol% 236

EtO2C

O

N

OCO2Me

CH2Cl2

254 (R,R)-236

N 255 84% yield, >99% ee

(R,R)-236 + Eu(fod)3

CO2Et O N 256 85% yield, 68% ee

Controlling enantioselectivity at the enol centre alone can be achieved with special reagents and a very complex catalyst. An allylic carbonate with a fluorinated esterifying group allylates a prochiral enolate derived from i-propyl cyanopropionate catalysed by Pd, Rh, and the chiral Fe ‘TRAP’ ligand 257, gives excellent results. The explanations for these last two examples are complicated and you are referred to the papers if you want to know more.60

596

26 Asymmetric Induction III Asymmetric Catalysis: Formation of C–H and C–C Bonds

O

Ar2P O

CF3 O

Fe

Fe

O PAr 2

257; Ar = p-MeOC 6H4

NC

O Me

CF3

Rh(acac)(CO)2

O

Pd(cp)(π-C3H5) ligand 256 THF, –40 ˚C 6 hours

O Me

CN

258; 93% yield, 99% ee

Summary This is a very big and growing subject. There are many catalytic reactions, such as the phase-transfer catalysed alkylation of enolates, that we have no space to discuss here. That particular one is treated in the workbook.

References General references are given on page 893 1. I. Ojima (ed.) Catalytic Asymmetric Synthesis, Second Edition, Wiley, New York, 2000. 2. A. J. Birch and D. H. Williamson, Org. React., 1976, 24, 1. 3. E. J. Corey and B. E. Roberts, J. Am. Chem. Soc., 1997, 119, 12425. 4. L. Horner, H. Siegel and H. Büthe, Angew Chem., Int. Ed., 1968, 7, 942. 5. W. S. Knowles and M. J. Sabacky, Chem. Commun., 1968, 1445. 6. T.-P. Dang and H. B. Kagan, Chem. Commun., 1971, 481; H. B. Kagan and T.-P. Dang, J. Am. Chem. Soc., 1972, 94, 6429. 7. T. Ohkuma, M. Kitamura and R. Noyori, in I. Ojima, Asymmetric, pages 1-110 and 801–856. 8. R. Noyori, I. Tomino, Y. Tanimoto and M. Nishizawa, J. Am. Chem. Soc., 1984, 106, 6709; R. Noyori, I. Tomino, M. Yamada and M. Nishizawa, J. Am. Chem. Soc., 1984, 106, 6717. 9. K. Rossen, S. A. Weissman, J. Sager, R. A. Reamer, D. Askin, R. P. Volante and P. J. Reider, Tetrahedron Lett., 1995, 36, 6419. 10. H. Takaya, T. Ohta, S. Inoue, M. Tokunaga, M. Kitamura and R. Noyori, Org. Synth., 1995, 72, 74. 11. M. Kitamura, M. Tokunaga, T. Ohkuma and R. Noyori, Org. Synth., 1993, 71, 1. 12. H. Kawano, Y. Ishii, M. Saburi and Y. Uchida, J. Chem. Soc., Chem. Commun., 1988, 87. 13. T. Ohta, T. Miyake, N. Seido, H. Kumobayashi, S. Akutagawa and H. Takaya, Tetrahedron Lett., 1992, 33, 635. 14. T. Ohkuma, M. Koizumi, H. Doucet, T. Pham, M. Kozawa, K. Murata, E. Katayama, T. Yokozawa, T. Ikariya and R. Noyori, J. Am. Chem. Soc., 1998, 120, 13529; H. Doucet, T. Ohkuma, K. Murata, T. Yokozawa, M. Kozawa, E. Katayama, A. F. England, T. Ikariya and R. Noyori, Angew. Chem., Int. Ed., 1998, 37, 1703. 15. E. J. Corey, R. K. Bakshi, and S. Shibata, J. Am. Chem. Soc., 1987, 109, 5551; E. J. Corey, R. K. Bakshi, S. Shibata, C.-P. Chen and V. K. Singh, J. Am. Chem. Soc., 1987, 109, 7925; E. J. Corey and C. J. Helal, Tetrahedron Lett., 1996, 37, 4837; Review: E. J. Corey and C. J. Helal, Angew. Chem., Int. Ed., 1998, 37, 1987. 16. J. Y. L. Cheung, R. Cvetovich, J. Amato, J. C. McWilliams, R. Reamer and L. DiMichaele, J. Org. Chem., 2005, 70, 3592. 17. Z. G. Hajos and D. R. Parrish, J. Org. Chem., 1974, 39, 1612, 1615. 18. F. R. Clemente and K. N. Houk, Angew. Chem., Int. Ed., 2004, 43, 5766; see also T. Bui and C. F. Barbas, Tetrahedron Lett., 2000, 41, 6951. 19. B. List, R. A. Lerner and C. F. Barbas, J. Am. Chem. Soc., 2000, 122, 2395. 20. A. B. Northrup and D. W. C. MacMillan, J. Am. Chem. Soc., 2002, 124, 6798. 21. W. Notz and B. List, J. Am. Chem. Soc., 2000, 122, 7386.

26 References

597

22. T. D. Machajewski and C.-H. Wong, Angew. Chem., Int. Ed., 2000, 39, 1352. 23. K. A. Ahrendt, C. J. Borths and D. W. C. MacMillan, J. Am. Chem. Soc., 2000, 122, 4243; N. A. Paras and D. W. C. MacMillan, J. Am. Chem. Soc., 2001, 123, 4370. 24. N. A. Paras and D. W. C. MacMillan, J. Am. Chem. Soc., 2002, 124, 7894. 25. J. F. Austin and D. W. C. MacMillan, J. Am. Chem. Soc., 2001, 123, 1172. 26. Y. Iwabuchi, M. Nakatani, N. Yokoyama and S. Hatakeyama, J. Am. Chem. Soc., 1999, 121, 10219. 27. A. B. Northrup and D. W. C. MacMillan, J. Am. Chem. Soc., 2002, 124, 2458. 28. E. J. Corey, R. Imwinkelried, S. Pikul and Y. B. Xiang, J. Am. Chem. Soc., 1989, 111, 5493; S. Pikul and E. J. Corey, Org. Synth., 1993, 71, 30. 29. D. Seebach, R. Dahinden, R. E. Marti, A. K. Beck, D. A. Plattner and F. N. M. Kühnle, J. Org. Chem., 1995, 60, 1788; D. Seebach, R. E. Marti and T. Hintermann, Helv. Chim. Acta, 1996, 79, 1710. 30. D. A. Evans, K. A. Woerpel, M. R. Hinman and M. M. Faul, J. Am. Chem. Soc., 1991, 113, 726. 31. D. A. Evans, J. A. Murry, P. von Matt, R. D. Norcross and S. J. Miller, Angew. Chem., Int. Ed., 1995, 34, 798. 32. M. P. Doyle in Ojima, Asymmetric, chapter 5 pages 191–228. 33. T. Niimi, T. Uchida and T. Katsuki, Tetrahedron Lett., 2000, 41, 3647. 34. T. Aratani, Pure Appl. Chem., 1985, 57, 1839. 35. F.-C. Shu and Q. Zhou, Synth. Commun., 1999, 29, 567. 36. S. J. Dolman, E. S. Sattely, A. H. Hoveyda and R. R. Schrock, J. Am. Chem. Soc., 2002, 124, 6991; J. J. Van Veldhuizen, S. B. Garber, J. S. Kingsbury and A. H. Hoveyda, J. Am. Chem. Soc., 2002, 124, 4954. 37. P. Stutte in Chirality in Industry, Eds A. N. Collins, G. N. Sheldrake and J. Crosby, Wiley, Chichester, 1992, page 341. 38. H. Wynberg and E. G. J. Staring, J. Am. Chem. Soc., 1982, 104, 166; J. Org. Chem., 1985, 50, 1977. 39. H. Wynberg and E. G. J. Staring, J. Chem. Soc., Chem. Commun., 1984, 1181. 40. T. Fujisawa, T. Ito, K. Fujimoto, M. Shimizu, H. Wynberg and E. J. G. Staring, Tetrahedron Lett., 1997, 38, 1593. 41. H. Yanagisawa, S. Ishihara, A. Ando, T. Kanazaki, S. Miyamoto, H. Koike, Y. Iijima, K. Oizumi, Y. Matsushita and T. Hata, J. Med. Chem., 1987, 30, 1984. 42. R. Tennyson and D. Romo, J. Org. Chem., 2000, 65, 7248. 43. K. Mikami, M. Terada, S. Narisawa and T. Nakai, Org. Synth., 1993, 71, 14; K. Mikami and T. Nakai in Ojima, Asymmetric, chapter 8C pages 543-568; K. Mikami, M. Terada and T. Nakai, J. Am. Chem. Soc., 1989, 111, 1940; J. Am. Chem. Soc., 1990, 112, 3949. 44. A. Lei, M. He, S. Wu and X. Zhang, Angew. Chem., Int. Ed., 2002, 41, 3457. 45. G. E. Keck and D. Krishnamurthy, Org. Synth., 1998, 75, 12. 46. J. W. Bode, D. R. Gauthier and E. M. Carreira, Chem. Commun., 2001, 2560. 47. A. S. Thompson, E. G. Corley, M. F. Huntington and E. J. J. Grabowski, Tetrahedron Lett., 1995, 36, 8937. 48. D. Boyall, D. E. Franz and E. M. Carreira, Org. Lett., 2002, 4, 2605. 49. N. K. Anand and E. M. Carreira, J. Am. Chem. Soc., 2001, 123, 9687. 50. T. Shibata, J. Yamamoto, N. Matsumoto, S. Yonekubo, S. Osanai and K. Soai, J. Am. Chem. Soc., 1998, 120, 12157. 51. T. Shibata, S. Yonekubo and K. Soai, Angew. Chem., Int. Ed., 1999, 38, 659. 52. D. G. Blackmund, C. R. McMillan, S. Ramdeehul, A. Schorm and J. M. Brown, J. Am. Chem. Soc., 1999, 121, 10103. 53. see the review by B. M. Trost and C. Lee in Ojima, Asymmetric, pages 593–649. 54. B. M. Trost and R. C. Bunt, J. Am. Chem. Soc., 1994, 116, 4089. 55. B. M. Trost, L. Li and S. D. Guile, J. Am. Chem. Soc., 1992, 114, 8745. 56. S. Kudis and G. Helmchen, Tetrahedron, 1998, 54, 10449. 57. B. M. Trost, M. G. Organ and G. A. O’Doherty, J. Am. Chem. Soc., 1995, 117, 9662. 58. J.-P. Genet and S. Grisoni, Tetrahedron Lett., 1988, 29, 4543; N. Kardos and J.-P. Genet, Tetrahedron: Asymmetry, 1994, 5, 1525. 59. B. M. Trost, K. L. Sacchi, G. M. Schroeder and N. Asakawa, Org. Lett., 2002, 4, 3427. 60. M. Sawamura, M. Sudoh and Y. Ito, J. Am. Chem. Soc., 1996, 118, 3309.

27

Asymmetric Induction IV Substrate-Based Strategy Introduction to Substrate-Based Strategy Chiral Carbonyl Groups The asymmetric Strecker reaction Chiral Enolates from Imines of Aldehydes: SAMP and RAMP Chiral Enolates from Amino Acids Schöllkopf’s bislactim ethers Other chiral glycine enolates: Williams’s method Seebach’s relay chiral units Chiral Enolates from Hydroxy Acids Seebach’s relay chiral units Chiral Enolates from Evans Oxazolidinones Application of asymmetric alkylation with Evans auxiliaries Aldol Reactions with Evans Oxazolidinones The syn aldol reaction with boron enolates Anti-aldols by Lewis acid-catalysed reactions with Evans oxazolidinones Chiral Auxiliaries Asymmetric aldol, conjugate addition, and Diels-Alder reactions compared The Asymmetric Diels-Alder Reaction Introduction Diels-Alder reactions with Oppolzer’s chiral sultam Diels-Alder reactions with pantolactone as chiral auxiliary Chiral auxiliaries attached to the diene Improved Oxazolidinones: SuperQuats Asymmetric Michael (Conjugate) Additions Michael additions with 8-phenylmenthyl esters of unsaturated acids Chiral auxiliaries attached elsewhere in asymmetric Michael additions Other Chiral Auxiliaries in Conjugate Addition The Evans oxazolidinones Chiral sulfoxides Asymmetric Birch Reduction Birch reduction of benzene Asymmetric Birch reduction of heterocycles

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

600

27 Asymmetric Induction IV Substrate-Based Strategy

Introduction to Substrate-Based Strategy In the last few chapters we have started with a prochiral molecule having enantiotopic features: the faces of an alkene for instance - and differentiated these faces with enantiomerically pure reagents or catalysts. In this chapter we explore the alternative approach. The enantiotopic faces are transformed into diastereotopic faces by the covalent attachment of a chiral auxiliary. It seems inevitable that this auxiliary must be stoichiometric but you will even see that ‘catalytic’ substrate methodology is starting to emerge. It also seems inevitable that two extra reactions will be needed: the attachment and the removal of the chiral auxiliary. In spite of this, one of the most important methods of asymmetric synthesis, centred around Evans’s amino acid-based chiral auxiliaries, uses this approach. O OR

R

R

O

N

1

O

O N

O

Ph

2

Ph

3

4

Thus the faces of the carbonyl group in the aldehyde 1 are enantiotopic but after attachment of a chiral amine they become diastereotopic in the imine 2. The faces of the enolate of the ester 3 are enantiotopic but those of the enolate of the amide 4, after attachment of the Evans phenylalanine-derived oxazolidinone, are diastereotopic. You may notice something else that is different in this chapter. It is much easier to offer reasonable explanations for substrate-style asymmetric induction because we have a better idea of the conformation the compound adopts during the reaction as the configuration of the product is a guide.

Chiral Carbonyl Groups The asymmetric Strecker reaction The Strecker reaction is a way to make α-amino acids 7 from aldehydes 1 with cyanide and ammonia. Cyanide adds to the unstable imine 5 to give the stable amino nitrile 6 that can easily be hydrolysed to the inevitably racemic α-amino acid 7. CN

CO2H

NH3

R

O

H2O

R

HCN

R

NH

NH2

R

NH2

HCN

1

5

6

7

One asymmetric Strecker reaction1 uses the enantiomerically pure amine 8, available as both enantiomers since it is easily prepared by resolution (chapter 24), as a way to make the imine 2 have diastereotopic faces. Cyanide addition is reasonably diastereoselective giving mainly the amino nitrile 9. CN R

O 1

+

H2 N

Ph 8

R

N 2

Ph

HCN

R

N H 9

Ph

27 Chiral Enolates from Imines of Aldehydes: SAMP and RAMP

601

The most likely explanation is that the imine prefers a Houk-style conformation 2a with the two marked H atoms in the plane and the cyanide adds 10 to the face of the C=N bond opposite the larger phenyl group.

H

H R

N

Ph

H

N

C

= R

2

N

R

Ph Me

CN

Me H

H N

HCN

R

Ph

2a

10

N H 9

Ph

An advantage of the substrate-based strategy is that the chiral auxiliary is still attached to the molecule after the induction reaction and so the product, 9 in this instance, can be purified as one diastereoisomer before the chiral auxiliary is removed. An example of the asymmetric Strecker reaction in practice is the synthesis of the ‘fat’ amino acid 16 having an adamantyl group to increase solubility in fats and hence permeability to cell membranes.2 After purification of the amino nitrile 13 hydrolysis gives the amide 14, hydrogenation removes (and sadly destroys) the chiral auxiliary, and a final hydrolysis gives the ‘fat’ amino acid 16. CN CHO

N

11 O

HCl

HCN

8

Ph

N H 13; 75% yield

12; 100% O

NH2 N H

Ph

OH

37% HCl

NH2

HCl, EtOH

NH2

90 ˚C 4 hours

15; 86% yield

14; 81% yield as .HCl

H2O

O

NH2

H2,Pd/C

Ph

16; 79% yield

Chiral Enolates from Imines of Aldehydes: SAMP and RAMP Aza-enolates derived from imines were introduced in chapter 10. It is easy to see that imines from chiral amines might well be used to make aza-enolates that would react asymmetrically with electrophiles. Among the most famous examples are the hydrazones SAMP and RAMP derived by Enders from proline.3 The starting material derived from natural (S)-(⫺)-proline 17 is called SAMP 18 and the one derived from unnatural (R)-(⫹)-proline is RAMP. The reactions of the two are identical except that they lead to products of opposite chirality.

17 (S)-L-(–)-proline

OMe N H

CO2H

N NH2

18; ( S)-(–)-SAMP 50-58% overall yield from (S)-L-(–)-proline on 75 g scale

Imines 20 formed from SAMP and enolisable aldehydes prefer an E-configuration across the C=N bond. Lithiation gives the aza-enolate 21 also with the E-configuration derived from the preferred conformation of 20. Reaction with an electrophile gives mostly the diastereoisomer 22.

602

27 Asymmetric Induction IV Substrate-Based Strategy

CHO

OMe

18 SAMP

N N

R

OMe

OMe N

N

1. LDA

N

R

Li

N

2. E

R

R

19

20

E 22

21; SAMP-azaenolate

The side chain OMe group is essential for the asymmetric induction as it complexes the lithium atom of LDA so that only one of the diastereotopic protons is removed 23 and holds the aza-enolate in one configuration 24 with Li bonded both to N and to O. Electrophilic attack then occurs anti to the lithium atom on the unhindered face of the aza-enolate 24 to give the product in the same conformation 22a.

i-Pr 20

LDA

i-Pr H

N

Li N

R

Me

Me O

Li N

N

Me

O

O R

N

R

H

H

N

N

E E

23

22a

24

This is an excellent and popular method but removal of the chiral auxiliary is rather difficult and must be carried out oxidatively to produce the aldehyde product 25A. The chiral auxiliary emerges as the N-nitroso compound 23 and this can be recycled. An alternative is alkylation (MeI) and acidic hydrolysis. Reductive removal gives the amine 25B and an intermediate 26 in the synthesis of SAMP. oxidative cleavage:

23 OMe N NO

ozone, CH2Cl2, –78 ˚C

OMe N N

or hydrolysis:

E R

CHO 25A

1. excess MeI, 2. 5M HCl, pentane

reductive cleavage:

R

1. catechol borane

OMe N H

2. Raney nickel

E

26

E NH2

R 22

25B

The electrophile E⫹ can be an alkyl halide or sulfate, an aldehyde to give aldol products, or an α,β-unsaturated ester when conjugate addition is preferred. Examples from simple alkylation show that the alkyl halide can be primary alkyl, allylic 33, and even an α-bromoester or γ-bromo-α,β-unsaturated ester 31. The original carbonyl compound that forms the chiral imine with SAMP or RAMP can be an aldehyde 27 or 29, a ketone (symmetrical 32 or blocked on one side 35), or an enone. Only the reagents and products are shown with oxidative [O] or hydrolytic [H 2O] workup. Notice that SAMP is used for the formation of either enantiomer of 28 by using different starting materials but that RAMP is used to enter the other enantiomeric series from 32.

27 Chiral Enolates from Amino Acids

CHO

CHO

CHO

Etl

Mel

[H2O]

[O]

(S)-28; 95% ee 71% yield

27 O

603

CO2Et

Br

(R)-28; 95% ee 65% yield

29 O

Br

O

CO2Et

RAMP

SAMP

+

+

[H2O]

OEt

[H2O]

OEt 31

(S)-30; >95% ee, 80% yield

CHO

32

(S)-34; >97% ee, 61% yield

33

Reactions with aldehydes give syn-aldol products 36 in moderate ee but the hydrazones are crystalline so, at the expense of yield, ee can be raised to 100%. Conjugate addition gives good ees but moderate yields and the first formed products, e.g. 37, can be converted into useful lactones, e.g. 38. O

O 1. SAMP

(S,S)-adduct 86% yield 72%ee

2. LDA 3. PhCHO

2. Hydrolyse

(S,S)-36, 35% yield, 100%ee

35 1. SAMP, 2. LDA

MeCHO

OH

1. crystallise hydrazones

NaBH4

MeO2C

3. MeO2C

CHO

37; 46% yield [H 2O]

O

O

(R)-38 35% yield from MeCHO >95%ee

Chiral Enolates from Amino Acids Schöllkopf’s bislactim ethers A simpler way to restrict the conformation of an enolate is to confine it in a heterocycle and an important group of chiral enolates come from various derivatives of amino acids. The first successful such compounds were Schöllkopf’s ‘bislactim ethers’ 41 derived from the diketopiperazines 40 formed when an amino acid such as alanine 39 condenses with itself.4 Treatment of 41 with butyl lithium creates a lithium enolate on one position in the ring: the methyl group in the other position keeps the chirality intact. Alkylation occurs selectively on the opposite side to the remaining methyl group 42 and hydrolysis releases a new tertiary amino acid 43 and one of the original alanines. H N

NH2 O

OH

O

Et3O BF4

MeO

N H 40; ‘diketo -piperazine’

O

39 (S)-(+)-alanine

N

OMe

MeO

N

R

42

N

41; ‘bislactim ether ’ CO2H

1. BuLi 2. RBr

OMe

N

3. H 3O

R

H2N 43

604

27 Asymmetric Induction IV Substrate-Based Strategy

A better version 47 has a larger isopropyl group, derived from valine 44, to differentiate the faces on the enolate, and no substituent at the enolate site. The starting material 46, prepared from valine 44 and glycine, is commercially available, as is its R-isomer and the derived bislactim ether 47.

H N

NH2

1. Et 3N

O

H2N

COCl2

HO

THF 40 ˚C

O

CO2Et

Cl

O

Cl 2. reflux, toluene

45

44; (S)-(+)-valine H N

O

N

Et3O BF4

N H 46; 91% from Val O

MeO

OMe

N

47; 82% yield

Treatment with BuLi generates chemoselectively the lithium enolate of the less substituted lactim and electrophiles attack the face opposite the branched isopropyl group. Selectivity is good: the purified diastereoisomers can be isolated in over 80% yield and hydrolysis requires only dilute aqueous acid as these are easily protonated imines. These examples show a benzylic and an allylic halide. The first 50a is unnatural (R)-phenylalanine and the second 50b is an unnatural amino acid.

Li 1. BuLi

OMe

N

OMe

N

47

2. RBr

MeO

MeO N R 49a; R = Bn, 81% yield 49b; R = cinnamyl, 90% yield

N 48 CO2H

0.25M HCl H2O

CO2H H2N

H2N

Ph

Ph 50a; ( R)-Phe (R)-50b 73% yield, 92% ee 89% yield, >95% ee

Conjugate addition is also possible with α,β-unsaturated esters and the interesting pyridine derivative (E)-51 gives a 90% yield of the conjugate addition product 52 that is 95% one diastereoisomer. Purification and hydrolysis gives the substituted proline derivative 54 via the amino acid 53. Many variations of this theme have been played with different electrophiles.

27 Chiral Enolates from Amino Acids

N

OMe H

1. BuLi

47

CO2Me

N

MeO 2.

605

N CO2Me

(E)-51

52

N

CO2Me H

HCl H2O

N

H2N

CO2Me

O

CO2Me

N H

N 53

54

Other chiral glycine enolates: Williams’s method Continuing the theme of pyridine in amino acids we take (S)-azatyrosine 56 as an example of Williams’s chiral glycine enolate equivalent.5 The only difference between tyrosine 55 and aza-tyrosine 56 is the nitrogen atom in the benzene ring yet the one is an amino acid found in protein and the other an antibiotic. The most appealing approach to aza-tyrosine is the alkylation of some asymmetric equivalent of glycine enolate 57 and, probably, a protected version of 58. OH

NH2 HO2C

OH

NH2 HO2C

55; ( S)-L-tyrosine in proteins

NH2

OH +

X

HO2C

N

56; ( S)-L-Azatyrosine antibiotic

57

N 58

The unusual amino acid 56 has been synthesised6 by alkylation with the silylated bromide 61 prepared from commercially available 59. You might notice the radical bromination step giving only moderate yields. OH

N 59; commercially available

OSiPh2t-Bu

t-BuPh2SiCl imidazole DMF

N 60; 100% yield

OSiPh2t-Bu

NBS AIBN CCl4

Br

N 61; 51% yield

The chiral glycine equivalent 62 is available from Glytech Inc., Fort Collins, Colorado. Its sodium enolate reacts with 61 to give a reasonable yield of one pure diastereoisomer of 63 from which the silyl group is easily removed to give 64.

606

27 Asymmetric Induction IV Substrate-Based Strategy

Ph

O

O

Ph

O

Ph

N

OSiPh2t-Bu

O

1. NaHMDS

Ph

2. 61

N CO2Bn

N

CO2Bn 63; 60% yield, one diastereoisomer

62 Ph

O

O

OH

TBAF THF

N

Ph

N

CO2Bn 64; 79% yield, one diastereoisomer

Removing the chiral auxiliary is more tricky. One hydrogenation cleaves both benzylic C–O bonds and releases 65. This is not isolated but hydrogenated under stronger catalysis in acidic solution to cleave the benzylic C–N bond and release aza-tyrosine 56 as its dihydrochloride. The chiral auxiliary is destroyed in this process. 1. H 2 (50 psi) Pd(OH)2/C, THF

Ph

HO

O

OH

OH

CO2H

64 Ph

N H

N

2. H 2 (50 psi) Pd/C, HCl/H 2O

65

N H2N 56; 64% yield (.2HCl)

Seebach’s relay chiral units Probably the most important and useful of all these amino acid-based chiral enolates are those of Seebach.7 The simplest is made from proline 17 simply by forming the N,O-acetal 66 with pivaldehyde (t-BuCHO). The lithium enolate 67 is alkylated diastereoselectively with various electrophiles E⫹ to give one diastereoisomer of 68. H

t-BuCHO

N H

CO2H

17; ( S)-L-Proline

pentane cat. CF 3CO2H

E 1. LDA

O

N O

t-Bu 66; 92% yield

–78 ˚C THF

OLi

N

2. E

O

t-Bu

O

N O

t-Bu 67

68

There is a lot to explain here. It looks very odd that syn-66 is preferred and even more so that alkylation of the enolate 67 occurs on the same face as the undoubtedly large t-butyl group. Both these issues matter as the original chiral centre in proline is destroyed in 67 and only the newly introduced chiral centre in 66 retains the stereochemical information from proline. This centre acts as a relay for the stereochemical information. Others call this a ‘memory’ effect. The acid-catalysed formation of the N,O-acetal 66 is under thermodynamic control (acetal formation is reversible) and the conformation 66a shows that the molecule folds about the necessarily cis ring junction and the t-butyl group prefers to be on the outside (or exo- face).8 The enolate 67 has a flattened conformation 67a (probably more flattened than this!) and its alkylation is under kinetic control. Attack is preferred on the outside, exo-face. Note that this happens to restore the original configuration at the ex-proline chiral centre.

27 Chiral Enolates from Hydroxy Acids

O

H

OLi

O

N

607

O

=

O E

O

O

H

O

E

N

t-Bu 66

N

N

H

H

H

66a

67a

68a

The examples below show that alkylation with a variety of groups gives good yields of the purified diastereoisomers shown (69–71). The last example 71 shows that hydrolysis gives the alkylated proline 72. As these are tertiary amino acids, there is little chance of racemisation. Me Br

67

MeI

O

N

O

N

67 O

O

t-Bu 69; 93% yield

t-Bu 70; 87% yield

Ph Ph

Br

O

N

67

Ph 48% HBr

CO2H

N H

O

t-Bu 71; 91% yield

72; 71% yield

Alkylated prolines can be incorporated into synthetic peptides as mimics for the β-turn found in the conformation of folded proteins. One such compound9 combines a spirocyclic proline derivative 77 with tyrosine. The allyl group in 74 is oxidatively cleaved to give 75 and eventually coupled to the nitrogen atom of the amide in 76 by a Mitsunobu reaction.

RNH2

70

O

N H

O

N

N

Cbz NHR 74

NHR 73

CHO

OsO4 NaIO4

Cbz

O NHR 75

OH NaBH4 deprotect

O

N Cbz

NHR 76

Ph3P DEAD

N H

NR O 77

Chiral Enolates from Hydroxy Acids Seebach’s relay chiral units Even more famous is Seebach’s related chemistry on hydroxy acids10 such as lactic and mandelic acids from the chiral pool (chapter 23). Mandelic acid 78 gives the cis pivaldehyde acetal 79 in 74% yield after a recrystallisation to remove a small amount of the anti isomer. The lithium

608

27 Asymmetric Induction IV Substrate-Based Strategy

enolate 80 reacts on the side opposite the t-butyl group to give 81. As with the amino acids, the original chirality is restored. Hydrolysis gives the new hydroxy acid 82.

t-BuCHO

CO2H

HO

Ph

Ph

Ph

78; ( S)-(+)mandelic acid

pentane cat. TsOH or CF3CO2H

1. LDA

O

O O

OLi

O

–78 ˚C THF

O

t-Bu

t-Bu 80

79; 24:1 diasts recrystallise 74% yield Ph

KOH MeOH, H 2O

PrI

Ph

O

O

10 minutes 60 ˚C

O

HO

CO2H

t-Bu 81

82

Lactic acid 83 does much the same thing though the stereoselectivity in forming the initial acetal 84 is not so good - as might be expected with the smaller methyl group.

HO

Me

Me

Me

t-BuCHO

CO2H

pentane cat. TsOH or CF3CO2H

83; ( S)-(+)lactic acid

2. E

O

O

t-Bu

OLi

O

–78 ˚C THF

O

Me

R

1. LDA

O

O

O

O

t-Bu

t-Bu 85

84; 4:1 diasts; 87% yield recrystallise twice to 96% de

86

The scope and diastereoselectivity of reactions with various electrophiles are shown below. The only weak point is the poor diastereoselectivity in aldol reactions with aldehydes. These methods have been widely used as they are robust and reliable. Nevertheless the methods we have so far described for chiral enolates are less significant than the Evans chiral oxazolidinones in the next section. EtI, allyl and benzyl bromides R

ketones: acetone cyclopent- and hexanone HO

Me

R

HO Me

R O

O O

77-82% yield 95-98% de

H Me

R O

O

t-Bu

aldehydes MeCHO, PhCHO

O t-Bu 83-87% yield >95% de

O

O O t-Bu

84-85% yield 82-84% de two aldol diastereoisomers

27 Chiral Enolates from Evans Oxazolidinones

609

Chiral Enolates from Evans Oxazolidinones The main Evans oxazolidinones are 87/88, derived from phenylalanine, and 89/90, derived from valine. All are available at a price - higher naturally for the unnatural R-isomers. Generally 87/88 are preferred as most of their derivatives are crystalline so that purification of diastereoisomers is easier.11 O HN

O O

O

HN

Ph

O

HN

O O

HN

O

Ph

87 (S)-(–)-4-benzyl2-oxazolidinone

88 (R)-(+)-4-benzyl2-oxazolidinone

89 (S)-(–)-4-isopropyl2-oxazolidinone

90 (R)-(+)-4-isopropyl2-oxazolidinone

The preparation of each is easy and this is the route to 87/88 as described by Evans11 in Organic Syntheses. Reduction of the free acid 91 with the borane-dimethylsulfide complex is easy if smelly and the rest is straightforward. The same auxiliary 87 can be used with any acid: acylation with the acid chloride gives the chiral derivative 93 ready for enolate formation. 1. BF 3.Et2O

H2N CO2H Ph

OH

H2N

2. BH 3.Me2S

(EtO)2CO

Ph

3. NaOH, H 2O

91; ( S)-(+)-Phe 165 g scale

K2CO3

92; phenylalaninol 73-75% after recrystallisation

O HN

O

O R

O

N

1. BuLi

Ph 87; oxazolidonone 78-79% after recrystallisation

2.

O

O

Ph 93; chiral acid derivative R = Me; 91-96%

R Cl

The benzyl group has three jobs to do: (i) force conformation 93a with R1 anti to N, (ii) Make the ‘cis’ or ‘Z’ enolate 94 selectively from this conformation, and (iii) direct the electrophile to the top face of the enolate to give 95. Chapter 4 reveals the reason why this enolate is called Z. Cleavage of the product 95 to give the free acid 96 is best done with LiOOH made from H2O2 and LiOH. This is the best nucleophile for attacking the ‘exo’ carbonyl without epimerisation through enolisation. O

O

O LDA

R1

N Ph

O

R

Li

O

N Ph

93a

94; ' Z'-enolate

O

O R2Br

1

O

H2O2 LiOH

1

R

N R2 Ph

O

O R1

OH R2

95

96

610

27 Asymmetric Induction IV Substrate-Based Strategy

If the other enantiomer is wanted, two strategies are available. Either the alkyl groups can be added in the reverse order, if this is mechanistically reasonable, or the oxazolidine 93c can be constructed from nor-ephedrine from the chiral pool (chapter 23). Hence both enantiomers of 96 can be made in good yield and high ee.

R2

1. LDA

N

O

O

O

O R2

N

O 2. R 1Br

O H2O2 LiOH

R1 Ph

Ph

95b

93b O

O R1

N

O

2. R 1Br

Me

O

O R1

1. LDA

ent-96

N

H2O2 LiOH

O

R2 Me

Ph

Ph

95c

93c

Application of asymmetric alkylation with Evans auxiliaries Modern treatments for arthritis includes inhibition of the collegenase enzymes that attack connective tissue in the affected joints. One such is the peptide analogue 97 clearly made from a tyrosine derivative and the unusual carboxylic acid 98 also containing a hydroxamic acid. This unusual functional group can be made from a carboxylic acid so the target is a diacid derivative of 98 with the two acids distinguished in some way. OMe H

O

H

C–N N H O

OH

amide

O

+

O HN

HN

OMe

O

H2N O

OH

OH 98

97; collegenase inhibitor

(S)-tyrosine derivative

The Evans alkylation method is ideal for this compound.12 The phenylalanine-derived auxiliary 87 is acylated to make the starting material 99. The sodium enolate is alkylated with t-Bu bromoacetate to give 100 with the expected stereoselectivity. Exo-cleavage gives 101 that has one free CO2H group and one protected as a t-Bu ester. This was used to make 98. O

O N

O

99

LiOH H2O2

N 2. Br

Ph

O

O

1. NaHMDS THF, –78 ˚C

O

CO2t-Bu

H

O OH CO2t-Bu

t-BuO2C Ph 100; 74% yield, >95% de

101; 91% yield 'enantiomerically pure'

27 Aldol Reactions with Evans Oxazolidinones

611

Aldol Reactions with Evans Oxazolidinones The syn aldol reaction with boron enolates The boron enolate of the starting material 102 reacts with an aldehyde such as PhCHO to give 103 with excellent selectivity: the induction from the oxazolidinone to the methyl group is near perfect and the relative stereochemistry of the aldol11 is ⬎500:1. Hydrolysis gives the hydroxy-acid 104 in nearly 100% ee and recovered auxiliary 87. OH

O

O N

1. Bu2B

O

OH

O

O

OTf

LiOH, H 2O2

N

Ph

2. Et 3N

O

Ph

O OH

THF, H 2O

3. PhCHO

Ph

Ph

104; 80-90% yield >99% ee + 89% recovered auxiliary 87

103; 84-93% yield 500:1 syn aldol selectivity >97% one diastereoisomer

102

The stereochemistry is more tricky to understand than that of the alkylations. Aldol reactions of the lithium enolates such as 94 are nowhere near so stereoselective as either alkylation or boron-mediated aldols. The key point is that lithium can coordinate to four different atoms at a time whereas a Bu2B group has only two spare coordination sites. The Z-boron enolate 105 is formed but coordination of the incoming aldehyde to the boron atom displaces the oxazolidinone C=O group and the left hand half of the molecule rotates about the C–N bond by 180⬚ to give 106. Why it should do this is open to question. Evans suggests that there is better (longer) conjugation as a result. Aldol reaction through a six-membered cyclic mechanism 106 gives 107, more usually drawn as 107a. Bu O

1. Bu 2BOTf

93a

R1

Bu B

O

N

2. Et 3N

R2 R1 R2CHO

Bu

O

B

Bu

O

Ph 105

106

R2

OH R1

O

N

O

Ph

HO

O

O

N

O

2 = R

O

O

O N

O

R1 Ph

Ph 107

107a

The stereochemistry of the aldol step 106 results from the chair-like conformation of the transition state 109a with R1 forced to be axial because it comes from the Z-enolate 108 and R2 choosing to be equatorial (see explanation of simple syn/anti aldols in chapter 21). The absolute stereochemistry comes from the blocking of the lower face of the enolate by the benzyl group of the chiral auxiliary 109b.

612

27 Asymmetric Induction IV Substrate-Based Strategy

H H

R1 R2CHO

105

Bu2B

O

Bu H

NR2

R2 O

NR2

B Bu

Bu H

O

O

B

O R1 109b

Bu

109a

O N

R2

R1

Z-enolate 108

O

H

Ph

Anti aldols by Lewis acid-catalysed reactions with Evans oxazolidinones Though either enantiomer of a syn-aldol can be made by using the right auxiliary in an Evans aldol reaction the anti aldols cannot be made this way. The addition of a Lewis acid catalyst transforms the situation.13 Using the valine-derived chiral auxiliary 89, the same Z-boron enolate 111 is used but the aldehyde is added in the presence of a threefold excess of the Lewis acid Et2AlCl. The product is predominantly one enantiomer of an anti-aldol 112. Bu O

O

O 1. Bu2B

R1

N

Bu B

O

OTf

R

O

O

O

OH R2CHO

1

N

2. Et 3N

O

R2

3 equivalents Et2AlCl

N

O

R1

63-86% yield

110

111; Z-boron enolate

112; anti-aldol

Instead of complexing to the boron atom, the aldehyde prefers the Lewis acid and the ‘open’ or ‘extended’ transition state 113 leading to the stereochemistry shown in a Newman projection as 114. In fact, as well as the anti-aldol 112, both enantiomers of syn-aldols are formed in small amounts - around 5% of 116 and traces of 117. Bu O

R2

R2CHO

R1

111 3 equivalents Et2AlCl

Bu B

O

N

R

O

H

O

N

R2

Al

O

H

O

O

HO 1

H

HO

H Et

O

O

R2 1

R

H

N

O

H

Et

Cl

113

114

112

With a ‘smaller’ Lewis acid (TiCl4) a slightly different ‘open’ or ‘extended’ transition state 115 is preferred leading predominantly to one syn-enantiomer 116. These variations on aldol reactions with the Evans auxiliaries allow the synthesis of almost any enantiomer in good yield. Bu 1. Bu 2BOTf

110

R2

2. Et 3N 3. R

2CHO

2 equivalents TiCl4

O

H Cl4Ti

Bu B

N 1

R

O

O

O

O

OH

O R2

N R1

OH O

R2

N R1

H 115

116; 87-94 parts of 50-72% yield

O

O

117

O

27 Chiral Auxiliaries

613

Chiral Auxiliaries Asymmetric aldol, conjugate addition, and Diels-Alder reactions compared Since conjugate (Michael) addition and Diels-Alder reactions use α,β-unsaturated carbonyl compounds, asymmetric versions of these reactions could use the auxiliaries that we have seen in aldol reactions in the form of 118 and 119. Diels-Alder reactions work very well with these unsaturated amides and also with amides 121 derived from Oppolzer’s chiral sultam14 120, prepared simply from camphorsulfonyl chloride. O

O R

N

O

O O R

N

O

NH S

118

119

S

O

O 120

Ph

R

N

O

O

O

121

There are other considerations than mere analogy. If R* represents some chiral auxiliary, a Diels-Alder reaction on 124 gives a product 125 with two new chiral centres, one, at C-2 in exactly the same place as in the aldol products such as 107. Michael (conjugate) addition to 124 on the other hand gives a compound 122 with only one new chiral centre (C-3) one atom further away from the chiral auxiliary. Another substitution pattern 128 still creates a new chiral centre at C-2 in the Diels-Alder adduct 129 but this time Michael addition does the same. However, the new centre is not created in the conjugate addition step, as that gives the enolate 127, but in the protonation of 127, an inherently less well controlled and probably reversible step. We shall therefore consider Diels-Alder reactions first and tackle the more troublesome Michael additions in the next section. R X

O

OH

X

O

H

R

R*

3

R

122

R*

R

1

R*

R 126

125 O

OH

3

2

R*

2

O

R

3

X

X

R*

R*

124

H 2

3 2

123 O

X

3

X 3

O 2

1

2

R*

R 128

R 127

R*

129

A study of the substituents on the oxazolidinone ring revealed that the highest diastereoselectivity in methylation of the lithium enolate of 122 was achieved when R ⫽ t-Bu. The normally used auxiliaries, with R ⫽ i-Pr and Bn respectively, performed less well, but the compound with R ⫽ Bn was the best compromise between availability and selectivity.15 1. LDA, –30 ˚C, THF

O

O N

O

O N

O

O

O O

N

+

2. MeI

R 130

R 131a

R 131b

O

R

131a:131b

Ph

2:1

i-Pr

6:1

benzy1

20:1

t-Bu

>100:1

614

27 Asymmetric Induction IV Substrate-Based Strategy

The Diels-Alder reaction with the analogous compound 132 needs Lewis acid catalysis and shows the expected regioselectivity 135. The predicted stereoselectivity is endo addition to the face of the diene opposite group R 133 to give the diastereoisomer 135a.

Et

Et

Et

Al O

O

H

O

O

Et Al

O

O O

O

Et2AlCl

N

O

N

–30 ˚C

N

R

R

R

133

132

O

O

H

CH2Cl2 toluene

R

N

O

134

135a

The isomer 135a is indeed favoured and the actual results for the four compounds are remarkably like those for the alkylation. The minor product 135b has the anti stereochemistry across the new six-membered ring but these two centres have the opposite stereochemistry relative to that of the chiral auxiliary. The message is that, in most cases, R ⫽ benzyl is the sensible choice but that if it performs poorly there ought to be a big improvement with R ⫽ t-Bu.

O

O N

O

O O

N

O +

N

–30 ˚C

R

CH2Cl2, toluene

132

R

135a

135a:135b

Ph

4:1

i-Pr

10:1

benzy1

17:1

t-Bu

68:1

O

Et2AlCl

R

R

O

O

135b

The Asymmetric Diels-Alder Reaction Introduction Most asymmetric Diels-Alder reactions have the chiral auxiliary attached to the dienophile as in the example we have just seen.14 One of the earliest was the blocked α-hydroxyenone16 136 that cyclises with electron-rich dienes to give adducts such as 139 and was used in a synthesis of pumiliotoxin 143 based on that of Overman.17 The weakness here is the cleavage of the auxiliary requires reduction to a diol and periodate cleavage with the inevitable destruction of the auxiliary. O O

CHO NHCO2Bn

OH 136

1. LiBH 4

NH ZnCl2, –78 ˚C toluene

OH

CO2Bn 137

2. NaIO 4 MeOH, H 2O (diol cleavage)

NH CO2Bn 138

27 The Asymmetric Diels-Alder Reaction

615

This time the Lewis acid is ZnCl2 and this holds the chiral dienophile in a fixed conformation 139. Reaction through the endo arrangement 140 gives one enantiomer of the product 137 in 67% yield and ⬎99% ee.

cis H

H

ZnCl2, –78 ˚C toluene

Zn

Zn

139

OH

H

CO2Bn

O

N H

O O

H

H

136 O

t-Bu

H O 140

NH CO2Bn

137

The rest of the synthesis of pumiliotoxin is straightforward involving a Horner-WadsworthEmmons olefination (chapter 15) and a hydrogenation that accomplishes reductive amination and the control of another chiral centre. This synthesis is also discussed in chapter 21. O

O

(MeO2P

H

O

H

Bu

Bu

H2, Pd/C

138 NaH

H

NHCO2Bn

HCl, EtOH

141 H

H

H

H

NH2 O

Bu

142

H

N H H H 143; (+)-pumiliotoxin C

Diels-Alder reactions with Oppolzer’s chiral sultam Oppolzer’s chiral sultam 120 is easily converted to the amide 144 with α,β-unsaturated acid chlorides and reacts with dienes with the Lewis acid EtAlCl2 as catalyst. The adduct 145 is essentially a single compound and LiOH hydrolysis gives the simplest of all Diels-Alder adducts 146 as a single enantiomer in good yield.18 EtAlCl2 –78 ˚C CH2Cl2

NH S O O 120; Oppolzer's chiral sultam

S O O 144

CO2H N

N O

S O O O 145; 81% yield, 99% de after recrystallisation

LiOH

146; 100% ee 82-84% yield after recrystallisation

616

27 Asymmetric Induction IV Substrate-Based Strategy

The Lewis acid holds the acylated sultam in conformation 147 and the diene approaches from the less hindered lower face away from the camphor cage. As both enantiomers of camphor are available, either enantiomer of Diels-Alder adducts can be prepared in this way. This one 146 was used in an asymmetric synthesis of shikimic acid 148, the biological precursor of aromatic rings.19

EtAlCl2 –78 ˚C CH2Cl2

N

144

146

S O O

1 2 CO2H

HO 6

?

3

4 HO 5

O

OH

Al

148; shikimic acid

147

Et

Only relative stereochemistry (chapter 21) now needs to be controlled and an iodolactonisation followed by elimination gives 149, epoxidation and elimination 151, a second epoxidation and then simple treatment of 152 with basic methanol gives the methyl ester 153 of the target molecule.

1. I 2, NaHCO3

O

O

O

O ArCO3H

1. DBU, Ph3P Me3SiBr

O

O

ArCO3H

146 2. DBU

2. 1M HCl

HO

O 150; 81% yield

149; 84% yield

151; 85% yield

O

O

MeOH Na2CO3

HO

O 152; 89% yield

153; methyl shikimate 96% yield

The mechanism of the last step is interesting. The lactone 152 cannot form an enolate as it would be at a bridgehead but after methanol has opened the lactone to give 154, the enolate 155 can form and immediately fragments. Notice that the original chiral centre, introduced in the asymmetric Diels-Alder reaction, is destroyed in this step after it was used to control the formation of the three OH groups. This synthesis is also discussed in chapter 21.

O

O

O

MeOH Na2CO3

HO

O 152

HO

O OMe

HO

CO2Me

HO OMe HO

HO

O 154

HO

O 155

OH 153

27 The Asymmetric Diels-Alder Reaction

617

Diels-Alder reactions with pantolactone as chiral auxiliary We saw pantolactone 156 acting as a reagent in chapter 26. Here it esterifies an unsaturated acid to control a Diels-Alder reaction. It has two main advantages. It is easy to use on a large scale and it is easily removed from the product. The simple acrylate derivative 157 reacts with cyclopentadiene to give essentially pure (by HPLC) adduct 158 in 81% yield after two recrystallisations.20

O

OH O

O

Cl

O

10 mol% TiCl 4

7:1 CH2Cl2, petrol

O

CH2Cl2, –24 ˚C

156; (R)-(–)- D pantolactone

O

O

O

Et3N

O

O

1 hour –10 ˚C

O 158; 81% yield after two recrystallisations

157; 88% yield 100 g scale distilled

Hydrolysis is simpler than those we have seen before and the separation of the crystalline DielsAlder adduct 159 from the soluble pantoic acid 160 simplicity itself. Acidification then recreates the chiral auxiliary 156 for the next reaction.

OH

1. 5M NaOH

158 275 g scale

1.1 L THF 550 mL MeOH 1.1 L H2O

OH CO2H

90-95 ˚C 2-3 hours

O

+ 2. 4M HCl

O

extract EtOAc

OH

O OH 159; crystallises 135 g; 89% yield 100% ee

160 pantoic acid in solution

156; D-pantolactone 122 g; 85% yield 100% ee

Chiral auxiliaries attached to the diene Most asymmetric Diels-Alder reactions have the chiral auxiliary attached to the dienophile as an ester or an amide. There is no such obvious place to attach an auxiliary to a diene but the chiral analogue 161 of Danishefsky’s diene has a C2 symmetric amine 164 attached as an enamine. Cycloaddition without a Lewis acid gives good selectivity in the formation of adduct21 162. Reduction and hydrolysis releases the enone 163 in reasonable ee.

RO

163; 79% yield 88% ee

RO

O

CHO

Ph

N

Ph

–10 ˚C toluene

Ph

N

CHO

1. LiAlH 4

Ph

2. 10% HF MeCN

OH

Ph 161

162; 71% yield R = t-BuMe2Si

H N

164

Ph

618

27 Asymmetric Induction IV Substrate-Based Strategy

Danishefsky’s diene 165 reacts with the same dienophile to give racemic 163. The hydrolysis of the adduct is much the same in both cases but protonation of the nitrogen atom 166 is easier. RO

RO

RO OH H

OMe

N

Ph

163

OH

Ph + 164

165

166

Improved Oxazolidinones: SuperQuats In spite of the excellence of all these auxiliaries in their various fields, the oxazolidinones remain the most widely useful. And they, like many of the others, still cause problems in the vital step of hydrolysis to remove the chiral auxiliary. Here are three examples22 where hydrolysis is carried out with the standard reagents: 4–8 equivalents of 30% H2O2, then 2 equivalents LiOH in 3:1 THF:H2O. The first two examples give significant amounts of unwanted endo hydrolysis while the last is partially epimerised at C-2.

R1

N

O

O

hydrolysis exo endo O O

H N

O

O N

O

OH

O

O 2

O

N

O

N3 Ph

O

Ph Ph

169; epimerisation 167; 91% yield >99% ee 168; 76% yield >99% ee 16% endo hydrolysis at C-2 during hydrolysis 6% endo hydrolysis

R2

One solution is to make the oxazolidinone ring more resistant to hydrolysis by packing it with substituents. Among the best examples are the ‘SuperQuats’ developed by Steve Davies at Oxford.23 Valine (the unnatural isomer in this case) is used to construct an oxazolidinone with two geminal methyl groups 171. We show a simple alkylation of the lithium enolate: note the high yield of both the final product after hydrolysis 174 and the chiral auxiliary 171 ready for recycling. There is no endo hydrolysis nor epimerisation.

BocHN

1. 4 x MeMgBr

HN

CO2Me

O

O

O O

1. BuLi

N

O

2. EtCOCl

2. t-BuOK, THF

170; ( R)-D-Boc-valine 172; 73% yield

171; 63% yield O

O

1. LiHMDS THF, –78 ˚C

N 2. BnBr

O

O O

Ph 173; 71% yield >95% de

LiOH THF/H2O

OH

+

HN

O

Ph 174; 98% yield >95% ee

171; 98% yield

27 Asymmetric Michael (Conjugate) Additions

619

Asymmetric Michael (Conjugate) Additions Michael additions with 8-phenylmenthyl esters of unsaturated acids Though some of the auxiliaries we have been discussing will initiate asymmetric Michael additions, there are two that are particularly effective and we shall discuss mainly those. The fi rst is based on a cyclohexane ring with an equatorial OH and an aromatic blocking group. The most famous is 8-phenylmenthol 175, made from the natural terpene pulegone24 (chapter 23) though later discoveries have shown that much simpler auxiliaries, even the bare bones 177 can work quite well.25 Ph

Ph OH

O

R OH

O

O

R O

176

175; 8-phenylmenthol

177

178

Conjugate additions of organocuprates as their BF3 complexes (RCu•BF3) add with good stereo-selectively to 176; R ⫽ Me even though the nearest stereogenic centre in 176 to the one being created has a 1,5 relationship, evident in the product 179. Two examples, 180 with an aryl and 181 with an alkyl nucleophile, show high selectivity ⬎99% ee in both cases.26 Other nucleophiles such as amines and thiols also give good results.27

O

RCu•BF3

Me

HO

Ph Me

176; R = Me O

R

ester hydrolysis

H

O

Ph

HO

Me O

179

H

Bu

H

(S)-(–)-180 75% yield >99% ee (R)-(–)-181 76% yield >99% ee

This and many other results, including the much lower diastereoselectivity when the cis alkene is used, suggest that π-stacking between the alkene and the benzene ring means that the conformation of 176 in reaction is 176a rather than 176b. The same conformation with the cis-alkene 182b would be destabilised. O

O

O

O

O R

O

R 176a

176b

182b

R

The same auxiliary gives high enantioselectivity in Diels-Alder reactions28 with cyclopentadiene and butadiene - it gives good yields of 146 used in a synthesis of sarkomycin.29 In a later section of this chapter you will see it used in asymmetric Birch reduction.

620

27 Asymmetric Induction IV Substrate-Based Strategy

Chiral auxiliaries attached elsewhere in asymmetric Michael additions Cyclic α,β-unsaturated carbonyl compounds cannot have a chiral auxiliary attached in the same way. We shall illustrate the many ingenious solutions to this problem with two heterocyclic examples. Both also show the effects of modifications of 8-phenylmenthol. The first is a pyridinium salt 184 and the chiral auxiliary is attached through an ester group to the nitrogen atom of the parent pyridine 183; R3Si ⫽ TIPS, i-Pr3Si. Addition of a Grignard reagent to the less hindered carbon atom next to N⫹ gives the enol ether 185 and hence the enone 186 on workup.30 Though many variations in the group R* were tried, the best results were with 8-phenylmenthyl itself (R*OH ⫽ 175). OMe

OMe SiR3

OMe SiR3

R*OCOCl

N

RMgX

R

N 183

O SiR3

N

SiR3

H H2O

R

N

CO2R*

CO2R*

CO2R*

184

185

186

The second is a dihydropyridone with the chiral auxiliary also attached to nitrogen31 through an ester group 187. The nucleophile is an allyl silane, requiring Lewis-acid catalysis for reaction (chapter 12). The chiral auxiliary is attached to the opposite end of the electrophilic enone system this time but at about the same distance from the prochiral electrophilic carbon. Though 187; R ⫽ Ph gave quite good results (92:8 in favour of 188), much better results were achieved with 187; R ⫽ 2-naphthyl (97:3) suggesting that some conformation such as 189, with good π-stacking making for a very hindered face of the enone, is responsible. O

O Me3Si

O

N

N

O

TiCl4

O

O

TiCl4

N

O

O

O R

TiCl4

R

187

188

189

The simpler auxiliaries we mentioned at the start of this section are effective in the addition of amines under pressure to crotonates. Of many auxiliaries tried, the simple trans cyclohexanol carrying a β-naphthyl substituent on a tertiary carbon, was the most effective.25

Ph2CHNH2

OH

190

O O 191

MeOH 20 ˚C, pressure

O O 192

NHCHPh2

The preferred conformation for the reaction is 193 with excellent π-stacking: the nucleophile adds from the front as drawn. These chiral auxiliaries are easily made as either enantiomer so

27 Other Chiral Auxiliaries in Conjugate Addition

621

both series can be entered. The chiral auxiliary and protecting group can be removed reductively to give the simple amino alcohol (S)-(⫹)-195 and this sequence is complementary to the Davies chiral amine addition to unsaturated esters discussed in chapter 26. O LiAlH4

O

HO

H2, Pd(OH) 2/C

HO

192 NHCHPh2

193

NH2

MeOH

(S)-(+)-194

(S)-(+)-195

Other Chiral Auxiliaries in Conjugate Addition The Evans oxazolidinones The original Evans auxiliaries have been used in a synthesis of the heart drug (⫹)-diltiazem 200 based on asymmetric conjugate addition of a sulfur nucleophile. The enone 197 was prepared by an aldol condensation with valine-derived 196 and anisaldehyde. This gave a 4:1 mixture of Z and E-197 but fortunately both isomers reacted with an excess of ortho-aminobenzene thiol and base to give about the same ratio of diastereoisomers of the adduct 198. Cyclisation with Me3Al removed the chiral auxiliary and diltiazem 200 then needs only the removal of the protecting group.32 This is a rare example of successful asymmetric Michael addition to a trisubstituted alkene. O

O N

O

O 1. LDA 2. ArCHO

O

OR

N

3. MsCl, Et 3N

NH2

O

RO

MeO

SLi NH2

4. DBU

196 NH2 S

O

O N

MeO

SH

197; 61% yield, 4:1 Z:E H O N OR

Me3Al

O

CH2Cl2

H N

OH

TiCl4

S

O

S

RO

198; 97% yield, 82:18 diastereisomeric ratio

199, 78% yield

OMe

200, 83% yield

OMe

Chiral sulfoxides Cyclic enones or unsaturated lactones inevitably have a fixed conformation (s-trans), but the chiral auxiliary cannot be attached to the carbonyl group. Chiral sulfoxides can be attached directly to the double bond (cf. enone synthesis strategy of chapter 5) and have been used chiefly by Posner33 based on work by Solladié34 in asymmetric Michael additions. Any sulfur-based functional group can be removed reductively with Raney nickel. Posner35 has used this strategy in the synthesis of the anti-tumour lignan (⫺)-podorhizon 201. Acylation of the lactone 203 with 202 would give podorhizon and 203 could be made by asymmetric Michael addition of 204 to the achiral butenolide 202. But where can the chiral auxiliary be attached?

622

27 Asymmetric Induction IV Substrate-Based Strategy

O

O

O MeO

MeO

X

O 1,3-diCO

MeO

MeO

O

O O

asymmetric cuprate addition

O Cu 205

+ MeO

MeO

202 O

O

O

O 204

O 203

O 201

The answer is at carbon as a sulfoxide 206 to make the butenolide chiral and to enhance the electron deficiency of the alkene, so that a Grignard reagent will do instead of a cuprate. After the sulfoxide is removed and the acylation complete, podorhizon is formed in good yield and 95% ee. O

p-Tol

O

O

1. ZnBr 2

S O

2.

MgBr

p-Tol

H

S

O O

(–)-206

Raney Nickel

203 70% yield

202; X = OCO2Et Koga's proceedure

201 (–)-podorhizon 95% ee

O O

O 207

O

A Lewis acid is needed to hold the sulfoxide in a fixed conformation 208: addition then occurs to the face of the alkene occupied by the lone pair and opposite to the aryl group.

O 1. ZnBr 2

(–)-206

p-Tol

Zn

O

O 2. RMgBr

S

p-Tol

S

H

O O

O R 208

207

Asymmetric Birch Reduction Birch reduction of benzene The normal Birch reduction is most interesting when applied to aromatic ethers 209 or acids 213. The addition of two electrons may make a dianion in which the charges keep away from the ether 210 but conjugate with the acid 214. Protonation of 210 gives the enol ether 211 and hence the non-conjugated enone 212. The dianion 214 has a proton which transfers to the less stable anion leaving the enolate 215 that can be alkylated to give 216. None of these compounds is chiral and there appears to be little scope for asymmetric induction.

27 Asymmetric Birch Reduction

623

Na, NH 3(l)

OMe

H3O

OMe

t-BuOH

209

O

OMe

t-BuOH

210

211

212

O CO2H

R

CO2H

Na, NH 3(l)

213

O

214

RBr

215

CO2H

216

Schultz’s remarkable work combines ether and acid in the same molecule and adds a chiral auxiliary (the same proline derivative 26 used in SAMP) to the acid.36 Both enantiomers of the starting material 217 are available from Aldrich at about the same price. Birch reduction with potassium in ammonia gives the chelated intermediate 218 that is alkylated on the face opposite the CH2OMe side chain to give 219 and hence 220 by hydrolysis of the enol ether.

OMe

OMe

N

N K, NH 3 THF, –78 ˚C

O

1 x t-BuOH

OMe

O

O

RBr

K

–78 ˚C

Me 218

217 OMe R

OMe

N R

6M HCl

N

O

O

OMe

O

219

220

An interesting version of iodolactonisation is used to remove the chiral auxiliary from 220. Iodine attacks the alkene with participation of the amide oxygen atom 221 to give a salt 222 that hydrolyses to the iodolactone 223 with recovery of the auxiliary 26 if required.

O 220

O

O OMe

OMe I

O 221

H2O

R

R

R

THF

O

N

N

I2, H 2O

I

O 222

O

I 223

624

27 Asymmetric Induction IV Substrate-Based Strategy

The iodolactone 223 can be fragmented in two different ways by different reagents to give the lactones 224 or 225 that bear little resemblance to the original benzene ring. I R O

LiOR

CO2R1

O

O

1

O

R LiOH

R

THF

O

H2O, THF

O 224

223

CO2H

O 225

Asymmetric Birch reduction of heterocycles 8-Phenylmenthol is the auxiliary in an asymmetric Birch reduction of pyrroles by Donohoe37 Lithium in ammonia does the reduction and the enolate is trapped with various alkyl halides. Hydrolysis of the esters 227 releases the enantiomerically enriched (78–90% ee) dihydropyrroles 228 in good yield. Furans give similar products with a C2 symmetric amine as auxiliary. This should become a general route to a variety of heterocycles.

Ph

R Ph

Li, NH 3, –78 ˚C

O

(MeOCH2CH2)2NH

N Boc

O

N RX

O

Boc

226

O 227

R

1. CF 3CO2H 2. NaOH, H 2O 3. (Boc) 2O

CO2H

N Boc

228

References General references are given on page 893 1. R. M. Williams, Synthesis of Optically Active Amino Acids, Pergamon, Oxford, 1989. 2. K. Q. Do, P. Thanei, M. Caviezel and R. Schwyzer, Helv. Chim. Acta, 1979, 62, 956. 3. D. Enders, P. Fey and H. Kipphardt, Org. Synth., 1987, 65, 173, 183. 4. U. Groth and U. Schöllkopf, Synthesis, 1983, 37; D. Pettig and U. Schöllkopf, Synthesis, 1988, 173. 5. R. M. Williams and M. Im, J. Am. Chem. Soc., 1991, 113, 9276. 6. S. R. Schow, S. Q. DeJoy, M. M. Wick and S. S. Kerwar, J. Org. Chem., 1994, 59, 6850. 7. D. Seebach, M. Boes, R. Naef and W. B. Schweizer, J. Am. Chem. Soc., 1983, 105, 5390. 8. Clayden, Organic Chemistry, chapter 33. 9. M. G. Hinds, J. H. Welsh, D. M. Brennand, J. Fisher, M. J. Glennie, N. G. J. Richards, D. L. Turner and J. A. Robinson, J. Med. Chem., 1991, 34, 1777. 10. D. Seebach and R. Naef, Helv. Chim. Acta, 1981, 64, 2705; G. Fráter, U. Müller, and W. Günther, Tetrahedron Lett., 1981, 22, 4221. 11. J. R. Gage and D. A. Evans, Org. Synth., 1990, 68, 77, 83. 12. R. P. Beckett, M. J. Crimmin, M. H. Davis, and Z. Spavold, Synlett, 1993, 137. 13. M. A. Walker and C. H. Heathcock, J. Org. Chem., 1991, 56, 5747. 14. W. Oppolzer, Angew. Chem., Int. Ed., 1984, 23, 876. 15. D. A. Evans, K. T. Chapman, and J. Bisaha, J. Am. Chem. Soc., 1988, 110, 1238. 16. S. Masamune, L. A. Reed, J. T. Davis, and W. Choy, J. Org. Chem., 1983, 48, 4441.

27 References

625

17. L. A. Overman, G. F. Taylor, C. B. Petty and P. J. Jessup, J. Org. Chem., 1978, 43, 2164. 18. W. Oppolzer, C. Chapuis and G. Bernardinelli, Helv. Chim. Acta, 1984, 67, 1397. 19. P. A. Bartlett and L. A. McQuaid, J. Am. Chem. Soc., 1984, 106, 7854. 20. T. Poll, A. Sobczak, H. Hartmann, and G. Helmchen, Tetrahedron Lett., 1985, 26, 3095; H. Chang, L. Zhou, R. D. McCargar, T. Mahmud and I. Hirst, Org. Process Res. Dev., 1999, 3, 289. 21. S. A. Kozmin and V. H. Rawal, J. Am. Chem. Soc., 1999, 121, 9562. 22. D. A. Evans, T. C. Britton, and J. A. Ellman, Tetrahedron Lett., 1987, 28, 6141. 23. S. D. Bull, S. G. Davies, S. Jones and H. J. Sanganee, J. Chem. Soc., Perkin Trans. 1, 1999, 387. 24. O. Ort, Org. Synth, 1987, 65, 203. 25. F. Dumas, B. Mezrhab, J. D’Angelo, C. Riche and A. Chiaroni, J. Org. Chem., 1996, 61, 2293. 26. W. Oppolzer and H. J. Löher, Helv. Chim. Acta, 1981, 64, 2808. 27. J. d’Angelo and J. Maddaluno, J. Am. Chem. Soc., 1986, 108, 8112. 28. E. J. Corey and H. E. Ensley, J. Am. Chem. Soc., 1975, 97, 6908. 29. R. K. Boeckman, P. C. Naegely and S. D. Arthur, J. Org. Chem., 1980, 45, 752. 30. D. L. Comins and L. Guerra-Weltzien, Tetrahedron Lett., 1996, 37, 3807. 31. M. Sato, S. Aoyagi, S. Yago and C. Kibayashi, Tetrahedron Lett., 1996, 37, 9063. 32. O. Miyata, T. Shinada, I. Ninomiya and T. Naitu, Tetrahedron Lett., 1991, 32, 3519. 33. G. H. Posner, Morrison, 2, 225. 34. M. Hulce, J. P. Mallamo, L. L. Frye, T. P. Kogan and G. H. Posner, Org. Synth., 1986, 64, 196. 35. G. H. Posner, T. P. Kogan, S. R. Haines and L. L. Frye, Tetrahedron Lett., 1984, 25, 2627. 36. A. G. Schultz, Chem. Commun., 1999, 1263. 37. T. J. Donohoe, P. M. Guyo and M. Helliwell, Tetrahedron Lett., 1999, 40, 435.

28 Kinetic Resolution Types of Reactivity The Water Wheel An unselective wheel A selective wheel S Values, Equations & Yields Standard Kinetic Resolution Reactions Chiral DMAP Epoxidation reactions Unwanted kinetic resolutions Enzymes Dynamic Kinetic Resolution Reaction of epichlorhydrin α-Acetoxysulfide Enzymes and metals together Hydrogenation Parallel Kinetic Resolutions Regiodivergent resolutions Double Methods

Desymmetrisation then kinetic resolution Resolution is the separation of enantiomers. We have seen resolution before in Chapter 22. A kinetic resolution is a resolution which works simply because, under the right conditions, one enantiomer of a racemic mixture reacts faster than the other.1

Types of Reactivity With a kinetic resolution, the substrate will already have a chiral centre. The question comes down to how much notice the reaction conditions take of this chirality. So for instance, allylic alcohol 1 could be acylated. Both enantiomers react at the same rate and the reaction pays no attention at all to the chiral centre. For a kinetic resolution to operate, that chiral centre would have to demand attention so that one enantiomer reacted and the other did not.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

628

28 Kinetic Resolution

OH

No Resolution

O O

1

2 OH

Kinetic Resolution

O

OH O

1

(R)-1

(S)-2

The reactions above illustrate another point. This simple acylation of a racemic alcohol can be turned, in principle, into a kinetic resolution. We shall see several reactions we have encountered in this book that also can be involved in kinetic resolutions. The above reaction did not create any extra chiral centres but that too is possible – even quite common – and naturally adds to the complexity of the reactions.

The Water Wheel A good visual depiction of kinetic resolution is a water wheel image. Imagine a tray containing black and white balls which have to run through the water wheel in order to get to the tray underneath The black and white balls represent R and S enantiomers and moving through the wheel is a transformation that yields the products in the tray underneath. This visualisation will also allow us to explain one of the more subtle and crucially important aspects of kinetic resolution—the degree of completion. R

S

Starting Materials

Products

An unselective wheel With an unselective wheel, the balls just flow randomly through the wheel and we get a 50/50 mix black and white products and, similarly, the starting materials are consumed in an unselective way. Unselective Catalyst R

Starting Materials

Products

S

28 The Water Wheel

629

Unselective Catalyst R

S

Starting Materials

Products

A selective wheel Imagine now a wheel which is selective for the black balls and prefers to let them through. The result is a product that is rich in black balls and a starting material that becomes increasingly rich in white balls. The wheel is not perfect, however, and occasionally lets a white ball through. Enantioselective Catalyst R

S

Starting Materials

Products

It is instructive to look at how rich the products and starting materials are in black and white balls as the reaction proceeds. When the reaction is 50% complete we have 80% ee of product and 80% ee of starting materials. One of the white balls has made it through to the product. Reaction at 50% completion

Enantioselective Catalyst R

S

Starting Materials 80% ee

Products

80% ee

It is a crucial feature of kinetic resolutions that if we want a very good enantiomeric excess in our starting materials then we will have to run the reaction past 50%. It is typical with kinetic resolutions

630

28 Kinetic Resolution

to find that the best quality material is what is left behind. If we continue to 75% completion then we can see that all of the black balls have been consumed and the remaining starting material has 100% ee. This is at the cost of the ee of the product of course which has fallen to 33% ee.

Reaction at 75% completion

Enantioselective Catalyst R

S

Starting Materials 100% ee

Products

33% ee

Keep this in mind. When you see a starting material with 98% ee in a paper, look to see how much conversion was needed to leave this sort of quality behind. If only 51% conversion was needed it is excellent, but if the workers needed to run the reaction to 90% conversion, then there can only be at most, 10% of this quality starting material – not so good! We look at this in more detail in the next section.

S Values, Equations & Yields With the water wheel it is not merely a question of one enantiomer reacting and the other one not. It is, rather, a question of one reacting quickly (kfast) and the other reacting slowly (kslow). We will see much more of something called the ‘s value’ later. This is the selectivity factor and is quite simple: s = kfast / kslow or, in other words, the relative rate. Consider two enantiomers of an alcohol being enantioselectively acetylated by some means. One reacts fast and one more slowly. OH

OAc

OH

kfast

(S)-3

OAc kslow

(S)-4

(R)-3

(R)-4

The graph shows several plots which show the % ee of the remaining alcohol 3 against the % level of conversion.2 This graph would be valid for any standard kinetic resolution. The perfect kinetic resolution where kfast / kslow = ∞ obviously gives 100% ee of the remaining alcohol at 50% conversion but it is instructive, and quite amazing, to note that the curve for k fast / kslow = 25 is nearly as good. It is possible to get nearly 100% ee of starting material even when kfast / kslow is as low as two, but the yield would be approaching 0% because the reaction has to be run until there is hardly any starting material left.3

28 Standard Kinetic Resolution Reactions

631

100 s = 25 is almost as good as s = ∞

s=∞

25 8

% ee of remaining alcohol

4

50 s=2

with s = 2 we approach 100% ee of starting material as we approach 0% yield 0 0

50

100

% conversion

Other symbols used in the realm of kinetic resolution include C, the level of conversion where 0 ⬍ C ⬍ 1. In a classical kinetic resolution where one enantiomer reacts faster than the other and the other is left behind then the following equation is true.4 This equation is important as relates three crucial factors—conversion, ee and selectivity.5

S=

In[(1 − C )(1 − ee)] In[(1 − C )(1 + ee)]

ee C = ee ′ 1 − C

The second equation tells us what we already know about kinetic resolutions from the graphical ‘water wheel’ description—in other words, you can’t have it all! A good ee means that you will have to compromise on yield and vice versa. But the better your s value, the less severe those compromises need to be. Ee’ is the enantiomeric excess of the product and ee is the enantiomeric excess of the remaining starting material. Remember that although the ee of starting material will increase as a reaction proceeds, the ee of the product must go down once conversion exceeds 50%.

Standard Kinetic Resolution Reactions Chiral DMAP Let us return to the reaction we mentioned at the very start of the chapter – the reaction of a chiral alcohol to form an ester. DMAP is used to catalyse the acylation of alcohols reactions by nucleophilic catalysis.

632

28 Kinetic Resolution

O O

N

O Me2N

O

6

NEt2 N OH

N

Ph

DMAP achiral

5

8

7

R

N

Chiral version of DMAP

DMAP itself is achiral but a chiral version would make the achiral reactive intermediate 6 chiral. Since the alcohol 7 reacts with this species, there is the possibility that one of the alcohol enantiomers will react more quickly than the other and there will be a kinetic resolution. All that needs is for a chiral version of DMAP to be developed. Because DMAP has two planes of symmetry, this takes some doing. One way that was developed by Spivey et al. was to use an axis of chirality.6 One plane of symmetry is removed because a naphthyl ring is attached on one side and not the other while the chiral axis differentiates the back and front faces of the pyridine ring 8. For the esterification of 9 the s factor was 27 which means good levels of ee in both the starting material and the product should be possible. Indeed 97% ee of the starting material can be achieved at around 56% conversion. OH

O

OH

O

O

O 5

O +

(+)-catalyst 1 mol% –78 ˚C toluene

(R)-9; 43% yield 97% ee

9

56% yield 73.7% ee

s = 27

An important consideration with any catalysed asymmetric reaction is the rate of the background reaction. In other words, the reaction of the isobutyryl anhydride 5 with the alcohol without the catalyst. This reaction will generate racemic material and so clearly it has to be slow (relative to the asymmetric reaction) if asymmetric catalysis is to be effective. The same alcohol 9 was kinetically resolved by Heathcock using porcine pancreatic lipase.7 There OH

O Cl3C

OH

O

O

O

+ porcine pancreatic lipase, heptane

9

10

(S)-9; 90-96% ee

is a striking contrast between the conditions used in the two reactions. The enzymatic reaction is conducted at 60 ⬚C and takes several days! The reactive enantiomer is esterified. Heathcock reveals that the S alcohol that remains at the end of the reaction contains 2–5% of the R enantiomer. This is upgraded by two recrystallisations to ⬎99.5% ee. Another chiral DMAP 11 has been developed by Fu.8 This contains neither a chiral centre nor an axis of chirality and is probably best described as having planar chirality. The s factor was improved from 14 to a very useful 43 simply by changing the solvent and reaction temperature The references contain useful sample experimental procedures.8, 9 Several alcohols including 12–15 have been resolved with this but in looking at the ees, remember to note the conversion necessary to achieve them. Clearly the acetylene 15 was more difficult to resolve.9

28 Standard Kinetic Resolution Reactions

633

With Ac2O and Et3N

N

OH N R Fe R

R 11

R

R

OH

OH

1% catalyst 0 ˚C t-amylOH

s = 14

s = 43 OH

OH

Cl 12 99% ee at 55% conversion

2% catalyst room temp. Et2O

Ph

Ph

15 95% ee at 86% conversion

14 96% ee at 58% conversion

13 98% ee at 56% conversion

While we are on the subject of chiral DMAP it is worth noting that even phosphines10 can be used to catalyse acylation reactions and some of the s values that can be achieved are quite spectacular. The phosphines are bicyclic compounds 16 and 18. The nature of the phosphine is very important and the focus of the work was not on enantiomeric excess but rather the all-important s value. The synthesis of 16 starts using a lactate derivative and makes cunning use of a lithium-binding chain to achieve the selectivity required in producing intermediate 17.11 The even better catalyst 18 was prepared from the same key intermediate 17. Me Me

Me Me

Me Me

H O

H

H H

P

H

Ph

Me

SO2

Me

H

P

Me

O 16

But

17

18

But

The reaction catalysed by 16 was an esterification with benzoic anhydride and the best kinetic resolution obtained (s = 67) with benzylic alcohol 19 was at – 40 ⬚C. A value of 67 is very good for a kinetic resolution. However, exceptional results were obtained using phosphine 18. This catalyst is effective with a variety of alcohols and conditions. The best result was obtained with 20, Ac2O and, crucially, catalyst which had been recrystallised to ⬎99.9% ee. In this instance s = 390! Me

OH

OH N

Me But

Ph

N Me

Me 19

20

Me 21

Vedejs points out that with these high s values, the importance of the enantiomeric excess of the catalyst becomes very important.12 For example, an s value of 145 obtained with catalyst with an ee of 99.6% is expected to improve to s = 204 with catalyst of 100% ee. With s values lower than 40 the effect is much less significant. The Fu and Vedejs catalysts are excellent and work with a range of substrates but good kinetic resolutions can be achieved even with some very simple chiral diamines. A catalyst derived from proline 21 was used at low temperature and low loading to resolve the secondary alcohol 22 very effectively.13 The use of the proline-derived catalysts has been extended to some primary alcohols.14

634

28 Kinetic Resolution

OH

OBz

OH

0.3 mol% 21

+

BzCl

+

Et3N molecular sieves – 78 ˚C

Ph (±)-22

Ph

Ph

49%, 96% ee

s = 160

48%, 95% ee

Epoxidation reactions Perhaps the most famous and widely used kinetic resolution is the Sharpless epoxidation of allylic alcohols5, 15 – a reaction we encountered16 in Chapter 25. This is where matters become marginally more complicated than the simplest kind of kinetic resolution because we are introducing more chiral centres to the molecules in addition to exploiting the ones that are already there. The tartrate complex is the same as the one we have previously encountered in chapter 25. Two enantiomers of allylic alcohol can form two diastereomeric reactive complexes with the single enantiomer of titanium complex. This is illustrated below with, in this instance, the chiral centre at the carbinol carbon of the allylic alcohol. The general allylic alcohol 23 is arranged with the double bond reaching round to the hydroperoxide ready for reaction. In one of the diastereomeric complexes the R group clashes sterically with the titanium complex. In the other complex the R group is able to occupy uncongested space. Not surprisingly, the enantiomer of allylic alcohol 23 that must react via the congested complex is the slow-reacting enantiomer. RS EtO2C

O

O

R1

O Ti

CO2Et O

(±)-23

R R1

RR

Ti

OH

R2

O

i-PrO i-PrO

R3

R HO

EtO2C

R3

O

R1

Fast-reacting enantiomer (S)-23

R1

Slow-reacting enantiomer (R)-23

R2

O O

HO

R3

R3

t-Bu R2 EtO the active Ti complex for asymmetric epoxidation

R R2

A compound which is nicely illustrative of this reaction (and often cited) is allylic alcohol17 24. With L-(⫹)-diisopropyl tartrate this alcohol has a fast-reacting (S) and a slow-reacting (R) enantiomer but what is not addressed in the above diagram is that, in principle, both enantiomers could give two different diastereomers 25 and 26. O

OH

O

OH

OH

kfast

(S)-24

98:2

(S)-25 O OH

OH

(S)-26 O

OH

kslow

(R)-24

(R)-25

38:62

(R)-26

28 Standard Kinetic Resolution Reactions

635

With the fast-reacting enantiomer there is good diasterO OH eoselection in this reaction too and the main product is 25 in O OH a 98:2 ratio. The slow-reacting enantiomer is less selective giving a 38:62 ratio. It is not uncommon for allylic alcohols in this context to be drawn in a different orientation – 25a is merely an alternative drawing of 25b. How much faster the fast reaction is depends on the nature of the 25a 25b catalyst. The selectivity increases from dimethyl- to di18 ethyl- to diisopropyl tartrate from 19 to 36 to 104. Before the dihydroxylation reaction burst onto the asymmetric scene, asymmetric epoxidation and associated kinetic resolutions were possibly the most popular methods of producing single enantiomers simply because they worked so well. The epoxidation could, for example, be used reliably in undergraduate laboratories. We will see Sharpless epoxidation reactions in the Double Methods section towards the end of the chapter. Interestingly, Sharpless’ other famous asymmetric method – dihydroxylation – has not found widespread use in kinetic resolution. This is probably because the AD is just too powerful or, to be anthropomorphic, too wilful. In other words, it is not sensitive to the chirality of the substrate and charges ahead and reacts with both enantiomers. That is not to say there are not examples of kinetic resolution with dihydroxylation,19 but they are more rare. However, the dihydroxylation is even more useful and much more general than the kinetic resolution of allylic alcohols by asymmetric epoxidation and was discussed in Chapter 25. A slightly complicated case of kinetic resolution of alcohols by asymmetric dihydroxylation is in the Double Methods section. Epoxides are a familiar sight in the world of kinetic resolutions. As well as being made by kinetic resolution, racemic epoxides can themselves be the substrates in a kinetic resolution. For example, the use of cobalt-salen complexes – something we shall see again in the dynamic kinetic resolution section – can be used to mediate the formation of enantiomerically pure oxazolidinones.20 Enzymes can be used to react with one enantiomer of epoxide.21 And enzymes are a good way to kinetically resolve compounds and can work under surprising conditions – supercritical CO2 for example.22

Unwanted kinetic resolutions Something to watch out for is unwanted kinetic resolutions. An example is found in the determination of enantiomeric excess by making derivatives like Mosher’s esters using an optically pure reagent. Consider the following. An alcohol 27 has been made in 60% ee – four parts S alcohol to one part R alcohol – but the ee has not yet been determined. Suppose that a reaction to make Mosher’s esters is done to 80% completion and that the R alcohol reacts very much faster than the S. 100% completion If R R or S If S reacts much reacts much reacts faster faster faster 80% completion

R

OH

1 part Mosher's esters

Ar

O 4 parts

S

Cl

OH

CF3

O

O

1

0

1

3

4

4

50% ee

100% ee

60% ee

Ph OMe

CF3

Ar Ph OMe

O

Ar O 27

CF3 Ph OMe

Ar Reality = 60% ee

implied ee

636

28 Kinetic Resolution

The result would be three parts S alcohol (now incorporated into the ester) to one part R. The implied ee is 50%. Perhaps worse is the case if the S enantiomer reacts very much faster. We would have four parts S alcohol to no parts R and the implied ee would be 100%! It is crucial that the reaction is taken to 100% completion so that every last bit of the starting alcohols show up in their derivatives. That way the 4:1 ratio is maintained and the real ee uncovered.

Enzymes The whole of the next chapter (Chapter 29) looks at enzymes and their application. Nevertheless, it is worth examining a simple case here in the context of kinetic resolution. The selectivity of enzymes is often so good that the reaction can be run to 50% completion so that the enantiomeric excess of both the product and the unreacted starting material is excellent. Lipase from Pseudomonas fluorescens will hydrolyse one of the enantiomers of the acetate 4 very effectively in water at OAc

OH

OAc

Pseudomonas fluorescens lipase

+

H2O, pH 7

(R)-3; 48% yield >99% ee

(±)-4

(S)-4; 48% yield >99% ee

neutral pH. Hence the (R)-enantiomer of alcohol 3 is produced in 48% yield and ⬎99% ee while the unreacted enantiomer is left behind in 48% yield and ⬎99% ee too! It is the (R)-enantiomer that reacts. So if we started with racemic alcohol (instead of racemic OH

OAc

OH

Pseudomonas fluorescens lipase

+ OAc

(±)-3

t-BuOMe

(R)-4; 45% yield >99% ee

28

(S)-3; 41% yield >93% ee

acetate) and ran the reverse reaction using the same enzyme, then the molecule that would be left behind would be the (S)-enantiomer of alcohol 3. This does indeed work.23 Interestingly, we can thus make either enantiomer of alcohol using the same enzyme. However, the yields and ees show us that the reverse reaction is not quite so good. The use of vinyl acetate as the acetylating agent is a common trick and cunning because the displaced vinyl alcohol 29 does not hang about but tautomerises into acetaldehyde 30 – thus removing it from the equilibrium and helping to drive the reaction forwards. The one drawback is that the increasing concentration of acetaldehyde is detrimental to the reaction (presumably by reducing the activity of the enzyme). Pseudomonas is found in the section below on Dynamic Kinetic Resolutions. We will meet Pseudomonas fluorescens again in the next chapter where we also see enzymes in kinetic resolutions with racemisation of starting material (dynamic kinetic resolution). O

O

29 +

O

+ R

OH

OH

28

H 30 O

R

O

28 Dynamic Kinetic Resolutions

637

Dynamic Kinetic Resolutions The biggest problem with a kinetic resolution is that, even with the best selectivity and thus product with 100% ee, your yield is limited to 50%. The unwanted enantiomer goes to waste. A dynamic kinetic resolution (DKR) gives us the opportunity to raise this theoretical maximum to 100%. In the simplest DKR, the starting material is racemised under the reaction conditions but only one of the enantiomers then goes on to react. The continual racemisation of starting material ensures that the unwanted enantiomer never builds up. This racemisation has an added bonus when it comes to selectivity. With a standard kinetic resolution, in becomes increasingly tricky for the reagent or catalyst to be selective for the desired enantiomer as we approach 50% completion because there is so much of the unreacted, undesired material competing for its attention. With a DKR, since there is no such build up, it is less difficult for a reagent to be selective at the end of the reaction. In Chapter 22 we saw resolution with racemisation applied to the synthesis of L-364,718. That could have been described as a dynamic resolution or possibly even a dynamic thermodynamic resolution but not a dynamic kinetic resolution as it did not depend on one enantiomer reacting faster than the other. Rather it depended on the thermodynamic stability of one crystalline form over the other.

Reaction of epichlorhydrin Racemic epichlorhydrin 32 can be dynamically kinetically resolved using a salen chromium complex 31. The enantiomeric excess of the product 33 is excellent at 97% and the yield 76%.24

H

H N

(S, S)

O

0.5 eq TMSN 3

N N3 O

Bu-t

2. 0.5 eq TMSN 3 slow addition

32

Cr t-Bu

1. 2 mol% 31

O

Cl

Bu-t

OSiMe3 Cl

t-Bu

N3

33; 76% yield, 97% ee

31

On first inspection of the reaction scheme it looks as though the epoxide is opened with azide and the process mediated by the chromium complex. But this would be a standard KR and give us a maximum yield of 50% – something else must be going on to interchange the two enantiomers of epichlorhydrin. Imagine epichlorhydrin opened by a chloride ion. We would have a species with a plane of symmetry 34 and formation of the epoxide on the other side of the molecule would transform one enantiomer into another. O

Cl (S)-32

OH Cl

Cl

34 plane of symmetry

O

Cl (R)-32

Something akin to this happens in the reaction above. In the normal course of events the azide complex 31 reacts with epichlorhydrin to give an intermediate which reacts with TMSN3 to regenerate the reagent and silylate the oxygen. But there is a side reaction which generates the chloride salen complex 35.

638

28 Kinetic Resolution

Cr O

Cl

Me3SiN3

O Cl

Cr

OSiMe3 Cl

N3

+

N3

N3 33

(S)-32

31

side reaction Cr 35

Cl

It is this complex which can add to epichlorhydrin and, in a reverse reaction that may happen from either side, racemise it 36. Note that under the reaction conditions the azide half of TMSN3 adds quickly but the Me3Si half slowly – the reactive enantiomer of epichlorhydrin is there to be had at the start but must be made by continuous racemisation in the second part of the reaction. A potential by-product from these reactions is the double azide 37 which the authors think looks suspiciously explosive. Cr

Cl

Cr

Cl

O

OSiMe3

O

36

Cl

35

N3

N3

37

Cl

CAUTION!

The azide 33 is readily transformed into the convenient reagent 38 which has been incorporated into a final intermediate 38 in the synthesis of the oxazolidinone antibiotic 39.24 1. MeOH, cat. TFA 2. H 2, PtO 2

OSiMe3 Cl

N3

K2CO3 MeOH

OAc Cl

3. Ac 2O, Et 3N

33

O

NHAc

NHAc 38

98% yield CbzN

CbzN N

F

N

O N H

O

1. BuLi, –78 ˚C

O

Ph

2. 2 equivs. 38 0 to 60 ˚C

F

N 39

O NHAc

α-Acetoxysulfide A synthesis of lamivudine in optically pure form requires the use of α-acetoxysulfide (R)-40.25 This can be produced in optically pure form by a kinetic resolution using the lipase from Pseudomonas fluorescens. This enzyme hydrolyses the enantiomer that we do not want and leaves behind the one that we do. The authors investigated a range of sulfide side chains and the acetal 40 that is used in the synthesis of lamivudine worked perfectly.26 The solvent is important. The use of chloroform instead of t-BuOMe gives low ees in the opposite sense. O

S HO O lamivudine

N

N NH2

28 Dynamic Kinetic Resolutions

OEt MeO2C

S

OEt

100 mg substrate 2 mg lipase 2 ml buffer 0.5 ml t-BuOMe

OAc (±)-40

639

OEt MeO2C

S

OEt

OEt

MeO2C

S

OAc (R)-40; 45% yield, >95% ee

30 ˚C, 2 hr

OEt

+ OH (S)-41

This is the result of kinetic resolutions that we have come to expect and although the selectivities are excellent (ee ⬎ 95%) the yield can never be more than 50%. In order to get more than 50% we would need to recycle the unreactive enantiomer.

R

H

R'SH

R

SR'

R

lipase

SR'

R

Ac2O DMAP

SR'

+ 43

O

OH

42

OAc

(±)-44

R

SR'

(±)-44

OAc

OH

(S)-45

(R)-44

py

OAc (±)-45

Racemic α-acetoxysulfides 45 can be made by the addition of a thiol 43 to an aldehyde 42 followed by acetylation of the resulting hydroxyl group with Ac2O.27 The idea that the lipase from Pseudomonas fluorescens might be used to mediate the acetylation instead, using vinyl acetate as the stoichiometric reagent, to give (S)-45 works. But we have still not achieved a dynamic kinetic resolution. We would need to return the unacetylated material (R)-44 to starting materials for this to happen. However, it was found that hemithioacetals decomposed during column chromatography on silica to give the starting thiol 43 and aldehyde 42. This led the workers to the idea of adding silica to the reaction mixture to promote reformation of the starting aldehyde and thiol. It worked.27 In the presence of silica gel the yield is substantially greater than 50% because the wrong enantiomer of the intermediate hemithioacetal can return to starting materials.

O

O

OAc

H

MeO

O methyl glyoxalate

HS

OSiEt3

50 mg methyl glyoxalate t-BuOMe 50 mg lipase 10 mg silica gel

46

MeO

S

OSiEt3

OAc (S)-47; 83% yield, 90% ee

Several thiols were looked at with 46 being one of several that gave high ee of product. Overall the enantioselectivities for the acetylation reactions are slightly higher than those for the hydrolysis reactions which the authors suggest is due to greater conformational rigidity of the enzyme in organic rather than in aqueous solvents. You may have noticed that since this reaction is concerned with the esterification of the (S)-enantiomer rather than the hydrolysis of the (S)-enantiomer we generate the (S)-acetate 47 and sadly not the one that is actually needed for the synthesis of lamivudine! The authors went on to use Pseudomonas fluorescens for the hydrolysis of diacetate 48 (a standard kinetic resolution reaction) to give a key intermediate of the correct enantiomer (R)-48. In this case a primary acetate is hydrolysed.28

AcO

S OAc (±)-48

OSiEt3

AcO

S OAc (R)-48

OSiEt3 + HO

S OAc

OSiEt3

640

28 Kinetic Resolution

Enzymes and metals together OH

(S)-3 racemise?

OH

(±)-3

In the above example, an enzyme for the kinetic resolution was combined with silica to provide the dynamic part of a dynamic kinetic resolution. The role of the silica is to racemise the hemithioacetal that does not get acetylated. The racemisation of most alcohols is unlikely to be so straightforward. Consider sec-phenethylalcohol 3; how would we racemise this? Silica is unlikely to do the job. Williams was in the unusual position for a worker in asymmetric synthesis of trying to find a catalyst that was good at racemisation.29 An ingenious solution is to use the transfer hydrogenation of acetophenone. If (S)-alcohol can be the hydrogen source for the reduction of acetophenone then, so long as there is no enantioselection of any kind, the products will be regenerated acetophenone (from the optically pure alcohol) and racemic alcohol from the achiral acetophenone. Numerous catalysts were investigated including those based on iridium and ruthenium. Two results with aluminium and rhodium are shown below. OH

OH

O Ph catalyst

(S)-(–)-3

catalyst

yield, ee

20 mol% Al(OPr-i)3

80% 0%ee

3 mol% Rh2(OAc) 4 + KOH + phenanthroline

73% 0%ee

(±)-3

The racemisation process was then combined with enzymatic kinetic resolution. The overall picture can be seen below.29 Racemisation is going on in the box and the lipase from Pseudomonas converts the (R)-enantiomer of alcohol into its acetate. At 60% conversion the product has 98% ee. The 60% may not seem all that impressive but in a standard kinetic resolution, the maximum ee that is possible at 60% conversion (product therefore contains 50% of one enantiomer and 10% of the other) is 67%. Another result gave 76% ee at 80% conversion. The maximum ee of product at this conversion with a kinetic resolution is a mere 32%! OH Ph

60% conversion 98% ee (theoretical maximum ee at 60% conversion is 67%)

OAc

Pseudomonas fluorescens

Ph

lipase OAc

O Ph

O Ph

2 mol% Rh2(OAc) 4 phenanthroline N

N

OH

phenanthroline Ph

Lipases have also been combined with palladium catalysts to provide the dynamic aspect of the kinetic resolution.30 In the example below the unreacted allyl acetate (S)-49 is racemised. The formation and reactions of Pd-allyl cations are discussed in chapter 18. OAc

Pseudomonas cepacia lipase

OH

OAc +

i-PrOH

(±)-49

72% yield, 98% ee Pd(0); Pd(PPh 3)4 + Pd(dppf)2

(S)-49

28 Parallel Kinetic Resolutions

641

Hydrogenation Something that we have not addressed is the rate at which racemisation occurs relative to the rate of the kinetic resolution. Clearly if the rate of racemisation is too slow, there will be a build up of the less reactive enantiomer. For the best selectivity the starting materials should remain racemic which means that the rate of racemisation must be faster than the fast kinetic resolution reaction. This is the case in the hydrogenation of β-ketoester 50.31 Racemisation is fast and occurs via the enol 51. The product has 99% ee. Most of the impurity is the wrong diastereomer but nevertheless the syn:anti is a healthy 96.4:3.6. The ligand on the ruthenium is a slightly modified version of BINAP. The phenyl rings of ordinary BINAP are replaced with xylyl rings.

O

O

H2 (R)-xylBINAP-Ru

OH

OH

O

O OMe

OMe

OMe kS

OH

NHCOPh 99% ee 96.4:3.6 syn:anti

NHCOPh (S)-50

O OMe

kinv

NHCOPh PAr 2

kinv > kS

PAr 2

NHCOPh 51

O

O

OH OMe

O

OH

O

xylBINAP-Ru

OMe

OMe

Ar = 3,5-Me2C6H3 (R)-3,5-xylylBINAP

kR

NHCOPh (R)-50

NHCOPh

NHCOPh

Parallel Kinetic Resolutions Kinetic resolutions are a bit like a big tin of assorted sweets at Christmas with an aunt who likes to eat the blue triangles. Initially, she has an easy time finding the blue triangles and eating them up—delicious. But as time goes on, those blue triangles become more scarce and thus harder to find amongst the red squares etc. Similarly, a reagent selecting an enantiomer from a racemic mixture finds it hard and harder to find the enantiomer it wants. Really what we need is an uncle who eats those red squares. Alternatively, think back to the image at the start of the chapter with the black and white balls for another analogy. Remember that the wheel is selective for white balls. Towards the end of the reaction, there are hardly any white balls left and the wheel, which doesn’t have perfect selectivity anyway, will find it increasingly hard to pick them out. And if all the white balls have gone, then the wheel, if it continues reacting, will be forced to react with the black balls. This is all bad news for selectivity. If there were another water wheel which was selective for black balls (but delivered them to a separate container) then the concentration of black balls would not increase. With a parallel kinetic resolution, one enantiomer reacts to give one product while the other reacts to give something else. The main advantage is that the increasingly steep uphill struggle is removed since the concentration of the desired enantiomer does not decrease relative to the one that is not wanted. The one that is not wanted is being reacted too – but to give something else. Ideally they should both react at a similar rate. One way to make sure that two enantiomers of starting material react at an equal rate is to use a racemic reagent. But, of course, this is totally useless as we shall just get a racemic

642

28 Kinetic Resolution

product at the end! However, with a reagent that is nearly racemic, as it were, we can achieve parallel kinetic resolution. What we mean by ‘nearly racemic’ should become clear and such a method was illustrated by Vedejs.32 The reagent is a chiral form of DMAP which has already been acylated. Note that the reagent is, in fact two optically pure reagents that are almost enantiomers of one another. The idea is that one reagent will react with one enantiomer of material and the other with the other and that the products will be separable. Each pseudo-enantiomer of reagent (53 and 56) delivers a different acyl group so that the ‘enantiomers’ are different compounds and can be distinguished. Additionally, the trichlorobutyl protecting group can be removed with zinc and acetic acid to give the alcohol (S)-9 and leave the fenchyl carbonate 57 intact. NMe2

NMe2

O O

Bu-t

N

OMe 52

N Cl3C

O

O

CCl3

Bu-t

O

OMe

OH

54; 49% yield, 86% ee

53 NMe2

9

O

NMe2 MgBr2 / Et3N

O N

Bu-t

N

OBn 55

O

O 56

O

Bu-t OBn 57; 43% yield, 93% ee

The main advantage of parallel kinetic resolution is that the selectivity value, s, need not be nearly so high to get good results. With a standard kinetic resolution, we would need s = 200 in order to obtain both 50% of starting material and product with 96% ee. A parallel kinetic resolution needs s = 49 to obtain the same results (there are two s values in a parallel kinetic resolution—one for each reaction, we assume here that they are equal). However, we can see that it is quite a business designing a parallel kinetic resolution and for many applications it simply would not be worth the bother if the s value is good. With a parallel kinetic resolution, the two reactions should not interfere with one another. In the example above, both reactions were acylations but the use of stoichiometric reagents meant that the right acylating group was attached to the right enantiomer. Even here, Vedejs does not rule out a small amount of leakage between the paths (once the chiral DMAP is liberated it can get in on the act with the other pathway). A PKR becomes more difficult if, in addition to the reactions being related, they are also catalytic. In the example below, one enantiomer of compound is acylated by a lipase and the other by a phosphine-catalysed process.10, 33 We want the lipase to catalyse the reaction between vinyl pivalate 58 and the R-alcohol, and the phosphine to catalyse the acylation of the S-alcohol using the polymer-bound anhydride 60. But, the lipase must not use the polymer-bound anhydride or the wrong enantiomer of alcohol will end up polymer-bound. Similarly, the phosphine must not react with the vinyl pivalate. Fortunately, the phosphine does not react with the vinyl pivalate and the lipase, which is insoluble, cannot interact with the polymer bound material. Also, the lipase will be kept away from the potentially damaging P-acylphosphonium intermediates. The conditions are described as a ‘Three-Phase System’. The three phases are the solution phase and the two separate insoluble phases of the polymer bound anhydride and the cross-linked lipase.

28 Parallel Kinetic Resolutions

643

O O O 58

OH

O

Bu-t

Bu-t

59 96% ee

Lipase

O polymer O s = 34

Phosphine

9 O

61 92% ee

O polymer O

O 60

The s-values for the lipase route and the phosphine route were not the same, but this is less important than the rates of enantiomer consumption being equal. Hence the relative amounts of lipase and phosphine were adjusted so that the rates of enantiomer consumption were equal. At the end of the reaction, one enantiomer of alcohol is bound to the polymer and the other has formed an ester in solution. They may be separated by filtration. Both reactions have products of higher ee’s than predicted for the simple kinetic resolution reactions at 50% conversion. Note that really quite different reagents were used to react with the two enantiomers. If you have got the idea by now that parallel kinetic resolutions are a complicated business then the next example will show just how simply they can be achieved. Eames set about using different (pseudo-enantiomeric) Evans’ auxiliaries to resolve an ester.34 However, the ‘auxiliaries’ are nothing of the sort in this reaction. On each side of the reaction, an enantiomerically pure oxazolidinone is reacted with a racemic ester 63 which is to be resolved. A crucial foundation to the work was to establish which chiral oxazolidinones show good selectivity. This was done with racemic oxazolidinones and racemic substrate. Using racemic oxazolidinone exposes the diastereoselectivity of the reaction without the complication of the concentration of the substrate enantiomers varying as the reaction proceeds. It also mimics the environment of the parallel kinetic resolution itself. Hence it was found that while valine-derived oxazolidinone 62 and phenylglycine-derived 67 gave good selectivity (for the syn diastereomer) norephedine-derived 66 did not. O HN

O

O 1. BuLi

O

2.

Ph

N

O

O

Ph

+

Me

Ph

O

O N

O

Me

OC6F5 Me

(±)-62

(±)-64 - anti

63

(±)-65 - syn

5 : 95

For the parallel kinetic resolution itself, (S)-62 and (R)-67 were used together with racemic ester 63 in an efficient and readily achieved kinetic resolution. Despite the similar appearance of 64 and 68 they could be separated. It is worth pointing out that the major syn products are complimentary to the anti products to be expected from a standard alkylation of an oxazolidinone like 69. O

O

HN

O

Me

Ph (±)-66

HN

O

O Ph

N

O

Ph (±)-67

Ph 69

O

644

28 Kinetic Resolution

O HN

O

O 1. BuLi

O

2.

Ph

N

O

Ph

+

O

N

Me

Ph

O

O

O

Me

OC6F5 Me

(±)-62

(±)-64 - anti

63

(±)-65 - syn

5 : 95

Regiodivergent resolutions Initially this sort of thing looks like one of those academic fancies that will not actually be of any use to a medicinal chemist working at the coal face of pharmaceutical production. But don’t give up just yet, later we come onto applications to regioselectivity in steroid chemistry. In this next example, one enantiomer of substrate reacts with the optically pure base in one way and the other enantiomer in another – hence the ‘divergence’. Before we get started with the reaction, the substrate needs careful stereochemical analysis.

O

O

OSiMe3

Me3SiCl

t-Bu

Ph H

t-Bu 70

70a

H

N

Ph

t-Bu

Li 71

72

With achiral ketone 70, an optically pure base 71 chooses between one enantiotopic proton and another to make the enolate enantioselectively. The choosing goes on within the compound – there is no enantiomer of the starting material. However, things are a little more complicated when we move to a racemic substrate. The protons on either side of the carbonyl of ketone 73 are not enantiotopic but diastereotopic. The enantiomeric proton can be found in the other enantiomer of compound. We might imagine, therefore, in a straightforward situation, that a chiral base would deprotonate one enantiomer of starting material and ignore the other completely. Thus we would have enantioselection and all would be well. With ketone 73 it is not quite so simple. H

H

O

O =

BocN H (R,R)-73

H O

H

Locally superimposable parts of the two enantiomers in the frames. =

BocN

H

H (R,R)-73 H

H O

BocN

BocN H (S,S)-73

H (S,S)-73

H

We know that enantiomers cannot be superimposed. However, if we compare (R,R)-73 and its enantiomer, we can see that a crucial part of each molecule can be superimposed. This local congruence in the molecules can be exploited. It is this region that is attractive to the chiral base and it corresponds to different parts in each enantiomer. Thus we can expect the regioselection of one enantiomer to be different from that in the other.

28 Regiodivergent resolutions

645

When a single enantiomer of (R,R)-73 was reacted with base 71, a 94:6 ratio of the regioisomeric enol ethers 74 was generated. When the other enantiomer (S,S)-73 was reacted with the same base 71 then the preference for the regioisomers was the other way round at 21:79. These numbers are not 6:94 as well because the molecule also has an inherent preference for which position it would like to be deprotonated. For the sake of simplicity we shall ignore this modest match/mismatch issue and move on. So, the question then is what would we expect to see if we had started with racemic 73. Regioisomer 74 would be generated in 94 parts from one enantiomer and in 21 parts from the other. We would therefore expect the ee to be 63%. Meanwhile, the other regioisomer 75 is generated in 79 parts from one enantiomer 75b and only 6 parts from the other 75a so we can expect an ee of 86%. The ees determined experimentally when racemic 73 was used were 60% and 83%. The relative amounts of 74 and 75 we can expect would be (94⫹21): (6 ⫹79) = 58:42. Experimentally35 it was 55:45. H

O

H

O

Predicted Results for racemate

BocN H H (R,R)-73 Me3SiCl –90 ˚C

Ph

H

N Li

H

NBoc

H

H

H (S,S)-73

Me3SiCl –90 ˚C

Ph

OSiMe3

Ph

H 74a H

NBoc

Me3SiO

OSiMe3

H

NBoc 75b

75a

Ph

N Li

Ph

21%

ee =

94 – 21 × 100 = 63% 94 + 21

79%

ee =

79 – 6 × 100 = 86% 79 + 6

74b H H

6%

BocN

Me3SiCl –90 ˚C

Ph

H

Me3SiO

94%

BocN

N Li

H

Broadly, what has been achieved here is one enantiomer reacting to give one product and the other enantiomer reacting to give the regioisomeric product. Hence the term - regiodivergent resolution. This is really another example of a parallel kinetic resolution. The implications are of more general use, of course, than the rather specialist example given here. Consider for instance the reaction of the optically pure steroidal ketone. Deprotonation with LDA and quenching with TMSCl gives a mixture of the two enol ethers 77 and 78 with moderate selectivity. Before we move on to using a chiral base in this situation, it is worth noting that the use of an achiral base, LDA, has uncovered an inherent preference (in a kinetically controlled reaction) of the molecule itself – this was something we ignored with ketone 73 but which was clearly present too.36 Anyway, by making use of an optically pure base, we can either improve the selectivity or overturn the molecules’ preferences. Both of these options are useful. Ph N

N

Me OBu-t

t-Bu

Li

H

base Me3SiCl

H

N Me

H

(R)-76 base 77 : 78 LDA 64 : 36 (R)-76 95 : 5 (S)-76 24 : 76

O

Me OBu-t

H

H steroidal ketone

H

H

H

H

+ H

Me3SiO

Me OBu-t

H 77

H

H Me3SiO

H 78

H

646

28 Kinetic Resolution

Double Methods If a selective process can be applied twice then we might expect the quality of the resulting material to be very high. In the next example, the selective process is applied first to produce mostly a single enantiomer. When applied the second time, it removes the small amount of the wrong enantiomer that was produced the first time round! In Chapter 25 we mentioned that desymmetrisations and kinetic resolutions were brothers. Both are used in each of the examples below.

Desymmetrisation then kinetic resolution The doubly allylic alcohol 79 is reacted under Sharpless epoxidation conditions.37 Selectivity is good because kfast /kslow is high at 104. This first reaction is not a kinetic resolution because we are not resolving anything and, indeed, the starting material is achiral (and prochiral). This first reaction produces mostly enantiomer 80a with a little of enantiomer 80b. With such good selectivity we would expect good ees even after the first round and so we do. At 50% conversion we have a 48% yield of 80a and already excellent ee of 99.4%. The second reaction converts each enantiomer 80a and 80b into the same double epoxide 81. Here is the key to the enhancement process – the double bond that the reagent found more attractive in the starting material 79 (and hence reacted with fast) is the one that remains unreacted in product 80b. Hence we might expect this undesired product 80b to react faster than 80a to give the double epoxide. At 99% conversion, the yield of 80a is 93% and the ee 99.96%. The reference contains plenty of the maths behind this rather complicated process.37

O

OH

fast

slow

O

OH slow

79; achiral

OH

80a OH

kfast / kslow = 104

fast

O 81; achiral

80b

O

If the enhancement of enantiomeric excess did not impress you much (though it should, 99.96% is very difficult to achieve in one go!) perhaps an example where the ee is a bit lower after the first round will make the point more effectively. The bistriflate 82 contains a biaryl bond with restricted rotation so that replacing one of the triflates with something else will give us a chiral molecule (83a and 83b are enantiomers). As above, the first reaction is a desymmetrisation and the second reaction a kinetic resolution upon the products from the first. Both reactions are palladium mediated cross-couplings of an aryl triflate and phenyl Grignard. An optically pure palladium complex 85 catalyses the formation of a biaryl bond and selects between the two enantiotopic triflates. The selectivity is less good than the example above with the fast reaction only 12 times faster than the slow reaction for the first stage. We would expect that the best enantiomeric excess in a product that could be obtained with a rate ratio of 12:1 would be 85% ee [(12⫺1)/(12⫹1) ⫻ 100]. With a slight excess of PhMgBr we observe 39% of product 83 and an ee of 85%.

28 Double Methods

85 PhMgBr

647

TfO

Ph FAST

SLOW

TfO

83a

OTf

rate 12 : 1

FAST

82

Ph

rate 5 : 1

SLOW

85 PhMgBr

N

Ph

Pd N

Ph

bisphenyl 84 Ph

Cl

OTf

Cl 83b

85

84

83 1.1 eq PhMgBr 39% 85% ee

0%

2.1 eq PhMgBr 87% 93% ee

13%

It is worth reiterating something here before we move on to the second reaction. With a kinetic resolution, the ee of a product is limited by the relative rates of the reactions. But the ee of the starting material left behind is not limited in this way. The ee can be increased by letting the reaction run a bit further to completion. There is a price to be paid in yield, of course, and you might find that to get your 98% ee you end up with only 2% yield but it can be done. In the second reaction (the reaction of 83a and 83b), the compound we want 83b is one of the starting materials. The relative rates of reaction are quite poor at 5:1 but we already have a head start in ee anyway. The wrong enantiomer 83a is eaten up in the second reaction With 2.1 eq we end up with an 87% yield of product in 93% ee with 13% of the doubly arylated product 84.38 If we wanted to increase this 93% further, we could run the reaction a bit further still. These double methods look quite esoteric but they are more widespread than one might imagine. They need looking out for because they are sometimes not even noted by the authors of the work. In Fürstner’s synthesis of (⫺)-balanol,39 which we looked at in chapter 25, the chiral centres in the seven membered ring originate in a Sharpless asymmetric epoxidation (see chapter 25). We can see that the starting material 86 is a double allylic alcohol of the kind we saw earlier 79. The enhancement of enantiomeric excess is certainly operating to give a product 87 of excellent ee. After a few manipulations, the opening of the epoxide with allyl amine and a ring closing metathesis, the key ring 88 has all the stereochemistry that it needs. OH

OH

(–)-DET t-BuOOH Ti(OPr-i)4 –20 ˚C

86 OBn

OBn

Cl

OH

94% from 87

NH2 O

BnBr

PCy3 Ph

OBn

OBn

Ru

N

Cl

Boc 100%

8 mol%

N3

OH

PCy3

Boc2O Et3N

O

87

OH

NH

OBn NaH

(PhO)2PON3

N Boc 88%

Ph3P, DEAD

N Boc 88, 91%

In the above examples enantiotopic reactive sites on either side of the compound were reacted before we moved onto more complicated matters. The next example is a variation on that theme

648

28 Kinetic Resolution

but this time we start with enantiotopic faces of a double bond. When triene 89 is reacted with OsO4 and (DHQD)2PYR and allowed to go to a high level of completion then the product can be obtained enantiomerically pure. However, if the same reaction is stopped at low conversion then the enantiomeric excess of the product is only 95%.

OsO4

OH

HO

(DHQD)2-PYR

89

90

The explanation40 is that there is a kinetic resolution going on—the minor enantiomer is being removed as the reaction proceeds. It reacts more quickly under the reaction conditions. We could leave the explanation at that but we shall not gloss over this point because it is worth consideration in a little more depth. In order to understand the nature of this process we need to know about the reactivity of diol 90. Both the chiral centres in diol 90 have the R configuration. We see perhaps the most revealing result when enantiomerically pure diol 90 is dihydroxylated with OsO4 using an achiral ligand – quinuclidine. This is important because it tells us how the substrate likes to react with osmium tetroxide. There is a 3:97 product ratio in favour of the meso compound 92 rather than the C2 symmetric compound 91. In other words, the molecule with two R chiral centres prefers the reaction that leads to the formation of two new chiral centres of opposite configuration – S.

OsO4

R HO

OH

quinuclidine, etc.

HO S

HO R R

R

OH HO

R

OH

S

+

R

OH HO

R

OH R

N

90

91 (C2 isomer)

3 : 97

92 (meso isomer)

So we see how the substrate likes to react but the reagents, of course, may have other ideas. The chiral ligand (DHQ)2PYR prefers to form two S chiral centres from a trans double bond of the kind we have in 90. The diene 90 wants that too so we have the matched case and a high selectivity of 0.3:99.7. OsO4

90

C2

91

0.3 : 99.7

92

meso

C2

91

75 : 25

92

meso

(DHQ)2PYR

90

OsO4 (DHQD)2PYR

However, (DHQD)2PYR, the pseudo-enantiomer of (DHQ) 2PYR (see chapter 25), would prefer to form two R chiral centres. The substrate and reagent are in competition here—the mismatched case. The selectivity is turned over to 75:25 in favour of the C2 symmetric product 91 this time. So the reagent wins but not outright; there is still much of the meso diastereomer formed. Now we can return to the original question which concerns the reaction of triene 89.

28 References

Major enantiomer

R HO

649

OH R 90

OsO4

89 (DHQD)2-PYR

HO R S HO

Minor enantiomer

OH S ent–90

reacts quickly

R

OH HO

S

OH S 92a

Triene 89 reacts to form the diol 90. A small amount of the enantiomeric diol, ent-90, is also formed. The crucial point is that this minor product is matched with the reagent and is reacted away, and quickly, to the tetraol 92a (92a = 92 but is drawn in a different way). The major product 91 is mismatched with the reagent and reacts more slowly. The first dihydroxylation (of 89) is the fastest of all three.

References General references are given on page 893 1. J. M. Keith, J. F. Larrow and E. N. Jacobsen, Adv. Synth. Catal., 2001, 343, 5-26. 2. This graph was produced with the kind assistance of Prof. Guy Lloyd-Jones 3. G. Balavoine, A. Moradpour and H. B. Kagan, J. Am. Chem. Soc., 1974, 96, 5152. 4. H. B. Kagan and J. C. Fiaud, Top. Stereochem., 1988, 18, 249 5. V. S. Martin, S. S. Woodard, T. Katsuki, Y. Yamada, M. Ikeda and K. B. Sharpless, J. Am. Chem. Soc., 1981, 103, 6237. 6. A. C. Spivey, T. Fekner and S. E. Spey, J. Org. Chem., 2000, 65, 3154. 7. P. D. Theisen and C. H. Heathcock, J. Org. Chem., 1988, 53, 2374. 8. J. C. Ruble, J. Tweddell and G. C. Fu, J. Org. Chem., 1998, 63, 2794. 9. B. Tao, J. C. Ruble, D. A. Hoic and G. C. Fu, J. Am. Chem. Soc., 1999, 121, 5091. 10. E. Vedejs, O. Daugulis, J. A. MacKay and E. Rozners, Synlett, 2001, 1499. 11. E. Vedejs and O. Daugulis, J. Am. Chem. Soc., 2003, 125, 4166. 12. R. F. Ismagilov, J. Org. Chem., 1998, 63, 3772. 13. T. Sano, K. Imai, K. Ohashi and T. Oriyama, Chem. Lett., 1999, 28, 265. 14. D. Terakado, H. Koutaka and T. Oriyama, Tetrahedron: Asymmetry, 2005, 16, 1157. 15. Y. Gao, R. M. Hanson, J. M. Klunder, S. Y. Ko and K. B. Sharpless, J. Am. Chem. Soc., 1987, 109, 5765. 16. T. Katsuki and K. B. Sharpless, J. Am. Chem. Soc., 1980, 102, 5974. 17. M. G. Finn and K. B. Sharpless, Eds., On the Mechanism of Asymmetric Epoxidation with TitaniumTartrate Catalysts, vol. 5 (Academic Press, Orlando, 1985). 18. B. E. Rossiter, Ed., Synthetic Aspects and Applications of Asymmetric Epoxidations, vol. 5 (Academic Press, Orlando, 1985). 19. M. S. VanNieuwenhze and K. B. Sharpless, J. Am. Chem. Soc., 1993, 115, 7864. 20. G. Bartoli, M. Bosco, A. Carlone, M. Locatelli, P. Melchiorre and L. Sambri, Org. Lett., 2005, 7, 1983. 21. I. V. J. Archer, Tetrahedron, 1997, 53, 15617. 22. T. Matsuda, T. Harada, K. Nakamura and T. Ikariya, Tetrahedron: Asymmetry, 2005, 16, 909. 23. S. M. Roberts, Ed., Biocatalysts for Fine Chemicals Synthesis (John Wiley & Sons, Chichester, 1999).

650

28 Kinetic Resolution

24. S. E. Schaus and E. N. Jacobsen, Tetrahedron Lett., 1996, 37, 7937. 25. M. F. Jones and C. M. Rayner, Tetrahedron Lett., 1995, 36, 6961. 26. J. Milton, S. Brand, M. F. Jones and C. M. Rayner, Tetrahedron: Asymmetry, 1995, 6, 1903. 27. S. Brand, M. F. Jones and C. M. Rayner, Tetrahedron Lett., 1995, 36, 8493. 28. S. Brand, M. F. Jones and C. M. Rayner, Tetrahedron Lett., 1997, 38, 3595. 29. P. M. Dinh, J. A. Howard, A. R. Hudnott and J. M. J. Williams, Tetrahedron Lett., 1996, 37, 7623. 30. Y. K. Choi, J. H. Suh, D. Lee, I. T. Lim, J. Y. Jung and M.-J. Kim, J. Org. Chem., 1999, 64, 8423–8. 31. M. Kitamura, M. Tokunaga and R. Noyori, J. Am. Chem. Soc., 1993, 115, 144. 32. E. Vedejs and X. Chen, J. Am. Chem. Soc., 1997, 119, 2584. 33. E. Vedejs and E. Rozners, J. Am. Chem. Soc., 2001, 123, 2428. 34. G. S. Coumbarides, M. Dingjan, J. Eames, A. Flinn, J. Northen and Y. Yohannes, Tetrahedron Lett., 2005, 46, 2897. 35. K. Bambridge, B. P. Clark and N. S. Simpkins, J. Chem. Soc., Perkin Trans. 1, 1995, 2535. 36. M. Sobukawa, M. Nakajima and K. Koga, Tetrahedron: Asymmetry, 1990, 1, 295. 37. S. L. Schreiber, T. S. Schreiber and D. B. Smith, J. Am. Chem. Soc, 1987, 109, 1525. 38. T. Hayashi, S. Niizuma, T. Kamikawa, N. Suzuki and Y. Uozumi, J. Am. Chem. Soc., 1995, 117, 9101. 39. A. Fürstner and O. R. Thiel, J. Org. Chem., 2000, 65, 1738. 40. H. Becker, M. A. Soler and K. B. Sharpless, Tetrahedron, 1995, 51, 1345.

29

Enzymes: Biological Methods in Asymmetric Synthesis Preliminary note on biological methods Introduction: Enzymes and Organisms Good and bad features of enzymes as catalysts Organisms: Reduction of Ketones by Baker’s Yeast Ester Formation and Hydrolysis by Lipases and Esterases Kinetic resolution with racemisation Enzymes versus whole organisms Desymmetrisation with lipases Immobilised enzymes in desymmetrisation Polymer-supported reagents and enzymes Effects of amines on lipases and esterases Other acylating enzymes Enzymatic Oxidation Asymmetric dihydroxylation of benzenes Epoxidation The Baeyer-Villiger reaction Baeyer-Villiger reactions with engineered baker’s yeast Nucleophilic Addition to Carbonyl Groups Asymmetric addition of cyanide to aldehydes Products derived from cyanohydrins Enzymatically catalysed aldol reactions Synthesis of syringolide using an enzyme-catalysed aldol reaction Engineered aldolase Practical Asymmetric Synthesis with Enzymes A calcium channel blocker: diltiazem Insecticides based on chrysanthemic acid esters Anti-fungal triazole-containing tetrahydrofurans Bristol-Meyers Squibb anti-psychotic agent BMS 181100 An Eli Lilly protein kinase inhibitor Lotrafiban: A GlaxoSmithKline drug to prevent blood clots Two desymmetrisations

Preliminary note on biological methods This chapter concerns methods totally different from anything else in this section. It is mainly about enzymes as catalysts but also deals briefly with whole organisms. These methods are outside

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

652

29 Enzymes

the experience of most organic chemists and we urge you to study the literature thoroughly before attempting any of them. Fortunately there are books dealing both with laboratory and commercial practice, including three particularly good ones.1

Introduction: Enzymes and Organisms Good and bad features of enzymes as catalysts Whether you use isolated enzymes or whole organisms, you are really employing one or more enzymes as catalysts in chemical reactions. Enzymes are the most refined catalysts in existence. They are usually highly specific: catalysing the reaction of only one compound at a specific reaction site with defined stereochemistry including enantioselectivity. They bring about almost incredibly large rate accelerations - 109 to 1012 is commonplace. They do so in reactions in dilute aqueous solutions at 37 ⬚C and at low concentrations. They do all this with inherently inefficient catalytic mechanisms - general rather than specific acid and base catalysis - that are successful because the protein binds the substrate so that the reagents (such as coenzymes, if any, or water) and the catalytic groups are in exactly the right positions to exert maximum synergistic effects. No wonder that organic chemists use them as catalysts in organic reactions. So why are chemical reagents rather than enzymes used for most reactions? Each enzyme virtue may be a vice in organic synthesis. High specificity is disastrous: who would want a reagent that would reduce just one ketone enantioselectively, however efficient it might be? Alongside large rate accelerations go small turnovers and long reaction times. Living things want to produce small amounts of many different compounds at a trickle over days or even years. Chemists want tonnes of compounds in a few hours. We don’t want to use low concentrations in water. We are not restricted to general acid or base catalysis: NaOH at 100 ⬚C hydrolyses esters by specific base catalysis much faster than the fastest enzymes. And the binding is a problem too: enzymes have molecular weights of tens or hundreds of thousands so that even 1 mole % catalyst might weigh more than the substrate! Now the boot is on the other foot - why should chemists use enzymes? This chapter will show that many enzymes are not very substrate specific and will for example reduce a large variety of ketones with essentially perfect enantioselectivity. Some of these enzymes catalyse reactions that are essentially impossible with chemical reagents. Some can be used in organic solvents at high temperatures. Practically they may be easy to use once someone else has worked out the details and, if they are immobilised on polymers, for example, they can be used again and again with a trivial workup since the starting material gives the product and nothing else. Genetic engineering allows simple organisms such as E. coli to promote, say, enantioselective Baeyer-Villiger oxidations of ketones. As with all synthetic methods, enzymes are used in some reactions with great success and are to be avoided in others. Their use will grow both in the laboratory and in the manufacturing plant.

Organisms: Reduction of Ketones by Baker’s Yeast The ordinary yeast used by bakers in their bread making is a valuable organism for the enantioselective reduction of unsymmetrical ketones.2 It is particularly efficient at the reduction of β-keto-esters such as ethyl acetoacetate 1. An Organic Synthesis procedure3 reveals that the true reagent is sucrose, which provides just one H atom, that plenty of yeast is required, and that 3–4 days are needed to make 20–30 g product 2 in only 85% ee.

29 Ester Formation and Hydrolysis by Lipases and Esterases

O

653

OH

200 g baker's yeast, 500 g sucrose

CO2Et

CO2Et 1.6 L water, 20–30 ˚C, 75–85 hours

2; 24-31 g; 85% ee

1; 40 g

The mnemonic for these reductions is based on the size of the two groups on the ketone. In this case 1 the ‘large’ group is CH2CO2Et and the ‘small’ group Me. With the keto-ester 3, though the ‘small’ group still appears to be Et rather than CH2CO2Et, in fact the other enantiomer (R)-4 is produced in low ee. The solution in such cases is the make the ester group very large, octyl being a popular choice 5. Now we get (S)-6 in good ee. O

OH baker's yeast

RS

sucrose

RL O

RS

OH

baker's yeast

CO2Et

Mnemonic for baker's yeast reduction

RL

OH CO2Et

sucrose

CO2Et

3

(R)-4; 40% ee

O

OH

baker's yeast

CO2Oct

OH CO2Oct

sucrose

CO2Oct

(S)-6; 99% ee

5

On a large scale, slow addition of the ketone and good aeration are essential to cut the reaction time and then kilos of 1 can be reduced in excellent ee in only 3-4 days.4 Under these conditions, each kg of product requires 1.8 kg ethyl acetoacetate 1, 3.9 kg yeast, and 3.5 kg sucrose. O CO2Et 1; 2.25 kg

5 kg baker's yeast, 1 kg sucrose slow addition of ketone to fermenter 40 L water, pH 3-4, 10 L air per minute 20-30 ˚C, 100 hours

OH CO2Et (S)-2; 1.29 kg; 56% yield 95-97% ee

On a large scale, smaller quantities of materials can be used if the sucrose is replaced with ethanol. Since the yeast is an organism there are other enzymes present and EtOH is converted enzymatically into the coenzyme NADPH, the true reducing agent, produced from sucrose in the reductions above. Aeration is again crucial and a bubble column reactor is used. Ethyl acetoacetate 1 is again efficiently reduced and so is hydroxyacetone 7 to give the useful diol5 8. O

112 g baker's yeast

1 60 g

104 g EtOH 1 L water, pH 3-4 25 L air per minute 30 ˚C, 62 hours

2; 40.5 g 56% yield 99.3% ee

112 g baker's yeast

OH 7; 40 g

18 g EtOH 1 L water, pH 3-4 3 L air per minute 30 ˚C, 38 hours

OH OH 8; 22 g; 56% yield 98.2% ee

Ester Formation and Hydrolysis by Lipases and Esterases Lipases and esterases are enzymes that hydrolyse esters in one direction or make them in the reverse direction. A wide variety of such enzymes is commercially available and you are advised to seek a close analogy before choosing one or another.6 If the alcohol 9 is the interesting part, these enzymes can be used to provide enantiomerically enriched alcohol or ester by either reaction.

654

29 Enzymes

O R1

enzymatic esterification of racemic alcohols enzymatic hydrolysis of racemic esters

OH racemic 9 O

R

1

O R2 racemic 10

R1

+

R1

R2

O (S)-10

OH (R)-9 O

+

R1

1

R

OH

O (R)-10

(S)-9

R2

On the other hand if the acid is the interesting part then the same two types of reaction can be carried out to produce enantiomerically enriched esters or acids as required. Both these reactions are equilibria. Both are also kinetic resolutions (chapter 28) and it is not usually possible to get both products in high ee. The key statistic is the rate ratio (‘E’) of hydrolysis or esterification of the two enantiomers. Since the molecules we shall be using are not natural substrates of lipases, the E value will vary considerably but, as you know from chapter 28, it does not need to be very large for acceptable results. This ‘E’ is often called kS/k R in chemical kinetic resolution and called the ‘s value’ in chapter 28.

asymmetric formation of esters from racemic acids

R1

asymmetric formation of acids by hydrolysis of racemic esters

R1

CO2H racemic 11

CO2R2 racemic 12

R1

R1

CO2R2 (S)-12

+

+ CO2H (S)-11

R1

R1

CO2H (R)-11

CO2R2 (R)-12

Important examples of this last type are the 2-aryl propionic acids, widely used as anti-inflammatory drugs. The active enantiomer of ibuprofen 14 can be formed by hydrolysis of esters 13 with the lipase from Candida cylindracea. The natural enzyme gives E = 10 but if it is refolded by precipitation from solution in 1:1 EtOH/Et2O with sodium cholate the E value rises to >100. The unnatural substrate 13 fits the refolded enzyme better than it fits the native enzyme.7

CO2R

13

Candida cylindracea lipase

CO2H

14; ( S)-Ibuprofen

Typical kinetic resolutions of the arylpropionic acids are those of flurbiprofen 16 and ketoprofen 18 with a cell-free extract of Pseudomonas fluorescens. Note the special ester (trifluoroethyl) selected for maximum efficiency and how successful that is: perfect ee in the hydrolysed products at close to 50% conversion.8 The unreacted esters 15 and 17 can of course be racemised by enolisation and added to the next resolution.

29 Ester Formation and Hydrolysis by Lipases and Esterases

655

cell-free extract of pseudomonas fluorescens

CO2CH2CF3

CO2H

pH 7.5 (NaOH), 20 hours 47% conversion 48% (R)-15 recovered

F

F (S)-16; ,43% yield, >99% ee

15 O

O

cell-free extract of pseudomonas fluorescens

CO2CH2CF3

CO2H

pH 7.5 (NaOH), 20 hours 48% conversion 48% (R)-15 recovered

(S)-18; 40% yield, >99% ee

17

In some cases two different enzymes may show opposite enantioselectivities. The lipase from Candida cylindraceae (CCL) and the esterase from pig liver (PLE) do so with racemic cyanohydrins9 such as 20. Note that the acetate is hydrolysed rather than the ethyl ester. Each enantiomer may be reduced in high yield to a single enantiomer of the amine 21. The literature value for the rotation of this compound is [α] D 20.8 and the measured value for (S)-21 was ⫺20.9 and for (R)-21 was ⫹20.6. Though this is not usually a reliable method of assessing enantiomeric purity, it is quite convincing here. Later in this chapter you will meet an alternative way to make the same sorts of compounds using different enzymes. OH

OAc

OH CCL

PLE

CO2Et NC (S)-19; 48% yield

60% conversion

CO2Et

NC

40% conversion

20

(S)-19

OH

OH

excess BH3.THF

H2N

excess BH3.THF

H2N

CO2Et

NiCl2.6H2O, i-PrOH

CO2Et NC (R)-19; 32% yield

CO2Et

(R)-19 NiCl2.6H2O, i-PrOH

(R)-21; 100% yield

(S)-21; 93% yield

Kinetic resolution with racemisation Clearly the efficiency of these reactions is improved if the starting material is racemising under the reaction conditions (see dynamic kinetic resolution in chapter 28). Ketorolac 23, another aryl propionic acid, has an acidic proton and its esters racemise at slightly alkaline pH (9-10). Is it possible to find an enzyme system that will operate in this pH range? First a lipase had to be found that would do the hydrolysis of the ester 22 in the right sense. A selection of lipases gave the wrong enantiomer (R)-23 except for porcine pancreatic lipase but that gave low ees.

CO2Me Ph

N

CO2H Ph pH 8

O (±)-22 Lipase from Mucor meihei Lipase from Pseudomonas spp. Lipase from Candida cylindracea Porcine Pancreatic Lipase

N O (R)-23 ketorolac 94% ee 90% ee 74% ee (S) 32% ee

CO2Me +

Ph

N O

(S)-22 90% ee 93% ee 94% ee (R) 4% ee

656

29 Enzymes

A selection of proteases did much better giving the right isomer of ketorolac (S)-23 in good yield and ee and one, the protease from Streptomyces griseus, did exceptionally well. This was at pH 8 - the normal pH for the enzyme. CO2Me Ph

CO2H Ph

N

N

CO2Me +

Ph

N

pH 8

O

O

O (S)-23 ketorolac

(±)-22 Protease from Aspergillus saitoi Protease from Bacillus subtilis Protease from Aspergillus oryzae Protease from Streptomyces griseus

96% ee 96% ee 70% ee >96% ee

(R)-22 80% ee 62% ee >96% ee >96% ee

With the protease from Streptomyces griseus at pH 9.7 racemisation occurred at a reasonable rate and the enzyme still operated efficiently, giving (S)-23 in good yield and only slightly reduced ee. One crystallisation gave 94% ee, acceptable for an anti-inflammatory.10 Streptomyces griseus protease

CO2Me Ph

pH 9.7 (Na2CO3/NaHCO3)

N 24 hours, 22 ˚C acidify to pH 2 to precipitate acid

O (±)-22

CO2H Ph

N O (–)-(S)-ketorolac 23 92% yield, 85% ee

Enzymes versus whole organisms We have discussed examples of pure enzymes - the lipases - and whole organisms - baker’s yeast and before we continue it is worthwhile to pause and consider the relative merits of the two approaches. Many of these points will be elaborated in the rest of the chapter. There are other methods, as you will see, that fit neither of these descriptions exactly such as engineered micro-organisms that have a particular enzyme over-expressed. There will be examples of these later. Meanwhile here are the main points: Cost: Pure enzymes are expensive while organisms such as yeasts are cheap. Individual cultures of microorganisms may also be expensive. Scale: Very little pure enzyme or bacteria may be needed in contrast to the large quantities of yeast. Specificity: A pure enzyme normally catalyses only one reaction while a living organism contains many enzymes and may catalyse many reactions. Reagents: Pure enzymes normally need no more than pH adjustment though some, say enzymes that catalyse redox reactions, need expensive co-factors. These can normally be supplied by a second enzyme the recycles the co-factor. Organisms need feeding - this is cheap but may require a lot of material. Work-up: Reactions with pure enzymes are clean to work-up though re-isolation of the enzyme will normally be preferred. Immobilised enzymes avoid this problem. The work-up of a reaction with a yeast is very messy indeed. Waste products: None with an enzyme, lots with a yeast but at least they are environmentally friendly.

Desymmetrisation with lipases These are hydrolyses of meso diesters rather than kinetic resolutions and can give quantitative yields of single compounds. The products are half acid, half ester and care must be taken to preserve this

29 Ester Formation and Hydrolysis by Lipases and Esterases

657

difference. One useful group is the Diels-Alder products such as those formed from butadiene and maleic esters 24. Hydrolysis with various lipases11 gives one enantiomer of the half ester 25. CO2R

various lipases

CO2R

Diels-Alder

CO2H

+ CO2R

CO2R

CO2R

24

25

The results are sensitive to the esterifying alcohol ROH. If R = Me 26, results are excellent (99% yield, ⬎98% ee) but fall off as R gets larger. Even with R = Et things are significantly worse (67% yield, 27% ee) and with R = i-Pr they are hopeless (5% yield, 2% ee). It does not do to move too far away from the natural substrates. Three and five-membered rings also give good results: in each case the ester that is hydrolysed is marked - this is not always what one would expect contrast 27 and 28. various lipases

CO2Me

CO2Me

CO2H

CO2H O

CO2Me

CO2H 27; 99% yield 100% ee

26

O CO2Me

CO2Me 29; 85% yield 82% ee

28; 90% yield 99% ee

Another group of meso diesters are the acyclic compounds 30 and 31. Pig liver esterase is good at forming the mono-ester 33 and this compound can be transformed by selective reduction into either 32 or 34. These lactones are enantiomers so that either series may be entered from the same meso starting material.12 In addition the precarious chirality of 33 is more secure in the lactones. R2 R1

CO2H CO2Me

30; usually >85% ee

R2 R1

O

CO2H

LiBH4

R CO2Me

O

H

EtOH

31; usually >70% ee

CO2H

R H

CO2Me

32

33

BH3

H

O

R 34

O

Immobilised enzymes in desymmetrisation The starting material for the next desymmetrisation comes from an interesting reaction first described in chapter 19. The Pd-catalysed attack of AcOH on the mono-epoxide (±)-35 gives the racemic monoester 36. In order to convert this to a single enantiomer it is first made into the meso diester13 37. 0.2 mol% (Ph3P)4Pd

O

HO

OAc

AcOH, THF

(±)-35

Ac2O

AcO

OAc

imidazole

36; 72-76% yield

37; 96-98% yield

Now the acetyl cholinesterase from the electric eel (EEAC) efficiently desymmetrises the diacetate to give the enantiomerically pure monoester 36. The enzyme can be used as a lyophilised powder or is available (from Sigma) immobilised on agarose beads. This form is easy to use and easy to isolate after the reaction. Other cyclic examples work well as does the most impressive open chain example 38 with an E-alkene.14

658

29 Enzymes

OH

18.6 mg EEAC, NaN 3

AcO

OAc

HO

NaH2PO3 buffer, pH 9 9-12 hours room temperature

37; 16 g

OAc OAc 38; 92%ee, 77% yield

36; 11.3 g, 96% ee 86-87% yield

The product 36 can be converted by simple chemistry into either enantiomer of the enone 40 (R = TBDMS). Notice that another enzyme, this time a plant lipase, is used to hydrolyse the acetate 41 efficiently. There is no asymmetric induction in this step as 41 is already a single enantiomer.15 HO

OAc

Et3N, DMAP

36 PCC O

RO

t-BuMe2SiCl

1. NaOMe MeOH

OAc

2. MnO 2 CH2Cl2

39; 77-80% yield

CH2Cl2

RO

O

(S)-40; 87-90% yield

CH2Cl2 wheat germ lipase

OAc

O

phosphate buffer

41; 83% yield

OH

O

t-BuMe2SiCl

OR

Et3N, DMAP CH2Cl2

42; 60% yield

(R)-40; 87-90% yield

Polymer-supported reagents and enzymes Polymer-supported enzymes have been combined with polymer-supported reagents in the synthesis of the bryostatins. The nitrone derived from the nitro compound 43 supported on a soluble aryl poly-ether polymer undergoes an efficient 1,3-dipolar cycloaddition with butenone to give the isoxazoline 44 and hence by reduction the racemic syn compound 45 required for the synthesis.16 N

PhNCO

O

NO2

O

O O

44; 93% yield

43 N

O

O

L-Selectride CH2Cl2

O OH syn-45; 93% yield racemic

The single enantiomer of 45 is produced by enantioselective acylation with vinyl acetate catalysed by Novozym 435, an immobilised lipase from Candida antarctica containing about 1% w/w enzyme on a macroporous polypropylic resin. You will appreciate that as both substrate and catalyst are on polymers, one at least must be soluble in the reaction mixture. The polymer is stripped from the substrate 45 with HF. N

OAc

(±)-syn-45

O

anhydrous toluene 70 ˚C, 4 hours 40% conversion

N

O

HF, H 2O

Novozym 435

O OH (+)-45; 96% yield; >99% ee

MeCN

HO OH (+)-46

29 Ester Formation and Hydrolysis by Lipases and Esterases

659

This is a kinetic resolution that works by enantioselective acylation of the unwanted enantiomer of the alcohol. The reaction is therefore an ester exchange and vinyl acetate is an efficient acetate transfer agent since the other product is ‘vinyl alcohol’ better known as the enol of acetaldehyde so the reaction is irreversible. N

O

OH +

OAc

O syn-45

N

Novozym 435

+ OH

O

O OAc

(–)-47

MeCHO

Effects of amines on lipases and esterases We end this section with two examples of the effects amines may have on esterase activity. The aggressive nitrogen nucleophile can obviously interfere with ester formation or hydrolysis. In designing a synthesis based on these enzymes we must find a symmetrical intermediate that might be desymmetrised. In the synthesis of the simple β-lactam 48, simple disconnections led to the symmetrical amino-diester 50, with the nitrogen atom protected. CO2Me NH O

C–N

HO2C

CO2Me

protect NH2

NH2

amide

48; β-lactam

MeO2C

CO2Me NHZ

esterify

49

50

Synthesis of a suitable substrate from citric acid 51 was straightforward17 and Ohno18 found that the Cbz-protected dimethyl ester 53 could be desymmetrised efficiently with pig liver esterase (PLE) or with even higher enantioselectivity, with the microbial enzyme from Flavobacterium lutescens (98% ee). The β-lactam 48 could be closed by the dipyridine-disulfide/Ph3P method. HO

CO2H

HO2C CO2H 51; citric acid

O

1. H 2SO4 fuming 2. HCl MeOH

CO2Me MeO2C 52; 700 g scale NHCbz

400 units PLE

NHCbz

1. NH 4 AcO Na(CN)BH3 2. Et 3N ClCO2CH2Ph

MeO2C

CO2Me 53

H

1. H 2, Pd/C

CO2Me 0.1M phosphate buffer pH 8, 25 ˚C, 1.5 hours

HO2C CO2Me 54; 93% yield

2. PySSPy Ph3P, MeCN

NH O (S)-(+)-48 84% yield; >96% ee

The selectivity of hydrolysis of this type of diester 55 (see also 31) by PLE depends on the substituents. The two enantiomers of 56 are formed as shown with various degrees of ee and you will notice that there is no easily discerned system. Thus R = Pr and R = i-Pr give opposite enantiomers as do amides such as 53 and the corresponding free amine.19 55

R

R PLE

MeO2C

CO2Me

HO2C

(S)-56; R = Bn, Bz, i-Pr Ph(CH2)2 and CO2Me RCONH

R or HO2C

(R)-56; R = Me, Et, n-Pr, AcO, CO2Me and NH2

660

29 Enzymes

Other acylating enzymes Amide formation and hydrolysis is of course also catalysed by enzymes and the recent report that the penicillin acylase from Alcaligenes faecalis catalyses an efficient kinetic resolution of racemic primary amines is very promising. A typical case is our old friend α-methylbenzylamine 57 (see chapter 22). One equivalent of an acyl donor 58 is needed and nearly 50% of the amide 59 can be isolated in excellent ee. The E value is 350 for this amine.20 H2N

Ph

+

NH2

Ph

penicillin acylase, water, pH 10 (KOH), 10–30 minutes

O (±)-57; α-methyl benzylamine

Ph

Ph

O 59; 47.3% yield 98.5% ee

one equivalent 58

58; acyl donor

H N

amide product precipitates

A selection of four amines shows the possible range. Some have aromatic substituents, some aliphatic and one 66 has an alcohol but no ester formation is observed. Though E varies from about 100 to over 1000, these numbers are all large enough to ensure excellent ee (see chapter 28) and the yields are all close to the maximum 50%. If the free amines are wanted, the same enzyme can be used to hydrolyse the amides. This process ensures essentially 100% ee as there is a further kinetic resolution in the hydrolysis. Cl

Cl H N

Bn

H2N

H2N

H N

Bn

Ph

Ph

O 60; E = >1000

O 61; 49% yield; 99.3% ee

Bn

H2N

62; E = 120

H2N

H N

O 65; 46.8% yield; 98.1% ee

64; E = 250

63; 45% yield; 96% ee

Ph

H N

Bn

Ph

O

OH 66; E = >100

OH 67; 43% yield; 98% ee

Enzymatic Oxidation Asymmetric dihydroxylation of benzenes When benzene falls on the soil it is processed (and so detoxified) by micro-organisms such as Pseudomonas putida by oxidation first to a diol21 then to catechol and finally to CO2. The diol 68 cannot be isolated as the dehydrogenase is too efficient but in 1970 Gibson discovered mutant F39D of Pseudomonas putida that lacks the dehydrogenase and produces the cis diol 68 in good yield.21 dioxygenase

OH

dehydrogenase

OH CO2

OH 68; cis diol

OH catechol

This mutant has been developed by Ley, Hudlicky and others into a practical method for the asymmetric oxidation of substituted benzenes to give the unstable diols best preserved as acetals. Even one substituent is enough to make the diol chiral and dihydroxylation normally occurs at

29 Enzymatic Oxidation

661

the 2,3-bond in the ring. Bromobenzene gives the diol 69 and hence the acetal 70 in 100% yield without isolation of the diol.22 Br

Br

Br

Pseudomonas putida

OH

OMe

+

O

TsOH

OMe

O 70; 100% yield

OH 69

The practical details23 are given in Organic Syntheses. One colony of Pseudomonas putida F39D from an agar plate is incubated in 50 mls broth at 30 ⬚C for 24 hours with arginine to induce the dioxygenase enzyme for the substrate provided. This is added to 500 mls broth with 10 mls PhCl and arginine. The broth contains phosphate, K(I), Mg(II), Ca(II), molybdate, Fe(II), N(CH2CO2H)3 as well as the trace metals Zn(II), Fe(II), Mn(II), Cu(II), Co(II), Na2B4O7, and EDTA. Incubation at 30 ⬚C for 48 hours gives the diol. The acetonide 70 has been used to make both enantiomers of pinitol 73. The rigid acetonide directs dihydroxylation or epoxidation to the bottom face of the ring to give 71 and 74. The epoxide rings in 72 and 74 can be opened regioselectively with methanol to give the usual trans-diaxial products such as 75. Carrying out the reactions in two different orders gives the two enantiomers24 (⫹)-73 and (⫺)-73. Br O

MNO

70

OH

O

OsO4

H2O

1. LiAlH 4

O

1. MeOH Al2O3

O

2. HCl H2O acetone

2. mCPBA

O

HO

HO

mCPBA

OH 71; 85% yield

CH2Cl2

OH 72; 86% yield

MeO

OH

HO

OH

OH (+)-73; pinitol OH

Br 1. MeOH Al2O3

O

O

2. LiAlH 4

O

2. HCl, H 2O acetone

OH

MeO

OH

OH (–)-73; pinitol

OH 75; 85% yield

74; 80% yield

HO

MNO, H 2O

O

MeO

O

1. OsO 4

All monosubstituted benzenes react with the same regio- and enantio-selectivity whether the substituent is halogen, alkyl, alkenyl, CO2H, CN, COMe, OMe or CF3. Thus toluene gives the diol 76 that can be produced at 3g/litre in the broth described above. It is also a commercial product. Me

Me

Me Pseudomonas putida

OH +

toluene

OH 76; 3 g per litre

OMe

O

TsOH

OMe

O 77

The acetonide 77 has been used in the synthesis of prostaglandins such as PGF2α 80. Once the aromaticity of the benzene ring is broken, further oxidation is easy and the diene 77 can be ozonised to the keto-aldehyde 79 that readily undergoes an intramolecular aldol to give the enone 79. It is more obvious that 79 can be a precursor to 80 when it is redrawn to show the possibility of adding the side chains by conjugate addition and enolate trapping22 (chapters 9 and 10).

662

29 Enzymes

Me O3

77

O

O

DME

O

OHC

CO2H O

O O

78; 70% yield

O

O

Al2O3

=

–60 ˚C EtOAc

O

O

HO

79; ~100% yield

OH 80; PGE 2α

Dihydroxylation of quinoline, to give 81, isoquinoline, to give 82, and dibenzothiophene to give 83 are less predictable but in each case it is a benzene rather than a heterocyclic ring, and the 2,3-bond next to a carbon rather than a heteroatom that reacts.25 The products are drawn to show the identity of the enantioselectivity. OH

OH

OH OH

OH

OH

N

S

N 81

82

83

Other organisms may give different results: benzoic acid gives the expected diol 85 with Pseudomonas putida but reacts at the bond bearing the CO2H group with Alcaligenes eutrophus to give26 84.

HO2C

CO2H

CO2H

OH OH

Pseudomonas putida

Alcaligenes eutrophus

OH

OH 84

85

Recent developments include the over-expression of the gene for the dihydroxylation and its transfer27 into the common and easily grown organism Escherischia coli to form recombinant JM 109 (pDTG601A). This organism creates diols 87 from many monosubstituted benzenes more conveniently. The closely related recombinant organism carries out the rearomatisation to give 88. R

86

R

R

Pseudomonas putida mutant 39D

OH

E. coli recombinant JM 109(pDTG601)

OH 87

OH E. coli recombinant JM 109(pDTG602)

OH 88

This method was used by Hudlicky in a synthesis of the alkaloid narciclasine. The diol 90 was trapped as the usual acetal and reacted as a diene with the nitroso ester MeO2C-NO in a hetero-Diels-Alder reaction to give 91. Suzuki coupling of the vinylic bromide with the right aryl boronic acid gave 92 and the N–O bond was reduced with Mo(CO) 6 to give the advanced intermediate 93 on the way to narciclasine28 94.

29 Enzymatic Oxidation

Br

663

Br

Br

E. coli JM109 (pDTG601A)

1. TsOH (MeO)2CMe2

OH

2. NaIO 4

Br

Br OH 90; 4g/litre >99%ee

89

Br

NH2CO2Me

Br

ArB(OH)2

O CO2Me 91; 70% yield

Pd(Ph3P)4 Na2CO3 EtOH

O

OH

O

OH

O

O N O

O O N

Mo(CO)6

O CO2Me

O

O

Ar

NHCO2Me

O

OH NH

O

OMe 92

O OH 94; narciclasine

93

Epoxidation There are all sorts of problems with epoxidation by micro-organisms and in general laboratory chemists prefer to use the Sharpless or Jacobsen epoxidations described in chapter 25. The ω-hydroxylase from Pseudomonas oleovorans does epoxidise aryl ethers of allylic alcohols with good selectivity and one product has been used in the synthesis of the β-blocker metropolol.29 However the organism requires gaseous hydrocarbons as carbon sources and the epoxide products poison it.

OAr

ω-hydroxylase

O

H

HO O

MeO

Pseudomonas oleovorans

H

i-PrNH

O

MeO

i-PrNH2

95

96; ~100% ee

97; metropolol; 98.4% ee

The Nippon Mining Company uses the soil micro-organism Nocardia corralina B-276 to produce a series of epoxides that are available commercially as single enantiomers. This micro-organism uses glucose as a carbon source and is not poisoned by the epoxides it produces. It also is very versatile as the selection 98 - 104 shows. The two epoxides 103 and 104 are especially notable as they are both made from hexa-1,5-diene.30 O

H

Alkyl 98; C 6 to C18 O

O

Me

Alkyl 99; C 6 to C11

H

O

H OR

O

O

H Ar

CF3 101

100; R = C 5 to C12 O

103

H

102

H

104

HO

O

The Baeyer-Villiger reaction The Baeyer Villiger reaction is an oxidation of ketones by peroxyacids that inserts an oxygen atom between the carbonyl group and the more substituted side of the ketone. The biological

664

29 Enzymes

version does the same thing and is particularly interesting to us when it creates asymmetry from a symmetrical ketone. The enzymes are mono-oxidases that are coupled to the reducing enzyme glucose dehydrogenase.31 O

The Chemical Baeyer-Villiger Rearrangement mCPBA

O2

6-phosphogluconate

NADPH

monooxidase

O

glucose dehydrogenase

H2O

O

NADP coenzymes

The Biological Baeyer-Villiger Rearrangement

glucose-6phosphate

Examples include simple achiral ketones such as 105 and meso compounds with some stereochemistry32 such as 107 that initially gives the seven-membered lactone 108 but this rearranges into the five-membered lactone 109. O O

O

O

105

O

O

O

O

HO

HO 107

106; 80% yield >98% ee

OH

H 108

109; 88% yield >98% ee

Nature has many surprises in the chemistry she does. One organism produces a scarcely credible result in the Baeyer-Villiger reaction. Acinetobacter calcoaceticus catalyses a kinetic resolution of the important bicyclic ketone 111. The chemical reaction gives the expected lactone 110 with migration of the more highly substituted carbon atom. The biological reaction does indeed give one enantiomer of the same lactone but the remaining material is not unreacted 111. Instead the other regioisomer 112 in the other enantiomeric series is formed in good yield and ee. Nature can accomplish reactions impossible to us, so far.33 The latest news34 on this reaction is that recombinant whole cells of E. coli expressing cyclohexanone mono-oxygenase are even more efficient. H

O

O

mCPBA

H Acinetobacter calcoaceticus

H

O

O

O

O

O

+

NCIB 9871

H (±)-110

H 110; 44% yield >95% ee

(±)-111

H 112; 42% yield >95% ee

Baeyer-Villiger reactions with engineered baker’s yeast The mono-oxygenase from Acinetobacter sp NCIB 9871 can be cloned and expressed in the same baker’s yeast that we met at the start of this chapter. This yeast carries out Baeyer-Villiger reactions and supplies its own NADPH from other enzymes in the yeast. A conventional kinetic resolution of cyclopentanones 113 gives single enantiomers of lactone and starting material.35 O

O engineered baker's yeast

R 113

O O

+

15C(pKR001)

R = Me R = n-Hex

R

R 114; 36% yield; 32 ee

115; 34% yield; 44% ee

116; 32% yield; >98% ee

117; 42% yield; >98% ee

29 Nucleophilic Addition to Carbonyl Groups

665

Whole cells of engineered E. coli BL21(DE3)(pMM4) desymmetrise cyclohexanones with substituents in the 4-position with variable results.36 Simple mono-substitution 118 and 120 gives good results but the enzyme is very sensitive to hydroxyl groups giving a lower yield with the tertiary alcohol 122 and poor ee with the secondary alcohol 124. O

O engineered E. coli

Et

O

Et H

engineered E. coli

Br

O

H

O

engineered E. coli

Et

Et HO

engineered E. coli

HO

O

HO H

HO 123; 54% yield 94% ee

122

H 121; 63% yield 97% ee O

120

O

O

Br

H 119; 91% yield 97% ee O

118

O

O

H 125; 61% yield 9% ee

124

Nucleophilic Addition to Carbonyl Groups Asymmetric addition of cyanide to aldehydes Oxynitrilase enzymes from various plants catalyse the addition of cyanide ion to aldehydes to give the cyanohydrins present in many plants. Amygdalin, the cyanohydrin of a trisaccharide, is perhaps the most notorious of these as it releases HCN from oil of bitter almonds and has been responsible for the death of people drinking old samples of the liqueur noyeau.37 The (R)-oxynitrilase from the almond Prunus amygdalis has unusually low substrate specificity and gives good yields of a variety of cyanohydrins 126 providing that the reaction medium has two phases to inhibit the uncatalysed addition of HCN to the aldehyde.38 O R

(R)-oxynitrilase, KCN, AcOH two-phase reaction mixture:

H

i-Pr2O, H 2O or EtOAc/Avicel

OH

H

R CN (R)-126

R = Ph; 95% yield, 99%ee R = 3-thienyl; 95% yield, >99%ee R = MeS(CH2)2; 98% yield, 96%ee R = Ph(CH2)3; 94% yield, 90%ee

The (S)-oxynitrilase from Sorghum bicolor produces the other enantiomer but only from aromatic aldehydes.39 The products can be directly converted into hydroxyacids 127 without loss of ee. O R

OH

(S)-oxynitrilase, KCN, AcOH

H

two-phase reaction mixture: i-Pr2O, H 2O

R

OH

H

H

CN

H2O

(S)-126; 78-97% yield 52-97% ee

R

H CO2H

(S)-127; 63-76% yield no loss of ee

These two methods involve the potential release of toxic HCN during the reaction and a cleaner as well as a safer method is to use acetone cyanohydrin 128 as a cyanide transfer reagent. The enzyme from the Brazilian rubber tree Hevea brasiliensis, called a hydroxynitrile lyase, catalyses cyanohydrin formation from aliphatic as well as aromatic aldehydes.40 OH + CN R 128

O

hydroxynitrile lyase

H

Na-citrate buffer, pH 4.0, 2 hours room temperature

OH

R = Ph; 94% ee R = n-Pr; 80% ee R CN R = i-Pr; 81% ee R = n-Bu; 84% ee (S)-126 H

666

29 Enzymes

The commercial process run by Solvay Duphar, based on the work of van der Gen,41 uses an aqueous extract of almond meal for (R)-126 and the sorghum enzyme for (S)-126. Again a twophase system (H2O/Et2O) is used and the rather unstable cyanohydrins, that may racemise by reversible loss of HCN, are stabilised by conversion to silyl ethers such as 129–131. The last example shows that direct is preferred to conjugate addition. OR 128 (R)-oxynitrilase

RCHO

two phases: H2O/Et2O

OH

O

H OR

OR

CN

H

R CN (R)-126

H O

H CN 130; R = TBDMS

O 129; R = TBDMS

CN 131; R = TBDMS

Products derived from cyanohydrins The obvious hydrolysis of the nitrile must be carried out carefully to avoid racemisation.39,42 Two successful ways are direct conversion into esters 134 in acid solution or a two step process to the acids 132 via the much less easily racemised amides 133. OH R

OH

conc HCl

H

CO2H (R)-132

100 ˚C

R

OH

conc HCl

H

20 ˚C

CONH2 (R)-133

H

OH

1. HCl, MeOH

R CN (R)-126

2. H 2O

R

H

CO2Me (R)-134

Conversion to other carbonyl compounds is more tricky because of the increased likelihood of enolisation. The protected cyanohydrins 136 can be reduced to the aldehydes 135 with 1.5 equivalents of DIBAL and cautious acidification. The more stable ketones 138 are best made in two stages.41 Additions of Grignard reagents gives the imine derivatives 137 that can be worked up in acid to give the ketones 133.

OTBDMS H R

OTBDMS H

1. rapid addition of DIBAL, THF, 20 ˚C

CHO

R

2. cautious acidification

135

OTBDMS H

MeMgI

OTBDMS H

H2SO4

R

CN

R NMgI 137

136

H2O, 0 ˚C

O 138; R = Ph; 80% yield, 92% ee

Secondary amines can be formed from the same starting materials 136 by tandem Grignard addition and borohydride reduction.43 The reduction shows good Felkin-Anh selectivity (chapter 21) and the silyl group can be cleaved from the products 140, as they are safe from racemisation, to give a range of amino alcohols 141. Compounds like these can also be made by asymmetric amino-hydroxylation (chapter 25).

OTBDMS H R1

CN 136

R2MgI

R1

OTBDMS H R2 NMgI 139

1. NaBH 4 2. HCl

R

1

OTBDMS H R2 NH2.HCl 140

OH

1. HF 2. HCl

R

1

H

R2

NH2.HCl 141

29 Nucleophilic Addition to Carbonyl Groups

667

Enzymatically catalysed aldol reactions The aldol reaction is extensively used in Nature as in the laboratory to make C–C bonds and some aldolases have been used in asymmetric synthesis. One of the most popular has been the fructose-6-phosphate aldolase44 from rabbit muscle, familiarly known as RAMA. The enzymatic reaction combines the enol from dihydroxyacetone phosphate (DHAP) 142 with glyceraldehyde-3-phosphate 143 in a diastereo- and enantioselective aldol reaction. PO in these diagrams means phosphate. O

O

PO

+

OH 142; DiHydroxyAcetone Phosphate (DHAP)

H

OH

O

FDP aldolase

PO

OP

OP OH OH 144; Fructose 1,6-diphosphate

OH 143; Glyceraldehyde 3-phosphate

The rabbit enzyme requires DHAP for the enol component but is more relaxed about the aldehyde. The azido-aldehyde 145 reacts cleanly with the formation of two new chiral centres 146 and the phosphate is removed by treatment with a second enzyme, a phosphatase.45 Then reduction of the azide gives the trihydroxy pyrrolidine 147 with excellent control over a third chiral centre, less than 10% of the epimer being formed. The ees of the compounds are not easily determined from the published work.

H2O, pH 6.8

N3

H

HO

2. acid phosphatase

HO

OH

O

1. DHAP, RAMA

O

H2

N3

Pd(OH)2/C

OH 146; 78% yield

145

OH

HO N H 147; 74% yield

The insect pheromone exo-brevicomin 151 has been made this way using the ketoaldehyde 148 as the electrophile with DHAP as the usual enol component.46 The reaction is entirely chemoselective: the ketone in 148 acts neither as enol nor electrophile. The product is a diol that one might think of making by an AD reaction (chapter 25). In acid solution it forms a cyclic acetal 150 having the skeleton of brevicomin. The rotation but not the ee of the brevicomin made this way is quoted. O OHC

148; 3 fold excess

1. DHAP RAMA

O

OH

HO

H2O, pH 6.3 2. acid phosphatase

O

O H

OH 149; 48% yield

OH

O O 150

O O 151; exobrevicomin

Synthesis of syringolide using an enzyme-catalysed aldol reaction Syringolide 152 is a natural product that elicits an immune response from soyabeans. A series of straightforward disconnections leads back to a tetraol 156 that looks like an aldol product from

668

29 Enzymes

DHAP and a very simple aldehyde. Other than doubts about the geometry of the alkene in 155 and chemoselectivity with all the OH groups, it is a simply designed synthesis.47

R

152 (–)-syringolide R = n-heptyl

OH H

O

O 1,1-diX

O

hemiacetal

H

O

conjugate addition

HO

CO2H 1,3-diCO

CO2H

R

O

HO

ester OH 154 O

HO

OH

aldol

OH

aldol

OH

HO

C–O O

1,3-diCO

HO

OH

O

R

HO

153

O

O

HO

O

O

HO

R

1,3-diX

R HO

O

O

O

H

155

OH

OH

HO

O

156

The solution to the chemoselectivity problem is minimal protection: a p-methoxybenzyl group for the hydroxyaldehyde and an acetal 158 for the diol unit in the product 157 of the enzyme-catalysed aldol reaction. The solution to the alkene geometry problem is going to be intramolecular trapping by the one remaining free OH group. O

O

O HO

OP

OR1 O R1 = p-MeOC6H4CH2

1. FDP aldolase 2. acid phosphatase

HO

OH

OR1 HO 157; 65% yield

Me2CO TsOH

OH

O

OR1

O

158; 67% yield

The Meldrum’s acid derivative 159 was chosen to represent the keto-acid electrophile in reaction with the carbonyl group of 158. One of the two carboxyl groups can be captured by the free OH group in 158 so that the geometry of the alkene is automatically correct in 160. Removal of the p-methoxybenzyl group by oxidation and the acetonide with acid allows conjugate addition and hemiacetal formation to give ( – )-syringolide 152 in one step from 160. O O

O

1. 158, THF reflux

R

O

O

2. SiO 2 hexane -EtOAc

O OR1

O 159

1. DDQ H2O/CH2Cl2 (83% yield)

O

O

R

O

160

R

OH H

O

O 2. TsOH H2O/acetone (55% yield)

H

O O

HO (–)-152

Three new chiral centres are created in this step. Removal of the p-methoxybenzyl group from 160 gives the free OH group that carries out conjugate addition 161. It approaches the underside of the lactone as the five-membered acetal is planar and the CH 2OH group is already on the underside. Protonation of the enol product is under thermodynamic control and may change in the next reaction. Removal of the acetal gives two more free OH groups only one being able to add to the ketone 163 to give the hemiacetal in 152. The stereochemistry of this centre is also under thermodynamic control and each addition ensures that each pair of five-membered rings have the more stable cis fusion in the final product 152.

29 Nucleophilic Addition to Carbonyl Groups

O [O]

O

R

O

O

H

O

O 161

152

O

O

OH

O

H

R HO

O

O

O

O

O

H

R

O

160

669

HO

162

163

Engineered aldolase A different aldolase has been over-expressed in E. coli and used by Chi-Huey Wong in his synthesis of epothilones. It is 2-deoxyribose-5-phosphate aldolase (DERA) and the natural reaction is the condensation of acetaldehyde as enol with glyceraldehyde-3-phosphate 164 as electrophilic component to give an aldol product 165 that is trapped as a hemiacetal 166. O H+

PO

OH

O PO

H

O

OH 164; glyceraldehyde -3-phosphate

OH

PO

CHO

HO 166; 2-deoxy-D-ribose -5-phosphate

OH 165

This enzyme is much less fussy than RAMA and will accept other enol components or electrophiles. In this case the β-hydroxyaldehyde 167 gives the six-membered cyclic hemiacetal 169 and the acetal 170 of unstable 2-hydroxypropanal gives one diastereoisomer of the aldol 171 by dynamic kinetic resolution and hence 172 in moderate yield.48

MeCHO

HO

DERA

Me

Me Me

167

OH 169; 48% yield

168

OMe

OH 1. Dowex (H +)

O

OH

CHO

H 2. MeCHO DERA pH 7.5

OH 170

OH

CHO

HO

H

MeO

O

OH

O

HO 172; 38% yield

OH 171

Both these products were used in the synthesis of epothilone A 173. The disconnection of the lactone is simple but the other is less obvious. The epoxide must come from a cis alkene and that can be made by a Suzuki coupling of a borane derived from 174 and the Z-vinyl iodide 175. As you will see, 174 was made from 169 and 175 from 172, unlikely as it seems. O S HO

N

RO

N OR CO2R macrolactonisation

O O

OH

O

173; epothilone A

S

I Suzuki

O

OR 174

175

670

29 Enzymes

The hemiacetal 169 was first oxidised to the lactone 176 and the lithium enolate allylated with the expected stereoselectivity to give 177 and then, by a series of reactions described in the workbook, into the aldehyde 178. O

O

O

Br2

O

OR

O

1. LDA

169 Me

BaCO3

2. allyl bromide

Me

OH 176; 62% yield

OH 177; 85% yield

178; R = p-MeOC6H4CH2

A stereoselective aldol reaction between the β-ketoester 179 at the less enolisable position and 178 creates the complete carbon skeleton of 174 with excellent stereoselectivity at the new OH group 180. This OH group was used to control the stereochemistry of an Evans 1,3 reduction (chapter 21/30) before it was oxidised to the required ketone 181: this is the same as 174 but has the protecting groups specified.

CO2t-Bu

RO

CO2t-Bu

1. NaH, 2. BuLi

H O

3. 178

1. Me 4NBH(OAc) 3 –30 ˚C, 83% yield 2. TBMSOTf 2,6-lutidine 100% yield 3. Dess-Martin 90% yield

O OH 180; 70% yield

179

RO

CO2t-Bu

OTBDMS O 181; R = p-MeOC6H4CH2

The other half was made from 172 by formation of the protected keto-aldehyde 183 followed by successive stereoselective Wittig-Horner and Wittig reactions to put in the two alkenes with control over stereochemistry. The Wittig-Horner is under thermodynamic control giving the trisubstituted E-alkene 185 while the Wittig is under kinetic control giving the Z-iodoalkene 175; R = Ac (chapter 15). 1. AcCl, pyridine, 95% yield 2. BF 3.Et2O, 84% yield

Swern

S

172 3. HS(CH 2)3SH, TiCl 4, 97% yield

OH S

DMSO (COCl)2

OAc 182

S

O Ph2P

N 184

1. BuLi –78 ˚C

S

O S

OAc

183; 95% yield S 1. Hg(ClO 4)2

S

N

2. 183

S

2. Ph 3P=CHI

175; R = Ac 60% yield (2 steps)

OAc 185; 88% yield

The Suzuki coupling works very well - the moderate yield being offset by the degree of convergence achieved. Removal of the acetate and t-butyl ester groups allows a very efficient Yamaguchi lactonisation (see workbook) and then removal of all the remaining protection groups gives the cis alkene 188 that can be epoxidised stereoselectively by dimethyldioxirane though only in 45% yield to give epothilone A 173. All the chiral centres in the final product, as well as some that were discarded en route, come from the original enzymatic aldol reaction.

29 Practical Asymmetric Synthesis with Enzymes

671

S 1. 9-BBN

RO

1. MeONa, MeOH 85% yield

N

181 2. 175; R = Ac [PdCl2(dppf)2] Ph3As, Cs 2CO3

2. TBDMSOTf 2,6-Lutidine 3. NaOH (cat.) MeOH

OAc CO2But O

OTBDMS 186; 65% yield S

RO

1. Yamaguchi lactonisation 85% yield

N OH CO2H

R

HO

O OTBDMS 187; 78% yield (last two steps)

173

O

2. DDQ, CH 2Cl2 99% yield

O

3. HF, pyridine 95% yield

OH 188

O

Practical Asymmetric Synthesis with Enzymes If you have reached this point in the chapter with the feeling that enzymes are all very well for the experts but that you are unlikely ever to use one, then this section is for you. Immobilized enzymes are now freely available from several companies and pharmaceutical companies use them on a large scale almost as a matter of routine. There are still problems of course and most asymmetric synthesis is not done with enzymes. In this section we set out to convince you that enzymes are practical reagents in organic solvents as well as in water and that practical minded chemists use them. An excellent review may convince you more.49 Recently a whole issue of the journal Organic Process Research and Development (2002, 6, issue 4, 420 ff.) was devoted to biocatalysis and the introductory article50 makes the point too.

A calcium channel blocker: diltiazem Calcium channel blockers, such as amlodipine introduced in chapter 22, are used to treat angina and hypertension. Diltiazem 189 was introduced by the Tanabe company in 1984 and is a simple derivative of the benzothiazepine 192 having two adjacent chiral centres on the heterocyclic seven-membered ring. OMe

OMe

S OAc

S

C–O C–N

N O

O Me2N

189; ditiazem

C–O C–N

O

N

Me2N

190

OMe

OH N H

O Me2N

S

O

Cl 191

192

Disconnecting the ring amide reveals a 1,2-difunctionalised compound 193 that can be made by the nucleophilic opening of the epoxide 195 with the anion of the thiol 194. To avoid chemoselectivity problems the corresponding nitro thiol is used. The chirality is already present in the epoxide which can be made by an asymmetric Darzens reaction or by Sharpless epoxidation. You may notice that the opening of the epoxide is, strangely a cis opening (see the workbook for details). Originally the resolution of an intermediate was used.51

672

29 Enzymes

OMe

NH2

C–N

1,2-diX

SH

S

192

OMe

NH2 + O

C–S

amide

RO2C

RO2C

OH 193

194

195

Enzymatic methods use lipase catalysed reactions on epoxides 195 or the adducts such as 193. Racemic trans methyl ester 195; R = Me can be hydrolysed by Candida cylindraceae lipase using mixed solvents MeCN, CH2Cl2, or cyclohexane and water while no hydrolysis occurred in pure water. Cyclohexane gave the best results: the ‘wrong’ ester was hydrolysed while the ‘right’ one remained behind in the organic layer making separation simple. Ten grams could be turned over in 1.5 hours at room temperature.52 Transesterification with chymotrypsin or lipases was less effective. OMe

OMe Candida cylidracea lipase

O

+ O

pH 7.4, H 2O cyclohexane

MeO2C (±)-trans-195; R = Me

OMe

O

MeO2C HO2C (2R,3S)-trans-195; R = Me (2S,3R)-trans-195; R = H 30-40% yield, 96-98% ee

Acylation of the adduct 196 of o-nitrophenylthiol and the epoxide (±)-195 with acetic anhydride or vinyl acetate in THF at room temperature catalysed by the lipase from Pseudomonas cepacia gave better results. Reactions with acetic anhydride are of course not possible in aqueous solution. The reaction took 2 days with Ac2O and 7 days with vinyl acetate but the yields and ee were excellent. At 50% conversion, 100% yield of the acetate product 197 and the other enantiomer of the unreacted alcohol 196 can be isolated.53 OMe

NO2

Pseudomonas capacia lipase

S

Ac2O or

OMe

NO2 S

OAc

RO2C OH (±)-196; R = Me

RO2C OAc (2S,3S)-197; R = Me 100% yield, >95% ee

Insecticides based on chrysanthemic acid esters The most important insecticides today are based on chrysanthemic acid: they are incredibly potent and can be used in small amounts only on the target pest. Chrysanthemic acid 198 is natural but the natural pyrethrin insecticides are too light sensitive to find much use. Chiral synthetic compounds such as cypermethrin 200 and deltamethrin 199 are used as single enantiomers.54 They are based on the cis acid while natural chrysanthemic acid is trans. X OH O 198; trans- chrysanthemic acid

X

O

O

O CN 199; X = Br, (1 R,cis,αS)-deltamethrin 200; X = Cl, (1R,cis,αS)-cypermethrin

29 Practical Asymmetric Synthesis with Enzymes

673

The synthetic compounds are obviously esters of a cyclopropane carboxylic acid and a cyanohydrin 202. Enantioselective transesterification of butanol and the acetate 201 of the cyanohydrin using immobilised lipases gives the required (S)-alcohol 202 and the unreacted enantiomer of the acetate (R)-201 easily racemised with Et3N. Reports using different lipases appeared at the same time: CHIRAZYME® L-6, the lipase from Pseudomonas immobilised on sephadex DEAE was used in i-Pr2O and the racemisation carried out with Et3N under reflux in the same solvent.55

BuOH, lipase

AcO

OPh CN (±)-201

AcO organic solvent 35 ˚C

OPh

+ HO

CN (R)-201

OPh CN (S)-202

Alternatively the Amano immobilised lipase P gave (S)-202 in 49% yield and ⬎99% ee with (R)-201. The mixture was separated by esterification with the chrysanthemic acid derivative giving the wanted ester and the easily separated unreacted acetate (R)-201 racemised with Et3N at 80 ⬚C in toluene. The acid was prepared in only 86% ee so the alcohol (S)-202 also acts as a resolving agent. Cypermethrin 200 is isolated in good yield and excellent diastereo- and enantiomeric purity after one recrystallisation.56 Cl

Cl (R)-201 + (S)-202

Cl

Cl

pyridine, 0 ˚C

O

Cl

O

OPh

O CN 200; X = Cl, (1R,cis,αS)-cypermethrin 82% yield, 99% de after recrystallisation

203; 86% ee

Anti-fungal triazole-containing tetrahydrofurans A major breakthrough in modern medicine has been the discovery of effective anti-fungal compounds. One family such as ketoconazole 204 or saperconazole 205 has a dioxolan core carrying the active imidazole or triazole functionality. Schering-Plough have developed a new type having an apparently simpler tetrahydrofuran (THF) core such as 206. This compound presents special problems for asymmetric synthesis. O X

O

O

O X

F

N

N

N Z

N 204; X = Cl, Z = CH, ketoconazole 205; X = F, Z = N, saperconazole

N O

R F

N

N N

R

N 206; SCH51048, anti-fungal compound

Initial attempts used the Sharpless epoxidation of the allylic alcohol 207 followed by addition of a carbon nucleophile and the triazole to give the ditosylate 208. The two tosylates are diastereotopic but all attempts at cyclisation led to the anti THF 209 as the major product.

674

29 Enzymes

OTs OTs

F

F

OTs

OH

F O

OH N

F

F

N

N 208; >98% ee

207

N

N

F

209; 60:40 anti:syn

N

A more positive discrimination between the two primary OHs might be made if a reagent-dominated reaction were used and an enzyme is ideal for this job. Various lipases could be used either in the acylation of the triol 210 or in the hydrolysis of the diester 212. In either case, the required monoester 211 could be formed in excellent ee. Novozyme (Novo Nordisk) SP 435 catalyses the monoacetylation of 210 in MeCN solution to give 211; R = Me in 95% yield and 96.6% ee after only 55 minutes.

OH

F

OH

F

N

OCOR

F OH

OH

OH F

OCOR

OCOR

OH

N

N

F

N 210; >98% ee

211

N

N

F N

212

N N

A simpler approach gave even better results. The prochiral diol 213 was cleanly monoacetylated by the same enzyme (98% yield, 98% ee) and iodo-etherification gave the required diastereoselectivity 215 with suitable functionality for the introduction both of the triazole and of the other side chain.57 OAc OH

F

OAc

F OH

OAc

F

OH

SP435

F

F I2

O I

F 214

213

215

Bristol-Meyers Squibb anti-psychotic agent BMS 181100 A related though simpler compound from Bristol-Meyers Squibb 216 is their anti-psychotic agent BMS 181100. It has two p-fluorophenyl groups joined through a C4-piperazine linker containing a chiral secondary alcohol. The compound is easily made58 by alkylation of an amine with the primary chloride 217. There are two opportunities from enzymatic resolution. F

F

OH

OH

N N

F

(R)-(+)-216; BMS 181100

F

+

Cl 217

N HN

29 Practical Asymmetric Synthesis with Enzymes

675

Racemic 217 could readily be acetylated with isopropenyl acetate catalysed by the Amano lipase PS 30. The resolution was successful only in organic solvents and heptane was the best. Any water reduced both the rate of the reaction and the ee of the product. Acetylation occurred on the wanted (R) enantiomer 218 but the product could be hydrolysed to (R)-217 with the same enzyme with added water. The reactions were carried out on a 100 g scale using 200 g of immobilised enzyme that could be filtered off and used again and again with little diminution in ee. This sounds like a lot of enzyme but it is a polymer-supported enzyme so the bulk is polymer. OAc

OAc

H2O

(S)-217 +

(±)-217 Amano PS-30 heptane

F

Cl (R)-218; 95-98% ee

(R)-217 100% ee

Amano PS-30 heptane

Alternatively, the acetate 219 of the final product (±)-216 could be enantioselectively hydrolysed with another Amano lipase GC-20, chosen after a survey of fourteen enzymes. At this stage both enantiomers of the drug were needed for evaluation and the unreacted (S)-acetate could be hydrolysed to (S)-215 without racemisation. Again organic solvent was necessary: E was 1 in pure water and maximum (>100) in toluene or CH2Cl2. F

OAc

Amano lipase GC-20

N N

pH 7 toluene water

(±)-219

F

(R)-(+)-216 BMS 181100 + (S)-219 47.6% yield 97.9% ee

Na2CO3

(S)-216 H2O MeOH

An Eli Lilly protein kinase inhibitor Protein kinase inhibitors are in the news as cancer drugs but Eli Lilley have a saturosporinone analogue 220 to help diabetes patients overcome problems such as sight loss from retinopathy. The bis-indole portion 221 does not concern us but the chiral fragment 222 does. Me Me

N

O

O

N

O

O X

2x C–N

N

X

N N H

N H

O NMe2 220: protein kinase inhibitor

221

O [NMe2] 222 X = leaving group [NMe2] protected

The near-symmetry of 222 suggests that writing OH or OR for the functionality 223 would be helpful. This reveals a 1,6 relationship between the two OH groups that might be made by oxidative cleavage of a cyclohexene 224. This in turn could be derived from a Diels-Alder reaction

676

29 Enzymes

providing that the CH2OR group is first replaced by a carbonyl group 225. This ester might be resolved enzymatically. HO

OH

1

6 5 4

2 3O

1,6-diO reconnect

hetero Diels-Alder

FGI

O

O

O

OR 224

223

CO2R 226

CO2R 225

OR

The commercial solution of ethyl glyoxylate 226; R = Et in toluene was heated in an autoclave with butadiene to give the adduct 225; R = Et in 50-60% yield, easily purified as the polymeric by-products are soluble in toluene. A survey of various enzymes and formulations showed that cheap aqueous B. lentus protease gave the best results. Only 90 mls of a 5% solution of the enzyme was needed to hydrolyse 150 g of 225; R = Et in eight hours. Under these conditions E is 36.6 and 35–40% yield of (S)-225; R = Et having ee ⬎ 99.5% could be extracted with toluene. The free acid 225; R = H remains in the aqueous layer. The trityl-protected bis-mesylate 227 was produced in good yield on a multi-kilogram scale with 96.7% ee and used to make the drug59 220. B. lentus protease

O

pH 7 buffer NaOH, H 2O

CO2Et 225; R = Et

1. O 3 CH2Cl2

1. LiAlH 4

O CO2Et

2. TrCl DMAP pyridine

O

2. NaBH 4

OTr

225; R = Et

MsO

224; R = Tr

3. MsCl Et3N

OMs

O 227

OTr

Lotrafiban: a GlaxoSmithKline drug to prevent blood clots During the scale-up of an asymmetric synthesis of lotrafiban 228, a drug for the prevention of thrombotic ‘events’, racemisation proved to be a problem. The synthesis used the chiral pool strategy (chapter 23), the fragment in the frame of the intermediate 229 being derived from aspartic acid. O

Me O

N

Me N

HO2C

HO2C

HN

HN N NH2 O 228; SB 214857 Lotrafiban

CO2But 229

Since they already had a good synthesis of racemic ester, a late stage in the synthesis, an enzymatic hydrolysis looked a good bet. In the event hydrolysis with Candida antarctica lipase was superlative: hydrolysis gave only the required enantiomer and stopped cleanly at 50% conversion. Various supported versions were tried and the best was the Boehringer L-2 preparation in which the enzyme is covalently bound to a cross-linked resin making it stable in aqueous solution. In the pilot plant 76 kgm of racemic methyl ester 230 react in 2.5 hours with 6 kg of supported enzyme. Benzyl chloroformate in CH2Cl2 is added and the unreacted ester (R)-230,

29 References

677

now Cbz derivatised, remains in the CH2Cl2 layer (and is later recycled by racemisation) while the product is in the aqueous layer. The enzyme could be reused at least 100 times without loss of activity.60

Two desymmetrisations Me

O

N Candida antarctica lipase Boehringer L-2

MeO2C HN

(R)-230 + N NH O (±)-230

pH 6.2, 30 ˚C water 1.5 M NH3

(S)-228 42% yield 99% ee

We end this chapter with two much simpler reactions - a couple of desymmetrisations - each special in their own way. The first is amazingly efficient, using a lipase/esterase from a thermophilic organism supplied as a recombinant protein by the Diversa corporation. The symmetrical Diels-Alder adduct 231 is rapidly and perfectly desymmetrised to the monoacid61 232 at 70 ⬚C. CO2Me

CO2Me

ESPESL (Diversa)

231 phosphate buffer 70 ˚C

CO2Me

CO2H

232 98% yield >99% ee

The second is even more remarkable. The amlodipine type of dihydropyridine drugs are chiral but only just so. In some cases the only distinction is a methyl ester on one side and an ethyl ester on the other. The symmetrical compound 234 can be easily made by the Hantzsch pyridine synthesis and desymmetrisation looks as good as anything in attempting an asymmetric synthesis. Work from the Amano company using their own lipase AH established, almost incredibly, that one enantiomer is formed in i-Pr2O but the other in cyclohexane (both solvents are saturated with water).62 One problem with enzymes is generally that they can give only one enantiomer. With an unnatural substrate in unnatural media this may not be the case! NO2

NO2 Amano lipase AH

CO2i-Pr

HO2C Me

N Me H (S)-233

NO2 Amano lipase AH

i-PrO2C i-Pr2O saturated with water

Me

CO2i-Pr N Me H (±)-234

i-PrO2C cyclohexane saturated with water

Me

CO2H N Me H (R)-233

References General references are given on page 893 1. Collins, Chirality in Industry, K. Faber, Biotransformations in Organic Synthesis, Springer, 4th edition, 2000; Drauz and Waldmann, Enzyme Catalysis in Organic Synthesis. 2. Review: C. J. Sih, Angew. Chem., Int. Ed., 1984, 23, 570. 3. D. Seebach, M. A. Sutter, R. H. Weber, and M. F. Züger, Org. Synth., 1985, 63, 1.

678

29 Enzymes

4. B. Wipf, E. Kupfer, R. Bertazzi and H. G. W. Leuenberger, Helv. Chim. Acta, 1983, 66, 485; see also V. Crocq, C. Masson, J. Winter, C. Richard, G. Lemaitre, J. Lenay, M. Vivat, J. Buendia and D. Prat, Org. Process Res. Dev., 1997, 1, 2. 5. T. Kometani, H. Yoshii, and R. Matsuno, J. Mol. Cat., B: Enzymatic, 1996, 1, 45. 6. Review: C. Carrea and S. Riva, Angew. Chem., Int. Ed., 2000, 39, 2226. 7. S. H. Wu, Z.-W. Guo and C. J. Sih, J. Am. Chem. Soc., 1990, 112, 1990. 8. I. Kumar, K. Manju and R. S. Jolly, Tetrahedron: Asymmetry, 2001, 12, 1431. 9. Y. Lu, C. Miet, N. Kunesch and J. Poisson, Tetrahedron: Asymmetry, 1990, 1, 703. 10. G. Fülling and C. J. Sih, J. Am. Chem. Soc., 1987, 109, 2845. 11. Review: E. Santaniello, P. Ferraboschi, P. Grisenti and A. Manzocchi, Chem. Rev., 1992, 92, 1071. 12. work of M. Ohno, reviewed by J. B. Jones, Tetrahedron, 1986, 42, 3351. 13. D. R. Deardorff and D. C. Myles, Org. Synth., 1989, 67, 114. 14. D. R. Deardorff, C. Q. Windham and C. L. Craney, Org. Synth., 1996, 73, 25. 15. L. A. Paquette, M. J. Earle and G. F. Smith, Org. Synth., 1996, 73, 36; L. A. Paquette and T. M. Heidelbaugh, Org. Synth., 1996, 73, 44. 16. J. A. López-Pelegrín, P. Wentworth, F. Sieber, W. A. Metz and K. M. Janda, J. Org. Chem., 2000, 65, 8527. 17. R. F. Borch, M. D. Berstein and H. D. Durst, J. Am. Chem. Soc., 1971, 93, 2897. 18. M. Ohno, S. Kobayashi, T. Iimori, Y.-F. Wang and T. Izawa, J. Am. Chem. Soc., 1981, 103, 2405; T. Iimori, Y. Takahashi, T. Izawa, S. Kobayashi and M. Ohno, J. Am. Chem. Soc., 1983, 105, 1659. 19. K. Drauz and H. Waldmann, Enzyme Catalysis in Organic Synthesis, Vol I, 1995, p 178-195. 20. D. T. Guranda, L. M. van Langen, F. van Rantwijk, R. A. Sheldon and V. K. Svedas, Tetrahedron: Asymmetry, 2001, 12, 1645. 21. D. T. Gibson, M. Hemsley, H. Yoshioka and T. J. Mabry, Biochemistry, 1970, 9, 1626; Reviews: G. N. Sheldrake in Collins, Chirality in Industry, Vol I, 1992, p127; T. Hudlicky, D. Gonzalez and D. T. Gibson, Aldrichim. Acta, 1999, 32, 35. 22. S. V. Ley and F. Sternfeld, Tetrahedron, 1989, 45, 3463; S. V. Ley, M. Parra, A. J. Redgrave and F. Sternfeld, Tetrahedron, 1990, 46, 3463; T. Hudlicky, H. Luna, G. Barbieri and L. D. Kwart, J. Am. Chem. Soc., 1988, 110, 4995. 23. T. Hudlicky, M. R. Stabile, D. T. Gibson and G. M. Whited, Org. Synth., 1999, 76, 77. 24. T. Hudlicky, J. D. Price, F. Rulin and T. Tsunoda, J. Am. Chem. Soc., 1990, 112, 9439. 25. D. R. Boyd, R. A. S. McMordie, H. P. Porter, H. Dalton, R. O. Jenkins and O. W. Howarth, J. Chem. Soc., Chem. Commun., 1987, 1722. 26. A. M. Reiner and G. D. Hegeman, Biochemistry, 1971, 10, 2530. 27, D. T. Gibson, J. R. Koch and R. E. Kallio, J. Biol. Chem., 1989, 264, 14940; V. B. Bui, T. Hudlicky, T. V. Hansen and Y. Stenstrom, Tetrahedron Lett., 2002, 43, 2839. 28. D. Gonzalez, T. Martinot and T. Hudlicky, Tetrahedron Lett., 1999, 40, 3077. 29. S. W. May and R. D. Schwartz, J. Am. Chem. Soc., 1974, 96, 4031. 30. K. Furuhashi in Collins, Chirality in Industry, Vol I, 1992, p167, Drauz and Waldmann, Vol II, 1995, p 686. 31. Drauz and Waldmann, Vol II, 1992, p 755. 32. J. M. Schwab, W.-B. Li, and L. P. Thomas, J. Am. Chem. Soc., 1983, 105, 4800. 33. V. Alphand, A. Archelas, and R. Furstoss, J. Org. Chem., 1990, 55, 347; Biocatalysis, 1990, 3, 73. 34. P. A. Bird, C. A. Sharp and J. M. Woodley, Org. Process Res. Dev., 2002, 6, 569. 35. M. M. Kayser, G. Chen and J. D. Stewart, J. Org. Chem., 1998, 63, 7103. 36. M. D. Mihovilovic, G. Chen, S. Wang, B. Kyte, F. Rochon, M. M. Kayser and J. D. Stewart, J. Org. Chem., 2001, 66, 733. 37. ‘Bitter Almonds,’ In the Teeth of the Evidence, Dorothy L. Sayers, Gollancz, 1939. 38. F. Effenberger, T. Ziegler and S. Förster, Angew. Chem., Int. Ed., 1987, 26, 458. 39. F. Effenberger, B. Hörsch, S. Förster and T. Ziegler, Tetrahedron Lett., 1990, 31, 1249; U. Niedermeyer and M.-R. Kula, Angew. Chem., Int. Ed., 1990, 29, 386. 40. N. Klempier, H. Griengl and M. Hayn, Tetrahedron Lett., 1993, 34, 4769. 41. J. Brussee, E. C. Roos and A. van der Gen, Tetrahedron Lett., 1988, 29, 4485; P. Zandberger, J. van der Linden, J. Brussee and A. van der Gen, Synth. Commun., 1991, 21, 1387.

29 References

679

42. B. B. Corson, R. A. Dodge, S. A. Harris and J. S. Yeaw, Org. Synth. Coll., 1932, 1, 336; C. R. Noe, Chem. Ber., 1982, 115, 1591. 43. J. Brussee, F. Dofferhoff, C. G. Kruse and A. van der Gen, Tetrahedron, 1990, 46, 1653. 44. J. M. Schwab, W.-B. Li, and L. P. Thomas, J. Am. Chem. Soc., 1983, 105, 4800. 45. R. R. Hung, J. A. Straub and G. M. Whitesides, J. Org. Chem., 1991, 56, 3849; see also K. K.-C. Liu, T. Kajimoto, L. Chen, Z. Zhong, Y. Ichikawa and C.-H. Wong, J. Org. Chem., 1991, 56, 6280; T. Kajimoto, K. K.-C. Liu, R. L. Pedersen, Z. Zhong, Y. Ichikawa, J. A. Porco and C.-H. Wong, J. Am. Chem. Soc., 1991, 113, 6187. 46. M. Schultz, H. Waldmann, W. Vogt and H. Kunz, Tetrahedron Lett., 1990, 31, 867. 47. R. Chênevert and M. Dasser, J. Org. Chem., 2000, 65, 4529. 48. J. Liu and C.-H. Wong, Angew. Chem., Int. Ed., 2002, 41, 1404. 49. G. Carrea and S. Riva, Angew. Chem. Int. Ed., 2000, 39, 2226. 50. V. Gotor, Org. Process Res. Dev., 2002, 6, 420. 51. J. Saunders, Top Drugs, page 23. 52. A. Gentile, C. Giordano, C. Fuganti, L. Ghirotto and S. Servi, J. Org. Chem., 1992, 57, 6635. 53. L. T. Kanerva and O. Sundholm, J. Chem. Soc., Perkin Trans. 1, 1993, 1385. 54. J. Martel in Collins, Chirality in Industry, Vol 1, chapter 4, page 87. 55. A. Fishman and M. Zviely, Tetrahedron: Asymmetry, 1998, 9, 107. 56. J. Roos, U. Stelzer and F. Effenberger, Tetrahedron: Asymmetry, 1998, 9, 1043. 57. B. Morgan, D. R. Dodds, A. Zaks, D. R. Andrews and R. Klesse, J. Org. Chem., 1997, 62, 7736. 58. R. L. Hanson, A. Banerjee, F. T. Comezoglu, K. D. Mirfakhrae, R. N. Patel and L. J. Szarka, Tetrahedron: Asymmetry, 1994, 5, 1925. 59. J.-C. Caille, C. K. Govindan, H. Junga, J. Lalonde and Y. Yao, Org. Process Res. Dev., 2002, 6, 471. 60. T. C. Walsgrove, L. Powell and A. Wells, Org. Process Res. Dev., 2002, 6, 488. 61. Y. Kashima, J. Lui, S. Takenami and S. Niwayama, Tetrahedron: Asymmetry, 2002, 13, 953. 62. Y. Hirose, K. Kariya, I. Sasaki, Y. Kurono, H. Ebiike and K. Achiwa, Tetrahedron Lett., 1992, 33, 7157.

30

New Chiral Centres from Old — Enantiomerically Pure Compounds & Sophisticated Syntheses — The Purpose of this Chapter New Chiral Centres from Old Creating New Chiral Centres with Cyclic Compounds Conformational control Cyclisation: ring-closing reactions Thermodynamic control in formation of cyclic compounds Cyclic intermediates Stereochemical control in folded molecules Stereochemical Transmission by Cyclic Transition States Sigmatropic rearrangements Mapp and Heathcock’s synthesis of myxalamide A Control of Open Chain Stereochemistry Felkin-Anh control Houk control Reactions of allyl silanes with electrophiles New chiral centres by 1,3-control Aldol reactions 1,3-Anti induction Introduction to 1,4- 1,5- and Remote Induction Remote control by substrate-controlled induction 1,4 -Syn induction in the aldol reaction 1,4-Control by sigmatropic shifts 1,5-Induction by the aldol reaction The Asymmetric Synthesis of (⫹)-Discodermolide Making discodermolide on a large scale

The Purpose of this Chapter We have already seen in earlier chapters how one chiral centre can lead to another. We have seen this in the context of racemic compounds and in the context of optically pure ones. So for instance, new chiral centres are made when unsaturated acid (±)-2 reacts to form the iodolactone 1. If the acid 2 is racemic, both enantiomers of one diastereoisomer of the lactone 1 are formed but if 2 is optically pure, only one is formed. The existing chiral centre in 2 controls the two new ones in 1. The relative stereochemistry in the racemic product (±)-1 is exactly the same if we start with optically pure material.

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

682

30 New Chiral Centres from Old

O

O

O O

O

CO2H

I2

+

CO2H

NaHCO3

I

O I2 NaHCO3

I (±)-1

(±)-2

I

optically pure 2

optically pure 1

There is a clear overlap in concept between this chapter and chapter 21 but the purpose of this chapter is to explore how new chiral centres are made from old ones in some more sophisticated molecules, to look at some strategies we have not seen before, and to see how things are different when the starting material is a single enantiomer. We shall revisit the principles of chapter 21 so return there if you are in doubt about any of them. We shall use the asymmetric methods of chapters 22–29 and consider what advantage there is in having both substrate and reagent as single enantiomers. But first we shall look at an application of iodolactonisation in asymmetric synthesis. We pointed out in chapter 27 that Schultz’s asymmetric Birch reduction can be developed with iodolactonisation to remove the chiral auxiliary and set up new chiral centres. Now we shall see how he applied that method to alkaloid synthesis.1 The fi rst reaction is the same as in chapter 27 but the alkyl halide is now specified: this gave diastereomerically pure acetate in 96% yield and hydrolysis gave the alcohol 4. Mitsunobu conversion of OH to azide and enol ether hydrolysis gave 5, the substrate for the iodolactonisation. Iodolactonisation not only introduces two new chiral centres but cleaves the chiral auxiliary, as described in chapter 27. Reduction of the azide 6 to the amine with Ph3P leads to the imine 7 by spontaneous ring closure. OH OMe

O

2.

N

1. DEAD, Ph 3P

OAc

N

Br 3. KOH, MeOH

OMe 3 N3

(PhO)2PO.N 3 2. HCl, MeOH

OMe 4 O

OMe

O

OMe

O

1. K, NH 3, t-BuOH

O

I

I2

O

I

Ph3P

O

N THF H2O

O

THF reflux

O

5

N3

N

6

7

Acylation with 8 moves the imine to an enamine 9 and opening the lactone with the lithium alkoxide of benzyl alcohol gives the epoxide 10 ready for the key cyclisation. O Br

Br

O

7

Cl

O O 8

Et3N CH2Cl2

O

I

BnOH

N

O O 9

CO2Bn

Br

O

O

BuLi THF

O N

O O 10

30 Creating New Chiral Centres with Cyclic Compounds

683

This synthesis was designed around a radical cyclisation to close a six-membered ring and create yet another two new chiral centres. The radical may be induced either by Bu3SnH or photochemically and cyclises in a 6-endo-trig manner from the top face of the enamine giving cis 5/6 and trans 6/6 ring junctions. The original six-membered ring that was the benzene ring in 3 now has five chiral centres. One was created in the original asymmetric Birch reduction but the other four were introduced two by two using iodolactonisation and cyclisation as diastereoselective reactions. Product 12 has the skeleton and most of the functionality of the lycorine alkaloids. O

O CO2Bn

Bu3SnH

H

O

10

CO2Bn

O

AIBN

H

N

O

O

O 11

N

O 12

New Chiral Centres from Old The starting materials in this chapter will already be single enantiomers and will have been made so by one of the methods we have discussed at length over the last eight chapters: Resolution (chapter 22), Chiral Pool (23), Asymmetric Reagents (24) Asymmetric Catalysis (25 and 26), Substrate-Based Strategy (27), Kinetic Resolution (28) or Enzymes (29). These strategies will be identified by chapter titles and you are invited to check back if such descriptions are obscure or to check the references for more detail. The starting materials will generally have one or two chiral centres and methods for transmitting this information across the molecule to form new chiral centres will be based on the discussion in chapter 21 and will involve: • • • •

Use of Rings: preformed rings, conformational control, ring formation, folded molecules. Cyclic Transition States: chelation control, reagent delivery, rearrangements. Open Chain Stereochemistry: 1,2-control by Felkin-Anh and Houk transition states. More Distant Control: 1,3-, 1,4- and remote control.

It will not be possible to separate these methods totally as many syntheses involve several of them sequentially or together but we shall usually introduce an area with a simple example to set the scene.

Creating New Chiral Centres with Cyclic Compounds Rings have fewer conformations than open chains. This is as important in transmitting asymmetry as it was in chapters 20–29 in creating asymmetry. A simple illustration is Bailey’s synthesis of cis-5-hydroxypipecolic acid 15 by a chiral pool strategy from natural (S)-glutamic acid 13. Bailey’s original idea was to cyclise the α-chloroketone 14 but this was not successful, even when NH2 and CO2H were protected.2 O

HO2C H2N

HO ?

CO2H

13; ( S)-glutamic acid

Cl

H2N 14

CO2H

N CO2H H 15; cis-5-hydroxy pipecolic acid

684

30 New Chiral Centres from Old

The corresponding protected alcohol 17 would cyclise but reduction of 16, though giving a good yield, was totally non-stereoselective, giving cis- and trans-17 in a 1:1 ratio. This is not surprising: drawn out as a straight chain compound 16 has a 1,4-relationship between the existing chiral centre and the prochiral carbonyl group. This is usually too far for effective induction. O Cl

OH 2

1

3

4 CO2Me

NaBH4

OH

Cl

CO2Me + Cl

CO2Me

MeOH

NHCbz 16

cis-17

NHCbz 96% yield; 1:1

NHCbz

trans-17

Cyclisation of the mixed (silylated) alcohols 17 and oxidation to the ketone 18 allowed a comparison between reduction of the open chain 16 and cyclic 18 ketones. Reduction of 18 gave just one diastereoisomer cis-19 and so, after deprotection, the target molecule. Later improvements allowed a shorter route to 18 by rhodium-catalysed carbene cyclisation.3

17 1:1 cistrans

NaBH4

N

3. CrO 3, acetone 4. HF, H 2O, MeCN

CO2Me

CO2Me

1. R 3SiCl, base 2. NaH, DMF

O

Cbz

MeOH

HO

N

1. NaOH H2O, MeOH

15 Cbz

2. H 2, Pd/C

cis-19; 93% from 18

18

The high selectivity is due to the single conformation of 18 in contrast to the many conformations available for 16. The CO2Me group is axial to avoid eclipsing the planar Cbz group that is conjugated with the planar nitrogen atom. Reduction with the small reducing agent NaBH4 gives, as usual, the equatorial alcohol 19. Conformations of six-membered rings are particularly well defined but all rings have better defined conformations than the corresponding open chain compounds.

Conformational control In chapter 27 we recommended 8-phenylmenthol 21 as a chiral auxiliary. The three chiral centres in 8-phenylmenthol are made by from the terpene pulegone 22 (chiral pool) having only one chiral centre. The new chiral centres have 1,3- and 1,4-relationships to the centre already present in pulegone and the key to their successful control is conformational analysis. 8-Phenylmenthol has all three substituents equatorial.

OH OH 20; (–)-menthol

Ph

Ph OH

21; (–)-8-phenylmenthol

O conformation of 21

22; (+)-pulegone

Copper-catalysed conjugate addition of PhMgBr to pulegone gives the enolate 23 with no new chiral centres but protonation gives a 55:45 mixture of the cis- and trans-ketones 24. This is disappointing as we might have expected greater selectivity for the all-equatorial trans-24. In fact equilibrium was not achieved in acid and a separate treatment with KOH in ethanol under reflux converts the 55:45 mixture into an 87:13 mixture.4

30 Creating New Chiral Centres with Cyclic Compounds

PhMgBr

O 22; (+)-pulegone

685

2M HCl

CuBr Et2O –20 ˚C

+ Ph

Ph

Ph

O

O

trans-24

23

O

cis-24

The mixture is reduced with sodium in i-PrOH to give the same proportions (87:13) of two alcohols 21 and 25. This thermodynamically controlled reduction is totally stereospecific, giving the equatorial alcohol in both cases. Presumably the equilibration of the ketone 24 is less successful because the large PhCMe2⫺ group eclipses the carbonyl group when it is equatorial. The two alcohols are easily separated as their chloroacetyl esters.

Na

KOH

trans-24 cis-24

Ph

EtOH reflux

i-PrOH

O

+ Ph

trans-24 (87:13 mixture with cis-24)

Ph

OH

OH

21

25

Cyclisation: ring-closing reactions Schultz’s lycorine 12 synthesis started with the creation of asymmetry (one centre) in one ring and built two other rings onto that ring by cyclisation reactions. This is a good strategy and can be applied to much simpler molecules such as Merck’s anti-schizophrenia drug MGS002836. The first centre was introduced by asymmetric allylic alkylation of the palladium allyl cation complex 28 from the racemic acetate 26 using the Trost ligand 29 to provide the asymmetry.5 OAc +

Pd

26

Cl Pd Cl

O

L*Pd

O HN

NH

27

28

29; L* Trost ligand

PPh2 Ph2P

The nucleophile is the enolate anion from ethyl 2-fluoroacetate 30 formed with NaH in the presence of n-Hex4N⫹Br⫺. The complex 28 would be achiral except for the ligand which directs the nucleophile to one end of the allyl cation on the opposite face to the palladium. With only 1 mol% of the catalyst up to 96% ee was achieved though only one centre in 31 is controlled. O OAc

F

O CO2Et

1 mol% Allyl-Pd-Cl Trost ligand 29

+ 26

CO2Et H

F 30

NaH, n-Hex 4N

Br

31 89% yield 94-96% ee

This is not important as deacylation destroys that centre anyway 32. Epoxidation followed: the best route was bromohydrin formation with NBS in water followed by elimination with DBU and gave an 8:1 selectivity for epoxidation on the opposite face to the substituent 33.

686

30 New Chiral Centres from Old

O F

H

F

CO2Et NaOEt

H

CO2Et

1. NBS, H 2O, acetone

H

H

2. DBU, CH 2Cl2

EtOH

O

32; 100% yield

31

H

F

CO2Et

33; 85% yield (8:1)

Now comes the critical cyclisation: treatment with a stronger base (LiHDMS) creates the prochiral enolate centre which cyclises onto the epoxide. The three centres on the ring are controlled: two are not involved and the third is stereospecifically inverted by the intramolecular SN2. The new fluorinated centre on the three-membered ring 34 is controlled stereoselectively with the larger CO2Et group on the outside (convex or exo face of the folded molecule - see below). Further reactions involve the oxidation of the alcohol 34, bromination of the ketone, again on the outside of the folded molecule 35, and introduction of an amino acid to give Merck’s potential schizophrenia drug 36. Each chiral centre was developed from the original asymmetric allylation. F

H H

CO2Et

F

CO2Et

CO2Et

F

H LiN(SiMe3)2

O

CO2H

H

O

NH2

O

34; 90% yield

33

H

H

2. Br 2, CH 2Cl2 (100%)

HO

CO2H

H 1. IBX (96%)

H

Et3Al, THF –78 ˚C

F

35

Br

36; MGS0028

Thermodynamic control in formation of cyclic compounds The rather complex anti-fungal compound itraconazole (Sopranox) 37 has two triazole rings each substituted with chiral units. They are very far apart and each must be made separately.6 37; Itraconazole (Sporanox) Anti-fungal

Cl

Cl N N

O

O O

N

O

N

N

N

N N

Chemists at Sepracor appreciated that the left hand unit 38, being an acetal, could be made only by the usual acetal formation from a ketone 39 and a diol 40. Note that the diol is precariously chiral as if R ⫽ H, the compound has a plane of symmetry. Cl

Cl

1,1-diX

Cl

Cl O

N N

N

O

trans-38

HO

OR

+

acetal

N OTs

N

N

O 39

HO 40

30 Creating New Chiral Centres with Cyclic Compounds

687

Diol derivatives 40 can be made from protected glyceraldehyde (chiral pool strategy) and the obvious choice for R is Ts so that the new C–O bond needed for itraconazole can be made by alkylation of the phenol. Reaction of the diol 40; R ⫽ Ts with the ketone 39 in acid solution gave a mixture of trans- and cis-38 in an acceptable 3:1 ratio. The preferred diastereoisomer has the benzene ring on the opposite side to the CH2OH group. The triazole ring is further away and so less hindering. Trans-38 can easily be separated from cis-38 by crystallisation and the overall yield (51%) from 41 is excellent. Cis-38 can of course be recycled as acetal formation is reversible: an advantage of thermodynamic control. CHO

O

HO

1. NaBH 4

OTs

2. TsCl, pyridine

O

3. HCl, acetone

HO 40; R = Ts

41

39

crystallise HOTs salt

trans-38 + cis-38

trans-38

H

51% yield from 41

3:1 ratio

Cyclic intermediates The conversion of prochiral centres in cyclic intermediates into fully blown chiral centres is a profitable way to transmit stereochemical information. The ant trail pheromone monomorine 42 when re-drawn so that we can see all three hydrogen atoms on the same surface of the molecule 42a, is an ideal candidate for synthesis by reductive amination. One of the three centres must be made asymetrically and then the other two set up by reduction of imines formed from that amine and two ketones as in 43. H

H reductive amination

N

N H

42; monomorine

O H Bu

Bu

O

H2N

42a

43

The synthesis we are going to describe is particularly ingenious. A Stetter reaction (chapter 14) with the Diels-Alder adduct 44 (it doesn’t matter that this is racemic and a mixture of exo/endo isomers) and the enone 45 provides the 1,4-diketone in 85% yield and hence, by retro-Diels-Alder, the achiral enedione 47 in 81% yield.7

O

O

Stetter

O heat

+ thiazolium salt

Bu CHO 44

O

O 46

45

47

The amine 49, made from one enantiomer of propylene oxide by simple reactions, is the source of the chirality. A cross-metathesis with 47, using the Grubbs catalyst 50 gave a surprisingly excellent yield (89%) of the complete carbon skeleton 49 of monomorine. O

MesN Bu

47

O propylene oxide

NHCbz 48

Grubbs catalyst

NHCbz 49

O

Cl Ru Cl O

NMes Grubbs catalyst 50

688

30 New Chiral Centres from Old

Catalytic hydrogenation of 49 produces monomorine in one step and in 75% yield. First the alkene must be reduced to give the molecule the flexibility to cyclise. Hydrogenation removes the Cbz group and the amine cyclises onto the nearer ketone to give the imine 51 which is reduced from the less hindered face opposite the Me group to give 52. Now cyclisation to the second ketone produces the imine salt 53 and a second reduction from the same side gives (⫹)-monomorine. The first reduction of 51 is totally stereoselective but the second gives about 15% of the other epimer. Since the origin of the stereochemistry is only a methyl group, this is remarkably good. H

H H2, Pd/C

43

N

MeOH

NH

Bu

O 51

(+)-42 75% yield

N Bu

O

H

Bu

52

53

Other prochiral units that can be trapped in rings are enolates. One famous application is the alkylation of β-hydroxy acid derivatives available from the chiral pool (chapter 23), by asymmetric aldol reactions (chapter 27) and by asymmetric reduction of β-keto-esters either by catalytic hydrogenation (chapter 26) or by enzymes (chapter 29). Fráter found that the alcohol 55, from the baker’s yeast reduction of the ketoester 54, formed the enolate 56 held in shape by chelation. Alkylation occurred on the top face.8 This is not so much because the OH is down in 55 as that the methyl group is down in 56. O

OH

baker's yeast

O

CO2Et

Li

O

OH CO2Et

RBr

2 x LDA

CO2Et

54

Li

OEt

55

R 57

56

Seebach has done similar work with malic acid. A detailed account9 of the allylation of the enolate 59 formed preferentially on the unsubstituted carbon atom with allyl bromide giving a 92:8 ratio of diastereoisomers 60 is in Organic Syntheses.

OH EtO2C

Li

O

O

OH

Li

Br

EtO2C

58; diethyl malate

CO2Et

EtO2C

CO2Et OEt

60; 70% yield

59

Fráter quotes an interesting example 62 where NaBH4 reduction gave the all equatorial trans61 (racemic of course) while baker’s yeast reduction gives one enantiomer of the cis compound 61 evidently by reagent controlled dynamic kinetic resolution (chapter 28). Allylation of the enolate of cis-61 goes with the expected anti selectivity.10 O

OH CO2Et

NaBH4

CO2Et

baker's yeast

OH

OH CO2Et 2.

(±)-trans-61

62

(+)-cis-61

CO2Et

1. LDA Br

63; 95% yield

30 Creating New Chiral Centres with Cyclic Compounds

689

If the β-hydroxy acid derivative is a lactone, the chain is turned the other way and no chelation is possible. The lactone 65 can be prepared from malic acid 64 (either enantiomer) and the dilithium derivative 66 cannot be chelated. Instead the solvated (THF is the solvent) LiO group blocks the top face of the almost planar enolate. The selectivity is anti as before but when the chain is opened out 68 by hydrolysis of the lactone it is clear that the other diastereoisomeric relationship to that in 60 has been made.11 O

O

OLi

R

R

OH CO2H

HO2C

O HO

64; ( S)-(–)-malic acid

THF –78 ˚C

CO2H

RBr

2 x LDA

O

O 66

65

HO

OH

HO

LiO

67

68

We shall illustrate this with Prestwich’s synthesis11 of the natural anti-cancer compound (⫺)aplysistatin 69. This interesting compound with a seven-membered cyclic ether and a bromine atom proclaiming its marine origin is probably made in nature by a polyolefin cyclisation and that strategy was used in the synthesis. The alkene must be removed 70 before the cyclisation can be reversed. The stereochemistry of the extra centre in 70 is unimportant until the realisation that it can be made by alkylation of a hydroxylactone 71 whose stereochemistry must be anti. H

H

O

O

O

FGI

H

polyolefin cyclisation

O

HO O

O

Br

H

Br

H

O H

H

69; (–)-aplysistatin

70

O

71

The (R)-(⫹)-isomer of malic acid 72 must be used and the lactone 73 is the enantiomer of 65. The alkylating agent is the terpene derivative homogeranyl iodide and the selectivity is anti as expected. The yields in this synthesis are not wonderful but it is short and biomimetic. O 1. 2 x LDA, THF, –78 ˚C

OH CO2H

HO2C

O

71 2. I

HO

72; ( R)-(+)-malic acid

73

The cyclisation uses the convenient brominating agent 74 and the preferred conformation for cyclisation has an all chair arrangement of the side chain 75. By this means the two chiral centres in 71 induce three more in 70. The remaining double bond is introduced by selenium chemistry (chapter 33). In this process the chiral centre introduced by alkylation of 73 is destroyed but it has played its part in controlling the conformation 75. O Br

Br

OH

H

H

O

"Br "

O

O

74 O Br

Br

75; preferred chair-like transition state for cyclisation

Br

H H

anti-70

O

690

30 New Chiral Centres from Old

Stereochemical control in folded molecules When two small rings are fused together, the molecule is folded rather like a half-opened book. When more rings are added, the molecule may adopt a bowl shape.12 In both cases the molecule has a convex or exo face on the outside that reagents find easy to reach and a concave or endo face on the inside that is much more difficult to reach. Astericanolide 76 has two five- and one eight-membered rings and is bowl-shaped as all four hydrogen atoms at the ring junctions are on the same side. Paquette13 decided to close the eight-membered ring by ring-closing metathesis (RCM, chapter 15) and so removed the ketone, the remote chiral centre (C-7) and the lactone carbonyl 77 as a preliminary. Disconnection gave 78 as the substrate for metathesis, the two arrowed carbons being destined for cyclisation.

O

7

H

H

H

FGI

H

H

H

H

RCM

H

O O H 76; (+)-astericanolide

O

O

H

H

77

78

The 5/5 fused system in 78 is folded but has the side chain on the inside: it may be possible to force it inside by adding the hydrogen atom last. The synthesis starts with a vinyl-lithium reagent derived from 79 (chapter 16): reaction with a single enantiomer of menthyl p-tolylsulfinate gave the sulfoxide 80. Michael addition of the propargyl alcohol 81 was not very efficient but could be persuaded to give one diastereoisomer of the adduct 82 in 38% yield. As we shall see in chapter 36, this sacrifice in yield is worthwhile when a tandem reaction occurs. Here Michael addition of the alcohol has been followed by Michael addition of the resulting anion to the acetylenic ester. O Br

O

O

1. BuLi

CO2Me

O S

O

Tol

2. R*SO 2Tol

O K2CO3

+

CO2Me

R*

THF

R* = menthyl

O HO 80; 77% yield

79

H 81

82

Hydrogenation with Raney nickel removes the chiral auxiliary and, as we had hoped, hydrogenates the alkene from the less hindered exo-face, pushing the side chain inside the folded molecule 83. The ketone is converted into the only possible vinyl triflate for Stille coupling with a vinyl stannane to give the diene 84. O

CO2Me

H

H

H2, Raney Ni

KHMDS PhNTf2

TfO

CO2Me

H

H

CO2Me

H

H

Bu3Sn

82 MeOH

O H 83; 88% yield

THF, –78 ˚C

O H 98% yield

Pd2(dba)3 CHCl3 LiCl, THF

O H 84; 98% yield

The next few stages to produce the metathesis substrate 78 are straightforward and involve the creation of no new chiral centres. The metathesis was very successful with the original Grubbs’ catalyst 85 and gave the cyclic diene 86. Now the two carbonyl groups and the last two chiral

30 Stereochemical Transmission by Cyclic Transition States: Sigmatropic Rearrangements

691

centres must be added. Porphyrin-sensitised (TPP ⫽ tetraphenylporphine) photochemical addition of singlet oxygen by an ene reaction 87 occurred also on the exo-face and gave, after reduction of the hydroperoxide, the doubly allylic alcohol 88. H O HO

O

PCy3 Cl Ru Cl PCy3 Ph

78

O2 TPP

H H

H H

hν CH2Cl2

O H 86; 93% yield

85; Grubbs catalyst

H

LiAlH4

H

H

O

O H 88; 67% yield

87

This new centre does not appear in the natural product 76. It is oxidised to the ketone and both alkenes reduced catalytically, both hydrogen molecules being added from the exo-face to set up both chiral centres in one step. The THF is finally oxidised to the lactone with Ru(IV). After the initial Michael addition, controlled by the sulfoxide, every chiral centre depends on the folded or bowl-shaped molecule. H O

O

Dess-Martin periodinane

H

H

88

H2, Pd/C

H H

O RuCl3

H H

EtOH

O

O H 90; 67% from 88

89

H

NaIO4

H H

CCl4 MeCN H 2O

O H 76; 63% yield

O

Stereochemical Transmission by Cyclic Transition States: Sigmatropic Rearrangements Rearrangements are more commonly used to move chirality around a molecule without actually creating new centres. That new centres can be made is illustrated clearly by Mulzer’s Ireland-Claisen approach to indolizidine alkaloids14 such as (⫺)-petasinecine 98 from proline (chiral pool). Proline is converted into the allylic alcohol 91 which is acylated to give the substrate 92 for an Ireland-Claisen rearrangement. O

H OH N

Boc 91

O

H O

Ph

O

Cl pyridine

N

OBn

Boc 92; 98% yield

Treatment with LiHMDS [LiN(SiMe3) 2] gives the Z-enolate 93 because of chelation and hence the silyl enol ether 94 that undergoes the [3,3] sigmatropic rearrangement to give one diastereoisomer of 95 with two new chiral centres both controlled by the original one from proline.

692

30 New Chiral Centres from Old

Li Bn

Me3SiCl

BnO

O

H

LiHMDS

OSiMe3

O

O

N

N

Boc 93

BnO

O

H

OSiMe3

H

O

H

[3,3]

N

Boc

H Boc

94

95

Removing the Me3Si and Boc groups and redrawing the structure reveals the amino acid 96 that cyclises in acid solution to bicyclic 97 having the skeleton of the indolizidine alkaloids. Further steps lead to (⫺)-petasinecine 98. This sequence looks involved but in fact only three intermediates are isolated: 92, 96, and one between 97 and 98. H

OBn

CF3CO2H

95 BuOH

OH

H

H

OBn

N

CO2H

NH

O 97; 82% from 92

96

OH

N O 98

A more complicated example is the synthesis of (⫺)-hastanecine by Hart and Yang.15 (R)-(⫹)Malic acid 99 was used as a chiral pool starting material and converted into the anhydride 100 by straightforward steps. Reaction with the amine 101 gave the pyrrolidine 102.

HO2C

OH

OAc

O

1. amine 101 BnO

O

HO2C

NH2

2. AcCl, CH 2Cl2

BnO

O 100

99; ( R)malic acid

OAc

O N

O 102; 81% yield

Reduction of the anhydride gave 103 with complete regioselectivity in favour of the carbonyl next to the acetate but without stereoselectivity. This doesn’t matter as treatment with acid initiated a rearrangement - cyclisation sequence giving a the alcohol 104 after hydrolysis of its formate. This time there is complete stereoselectivity in the creation of three new chiral centres.

MeOH

H

OAc

1. HCO 2H, 25 ˚C

NaBH4

102

BnO

OAc

HO BnO

N

2. NaOH, H 2O, MeOH

O 103; 83% yield

HO

N O 104; 73% yield

Loss of water from 103 gives the iminium salt that undergoes the [3,3]-sigmatropic rearrangement 105 to give 106 with two of the new chiral centres already fixed. An ionic cyclisation 106 fixes the third and the cation 107 picks up water to give 104 or formic acid to give its formate ester.

30 Stereochemical Transmission by Cyclic Transition States: Sigmatropic Rearrangements

BnO

BnO

BnO

OAc –H2O

693

H

OAc

H

OAc

[3,3]

103 N

N

N O

O

O

105

106

107

The stereochemistry of these steps is of course controlled by the original chiral centre derived from malic acid. The acetate prefers to adopt a pseudo-equatorial position in the cation 105 and the conformational drawings 105a–107a show that a chair conformation for the [3,3] step with equatorial substituents gives the first two new chiral centres and a pseudo-equatorial position for the isopropyl substituent in the ionic cyclisation gives the third. H

H OBn

H OBn

[3,3]

N

OBn

N

N

OAc

OAc

H 105a

OAc

H

H 106a

H 107a

Mapp and Heathcock’s synthesis of myxalamide A These [3,3]s and the related [2,3] sigmatropic rearrangements were used in the synthesis of the far more complicated anti-fungal and anti-viral myxalamide A 108. At first sight the worst problem seems to be the chain of trans and cis alkenes dominating the structure. There is also some three-dimensional stereochemistry. At the right hand end is an amide of alaninol that will cause no problem but at the other end is a section with three chiral centres. This piece was disconnected at the first alkene and the necessary aldehyde functional group replaced by an alcohol 109 to avoid racemisation by enolisation. It is now clear that, while two of the chiral centres are adjacent, the third has a 1,4 relationship with the OH group and, worse, is separated by a trans double bond. At first sight it appears impossible to control this relationship. But look and wonder: a combination of [2,3] and [3,3] shifts succeeds.16 OH 4

O

aldol? Wittig?

NH HO

1

HO

FGI

108 myxalamide A

2

OH

109

The synthesis started with an Evans’ aldol between the phenylalanine-derived oxazolidinone 110 and the aldehyde 111. Reductive cleavage of the syn-aldol 112 gave the diol 113 in excellent yield. This type of aldol reaction is discussed in detail in chapter 27. You may be impressed by the near perfect stereoselectivity but unimpressed that 113 is the wrong diastereoisomer. Look and wonder. Bn

Bn 1. Bu 2BOTf, Et 3N

LiBH4

t-BuS

N

t-BuS

N

Et2O H2O

2.

O

O 110

t-BuS 111

CHO

OH 112

O

O

OH 113; 90% from 110

OH

694

30 New Chiral Centres from Old

Protection of the primary and acylation of the secondary alcohol prepares the way for an Ireland-Claisen rearrangement. The E-enolate is produced and the [3,3] sigmatropic rearrangement transmits the chirality across the alkene to set up two new centres. The mechanism of the Ireland-Claisen rearrangement is described above 94 and occurs suprafacially across the backbone of the molecule through a chair-like transition state. We hope you agree both with the relative stereochemistry of the new centres and the E stereochemistry of the new alkene. 1. LDA, THF 1. t-BuMe2SiCl Et3N, imidazole 88% yield

t-BuS

OR

113

3. room temperature

OR

MeO2C

O 2. EtCOCl pyridine 92% yield

2. t-BuMe2SiCl

4. K 2CO3, H 2O

t-BuS

5. CH 2N2

O 114; R = TBDMS

115; R = TBDMS, 86% yield

The ester group is transformed into an ethyl group in simple fashion 116 and the position and stereochemistry of all the chiral centres corrected by an Evans-Mislow rearrangement. Oxidation to the sulfoxide 118 allows a [2,3] sigmatropic shift across the top face of the molecule giving an unstable sulfenate ester 119 from which sulfur is removed by a phosphite ‘thiophile’ to give the alcohol 117 with the correct anti-stereochemistry. The alkene returns to its original position and stereochemistry. The rest of the synthesis involves the construction of the polyene chain.16 OR

1. mCPBA 2. (MeO) 3P MeOH

t-BuS 116; R = TBDMS mCPBA

OR OH 117; R = TBDMS MeOH

OR

[2,3]

OR

suprafacial

t-Bu

S

O

118; R = TBDMS

t-BuS

O P(OMe)3

119; R = TBDMS

Control of Open Chain Stereochemistry In chapter 21 we emphasise that the most reliable method of controlling one chiral centre by another in open chain compounds is by Felkin-Anh orbital control of nucleophilic attack on a carbonyl group next to a chiral centre. We shall start this discussion with that method and follow with the next most reliable - Houk control of electrophilic attack on alkenes also next to a chiral centre. These are of course both 1,2-control and we shall deal with that before discussing 1,3-, 1,4- and remote control.

Felkin-Anh control We start with simple but important example. The search for a manufacturing method for Janssen’s anti-cancer drug 120 as one enantiomer of one diastereoisomer in high yield led to the racemic amino-ketone 123 easily made from 121 via a Friedel-Crafts acylation and a substitution with dimethyl amine.17

30 Control of Open Chain Stereochemistry

695

N O

N Cl

N

Cl

N

NMe2 N H 120; Janssen's R116010 S

S

AlCl3, CH 2Cl2

N H 121

O

O Me2NH

N S

Cl

N H (±)-122

N

H2O i-PrOH

S

NMe2

N H (±)-123

Dynamic kinetic resolution (chapter 28) of 123 with (⫹)-di-p-toluoyltartaric acid equilibrated the easily enolisable chiral centre and precipitated a 90% yield of the salt of the (S)-amine having 80% ee. Reduction with NaBH4 in i-PrOH was very diastereoselective giving 99:1 syn:anti-124. O

OH NaBH4

N

(±)-123

i-PrOH NaOH

NMe2 N H (S)-123 as di-p-toluoyl-tartrate S

N NMe2 N H (S,S)-124; 80% yield, 80% ee S

This is Felkin-Anh control: the NMe2 group sits at right angles to the carbonyl group and nucleophilic hydride approaches at the Bürgi-Dunitz angle alongside the H atom 125. The synthesis is completed by displacement of OH with an excess of imidazole using CDI (carbonyl-di-imidazole) 126 to activate the alcohol. Participation by the NMe2 group is responsible for the retention of configuration. O

O

Me

(S,S)-124

Me2N

N Ar 125

H

N

N

N

(S,S)-120 imidazole EtOAc

126

Even one methyl group can be enough to give reasonable Felkin selectivity. A vital step in Chi-Huey Wong’s epothilone synthesis18 is the addition of the double enolate 128 to a single enantiomer of the syn,anti-aldehyde 129 to give good diastereoselectivity (8:1) at the new chiral centre C-5. Though this is an aldol reaction of sorts, it is not a stereoselective aldol in the usual sense (chapter 3) but Felkin control in attack on the α-chiral aldehyde 129. 1. NaH 2. BuLi

Ot-Bu

CO2t-Bu O 127

OLi O 128

5

CO2t-Bu OR

DERA an aldolase

CHO OR 129

OH

O

130; 70% yield, 8:1 diastereoisomers

696

30 New Chiral Centres from Old

Since Felkin control has been well established in chapter 21, we shall simply give a couple of examples and two warnings before moving on to the less familiar Houk control. The first warning is that chelation control, particularly when metals such as magnesium are involved, may reverse the normal sense of induction. The second is that Felkin (or chelation) control cannot be guaranteed even when there is a large electron-withdrawing group next to the carbonyl group. In the original synthesis19 of the HIV protease inhibitor viracept by Aguron Pharmaceuticals the key intermediate 133 was made from serine (chiral pool) 131 via the simple ketone 132. Reduction of 132 with NaBH4 gave mostly the stereochemistry required 133. The Felkin conformation is 134 with the large (and electronegative) NHCbz group orthogonal to the carbonyl group and nucleophilic hydride approaching alongside the small H atom. Yet the stereoselectivity is poor and this was not the route chosen for the manufacture of the drug.20 NH2

NHCbz

NHCbz

Cl

PhS

NaBH4

CO2H

Cl

OH

PhS

131; ( S)-serine

Cl THF H2O

O

PhS

132

NHCbz

OH

H

133

O 134

The anti-tumour antibiotic (⫹)-FR900482 135 is particularly interesting as it contains a hemiacetal. Disconnecting this reveals an eight-membered ring with stereochemistry 136. Martin21 decided to make this ring by metathesis so that a protected version of 137 becomes an intermediate. OCONH2

OH

OH

OHC

O

N

OCONH2 O

OH

OCONH2 O

OH

1,1-diX

NH

NH hemiacetal

OHC

N

OHC

N

HO 136

135; FR900482

HO 137

Now stereochemistry must be controlled in open chain chemistry. They were able to make a single enantiomer of 138 by desymmetrisation with a lipase (chapter 29). Reaction with vinyl Grignard gave one diastereoisomer of 139a whose silyl ether 139b underwent excellent metathesis with a Schrock Mo catalyst to give 140 and hence FR900482. They say disarmingly that ‘the stereochemical outcome…may be rationalised either…(by)…chelation control or by the FelkinAnh model’. Take your choice! It may appear that this new centre is unnecessary but it is in fact vital to control the formation of the aziridine. OMPM

OBn

CHO

OBn

Boc

138; MPM - p-MeOC6H4CH2

OMPM OR

OBn Schrock Mo metathesis catalyst

BrMg

N OBn

OMPM OR

N OBn

Boc 139a; R = H

139b; R = SiMe 2But

N OBn

Boc 140

30 Control of Open Chain Stereochemistry

697

Houk control Houk control concerns electrophilic attack on alkenes, enolates and the like. The alkylation of enolate 56 would be an example if it were not held in a ring by chelation. It can in fact be difficult to tell whether chelation is involved or not with many enolates and the outcome of the reaction may tell which. Chamberlin’s asymmetric preparation of both pyrrolidine 2,3dicarboxylic acids 141 and 142 from natural aspartic acid illustrates this perfectly. The key to the stereochemical control is the very large protecting group 9-phenylfluorenyl- 143 introduced by Rapoport.22 CO2H

CO2H

CO2H and

H2N

CO2H

N H

(S)-aspartic acid

CO2H

N H

141

CO2H Ph 143: PhFl group

142

Aspartic acid is protected as the dimethyl ester 144 with the nitrogen atom carrying one benzyl group and one PhFl- group. Potassium or lithium enolates, made with K/LiN(SiMe3)2, were alkylated with allyl iodide. The lithium enolate gave a 23:1 mixture (96% yield) favouring syn-145 but the potassium enolate gave a 10:1 mixture (94% yield) favouring anti-146. These were converted into 141 and 142 by ozonolysis, deprotection and reductive amination. CO2Me

CO2H Bn H2N

N

CO2Me

(S)-aspartic acid

N

I

PhFl

CO2Me +

Bn

2.

CO2H

CO2Me

1. MN(SiMe 3)2

Bn

CO2Me

PhFl syn-145

144

N

CO2Me

PhFl anti-146

The lithium enolate is expected to have the E(enolate) configuration 147. Chelation is impossible and it adopts the Houk conformation 147a with the H atom on the inside eclipsing the π-bond. The enormous protected amine forces the allyl bromide to the opposite face. The potassium enolate prefers the chelated structure 148 and the same group directs the allyl iodide to the bottom face. OLi OMe Bn

N

CO2Me

PhFl 147; E-Li enolate

O Li OMe

MeO2C

Bn

N

MeO PhFl

H FlPh

147a; Li enolate in Houk conformation

Bn

O K

N H

O

148; chelated potassium enolate

MeO

In these examples the sheer size of one of the groups controlled the transmission of stereochemical information. If the directing group is a silicon atom, electronic factors are important too.23 Alkylation of the enolate from 149 was used in a synthesis of tetrahydrolipstatin.24 There is no question of chelation here as no chelating group is present. The Houk conformation 151 of the lithium enolate has H inside and the Me3Si group directs the alkylation to the bottom face.

698

30 New Chiral Centres from Old

R2

CO2Bn R1

Me3Si

1. LDA

SiMe3

CO2Bn

2. n-C6H11I

R1 149;

R1

= n-C11H23

SiMe3

150;

R2

O Li OBn

H R1 151

= n-C6H11

Reactions of allyl silanes with electrophiles We shall see in this section that allyl silanes can direct the transfer of chirality through Houk conformations without the need to form an enolate. The typical reaction of an allyl silane with an electrophile (chapter 12) is at the remote atom of the alkene with loss of the silyl group. This transfers, but does not create, chirality. So the allyl silane 152 reacts with formaldehyde and a Lewis acid to give the homoallylic alcohol 153 with no loss of ee. The silyl group has gone, the alkene is transposed, and the sense of the SE2´ reaction is anti. All this is explained by the Houk conformation 154 with the C–Si bond able to interact with the alkene to raise the energy of the p-orbital and direct both the regio- and the stereoselectivity.25 Ph

CH2O

Ph

Me3Si

Me

TiCl4

SiMe3

H OH 153; 86% ee

152; 85% ee

Ph

154

To create new chirality we need a prochiral electrophile. Single enantiomers of functionalised allyl silanes 155 are made by kinetic resolution with a lipase.26 Reaction of 155 with an aldehyde and a Lewis acid gives the syn and anti homoallylic alcohols 156 and 157. CO2Me BnO

OH

CHO

BnO

CO2Me

OH

+ BnO

5

CO2Me 5

Et2O-BF3

SiMe3

SiMe3 syn-156

155

SiMe3 anti-157

With BF3, the syn product 156 predominates by 87:13 but with MgBr2, the anti 157 is even more dominant (92:8). Both products have the same stereochemistry at C-5, determined by the Me3Si group as in 154. The mechanism 158 forces the electrophile to approach the alkene from the top face - opposite the Me3Si group - but a detailed transition state such as 159 is needed to explain the new chiral centre at C-6. Panek suggests that 159 is the anti-peri-planar transition state and that MgBr2 adopts a ‘synclinal’ transition state. Me CO2Me

F3B O

SiMe3 R 158

H

O

syn-156 F3B

SiMe3

R H

H

R

159

The electrophile, the substitution pattern, and the silyl group can all be varied: acetals and amines combine to form imines such as 162 under Lewis acid catalysis and these react in the same

30 Control of Open Chain Stereochemistry

699

way with allyl silanes such as 160 to give good yields of carbamates 161 with excellent stereoselectivity.27 OMe

CO2Me

Ph

NHCO2Me CO2Me

Ph

H2N CO2Me Et2O.BF 3

SiMe2Ph

H

OMe

Ph

N

161; 94% yield >30:1

160

CO2Me

162

An impressive application is the total synthesis28 of the anti-tumour antibiotic (⫹)-macbecin I 163. The amide disconnection is obvious and the decision to make the amino quinone from an aromatic nitro compound 164 sensible. Less obvious is the synthon 165 as the anion at C-14 could be an allyl silane but the further disconnections of 165 all correspond to allyl silane chemistry. OMe

O O MeO

1

MeO

15

N H

O

1

NO2

OMe OMe 164

O

OMe

12

HO2C

15

OMe 14

O

7

MeO

OH

12

5

7

NH2 OMe 165

163; (+)-macbecin I

Three reactions of single enantiomers of allyl silanes were used to add each section of 165. First the allyl silane 166 was coupled to the acetal 164 with Me3SiOTf as the Lewis acid. The silyl group at C-12 controls the stereochemistry of both new chiral centres and the new alkene. The product is transformed into the aldehyde 168 for the next reaction. OBn

OMe

OMe

OBn

164

12

CO2Me 14

12

14 Me3SiOTf

Ar

Ar

CO2Me

15

SiMe2Ph

CHO

OMe

166

167

168

Now coupling to the simpler allyl silane 169 with the same Lewis acid but also with MeOSiMe3 gives the new alcohol as its methyl ether 170. Once again control is from silicon as the aldehyde has no neighbouring chiral centre to exert Felkin control. New chiral centres at C-10 and C-11 are created. OMe

OMe 8

CO2Me

10

SiMe2Ph 169

168 Me3SiOTf Me3SiOMe

Ar

12

10 8

CO2Me

171

OMe 170

Again the product 170 is converted into an aldehyde 171 for the third allyl silane 172 reaction. This adds two more chiral centres by Houk control 173 and completes all seven chiral centres in the fragment 165. The diene side chain is added but this involves no new chiral centres

700

30 New Chiral Centres from Old

and the synthesis of (⫹)-macbecin I 163 was completed by reduction, deprotection and amide formation.28 OMe

OMe

OMe

8

Ar

CHO

8

Ar

7

Me3SiOTf Ar1CH2OSiMe3

OMe 171

MeO

OAr 1

OMe

OMe

SiMe2Ph 172

5

OMe 173

New chiral centres by 1,3-control 1,3-Control - the transmission of stereochemical information to the next-but-one carbon atom - is inherently more difficult than 1,2-control and almost impossible unless the molecule is held in some fixed conformation, usually by a ring (we shall not deal with that aspect as we have already discussed it at length), chelation, a cyclic intermediate, or a cyclic transition state. The next stage in the synthesis of tetrahydrolipstatin, the hydroboration of the alkene in 150 appears to involve 1,3-control in an open chain compound. However, the attack of 9-BBN on the alkene, which does indeed occur anti to the SiMe3 group 174, really occurs through a cyclic mechanism 176 involving 1,2 as well as 1,3-control. The B and H atoms add stereospecifically cis but stereoselectively anti to the SiMe3 group.24 The oxidation of the borane 174 to give the alcohol 175, occurs stereospecifically with retention (chapter 17). R2 9-BBN

150

R2B

R1 = n-C11H23 R2

R2

H2O2 NaOH

HO

CO2Bn R1

= n-C6H11

CO2Bn R1

SiMe3

SiMe3

SiMe3

R1

E R2B

H

R2 176; E = CO 2Bn

175; >95:5 anti:syn

174

H

The proline-catalysed aldol reaction (chapter 26) of the heterocyclic ketone 177 with benzaldehyde gave the anti-aldol 178. Reduction by the Prasad method (chapter 21) gave the anti,antidiol 179 and so, by desulfurisation with Raney Ni, the anti,anti-diol 179. The reduction could be controlled by the neighbouring chiral centre but the Prasad method uses 1,3-control by chelation with BEt2 and external axial delivery of nucleophilic hydride from NaBH4 181. This is 1,3-control via a cyclic intermediate.29 O

O PhCHO proline

OH

OH Ph

Ph

Raney Ni EtOH, THF

OH

Et Ph

pH 5.2

DMF

S 177

OH

OH

Et2BOMe NaBH4

S 178; 92% yield

H3B

H

O

Et B

O Ph

S 179

180; 94% yield S 181

The alternative method of reduction of such hydroxyketones is the anti-selective Evans’ intramolecular delivery of nucleophilic hydride through a cyclic transition state. In Janda’s bryostatin synthesis, the diol monoester 182, produced by kinetic resolution with an immobilised lipase (chapter 29) is reduced to the anti-diol 183 by acetoxy-borohydride. The hydride delivery is intramolecular through a chair transition state 184. This is 1,3-control via a cyclic transition state (chapter 21).30

30 Control of Open Chain Stereochemistry

O

OH

OH

Me4N B H(OAc) 3

R

R1

OH

O AcO

R

MeCN, AcOH

701

(–)

OAc

OAc

182

H

B OAc

183

184 R Evans' anti1,3-selective reduction O (–)

Apart from cases like these, we shall look for special methods to exercise 1,3-control. The importance of 1,3-control is immediately obvious from the structure of important cholesterollowering drugs like Lipitor (atorvastatin) 185. The syn-1,3-diol is a challenge but might be made from the amine 186 or the nitrile 187. OH

OH

OH

OH CO2H 186

CO2H

N

H2N

PhNHCO 185 atorvastatin F (Lipitor)

Ph

OH

OH CO2H

NC

187

One successful approach is to use the enzyme DERA (chapter 29) to combine two molecules of MeCHO and one of ClCH2CHO in a double aldol reaction. Each aldol creates a new chiral centre: the first is controlled entirely by the enzyme but the second, using the chiral aldehyde 189 as the electrophile, could be affected by 1,3-control.31 As might be expected for an enzyme-catalysed reaction, the catalyst dominates and the same stereoselectivity is found 190. The product is isolated in the form of the lactol 191.

CHO

OH

MeCHO DERA

CHO

H2O

Cl

OH

MeCHO DERA

OH

Cl

O

OH

CHO

H2O

Cl 188

Cl 189

OH 191

190

Transformation into a protected form of 187 is straightforward: oxidation with household bleach gives the lactone 192, displacement with cyanide ion, acetal formation and esterification with trimethylsilyl diazomethane gives 193. The synthesis of atorvastatin from 193 was already known. The next section discusses 1,3-control in non-enzymatic aldol reactions. NaOCl

Cl

O

O 1. NaCN, DMF, H 2O, 40 ˚C

191

O

HOAc H2O

2. Me 2C(OMe)2, H

OH 192; 45% from 188 >99.9% ee, >99.8% de

NC

O CO2Me

3. Me 3SiCHN2

193; protected 187; 48% from 192

Aldol reactions We have already seen how relative stereocontrol may be achieved in aldol reactions at the positions labelled 1 and 2 in 194 (chapters 4 and 27). One of these chiral centres is formed from the aldehyde electrophile and the geometry of the double bond of the enolate determined whether we got anti or syn geometry (chapters 4 and 21). The absolute stereochemistry at these centres could be controlled by a variety of methods (chapters 23–29), including the use of a chiral auxiliary (chapter 27).

702

30 New Chiral Centres from Old

O

O

OH

OH

OPMB

O

OH

2 1

BnO

3

TBSO

195

OH 5

R

R

1

1

194

O

4

1

196

197

Aldol reactions have become very much more sophisticated in the level of control that may be achieved. The types of control we are about to discuss do not depend upon double bond geometry. In each case from 1,3 195 to 1,5-related stereocentres 197 the newly forming alcohol is one of the stereocentres and the other may be either in the incoming enolate fragment 196, 197 or already present in the aldehyde 195. So the chiral centres are on either side of the ketone in 196 and 197 but both on the same side of the ketone in 195.

1,3-Anti induction Good levels of 1,3-anti induction can be achieved with aldehyde 199 when it is attacked by silyl enol ether 198. The aldehyde hydroxyl is the chiral centre and p-methoxybenzyl (PMB) is the best protecting group. Chelating Lewis acids (Ti, Li) are less effective than the combination of a silyl enol ether and BF3 as the Lewis acid. The nature of the enol derivative is important and good selectivity is not achieved when an enol borinate is used.32 We shall see in a moment that this lack of influence can be useful too. O

OSiMe3 +

OPMB

O

OPMB

F3B H O

H

dipole H R

199; PMB = p-MeO-benzyl

198

OH

BF3•Et2O

200; 92:8 anti : syn

Nu

H

H

OPMB

dipole 201

Evans suggests that an extended transition state is best with a Felkin-like orthogonal orientation of the chiral substituent to the carbonyl group. Conformation 201 shows an orientation in which the dipoles of the C⫽O and OPMB groups are as far apart as possible and with the nucleophile approaching alongside an H atom at the Bürgi-Dunitz angle. There are other possible explanations.33 We referred above to a synthesis of bryostatin that contained a reduction controlled by a 1,3relationship. Evans’ synthesis34 contains a 1,3-selective aldol as well as a 1,3-controlled reduction The aldehyde 202, made by an asymmetric aldol reaction, was combined with the double silyl enol ether of methyl acetoacetate to give, as expected, the anti-aldol 203. However, the only Lewis acid that gave this good result was (i-PrO)2TiCl2 and not BF3 thus emphasising the rather empirical aspect of this type of control. Evans’s own 1,3-controlled reduction gave the anti,anti-triol 204 that was incorporated into bryostatin. Me3SiO

PMBO Ph

OSiMe3

CHO

OMe (i-PrO)2TiCl2

Ph 202; PMB = p-MeO-benzyl

OH

O CO2Me

Ph

anti-203; 83% yield, 94:6 anti:syn PMBO

Me4N B H(OAc) 3

PMBO Ph

OH

OH

Ph

CO2Me

AcOH, MeCN

Ph

anti,anti-204; 92% yield

30 Control of Open Chain Stereochemistry

703

These aldols have all had just one chiral centre in the starting material. Should there be more than one, double diastereomeric induction produces matched and mismatched pairs of substrates and reagents, perfectly illustrated by the Evans’ aldol method applied to the syn and anti aldol products 205 themselves derived from asymmetric aldol reactions. The extra chiral centre, though carrying just a methyl group, has a big effect on the result. The absolute stereochemistry of the OPMB group is the same in both anti-205 and syn-205 but the stereoselectivity achieved is very different. The matched case favours Felkin selectivity as well as transition state 201 but, with the mismatched pair, the two are at cross purposes. It is interesting than 1,2-control does not dominate in this case.33 OPMB

O

198

OH

OPMB

O

OHC

OPMB

+

Et2O.BF 3

syn,anti-206

anti-205

O

OPMB

OH

anti,anti-206

98:2 OPMB

O

198

OHC

OH

OH

OPMB

+

Et2O.BF 3

syn-205

syn,syn-206

anti,syn-206

56:44

So far we have looked at the enolates of methyl ketones e.g. 198. This has been deliberate in order to decouple these more remote selectivity issues from those of double bond geometry that we would also encounter if we were making enolates from, say, ethyl ketones such as 207. In fact, although we have seen good selectivity, it is easier to obtain selectivity with the added congestion that arises with ethyl ketones. So, for instance the reaction of methyl ketone 198 leads to an anti selective product in a ratio of 94:6. Contrast this with the reaction of ethyl ketone 207 where there are four possible products rather than two and where the selectivity is 95:5 – the ‘95’ being the major isomer 209 and the ‘5’ being the sum of all the others! Note that the selectivity is observed without the use of chiral ligands.35 O

OBR2 R2BOTf

BnO

O PrCHO

BnO 207

OH

95 : 5 Pr anti,anti: the rest anti,anti-209; 72% yield

BnO

E-208: an E-enolate

The trans double bond in 208 is responsible for the anti relationship between the alcohol and the adjacent methyl group. In fact, in a study with a variety of aldehydes, the most abundant of the three minor isomers also has that anti relationship. If we were to have a cis double bond in the enol borinate instead then we would expect to observe syn relationships here and that is exactly what happens. That is 1,2-control. Now, whether this local syn arrangement is anti or syn to the more remote chiral centre in the enol borinate (1,3-control) can be controlled by the use of one enantiomer or other of a chiral boron reagent. The cis enol borinate with achiral boron reagent does not have much will of its own and leads to a 1:1.2 ratio of syn, syn: anti, syn. OBR2

O

OH

O

OH

PrCHO

BnO

Z-208: a Z-enolate

BnO

Pr

syn,syn-209

+

1:1.2

BnO

Pr

anti,syn-209

704

30 New Chiral Centres from Old

This is an advantage in that the enantiomers of Ipc (chapter 24) can exert their influence. Since both enantiomers of Ipc are available, we can make the diastereomers of the aldols 209 too using (⫹)- or (⫺)-(Ipc)2BOTf to make the boron enolates.36 And of course if (⫹)-(Ipc)2BOTf gives syn,syn-209, (⫺)-(Ipc)2BOTf gives anti,syn-209. O

OH

OB(Ipc)2

(–)-Ipc used

BnO

BnO

OHC

anti,syn-210

O

(+)-Ipc used

OH

BnO

OHC

Z-208; R = Ipc

syn,syn-210

cis

syn

syn

7 : 93 anti, syn : syn, syn

93 : 7 anti, syn : syn, syn

Introduction to 1,4- 1,5- and Remote Induction Hepatitis C is the one that causes liver cancer and more effective treatments are urgently required. One approach is to discover specific protease inhibitors that should prevent replication of the virus. The Boehringer company37 have found BILN 2061 211 which they have made by the strategy outlined in 211. 212

S

S

NH MeO

N

O

C–N amide H N O

NH MeO

N SN2 ether

N

N

H2N

C–N amide H N

N

OH CO2H

O O

CO2H

HO 213

H2N

CO2H

N H

CO2H 214

O

C=C metathesis 211; BILN 2061 protease inhibitor

215

The achiral aromatic part 212 can easily be removed. The rest of the molecule has five chiral centres in three groups and the best strategy is to assemble each group separately. The amino acid fragment 215 can be made by asymmetric catalytic reduction (chapter 26). Hydroxyproline 213 is from the chiral pool and the cyclopropane 214 was made as a racemate and resolved. Linking these three fragments by stereospecific reactions (note that two inversions will be needed during ether formation with 212 to keep the trans arrangement across the ring) or by reactions that do not affect the chiral centres ensures that the final product has the correct absolute and relative stereochemistry. Another popular strategy is to use a reagent-controlled reaction to add any further centres independently of those already present, trusting in the remoteness of the relationship to prevent any induction. A simple example is the anti-fungal agent (⫺)-PF1163B 216, a compound from a Streptomyces species. Again a metathesis was added to the obvious disconnections (ester and amide) to give an amino acid fragment 218, an achiral fragment 219, and the interesting unsaturated alcohol 217 having a 1,4 relationship between the two chiral centres.38

30 Introduction to 1,4- 1,5- and Remote Induction

O

O C–O ester R

O

705

O

OH

HO2C

NMe C–N amide

C=C metathesis Me 216; R = n-C 5H11; (–)-PF1163B

NMe 218

OH

R

O

OH

HO2C

Me 219

217; R = n-C 5H11

(S)-Citronellene 220 was available from the chiral pool so an oxidative cleavage gave the aldehyde 221 and the n-pentyl group could be added as an organozinc reagent by the method of Kobayashi39 using the powerful C2 symmetric catalyst 222. The catalyst-dominated reaction ignores the chiral centre already present and gives excellent ee (98%). OH

R 1. mCPBA (90% yield)

0.08 equiv catalyst 222

2. H 5IO6 (63% yield)

(i-PrO)4Ti Et2O

Me 220; ( S)-citronellene

NHTf

(C5H11)2Zn

CHO

Me

NHTf Me 217; R = n-C 5H11 50% yield, 98% ee

221

222 catalyst for organo-Zn addition

These strategies are often chosen when relationships are 1,4- or more remote. In 211 and 216 the relationships between the groups of chiral centres are 1,4 at best. However, it is possible to control 1,4-relationships, and even a few that are more remote, by induction (that is stereoselective reactions) from existing centres and that is the concern of this part of the chapter. We shall first look at some relatively simple methods and then return to the aldol reaction.

Remote control by substrate-controlled induction There are probably more different reagents for reducing carbonyl groups than for any other reaction and some of these are very effective in controlling stereochemistry.40 Many of these methods use reagent control but we are concerned with those that induce stereochemistry from an existing centre. The most famous example is the secondary alcohol in the prostaglandin side-chain. This is particularly difficult as the prochiral ketone in it is separated from the existing chiral centre by a trans alkene.41 In the event,42 the achiral reducing agent 224, a bulky version of DIBAL, gave a 92:8 ratio of diastereoisomers of 225. O

O

O O

10 equivs 224

R HO

O 223; R = n-C 5H11

O

toluene –78 ˚C

t-Bu

t-Bu

R HO

OH 225; R = n-C 5H11

Al(i-Bu)2

Me 224

706

30 New Chiral Centres from Old

The requirement for such a large excess of 224 suggests that a second molecule binds to the OH group thus shielding one face of the ketone from reduction 226. Though asymmetric reduction would now be preferred to this method it is remarkable that such a remote centre (and the OH is five atoms from the carbonyl group) exerts such a powerful effect. A more typical early example is Kishi’s polyether antibiotic syntheses using a sequence of epoxides made by controlled oxidation. The sequence starts with a somewhat similar reduction: LiAlH4 with the diamine 229 reduces ketone 227 highly stereoselectively in a Felkin sense 230 to give almost exclusively anti-228 which was resolved as a urethane (chapter 22) to give the enantiomer shown.43 O LiAlH4

O O H 226

O

R2 Al

R OAr

O

OH

229

MeO

MeO

228; 97% yield, >10:1 anti:syn

227 Me

H N

O

Me

H Ar 230

229

Me

H

R = Ar

H

R

N H

OH

H

R

H

Ar 228

OH 228

The hydroxyl group now directs the epoxidation of the nearer alkene using a vanadium directed delivery of t-BuOOH that gives an 8:1 selectivity. This resembles the last example being 1,4-control from OH though there is a nearer (1,3) chiral centre. The epoxide was not isolated but cyclised in situ to the THF 231. This releases a new OH group that directs a second epoxidation and cyclisation with the same reagents to the bis-THF 232. The stereochemistry of the THFs makes it clear that the OH is a 1,4-syn directing group for the epoxidations.

1. t-BuOOH VO(acac) 2 NaOAc

228

1. t-BuOOH VO(acac) 2 NaOAc

Ar

2. HOAc

O

Et

2. HOAc

OH

231; 75% yield, 8:1 diasts

Ar

O

Et H

O

Et

OH

232; % yield, 5:1 diasts

1,4 -Syn induction in the aldol reaction When we discussed how E-enolates of ethyl ketones such as 207 gave 1,3-control in the aldol reaction, we noted that there was 1,4-control too. Paterson did the same reaction on the corresponding methyl ketones and found that the lithium enolate (M ⫽ Li in 234) was unselective. The boron enolate with an achiral group 9 (M ⫽ dicyclohexyl-B) was selective giving 88:12 syn:anti-235 in 84% yield but with a chiral group [M ⫽ (⫺)-(Ipc)2B] the stereoselectivity was significantly better35 (92:8). O

OM

O

OH

PrCHO

BnO

BnO (+)-233

Pr + anti-235

BnO 234

syn-235

30 Introduction to 1,4– 1,5- and Remote Induction

707

This must be entirely 1,4- control. As these adducts 235 are barely distinguished by 13C NMR and could not be separated even by HPLC, high diastereoselectivity is essential. One explanation is a Houk-like arrangement of the enolate but with methyl ‘inside’ (as the large groups on boron allow only H to be nearby) with the aldehyde arranged with Pr and BR2 ‘trans’ 236. O B

Pr

H

Ipc Ipc

O H OBn 236

1,4-Control by sigmatropic shifts Some pages back we described part of a synthesis of macbecin 163 by allyl silane chemistry. An alternative approach uses sigmatropic rearrangements to transfer and create new chiral centres.44 The starting material gives a [2,3] Wittig rearrangement 238 after treatment with base. Centres 2 and 3 in the product 239 survive intact. Centre 6 is transposed from 4 by the rearrangement and centre 7 is created. The new alkene is E and all this is controlled by the chair-like transition state. This product 239 has two families of adjacent chiral centres 1,4-related across the new E-alkene.

N

N

O

O O

O

LDA

N

[2,3]

7

O

3 6

4

MOMO 237

OBn

MOMO 238

OH

OBn

1 2

MOMO

OBn

239

Further transformations involving the Felkin style creation of a new chiral centre at C-8 allow a second [2,3]-rearrangement after the exchange of the Bu3Sn group for Li. Another trans alkene is formed and the centre at C-8 transposed to C-10. Bu3Sn

O MeLi

239

7

3 6

8

1 2

OMe

MOMO

7 10

[2,3]

OBn

3 6

OH

OMe

240

1 2

MOMO

OBn

241

This strategy is generally more efficient than the direct creation of new centres by, for example, nucleophilic addition of organometallic compounds to carbonyl groups separated by an alkene from even such a powerful directing group as an amine. The cis-alkene Z-242 reacts with a range of organolithiums (R2 ⫽ alkyl or aryl) giving syn-243 with usually ⬎90:10 selectivity. Direct coordination of Li to the amino group in a Houk-like arrangement 244 explains this result.45

R1

CHO NBn2

Z-242

R2Li –100 ˚C

R2

R1 Bn2N HO

Z-syn-243

Bn2N R1

Li H O 244

R2 H

708

30 New Chiral Centres from Old

The trans-alkene E-242 also gives reasonable stereoselectivity for the syn adduct E-syn-243 but requires an organocuprate, HMPA in the solution, and the product must be trapped with Me3SiCl to get between 80:20 and 90:10 selectivity. One possible explanation is that the alkene-Cu π-complex 245 bridges the longer gap between the amino group and the aldehyde. R2

OSiMe3 R1

CHO

(R2)2CuLi, –78 ˚C

R1

HMPA, Et 3N, Me 3SiCl

NBn2

O H

R1

H

E-syn-243

E-242

SiMe3

Bn2N Cu

R2 NBn2

R2

245

1,5-Induction by the aldol reaction Highly 1,5-diastereoselective reactions can be achieved when the β-hydroxy group is present in the boron enolate of a methyl ketone such as 246 instead of the aldehyde. Notice that we already have a regioselectivity issue here – the Bu2BOTf/i-Pr2NEt combination selectively reacts with the methyl group and not the methylene group. Reaction with dihydrocinnamaldehyde gives an aldol 247 with 95:5 anti:syn selectivity. In interesting contrast to the 1,3-diastereoselective aldols 200 above, silyl enol ethers give no 1,5-diastereoselectivity at all.32 Selectivity can also be switched off by replacing the methoxybenzyl protecting group (PMB) with a silyl protecting group (TBDMS). PMBO

O

PMBO

1. Bu 2BOTf, i-Pr2NEt

O

OH Ph

2. OHC Ph

247; 95:5 anti : syn

246

The non-selective reactions that result if boron enolates are used in the 1,3 case or if silyl enol ether are used for a 1,5 case means we can choose to switch off selectivity that might otherwise clash (mismatch). Boron enolate 248 combines with aldehyde (S)-249 and, in another reaction, its enantiomer (R)-249. As we do not want the chirality of the aldehyde to be in control, boron enolates are what we need. The 1,5 anti relationship is controlled in both cases but the 1,3 anti or syn relationship develops by default and not by control. H

OR

OTr

1,3 OFF 1,5 ON

O

OTBS O (S)-249

O

OTr

O

O

OTr

OBBu2 H

248

O

OTBS O

3

OH

OTBS O

anti 5

OR

Ph

O

1

O

OR

250; 91:9 85% yield

Ph O

syn

anti 5

O

O

anti 1

3

OH

OTBS O

OR

(R)-249

Ph

251; 91:9 85% yield

Alternatively we can combine the chiral nucleophiles 252 and 253 and a chiral aldehyde 254, keeping the chiral aspects of the molecules the same in each reaction, but changing the nature of the enolate to suit us. Hence 1,5-induction operates when boron enolate 253 is used but 1,3induction operates with the silyl enol ether 253.

30 The Asymmetric Synthesis of (+)-Discodermolide

709

syn

anti

TrO O

O

5

TrO

OBBu2

O 1,3 OFF 1,5 ON

Ph 252

H

O Ph

O

OTES

Ph 253

1,3 ON 1,5 OFF

OH

OPMB OBn

255; 96:4 81% yield

syn

(S)-254

O

3

OPMB OBn

TrO O

O

1

5

TrO O

O Ph

anti 1

O

OH

3 OPMB OBn

256; 82:18 66% yield

Note the strategy in these two schemes: one enantiomerically pure fragment is being combined with another enantiomerically pure fragment and, where they join, conditions are selected to determine which of these fragments controls the new chiral centre. This is a substrate-based strategy in all respects. One thing we have not done is to use an external chiral factor, such as a chiral boron reagent, and we will see this strategy now. We have already mentioned that the use of a silyl protecting group is not effective for 1,5-anti induction. So, if the boron enolate from 258 reacts with methacrolein, the resulting selectivity (71:29) is not too bad – though not as good as with a benzyl protecting group 257. An alternative method of obtaining good stereocontrol, if we want to keep silyl protection, is to introduce reagent control and replace the achiral boron chloride with (⫺)-Ipc2BCl. This improves selectivity markedly with no drop in yield (which is 80% in both cases). OTES OP

O

OTES OP 2.

BnO 257; P = PMB 258; P = TBDMS

O

OH

1. (c-Hex) 2BCl, Et 3N O

H

1. (–)-Ipc 2BCl, Et 3N

BnO 259; P = PMB 91:9 anti:syn 260; P = TBDMS 71:29 anti:syn 260; P = TBDMS 90:10 anti:syn

We have already collected some powerful tools for use in stereocontrolled aldol reactions, but we have not finished. We shall see now in Paterson’s synthesis of (⫹)-discodermolide, how reagent control is used not to enhance the intrinsic substrate selectivity, but to overturn it. The aldol reaction is undoubtedly one of the most powerful ways of making carbon-carbon bonds and nature thinks so too. There are numerous natural products that are replete with 1,3 related oxygen functionality. Many of these are acetate or propionate-derived in nature. The methods detailed above developed from studies into the syntheses of these natural products. The manipulations of chiral ethyl ketones of this kind are of particular interest when it comes to natural products that are polypropionate-derived.

The Asymmetric Synthesis of (⫹)-Discodermolide As one might expect, there are very many aldol reactions to be found in Paterson’s synthesis46 of discodermolide 261. The synthesis involves 23 steps (in the longest linear sequence) and the overall yield is 10.3%. If we naïvely equate the yield with those 23 steps then we have an average yield in every step of around 90%. Several other groups are interested in the synthesis of discodermolide.47,48

710

30 New Chiral Centres from Old

b HO O

a

O

OH

O

OH

O

261; (+)-discodermolide

NH2

OH

Although the molecule contains thirteen chiral centres, the synthesis starts with only three. They are ethyl ketones 262, 263 and 264. Discodermolide is retrosynthetically divided 261 into three sections and all of the chiral centres in each of these sections derive from one of the ketones. Ketones 262 and 263 differ only in the nature of the protecting benzyl group.

BnO

PMBO

OBz

O

O (S)-264

O

(S)-262

(S)-263

We shall not go through every detail of the synthesis – you are directed to the paper for that – but we will highlight three details. Firstly, there is the reaction of 262 which (in one pot!) is transformed into the advanced intermediate 265 in 86% yield and ⬎97% diastereomeric excess. Secondly, there is the late-stage coupling of fragments (the C–C bond made is disconnection ‘b’ in 261) and finally there is the overturning of selectivity by reagent control that we alluded to earlier.

BnO

aldol

stereoselective

MeCHO

reduction

O

BnO OH OH 265; 86% yield, >97% ds

(S)-262

So, to kick off, ketone 262 is reacted to form trans boron enolate 266 that combines with acetaldehyde to give the intermediate 267. The 1,2-anti stereocontrol comes from the enolate geometry and the chiral centre of the original ketone imposes the 1,3-anti control in the way we have seen already in this chapter. The intermediate 267 is reduced without working the reaction up and the boron removed to give 268. 1,3-anti 1,2-anti MeCHO

BnO O

BR2

E-266; R = cyclohexyl

BnO O

B R2 267

O

1. LiBH 4 –78 ˚C 2. H 2O2 NaOH

BnO OH

OH

268

The newly formed chiral centre bearing the alcohol in 266 determines the new 1,3-anti relationship that develops as the carbonyl is reduced by LiBH4. Hydride attacks axially 269 (as in Chapter 21).

30 The Asymmetric Synthesis of (⫹)-Discodermolide

711

The alcohol that resulted from the aldol reaction is oxidised later (on the way to 270) and so one chiral centre is destroyed after it has done its work. Hydride arriving with axial attack R

Recently formed H chiral centre H O B R MeMe O R 269

268

MeO2C TBDMSO

O

270

With all this boron enolate chemistry it is easy to forget some of the older enolate methods that are still available. The use of an ester containing a hindered phenolic group has been seen before in Chapter 21 to control the formation of trans enolates so that anti aldol products may be produced. It is used here to couple two fragments in just that way. Ester 271 reacts with the hindered base LiTMP (in combination with LiBr) to give trans enolate 272 which was combined with aldehyde 273 to give49 the aldol adduct 274.

O

PMBO

LiTMP

OAr

PMBO

RO

LiBr

O

OLi OH 272; trans-enolate

OH 271

ArO 272

H O

REDRAW

syn RO

O

syn

PMBO

OH

OPMB 273

O

O OH anti-275

OH

274

OPMB

OH 'syn'-275

The syn selectivity is controlled by the double bond geometry of the trans enolate 272 whereas the more remote aspects of the stereocontrol are controlled by the molecules themselves. Just a note at this point: we normally associate trans enolates with anti aldol products – the product observed is called syn only because of the way it is drawn, there is nothing unusual here 275. The chiral centres in the enolate 272 are too remote to be effective and it is those in the aldehyde 273 that control the more remote aspects of stereochemistry. If we want to explain this selectivity, we need to have a reason for why one diastereotopic face of the aldehyde is attacked and not the other. As might be expected, the aldehyde reacts with Felkin-Anh selectivity as described earlier in this chapter. Our final highlight in the discodermolide synthesis is the use of reagent control to get what we want and not what the molecules want. The combination of enol borinate 276 and an aldehyde 277 featuring a cis double bond, led to the formation of aldols anti- and syn-278. A model study had shown that the inherent selectivity with these enals could be improved by the use of (⫺)Ipc groups on boron instead of cyclohexyl but it is not the selectivity that was wanted. (⫹)-Ipc groups were able to turn around the selectivity and improve the yield of the reaction while they were at it.50

712

30 New Chiral Centres from Old

HO

O

HO

BR2 MeO2C

H

O

R' TBDMSO

MeO2C

R'

O

MeO2C TBDMSO

277

OTBDMS anti-276

TBDMSO

R'

O TBDMSO

+ TBDMSO

anti-278

syn-278

R = c-Hex (67% yield) R = (+)-Ipc (87% yield)

88:12 16:84

The tricky thing with a mismatched situation is that you do not know whether you will win (with reagent control) or whether the molecules will get their way (substrate control) but either way the result will be the product of a conflict. It is all the more impressive, therefore, when good levels of reagent control can be achieved in a mismatched situation.

Making discodermolide on a large scale Discodermolide is a promising anti-cancer drug and the Novartis company decided to blend the best parts of the published academic syntheses with their own ideas and make 60 grams of the compound for clinical trials. This was a decision necessitated by the very meagre supply from natural sources but a bold decision too as a large scale synthesis of such a complex molecule was without precedent. Discodermolide has three Z-alkenes and 13 chiral centres with, at first sight, no particular pattern. Smith noticed47 that there were three repeating triads of centres marked with circles in 261a and that each of these might be made from a ‘common precursor’ 279.

OMe

HO PMBO O

OH

O

O

N

O OH

OH

NH2

O

279; Amos Smith's 'common precursor'

261a; (+)-discodermolide

OH

Me

The Novartis company decided to use 279 as starting material for three fragments 280, 281 and 282 and to use Paterson’s aldol methodology (described above) to link them together and complete the synthesis.51 In the event, they modified the original plan in many ways to make the synthesis practical. Key points were the production of 279 without chromatography, the conversion of 279 to 281 via all crystalline intermediates, a Suzuki coupling (chapter 18) and a Still-Gennari Z-alkene synthesis (chapter 15) and many other reactions already treated in this book. OMe I

PMBO TBDMSO 280

O

O

OTBDMS

Ar 281; R = p-MeOC6H4

N O

OR

O

282; R = TBDMS

Me

30 References

713

Though improvements continue to appear, such as a revised route to the lactone portion,52 and no doubt will continue to appear, this is a staggering achievement involving 43 chemists and you are particularly directed to the fifth paper in the series.51 Commenting on the synthesis as a whole they say: ‘The success of this project and its chemistry paves the way for other, perhaps even more complex, natural products to be prepared for early-phase clinical evaluations and sends a positive message to both the isolation and synthetic academic community and possibly other pharmaceutical companies that; “your work need not just be of academic interest” and it may be worth taking a few risks.’

Conclusion Creating more chiral centres with enantiomerically pure rather than racemic starting materials is different in some important ways. 1. Remote diastereomeric relationships can be controlled by using enantiomerically pure components when it is impossible to control them by diastereoselective reactions. 2. Double diastereodifferentiation (using enantiomerically pure reagents and substrates) can enhance diastereoselectivity if we have a matched case, but can erode it if we have the mismatched case. 3. It is important to check the ee of the first product formed with extra chiral centres. There are three considerations here: (a) The starting material may racemise under the conditions of a reaction. (b) Coupling a starting material that is racemic or of low ee to an enantiomerically pure starting material may allow effective resolution of the product. (c) Once two or more chiral centres are in place, the enantiomeric purity of the molecule can be checked by ordinary NMR spectra as epimerisation leads to a diastereoisomer and simultaneous epimerisation of two or more centres without forming another diastereoisomer is unlikely (though not unknown).

References General references are given on page 893 1. A. G. Schultz, M. A. Holoboski and M. S. Smyth, J. Am. Chem. Soc., 1996, 118, 6210. 2. P. D. Bailey and J. S. Bryans, Tetrahedron Lett., 1988, 29, 2231. 3. D. R. Adams, P. D. Bailey, I. D. Collier, J. D. Heffermann and S. Stokes, Chem. Commun., 1996, 349. 4. O. Ort, Org. Synth., 1983, 63, 203. 5. F. Zhang, Z. J. Song, D. Tschaen and R. P. Volante, Org. Lett., 2004, 6, 3775. 6. G. J. Tanoury, R. Hett, S. Wilkinson, S. A. Wald and C. H. Senanayake, Tetrahedron: Asymmetry, 2003, 14, 3487. 7. S. Randl and S. Blechert, J. Org. Chem., 2003, 68, 8879. 8. G. Fráter, Helv. Chim. Acta, 1979, 62, 2825, 2829. 9. D. Seebach, J. Aebi and D. Wasmuth, Org. Synth., 1983, 63, 109. 10. G. Fráter, Helv. Chim. Acta, 1980, 63, 109. 11. H.-M. Shieh and G. D. Prestwich, J. Org. Chem., 1981, 46, 4319; A. R. Chamberlin and M. Dezube, Tetrahedron Lett., 1982, 23, 3055; H.-M. Shieh and G. D. Prestwich, Tetrahedron Lett., 1982, 23, 4643. 12. Clayden, chapter 33, especially pages 864–868. 13. L. A. Paquette, J. Tae, M. P. Arrington and A. H. Sadoun, J. Am. Chem. Soc., 2000, 122, 2742. 14. J. Mulzer and M. Shanyoor, Tetrahedron Lett., 1993, 34, 6545.

714

30 New Chiral Centres from Old

15. D. J. Hart and T.-K. Yang, J. Chem. Soc., Chem. Commun., 1983, 135; J. Org. Chem., 1985, 50 235. 16. A. K. Mapp and C. H. Heathcock, J. Org. Chem., 1999, 64, 23; this synthesis is discussed at length in Bittner et al., Organic Synthesis Workbook II. 17. W. Aelterman, Y. Lang, B. Willemsens, I. Vervest, S. Leurs and F. De Knaep, Org. Proc. Res. Dev., 2001, 5, 467. 18. J. Liu and C.-H. Wong, Angew. Chem. Int. Ed., 2002, 41, 1404. 19. S. W. Kaldor, V. J. Kalish, J. F. Davies, B. V. Shetty, J. E. Fritz, K. Appelt, J. A. Burgess, K. M. Campanale, N. Y. Chirgadze, D. K. Clawson, B. A. Dressman, S. D. Hatch, D. A. Khalil, M. B. Kosa, P. P. Lubbehusen, M. A. Muesing, A. K. Patick, S. H. Reich, K. S. Su and J. H. Tatlock, J. Med. Chem., 1997, 40, 3979. 20. M. Ikunaka, Chem. Eur. J., 2003, 9, 379. 21. I. F. Fellows, D. E. Kaelin and S. F. Martin, J. Am. Chem. Soc., 2000, 122, 19781. 22. J. M. Humphrey, R. J. Bridges, J. A. Hart and A. R. Chamberlin, J. Org. Chem., 1994, 59, 2467. T. F. Jamison, W. D. Lubell, J. M. Dener, M. J. Krisché and H. Rapoport, Org. Synth., 1993, 71, 220. 23. I. Fleming, A. Barbero and D. Walter, Chem. Rev., 1997, 97, 2063. 24. I. Fleming and N. J. Lawrence, Tetrahedron Lett., 1990, 31, 3645. 25. T. Hayashi, M. Konishi, H. Ito and M. Kumada, J. Am. Chem. Soc., 1982, 104, 4962; T. Hayashi, M. Konishi and M. Kumada, J. Am. Chem. Soc., 1982, 104, 4963. 26. R. T. Beresis, J. S. Solomon, M. G. Yang, N. F. Jain and J. S. Panek, Org. Synth., 1998, 75, 78. 27. J. S. Panek and N. F. Jain, J. Org. Chem., 1994, 59, 2674. 28. J. S. Panek and F. Xu, J. Am. Chem. Soc., 1995, 127, 10587. 29. D. E. Ward and V. Jheengut, Tetrahedron Lett., 2004, 45, 8347. 30. J. A. Lopez-Pelegrin, P. Wentworth, F. Sieber, W. A. Metz and K. D. Janda, J. Org. Chem., 2000, 65, 8527. 31. W. A. Greenberg, A. Varvak, S. R. Hanson, K. Wong, H. Huang, P. Chen and M. J. Burk, Proc. Nat. Acad. Sci., 2004, 101, 5788. 32. D. A. Evans, P. J. Coleman and B. Côté, J. Org. Chem., 1997, 62, 788. 33. D. A. Evans, J. L. Duffy and M. J. Dart, Tetrahedron Lett., 1994, 35, 8537; D. A. Evans, M. J. Dart, J. L. Duffy and M. G. Yang, J. Am. Chem. Soc., 1996, 118, 4322. 34. D. A. Evans, P. H. Carter, E. M. Carreira, J. A. Prunet, A. B. Charette and M. Lautens, Angew. Chem. Int. Ed., 1998, 37, 2354. 35. I. Paterson, J. M. Goodman and M. Isaka, Tetrahedron Lett., 1989, 30, 7121. 36. I. Paterson and M. A. Lister, Tetrahedron Lett., 1988, 29, 585. 37. A.-M. Faucher, M. D. Bailey, P. L. Beaulieu, C. Brochu, J.-S. Duceppe, J.-M. Ferland. E. Ghiro, V. Gorys, T. Halmos, S. H. Kawai, M. Poirier, B. Simoneau, Y. S. Tsantrizos and M. Llinàs-Brunet, Org. Lett., 2004, 6, 2901. 38. F. Bouazza, B. Renoux, C. Bachmann and J.-P. Gesson, Org. Lett., 2003, 5, 4049. 39. H. Takahashi, T. Kawakita, M. Ohno, M. Yoshioka, and S. Kobayashi, Tetrahedron, 1992, 48, 5691; P. Knochel, Synlett., 1995, 393. 40. H. Sailes and A. Whiting, J. Chem. Soc., Perkin Trans. 1, 2000, 1785. 41. H. J. Mitchell, A. Nelson and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1999, 1899. 42. Y. Tsuda, T. Sano, J. Taga, K. Isobe, J. Toda, S. Tagaki, M. Yamaki, M. Murata, H. Irie and H. Tanaka, J. Org. Chem., 1979, 94, 1358; S. Iguchi, H. Nakai, M. Hayashi, H. Yamamoto and K. Maruoka, Bull. Chem. Soc. Jpn., 1981, 54, 3033. 43. T. Fukuyama, B. Vranesic, D. P. Negri and Y. Kishi, Tetrahedron Lett., 1978, 2741. 44. S. J. Coutts, M. D. Wittman and J. Kallmerten, Tetrahedron Lett., 1990, 31, 4301; S. J. Coutts and J. Kallmerten, Tetrahedron Lett., 1990, 31, 4305. 45. M. T. Reetz, F. Wang and K. Harms, J. Chem. Soc., Chem. Commun., 1991, 1309. 46. I. Paterson, G. J. Florence, J. P. Scott and N. Sereinig, J. Am. Chem. Soc., 2001, 123, 9535. 47. A. B. Smith, T. J. Beauchamp, M. J. LaMarche, M. D. Kaufman, Y. P. Qiu, H. Arimoto, D. R. Jones and K. Kobayashi, J. Am. Chem. Soc., 2001, 123, 8654. 48. D. T. Hung, J. B. Nerenberg and S. L. Schreiber, J. Am. Chem. Soc., 1996, 118, 11054. 49. P. L. Hall, J. H. Gilchrist and D. B. Collum, J. Am. Chem. Soc., 1991, 113, 9571. 50. I. Paterson and G. J. Florence, Tetrahedron Lett., 2000, 41, 6935.

30 References

715

51. S. J. Mickel, D. Niederer, R. Daeffler, A. Osmani, E. Kuesters, E. Schmid, K. Schaer, R. Gamboni, W. Chen, E. Loeser, F. R. Kinder, K. Konigsberger, K. Prasad, T. M. Ramsey, O. Repic, R.-M. Wang, G. Florence, I. Lyothier and I. Paterson, Org. Process Res. Dev., 2004, 8, 122. 52. I. Paterson, O. Delgado, G. J. Florence, I. Lyothier, M. O’Brien, J. P. Scott and N. Sereinig, J. Org. Chem., 2005, 70, 150; A. B. Smith, B. S. Freeze, M. J. LaMarche, T. Hirose, I. Brouard, P. V. Rucker, M. Xian, K. F. Sundermann, S. J. Shaw, M. A. Burlingame, S. B. Horwitz and D. C. Myles, Org. Lett., 2005, 7, 311; S. J. Mickel, R. Daeffler and W. Prikoszovich, Org. Process Res. Dev., 2005, 9, 113.

31 Strategy of Asymmetric Synthesis This chapter examines the asymmetric synthesis of compounds that have been made by several different methods and assesses the factors that make one synthesis the most favourable. It also revises material from the rest of this section (chapters 20–30)

Introduction: Strategy of Asymmetric Synthesis PART I – (R) AND (S)-2-AMINO-1-PHENYLETHANOL Methods rejected for development Methods considered for development and tried experimentally but rejected A successful large scale asymmetric synthesis by resolution PART II – (2S,4R)-4-HYDROXYPIPECOLIC ACID A simple heterocyclic acid needed for an anti-HIV drug A chiral pool synthesis from aspartic acid Choosing a new reaction to solve the stereochemistry problem Making the new reaction asymmetric PART III – GRANDISOL AND SOME BICYCLO [3.2.0]HEPTAN-2-OLS A bicyclic insect attractant used in agriculture Chiral Pool Syntheses from Other Terpenes An attempt from linalool An attempt from citronellol using carbene chemistry Asymmetric Synthesis Asymmetric addition of an organo-zinc to an aldehyde Synthesis by asymmetric dihydroxylation of a non-conjugated diene A Disappointment and a Resolution PART IV – METALLOPROTEINASE INHIBITORS A group of compounds that protect cartilage from enzymatic degradation Asymmetric alkylation of Evans’s chiral enolates Asymmetric synthesis from an enantiomerically enriched hydroxy-acid PART V – CONFORMATIONAL CONTROL AND RESOLUTION: KINETIC OR NOT? A tetrahydropyran that inhibits leukotriene biosynthesis Asymmetric synthesis of 2-methyl-tetrahydropyran-4-one by kinetic resolution PART VI – ASYMMETRIC DESYMMETRISATION OF A DIELS-ALDER ADDUCT Ifetroban sodium: a thromboxane receptor antagonist A laboratory synthesis starting with a Diels-Alder reaction Desymmetrisation of a symmetrical anhydride with a chiral Grignard reagent Laboratory and process routes compared PART VII – ASYMMETRIC SYNTHESIS OF A BICYCLIC β-LACTONE Lactacystin: a natural proteasome inhibitor Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

718

31 Strategy of Asymmetric Synthesis

Strategies used in the synthesis of lactacystin, the β-lactone, and analogues PS-519: A lactacystin analogue A chiral pool synthesis of lactacystin from glucose Chiral pool syntheses of lactacystin from amino acids The reagent strategy: asymmetric epoxidation and asymmetric allylboration Enzymes as reagents in the synthesis of lactacystin analogues Syntheses of 3-hydroxyleucine by other strategies A large-scale synthesis of lactacystin analogues using AD and an asymmetric aldol reaction

Introduction: Strategy of Asymmetric Synthesis In the last seven chapters we have introduced the main approaches to the synthesis of single enantiomers. In this chapter we gather together those ideas and consider why anyone should choose one of these approaches rather than another. We shall look at simple and not-so-simple compounds that have been made by different routes. We shall examine why chemists rejected some of those routes and preferred others. Some compounds will be needed for laboratory syntheses on a small scale: others for commercial production on a large scale and different criteria will apply to these two situations. These are strategic considerations and the main point will be, as it always is, to get high yields in few steps with maximum selectivity. Here though the enantiomeric excess achieved by the different methods will also be important. We start with a very simple compound. More complicated molecules are considered in detail in Nicolaou and Sorensen, a book we highly recommend.

PART I – (R) AND (S)-2-AMINO-1-PHENYLETHANOL This simple compound 2 was needed by Novartis Pharma Inc. (Basel) for the preparation of a new class of anti-diabetic agents 1. Both enantiomers were needed on a large scale to explore the relationship between stereochemistry and biological activity.1 OH

OH

H N Ph

reductive amination

NH2

+

O Ph Alk

Alk 1; anti-diabetic drugs

2; 2-amino-1-phenylethanol

Methods rejected for development The first step is to search the literature. Many methods had already been published but four of these were immediately discarded as being unsuitable for scale-up. Asymmetric reduction (chapter 26) of the amino-ketone 3 required ligands that were not available and would themselves have to be the subject of asymmetric synthesis. Then pressures greater than 50 atmospheres were needed to reduce the rather stable carbonyl group and even then ees were about 93% at best and usually much lower. Finally the turnover was slow so a lot of catalyst would be needed to make an efficient process. An alternative to chemical reduction would be baker’s yeast (chapter 26). This gave only one enantiomer and required high dilution to be effective. Even then there was a very low throughput. O

O

H2/cat

NH2

Ph 3

OH

baker's yeast

2

NH2

Ph 3

NH2

Ph (R)-2

31 Introduction: Strategy of Asymmetric Synthesis

719

The racemic amino alcohol 2 was easy to prepare so one obvious choice is an enzyme-catalysed acylation or deacylation process (chapter 26). These worked reasonably well. Deacylation of the ester-amide 6 gave the amide 5 which was difficult to hydrolyse. Acylation gave the monoester 4 but this equilibrated rapidly with the same amide 5. Both gave the same enantiomer of the amino alcohol and of course the maximum yield is only 50%. R O

OH

O

NH2

Ph

R

OH

enzyme

OCOR

enzyme

O

NHCOR

NH NH2

Ph rac-2

Ph

Ph

4

5

rac-6

Finally they rejected the crystallisation of a conglomerate. It was known that the salt between the amine 2 and m-aminobenzoic acid formed a conglomerate that could be crystallised with seeding of either enantiomer (chapter 22) but only from the rather expensive solvent THF. In any case the yield was only moderate and the ees too low to make this method worthwhile.

Methods considered for development and tried experimentally but rejected Two methods were promising enough for serious consideration but were eventually rejected on practical grounds. They both amount to chiral pool strategies (chapter 25) though the first starting material, styrene oxide 7, is not a natural product. Both enantiomers of 7 are commercially available. NH2 OH

O

NH3

Ph

MeOH

7; (R)-(+) styrene oxide

OH NH2

HCl

Ph (R)-2; 70% yield

NH3

Ph

Cl

crystallise

NaOH, H 2O

EtOAc

toluene

(R)-2 Ph 43% yield 99% ee

OH 8; main impurity

(R)-2.HCl

Treatment of styrene oxide with ammonia in methanol gave mostly (R)-2 by nucleophilic attack at the less hindered end of the epoxide. However, some of the regioisomer 8 was formed by attack at the benzylic carbon and this proved difficult to remove. Crystallisation of the salt 2.HCl and neutralisation gave the best samples (99% ee) of (R)-2 but only in 43% yield and still contaminated with about 3.5% of the impurity 8. This is not good enough. Mandelic acid 9, available as both enantiomers, was used to make the other enantiomer, (S)-2, for a change. The cyclic acetal ester 10 reacted with ammonia to give a good yield of the amide 11 but with significant racemisation. Worse, reaction with LiAlH4 gave incomplete reduction even with a large excess of reducing agent so that the product, (S)-2, was contaminated with large amounts (about 30%) of unreacted amide 11. Complete reduction could be achieved with a large excess of borane to give 90% (S)-2, without racemisation, but this is not practical for scale-up.

O

OH

OH

O O

Ph

CO2H

9; (S)-(+)mandelic acid

H2SO4

Ph O 10; 80% yield

NH3 EtOH

Ph

OH NH2

O 11; 87% yield

LiAlH4 THF

Ph

NH2

(S)-2; 60% yield 93% ee

11 + 30% yield

720

31 Strategy of Asymmetric Synthesis

A successful large scale asymmetric synthesis by resolution Initially resolution looked the least promising of all strategies. Resolution with tartaric acid was known but the best that could be done was a 3.7% yield of material with 63% ee. Trials with ten other acids gave better results, e.g.: (figures in brackets are ees after one recrystallisation) camphoric acid (38%), pyroglutamic acid (67%), Boc-alanine (69%). However the best resolutions were with acyl derivatives of tartaric acid and the very best was di-p-toluoyl tartaric acid (12; DPTTA). The salt from technical grade racemic 2 with DPTTA was crystallised from aqueous iso-propanol on 100 g scale giving eventually 80% yield of (R)-2 with 99.8% ee. This has now been scaled up to 9 kg and is the method chosen for production. CO2H

HO2C O

O

O

O

12; di-p-toluoyl tartaric acid (DPTTA)

PART II – (2S,4R)-4-HYDROXYPIPECOLIC ACID A simple heterocyclic acid needed for an anti-HIV drug The HIV protease inhibitor palinavir 13 (from Bio-Méga/Boehringer, Québec) is a complex molecule that can be disconnected simply to five components. Two of these (14 and 18) are simple achiral aromatic compounds and two can be derived from the chiral pool (chapter 25): the amino acid valine 15 and an epoxide 16 derived from phenylalanine. The fifth 17 is the subject of this section. This compound has two chiral centres so we must first produce the right (syn-) diastereoisomer and then the right enantiomer. Ph H N

O

H N

O

N

N H

O

13; palinavir an HIV protease inhibitor

N OH O

N C-N

amide

C-N

H2N

amide

C-C

1,2-diX

C-O

X

CO2H

Ph

CO2H

ether

HN N 14

H2N

CO2H 15; valine

N

O 16

17

OH

18

A chiral pool synthesis from aspartic acid Though (2S,4R)-4-hydroxy pipecolic acid 17 is a natural product it cannot be obtained in sufficient quantity from nature. A chiral pool synthesis from aspartic acid 19 is known involving

31 Introduction: Strategy of Asymmetric Synthesis

721

a Pd-catalysed coupling of a vinyl stannane to an acid chloride.2 This works well, but stoichiometric toxic tin is not acceptable in production. O

O

O O

HO

HO F3C

H2N

1. SOCl 2

CF3

CO2H

O

HN

2.

O

19; (S)-Laspartic acid

F3C

O

O

N O F3C

F3C

CF3 20

CF3 21

OH

H2O

K-Selectride

i-PrOH

[K(s-Bu)3BH]

CF3 22

O

BnPdCl(PPh3)2

O

Et2O.BF 3

O

HN

SnBu3

N

CO2H

N

H

CO2H

H (2S,4R)-17

23

An important point from this synthesis is the transmission of chiral information (chapter 30) from the acid group in 23 to the OH group in 17 by reduction with the bulky reagent K-selectride. This sterically demanding reagent prefers an equatorial approach (chapter 21) and we can therefore deduce that the CO2H group in 23 and both the CO2H and OH groups in 17 are axial. The CO2H group adopts an axial conformation in 23a because the planar and conjugated (to N) Boc group would eclipse an equatorial CO2H group. This information is vital for other syntheses. OH CO2H

O K-Selectride

N Boc

23a

H atom added equatorial

H

[K(s-Bu)3BH]

N Boc

CO2H

17a

Choosing a new reaction to solve the stereochemistry problem A simple reaction that creates both chiral centres in one step was already in the literature.3 It is a cross between a Mannich and a Prins reaction involving the condensation of an unsaturated amine 24 with glyoxylic acid (25; available as an aqueous solution - fortunately the reaction works in water) to give a heterocyclic bridged lactone 26 with the right relative stereochemistry (but of course racemic). OH

O CHO + NH R 24

CO2H 25 glyoxylic acid

heat H2O MeCN

O N

6M HCl reflux

N

R 26

R

CO2H 27

722

31 Strategy of Asymmetric Synthesis

Presumably the aldehyde forms an iminium ion 28 with the amine and this is attacked by the alkene. One way to draw this would be 28 → 29 but it looks better to have concerted attack by the CO2H group on the cation as it is formed 30. At all events, the lactone 26 must have a 1,3-diaxial bridge 26a.

HO

O

O

O

26 N

CO2H

N

R 28

CO2H

R

R

N

R 29

30

N 26a

Making the new reaction asymmetric The workers at Bio-Méga/Boehringer turned this promising reaction into an asymmetric synthesis.4 They first tried resolution of 26 with R ⫽ benzyl as the group ‘R’ on nitrogen must be removable. The amine 24; R ⫽ Bn could be made in only 41% yield - dialkylation was a problem. The new reaction with glyoxylic acid worked well and the lactone 26; R ⫽ Bn could be resolved with a 55% amount (see chapter 22) of a bromocamphor sulfonic acid 31 but the yield was only 30% making it 12% overall from but-3-en-1-ol (see below) and the ee was not perfect (95.4%). However, they did find that the lactone could be hydrolysed to 27; R ⫽ Bn without loss of stereochemical integrity. O HO3S

H

26 + R = Bn

1. crystallise salt from MeOH/EtOH/EtOAc (30.8% yield)

O

6M HCl

N

2. NH 4OH

Br

reflux

27

Bn 26; R = Bn 30% yield, 95.4% ee

O 31

Their method of choice was part asymmetric induction and part resolution. The benzyl group was replaced by the chiral α-methylbenzyl group. This reduced dialkylation in the formation of the allylic amine 32 and allowed modest asymmetric induction in the formation of the lactone: 33 was 60:40 ratio of diastereoisomers. Crystallisation of the salt of 33 with the same sulfonic acid 31 completed the separation of one diastereoisomer and this gave 17 in good yield and essentially perfect ee. Note that the yield of pure 33 (41%) is actually quite good as the crude sample contained only 60% of the right diastereoisomer. The α-methylbenzyl group is removed by hydrogenation but its chirality is destroyed as it gives ethylbenzene. Recent work (discussed in the Workbook) at Chirotech leads to both enantiomers of both diastereoisomers.5 O

O

CHO 1. TsCl, Et 3N

OH

2. Ph

CO2H

NH

NH2

Ph 32

heat H2O MeCN

O N Ph 33; 67-72% yield 60:40 diasts

1. crystallise sulfonate salt with 31

O

2. NH 4OH

N Ph 33; 41% yield 96.4% ee

1. H 2/Pd(OH)2

2. 6M HCl reflux 3. ion-exchange resin

17; 88% yield 99.8% ee

31 Chiral Pool Syntheses from Other Terpenes

723

PART III – GRANDISOL AND SOME BICYCLO [3.2.0] HEPTAN-2-OLS A bicyclic insect attractant used in agriculture A modern approach to dealing with insect pests that wreck crops is to use specific attractants to lure the pest to its doom while offering no threat to beneficial creatures. Compounds such as grandisol 34 are not themselves toxic but they are irresistible to specific insect pests. Grandisol, marketed as ‘grandlure’, an attractant for male cotton boll weevils, is one of a family of unusual monoterpenes containing a four-membered ring. Others are lineatin 35, an attractant for ambrosia beetles that bore into conifers, filifolone 36, and raikovenal 37 (a sesquiterpene). H

H

H

H

O

H CHO

O O

OH

HO

H

34; (+)-grandisol

35; (+)-lineatin

H

36; (–)-filifolone

37; raikovenal

Chiral Pool Syntheses from Other Terpenes An attempt from linalool This subject has recently been reviewed6 and we shall concentrate on the different approaches to grandisol outlined in that review. Many of these approaches can clearly be used for the other compounds. An early approach was a chiral pool strategy from the simpler and commercially available terpene (R)-(⫺)-linalool 38. Though this did not lead to a useful synthesis, it pointed the way to effective methods, particularly the Cu(I)-catalysed photochemical closure of the four-membered ring. OH OH

1. HMPT, heat 2. O 3

H

HO hν

6 steps

O

O

H

H +

CuOTf

38; 3R-(–)-linalool

3. Zn 4. Jones

39

40

HO2C 41

3:7 ratio 42

The photochemical closure by a 2 ⫹ 2 cycloaddition of the two alkenes in 39 gave one diastereoisomer of 40. The cis ring junction is inevitable and the rings fold so that the methyl group at the third centre is exo to the fold though the Cu(I) catalyst may be important for this. The bicyclic ring system in 40 makes it a useful precursor for compounds 35–37. The next step is an unusual elimination catalysed by HMPT [(Me2N)3P]. The main product is the cyclic alkene 44 that gives the required intermediate 42 after ozonolysis but significant amounts of the exo-methylene compound 43 are also produced. H 40

H 2. O 3

1. HMPT

+

heat

42 3. Zn 4. Jones

43

3:7 ratio

44

724

31 Strategy of Asymmetric Synthesis

What really makes this synthesis useless is that it would give the wrong enantiomer of grandisol (see the ring junction stereochemistry). While (R)-(⫺)-linalool is available in 100% ee, the (S)-(⫹)-enantiomer is both more expensive and available in only 64% ee.

An attempt from citronellol using carbene chemistry Citronellol 45 is a useful source of an isolated unfunctionalised chiral centre (chapter 23). It can be transformed into the ester 46 and then into the diazosulfone 47 by standard chemistry.7 Now comes the interesting bit: an alternative to photochemistry for four-membered ring formation. Rhodium carbene complexes can be made from diazocompounds and insert into CH bonds, particularly if a five-membered ring can be formed. This one inserts into the chiral centre of 47 and does so with retention as expected. The centre might appear to be inverted because it is drawn differently in the two compounds. The chain is kinked downwards at the chiral centre in 47 but upwards in 48. Conversion of the MeO group into a leaving group and cyclisation give the [3.2.0] bicyclic system with the right absolute stereochemistry but still having the sulfone attached. Compound 50 was converted into grandisol but the synthesis is very long (11 steps from 46) and contains an elimination step like that in the previous synthesis (from 40) but even less regioselective. Try again. O OH

N2

CO2Me

1. PhSO 2CH2Li

OMe

45; ( R)-(+)-citronellol

OMe

2. NaN 3, HOAc

46; 81% yield O

47 O

H

O

H

SO2Ph

Rh2(OAc) 4

SO2Ph

SO2Ph

SO2Ph

I

OMe 48; 60% yield

49

50

Asymmetric Synthesis Asymmetric addition of an organo-zinc to an aldehyde There are many examples of asymmetric addition of organo-metallic reagents to aldehydes (chapter 27) and the one that works here uses Seebach’s TADDOL 51 as a ligand for zinc.8 The TADDOL 51 is prepared from tartaric acid and used as its titanium (IV) complex to catalyse the addition of the organozinc compound to acrolein. After photochemical cyclisation in the presence of Cu(I), the bicyclic alcohol 53 was isolated in good yield and ⬎98% ee. Ar H O O

Ar O

OH H Ar

Ar

51; a TADDOL Ar = 2-naphthyl

HO

HO

1. ZnCl 2 2. Ti(OPr i)4

OH MgBr

TADDOL 51

hν

3. –30 ˚C acrolein

CuOTf

52; 51% yield ee >98%

H

53; 95% yield

31 A Disappointment and a Resolution

725

The remainder of the synthesis involved destruction of the original chiral centre from 52 and regioselective elimination. Several more steps were needed to give optically pure grandisol. The most interesting are the regioselective oxidation of the silyl enol ether 55 and the ingenious elimination that avoids formation of the exo-methylene compound 43 by an unusual base-catalysed elimination on a tertiary ether 57. This process is too long for further development. O

Me3SiO

H

H

1. O 3

1. LDA

2. NaBH 4

2. Me 3SiCl

3. H +

Cr(VI)

53

O

H

O

H 1. LDA

O

2. H +

1. TsOH (51%)

43 64% yield

57

56; 73% yield

55; 94% yield

54; 69% yield

1. MeLi (71%)

Synthesis by asymmetric dihydroxylation of a non-conjugated diene Since the alcohol 39 is already known as a precursor for the wrong enantiomer of grandisol, an asymmetric synthesis of the right enantiomer would complete an efficient synthesis. The diene 58 is an ideal substrate for a Sharpless asymmetric epoxidation - reaction occurs only at the allylic alcohol (chapter 25). Two stage conversion of the OH into an iodide and treatment with zinc leads to an elimination 60 and the correct enantiomer of the tertiary alcohol 39 for conversion to grandisol.9 O

1. (+)-DIPT, t-BuOOH

OH

OTs

Ti(Oi-Pr)4 etc

Zn/Cu NaI

2. (MeO) 3P, TsCl, DMAP, Et 3N

59; 94% yield, >95% ee

58

O ZnI OH (S)-(+)-39, 80% yield, >95% ee

60

A Disappointment and a Resolution All the methods so far described use either a photochemical cycloaddition or a transition metal carbene complex. Neither is well adapted to large scale production. The workers at Bologna7 sought a new method. They found it in a thermal cycloaddition of a ketene and an alkene. The mixed anhydride 62 gave a single diastereoisomer of the bicyclic ketone 63 in 82% yield simply on heating. No UV light, no metals. OAc O

OH CO2H

Ac2O

OAc

AcOK

61

H

62

O

reflux 4 hours

63; 82% yield

They then tried the same reaction on a single enantiomer of the alcohol 61 but to their great disappointment found only racemic product. It appears that this is indeed a 2 ⫹ 2 cycloaddition of the ketene 64 and this is an achiral molecule. However, reduction of the ketone with LiAlH4 gave

726

31 Strategy of Asymmetric Synthesis

only the endo alcohol 65 (the reagent approaches on the outside of the folded molecule) and this could be resolved efficiently with camphanic acid chloride 66. H

O

H

O

C

OH

LiAlH4

O

THF –65 ˚C to RT

racemic 63

64

O COCl 66; (1S)-(–)-camphanic acid chloride

65; 97% yield

The rest of the synthesis is straightforward but you should notice the catalytic Ru(III) oxidation used both on 65 and 67, the periodate cleavage to replace ozonolysis in the formation of 68 and the Peterson reaction (chapter 15) used to make the alkene 69. This method was used for the synthesis of quantities of both enantiomers of grandisol 34. H

H

OH

O

H

TPAP

1. NH 2NH2

RuCl3/NaIO4

NMO

2. KOH

t-BuOH/H2O

(+)-65

(+)-63

67

O H

H

1. SiMe3

H

MgCl

HO2C

2. SOCl 2

68; 71% yield

LiAlH4

HO2C

Et2O

69; 81% yield

HO (+)-34; 82% yield

PART IV – METALLOPROTEINASE INHIBITORS A group of compounds that protect cartilage from enzymatic degradation Arthritis attacks all of us who live long enough. The stiffness in the joints arises because the collagen that lubricates them is eaten away by over-enthusiastic zinc-dependent enzymes known as metalloproteinases. One possible therapy for arthritis is inhibition of these enzymes by compounds such as ‘Trocade’ 70 and ‘Marimastat’ 71.

O

O

HO

N N H

N H

O

O N O N Me

70; Ro 32-3555 'Trocade' Cartilage Protective Agent

H N

HO

O NHMe

OH

O

71; BB-2516 'Marimastat' Cartilage Protective Agent

31 A Disappointment and a Resolution

727

These drugs are clearly peptide mimics having a genuine amide portion to the right as drawn, a hydroxamic acid to the left and a core that binds zinc and acts as a transition state mimic for the peptide cleavage. They have a C–C bond instead of the amide linkage in the natural substrate so that the drug cannot be hydrolysed by the metallopeptidase. This is easily seen if we disconnect 71a to show the amino acid tert-leucine 73 and a derivative of succinic acid (butanedioic acid) as the core 72. A similar core is present in trocade 70. We shall discuss the asymmetric synthesis of both core succinates. O O HO N H

H2N

O

H N

2 x C–N

NHMe OH

HO2C

CO2H

amides

NHMe

+

OH

O

73; derivative of tert-leucine

72; chiral succinate derivative

71a

Asymmetric alkylation of Evans’s chiral enolates The succinate core can be prepared by alkylation of a simple carboxylic acid (74a or b) with a bromoacetate using the Evans oxazolidinone chiral auxiliary derived from phenylalanine (chapter 27). The branched alkyl group must be present in the substrate for alkylation and the second acid group added in the alkylation so that there is no competition between two carbonyl groups during enolate formation. After hydrolysis (91% yield) the alkylated succinic acid 77 has ⬎95% ee. Notice that the two acid groups are differentiated by this procedure.10

O

O CO2H

CO2H

74a

1. NaHMDS THF, –78 ˚C CO2t-Bu

O

Ph 75

74b O

O

2. Br

N

N

O

LiOH H2O2

t-BuO2C

t-BuO2C 76

Ph

CO2H 77

Direct alkylation of the enolate of 77a; R ⫽ cyclopentyl with the alkyl halide 78 was not very diastereoselective so an alternative route was used to allow equilibration to the more stable isomer.11 Carboxylation of the enolate gave 79a and alkylation of this malonate gave a good yield of 80. The benzyl esters were cleaved by hydrogenation and the malonate decarboxylated to give the required anti isomer of 81 in a 4:1 ratio with its syn diastereoisomer. Conversion into ‘Trocade’ is straightforward.

728

31 Strategy of Asymmetric Synthesis

R 1. 2.02 x LDA 2. ClCO 2Bn

t-BuO2C

CO2H 77a or b

3. BnBr Cs2CO3

t-BuO2C BnO2C

R 1. NaH

t-BuO2C

2. O

Br

CO2Bn

CO2Bn

2. heat

O

N

t-BuO2C

CO2H

1. H 2, Pd/C

O N

N O

BnO2C 79a or b

N

O

O N

N

Me 78

Me

Me 80

81

Asymmetric synthesis from an enantiomerically enriched hydroxy-acid An alternative route to 79b involved nucleophilic displacement from an enantiomerically pure hydroxy acid 82 with a malonate enolate. The OH group was activated as a triflate 83. This method depends on the availability of optically pure hydroxyacid.12 Notice that the ‘wrong’ enantiomer of 82 must be used as inversion occurs during the SN2 reaction leading to 79b.

1. BnBr, Et 3N EtOAc, reflux 2. Tf 2O, pyridine

OH HO

NaH

OBn

BnO2C

TfO

BnO2C

OBn

t-BuO2C

O

O 82; 2-(R)-hydroxyacid

t-BuO2C

83

O

79b

However intermediate 81 is made, the rest of the synthesis involves coupling the free acid to piperidine using a carbodiimide to give 84, removing the t-butyl group to free the other acid, coupling that acid with a protected (O-benzyl) hydroxylamine and removing the protecting group.

t-BuO2C

N

HN

81 EDAC (carbodiimide) HOBt

O

O N O N Me 84

1. Et 3N Me3SiOTf 2. BnONH 2 EDAC HOBt 3. H 2, 5%Pd/C

70 Ro 32-3555 'Trocade' Cartilage Protective Agent

31 A Disappointment and a Resolution

729

PART V – CONFORMATIONAL CONTROL AND RESOLUTION: KINETIC OR NOT? A tetrahydropyran that inhibits leukotriene biosynthesis The leukotrienes are implicated in various inflammatory conditions such as rheumatoid arthritis and inhibiting their biosynthesis is potentially a useful therapy. Some time ago ICI (AstraZeneca now) found that some THPs (TetraHydroPyrans) inhibited the enzyme 5-lipoxygenase in the pathway to leukotrienes and they concentrated on achiral compounds such as 85 (known as ICI D2138). However it soon became clear that an extra methyl group, as in 86, enhanced activity. This small change makes the molecule chiral by introducing two chiral centres. It was necessary to find which diastereoisomer (syn or anti) was more active and which enantiomer of that diastereoisomer was better.13 F

F

OMe

OMe

O

O O

O

O O

N

N

Me

Me

Me 85; ICI D2138

86

The synthesis of 85 had involved the addition of a suitable organo-Li or Grignard reagent made from 87 to the heterocyclic ketone 88 and it seemed sensible to adopt the same strategy for 86. So two questions were asked: would it be better to start with a single enantiomer of 90 and what was the diastereoselectivity of RLi and/or RMgBr additions to 90? F

F 1. BuLi or Mg

ArO

O

OH

2. O

Br

O

Me

ArO

O O

87

89

88

90

Answering the second question first, they found that the close relative 91 added to racemic 90 to give predominantly the anti diastereoisomer of 92 (3:1) while the Grignard reagent gave more of the syn isomer (2:1). The aryl-lithium prefers equatorial while the Grignard reagent prefers axial addition. F

Me O

1. BuLi or Mg 2. 90

BnO

Br 91

Ar

Me O

BuLi - 1:3 HO Mg - 2:1

OH

Ar

syn-92

anti-92

730

31 Strategy of Asymmetric Synthesis

Asymmetric synthesis of 2-methyl-tetrahydropyran-4-one by kinetic resolution Now it is worth making enantiomerically enriched 90. One method already in the literature14 involved reduction of racemic 90 with horse liver alcohol dehydrogenase. This is an enzymatic kinetic resolution (chapters 28 and 29) and at 50% reduction the products are 31% unreacted ketone 90 in good ee, 33% of one enantiomer of the anti-alcohol 93 in perfect (100%) ee, and a trace of the syn-alcohol 93. Horse liver alcohol dehydrogenase

90

O

Me

HO

Me

+

NAD+, pH7, 20 ˚C 50% completion

HO

Me

+

O

O

(S)-(+)-90 31% yield, 85% ee

O

(2R,4S)-(–)-anti-93 33% yield, 100% ee

(2S,4S)-(+)-syn-93 2% yield, 36% ee

Oxidation of (⫺)-anti-93 with PCC gave enantiomerically pure (⫹)-92 and addition of the organo-Li or -Mg compounds would give the two diastereoisomers in this series. As you see the yield of enantiomerically pure 93 is only 33% and, as is often the case with a kinetic resolution, the unreacted ketone has a lower ee. An alternative and more ingenious method gave all the stereochemical information required.13 The racemic dienol 94 was subjected to Sharpless asymmetric epoxidation (chapter 25).15 This is another kinetic resolution run to about 50% completion. Using an excess of di-isopropyl tartrate (DIPT, 1.5 equivalents) one enantiomer of the alcohol (R)-94 remained (72% ee) and one enantiomer of one diastereoisomer of the epoxide 95 (⬎95% ee) was formed. Once again the unreacted starting material 94 has a lower ee than the enantioselectively formed product 95. OH

OH

OH

t-BuOOH, Ti(O i-Pr)4

+

1.5 equivalents (+)-di-isopropyl tartrate, –20 ˚C

(R)-(+)-94 23-38% yield 72%ee

94

O (–)-95 27-33% yield >95%ee

Alternatively, the ‘wrong’ enantiomer of DIPT was used to give the ‘right’ enantiomer of unreacted allylic alcohol (S)-(⫺)-94 (of course also in 72% ee) and this was epoxidised using a catalytic amount (0.2 equivalents) of DIPT. This gave the ‘right’ enantiomer of the epoxide. By these means, good yields of both enantiomers of 95 could be formed and both could then be converted into the various stereoisomers of 86.

OH t-BuOOH, Ti( i-OPr)4

94

1.5 equivalents (–)-di-isopropyl tartrate, –20 ˚C

(+)-95 +

t-BuOOH, Ti( i-OPr)4

(S)-(–)-94 23-38% yield 72%ee

0.2 equivalents (+)-di-isopropyl tartrate, –20 ˚C

(–)-95 75% yield 92%ee

31 A Disappointment and a Resolution

731

The epoxide 95 was reduced with REDAL [NaAlH(OCH2CH2OMe)2] to give the diol 96. Further simple chemistry gave the same (S)-(⫹) enantiomer of 90 as was left unreacted in the enzymatic kinetic resolution. This enantiomer was converted into the two tertiary ethers 86 in this series. OH

OH

REDAL

95

1. O 3, MeOH

1. Et 3SiH Me3SiOTf

2. EtOH, HCl

2. CrO 3

HO 96

EtO

(S)-(+)-90 >95% ee

O 97

The other enantiomer (R)-(⫺)-90 was obtained by resolution of racemic syn-93 via the α-methylbenzylamine salt of the phthalate 98. After chiral HPLC (chapter 22) the ester was hydrolysed and oxidation gave (R)-(⫺)-90 from which the two remaining stereoisomers of 86 could be made. HO

1. phthalic anhydride

O

2. (R)-(+)amine

Me Ph

O NH3

CO2

O

O

98

rac-syn-93

The purpose of this investigation was analytical rather than synthetic and it was established that the (2S,4R) enantiomer of 86, one of the enantiomers with equatorial aryl and equatorial methyl groups, was the most potent compound.13 F

Me O

Me O MeO O O

N Me

Q

O OMe

(2S,4R)-86

(2S,4R)-86; Q = quinolone-5-yl

PART VI – ASYMMETRIC DESYMMETRISATION OF A DIELS-ALDER ADDUCT Ifetroban sodium: a thromboxane receptor antagonist Thromboxanes are human metabolites vital for blood clotting but too much of them is dangerous for patients with heart, lung, or kidney disease. They are also very unstable and the search for stable, orally available, thromoboxane antagonists at the Bristol-Meyers Squibb company revealed the useful compound ifetroban16 sodium 99. It contains three rings: an ortho-disubstituted benzene ring, an oxazole, and a bicyclic ether.

732

31 Strategy of Asymmetric Synthesis

O2C

Na

99 Ifetroban sodium

O

orally bioavailable thromboxane A2 receptor antagonist

O

N O

HN

A laboratory synthesis starting with a Diels-Alder reaction The Diels-Alder reaction between furan and maleic anhydride is reversible and gives the more stable exo-adduct 100 (also commercially available). This compound contains the bicyclic ring system of ifetroban but is achiral and the key problem is to disrupt symmetry of the adduct 100 in a controlled way. The original synthesis used to make the drug17 converted the adduct 100 into the menthyl acetal 101 as a single enantiomer in four steps and then into the carboxylic acid 102 in another six steps. This strategy amounts to a resolution as each menthol adduct could be isolated by crystallisation of a diastereoisomeric mixture in only about 30% yield.

O O

Ar

O O

O O 100

CO2H O

O 101

102

The problem with this synthesis is its great length (23 steps altogether) caused by the repeated changes of oxidation level of the functional groups. The overall yield is less than 3% but even so it was used to make some 20 kg of ifetroban sodium. As the compound became a more likely drug candidate, a more efficient synthesis was clearly needed.

Desymmetrisation of a symmetrical anhydride with a chiral Grignard reagent Since the starting material for all syntheses of ifetroban is bound to be the achiral Diels-Alder adduct 100, one good strategy is to desymmetrise this compound or the saturated version 103 with a chiral nucleophile. One that works is the oxazolidinone 104 derived from ephedrine.18 [You have seen ephedrine used as a chiral auxiliary in chapter 27]. Reaction occurs predominantly at one of the enantiotopic carbonyl groups to give the keto-acid 105. This is immediately reduced and the chiral auxiliary removed to give the lactone 106. Though only one centre is marked, five new stereogenic centres are formed in 106, one by the reduction and the others by desymmetrisation. The sequence is very selective, giving lactone 106 in 99.2% ee after recrystallisation. The aldehyde is used to add the side chain 107 by a Wittig reaction and the carboxylic acid is used to add the oxazole. The problem is that the yield is not very good. The lithium derivative of 104 gives only 50% yield and even the Grignard reagent gives 65% at best.

31 A Disappointment and a Resolution

733

O BrMg 100

O O

O

N

103

Me

O

N

Ph

104

Me

Me

105

1. NaBH 4 2.

O

O Ph

O CO2H

Me

Wittig

H+

O O

O

CHO O

O 106

O

CO2Me 107

An alternative approach17 is to attach a chiral auxiliary, (phenyl ethylamine, available by resolution, chapter 22, and used repeatedly in chapters 22–28, to the anhydride 100 in the form of a chiral imide 108. This auxiliary is very close to the two carbonyl groups of the anhydride in 108 and directs the achiral Grignard reagent 109 to one of them to give adduct 110 immediately reduced and hydrolysed to the lactone 106. One recrystallisation gives 106 in ⬎98% ee. This time a Horner-Wadsworth-Emmons reaction (chapter 15) adds the side chain as a conjugated unsaturated acid reduced to the saturated side chain in 111, a common intermediate in most syntheses of ifetroban. The formation of the oxazole and the rest of the synthesis is described by R. N. Misra and the Bristol-Meyers Squibb team.16 O O 100

H2N

N

Ph

O Me

O

O

Me

O 109 1. DBU

2. 1M NaOH EtOH 3. 10% HCl

O

(MeO)2P

1. NaBH 4

O

CHO

O

N

O O

108; 80-95% yield crystallised

O

OH

Ph + BrMg

(99% yield)

Me 110

HO2C

CO2Me

2. H 2, Pearlman's catalyst(86% yield)

106; 89% yield, 88% ee crude >98% ee recrystallised

O

Ph

CO2H O 111

Laboratory and process routes compared A simple comparison between the laboratory route used by the medicinal chemists and the route devised by the development chemists at Bristol-Meyers Squibb for the production of 99

734

31 Strategy of Asymmetric Synthesis

shows how important is this efficient asymmetric synthesis. The laboratory route used 23 steps and gave 3% overall yield with the asymmetry coming from a resolution. The process route, summarised below, has 12 steps, gives 28% overall yield and derives asymmetry from a substrate-attached chiral auxiliary. You may notice that the oxazole is assembled from the chiral pool (serine, see chapter 25) even though the chirality is lost: serine is just a convenient starting material. O H2N

Ph

O

O

Br

Me

CHO

O

1 step 1 step

O O

Br N

H2N

OH

O

Ph O

O

HO

O

Me

108

109

(S)-L-serine 3 steps

3 steps

MeO2C

O H2N

N H CO2H

HO

O

4 steps

Na

O2C

N O

O

O

99; Ifetroban sodium

HN

PART VII – ASYMMETRIC SYNTHESIS OF A BICYCLIC β-LACTONE Lactacystin: a natural proteasome inhibitor Proteins in the human body have a limited lifetime. They are earmarked for degradation with a marker ‘ubiquitin’ and hydrolysed by proteasomes. Though this is a natural and necessary sequence of events it can get out of hand and proteasome inhibitors could provide treatments for heart diseases, asthma, and arthritis. Lactacystin 112 is a proteasome inhibitor from a soil micro-organism.

31 A Disappointment and a Resolution

735

It works by the spontaneous hydrolysis of the thiol ester to give N-acetyl cysteine 114 and the true inhibitor, the β-lactone 113. This β-lactone has been the target of many syntheses using diverse strategies and these form the last section of this chapter. Notice that the stereochemistry of the OH and CO groups in the β-lactone is the same as that in lactacystin itself since the β-lactone is both formed from 112 and reacts with 114 to give 112 by nucleophilic attack at the carbonyl group. The biology and chemistry of these compounds is summarised in an excellent review.19 There are many syntheses - some very long - and we shall concentrate on the origin of chirality in each. O

O

112 (+)-Lactacystin proteasome inhibitor (inhibits protein degradation)

NH H O

S

thiolester hydrolysis

+

OH

O

CO2H O

O 113 (+)-Lactacystin β-lactone: the active proteasome inhibitor

H N CO2H O

H N

H

OH

HO

SH

NH spontaneously in vivo

114 N-acetyl cysteine

Strategies used in the synthesis of lactacystin, the β-lactone, and analogues The thiolester portion is always derived from natural cysteine so there is a chiral pool element present in all syntheses. Because there are four consecutive chiral centres in the main portion of the molecule, just about every strategy has been employed by someone: more chiral pool, chiral reagents and catalysts, chiral auxiliaries attached to substrate or reagent, and sometimes combinations of these.

A chiral pool synthesis of lactacystin from glucose An early synthesis of lactacystin, not via the β-lactone, used glucose as starting material.20 Disconnection to 115 (the vinyl group can be converted into CO2H and the benzyl ether into CHO) gives a structure that can be redrawn as 116 and then, with incorporation of an extra C atom (to be removed in the synthesis by diol oxidation) derived from 117. OH Me

1 CHO

N R

112/113

Me

2

OBn HO

HO 115

HO

HN

2 HN

1

R OBn

3

Me

HO

116

R OBn

O 3 117

It was already known that glucose could be converted into the cyclic ether 118 and a [3,3] sigmatropic rearrangement (Overman style, see chapter 19) on the allylic imidate ester 119 should give 120, a derivative of 117. The ‘should’ has to be there because it is not possible to forecast the stereochemical outcome with certainty. Me

O O

OH

?

OBn

Me

O

O

CCl3

O O

O

118

Me

O

HN

?

O

OBn

O OBn 119

COCCl3

NH 120

736

31 Strategy of Asymmetric Synthesis

In the event, a 1:1 mixture of E:Z isomers of 119 gave a 4.8:1 mixture of 120 and its diastereoisomer at the new centre in 60% yield. This is a remarkable level of control but later syntheses do better. Deprotection, oxidative cleavage and cyclisation gave the pyrrolidine 121. O 1. Jones oxidation

1. TFA, H 2O

120

115; R = COCCl 3

2. NaIO 4, MeOH

Me

NH

2. NaBH 4

OBn HO 121

There was a further problem in the addition of i-PrMgBr to the aldehyde (R ⫽ t-BuMe2Si-) derived from 121. The chemo- and stereoselectivity were poor and the product contained 35% 123a, 30% 123b, and 21% alcohol from the reduction of 122. Though the unwanted products could be recycled: 123a by oxidation and re-reduction and the alcohol by oxidation to 122 it is clearly possible to do better. O

O

Me

NH

i-PrMgBr

Me

O Me

NH

CHO RO

OH

RO 122

NH

+

121

OH

RO

123a

123b

Chiral pool syntheses of lactacystin from amino acids Amino acids are attractive starting materials for lactacystin and there are several published syntheses using this type of chiral pool strategy. Baldwin21 used glutamic acid 124 with the idea of using the known Seebach-style intermediate 125 (chapter 24) with a relay chiral centre and destroying the original centre to give a pyrrole 126 that is also an extended enolate. An aldol reaction of 126 with i-PrCHO might be controlled to give the correct aldol product 127. O

OR

O

O Ph

Ph OH

Me

N

NH2

?

N O

HO

H

OH 124 glutamic acid

Me

aldol reaction of extended enolate

O

O

H

Ph

N

O

125

126

127

The conversion of 125 into the pyrrole 126 uses reactions from earlier chapters. Methylation is not stereoselective but that doesn’t matter as selenenylation, oxidation, and elimination gives a single internal alkene – this chemistry will be explained in chapter 33. Now silylation destroys the original chiral centre from glutamic acid but keeps the relay centre. O 125

1. LDA

O Ph

Me

N

2. MeI

1. LDA

Ph

Me

N

2. PhSeBr

O

base, CH 2Cl2

O

3. O 3, heat

t-BuMe2SiOTf

H

H

128; 95% yield

129; 79, 87% yields

Me

OSiMe2t-Bu Ph N O 130; 89% yield

31 A Disappointment and a Resolution

737

The aldol reaction proved difficult but a Lewis acid (SnCl4) catalysed reaction at low temperature gave a 9:1 ratio of aldols in favour of 127 which could be isolated in 55% yield after chromatography. Notice that the aldehyde has added to the same face as that ‘blocked’ by the phenyl group, as we should expect from chapter 24. The alkene in 127 must now be removed, the remaining OH group introduced, and the last two chiral centres in the pyrrolidine ring controlled. These all present problems but the extra OH is the worst. Baldwin found an ingenious solution: two OHs were introduced diastereoselectively by a racemic dihydroxylation of the acetate 131. The two OHs go on the right side, opposite the Ph and aldol groups, and the diol 132 is cyclised to the thiocarbonate 133 (Im ⫽ imidazole). O

O Ph

Me

N

Ac2O

127

Me

AcO

N

HO

OsO4

O

pyridine room temperature

O Ph

NMO acetone H2O

131; 99% yield

Im2CS

O

HO AcO

THF

Ph

Me O

N O

S

O AcO

132; 87% yield

133; 91% yield

The thiocarbonate is used to get rid of the unwanted OH group in a radical reaction using Bu3SnH to initiate a radical chain The methyl centre is epimerised in this operation but treatment with base gives the right diastereoisomer 135 and also hydrolyses the acetate. The original chiral auxiliary can now be removed by hydrogenation and 136 should be close enough to lactacystin for you to see that the synthesis can be completed. O

O

O Ph

133

Bu3SnH AIBN toluene reflux

Me HO AcO

Ph NaOH Me

N O

H2, Pd/C

N

MeOH

O

HO HO

Me HO HO

135; 94% yield

134; 94% yield

NH

HCl, MeOH

OH

136; 87% yield

Corey’s earlier synthesis22 uses a related strategy from serine involving a Seebach-style relay chiral centre and two stereoselective aldol reactions. The aldol reaction with i-PrCHO is carried out first and the pyrrolidine ring assembled later with the second aldol reaction. Serine 137 is converted to the cyclic t-BuCHO derivative 138. An aldol reaction with the enolate of 138 and i-PrCHO gives mostly the wanted aldol 139. The aldehyde adds to the face opposite to the t-butyl group. Ph Ph NH2 HO

CO2H

137; (S)-serine

1. LDA, LiBr

O

N

CO2Me 138; 9:1 cis:trans

O

N

2. i-PrCHO

MeO2C

OH

139 77% yield >98% this isomer

This oxazolidine ring is hydrolysed to give an open chain diol that can be selectively protected with the large t-BuMe2Si group 140. Treatment with formaldehyde now gives a different oxazolidine

738

31 Strategy of Asymmetric Synthesis

141 and hence the aldehyde 142 ready for the next aldol reaction. This aldehyde cannot be enolised so it is a good electrophile. Ph 1. TfOH, MeOH 91% yield

139

TBDMSO

HN

Ph

TsOH

2. TBDMSCl 97% yield

N

(CH2O)n

OH

Ph

2. Swern

MeO2C

MeO2C

TBDMSO

140

N

1. LiBH 4

O

O

OHC TBDMSO

141; 96% yield

142; 85% yield

Heathcock’s diastereoselective aldol method (chapter 4) gave the required anti-aldol 144 but only in low yield. Removal of the N-benzyl group allowed the amine to cyclise onto the aryl ester and the bicyclic compound 145 of fixed conformation is formed. Notice that all the substituents are on the outside (exo face) of the folded molecule (chapter 21). The further development of 145 into lactacystin can be found in the paper.22 Ph

O N

O

O

1. LDA 2. 142

O

H2, Pd/C

ArO2C

Me

OR N O

HO HO

OR

anti-144; 48% yield

143

O

Me

EtOH 23 ˚C

OH

O

N

H

H H conformation of 145

OR

145; 87% yield

The reagent strategy: asymmetric epoxidation and asymmetric allylboration We move into approaches that are more sophisticated in several ways. The synthesis by Omura and Amos B. Smith23 was partly inspired by the biosynthesis of lactacystin and partly by an attempt to make other diastereoisomers. Their key intermediate is the non-proteinogenic amino acid hydroxyleucine 147 and they were able to make both enantiomers of both diastereoisomers of this compound by asymmetric synthesis. O amide and aldol

NH

NH2

H2N

H

OH OH OH

HO O

O

= HO2C OH

OH

(2R,3S)-(+)-147 3-Hydroxyleucine

146

Sharpless asymmetric epoxidation (chapter 27) of the allylic alcohol 148 with cumyl hydroperoxide and (⫹)-di-isopropyl tartrate (DIPT) gave the epoxide 149 in excellent yield and enantiomeric purity. The other enantiomer could be made simply by using the other enantiomer of DIPT. The problem is now: how to react the epoxide regiospecifically with a nitrogen nucleophile? NH2 OH

HO

Ph

148

O Ti(Oi-Pr)4 (+)-DIPT

HO

?

O 149; 82% yield, >97% ee

HO2C OH (2R,3S)-(+)-147

31 A Disappointment and a Resolution

739

A cunning solution was to attach the nucleophile to the OH as an isocyanate and to use the anion 150 to react intramolecularly with the epoxide 149. Initially the anion must cyclise to the nearer end of the epoxide to give 151 but even under the reaction conditions 151 starts to isomerise to 152. This isomer 152 is not the alternative product from 150: it is formed by attack of the secondary alcohol on the C⫽O group of 151. Further treatment with NaH completes the isomerisation to 152. Ph NaH

149

O

PhNCO

HO

O

N

Ph N

NaH

O

O O

OH

150

Ph

THF heat

151 (5:1 with 152)

N O

O 152; 75% yield from 149

Oxidation and esterification give one diastereoisomer of protected 3-hydroxyleucine 153 and epimerisation in base gives another 154. It is the latter that is required for the synthesis of lactacystin but either enantiomer of either diastereoisomer can be made by this sequence. The overall yield of natural 3-hydroxyleucine 147 is about 50% from the allylic alcohol 148. MeO2C 1. Jones (100% yield)

152

2. CH 2N2 (87% yield)

Ph

HO2C KOH

N

Ph

EtOH

HO2C 1. KOH

N

O

O

O

2. H 2 Pd(OH)2

O anti-154; 96% yield

153

H2N OH (2R,3S)-(+)-147 98% yield

A few small changes give the oxazoline needed for the stereoselective aldol reaction with formaldehyde. There is no syn/anti aldol question here: the electrophile simply adds to the face of the more or less planar enolate opposite the isopropyl group. Now there was a severe difficulty in an apparently trivial reaction: oxidising the primary alcohol 156 to the aldehyde 157. The advantage that the aldehyde is not enolisable is a disadvantage too as nucleophiles may attack and remove the CHO group. The problem was solved by careful choice of reagents (Moffat oxidation with DMSO, DCC and CF3CO2H). The aldehyde 157 was not isolated but used immediately in the next step. OH

MeO2C 1. HCl, MeOH (100% yield)

147

1. LHMDS

N

2. PhC(OMe) 3

O Ph 155

2. CH 2O

CO2Me

OHC

CO2Me

?

N

N O

Ph 156; 85% yield >98% this diastereoisomer

O Ph 157

Brown’s crotyl borane 158 (chapter 24) provides reagent control through a chair-like sixmembered ring in the formation of 159. Oxidation of the alkene (ozone with oxidative workup) gives the free acid marked in 160 and on removal of the benzylic group the freed amine (circled) cyclises to it to give the pyrrolidone required (147 as its methyl ester). The synthesis can easily be completed from there.

740

31 Strategy of Asymmetric Synthesis

OH

OH

CO2Me

HO2C

157

Ipc2B

O HCO2H Pd, NH 3

NH

N

158

CO2Me

N O

H

O

Ph 159

HO

Ph 160

OH CO2Me

(2R,3S)-(+)-147 (methyl ester)

Enzymes as reagents in the synthesis of lactacystin analogues Many developments have come from Corey’s laboratories in the lactacystin area since the synthesis described above.24 One strategy we have not mentioned before is the use of an enzyme, pig liver esterase (chapter 29) in the selective hydrolysis of one of two enantiotopic ester groups in the malonate 161. An MeS group is used to block enolisation and prevent racemisation of the product 162. The mono ester 162 is initially formed in 67% ee improved by one crystallisation of the quinine salt to 95% ee. The pyrrolidine ring 164 is made in an unusual way by first forming the amide 163 and then cyclising the diester by carbonyl condensation. The new chiral centre in 164 is not controlled but disappears in the next step. O O Me MeS

CO2Me CO2Me

CO2H

1. PLE, pH 7.3

Me

2. crystallise quinine salt

MeS

161

1. (COCl) 2 2. RNH 2

CO2Me

PMB

Me

N

Me MeS

162; 62% yield 95% ee

LDA

CO2Me CO2Me 163; 99% yield

N

MeS

PMB

O CO2Me 164; 93% yield 1:1 diastereoisomers

Now a remarkably simple aldol reaction with formaldehyde (compare 156 to 165) needing only DBU as base gives mostly (9:1) the correct diastereoisomer 165. Easier to explain is the stereoselective reduction of the ketone with acetoxy borohydride to give 166: the primary alcohol replaces one of the acetates and directs the H atom to the bottom face of the ketone. Juggling the protecting groups gives 167 ready for the introduction of the remaining chiral centres. O

O

Me 164

DBU

N

MeS

CH2O, H 2O –78 ˚C

O

PMB OH

O

Me NaBH(OAc) 3

MeS

CO2Me

HO

165; 90% yield

N

Me

PMB OH

N

MeS

CO2Me

TBDMSO

166; 95% yield, 99% ee

PMB OH

CO2Me 167

First, the MeS group was removed with Raney nickel giving the correct stereochemistry at the new centre in 168. The stereochemistry actually comes from the protonation step and presumably arises because EtOH donates an H atom to the less hindered side of the desulfurised enolate opposite the OTBDMS group. Then oxidation with the Dess-Martin periodinane gives the vital aldehyde 169. O

O

Me 167

Raney Ni EtOH

H

N

PMB OH

Dess-Martin periodinane

Me H

N

PMB CHO

TBDMSO

CO2Me

168 (10:1 diastereoselectivity)

TBDMSO

CO2Me

169; 78% yield from 167

31 A Disappointment and a Resolution

741

In previous syntheses this last centre has usually been added in an aldol reaction and here too a carbon-carbon bond is made as the centre is introduced. But this time it is the addition of isopropenyl Grignard reagent in the presence of Me3SiCl that is required to give 171. The high stereochemical control suggests a chair-like intermediate 170 with both carbonyl groups coordinated to a magnesium atom and the isopropenyl group being transferred to the top face of the CHO group as drawn. Further adjustments lead to the hydroxyacid 172 which was, in this synthesis, transformed into the β-lactone 113 on the way to lactacystin 112. The isopropenyl Grignard was used as isopropyl Grignard annoyingly reduced the aldehyde rather than adding to it, as in the reaction of 122. H OMe O Mg O

PMB

BrMg

N

O

169 Me3SiCl

Me

O

O

Me

Me

PMB N

H

PMB N

H

OR

TBDMSO MeO2C OSiMe3

170

171

HO HO2C OH 172

This synthesis includes the formation of the β-lactone using a phosphorus-based coupling reagent ‘BOP-Cl’ 173 bis(2-oxo-3-oxazolidinyl)phosphinic chloride. Notice the preferential formation of the fused rather than the spirocyclic β-lactone and that cyclisation occurs with retention so the phosphorus reagent 173 must activate the CO2H group. The p-methoxybenzyl (PMB) group is removed by oxidation with Ce(IV) to give 113. This approach from the enantiomerically pure 165 is ideally suited for the production of analogues and Corey has made many analogues to probe the origin of the biological activity.24 O

O O

O N

O

Cl P

N

O

Me 172

O

N

H

Et3N, CH 2Cl2

173; BOP-Cl bis(2-oxo-3-oxazolidinyl) phosphinic chloride

Me

PMB Ce(NH4)2(NO3)6

NH

H

MeCN, H 2O

O

O

O OH 174; 90% yield

O OH 113; 62% yield

Syntheses of 3-hydroxyleucine by other strategies A leading group in lactacystin research is led by Panek at Boston University. They have produced two further approaches to a key intermediate, 3-hydroxyleucine 147. One uses a chiral auxiliary strategy.25 Phenylglycinol 175 is available as either enantiomer. It can be alkylated chemoselectively on nitrogen to give the amino ester 176. Trapping this with Ph2CHCHO produces the Seebach-style oxazoline 177 with a cis relationship between the old and the new (relay) chiral centres in 100% yield. The key step is a stereoselective aldol of the lithium enolate of 177 with i-PrCHO. Reaction occurs entirely on the face of the enolate opposite the two large groups (as drawn in 178) and very much in favour of the anti-aldol product 178. In this aldol reaction, the original chiral centre is not removed and the relay centre reinforces the stereoselectivity while holding the molecule in a fixed conformation. The rest of the synthesis is simply the removal of the various auxiliaries and protecting groups.

742

31 Strategy of Asymmetric Synthesis

OH

OH

Ph

O

Ph

N

Ph

OHC

Ph

MeO2C

NH2

175 (S)-phenylglycinol O

Ph

N

Ph

Br

Ph

NH

Et3N, THF, 0 ˚C

Ph

Ph

1. LDA, –78 ˚C

MgSO4, CH 2Cl2

MeO2C 176; 92% yield

2. i-PrCHO

MeO2C 177; 100% yield

OH NH2

Ph

HCO2H

Ph

NH

H2, Pd/C

MeO2C

OH (2R,3S)-(+)-180 3-Hydroxyleucine methyl ester

MeO2C

OH 178; 90% yield >30:1 anti:syn aldol

(2R,3S)-(+)-147 3-Hydroxyleucine

MeO2C

THF/H2O

OH 179; 90% yield

The other uses a catalytic process, Sharpless’s aminohydroxylation26 (chapter 25). By skilful choice of the esterifying group, the amino hydroxylation of the unsaturated ester 181 could be controlled to give the required regioisomer 182 in high ee after two recrystallisations from aqueous ethanol (it was initially 87%). CbzNH 5 mol% K2[OsO2(OH)4]

O

5 mol% (DHQ)2AQN = 1,4-bis(dihydroquininyl)anthraquinone

O

Br

O

CbzNCl.Na, PrOH, H 2O

O

Br

181

OH

182; 60% yield, >99% ee

The methyl ester of the ‘wrong’ diastereoisomer of 3-hydroxyleucine 183 required merely deprotection of 182 and ester exchange best catalysed by Ti(IV). Incorporation into the oxazoline 184 prepares the way for the aldol reaction with formaldehyde that inverts the stereochemistry at the ester centre and leads to their lactacystin synthesis.27

182

2. H 2, Pd/C MeOH

N

PhC(OMe)3

MeO2C

183; 100% yield

Ph

CO2Me N

1. LHDMS

TsOH

OH

HO

CO2Me

NH2

1. MeOH Ti(OPr-i)4

O

184; 85% yield

2. CH 2O THF, –78 ˚C

Ph

O 185; 85% yield

A large-scale synthesis of lactacystin analogues using AD and an asymmetric aldol reaction We end with a large scale synthesis by LeukoSite Inc. of Cambridge Mass.28 designed to be used for analogues of natural lactacystin. They started with the asymmetric dihydroxylation (chapter 25) of the unsaturated ester 186 (cf. compounds 148 and 181). This gave a 94% yield on a small scale with AD-mix-α but could not easily be adapted to a larger scale because of the excessive quantities of salts in over large volumes of solvent. They preferred a modification from PharmaciaUpjohn at Uppsala using the organic re-oxidant N-methylmorpholine-N-oxide (NMO) instead of potassium ferricyanide, no potassium carbonate and a more concentrated solution.29 This gave 187 in only 60% yield but ⬎99% ee after two recrystallisations. All three Sharpless methods have now been used in lactacystin synthesis.

31 A Disappointment and a Resolution

743

187 60% yield after recrystallisation >99% ee

OH

1 mol% K2OsO2(OH)4 1 mol% (DHQ)2PHAL

MeO2C

1.6 equiv NMO

MeO2C

OH

H2O/t-BuOH

186

Now one of the two OH groups must be replaced by NH2 with retention of configuration. This can be achieved by the ortho-ester method like that summarised in chapter 23 for propylene oxide. The intermediates 188 and 189 can be isolated but the process works better if 189 is hydrogenated to give the oxazoline 184 in 89% yield over the three steps. Note that the stereochemistry of 189, and hence 184, comes from a double inversion at the position α to the ester group.

Br

1. PhC(OMe) 3 BF3.OEt2

187

MeO2C

N3 NaN3

MeO2C

2. MeCOBr Et3N, 0 ˚C

H2

MeO2C

DMSO

OCOPh

N

Pd(OH)2/C

O

OCOPh

188 (98% yield)

Ph 184; 89% yield

189 (85% yield)

The formation of 184 from 189 needs some explanation. Hydrogenation converts the azide into an amino group to give 190 and this is in equilibrium with 192 by transfer of the benzoyl group from O to N via the tetrahedral intermediate 191. Dehydration of 191 gives the stable heterocycle 184. NH2 H2

189

Ph

MeO2C

MeO2C

Pd(OH)2/C

O

O

HN

O

O

Ph

Ph 190

NH

TsOH toluene

184

MeO2C OH

OH 191

192

The oxazoline 184 provides an attractive approach to lactacystin as it is a protected form of 3-hydroxyleucine. The other half of the molecule was made in the LeukoSite synthesis by a very different method: the alkylation of an Evans’ chiral auxiliary. This was chosen partly because they wished to vary the alkyl group on the pyrrolidone ring and we use the propyl compound as example. The phenylalanine derived oxazolidinone 193 (chapter 27) was acylated and then the titanium enolate of 194 was alkylated to give 195 with very high selectivity and the chiral auxiliary removed to give the simple acid 196.

O

NH O 193

Ph

Ph

Ph

1. BuLi 2. BuCOCl

1. TiCl 4 i-Pr2NEt

O

N

O O 194; 97% yield

2. BOMCl (BnOCH2Cl)

LiOH.H2O

O

N

OBn

O O 195; 85% yield pure after recystallisation

H2O2

HO2C

OBn

196; 90% yield

Coupling with Et2N using TBTU [2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate], removal of the benzyl group and oxidation with the Dess-Martin periodinane

744

31 Strategy of Asymmetric Synthesis

gave the aldehyde 199 in good yield. This sequence from 193 gave optically pure aldehyde 199 in 70% yield without any chromatography.

Et2NH i-Pr2NEt

196

Et2N

OBn

Et2N

H2, Pd(OH) 2/C

Dess-Martin

OH

Et2N

O

TBTU

O 197; 98% yield

O 198; 100% yield

O H 199; 96% yield

Now the two parts, 184 and 199, must be linked in a controlled aldol reaction. The lithium enolate 200 reacts with the aldehyde in 199 in the presence of excess Lewis acid Me2AlCl to give one diastereoisomer of 201 as the only product. Presumably the aluminium coordinates the tertiary amide and aldehyde oxygen atoms to hold the aldehyde in one Felkin conformation while it is attacked by the lithium enolate from one face only as sketched in 202. Al O OMe

LiO

2. Me 2AlCl 1. LiHMDS

184 –80 ˚C

E

Pr

2.2-2.4 equiv

N

LiO NR2

O

1.25 x 199

O

Et2NOC OH

Ph 200

N

H

H

O

OMe

N

Ph 201; E = CO 2Me

202

O

Ph

Any attempts to purify 201 by chromatography led to reverse aldol reactions and it was better not to isolate 201 but just to wash it with aqueous bisulfite before removing the oxazoline and allowing the amine 203 to cyclise to form the pyrrolidine 204. An X-ray structure of 204 confirmed that the stereochemistry was correct. O

201 washed with NaHSO3

E

55 psi H2

NH

Et2N

Pd(OH)2/C 1:10 AcOH/H2O

O

room temp

OH

NH2

H

OH

OH CO2Me 204; 79% yield from184 recrystallised, X-ray HO

203; E = CO 2Me

Hydrolysis of the ester with NaOH and β-lactone formation gave the propyl analogue 206 of 113 in ⬎20% yield from simple starting materials over ten steps. O 203

NaOH H2O

O Et3N

NH

O

H

4 ˚C

OH CO2H

HO 205

Cl

75% yield recrystallised

NH H

O

OH

O O 206

10 steps from i-PrCHO >20% yield

31 References

745

References General references are given on page 893 1. O. Lohse and C. Spöndlin, Org. Process Res. Dev., 1997, 1, 247. 2. A. Golubev, N. Sewald and K. Berger, Tetrahedron Lett., 1995, 36, 2037. 3. S. J. Hays, T. C. Malone and G. Johnson, J. Org. Chem., 1991, 56, 4084. 4. J. Gillard, A. Abraham, P. C. Anderson, P. L. Beaulieu, T. Bogri, Y. Bousquet, L. Grenier, I. Guse and P. Lavalée, J. Org. Chem., 1996, 61, 2226. 5. R. C. Lloyd, M. E. B. Smith, D. Brick, S. J. C. Taylor, D. A. Chaplin and R. McCague, Org. Process Res. Dev., 2002, 6, 762. 6. E. Marotta, P. Righi, and G. Rosini, Org. Process Res. Dev., 1999, 3, 206. 7. H. J. Monteiro and J. Zukerman-Schpector, Tetrahedron, 1996, 52, 3879. 8. K. Langer and J. Mattay, J. Org. Chem., 1995, 60, 7256. 9. E. Balmer, A. Germain, W. P. Jackson and B. Lygo, J. Chem. Soc., Perkin Trans. 1, 1993, 399. 10. R. P. Beckett, M. J. Crimmin, M. H. Davis and Z. Spavold, Synlett, 1993, 137. 11. K. F. McClure and M. Z. Axt, Bioorg. Med. Chem. Lett., 1998, 8, 143. 12. M. J. Broadhurst, P. A. Brown, G. Lawton, N. Ballantyne, N. Borkakoti, K. M. K. Bottomley, M. I. Cooper, A. J. Eatherton, I. R. Kilford, P. J. Malsher, J. S. Nixon, E. J. Lewis, B. M. Sutton and W. H. Johnson, Bioorg. Med. Chem. Lett., 1997, 7, 2299. 13. G. C. Crawley, M. T. Briggs, R. I. Dowell, P. N. Edwards, P. M. Hamilton, J. F. Kingston, K. Oldham, D. Waterson and D. P. Whalley, J. Med. Chem., 1993, 36, 295. 14. J. A. Haslegrave and J. B. Jones, J. Am. Chem. Soc., 1982, 104, 4666. 15. W. R. Roush and R. J. Brown, J. Org. Chem., 1983, 48, 5093. 16. R. N. Misra, B. R. Brown, P. M. Sher, M. M. Patel, S. E. Hall, W.-C. Han, J. C. Barrish, D. M. Floyd, P. W. Sprague, R. A. Morrison, R. E. Ridgewell, R. E. White, G. C. DiDonato, D. N. Harris, A. Hedberg, W. A. Schumacher, M. L. Webb and M. L. Ogletree, Bioorg. Med. Chem. Lett., 1992, 2, 73. 17. R. H. Mueller, S. Wang, P. D. Pansegrau, J. Q. Jannotti, M. A. Poss, J. K. Thottathil, J. Singh, M. J. Humora, T. P. Kissick and B. Boyhan, Org. Process Res. Dev., 1997, 1, 14; see also J. C. Barrish, J. Singh, S. H. Spergel, W.-C. Han, T. P. Kissick, D. R. Kronenthal and R. H. Mueller, J. Org. Chem., 1993, 58, 4494. 18. S. D. Real, D. R. Kronenthal and H. Y. Wu, Tetrahedron Lett., 1993, 34, 8063. 19. C. E. Masse, A. J. Morgan, J. Adams and J. S. Panek, Eur. J. Org. Chem., 2000, 2513. 20. N. Chida, J. Takeoka, N. Tsutsumi and S. Ogawa, J. Chem. Soc., Chem. Commun., 1995, 793. 21. H. Uno, J. E. Baldwin and A. T. Russell, J. Am. Chem. Soc., 1994, 116, 2139. 22. E. J. Corey and G. A. Reichard, J. Am. Chem. Soc., 1992, 114, 10677. 23. T. Nagamitsu, T. Sunazuka, S. Omura, P. A. Sprengeler and A. B. Smith, J. Am. Chem. Soc., 1996, 118, 3584. 24. E. J. Corey, W. Li, and T. Nagamitsu, Angew. Chem., Int. Ed., 1998, 37, 1676; E. J. Corey, W. Z. Li, T. Nagamitsu and G. Fenteany, Tetrahedron, 1999, 55, 3305. 25. J. S. Panek and C. E. Masse, J. Org. Chem., 1998, 63, 2382. 26. A. J. Morgan, C. E. Masse and J. S. Panek, Org. Lett., 1999, 1, 1949. 27. J. S. Panek and C. E. Masse, Angew. Chem., Int. Ed., 1999, 38, 1093. 28. F. Soucy, L. Grenier, M. L. Behnke, A. T. Destree, T. A. McCormack, J. Adams and L. Plamondon, J. Am. Chem. Soc., 1999, 121, 9967. 29. L. Ahgren and L. Sutin, Org. Process Res. Dev., 1997, 1, 425.

Section E: Functional Group Strategy 32. Functionalisation of Pyridine ............................................................................................ 749 33. Oxidation of Aromatic Compounds, Enols and Enolates ............................................... 777 34. Functionality and Pericyclic Reactions: Nitrogen Heterocycles by Cycloadditions and Sigmatropic Rearrangements ......................................................... 809 35. Synthesis and Chemistry of Azoles and other Heterocycles with Two or more Heteroatoms.................................................................. 835 36. Tandem Organic Reactions ............................................................................................... 863

32 Functionalisation of Pyridine This chapter examines functionalisation strategy - the addition of functional groups to a pre-formed carbon skeleton. It revises previous sections in the context of some difficult reactions.

Introduction PART I – THE PROBLEM Low reactivity of pyridine compared with benzene at carbon Reaction at nitrogen prevails Formation of stable complexes with electrophiles N-Nitro Heterocycles as Nitrating Agents PART II – TRADITIONAL SOLUTIONS : ADDITION OF ELECTRON-DONATING SUBSTITUENTS Amino and alkoxy groups: synthesis of flupirtene Regioselectivity in Electrophilic Substitution with Electron-Donating Groups The Anti-Tumour Antibiotic Kedarcidin Halogenation and Metallation Pyridine N-Oxides in Electrophilic Substitution Synthesis of nifluminic acid and omeprazole Ortho-Lithiation of Pyridines Diazines The Halogen Dance Tandem Double Lithiation: The asymmetric synthesis of camptothecin Tandem Lithiation of Pyridine N-Oxides and Nucleophilic Substitution PART III – SURPRISINGLY SUCCESSFUL DIRECT ELECTROPHILIC SUBSTITUTIONS Sulfonation and halogenation at the 3-position in strong sulfuric acid The reaction with SOCl2: reaction at the 4-position, synthesis of DMAP PART IV – SUCCESSFUL NITRATION OF PYRIDINE Nitration in the 3-position with N2O5 and a sulfur nucleophile Scope, limitations and mechanism Sulfonation of Pyridines Extension by Vicarious Nucleophilic Aromatic Substitution Synthesis of Imidazo[4,5-c]pyridines PART V – APPLICATIONS The Baeyer anti-cancer drug BAY 43-9006 Anatoxin analogues Epibatidine analogues Polycyclic compounds Katritzky’s improvement of Bakke nitration

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

750

32 Functionalisation of Pyridine

Introduction A theme of this section is making reactions work. Usually that means making reactions that already work quite well work much better. But we start with a reaction - electrophilic aromatic substitution on pyridines and related heterocycles - that doesn’t really work at all. We shall see why not and how the reaction can be successfully realised. This involves some serious mechanistic thinking and questions of regioselectivity are vital to the solution. This is new chemistry with a great future.

PART I – THE PROBLEM Because it is such a familiar reaction, it can be forgotten that attack by electrophiles on benzene 1 is difficult because it gives an intermediate cation 2 which has lost the aromaticity of the starting material. This is, after all, why Lewis acids are needed for many such reactions. The product 3 has regained aromaticity but this is no help to the rate-determining step, the attack of the electrophile and the formation of 2. That the reaction occurs at all is because 2 is a delocalised pentadienyl cation and has some intrinsic stability as a cation. Successful electrophilic substitution of benzene by HNO3/H2SO4, fuming H2SO4, RCOCl and AlCl3, or halogens with various Lewis acids is also testimony to the extreme reactivity of these combinations of reagents. H

E

E

E slow

fast

1

2

E = NO2; HNO 3 + H2SO4 E = Cl, Br, I; Hal 2 + Lewis acid E = COR; RCOCl + AlCl 3

3

Pyridine has all the disadvantages of benzene plus some special ones of its own. Attack at the 2- or 4-positions 4 gives intermediate cations such as 5 which have part of the positive charge delocalised onto nitrogen 5c. This would be no bad thing if the nitrogen atom were tetravalent like stable NH4⫹, but it is actually divalent like the unknown, electron-deficient, and presumably very unstable NH2⫹.

E N

N 4

5a

H

N

E

5b

H E

N 5c

H E

Attack at the 3-position 6 is not so bad as the intermediate cation 7 is no longer delocalised onto the electron-deficient nitrogen atom but is not as stable as the benzene intermediate 5 and the slow step is very slow because the HOMO of pyridine is lower in energy than the HOMO of benzene. H

E

H E

H E

E

slow

N 6

N 7a

N 7b

N 7c

32 N-Nitro Heterocycles as Nitrating Agents

751

So far electrophilic attack on pyridines sounds only a bit more difficult than that on benzenes but there is worse to come. The reagents for nitration, halogenation and so on all require strong protic or Lewis acid catalysts and these go for the pyridine nitrogen atom fi rst. Attempted nitration would have to occur on the cation 8, making the intermediate an impossibly unstable dication 9.

H

NO2 HNO3

N

H2SO4

NO2 NO2

N

N

H 8

H 9

?

N 10

N-Nitro Heterocycles as Nitrating Agents In fact, nitration of pyridine in the absence of protic or Lewis acid catalysts occurs at nitrogen.1 The reagent has to be the rather unstable and moisture sensitive salt NO2⫹ BF4⫺. This is commercially available, though expensive. The product is also not very stable but can be isolated as the tetrafluoroborate salt 11 as BF4⫺ is non nucleophilic. This compound 11 is itself a nitrating agent and can be used to nitrate benzenes (but obviously not pyridine) simply by mixing in acetonitrile. The products 12 have the usual substitution pattern for electrophilic substitution. Olah has called this process ‘transfer nitration’.

NO2

NO2 BF4

+

N

R MeCN

N

R

NO2 BF4 11; 91% yield

12; >90% yield

A better method for the same reaction (nitration of benzenes) is to use the product of nitration of pyrazole2 13. When the nitration is carried out in the absence of strong acid, the product is the N-nitro compound 14, a stable neutral molecule because a proton can be lost from the other nitrogen atom. This compound is stable but reacts with the Lewis acid BF3 to give an intermediate 15 that nitrates benzene efficiently. It is probable that the C-nitration of pyrazoles also takes place via such an intermediate as 15, see below.

HNO3, Ac 2O

N N H 13

HOAc

N NO2 14

benzene

N

N BF3

N NO2 15

NO2

BF3 89% yield

Pyridine can also be used as an efficient catalyst for the bromination of benzene (see Vogel, page 860) through reaction at the nitrogen atom 16 while the crystalline reagent pyridinium

752

32 Functionalisation of Pyridine

bromide perbromide 17 is a convenient way of brominating alkenes. These are convincing illustrations that pyridine does not react with bromine at carbon.

Br

Br2

Br N

pyridine catalyst

16 Br2 HBr

N

Br

cyclohexene

N

Br

H

Br

Br

Br

17

18

Direct electrophilic substitution of pyridines is not, in general, a useful reaction. Only 3-substitution is remotely possible as the intermediate 7 at least does not attempt to put the positive charges on the same atoms! The maximum yield of 3-nitro-pyridine by direct nitration is 5%, normally 3% is expected. If we want to carry out nitration on the pyridine ring, some means must be found better than normal electrophilic substitution. We shall now explore various ways in which this simple reaction can be achieved either by subterfuge or mechanistic ingenuity. The same remarks apply to other electron-deficient six-membered aromatic heterocycles such as pyrimidine 20 and quinoline 22. Solutions to these problems are needed as many drugs and natural products contain these rings. The Pharmacia-Upjohn anti-AIDS drug PNU-142721 19 contains pyridine and pyrimidine rings while the coenzyme methoxatin 21, that allows bacteria to live off methane as their sole carbon source, contains an oxidised quinoline ring. All three rings are heavily substituted and preparation of these compounds needs regiochemically controlled substitution around the rings.

HO2C

O

NH S

N Me

H

N N

NH2 19; PNU-142721

Cl

CO2H

N N

O

N

CO2H

N

O 20; pyrimidine

21; methoxatin

22; quinoline

PART II – TRADITIONAL SOLUTIONS: ADDITION OF ELECTRON-DONATING SUBSTITUENTS The analgesic Flupirtine 23 is a simple pyridine with three substituents at the 2, 3, and 6 positions. Removal of the amide shows the core 24. The 4-fluorobenzylamine 25 could be added by nucleophilic substitution (easier in pyridines than in benzenes) and we shall delay the choice of the leaving group (X in 26) for the moment. The only amino group we could conceivably add by nitration is the one in the 3-position so we might continue by FGI (reduction) and C⫺N disconnection 27.

32 Regioselectivity in Electrophilic Substitution with Electron-Donating Groups

H N

N H F

N

NH2

OEt C–N

O NH2

C–N

N H

amide

F

23; Flupirtene NH2

753

NH2

N

NH2

24 NO2

FGI

nitration

+ X

F 25

N 26

N 27

X

NH2

NH2

X

N 28

NH2

The nitro group will help the nucleophilic aromatic substitution so now we must choose the leaving group X in 27. It need not be a good leaving group - the slow step in nucleophilic aromatic substitution is the addition of the nucleophile -so as long as X is a tolerable leaving group, it will do. It should promote electrophilic attack by NO2⫹ in the 3-position and 28; X ⫽ OMe should be available commercially. Putting X ⫽ OMe fulfils all these conditions. Here is the synthesis.3 NO2

NO2 HNO3

MeO

N

NH2

H2SO4

28; X = OMe

25

MeO

Ar

NH2

N

27; X = OMe

N N NH2 H 29; Ar = 4-F-C 6H4

1. H 2 Raney Ni

23 2. ClCO 2Et

Nitration of 28; X ⫽ OMe is successful because both electron-donating OMe and NH2 groups activate the 3-position and the intermediate cation 31 is delocalised over both N and O atoms. We could have used lone pairs on either the OMe or NH2 groups to draw the mechanism 30. The reaction occurs in spite of the pyridine ring rather than because of it. The position of attack - C3 or C5 - is not easily predicted and searching the literature or trial and error is needed. H NO2

(+) H NO2

NO2 27

MeO

N 30

NH2

MeO

N 31a

NH2

MeO (+)

N

NH2 (+)

31b

Regioselectivity in Electrophilic Substitution with Electron-Donating Groups You might think that nowadays there would be no doubt about the results of such reactions as the nitration of 28 since NMR would quickly distinguish the possible products. This is not so as a notorious case shows. Though it was known that reaction of 33 with iodine in aqueous sodium carbonate gave the 2-iodo compound 34 in good yield, it was claimed in 1990 that a more complex reagent gave the 6-iodo isomer 32. Reinvestigation showed that 34 was the only product though the alternative 32 could be formed in moderate yield at higher temperature. It was also known that chlorination gives the 2-chloro compound 35. The preference for the site next to nitrogen (and therefore para to the OH group) may be due to steric crowding of the solvated oxyanion in water.4

754

32 Functionalisation of Pyridine

?

OH

OH

OH

OH

I2

NaI, NaOCl

Cl2

33 I

MeOH, H 2O

N

K2CO3 H2O

N

32 (6-iodo)

N

K2CO3 H2O

I

N

34 (2-iodo)

33

Cl 35

The synthesis of the drug PNU-142721 19 required a series of electrophilic substitutions on 3-hydroxy pyridine5 33. Iodination of the chloride 35 went in the previously inaccessible 6position to give 36. Now iodine-lithium exchange and reaction with acetaldehyde gave the secondary alcohol 37 and finally iodination gave the tetrasubstituted pyridine 38. Each new substituent is added as an electrophile in a specific position. Only the acetaldehyde had no choice. I HO K2CO3

HO

1. LiH 2. BuLi

I2, H 2O

35 Cl

3. MeCHO

I

N

HO I2, H 2O

Cl

OH

N

36; 71% yield

K2CO3

Cl

OH

N

Me

Me

37; 62% yield

38; 62% yield

Installation of the furan ring requires two extra carbons provided by a Sonagashira coupling with trimethylsilyl acetylene to give 39. Cu(I)-catalysed closure of the furan ring follows and now the role of the chlorine atom on the pyridine ring is revealed. It is there to block the most reactive position and, its job complete, it can now be removed by hydrogenation. SiMe3 Me3Si (Ph3P)2PdCl2 CuI, Et 3N

CuI Et3N

HO

35

cyclohexene

O

EtOH

CH2Cl2, THF

Cl

OH

N

Cl

OH

N Me

Me 39; 81% yield

O

20% Pd(OH)2 MeOH reflux

40; 86% yield

OH

N Me

41; 77% yield

Completion of the synthesis of PNU-142721 19 requires introduction of asymmetry and coupling with a pyrimidine thiol. Asymmetry was introduced in two ways. The alcohol 41 was enantioselectively acylated with trifluoroethyl butyrate catalysed by porcine pancreatic lipase (chapter 29) or else a Swern oxidation gave the ketone that could be enantioselectively reduced with (⫺)-Dip chloride (chapter 24). Clean inversion from (S)-41 to the chloride (R)-42 was achieved only with the Mitsunobu-like reagent Ph3P/CCl4. Finally, coupling with the thiolate anion of the pyrimidine 43 gave a second clean inversion.5 One recrystallisation improved the ee of the drug 19 from 97.6% to over 99%.

O

HS

O

N

Ph3P

OH

N Me (S)-41

H

CCl4

Cl 1. NaH, DMF

Cl

N Me

H

(R)-42; >97% ee 75–89% yield

N

2. ( R)-42

NH2 43

(–)-(S)-19 97.6% ee 89% yield

32 The Anti-Tumour Antibiotic Kedarcidin

755

The Anti-Tumour Antibiotic Kedarcidin Another incorrect structure was revealed by synthetic work aimed at structure 44, alleged to be a part structure (the ‘chromophore’) of the antitumour antibiotic kedarcidin. The synthesis was successful but revealed that the structure was wrong as the compound is a β-amino acid and not an α-amino acid as had been assumed.6 i-PrO

OH

i-PrO O

MeO HN

OMe Cl

OH O

MeO CO2Me

HN

OMe

N

CO2Me

N Cl

HO

OH 44; alleged kedarcidin chromophore (1993)

45; kedarcidin chromophore (1997)

The obvious compounds to make are the pyridine amines for coupling to the undisputed naphthalene fragment. Both were made from the same pyridine 35 that featured in the last section. Iodination and protection gave two iodo-pyridines 46; R⫽TBDMS and MOM. The first was combined with Jackson’s d3 reagent 47 (chapter 13) in a Negishi coupling to give 48 and the second in a Heck reaction to give the unsaturated ester 49 to which Davies’s chiral lithium amide (chapter 24) could be added to give a 97:3 ratio of 50 and its diastereoisomer. 1. t-BuOCl, NaI

Cl

N

I

NHBoc

IZn

H2O, MeCN

+

35 2. TBDMSCl imidazole

RO 46; R = TBDMS 71% yield

CO2Me

Cl

1. PdCl 2(PPh3)2 MeCONMe2 2. HF, NaF, pH5

N

CO2Me

HO

47; d 3 reagent

48; 73% yield Ph Ph

46 R = MOM 75% yield from 35

CO2t-Bu

Cl

NHBoc

CO2t-Bu

N

BnN

BnN Cl

CO2t-Bu

N

Li Pd(OAc) 2, Et 3N DMF, 60 ˚C

RO 49; R = MOM; 90% yield

THF, –80 ˚C

RO 50; R = MOM; 83% yield

Deprotection of 48 gave the amine needed for the synthesis of 44: this proved not to be the natural product. The synthesis of the correct structure6 45 had a curious twist: removal of the benzyl groups from 50 resulted in dechlorination 51 and the chlorine was reinstated 52 with another electrophilic substitution having the same regioselectivity as that of 35.

50 R = MOM

NHCOCF3

1. H 2, Pd(OH) 2/C MeOH, HOAc, H 2O 2. CF 3CO2Et Et3N, MeOH 3. t-BuOH, TsOH

N

NHCOCF3

CO2t-Bu

Cl

N

NCS DMF

HO 51; 68% yield

HO 52; 82% yield

CO2t-Bu

756

32 Functionalisation of Pyridine

Halogenation and Metallation To summarise these last two examples: with even one electron-donating group on the pyridine ring, halogenation becomes a useful reaction. Metallation with lithium or palladium-catalysed couplings with alkenes (Heck), organo-zinc compounds, including those with β-hydrogens (Negishi), or boronic acids (Suzuki) make new C⫺C bonds while Buchwald-Hartwig chemistry makes new bonds to O or N atoms. These are obviously very versatile sequences.

Pyridine N-Oxides in Electrophilic Substitution This is all very well if your target molecule happens to have electron-donating substituents on the pyridine ring, but what if it hasn’t? The traditional solution is to use a pyridine N-oxide 53 which simultaneously blocks the lone pair on nitrogen and provides a strongly electron-donating substituent. Pyridines are easily oxidised to N-oxides with H2O2 (another instance of electrophilic attack at N) and the substituted N-oxides 54 are easily reduced back to the pyridines 55 with trivalent phosphorus compounds such as PCl3.

N

E

E

H2O2

PCl3

E

O

N O 53

N

N 54

55

Nitration can be carried out with the usual reagents and occurs in the 4-position (the 2- and 6-positions being sterically hindered by the solvated oxido group). The electrons from oxygen are used in the reaction 56 and the nitrogen atom retains its positive charge throughout but the intermediate 57 is only a mono-cation unlike 9. This is a good synthesis of 4-nitropyridine 59. O

H

N O

O

N

NO2 NO2

O

N

56

O

NO2

PCl3

N

57

N 58

59

The deoxygenation of the product can be developed in synthetically useful ways. If there is an electron-withdrawing group at the 3-position and PCl3 is used, deoxygenation is accompanied by chlorination at the 2-position 62. This new chlorine is activated towards nucleophilic aromatic substitution both by its position on the pyridine ring and by the electronwithdrawing group at the 3-position and can be replaced by O, N, S, or P nucleophiles 63. Pyridines such as 61 are easily made from available nicotinic acid, pyridine-3-carboxylic acid 60. O OH N 60; nicotinic acid

R

O

R

PCl3

Cl

N

N 61

O

O

O

62

R

Nu

Nu

N 63

This result suggests that the initial interaction of the N-oxide with PCl3 is nucleophilic attack by O⫺ 64. The intermediate can then either lose POCl3 65 or add chloride. When nucleophilic attack

32 Pyridine N-Oxides in Electrophilic Substitution

757

by chloride 67 occurs, it is the most electrophilic atom which is attacked - the one between the pyridine nitrogen and the electron-withdrawing group. Aromaticity is restored by the loss of a proton from the 2-position 68. This is a sort of nucleophilic substitution occurring after electrophilic substitution. The leaving group (Cl2PO⫺) is not at the site of nucleophilic attack. E

E

O

O R

N

N

O

O

PCl3 64

N

55

65

R

O

PCl3

66

H

Cl

N

O

PCl3

O

N

Cl

O

PCl2

67

R 62

PCl2 68

Good use is made of this chemistry in the synthesis of nifluminic acid 69, an analgesic with a pyridine ring. Disconnection of the C⫺N bond suggests a 2-chloropyridine 70 starting material easily derived from nicotinic acid 60 and a simple aromatic amine 72 available by functionalisation of trifluorobenzene 74. 69 Is actually available as niflumic acid from Aldrich. CO2H

CO2H

CO2H

CO2H N N

NH

Cl

N O 71

70

C–N

N

NH2

60

NO2

CF3 69 nifluminic acid

CF3

CF3 72

CF3

73

74

The only mild surprise in the synthesis is that the PCl3 reaction turns the acid into an acid chloride 75 which must be hydrolysed before the amine is added, otherwise an amide would be formed.7 COCl

1. HNO 3 H2SO4

74

H2O2

72 2. H 2/Pd/C HOAc

PCl3

1. H 2O

71

60

N

Cl

2. 72

TM69 nifluminic acid

75

These N-oxides can also be used to functionalise side chains in pyridines 76. The reaction is rather similar to the PCl3 reaction in that acetic anhydride attacks the oxygen atom of the N-oxide and the released acetate ion can deprotonate 77 at an alkyl group in the 2-position. The resulting enamine 78 can sacrifice the weak N⫺O bond between the now neutral heteroatoms and form a much stronger C⫺O bond by a [3,3] sigmatropic rearrangement to give 79.

N O

76

Ac2O

H

N O

OAc

[3,3]

N O

77

O 78

N OAc

O 79

758

32 Functionalisation of Pyridine

This chemistry, preceded by an electrophilic substitution on the N-oxide, finds a most important application in the synthesis of omeprazole, Astra(Zeneca)’s anti-ulcer drug and one of the best selling medicines of recent years. Omeprazole 80 is a sulfoxide and has a 2,3,4,5-tetrasubstituted pyridine linked through the sulfur atom to a benzimidazole.8 OMe Me

Me

N

O S

N HN

80 omeprazole

OMe

The sulfoxide 80 can be made from the sulfide 81 and an obvious C⫺S disconnection suggests two much simpler heterocyclic starting materials, a pyridine 82 and a benzimidazole 83. It is the pyridine 82 which concerns us now.

Omeprazole: Analysis 1 OMe Me

OMe Me

FGI

C–S

80

S

N

oxidation

Me

Me

OMe

N + HS

N

Cl

N HN

N H

OMe

81

82

83

Using the chemistry we have just described, the CH2Cl group in the 2-position could be introduced by functionalisation of the trimethyl pyridine N-oxide 86 since only the methyl group in the 2-position should be affected. The group in the 4-position might be introduced by electrophilic attack on the N-oxide but there is no good reagent for “MeO⫹” as we shall see in the next chapter. Nitration of the N-oxide 86 works well, however, and nucleophilic substitution is so good on pyridine N-oxides that this is a reasonable strategy.

Omeprazole: Analysis 2 OMe Me

NO2 Me

N O 84

Me

FGI

Me

Me N O 85

Me

C–N

Me FGA Me

Me N O 86

Me

Me N 87

Me

C–C

Me

Me N 88

3,5-Dimethylpyridine 88 is available and the methyl group in the 2-position can be added by the remarkable reaction in which MeLi appears to displace a hydride ion. It is not obvious that the nitration of 86 will occur in the 4- rather than the 6-position but O⫺ is a bigger substituent than it appears to be because of solvation. The published synthesis gives only “NO2⫹” as the reagent and does not say how it was formed.

32 Ortho-Lithiation of Pyridines

759

Omeprazole: Synthesis 1 The difficult nucleophilic displacement of the NO2 group by methoxide must be carried out while the N-oxide is still present. The leaving group in this reaction is nitrite ion, NO2⫺, which is reasonably stable but you will recall that merely tolerable leaving groups are acceptable in a reaction in which the addition step is rate-determining providing that they accelerate that step. The nitro group certainly does that. The N-oxide has now completed its duties and functionalisation of the side chain (2-Me) by the acetic anhydride reaction follows to give 89. OMe

NO2 Me

Me

MeLi

Me 1. H 2O2

N

Me

OMe Me

N

2. NO 2

Me

MeOH

N

Me

Me

N

O 84

O 85

87

Me Ac2O

MeO

88 THF

Me

Me

OAc 89

Acetate is not a good enough leaving group for an SN2 displacement (note the difference between SN2 and aromatic nucleophilic substitution) so it must be transformed into the chloride 82 before reaction with the thiolate anion of 58.

Omeprazole: Synthesis 2 OMe

OMe SOCl2

89

Me

Me

83

mCPBA

Me

Me

81 NaOH

N

Cl

62

N 80 omeprazole

O S

N HN

OMe

The final oxidation of achiral sulfide 62 to sulfoxide 80 with achiral mCPBA must of course produce racemic omeprazole. Omeprazole is actually a pro-drug - it is transformed enzymatically into the active compound - and in this process the chirality at sulfur is lost. The two enantiomers have nearly the same biological effects though the more readily absorbed (S)-omeprazole is now marketed as esomeprazole.

Ortho-Lithiation of Pyridines Halogenation and replacement of halogen (Br or I) by a metal is a useful way to extend the scope of electrophilic substitution on pyridines already having an electron-donating substituent (usually based on O or N). Another valuable way is ortho-lithiation (chapter 7), that is replacement of a hydrogen atom by lithium ortho to a similar substituent. The requirements are not quite the same. The substituent must be an ortho-director: it must acidify the ortho proton and/or direct the base (usually BuLi) to the ortho position. The simplest ortho directors are F and OMe but the best are oxazolines, amides, and carbamates. There are more details of these processes on benzene rings in chapter 7. If one of these groups is in the 4-position 90 or 93 lithiation 91 is unambiguous and any electrophile (E⫹) that reacts with an aryl lithium reacts in the 3-position 92 and 94.

760

32 Functionalisation of Pyridine

OMe

OMe H

OMe Li

BuLi

CONR2 E

E

CONR2 E

1. BuLi 2. E

N 90

N

N 92

91

N 93

N 94

The relative acidities of the positions around the pyridine ring9 without any assistance from an ortho-director are 700:72:1 for positions 4:3:2. It may seem odd that the position next to nitrogen is least acidic but the sp2 lone pair on nitrogen has an antibonding interaction with the C-Li bond 97. Though BuLi may be directed by the nitrogen 96 to C-2 the unstable product 97 soon equilibrates to the more stable C-4 Li derivative 98.

700

H

Li H

72 N

N

H

H

Li

1

95; acidity ratios

N

Li anti-bonding

Bu

96

N

97

98

Regiospecificity demands ortho-direction. A surprisingly effective group is a carboxylic acid CO2H that acts as its lithium derivative CO2Li. One equivalent of BuLi is used to make this 100 from, say picolinic acid 99, and three equivalents of LiTMP to make the C-Li derivative 101. Reaction with electrophiles such as CO2 or carbonyl compounds occurs at C-3 102. Li

3x LiTMP

BuLi

CO2H

CO2

O N

CO2H

N

99

CO2Li

N

100

CO2H

N 101

OLi

102; 85% yield

Iso-nicotinic acid 103 must also react at C-3 but nicotinic acid 105, R ⫽ H shows the regioselectivity expected from the relative acidity of C-1 and C-4 and gives substitution at C-4 whether C-1 is substituted 105, R ⫽ Cl, or unsubstituted 105, R ⫽ H. The same result is found for amides or oxazolines as ortho-directors.10 CO2H

CO2H 1. BuLi

CO2H

65% yield

CO2H

2. LiTMP 3. CO 2

2. LiTMP 3. CO 2

N 103

CO2H 1. BuLi

N

73% yield

104

N

R 105

1. BuLi 2. LiTMP 3. CO 2 69% yield

CO2H N

Cl 106

This behaviour also occurs with quinolines, e.g. 107 and 109. BuLi cannot be used as it attacks quinolines as a nucleophile. This problem is also found with other heterocycles, particularly those with more than one heteroatom.11

32 The Halogen Dance

761

CO2H CO2H

OMe

OMe CO2H

CO2H

1. LiTMP

1. LiTMP

2. CO 2

N

108; 80% yield

107

CO2H

N

N

2. CO 2

N

CO2H

110; 60% yield

109

Diazines The three diazines, pyridazine, pyrimidine and pyrazine, are metallated well providing there is good ortho-direction to prevent nucleophilic addition. The pyridazine (OMe as ortho-director) 111 and pyrazine 113 derivatives can be metallated12 and their zinc derivatives coupled with aromatic rings (Negishi coupling) to give 112 and 114. It is unusual to see Cl in 113 acting as such an efficient ortho-director. OMe

OMe 1. BuLi 2. ZnCl 2

N N OMe

N

N

N

3. 2-bromo pyridine 4% Pd(PPh3)4

OMe 112; 77% yield

N

3. PhI 4% Pd(PPh3)4

N

111

Cl

N 1. BuLi 2. ZnCl 2

Cl

N

114; 99% yield

113

N-Formyl amines such as DMF make good electrophilic CHO donors and in reaction with the pyrimidine provide a tandem double coupling of electrophilic formylation followed by nucleophilic substitution by the released amine anion.13 The OMe group is the ortho-director. OMe N Cl

OMe

N

N

Li

N

1. LiTMP

Cl

O

N

115

OMe

OMe

OHC

CHO

N Cl

N

N

116

CHO

N

O

117

N 118

The Halogen Dance The iodo-fluoro-pyridine 119 is lithiated by LDA to give the only possible lithium derivative 120. But iodine is a poor ortho-director and the compound equilibrates by the ‘halogen dance’ to put the Li atom next to one of the best ortho-directors: fluorine.14 Presumably one molecule of 120 removes I from 119 to give 123 from which the 3-I atom is removed by another molecule of 120 initiating a chain.

Li I

I Li

119

F

I

Me MeI

LDA

N

I

I

I

N 120

F

N 121

F

N

F 122

N 123

F

762

32 Functionalisation of Pyridine

This product was used by Comins in his synthesis of mappicine. The antiviral alkaloid is the alcohol formed by borohydride reduction of mappicine ketone 124. Strategic disconnections 124 split the molecule into a quinoline 125 and a pyridone 126 and correspond to a simple SN2 and a Heck reaction. If the Heck reaction is performed last it will be intramolecular and hence regioselective. O

O

Heck reaction Br

N

Br

N

N

N O H 126 the pyridone

O nucleophilic substitution

125 the quinoline

124; mappicine ketone

Each component is made by lithiation of a heterocycle. Chlorine-directed lithiation of the commercially available quinoline 127 goes at C-3 and reaction with formaldehyde gives the primary alcohol15 128. Reaction with PBr3 replaces both OH and Cl with Br to give 125. Cl

N

Cl

N

1. LDA

N

PBr3

Br

OH

2. CH 2O

127

Br 125

128

The pyridine 122 reacts with BuLi by I/Li exchange: capture of the lithium derivative with propanal and oxidation gives the ketone 130 in the heterocyclic equivalent of a Friedel-Crafts acylation. The fluorine atom that was originally present to initiate the halogen dance from 120 to 121 can now be hydrolysed to give the pyridone 126. O

HO

I Me

Me

Me

1. BuLi 2. EtCHO

N

O

F

N

N

F

129; 84% yield

122

Me HCl

PCC

F

130; 84% yield

H2O

N O H 126; 92% yield

The quinoline 125 and the pyridone 126 are coupled by a base-catalysed SN2 reaction. Formation of the anion of the pyridone ensures attack at nitrogen (chapter 35). It only remains to close the ring by a Heck reaction.16 O

O Me

t-BuOK

126

125

N

Pd(OAc) 2 KOAc

Br N

N 131

Bu4N Br MeCN

O 132; 81% yield

O

124 57% yield

32 Tandem Double Lithiation: The asymmetric synthesis of camptothecin

763

Tandem Double Lithiation: The asymmetric synthesis of camptothecin The closely related camptothecin 133 is more complicated because it is chiral. Close analogues (topotecan and irinotecan) are in use as anti-cancer drugs. As these compounds differ only by substitution on the quinoline, synthesis of the enantiomerically pure pyridone 135 is obviously of prime importance. O

O HO

Heck reaction

O

HO

O

I

N

Cl

N

N

N O H 135 the pyridone

O nucleophilic substitution

134 the quinoline

133; ( S)-camptothecin

Comins, synthesis of 135 uses an ingenious double lithiation strategy. Treatment of 2-methoxy-pyridine 137 with the very hindered lithiating agent mesityl-lithium 136 gives clean ortho-lithiation 138. Addition of Me2NCHO would give the 3-aldehyde but Comins uses the N-formyl diamine 139 to give, before work-up, the lithium alkoxide 140 with the diamine still attached. Me2N

Li Me

Me

Li

+

+ Me N

Me

OMe

N

137 2-methoxy pyridine

136; mesityl -lithium

Me2N

OMe

Me

N

N

OLi

CHO N

138

139

OMe 140

A molecule of BuLi is now guided to C-4 of 140 by the chelating diamine to give 141. Reaction with iodine, reduction, demethylation and acetal formation give 143 in reasonable overall yield after a short sequence from 137. Me2N Li BuLi

140

I

I N

Me

1. I 2

OLi

2. H 2O NaBH4 CeCl3

N OMe 141

OH N

OMe

142; 46% yield

Me3SiCl NaI (CH2O)n MeCN

O N

O

143; 87% yield

Now the asymmetric addition of the TCC 145 ester of 2-keto butyric acid (substrate strategy, see chapter 27) to the lithium derivative of 143, prepared by I/Li exchange, gave the alkoxide 144 and hence, on work up in acidic solution, the pyridone17 135. Coupling to the quinoline 134 was achieved by the same SN2-Heck sequence used in the synthesis of 124.

764

32 Functionalisation of Pyridine

O LiO

1. BuLi

OTCC

143 O

2.

HCl H2O

i-PrOH

OH

135 60% yield 93% ee

Ph

O OTCC

145 TCC alcohol

N O 144

Tandem Lithiation of Pyridine N-Oxides and Nucleophilic Substitution We have seen how pyridine N-oxides can be used to promote electrophilic oxidation at C-2,4, and 6. They also promote ortho-lithiation and nucleophilic substitution making them very versatile intermediates. This is well illustrated in Quéguiner’s synthesis of the antibiotic caerulomycins. 2,2’-Bipyridyl, available as a ligand for many metals, is easily oxidised to its monoxide 146. The synthesis starts with two electrophilic substitutions. Lithiation occurs ortho to the N-oxide: quenching with BrCN gives a good yield of the bromide 147 and a conventional electrophilic nitration occurs para to the N-oxide. NO2 KNO3

N

N

Br

N

80 ˚C 20 hours

2. BrCN

O

H2SO4

N

1. LDA

O

Br

147; 70% yield

146

N

N O

148; 66% yield

Next a nucleophilic substitution: the displacement of the poorish leaving group nitrite by methoxide to give 149. The N-oxide had done its work so it is removed with PBr3 and a Negishi coupling with MeZnCl completes the skeleton of the caerulomycins18 151. Oxidation of the methyl group to an aldehyde with SeO2 or phenyl seleninic anhydride gives caerulomycin E: its oxime is caerulomycin A.

OMe MeONa MeOH

145

OMe

OMe MeZnCl Pd(PPh3)4

PBr3

Br

N

N CHCl3 reflux

Br

N

N

O 149; 84% yield

150; 90% yield

THF reflux

Me

N

N

151; 90% yield

A limited use of direct electrophilic substitution combined with halogen/lithium exchange or ortho-lithiation allows us to make a variety of pyridines and quinolines. When the N-oxides are used as well, the scope is even wider. Nucleophilic substitution extends this still further. In the next chapter (33) we shall see that electrophilic oxygen can also be added to pyridines. What is missing is the direct formation of 3-substituted pyridines: this is dealt with in the next section.

32 Tandem Lithiation of Pyridine N-Oxides and Nucleophilic Substitution

765

PART III – SURPRISINGLY SUCCESSFUL DIRECT ELECTROPHILIC SUBSTITUTIONS When, as in the omeprazole and caerulomycin syntheses, an N-oxide has a number of functions, it is worth building in this extra functionality. Only 4-substituted pyridines can be made from the N-oxide: the 3-isomers must be made available by direct electrophilic substitution on pyridines themselves and fortunately some direct methods do work, though it is not immediately clear why they do! Direct sulfonation19 and bromination20 both work well in very strong sulfuric acid, giving the 3-sulfonic acid 152 and 3-bromopyridine 153 in good yield. SO2OH

very concentrated H2SO4 catalytic HgSO4 220 ˚C, 24 hours

N 152; 70% yield

Br

Br2, 130 ˚C very concentrated H2SO4

N

N 153; 86% yield

It seems strange that reactions are successful under very strongly acidic conditions when the pyridine will be completely protonated: one clue to this mystery is that “very strong sulfuric acid” is made by adding SO3 to concentrated sulfuric acid. Another is the mechanism of an important reaction of pyridine with SOCl2 which gives a double pyridinium salt21 158. Initial electrophilic attack at nitrogen 154 is followed by nucleophilic attack by a second pyridine at the 4-position 155. Loss of a proton 158 then re-aromatises the first pyridine and the remains of the SOCl2 must disappear as SO(?) and chloride ions.

N N N

N

N

H

2Cl

1. SOCl 2

N

O

2. EtOH

Cl

S

Cl O

154

N

N

S

S

Cl

155

O

N

N

H

Cl

156

157

158

The final product 158 is useful because the pyridinium group can be displaced by nucleophiles. An important example is the reaction with DMF which gives22 the acylation catalyst DMAP (pronounced ‘D-map’, 4-DiMethylAminoPyridine) 159, and, presumably CO and HCl. NMe2 1. SOCl 2

Me2N-CHO

158 N

2. EtOH

140–160 ˚C

N

159 DMAP DiMethyl Amino Pyrdidine

If the double salt 158 is heated with sodium sulfite (Na2SO3) in water, aromatic nucleophilic substitution 160 leads to the sodium salt of pyridine-4-sulfonic acid 162 or to the sulfonic acid itself 163 after acidification. These are nucleophilic substitutions following electrophilic attack at nitrogen.

766

32 Functionalisation of Pyridine

O

O

N O

S

N

SO2O

SO2OH

N

N

162

163

O

O

N H

O S

N H 160

161

The products of many of the reactions in this section, e.g. 152, 153, and 163, look as though they have been made by straightforward electrophilic substitution on the pyridine ring. But they haven’t. In the next section we explore electrophilic substitutions that really work.

PART IV – SUCCESSFUL NITRATION OF PYRIDINE We now come to the successful nitration of pyridines in the 3-position by an indirect procedure.23 The reagent is N2O5, dinitrogen pentoxide, a kind of “nitric acid anhydride”, a white crystalline molecular solid 164 that ionises in polar solvents such as nitromethane to NO2⫹ and NO3⫺. Treatment of pyridine with N2O5 in nitromethane followed by aqueous work-up with any of a variety of sulfites (SO2, Na2SO3 or NaHSO3) gives high yields of 3-nitro-pyridine 165. The contrast with previous methods could hardly be greater. The maximum yield from direct nitration is 5% - here we get 77% of 3-nitro pyridine 165. O

O

O

O

N

N

N

N

NO2

1. N 2O5, MeNO 2

O

O O 164 N2O5 white crystalline solid

O

O

O N

2. NaHSO 3, H 2O

N 165; 77% yield

in polar solvents

Reaction at pyridine is actually preferred to reaction at benzene - easily illustrated by the fate of 4-phenyl pyridine 167 under the two sets of conditions. Traditional nitration occurs only on the benzene ring 166 with the (presumably protonated) pyridine ring acting as a meta-directing deactivating substituent. With N2O5, nitration of the (presumably unprotonated) pyridine ring is preferred and occurs in the 3-position 168.

NO2 N

H2SO4

166

1. N 2O5, MeNO 2

HNO3

N

2. NaHSO 3, H 2O

167

N

NO2 168

You may have partly guessed what is actually happening from the earlier parts of the chapter. Electrophilic attack occurs at nitrogen and the N-nitro pyridinium salt 169 is the product from the first stage. If the reaction is carried out at low temperature, this salt, similar to the fluoroborate 11 already known from the work of Olah, can be isolated.24 The role of the bisulfite ion is as nucleophile. Attack actually occurs reversibly in the 2- and 4-positions but it is the former intermediate 171 that leads to the final product 165. If bisulfite ion is added to the salt 169 in D2O, the proton

32 Tandem Lithiation of Pyridine N-Oxides and Nucleophilic Substitution

767

NMR of compound 171 can be seen. You may already have guessed that the role of bisulfite was as a nucleophile, but what happens next? SO2OH O

N2O5 MeNO2

N

N

N

NO3

S

HO

O

169

N

NO2

NO2

NO2

SO2OH +

N

170

NO2 172

171

The bond between the nitro group and the pyridine nitrogen atom in 171 is a weak N⫺N σbond and it is advantageous for this to be transformed into a stronger C⫺N σ-bond. Intermediate 172 cannot easily achieve this but intermediate 171 can undergo a favourable [1,5] sigmatropic NO2 shift 173. Loss of bisulfite 174 then gives the product 165. 4 3

H

5 [1,5]

NO2

NO2

2 1N NO2 173

SO2OH

NO2 shift

N

N

SO2OH 174

165

The [1,5] sigmatropic shift may look unlikely at first but it should be compared with the [1,5] shifts familiar from cyclopentadiene chemistry. The nitro group passes suprafacially across the gap in the delocalised π-system of 173. It is quite likely that the nitration of many pyrazoles goes through a similar sigmatropic rearrangement. If nitration occurs first at nitrogen 14, the nitro group can then be transferred to carbon 175 (or to the other nitrogen atom!) by a [1,5]NO2 shift to give 176 and this gives the product 177 by a [1,5]H shift. Even the tautomerism of pyrazoles, such as the interconversion of 177 and 178, could be a [1,5]H shift.25 NO2 [1,5]NO2

[1,5]NO2

N N

N

NO2 14

H

N NO 2

NO2

[1,5]H

NH

N

N N H 178

N

N

175

NO2

[1,5]H

176

177

This mechanism via 173 and 174 explains some puzzling results. Nitration of 4-acyl pyridines 179 occurs in excellent yield though electrophilic substitution would be disfavoured. Both the nucleophilic addition of NaHSO3 (181 to 182) and the [1,5] shift (182 to 183) are probably accelerated by the presence of a ketone in the 4-position.

O

O

O 1. N 2O5 MeNO2

179

O H

NO2

2. H 2O NaHSO3

N

O

N 180; 75% yield

N

N

NO2 181

NO2 182

SO2OH

N

NO2 SO2OH

183

768

32 Functionalisation of Pyridine

No reaction at all occurs with 3-acyl pyridines26 86. With this substitution pattern, we already know that nucleophilic attack occurs between the carbonyl group and the nitrogen atom to give 186 and hence 187 by a [1,5]NO2 shift. No elimination of bisulfite would now be possible and this is a dead end. O

O

O O2N

O

[1,5]

N 184; no reaction

N

N

NO2 185

NO2 186

SO2OH

N

SO2OH 187

This work opens up important developments. • Simple 3-nitro-pyridines and, by reduction, 3-amino pyridines are now readily available as starting materials. • Nitration of compounds having pyridine and benzene rings can be controlled to give chemoselective reaction at either ring. There will obviously be many applications in the future as the key reagent, N2O5, can be manufactured on a 100 tonne scale from N2O4 and nitric acid (HNO3).

Sulfonation of Pyridines Some reactions thought of as normal electrophilic substitution may in fact proceed by this mechanism. One example might be the sulfonation of 2-hydroxy-nicotinic acid used in the synthesis of a component27 190 of Pfizer’s successor to Viagra 188. MeO

MeO

O N

HN

O N

N

O S O

N

2x C–N

N

O

C–O N–S

N 189

OH

N

O

SO2OH

N

N

Et 188; Pfizer′s UK-357,903

OH O

O S

N

OH

O

N

Et 190

Sulfonation of 2-hydroxy-nicotinic acid 191 with sulfuric acid does not work but 30% oleum, that is sulfuric acid containing a 30% excess of SO3, gives a 90% yield of 190. This might easily occur by N-sulfonation 192, addition of some nucleophile 193, [1,5] shift 194 and elimination. Note that the position between CO2H and N in 192 is blocked by the OH group so addition must occur on the other side in contrast to 184. This compound will also be discussed in chapter 33.

32 Synthesis of Imidazo[4,5-c]pyridines

OH

OH CO2H

N

OO2S

769

OH

OH CO2H

N

OO2S

CO2H

N

190 Nu

Nu 191

CO2H

N

192

SO2OH 194

193

Extension by Vicarious Nucleophilic Aromatic Substitution Vicarious nucleophilic substitution is the addition of nucleophiles carrying their own leaving group at an unsubstituted position on an aromatic ring. Since the Bakke nitration makes 3-nitropyridines available, reactions with amine nucleophiles X–NH2, where X is a leaving group, at unsubstituted positions become possible.28 One such nucleophile is hydroxylamine. Attack occurs next to nitrogen, but para to the nitro group 195. Dehydration 196 and protonation of the anion 197 at the imine nitrogen gives the amine products 198. A base (KOH is best) is also needed and R can be H, Alk, or Ar. O

R

N

OH NH2

R O HO

N

O N

H N H

195

R

O N

O

N

HN

196

R NO2

O

N

H2N

N

197

198

An alternative nucleophile is the amino-triazole 200 and this works better in some cases, notably the isoquinoline 199. Amination again occurs para to the nitro group 201. NO2

NO2 +

N N

N 199

t-BuOK

N

NH2

201 N 65% yield

DMSO

NH2

200

Synthesis of Imidazo[4,5-c]pyridines If, on the other hand, an amino group can be introduced ortho to the nitro group, reduction of 202 or 203 makes available the starting materials for fused heterocycles. One example is the synthesis of imidazo[4,5-c]pyridines 205 from diamino pyridines such as 204. NH2

NO2 or

N

NO2 202

N

NH2 203

H N

NH2

[H]

?

N

NH2 204

N

R N 205

A published synthesis of 3,4-diaminopyridine29 204 starts with nitration of ‘4-hydroxypyridine’ 207. This is a difficult reaction in spite of the electron-donating OH group because the compound is mostly a pyridone 206. Nevertheless a 74% yield of 208 can be achieved with a mixture of fuming nitric and sulfuric acids.

770

32 Functionalisation of Pyridine

O

OH

HN

N 206

OH

fuming HNO3

N

fuming H2SO4

O HN

NO2

207

NO2

208

209

The product 208 may be as drawn or it too may be a pyridone 209 but in any case it reacts with POCl3 to give the chloro-compound 210 and then by nucleophilic substitution with amines such as benzylamine to give, after reduction, the diamine 212 from which the benzyl group could be removed to give 204.

208 or 209

Cl

POCl3

H N

BnNH2

N

N

NO2 210; 94% yield

Ph

H N

H2 5% Pt/C

[H]

204 N

NO2 211

Ph

NH2 212

Another method30 uses nitration of available 4-aminopyridine 213 to give unstable 214 and then by thermal rearrangement, the nitro compound 202 that can be reduced to 204 in good yield. NH2

H N

HNO3 H2SO4

N

NH2 heat

NO2

N

N

213

H2, 5% Pd/C MeOH

204 98% yield

NO2 202; 72% yield

214

Neither of these methods is suitable for large scale production but the Bakke method31 works well with a significant modification. The amino group in 213 must be protected as an amide 215 to prevent formation of 214. Then nitration and reduction to give 217 (half-protected 204) work very well. H N

Ac2O Et3N

213 MeCN

O

N

H N

1. N 2O5 MeNO2 2. H 2O NaHSO3

O NO2

N

room temperature 20 hours

216; 72% yield

215; 97% yield

H N

H2, MeOH 5% Pd/C

O NH2

N

217; 97% yield

If the diamine 204 is wanted, hydrolysis of the amide in 217 is the answer. But if the more interesting imidazolo-pyridine 205 is wanted, acid catalysed cyclisation of the amine onto the amide gives the imidazole 205; R ⫽ Me directly. Of course if R needs to be something else, the original acylation of 213 should be done with that anhydride (RCO)2O. H N

NH2 N

NH2

204; 96% yield

2M HCl 5 hours 70 ˚C

N

O

NH2 217

H N

37% HCl 1-butanol reflux 1 hour

N

Me N

205; R = Me; 90% yield

The previous synthesis29 ended at the half-protected diamine 212. The N-benzyl group can of course be removed by hydrogenation to give 204 but cyclisation of 212 with an acid anhydride gives 218, a part-protected version of 205.

32 Synthesis of Imidazo[4,5-c]pyridines

NH2 N

Ph

H N

H2 catalyst

N

NH2

Ph

N

Ac2O AcOH

NH2 212

204

771

Me

N

N 218

The type of molecule to be made from these compounds is illustrated by ABT-491, Abbott laboratories inhibitor32 of platelet aggregation factor 219. The imidazolo-pyridine segment is the very one we have been discussing (205; R ⫽ Me) with the added complication that the rest of the molecule must be added to one of the two nitrogen atoms of the imidazole and not to the other.

O N N

N CONMe2

N

219 ABT-491 (Abbott Laboratories)

F

The complete synthesis starts with a Friedel-Crafts reaction between a benzoyl chloride 220 and an indole 221. Replacement of the chloride in 222 by an amino group is achieved by SN2 with azide and reduction with Ph3P. Nucleophilic substitution on the chloro pyridine 224 gives an intermediate 225 having imidazole connectivity problem solved. All that was needed was the correct nitro pyridine, easily made by Bakke nitration on 4-chloropyridine. The imidazole is closed by reduction and acetylation and a Sonigashira coupling inserts the acetylene. This is an impressively short and convergent synthesis with only one protecting group - the Me3Si on the acetylene. Br

O

Br

O

AlCl3

Cl

room temperature

N F 220 Br

1. NaN 3 DMF

CH2Cl2

+

CONMe2 221

Cl

2. Ph 3P THF, H 2O

N CONMe2 222 Br

O

F

Cl

O

NO2 +

Et3N

Cl

MeCN

N

N

N

CONMe2 223

F

NH2

CONMe2 224

Br

NO2

F 225

N

O 1. Pd(PPh 3)4, toluene

N

Me3Si

1. H 2, Pt/C 2. Ac 2O AcOH

H N

N

N CONMe2

F 226

N

SnMe3

2. KF, H 2O, DMF

219 ABT-491

772

32 Functionalisation of Pyridine

PART V – APPLICATIONS In this section we discuss the synthesis of various substituted pyridines by methods from earlier in the chapter and their applications in the synthesis of larger drug molecules.

The Baeyer anti-cancer drug BAY 43-9006 This relatively simple compound 231 requires the 4-Cl picolinic acid chloride 227 easily made in high yield by the SOCl2 reaction described earlier. This has the advantage of simultaneously turning the acid into the acid chloride. Formation of the amide 228 and then the diaryl ether is followed by coupling with an isocyanate 230 to make the unsymmetrical urea 231. This large scale synthesis is planned for manufacture.33 CO2H N

CONHMe

COCl SOCl2 DMF

N

MeNH2 MeOH THF

Cl

227; 89% yield

99

F3C

N

CONHMe H2N

N

K2CO3 DMF

Cl

O 229; 87% yield

228; 88% yield

229 + CH2Cl2

CONHMe

H N

H N

F3C

NCO

Cl

p-aminophenol t-BuOK

N

O

Cl

O

231; BAY 43-9006, 92% yield

230

Anatoxin analogues The heterocycles may be pyridine, pyridazine, pyrimidine (positional isomers) and the organozinc reagents 233 were made from the bromo-compounds. Six compounds such as 234 were made by Negishi coupling.34 O O

N

H N

OTf

N

1. Pd(PPh 3)4

+

N

BrZn 232

2. HCl, H 2O

234; anatoxin analogue

233

Epibatidine analogues (anti-smoking) The starting materials were prepared by electrophilic and nucleophilic substitution on activated pyridines such as 235. The essential boronic acid 238 needed lithiation.35 I

NaNO2

I2, H 5IO6

N

NH2 235

H2SO4 H2O, AcOH

N 236

NH2

HF pyridine

I

(HO)3B 1. BuLi

N 237

F

2. ( i-PrO)3B

N

3. NaOH

F

238

Suzuki coupling of such boronic acids with the triflate 240 of the protected ketone 239 gives bicyclic compounds such as 241 easily reduced to epibatidine analogues such as 242.

32 Synthesis of Imidazo[4,5-c]pyridines

Boc

Boc

N

1. (Me 3Si)2NNa

O

F

Boc

N

N

238

OTf

2. PhNTf 2

239

773

N

Pd(PPh3)4 LiCl, Na 2CO3

240

241

Alternatively, coupling of the boronic acid from 236 gives 242 that can be modified by electrophilic substitution afterwards and further Suzuki coupling used to give analogues36 Ph

NH2

H N

1. Br 2, HOAc

X

H N

N 2. PhB(OH) 2 Pd(OAc) 2 etc.

242

N

243

epibatidine analogues: X = F, NH 2, NMe 2

Polycyclic derivatives Lateral lithiation of 244 and intramolecular electrophilic substitution with iminium ions derived from the partly reduced phthalimide 247 in acidic solution leads to the polycyclic derivatives37 such as 248. The 4-Me compound cyclises with some regioselectivity favouring para to OMe.

MeO

N 244

1. BuLi

phthalimide

2. CH 2O

Ph3P, DEAD

MeO

N 245; 60% yield

OH

O MeO N 246; 90% yield

N O

OH 1. LiEt 3BH 2. HCl, H 2O

TsOH

MeO

N

N

benzene

O

N

MeO

247; 90% yield

O

N 248; 70% yield

Katritzky’s improvement of Bakke nitration The main problem with the Bakke procedure is the need to prepare N2O5. Katritzky’s group has avoided this by making N2O5 in situ from trifluoroacetic anhydride and nitric acid. Pyridine gives 83% yield of 165, better than 77% by Bakke. Substituted pyridines such as 249 (76% rather than 15) and 251 (86% rather than 70) also give better yields, much better with 249. O + N 1 part

N + O OH F3C 2.5 parts conc. nitric

NO2

O

O O

1. mix, 0-24 ˚C

CF3

2. Na 2S2O5 H2O, 0 ˚C

6 parts Me

Cl

N 249

O2N

Cl

N 250; 76% yield 15% Bakke

N 165; 83% yield Me NO2

N 251

N 252; 86% yield 70% Bakke

774

32 Functionalisation of Pyridine

References General references are given on page 893 1. G. A. Olah, S. C. Narang, J. A. Olah, R. L. Pearson and C. A. Cupas, J. Am. Chem. Soc., 1980, 102, 3507. 2. G. A. Olah, S. C. Narang and A. P. Fung, J. Org. Chem., 1981, 46, 2706. 3. W. Orth, J. Engel, P. Ernig, G. Scheffer and H. Pohle, Ger. Pat., 3,608,762; Chem. Abstr., 1987, 106, 50057. 4. P. W. Sheldrake, L. C. Powling and P. K. Slaich, J. Org. Chem., 1997, 62, 3008. 5. D. G. Wishka, D. R. Graber, E. P. Seest, L. A. Dolak, F. Han, W. Watt and J. Morris, J. Org. Chem., 1998, 63, 7851. 6. S. Kawata, S. Ashizawa and M. Hirama, J. Am. Chem. Soc., 1997, 119, 12012. 7. E. C. Taylor and A. J. Crovetti, J. Org. Chem., 1954, 19, 1633; C. Hoffmann and A. Faure, Bull. Soc. Chim. Fr., 1966, 2316. 8. Saunders, Top Drugs, chapter 5, page 37. 9. J. Clayden, Lithium, pages 61–70. 10. F. Mongin, F. Trécourt and G. Quéguiner, Tetrahedron Lett., 1999, 40, 5483. 11. A.-S. Rebstock, F. Mongin, F. Trécourt and G. Quéguiner, Tetrahedron Lett., 2002, 43, 767. 12. A. Turck, N. Plé, A. Leprêtre-Gaquère and G. Quéguiner, Heterocycles, 1998, 49, 205. 13. N. Plé, A. Turck, F. Bardin and G. Quéguiner, J. Heterocycl. Chem., 1992, 29, 467. 14. P. Rocca, C. Cochennec, F. Marsais, L. Thomas-dit-Dumont, M. Mallet, A. Godard and G. Quéguiner, J. Org. Chem., 1993, 58, 7832. 15. D. L. Comins, M. F. Baevsky and H. Hong, J. Am. Chem. Soc., 1992, 114, 10971. 16. D. L. Comins and J. K. Saha, J. Org. Chem., 1996, 61, 9623. 17. D. L. Comins and J. M. Nolan, Org. Lett., 2001, 3, 4255; D. L. Comins and J. M. Salvador J. Org. Chem., 1993, 58, 4656. 18. F. Trécourt, B. Gervais, O. Mongin, C. Le Gal, F. Mongin and G. Quéguiner, J. Org. Chem., 1998, 63, 2892. 19. S. M. McElvain and M. A. Goese, J. Am. Chem. Soc., 1943, 65, 2233. 20. H. J. den Hertog, L. van der Does and C. A. Landheer, Rec. Trav. Chim., 1962, 81, 864. 21. R. F. Evans, H. C. Brown and H. C. van der Plas, Org. Synth. Coll., 1973, V, 977. 22. G. Höfle, W. Steglich and H. Vorbrüggen, Angew. Chem., Int. Ed., 1978, 17, 569. 23. J. Bakke, I. Hegborn, E. Øvreeide and K. Aaby, Acta Chem. Scand., 1994, 48, 1001. 24. J. M. Bakke and E. Ranes, J. Chem. Soc., Perkin Trans. 2, 1997, 1919; J. M. Bakke, H. Svensen and E. Raines, J. Chem. Soc., Perkin Trans. 2, 1998, 2477. 25. J. W. A. M. Janssen, C. L. Habraken and R. Louw, J. Org. Chem., 1976, 41, 1758. 26. J. M. Bakke and E. Ranes, Synthesis, 1997, 281. 27. D. J. Dale, J. Draper, P. J. Dunn, M. L. Hughes, F. Hussain, P. C. Levett, G. B. Ward and A. S. Wood, Org. Process Res. Dev., 2002, 6, 767. 28. J. M. Bakke, H. Svensen and R. Trevisan, J. Chem. Soc., Perkin Trans. 1, 2001, 376. 29. G. C. Wright, J. Heterocycl. Chem., 1976, 13, 601 modified by D. M. Houston, E. K. Dolence, B. T. Keller, U. Patel-Thombre and R. T. Borchardt, J. Med. Chem., 1985, 28, 471; M. F. Reich, P. F. Fabio, V. J. Lee, N. A. Kuck and R. T. Testa, J. Med. Chem., 1989, 32, 2474. 30. H. Dvorakova, A. Holy, I. Votruba, and M. Masojidkova, Coll. Czech. Chem. Commun., 1993, 58, 629. 31. J. M. Bakke and J. Riha, J. Heterocycl. Chem., 1999, 36, 1143. 32. M. L. Curtin, S. K. Davidsen, H. R. Heyman, R. B. Garland, G. S. Sheppard, A. S. Florjancic, L. Xu, G. M. Carrera, D. H. Steinman, J. A. Trautmann, D. H. Albert, T. J. Magoc, P. Tapang, D. A. Rhein, R. G. Conway, G. Luo, J. F. Denissen, K. C. Marsh, D. W. Morgan and J. B. Summers, J. Med. Chem., 1998, 41, 74. 33. D. Bankston, J. Dumas, R. Natero, B. Reidl, M.-K. Monahan and R. Sibley, Org. Process Res. Dev., 2002, 6, 777. 34. C. G. V. Sharples, G. Karig, G. L. Simpson, J. A. Spencer, E. Wright, N. S. Millar, S. Wonnacott and T. Gallagher, J. Med. Chem., 2002, 45, 3235.

32 References

775

35. F. I. Carroll, F. Liang, H. A. Navarro, L. E. Brieaddy, P. Abraham, M. I. Damaj and B. R. Martin, J. Med. Chem., 2001, 44, 2229. 36. F. I. Carroll, J. R. Lee, H. A. Navarro, L. E. Brieaddy, P. Abraham, M. I. Damaj and B. R. Martin, J. Med. Chem., 2001, 44, 4039 and F. I. Carroll, J. R. Lee, H. A. Navarro, W. Ma, L. E. Brieaddy, P. Abraham, M. I. Damaj and B. R. Martin, J. Med. Chem., 2002, 45, 4755. 37. M. A. Brodney and A. Padwa, Tetrahedron Lett., 1997, 38, 6153. 38. A. R. Katritzky, E. F. V. Scriven, S. Majumder, R. G. Akhmedova, A. V. Vakulenko, N. G. Akhmedov, R. Murugan and K. A. Abboud, Org. Biol. Chem., 2005, 3, 538.

33

Oxidation of Aromatic Compounds, Enols and Enolates This chapter examines the addition of functional groups to aromatic rings and to carbonyl groups via enols or enolates. It revises previous sections.

Introduction PART I – ELECTROPHILIC SUBSTITUTION BY OXYGEN ON BENZENE RINGS The Diazotisation Approach Synthesis of a successor to Viagra The Friedel-Crafts and Baeyer-Villiger Route Avoidance by choice of oxygenated starting materials Oxidation through Lithiation and Ortho-Lithiation Hydroxylation of Pyridines by ortho-Lithiation Synthesis of Atpenin B Introducing OH by Nucleophilic Substitution PART II – OXIDATION OF ENOLS AND ENOLATES Direct Oxidation without Formation of a Specific Enol Selenium dioxide Nitrosation with nitrites Nitrosation with stable nitroso compounds Indirect Oxidation with Formation of a Specific Enol: Enone Formation Pd(II) oxidation of silyl enol ethers Bromination of enols in enone formation Sulfur and selenium compounds in enone formation Asymmetric Synthesis of Cannabispirenones Oxidation of Enones: Epoxides and the Eschenmoser Fragmentation PART III – ELECTROPHILIC ATTACK ON ENOL (ATE) S BY OXYGEN The Problem Competition with Baeyer-Villiger rearrangement and oxidative cleavage Unexpected Success with the “Obvious” Reagents Unique successful hydroxylations with ozone and mCPBA First Successful Method: Epoxidation of Silyl Enol Ethers (Rubottom Oxidation) Method, mechanism, and isolation of intermediates Chemo- regio- and stereoselectivity: synthesis of fused γ-lactones Hydroxylation of Amino-ketones via Silyl Enol Ethers Synthesis of vinca alkaloids and model compounds

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

778

33 Oxidation of Aromatic Compounds, Enols and Enolates

Second Successful Method: Hydroxylation with MoOPH The reagent: hydroxylation of lithium enolates Stereoselectivity (substrate control) Hydroxylation of steroids and amino acids Third Successful Method: Hydroxylation with N-Sulfonyl Oxaziridines The reagents and the mechanism Hydroxylation of sodium and potassium enolates: side reaction with lithium Substrate controlled stereoselectivity: asymmetric hydroxylation Asymmetric hydroxylation with camphor sultam derivatives Asymmetric synthesis of tetracycline precursors Synthesis of a calcium channel opening drug Asymmetric synthesis of buproprion Summary

Introduction This chapter is also about functionalisation. But it deals with the addition of a difficult electrophile (‘RO⫹’) to familiar nucleophiles: aromatic compounds, including the pyridines of chapter 32, and enols and enolates. As well as the difficulties of creating suitable reagents that control chemo- and regioselectivity, stereoselectivity and asymmetric induction are important.

PART I – ELECTROPHILIC SUBSTITUTION BY OXYGEN ON BENZENE RINGS The one large gap in electrophilic substitution reactions on benzene rings 1 is the absence of simple methods to introduce oxygen with an electrophilic reagent 2. You were advised in the Disconnection Textbook (page 21) to have the oxygen atom present in the starting material 3 and achieve the same result by adding an electrophile 4. OH

OH

[HO ]

OH E

?

R

R

E

1

2

3

4

The Diazotisation Approach The traditional approach was a nitration, reduction, diazotisation and hydrolysis sequence of four reactions. Though the OH group is actually added by a nucleophilic substitution on the diazonium salt 7 the nitrogen atom was originally added as an electrophilic nitro group 5.

NO2 1

HNO3

NH2

H2, Pd/C

H2SO4

N2

NaNO2 HCl

R

R 5

R 6

7

H2O

2

33 The Diazotisation Approach

779

You might think such an apparently cumbersome approach is now no longer used but it is. In a synthesis of biologically active compounds related to epibatidine, the nitropyridine 9, available by chemistry described in chapter 32, was converted into the key intermediate 14 by a sequence of reactions including the removal of the OH group already present and the introduction of a new one by diazotisation and so on.1 HNO3

O2N

H2SO4

N H

O2N

I2

N

O 8

OH

H2N

K2CO3 DMF

quinoline

N

H2O AcOH

N

OH

N

10; 45% yield

I

AcO

1. NaNO 2 HBF4, H 2O

Cl

11; 78% yield

I

HO

I

2M KOH

2. Ac 2O

Cl

I

POCl3

9

Fe powder

O2N

I

N

12; 68% yield

H2O

Cl

N

13; 72% yield

Cl

14; 63% yield

Synthesis of a successor to Viagra In the large scale synthesis of the successor to Viagra described in the last chapter a short synthesis of the pyridine 23 replaced a much longer laboratory synthesis that used a similar sequence. The amino group in available 2-aminopyridine 15 directed sulfonation and bromination to give 17 and then, its work done, was replaced by the OH group 18. NH2

NH2 SO3

N

N

H2SO4

NH2 N

Br2

140 ˚C

NaNO2 HCl H2O

H2O

O S

O S

OH

O 16; 80% yield

15

OH Br

OH

O 17; 70% yield

Cl Br

N

PCl5

Br

N

POCl3

O S

O S

OH

Cl O 19; 48% yield

O 18; 82% yield

The conversion of the sulfonic acid to the sulfonyl chloride 19, necessary for the formation of the piperazine amide 20, also replaced the OH group by Cl. This was turned to advantage in introducing the OEt group by nucleophilic substitution before Pd(0)-catalysed carbonylation gave the carboxylic acid 23. This eight step route2 giving 11% yield, and containing a hazardous high temperature reaction with carbon monoxide, contrasts with the five step large scale route described in chapter 32 giving ⬎50% yield. OEt

Cl

19

Et3N CH2Cl2 HN NEt

Br

N

O S O

OEt Br

N

N

NaOEt

CO, Pd(0)

EtOH

EtOH, Et 3N 120 ˚C

O S

N

O NEt

20; 90% yield

N NEt

21; 66% yield

OEt CO2Et

CO2H

N

NaOH EtOH

O S O

O S

N

O NEt

22; 95% yield

N NEt

23; 88% yield

780

33 Oxidation of Aromatic Compounds, Enols and Enolates

The Friedel-Crafts and Baeyer-Villiger Route About the only simple way to introduce an oxygen atom by electrophilic substitution is to follow a Friedel-Crafts acylation of 1 using, say acetyl chloride to give 24, with a Baeyer-Villiger oxidation with a peroxyacid. The aromatic ring will normally migrate in preference to the methyl group and the product 25 will be formed by the overall introduction of an acetate with the regioselectivity of an electrophilic substitution. O

O

OAc mCPBA

Cl AlCl3

R

OH NaOH H2O

R

1

R

R

24

25

26

Avoidance by choice of oxygenated starting materials You are usually advised when making a complex aromatic compound to have any oxygen atoms (directly joined to the benzene ring) already present at the start. In their synthesis of the lichen metabolite dechlorolecideoidin 27, McEwen and Sargent3 chose the fully substituted benzene 29 as precursor for the right hand ring. Me

O

Cl

Cl

O

Cl OH

HO OH

Cl O

OH

B-V

CO2Me

HO CO2Me

HO

HO

2 x C–O

OHC

CO2Me

Me

Me 27; dechlorolecideoidin

Me

28

29

A search for available starting materials with both oxygen atoms present revealed methyl orsellinate 31 so the chlorine could be disconnected next to give 30. Regioselectivity was confined to the formylation of 31 and they did indeed have difficulty with that step. Cl OH

HO

OH

HO

CO2Me

OHC

SEAr

HO

OH

C–C

C–Cl

CO2Me

OHC

Me 29a

CO2Me

SEAr

Me 30

Me 31

Methyl orsellinate 31 was differentially protected and formylated to give 33 after removal of the isopropyl group with TiCl4. Benzylation and oxidation gave the benzoic acid 34 ready for a surprising conclusion to the synthesis. i-PrO 31

i-PrBr K2CO3

OH

CO2Me Me 32; 75% yield

1. (MeO) 2SO4 K2CO3 2. SnCl 4 Cl2CHOMe 3. TiCl 4

HO

OHC

OMe

CO2Me Me 33; 53% yield

1. BnBr K2CO3 2. NaClO 2 NH2SO2OH H2O, dioxan

BnO

OMe

HO2C

CO2Me Me 34

33 Hydroxylation of Pyridines by ortho-Lithiation

781

Friedel-Crafts acylation of 35 with 34 was brought about with trifluoroacetic anhydride and the benzyl groups removed to give 36. The last oxygen was now introduced by an interesting oxidation with K3Fe(CN) 6. The spirolactone 37 was formed in good yield and, on heating and removal of the OMe group, gave dechlorolecideoidin 27. The interesting chemistry as well as the selectivity in this synthesis is explored in the workbook. Me Cl BnO

Cl

2. H 2, HCl Pd/C EtOAc

HO

OBn

O

Me

1. 34 (CF3CO)2O CH2Cl, 0 ˚C

Me

Me

O Me

CO2Me

CO2Me Fe(III) Cl O O

OMe

H2O K2CO3

OMe O

HO

O

H H 36

35

37

Oxidation through Lithiation and Ortho-Lithiation Probably the best modern method for introduction of OH by electrophilic aromatic substitution is lithiation, reaction with a boronate ester, and oxidation.4 These are the same boron compounds that are used in Suzuki coupling (chapter 18) and are made the same way. In this example, selective mono-lithiation by Br/Li exchange on available tribromoanisole 39 (easily prepared by bromination of anisole or phenol) occurs ortho to the MeO group and reaction of aryl-lithium 39 with trimethyl borate gives the boronic ester 40. Peroxyacids such as peracetic acid are usually used for the final oxidation. OMe Br

OMe Br

Br BuLi

OMe

OMe Br

Li

B(OMe)2

B(OMe)3

Br 38

Br 39

Br

OH

MeCO3H

Br 41; 87% yield

Br 40

The oxidation is very similar to that used after hydroboration to make alcohols (chapter 17) and involves a migration of the aryl group from B to O as the weak O⫺O bond of the peroxy group is broken. It also closely resembles a Bayer-Villiger oxidation. The boronic acids are as good as the esters: PhMgBr reacts with B(OMe)3 followed by 10% HCl to give PhB(OH)2 oxidised without isolation to phenol5 by H2O2 in 78% yield.

MeO

OMe B

MeCO3H

ArB(OMe)2

O

O

O O

R

B(OMe)2

H2O

ArOH

R

Hydroxylation of Pyridines by ortho-Lithiation The basic ortho-lithiation strategy for electrophilic substitution on pyridines was explained in the last chapter. Now we need to explore the more powerful methods available when we add this boron-mediated oxidation. An example of control by ortho-lithiation in the synthesis of caerulomycin should reveal the possibilities. The pyridine 42 is lithiated in the 2-position because

782

33 Oxidation of Aromatic Compounds, Enols and Enolates

an amide is a stronger ortho-director than a simple ether. Negishi coupling (chapter 18) of the Zn derivative with 2-bromo-pyridine gives the dipyridine 43 in good yield. A second ortho-lithiation must now occur at the 5-position and bromination gives 44. OMe

OMe OCONi-Pr2

OMe OCONi-Pr2

1. s-BuLi 2. ZnCl 2

1. PhLi

3. 2-Br-pyridine 1 mol% Pd(PPh3)4

N

OCONi-Pr2

Br

N

N

2. BrCN

43; 77% yield

42

N

N

44; 81% yield

Treatment of 44 with BuLi would lead to Br/Li exchange, but treatment with LDA lithiates the remaining position (C-6) on the ring 45 then the halogen dance (chapter 32) gives the more stable lithium derivative 46 with Li ortho to the OMe group. Protonation gives 47 in good yield.6 OMe LDA

OMe OCONi-Pr2

Br

OMe OCONi-Pr2

Li

44 Li

N

N

Br

N

N

45

OCONi-Pr2

H

H2O

Br

46

N

N

47; 85% yield

Completion of the synthesis of caerulomycin C starts with a very efficient second Negishi coupling to give 48, but then requires oxidation to an aldehyde, oxime formation, and removal of the original ortho-directing amide. The yield of 49 was unsatisfactory. OMe H 47

MeZnCl 1 mol% Pd(PPh3)4

Me

OMe OCONi-Pr2 1. amide hydrolysis 2. DDQ 3. NH 2OH

N

N

H H

OMe N

N N

HO

49; 99% de

107

NH2 HO2C

SO2 109

R

110

An ingenious application was in the asymmetric synthesis of coniine15 114, the alkaloid in hemlock by which Socrates met his death. The sultam amide 111 incorporates a protected ketone so that acidic hydrolysis forms the cyclic nitrone 112 by capture of that ketone. Reduction and base-catalysed cleavage of the amide gives coniine 114.

O

O N

1. NaN(SiMe 3)2 2. 101

N

N

3. HCl, H 2O

SO2

SO2

O 111

H2, Pd/C

N SO2

1. NaH, toluene 2 hours reflux

N

2. i-Bu2AlH

O OH 113; 92% yield

O O 112; 72% yield

109 + 83% yield

N H 114; (–)-coniine 56% overall yield

Much simpler, though not enantioselective, examples have been reported recently by Hisashi Yamamoto.16 Lithium or tin enolates 116 react with nitrosobenzene to give good yields of hydroxylamines 117. O

OM

Ph 115

O

N Ph

Ph 116; M = Li or SnBu 3

O

Ph

117; M = Li 70% yield

OH N

Ph

M = SnBu3 96% yield

788

33 Oxidation of Aromatic Compounds, Enols and Enolates

One interesting aspect of this reaction is that the acid-catalysed variant using silyl enol ethers 118 and Lewis acids provides a good way to oxidise the enol as the electrophilic end of the nitroso group switches to the oxygen atom. The product 120 is an N,O-disubstituted hydroxylamine. As the N⫺O single bond is rather easily cleaved reductively, this route promises to be a good way to make α-hydroxyketones.17

O

OSiMe3

O PhNO

O

Et3SiOTf

118

119

N H

Ph

120; 94% yield

Indirect Oxidation with Formation of a Specific Enol: Enone Formation Pd(II) oxidation of silyl enol ethers More regioselective methods are available if a specific enol equivalent (often the silyl enol ether 122) can be used in the oxidation. These methods are usually used to introduce a conjugated alkene next to the carbonyl group. Oxidation of a silyl enol ether with Pd(II) does this job well.18 O

OSiMe3

O Pd(OAc) 2

R2

R1

R2

R1

121

R2

R1

122

123

Regioselective oxy-palladation of the enol derivative gives 124 or 125, either of which might lose palladium in a β-elimination that can only go one way. The problem is that the palladium comes out as Pd(0) but must go in as Pd(II) so the reaction is not intrinsically catalytic. The answer is to add a stoichiometric oxidant, such as a quinone, to complete the cycle.

AcO 122

O

OSiMe3

H β-elimination

Pd(OAc) 2 oxypalladation

R1

R2 PdOAc 124

R1

H

R2 PdOAc 125

Pd

OAc

Pd(0)

123

An excellent example is the synthesis of mevinolin 126, an HMG-CoA reductase inhibitor that lowers the cholesterol level, by Hirama and Iwashita.19 Their idea was to use a Diels-Alder reaction to close both rings in one step and they therefore disconnected various peripheral fragments to leave the simpler ketone 127. Reversing the Diels-Alder reveals 128 and the starting material was actually one diastereoisomer to ensure stereochemical control in the ring closure.

33 Indirect Oxidation with Formation of a Specific Enol: Enone Formation

HO

O 3

O

3 1 O

O

OBn 1

5 O

H

R

DielsAlder

FGI

O

O

O

5

789

H X 128

X

126; mevinolin

127

OTBDMS

O

O

=

128

OBn O

The Diels-Alder was successful20 129 but now the methyl group in the SW corner needs to be added by conjugate addition and that requires an enone. Oxidation of the silyl enol ether 130, prepared by kinetic enolate formation with LDA, with catalytic Pd(OAc)2 and benzoquinone as the stoichiometric oxidant gave the required enone 131 in 57% yield. Addition of Me2CuLi was stereoselective to give 132 and the synthesis of mevinolin completed. O

128

H

R

TBDMSO

H

1. LDA 2. t-BuMe2SiCl

heat

O

R

H

O

R

Pd(OAc) 2

Me OTBDMS

OTBDMS

R

Me2CuLi

quinone

OTBDMS 129

H

OTBDMS 132; 91% yield

131

130

Bromination of enols in enone formation The oldest but still occasionally useful method for enone formation from saturated ketones involves bromination and elimination of HBr from the bromoketone 133. The problem here is that the proton to be removed (Hb in 133) in the elimination is not the most acidic proton in the molecule: Ha is much more acidic. O

Hb

O

O

Br2

R2

R1 H 121

base

R1

Ha

R2 Br 133

R2

R1 123

Nevertheless this method can work well if hindered or non-nucleophilic bases (such as DBN and DBU) are used. The dienone 134 was needed for a study of migrating groups in the dienonephenol rearrangement.21 The bold proposal was that double elimination from dibromoketone 135

790

33 Oxidation of Aromatic Compounds, Enols and Enolates

would allow the ketone 136 to be used and that this might in turn be made by double conjugate addition of a cuprate to dienone 137. Direct oxidation by a quinone or SeO2 would give this from the enone 138, easily made by Robinson annelation. O

O Br

O

O

O

Br

FGI

FGA?

FGI

2 x C-C conjugate addition

134

135

136

137

138

The enone 138 could be oxidised to dienone 137 with the quinone DDQ 139. Bromination of 136 rather surprisingly gave a single crystalline isomer 140 in 90% yield: NMR revealed the stereochemistry shown. The double dehydrobromination was carried out with calcium carbonate in DMF. O Cl

O CN

Cl

138

CN 136

O 139; DDQ

DDQ 139

Br2 HOAc

137

140

Me2CuLi

CaCO3 DMF

136

Br

Br

134 140; 90% yield

This synthesis was well worthwhile as the dienone-phenol rearrangement of 134 gave the phenol 144 showing that the ethyl group migrates better than methyl. The dienone rearranges to 142 and a second migration is needed before a cation is formed 143 that can give an aromatic compound by loss of a proton. In each rearrangement there is a choice between a methyl and an ethyl group.

OH

OH

141

142

OH

H

OH

H

134

143

144

Sulfur and selenium compounds in enone formation Probably the most popular method to introduce an alkene is by regioselective functionalisation with a sulfur or selenium reagent, oxidation, and elimination. The α-RS or α-RSe compound 145 is made by sulfenylation or selenenylation of a specific enol equivalent. Oxidation requires only mild reagents such as H2O2, and the elimination follows a cyclic syn mechanism 146 so the lack of acidity in Hb is unimportant. The elimination (a reverse ene reaction) is highly E-selective as the substituents R1CO and R2 prefer an anti conformation in the transition state. This resembles the high E-selectivity in the [3,3]-sigmatropic rearrangements discussed in chapter 15. S(e) means either Se or S.

33 Indirect Oxidation with Formation of a Specific Enol: Enone Formation

R2

O

O

O

[O]

121

heat

Hb

R1

R2

R1

791

RS(e)

RS(e) 145

R2

R1

+ PhS(e)OH

O (E)-123

146

Sulfenylation of lithium enolates with PhS-SPh or of silyl enol ethers with PhS-Cl allows any α-PhS carbonyl compound to be made regioselectively, e.g. 149 and 152 from 147 (see chapter 5). Oxidation with sodium periodate gives the sulfoxide without over-oxidation to the sulfone, but elimination requires reasonably high temperatures (about 120 ⬚C for MeSO but only about 50 ⬚C for PhSO). Together with the unpleasant by-products, the results of disproportionation of unstable PhSOH, this has led to a preference for the selenium version of the reaction, though we must admit that the by-products are even more offensive.22 O

OSiMe3 Me3SiCl

PhS

O

O 1. NaIO 4

PhSCl

2. heat

Et3N

147

148

149

150

LDA

O

OLi

O

O SPh NaIO4

PhS-SPh

151

O

SPh

152

heat

153

154

The selenium version of this reaction offers the advantages that over-oxidation is no problem and that the elimination of the selenoxides occurs at room temperature or below so that the oxidation and elimination normally occur as a single step.23 A simple example is the preparation of another starting material 158 for a dienone-phenol rearrangement.24 The lithium enolate of spirocyclic ketone 155 reacts with PhSeCl to give 156 and oxidation with H2O2 gives the dienone 158 directly, with the selenoxide 157 as an intermediate. The overall yield is 83%. O

O

H2O2

2. PhSeCl

155

O

O

1. LDA

SePh

SePh 156

157

O

158

A general problem that is neatly solved by these methods is that of making exo-methylene lactones such as 167. These compounds are cytotoxic and both carcinogenic and anti-cancer compounds are found in this class. If the cis-fused lactone 159 is converted into its lithium enolate 160 we have a classic case of a folded molecule and attack by an electrophile occurs on the exo face, that is the same face as the ring-junction hydrogen atoms. Methylation gives 161. The lithium enolate of this molecule 162 again has a flat five-membered ring and a second electrophile also adds on the exo face, say to give 163. In this compound the PhSe group and the ring junction H are syn to each other so elimination on the selenoxide tends to give the more stable but less interesting unsaturated lactone 164.

792

33 Oxidation of Aromatic Compounds, Enols and Enolates

H

H O

H

LDA

MeI

OLi O

O H 159

O H 161

H 160

H

SePh

H PhSe-SePh

OLi

O

or PhSeCl

O

H2O2

O

O

H 162

LDA

O

O

H 163

164

If we want the exomethylene lactone 167, we must simply add the electrophiles in the reverse order.25 The PhSe group is therefore anti to the ring junction H atom 166 and elimination on the selenoxide must give 167. SePh

H 159

H

H SePh O

1. LDA

1. LDA

O

2. PhSe-X

2.MeI

H2O2

O

O

O

O

H 165

H 166

H 167

Asymmetric Synthesis of Cannabispirenones The selenium oxidation can even be a source of asymmetry. Both enantiomers of cannabispirenones 170 occur naturally in different plants and the racemic enone is made in two steps from the achiral ketone 168. You will see that 170 is chiral only by virtue of the alkene in the ring. O MeO

MeO

O

MeO

MeO

CHO

MeO

MeO

168

169

170

Hydrogenation of 170 gives the achiral saturated ketone 171 that can be desymmetrised by a chiral base. The lithiated diamine 172 gives mostly the lithium enolate 173 and hence, by addition of selenium and oxidative elimination, the enone (R)-170. Different bases give the other enantiomer.26 O

O

OLi Me

Li N

MeO

MeO

N

MeO

Ph Me

1. PhSeBr 2. H 2O2

172

MeO

MeO 171

173

MeO (R)-170; 84% ee

33 Oxidation of Enones: Epoxides and the Eschenmoser Fragmentation

793

Oxidation of Enones: Epoxides and the Eschenmoser Fragmentation Once the enone is prepared, further oxidation becomes easier. The epoxide can be formed at the alkene with the nucleophilic oxidant H2O2 /NaOH. The mechanism involves a discrete enolate intermediate 176 so the stereochemistry of the alkene is lost and E- and Z-enones 174 usually both give the anti-epoxide 177.

O

O 2

1

R

R

NaOH

O

O OH

H2O2

R1

R

(E)-174

2

O

1

R

R

175

OH

O

2

O R2

R1

176

177

This reaction gives some stereoselectivity with pulegone 178 as the epoxide 179 is formed in an approximately 2:1 anti:syn ratio,27 and chemo-selectivity as carvone 180 gives just one epoxide 181 as a mixture of diastereoisomers: the electron-rich alkene is not epoxidised.28 O H2O2

H2O2

NaOH

NaOH

O 178; ( R)-(+)-pulegone

O

O

O 179; 2:1 anti:syn

O

180; ( R)-(+)-carvone

181; 92% yield

Further reactions on these compounds lead to other oxidised products in which the lack of stereochemical control in the epoxidation is unimportant, so, for example isophorone oxide rearranges with various catalysts to the cyclopentanone 182 (80% yield) while both isomers of pulegone oxide 179 gives the cycloheptadione29 183 (78% yield). Exhaustive methylation of the extended enolate produced by reduction of 181 gives 184 in good yield.28 OHC

O H2O2

179

O

NaOH

O isophorone

181 O

O

O isophorone oxide

182

183

Li MeI NH3(l) THF

O 184; 80% yield

The most dramatic of all these reactions is the Eschenmoser fragmentation30 - a general route to alkynyl aldehydes 189 and ketones. The tosylhydrazone 186 of an epoxyketone 185 gives an anion on treatment with base that fragments once 187 to give an alkene and again 188 in the reverse direction 188 to give the alkyne. Ts O

N

NHTs

TsNH.NH2

N

N

base

O 185

N

NTs

O

O 186

187

188

O

CHO 189

794

33 Oxidation of Aromatic Compounds, Enols and Enolates

The ‘disconnection’ corresponding to a fragmentation must of course be a reconnection and there is a choice here 191 when reconnecting the alkyne to the aldehyde 190 or ketone 192. Whether you draw the intermediates or not you should get back to the same two enones as potential starting materials. In this case either method works and you may choose the enone you prefer, or that you find easier to make. O

H

O

b reconnect

H Eschenmoser fragmentation

a reconnect

a b

CHO

190

H

Eschenmoser fragmentation

191

192

The rest of this chapter is about attempts to introduce electrophilic oxygen onto enolates. The option of acylation followed by Baeyer-Villiger rearrangement is not available as the starting materials themselves contain carbonyl groups and react with peroxy acids. The alternative solutions will lead us into interesting chemistry. We shall omit unreliable older solutions involving oxygen gas, Pb(OAc) 4 etc.

PART III – ELECTROPHILIC ATTACK ON ENOL(ATE) S BY OXYGEN The Problem Competition with Baeyer-Villiger rearrangement and oxidative cleavage The idea is to be able to react a carbonyl compound 193 (or some specific enol derivative of it) with some reagent for the synthon “HO⫹” so that a new C⫺O bond is formed on the next atom 192. Sometimes no stereochemistry will be involved but we should also like to be able to create single enantiomers. O

OH

O

O OH

?

OH

193

The most obvious thing to do is to react the carbonyl compound with a peroxyacid. After all, double bonds react to give epoxides and enols are more nucleophilic than double bonds. Unfortunately, peroxyacids are nucleophilic as well as electrophilic and form hemiacetals with carbonyl compounds which decompose by the C to O migration 195 known as the Baeyer-Villiger rearrangement. O ArCO3H

OH O

O O

O Ar

R 194

O

R 195

196 R

33 Unexpected Success with the “Obvious” Reagents

795

Electrophilic oxygen is introduced at the carbon atom next to the old carbonyl group but the C⫺C bond linking these atoms is lost so the result is not α-hydroxylation. Oxidative cleavage of the enol alkene bond is going generally to be a problem in this section and is best illustrated by the useful ozonolysis of silyl enol ethers 198 and 199 of the same ketone 194. The regiospecific formation of silyl enol ethers is described in chapter 3.

HO2C

R

O

O

OSiMe3 1. O 3 MeOH

OSiMe3 R

R

R

HO HO2C

1. O 3 MeOH

R

2. NaBH 4

2. Me 2S

197; R = Me 90% yield

194

198

200; R = Me 94% yield

199

Reduction and cyclisation of ketoacid 197 would give lactone 196 while cyclisation of 200 would give the isomeric lactone. This second method is therefore controllable whereas in the Baeyer-Villiger rearrangement, only the more highly substituted group migrates. However, the Baeyer-Villiger does go with retention at the migrating group whereas there is no stereochemical control in the ozone reactions.

Unexpected Success with the “Obvious” Reagents Unique successful hydroxylations with ozone and mCPBA Two remarkable observations need discussion at this stage as they show that the two methods we have dismissed sometimes work. Attempted ozonolysis of the silyl enol ether 202 derived from camphor 201 gave instead a quantitative yield of the silyl ether of hydroxycamphor31 203.

O

OSiMe3

O

O3 MeOH

201; camphor

202

203 100% yield

OSiMe3

Though the product 203 is a mixture of diastereoisomers, it does at least show that even ozone sometimes attacks an enol as an oxygen electrophile. The second example is even more remarkable. Attempted epoxidation of the tricyclic ketone 204 gave a mixture of four compounds: two epoxides 205 in a 7:1 ratio from mCPBA attack at the alkene and two regioisomeric lactones 206 from the Baeyer-Villiger rearrangement of the ketone. O

O

O

O

mCPBA

H

O O 204

CH2Cl2 room temperature

H

+

O 205; 59% yield 7:1 diastereoisomers

H

206; 25% yield mixture of lactones

796

33 Oxidation of Aromatic Compounds, Enols and Enolates

This compound was intended to be a model compound for a more complex ketone 207 to be used in the synthesis of the quassinoid natural products, some of which have anti-cancer properties, but, mindful perhaps of Woodward’s dictum that the only reliable model compound is the enantiomer of the real thing32 they tried the reaction with 207. With one equivalent of mCPBA they got just one diastereoisomer of the epoxide. O

O CO2Me

CO2Me mCPBA

H

O

CO2Bu-t

CH2Cl2 room temperature

H

O

O

O 207

CO2Bu-t

O

208; only product; 88% yield

For the purposes of this chapter, the more exciting result was that treatment with an excess of mCPBA converted 207 into 209 in quantitative yield: epoxidation is totally stereoselective and hydroxylation has also occurred at the easily enolised position between the ketone and the ester to give a single diastereoisomer of the alcohol 209. The stereoselectivities are remarkable enough but the mere fact the hydroxylation occurs so efficiently gives hope for the development of a more general method. Inspection of the structure of bruceantin 210, the most promising of the anticancer quassinoids, shows just how valuable this reaction turned out to be. The difficult tertiary alcohol has been inserted in exactly the right place with the right stereochemistry. O

207

OH CO2Me

HO

O

H

O O

CO2Me H

O

excess mCPBA CH2Cl2 room temperature

OH

?

OCOR

H

CO2Bu-t

HO

O

209; only product; 100% yield

H

H

O

O

210; bruceantin

First Successful Method: Epoxidation of Silyl Enol Ethers (Rubottom Oxidation) Method, mechanism, and isolation of intermediates It is a small logical step from the last section to attempt the epoxidation of silyl enol ethers 211. The result is slightly surprising: hydroxylation is successful but the product is the silyl ether 212 of the α-hydroxyketone.33,34 O

OSiMe3

O mCPBA

193

211

O OSiMe3

212

OH

33 First Successful Method: Epoxidation of Silyl Enol Ethers (Rubottom Oxidation)

797

The immediate product is presumably the epoxide 213 which opens in acetal fashion to give 214 and transfers silicon intramolecularly to give 212. The reaction 214 is formally a 5-endo-tet reaction and would not occur if carbon were being transferred. As silicon is the atom under attack, a pentacovalent intermediate can be formed and the requirement for a linear SN2 transition state no longer applies.

OSiMe3

Me3Si

O

O

O

SiMe3

OSiMe3

O

mCPBA

211

O

213

212

214

Support for this mechanism came from an unexpected quarter. In a synthesis of sterpuric acid 215, Paquette planned to put in the alkene at the end by a Julia olefination.35 Addition of MeLi to the ketone 216 and non-stereospecific elimination (chapter 15) looked a good strategy. The 1,2-relationship between OH and ketone in 216 is awkward and the decision was made to try a hydroxylation on the silyl enol ether of the parent ketone 217. H

Me

H

HO2C

FGI

Me OH Me 215; sterpuric acid

H

Me

HO2C

C–O

Me

Julia olefination

PhO2S

O

Me

enol oxidation

OH

Me

MeO2C

PhO2S

216

O

H

217

This decision was to have sensational results as treatment of the silyl enol ether 218 with buffered (NaHCO3) mCPBA gave an 86% yield of a crystalline compound whose structure, determined by X-ray, was 219 - the first stable epoxide of a silyl enol ether had been isolated. H 1. NaN(SiMe 3)2

217

Me

H

MeO2C

2. Me 3SiCl

mCPBA

Me PhO2S

NaHCO3 CH2Cl2 20 ˚C

OSiMe3

Me

MeO2C

218; 91% yield

Me O OSiMe3 219; 86% yield PhO2S

When the epoxide 219 was treated with even weak acid, or when the silyl enol ether 218 was epoxidised with unbuffered mCPBA, no epoxide could be found. Instead a 76% yield of the silyl ether 220 of the hydroxylated ketone 216 was formed. The stereochemistry of 220 is of course determined by that of 219 and the requirement for cis-fused four- and five-membered rings. Sterpuric acid 215 has the same stereochemistry. H

H

Me

MeO2C

PhCO2H

Me PhO2S 219

O OSiMe3

CH2Cl2 20 ˚C

Me

MeO2C Me

H

Me

HO2C Me

OSiMe3 PhO2S O 220; 76% yield

PhO2S 216

O

OH

798

33 Oxidation of Aromatic Compounds, Enols and Enolates

This approach can use the inherent regioselectivity of silyl enol ether formation (chapter 3) using kinetic or thermodynamic enolisation. Hence kinetic enolisation of enones (chapter 11) occurs on the α’ side leading to 2-Me3SiO-butadienes such as 222. Epoxidation of this silyl enol ether gives the unstable silyloxy ketone 223 which can be desilylated by fluoride ion and hence transformed into the hydroxyketone 225 or acetoxy ketone 224. These transformations are useful because the hydroxy ketones can be unstable34 (see below). O

OSiMe3 1. LDA

O OSiMe3

mCPBA

2. Me 3SiCl

222; 85% yield

221

223

O Et3NHF Et3N, Ac 2O

OAc

O Et3NHF

223

OH

CH2Cl2

224; 74% yield

225; 83% yield

Chemo- regio- and stereoselectivity: synthesis of fused γ-lactones A dramatic example of selectivity comes in the approach to helenolides by Lansbury and Vacca.36 Many sesquiterpenoids, such as 226, have five-membered lactones fused onto larger rings and conjugation with an exomethylene group gives some anti-cancer properties. The exomethylene group can be added by Wittig and other reactions (see selenium chemistry earlier in the chapter). Lansbury and Vacca decided to explore the strategy of adding the lactone to a protected hydroxyketone having all the rest of the stereochemistry in place. They wanted methods for the disconnection 227 to 228.

H

H

H Wittig

O

?

O etc.

O

HO

O

HO

226

O

RO 228

227

The idea was that the lactone 227 could come from 228 by the sequence: (i) stereo- and regioselective α-hydroxylation 229, (ii) Wittig style reaction on the ketone to add the extra two carbon atoms 230, (iii) stereoselective conjugate reduction of the double bond with the OH group directing the reagent to the bottom face of the alkene, and (iv) lactonisation.

H

H

H

H (iii) + (iv) ?

(ii) ?

(i) ?

OH 228

O

229

O

OH 230

COR

O O 227

33 Hydroxylation of Amino-ketones via Silyl Enol Ethers

799

The reasons for the choice of reagent for each reaction are instructive. Epoxidation of a silyl enol ether was chosen for the hydroxylation for two reasons. It was known that the potentially difficult (the α-atom on both sides is primary) regioselectivity of enolisation of ketones 228 could be solved by using LDA and HMPA. The lithium enolate is formed away from the quaternary ring junction. Then epoxidation of related alkenes was known to go on the bottom face and the stereochemistry of the epoxidation determines the stereoselectivity of the hydroxylation 232. H

H

1. LDA, HMPA 2. Me 3SiCl

H 3. mCPBA

OSiMe3

CH2Cl2 0 ˚C

O t-BuO 228; R = t-Bu

OSiMe3

t-BuO

O 232; 75% yield

t-BuO

231

The unstable silyl ether in 232 was replaced by a more stable acetoxy ketone 233 which was not isolated. A Horner-Wadsworth-Emmons reaction was used for the Wittig-style olefination (chapter 15). A nitrile 234 replaced the carbonyl group in 230 as it was found easier to reduce nitriles in conjugative fashion. H

H OSiMe3

(EtO)2P

O

NaH, DME

O

t-BuO

232

CN

OAc

OAc

DMAP CH2Cl2

t-BuO

H

O

Ac2O Et3N

CN

t-BuO

233

234; 90% yield

The reduction step was stereoselective with either LiAlH4 or LiBH4. In both cases, the hydride reducing agent first cleaved the acetate ester and then an intermediate “ate” complex transferred hydride ion intramolecularly in a cis fashion 235 to the unsaturated nitrile to give the correct stereochemistry 236. Finally cyclisation of the alcohol onto the nitrile to give the trans fused lactone 227 occurred in much the same way that nitriles are converted directly to ethyl esters with ethanol in acid solution. The overall yield of 227 was a respectable 40%. H

234

B H

t-BuO NC

1. NaOH EtOH

H

O

or 3-4 x LiBH4

235

H

H

1.3 x LiAlH4

OH

O 2. H

H CN

t-BuO 236; 92% yield

t-BuO 227; R = t-Bu 40% (over six steps)

O

Hydroxylation of Amino-ketones via Silyl Enol Ethers Synthesis of vinca alkaloids and model compounds It would be a limitation of any of these methods if they could not be used with heterocyclic nitrogen compounds. This is a serious matter because amines are easily converted into N-oxides as we saw in the last chapter. The synthesis of the alkaloid vindoline 237 by Langlois raised this question.

800

33 Oxidation of Aromatic Compounds, Enols and Enolates

Vindoline belongs to the vinca alkaloids, an important class of indole alkaloids whose dimers (such as vincadifformine) are anti-cancer compounds. The most difficult functionality in this compound is the tertiary α-hydroxyester and Langlois37 proposed to make this by hydroxylation of the related keto-ester 238. N

N

H

OH N O H CO2Me Me 237; vindoline

MeO

H

?

N

MeO

Me 238

H

O CO2Me

There were two related questions of chemoselectivity to be answered here: would the amino groups in 238 be too reactive to survive the oxidation with mCPBA and would the rather stable silyl enol ether of the β-ketoester be reactive enough to be epoxidised? Langlois decided to conduct a trial experiment with the simple amino-β-ketoester 239 that contains all these features. If this is successful, there will still be the question of stereoselectivity. The silyl enol ether 240 was duly formed in the most stable position between the two carbonyl groups and epoxidation with an excess of mCPBA gave a good yield of the required α-hydroxyester 243 via the epoxide 241 and the silyl ether 242 which was desilylated in aqueous base without isolation. Me

Me N

O

CO2Me

CO2Me 239

Me

Me OSiMe3

N

2.5 x mCPBA 0 ˚C CH2Cl2

N

OSiMe3

CO2Me

N

Na2CO3

O

241

240

Me O

N

O

H2O

CO2Me OSiMe3 242

CO2Me OH 243; 80% yield

The same conditions were applied to the vindoline precursor 238 and the α-hydroxyester 237 was formed in an even better 89% yield as a single diastereoisomer with the hydroxyl group going in syn to the nearer ring junction hydrogen atom. This puts it on the exo face of the locally folded molecule. The synthesis of vindoline was completed from here.38 The epoxidation of silyl enol ethers is the oldest of the methods in this chapter but it works well and is still very much in use today. The reagents are simple and easy to use, rather unlike the one we are now going to meet.

Second Successful Method: Hydroxylation with MoOPH The reagent: hydroxylation of lithium enolates MoOPH [pronounced “moof” and more correctly oxidoperoxymolybdenum(aqua)-(hexamethylphosphoric triamide)] is a molybdenum peroxide complex 245 that hydroxylates lithium enolates.39 HMPA Me2N HexaP Me 2N Methyl Phosphor- Me2N Amide 244

O

O O O MoOPH Mo Molybdenum O O Oxido O N Peroxy (Me2N)3P HMPA 245

O Me

N

N

246

DMPU Di Methyl Propylene Urea

Me

33 Second Successful Method: Hydroxylation with MoOPH

801

It can be made by a simple process but unfortunately has to contain a molecule of carcinogenic HMPA 244 for stability. Versions with the safer DMPU 246 instead are not so effective. The crystals are best prepared just before use and are pale yellow when pure.40 HMPA

MoO3 + 30% H2O2

MoO5.H2O.HMPA yellow crystals, 67%

1. dry in vacuum

MoO5.pyridine.HMPA

2. pyridine

245; MoOPH, yellow crystals

Stereoselectivity (substrate control) In a typical reaction, the lithium enolate of the carbonyl compound 247 is prepared in the usual way and reacted with yellow MoOPH to give the hydroxylated product 248 and blue molybdenum compounds so that the reaction is self-indicating. The best feature of the method is that it uses lithium enolates since the regioselective preparation of these intermediates is much easier to control than that of other metal enolates. Hence 2-phenylcyclohexanone 249 gives a single regioisomer of the hydroxyketone 250 from kinetic enolisation with good stereoselectivity. OEt

R

2. MoOPH –78 ˚C

O 247

Ph

OEt

R

O

O

OH

1. LDA THF, –78 ˚C

HO

1. LDA

Ph

2. MoOPH

O 248 58-85% yield

249

250; 100% yield 7:1 anti:syn

We have hinted in this chapter at the instability of some of these α-hydroxy carbonyl compounds and the hydroxylation of 249 unfolded some details of this problem. Compound 250 had been reported in 1960 from the hydrolysis of the ester 251 but the compounds produced by the two routes had different NMR spectra. The hydrolysis of the ester 251 actually gave a rearranged hydroxyketone 252. The same product was formed on treatment of 250 with KOH. The instability arises if the carbon atom with the OH group also has a hydrogen atom and can therefore enolise 255. The intermediate ene-diol (or its anion 254 or 253) can re-protonate at either end of the double bond to give different ketones. OH

O AcO

O

Ph

Ph

O

O

OH Ph HO

O

Ph

KOH

KOH

MeOH

MeOH

252

251

253

HO

Ph

H HO

254

255

Hydroxylation of steroids and amino acids The message is that you should not treat α-hydroxy carbonyl compounds with base if they can enolise on the hydroxyl side. Because MoOPH is not an epoxidising agent, there is good chemoselectivity if the molecule also contains alkenes. This steroid example 256 shows that good yields and high stereoselectivity can be obtained. Me O

Me O

256 H THPO

1. LDA, THF, –78 ˚C

H H

H

2. MoOPH, –78 ˚C

H THPO

OH H

257 75% yield

802

33 Oxidation of Aromatic Compounds, Enols and Enolates

There is stereoselectivity here, as in the last example, and both are determined by the substrate, that is the molecule being hydroxylated. The reagent MoOPH is large and it attacks the less hindered side of the substrate under kinetic control. Some control can be achieved in the interesting hydroxylation of aspartic acid enolates.41 The lithium enolate of the dimethyl ester of N-protected aspartic acid 259 gives high syn selectivity in hydroxylation next to only one of the two ester groups 260. CO2H

258 HO2C aspartic acid

CO2Me

MeO2C

259 NH2

HN

Ph

OH 1. LHMDS, THF, HMPA, –78 ˚C

259

2. MoOPH, –78 ˚C

CO2Me

MeO2C

NHR 259, N-protected aspartic 260; 95% yield, 11:1 syn:anti acid dimethyl ester

If 259 is first treated with BuLi and then turned into the lithium enolate and reacted with MoOPH, a good yield of the other diastereoisomer, anti-261 is formed. The first result is probably hydroxylation opposite the very large NHR group in a Houk conformation of the enolate (chapter 21) while anti-selective hydroxylation probably results from chelation control. OH CO2Me

MeO2C

NHR 259

1. BuLi, THF, –78 ˚C 2. LHMDS

CO2Me

MeO2C

3. MoOPH, –78 ˚C

NHR 261; 70% yield, 1:20 syn:anti

The popularity of MoOPH has waned with the discovery of the next type of reagents for hydroxylations. These N-sulfonyl oxaziridines should now probably be your first choice. In the 1986 volume of Organic Syntheses three recipes for hydroxylation of enolates stand side by side. Rubottom42 describes the formation of the kinetic lithium enolate of the enone 262 and the oxidation of the silyl enol ether 263 with mCPBA to give the α’ product 264 (see chapter 11 for the regioselectivity of such extended enolates). O

O

OSiMe3 1. LDA, DME, –15 ˚C

1. mCPBA, hexane, –15 ˚C

2. Me 3SiCl

2. Et 3NHF CH2Cl2

262

HO

263; 88-91%

264; 69-74%

Vedejs and Larsen43 describe the preparation of MoOPH and the hydroxylation of the lithium enolate of camphor 265 to give a good yield of a 5:1 ratio of endo:exo alcohols 266 while Moriarty44 gives an example of a method we have not discussed, the hydroxylation of potassium enolates with o-iodosobenzoic acid.

O

1. LDA

Ph

2. MoOPH

265; camphor

O

O

OH 266; 77% yield 5:1 endo:exo

KOH, MeOH 5-10 °C

IO

CO2H

MeO

OMe

CH2Cl2

Ph OH 267 65% yield

O

5% H2SO4

Ph OH 268 83% yield

33 Third Successful Method: Hydroxylation with N-Sulfonyl Oxaziridines

803

MoOPH is less effective than the boron method for oxidation of aromatic enolates too. In a similar comparison the pyridine 269 was lithiated and treated with oxygen, MoOPH, or the boronate sequence to give the hydroxypyridine 271. The results are very clear.45 OMe

OMe

OMe

reagent

O2 MoOPH

1. B(OMe) 3

PhLi

N

N 269

N

2. AcO 3H

Li

OH 271; 65% yield

270

yield of 271

29%

25%

Third Successful Method: Hydroxylation with N-Sulfonyl Oxaziridines The reagents and the mechanism These reagents contain a three-membered ring made up of C, N, and O atoms - an oxaziridine. The nitrogen atom must be a sulfonamide (usually a benzene sulfonamide) and the carbon atom usually has some substituent. The simplest such reagent 273 is formed by epoxidation of N-phenylsulfonyl benzaldehyde imine 272 with mCPBA or Oxone®, a persulfate potassium salt 2KHSO5. KHSO4.K2SO4. O O Ph

S

O O mCPBA

N 272

Ph

or Oxone®, K 2CO3

Ph

S

N Ph O 273

This compound 273 is a single diastereoisomer and one of the two stereogenic centres is the nitrogen atom. Nitrogen normally inverts too quickly to be stereogenic but in a threemembered ring and with an electronegative substituent, inversion through the unfavourable trigonal nitrogen atom is very slow. It reacts with sodium or potassium enolates to give products of α-hydroxylation.46 OH R2

R1

R1

1. Na or KHMDS

O 274

Me3Si R2

N

SiMe3

[Na or K]

2. oxaziridine 273

O 275

NaHMDS or KHMDS

The enolate attacks the oxygen atom of 273 in an unusual SN2 reaction at oxygen 276. The leaving group is the sulfonamide anion, which is why a sulfonamide is necessary. The anion returns 277 to form the original imine 272 and ejects the anion 278 of the α-hydroxy-carbonyl compound 275. O O

O O Ph

S

N

O

Ph

R2

Ph O

S

Ph 272

N

R2

R1

R1 276

O

O

277

O

R2

R1 O 278

275

804

33 Oxidation of Aromatic Compounds, Enols and Enolates

Hydroxylation of sodium and potassium enolates: side reaction with lithium Lithium enolates 279 do give α-hydroxy-carbonyl compounds but there is a significant side reaction that is a kind of aldol reaction on the imine product 272 through a six-membered cyclic transition state 280 like those we used to explain the stereoselectivity of aldol reactions in chapter 4. Hence the Na or K disilazide bases NaHMDS or KHMDS are usually used. Li OLi

R1

273

PhO2S

R2

R2

PhO2S

R1

Ph 280

279

Li O

O

N

R2

N R1

Ph 281

Substrate controlled stereoselectivity: asymmetric hydroxylation Though the reagent 273 does have stereochemistry it does not transmit it to the product and the stereoselectivity of the reaction is under substrate control. Folded molecules like the lactone 282 react on the outside as expected to give the exo-product 283. O

O

O

O

O

HO

O 1. KHMDS

O

2. 273

O H

H 282

283

Good asymmetric induction occurs with the Evans valine-derived chiral auxiliary (as in 284, see chapter 27) which reacts in alkylation style on the face of the enolate away from the i-Pr group. The products 285 give α-hydroxy-acids 286 on hydrolysis. O

O Ph

N

O

1. NaHMDS 2. 273

i-Pr 284

O

O Ph

N

O O

OH

1. purify by crystallisation 2. LiOH

i-Pr 285; 85% yield 95:5 diastereoisomers

Ph

OH OH 286

Asymmetric hydroxylation with camphor sultam derivatives For asymmetric hydroxylation of enolates it is necessary to provide more permanent asymmetry than is found in 273. The most successful reagents so far have been the (camphorsulfonyl) oxaziridines such as (⫹)-CSO 288. Since both enantiomers of camphor are available and the formation of Oppolzer’s chiral sultam (chapter 27) makes chemistry such as 265 → 287 familiar, these are reagents of some appeal. Derivatives with extra chlorine 289 or methoxy substituents are also available. Epoxidation of the imine 287 occurs exclusively on the lower face to give 288 and hence creates two new stereogenic centres (C and N in the oxaziridine ring) but it is the chirality of the bicyclic camphor system supported by the asymmetric position of the polar sulfone group that leads to successful induction. Hydroxylation of the simple ketone 290 with the chlorosultam 289 gives 291 with high ee.

33 Third Successful Method: Hydroxylation with N-Sulfonyl Oxaziridines

Cl N

N

1. NaHMDS

Ph

Ph

O O S O O O (+)-CSO 288 (+)-chlorosultam 289

S O

O

O Cl

N

805

2. (+)-289

S

O

O 287

Ph

Ph

OH 291; 84% yield 95.4% ee

290

Asymmetric synthesis of tetracycline precursors Strategies for producing optically active α-hydroxy-carbonyl compounds by hydroxylation with oxaziridines depend on whether enantio- or diastereo-selectivity is required. The synthesis of precursors for the antibiotic tetracyclines is a good illustration. These compounds, e.g. 292, have four rings A-D, hence “tetracyclines”. Ring D is usually aromatic but rings A–C have a variety of oxygenated functionality and stereochemistry. These two features come together at the A/B ring junction where there is a difficult tertiary α-hydroxy-carbonyl unit. OH

OH

O

D

C

B

OH

O

O NH2

A

H H NMe2 Me OH OH 292; oxytetracycline

OH

One strategy for tetracycline synthesis is to build a B/C precursor with suitable functionality for extension to ring A and with the tertiary α-hydroxy-carbonyl unit already in place. Two potential starting materials for different tetracyclines are 293 and 295 and both might be made by α-hydroxylation of enolates. H O OH

BnO C

BnO

O

OMe O

OH

OMe O CO2Me

H

CO2Me

?

?

B

OMe

H

C OMe

293

294

OMe

B

295

OMe

296

Both precursors 294 and 296 are chiral but the chirality of 296 lies at an enolisable centre between two carbonyl groups and the enolate, the intermediate in hydroxylation, is only prochiral. It makes sense therefore to use racemic 296 with the correct optically active sultam 297 to make 295 by asymmetric hydroxylation47 but to use optically active precursor 294 with racemic sultam 273 in diastereoselective hydroxylation48 (54% yield). OMe O 294

1. NaHMDS 2. O

293

O

296

1.KHMDS 2. 297

OH

OMe CO2Me N

OMe

S Ph

N 273

Ph O

O S O O 295; 70% yield, >95% ee (+)-dimethoxy-sultam 297 OMe

806

33 Oxidation of Aromatic Compounds, Enols and Enolates

Synthesis of a calcium channel opening drug The Bristol-Meyers Squibb calcium channel opener 298 can clearly be made by addition of some protected organometallic derivative 300 to the isatin 299. In fact this is easily done with a Grignard reagent from 301 with an OMe group at the phenolic position. In the synthesis (below) a preliminary treatment of 299 with NaH was needed before addition of the Grignard reagent H N

CF3

O C–C

H N

CF3

M

OH

O +

OH OH

Cl

O 298

protect OH

OMe

= Cl

299

Cl

Br

300

301

The chemists at B-MS were faced with the problem of determining the biological activity of both enantiomers of the drug. The availability of both enantiomers (from both enantiomers of camphor) of the Davies oxaziridine 288 made it worthwhile to remove the OH group from 302 and put it back again.49 Protected 302 was reduced with Et 3SiH and TFA, a reagent that reduces alcohols that easily form cations, to 303. Then this was enantiospecifically hydroxylated with each enantiomer of 288 to give, after deprotection with BBr3, the two enantiomers of the drug 298. H N

CF3

H N

CF3

O

O Et3SiH

1. Mg

301

Cl

O 1. KHMDS

H

OH CF3CO2H OMe

2. 299 + NaH

H N

CF3

OMe

2. (+)-288

Cl

BBr3 CH2Cl2

(+)-298 >95% ee

Cl (–)-304

303

302

OH OMe

Asymmetric synthesis of buproprion While a drug is being evaluated, it is necessary to prepare both enantiomers. When you know which enantiomer you want, you need an asymmetric (and preferably catalytic) synthesis. GlaxoSmithKline’s anti-depressant drug buproprion50 305 provides an excellent example. This simple compound is needed as the (R)-enantiomer and can obviously be made by SN2 displacement (with inversion) on some derivative of the (S)-alcohol 306. Asymmetric hydroxylation of the enolate of the simple aryl ketone 307 should answer. Buproprion – Analysis O Cl

C–N

NHt-Bu (R)-305

O

O Cl

asymmetric

inversion

OH (S)-306

Cl

hydroxylation of enolate

307

33 References

807

A chiral oxaziridine could have been used but the company prefers a catalytic method: the asymmetric dihydroxylation (chapter 25) of the silyl enol ether 308. Presumably the diol 309 is formed but the hemiacetal collapses on workup and the mesylate of (R)-306 allows displacement with t-BuNH2 and the synthesis of buprion 307. Buproprion – Asymmetric Synthesis OSiMe3 307

Me3SiCl base

AD

Cl

HO

OSiMe3 work-up

Cl

(S)-306

1. MsCl 2. t-BuNH2

(R)-305 buproprion

OH 308

309

Summary Electrophilic substitution of pyridines (chapter 32) and the hydroxylation of aromatic compounds and enol(ate)s (this chapter) are now useful and flexible reactions that can be made to work on a variety of compounds with all the necessary selectivity. Functionalisation as a strategy - that is the removal of functionality during analysis - can be difficult to find during the disconnection process as one naturally tends to be thinking of making rapid progress in simplification by carbon– carbon disconnections. The secret is to keep it in the mind alongside the FGA strategy and ask the question: would this synthesis be easier if I either removed (functionalisation strategy) or added (FGA strategy) a functional group to the target molecule? A comprehensive review51 of this topic appeared in 2003.

References General references are given on page 893 1. LaV. L. Brown, S. Kulkarni, O. A. Pavlova, A. O. Koren, A. G. Mukhin, A. H. Newman, and A. G. Hortl, J. Med. Chem., 2002, 45, 2841. 2. D. J. Dale, J. Draper, P. J. Dunn, M. L. Hughes, F. Hussain, P. C. Levett, G. B. Ward and A. S. Wood, Org. Process Res. Dev., 2002, 6, 767. 3. P. M. McEwen and M. V. Sargent, J. Chem. Soc., Perkin Trans. 1, 1981, 883. 4. K. Green, J. Org. Chem., 1991, 56, 4325. 5. M. F. Hawthorne, J. Org. Chem., 1957, 22, 1001. 6. See for example F. Mongin and G. Quéguiner, Tetrahedron, 2001, 57, 4059. 7. F. Mongin, F. Trécourt, B. Gervais, O. Mongin and G. Quéguiner, J. Org. Chem., 2002, 67, 3272. 8. F. Trécourt, M. Mallet, O. Mongin and G. Quéguiner, J. Org. Chem., 1994, 59, 6173. 9. Vogel, p. 629. 10. M. S. Newman and W. R. Reichele, Org. Syn. Coll., 1973, V, 1024. 11. B. L. Murr, G. B. Hoey and C. T. Lester, J. Am. Chem. Soc., 1955, 77, 4430; J. L. Harper and C. T. Lester, J. Org. Chem., 1961, 26, 1294. 12. Vogel, p. 630. 13. E. Müller, H. Metzger and D. Fries, Chem. Ber., 1954, 87, 1449. 14. W. Oppolzer, O. Tamura and J. Deerberg, Helv. Chim. Acta, 1992, 75, 1965. 15. W. Oppolzer, C. G. Bochet and E. Merifield, Tetrahedron Lett., 1994, 35, 7015. 16. N. Momiyama and H.Yamamoto, Org. Lett., 2002, 4, 3579. 17. N. Momiyama and H. Yamamoto, Angew. Chem. Int. Ed., 2002, 41, 2986; there is a correction on page 3313.

808

33 Oxidation of Aromatic Compounds, Enols and Enolates

18. Y. Ito, T. Hirao and T. Saegusa, J. Org. Chem., 1978, 43, 1011; B. M. Trost, Y. Nishimura, K. Yamamoto and S. S. McElvain, J. Am. Chem. Soc., 1979, 101, 1328; M. R. Roberts and R. H. Schlessinger, J. Am. Chem. Soc., 1979, 101, 7626; J. Tsuji, Palladium Reagents and Catalysts, Wiley, Chichester, 1995, page 104. 19. M. Hirama and M. Iwashita, Tetrahedron Lett., 1983, 24, 1811. 20. M. Hirama and M. Uei, J. Am. Chem. Soc., 1982, 102, 4251. 21. J. W. Pilkington and A. J. Waring, J. Chem. Soc., Perkin Trans. 2, 1976, 1349. 22. B. M. Trost, T. N. Salzmann and K. Hiroi, J. Am. Chem. Soc., 1976, 98, 4887. 23. H. J. Reich, J. M. Renga and I. L. Reich, J. Am. Chem. Soc., 1975, 97, 5434. 24. R. E. Zipkin, N. R. Natale, I. M. Taffer and R. O. Hutchins, Synthesis, 1980, 1035. 25. P. A. Grieco and M. Miyashita, J. Org. Chem., 1974, 39, 120. 26. M. Braun, B. Meyer and B. Féaux de Lacroix, Eur. J. Org. Chem., 2002, 1424. 27. W. Reusch and C. K. Johnson, J. Org. Chem., 1963, 28, 2557. 28. J. D. McChesney and A. F. Wycpalek, Chem. Commun., 1971, 542. 29. J. A. Elings, H. E. B. Lempers and R. A. Sheldon, Eur J. Org. Chem., 2000, 1905. 30. D. Felix, J. Schreiber, G. Ohloff and A. Eschenmoser, Helv. Chim. Acta, 1971, 54, 2896. 31. R. D. Clark and C. H. Heathcock, Tetrahedron Lett., 1974, 2027; C. H. Heathcock, C. Mahaim, M. F. Schlecht and T. Utawanit, J. Org. Chem., 1984, 49, 3264. 32. R. B. Woodward, Pure Appl. Chem., 1968, 17, 629 (see page 529). 33. A. Hassner, R. H. Reuss and H. W. Pinnick, J. Org. Chem., 1975, 40, 3427. 34. G. M. Rubottom and J. M. Gruber, J. Org. Chem., 1978, 43, 1599. 35. L. A. Paquette, H. Lin and J. C. Gallucci, Tetrahedron Lett., 1987, 28, 1363. 36. P. T. Lansbury and J. P. Vacca, Tetrahedron Lett., 1982, 23, 2623. 37. R. Z. Andriamialisoa, N. Langlois and Y. Langlois, Tetrahedron Lett., 1985, 26, 3563. 38. R. Z. Andriamialisoa, N. Langlois and Y. Langlois, J. Org. Chem., 1985, 50, 961. 39. E. Vedejs, D. A. Engler and J. E. Telschow, J. Org. Chem., 1978, 43, 188. 40. E. Vedejs and S. Larsen, Org. Synth., 1986, 64, 127. 41. F. J. Sardina, M. M. Paz, E. Fernández-Megía, R. F. de Boer and M. P. Alvarez, Tetrahedron Lett., 1992, 33, 4637. 42. G. M. Rubottom, J. M. Gruber, H. D. Juve and D. A. Charleson, Org. Synth., 1986, 64, 118. 43. E. Vedejs and S. Larsen, Org Synth., 1986, 64, 127. 44. R. M. Moriarty, K. Hou, I. Prakash and S. K. Arora, Org. Synth., 1986, 64, 138. 45. F. Trécourt, M. Mallet, O. Mongin, B. Gervais and G. Quéguiner, Tetrahedron, 1993, 49, 8373. 46. F. A. Davis and B.-C. Chen, Chem. Rev., 1992, 92, 919–934. 47. A. I. Meyers and K. Higashiyama, J. Org. Chem., 1987, 52, 4592. 48. F. A. Davis, A. Kumar and B.-C. Chen, Tetrahedron Lett., 1991, 32, 867. 49. P. Hewawasam, N. A. Meanwell, V. K. Gribkoff, S. I. Dworetzky and C. G. Biossard, Bioorg. Med. Chem. Lett., 1997, 7, 1255. 50. S. C. Stinson, Chem. Eng. News, Oct 23rd 2000, page 55. 51. B.-C. Chen, P. Zhou, F. A. Davis and E. Ciganek, Org. React., 2003, 62, 1.

34

Functionality and Pericyclic Reactions: Nitrogen Heterocycles by Cycloadditions and Sigmatropic Rearrangements This chapter considers the effect of nitrogen atoms on Diels-Alder reactions, [3,3]-sigmatropic rearrangements such as the Aza-Cope, group transfer reactions such as the Alder Ene reaction and other pericyclic processes

PART I–INTRODUCTION The Effects of Functionality on Pericyclic Reactions Cycloadditions: The Diels-Alder and the Alder ‘ene’ reactions [3,3] and [2,3]Sigmatropic rearrangements: the aza Cope rearrangement Other pericyclic processes giving nitrogen heterocycles PART II– CYCLOADDITIONS TO MAKE NITROGEN HETEROCYCLES Diels-Alder Reactions with Azadienes Problems with the Diels-Alder reactions with azadienes Stable 1-azadienes Stable nucleophilic 1-azadienes (HOMO used in cycloadditions) Stable electrophilic 1-azadienes (LUMO used in cycloadditions) Diels-Alder reactions with 2-azadienes Diels-Alder Reactions with Imines Intramolecular Diels-Alder Reactions with Azadienes Intramolecular Diels-Alder Reactions with Imines Intramolecular Diels-Alder Reactions with a Nitrogen Tether PART III–‘ENE’ REACTIONS TO MAKE NITROGEN HETEROCYCLES Intramolecular Alder ‘Ene’ Reactions with a Nitrogen Tether Intermolecular ‘Ene’-style Reactions with Oximes PART IV– [3,3] SIGMATROPIC REARRANGEMENTS The ‘Aza-Cope’ Rearrangement The Anionic ‘Aza-Cope’ Rearrangement PART V– OTHER REACTIONS Electrocyclic Reactions Ring Closing Olefin Metathesis The Pauson-Khand Reaction Metal-Catalysed Alkyne Trimerisation

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

810

34 Functionality and Pericyclic Reactions

PART I – INTRODUCTION The Effects of Functionality on Pericyclic Reactions Cycloadditions: The Diels-Alder and the Alder ‘ene’ reactions Our understanding of pericyclic reactions is ultimately based on Woodward and Hoffmann’s analysis of orbital symmetry. This chapter considers the consequences of inserting functionality into a pericyclic reaction so that the symmetry of at least one component is perturbed. For example, a nitrogen atom could be part of the diene in a Diels-Alder reaction in two ways: we could have a 1- or a 2- ‘azadiene’ (1 or 3). If the reaction is successful the product in both cases (2 or 4) would be a nitrogen heterocycle. Alternatively, we could use an imine 5 as the dienophile. Again, assuming that the reaction works, the product is a nitrogen heterocycle. In fact, as we shall see later, the influence of the added nitrogen atom is pernicious and none of these three reactions works very well with these simple compounds. This chapter will discuss what effects such functionality has on pericyclic reactions and how the difficulties can be overcome.1 R

R

R R

N

R

N

?

1

N

2

R

?

N 3

R

R

R R

N

4

R

?

N

5

R

6

To limit the scope of the discussion and also to give it a useful synthetic focus, we shall consider only those reactions that lead to heterocyclic nitrogen compounds. These products may be aromatic but will more usually be partly saturated so there will be a good deal of stereochemistry and some asymmetric synthesis. We shall therefore also discuss reactions in which the nitrogen atom is not part of the pericyclic process as such but influences it strongly by being part of a tether. The regioand stereo-selectivity in Diels-Alder reactions such as the cyclisation of 7 is largely controlled by the length and position of the tether so the functional group does influence the pericyclic process. The Alder ‘ene’ reaction 9 is a ‘group transfer’ reaction closely related to the Diels-Alder reaction but with a C–H bond in place of one of the alkenes in the diene. This reaction does not of itself close a ring so once again a nitrogen tether will be necessary if a heterocycle 10 is to be formed. R

R ?

R

R

R

R

?

NR

H

NR 8

NR

R

R 7

H

NR 9

10

[3,3] and [2,3]Sigmatropic rearrangements: the aza Cope rearrangement Typical [3,3] sigmatropic rearrangements include the Claisen and Cope rearrangements. The ‘aza-Cope’ is a [3,3] sigmatropic rearrangement with a nitrogen atom in one of two possible positions 11 or 13. In neither case is a ring of any kind formed in this reaction but in both cases the product is a rather unstable imine and further ionic reactions may lead to heterocycles.

34 Diels-Alder Reactions with Azadienes

R

[3,3]

N

R

R

N

11

12

811

N

R

[3,3]

13

N

14

The tandem reactions in which the aza-Cope is followed by an ionic C–C bond-forming second step are often carried out on anions or cations. Here is an anionic example 15 to 16 we shall be discussing later in the chapter. Some of the most dramatic [2,3] sigmatropic rearrangements have been of nitrogen ylids undergoing ring expansion to give medium rings. We shall discuss the ring expansion of 17 to give a nine-membered cyclic amine 18 by [2,3] rearrangement of an ylid later. H N

H N

R R

O

R

R

KH

N

R

18-crown-6

O

15

CO2R

R DBU

N

17

16

CO2R

18

Other pericyclic processes giving nitrogen heterocycles The chapter ends with a rather miscellaneous collection of reactions such as one example of an electrocyclic ring closure 19 . There are more important examples of ring closing olefin metathesis (RCM) and other organometallic reactions such as the Pauson-Khand reaction and co-trimerisation.

R

R

electrocyclic

N

19

N

RCM

20

N

N

R 21

R 22

PART II– CYCLOADDITIONS TO MAKE NITROGEN HETEROCYCLES Diels-Alder Reactions with Azadienes Problems with the Diels-Alder reactions with azadienes 1-Azadienes 24 are easily made from amines and enals such as 23 . The Diels-Alder reaction of 24 with maleic anhydride looks very attractive 25 and you might expect 26 to be the product.

O

Ph

Ph

N

N

O

Ph

O

N

PhNH2

O

O

O R 23

R 24

R 25

O R 26

812

34 Functionality and Pericyclic Reactions

Unfortunately the nitrogen atom lowers the HOMO energy of the diene 24 and it is not a good match with the LUMO of maleic anhydride. Instead the imine 24 is in equilibrium with the enamine 27 and, in conformation 27b, this has a high energy HOMO and so the product2 of the reaction is 28. The tautomerism between 24 and 27 may be especially easy as it can be drawn as a [1,5]H sigmatropic shift. This conflict between weakly electrophilic imines and their tautomers, the nucleophilic enamines, is one of the themes of this chapter.1 Ph

Ph

N

NH

NHPh O

NHPh O

O

24

O

R

R

R 27a

O O

R 28

27b

One solution to the imine/enamine equilibrium is to load the nitrogen atom with a removable acyl group. Popular choices are urethanes and sulfonamides. If the nitrogen atom is unsubstituted 32, the imine 33 is favoured by conjugation but if the nitrogen is substituted 31, it is forced to give the enamine 29. enamine H N R

CO2R N

ClCO2R

R

R

imine N

R

29

enamine NH2

R

NH

R

R

30

imine

31

CO2R N

ClCO2R

R

32

R

33

34

There is a further disadvantage in using simple Diels-Alder reaction of azadienes to make heterocycles. The products, e.g. 2 and 4 are themselves either imines or enamines and are inherently unstable and it is also necessary to stabilise them to make isolation of the heterocycle possible.

Stable 1-azadienes Among the few examples of simple 1-azadiene Diels-Alder reactions is a dihydropyridine synthesis using the stable azadiene 39 (prepared from cinnamaldehyde and aniline) with the dienophile 38 prepared from the isoxazole 35 by elimination. This is a reverse-electron-demand cycloaddition, the HOMO of the dienophile 38 combining with the LUMO of the azadiene 39 to give the cycloadduct 40 and hence the dihydropyridine 41 with complete regioselectivity and in very high yield.3 H

Li H

O

NC

H LDA

N Ph

–78 ˚C

O

O

Ph 36

H

Ph 37

Ph NC

H

–78 ˚C

Me3SiO Ph 38; 92% yield Ph

Ph NC

80 ˚C

Me3SiO

Ph 38

N Ph 39

H

Me3SiCl

35 NC

NC

H

N

Me3SiO

Ph

40

N

N

Ph

Ph

Ph 41; 98% yield

34 Diels-Alder Reactions with Azadienes

813

Other dienes, even cyclopentadiene, perform poorly with this dienophile 38 but the azadiene 39 that does so well here is a very special case. It is stabilised by the phenyl groups at both ends of the molecule and it cannot tautomerise into an enamine. More generally useful 1-azadienes must be stabilised by conjugating substituents that are definitely electron-donating or withdrawing so that the HOMO or the LUMO is unambiguously selected for cycloaddition.

Stable nucleophilic 1-azadienes (HOMO used in cycloadditions) Ghosez used hydrazones 35 for this purpose. The regioselectivity shows that the NMe2 group dominates the HOMO of the diene in reaction with the LUMO of dienophiles such as acrylonitrile. The arrows on the mechanism 44 show the nucleophilic end of the diene 43 interacting with the electrophilic end of acrylonitrile. The cycloaddition may or may not be concerted.4

CHO

NH2

N

N

NMe2

N

NMe2 44

NMe2 43

42

CN

CN

NMe2 45

Though the NMe2 group can be reductively removed from 45, it has generally been more valuable to convert the products into pyridines by elimination and oxidation. Reaction of the same diene 43 with the unsymmetrical naphthaquinone 46 gave a single regioisomer of 48 after treatment of the unstable adduct 47 with refluxing ethanol.5 O

OMe

O

43 room temperature

O

OMe

OMe

air EtOH

24 hours

N O

N

NMe2 O 47

46

O 48

If the substituent is OH instead of OMe, hydrogen bonding to one of the carbonyl groups of the naphthaquinone 49 reverses the coefficients in the LUMO and the opposite regioselectivity is found in the product 50. HOMO

O

O

LUMO room temperature 24 hours

N NMe2 43

N O

H 49

O

O

OH

50

In all these examples the 1-azadiene is stabilised by being a hydrazone, but most published examples cannot tautomerise to enamines as they do not have proton-bearing substituents in the right position.

Stable electrophilic 1-azadienes (LUMO used in cycloadditions) Sulfonamides are the most widely used electrophilic 1-azadienes, e.g. 51, and they react with electron-rich dienophiles such as enol ethers in reverse electron demand Diels-Alder reactions.6

814

34 Functionality and Pericyclic Reactions

The LUMO of the diene 51, activated by the imine, interacts 52 with the high energy HOMO of the enol ether to accelerate the reaction and determine the regioselectivity of the cycloadduct 53. LUMO

O

HOMO

NH2

N

N

SO2Ph

SO2Ph

SO2Ph

115 ˚C 48 hours

OEt

OEt

SO2Ph 53; 73% yield

52

51

N

Addition of another electron-withdrawing group to the azadiene (the ester in 54) lowers the LUMO energy even more and the reaction 55 can now be run at room temperature (though pressure helps) with perfect regioselectivity 57 and good endo:exo stereoselectivity from the interaction of the OEt group with the back of the diene 56. Ph

Ph

Ph

EtO2C

EtO2C

H

EtO2C

N

N

SO2Ph

SO2Ph

54

OEt

Ph H H H

O

N

Et

SO2Ph

55

25 ˚C

EtO2C

H H

24 hours

N

OEt

SO2Ph 57; 80% yield >20:1 endo:exo

56

More impressively, azadienes, such as 59, that could tautomerise to enamines 58 also give good yields of adducts 60 with high regioselectivity and endo stereoselectivity. 1-Azadienes are still not easy to use but enough methods are now available to make heterocycle synthesis by this route a practical proposition. H

OEt H N

OEt

110 ˚C

´

N

SO2Ph

58; enamine not formed

SO2Ph

N

48 hours toluene

SO2Ph

60; 79%, >20:1 endo:exo

59

Diels-Alder reactions with 2-azadienes The same problems arise with 2-azadienes and Ghosez7 has solved them by loading the dienes with strongly electron-donating or withdrawing groups. The imide 61 was doubly silylated to give a 2-aza-diene 62 that cannot tautomerise to an enamine. It adds regioselectively to an acetylenic ester to give a pyridone 64 after an acidic work-up that hydrolyses the silyl-imino ether of 63 and catalyses the aromatisation by loss of the other silyl ether. reflux CHCl3

RO

O TBDMSOTf

HN O 61

H

Et3N

N OR 62; R=TBDMS 86% yield

RO 1M HCl

N

CO2Me

CO2Me

OR 63; R=TBDMS

O HN

CO2Me

64; 64% yield

34 Diels-Alder Reactions with Azadienes

815

The regioselectivity of the addition is determined by the HOMO of the azadiene 62 which is dominated by the two electron-donating silyloxy groups. The nitrogen atom is almost irrelevant. The LUMO of the unsaturated ester is typical for any Diels-Alder reaction. TBDMSO HOMO of azadiene 62

N OTBDMS

CO2Me

LUMO of acetylenic ester

For a reverse electron-demand Diels-Alder with azadienes, Barluenga8 reacted stable silylated imines 65 of unenolisable aldehydes (R ⫽ Ar or cinnamyl) with acetylene dicarboxylic esters to give the 2-azadienes 67. CO2Me

CO2Me

toluene 25 ˚C Me3Si

N

CO2Me

MeO2C

CO2Me

SiMe3

CsF MeOH

N

R

MeO2C N

65

R 67; 94-97% yield

R 66

The imine nitrogen must attack the acetylene 68 and the resulting enolate desilylate the imine cation 69 to trap the diene as 70. No tautomerism to an enamine is possible. This intermediate is desilylated with fluoride to give the diene 67 in remarkably good yield. O OMe MeO2C

MeO2C

OMe

•

MeO2C

SiMe3 N

SiMe3

N

CO2Me

O

N

SiMe3

R

R

R 68

69

70

These dienes 70 have three electron-withdrawing groups, the imine and two esters, and react with electron-donating enamines, e.g. 71, to give pyridines 74, again in excellent yield, after an acidic work-up that tautomerises the initially formed imine 72 to the stable conjugated enamine 73 and then eliminates the amine used to make the enamine. The final oxidative aromatisation is spontaneous in air. MeO2C

MeO2C H

MeO2C 67 25 ˚C

N

MeO2C

R 72

N

MeO2C 2M HCl

HN

N

CH2Cl2

71

MeO2C H

R 73

N

THF 50 ˚C

N R 74; 95% yield

816

34 Functionality and Pericyclic Reactions

The regiochemistry of the cycloaddition is curious. The imine and the terminal ester groups direct to one end but the middle ester group dominates and directs the enamine to the other end. The diene LUMO evidently has its largest coefficient as shown, and seems again to ignore the imine, though whether this is a concerted cycloaddition or a two-step ionic reaction is uncertain. CO2Me MeO2C LUMO of 70

HOMO of 71 N

N R

Diels-Alder Reactions with Imines The other side of the coin is a Diels-Alder reaction between an ordinary diene and an imine.9 Here the problem is the instability of the imine. An ingenious solution was found by Kraus10 who distilled a THF solution of the trimer 75 into a flask cooled to ⫺78 ⬚C. The imine 76 was stable at that temperature and could be converted into the stable acylated enamine 77.

N ClCO2Me

distil with THF

N

N

N

75

Et3N –78 ˚C

MeO2C

76

N

77

When Danishefsky11 set out to synthesise ipalbidine 83, a Diels-Alder reaction of this imine 77 was an obvious approach. To get efficient reaction, it was necessary to use a diene 79 activated by two electron-donating ethers and to activate the imine with a Lewis acid. The orientation of 81 is as expected from the LUMO of the imine reacting with the HOMO of the diene.

1. LDA HMPA

OEt O

MeO

2. t-BuMe2SiCl

77

Ar

OEt

BF3 CH2Cl2

OTBDMS 79; 100% yield

78

N

N

Ar

EtO OTBDMS MeO

O 81; 40-45% yield

80

Conversion into ipalbidine 83 required reduction of the lactam 81 and demethylation of the aryl methyl ether 82. Though the yield in the cycloaddition was poor, this is a short synthesis.

N

Ar

LiAlH4

N

AlCl3

BBr3

N

CH2Cl2

O MeO 81

HO 82; 73% yield

83; 78% yield

34 Intramolecular Diels-Alder Reactions with Azadienes

817

Even simple imines can be used if the diene is activated enough and Danishefsky’s diene 84 is nucleophilic enough to combine with simple imines with Lewis acid catalysis. This example 85 could tautomerise to an enamine but the imine, especially as its ZnCl2 complex, reacts better with such a nucleophilic diene.12 OMe

OMe N

Me3SiO

Ph

N

ZnCl2, THF excess of imine room temperature 36-48 hours

Me3SiO

84

Ph

O 86; 69% yield

85

From the many examples of reactions with imines activated in various ways, we select the asymmetric synthesis of pipecolic acids by double diastereomeric induction using an activated imine 87 with two chiral auxiliaries - α-methylbenzylamine on the nitrogen atom and 8-phenylmenthol as esterifying group. Both these auxiliaries were discussed in earlier chapters (22 and 27). Acidcatalysed cycloaddition with normal dienes gives excellent asymmetric induction and reasonable yield. We show the matched case - the mismatched case is obviously worse.13

O Ph

N

N

O

CF3CH2OH, CF 3CO2H

Me

24 hours room temperature

Ph

87

OR*

O Ph Me 88; 53% yield >95% de

Intramolecular Diels-Alder Reactions with Azadienes One sure way to make the hetero Diels-Alder reaction go well and with regio-and stereoselectivity is to make it intramolecular.14 The highly reactive diene 90 is formed by elimination with fluoride ion from 89. The cycloaddition is rapid because it restores aromaticity. The regioselectivity is as expected from HOMO of dienophile plus LUMO of diene but is probably controlled by the tether.15 The stereochemistry certainly is: the azadiene is attached to the bottom face of the five-membered ring and is delivered to the bottom face of the dienophile to give the all-trans ‘aza-oestrone’ 91. F

SiMe3

RO

N

RO OR

O

OR

CsF

N NMe3 89

H

N

MeCN reflux

H 90

then H H2O

H H

91

Heathcock has used a remarkable cascade of reactions, including an intramolecular Diels-Alder reaction of a protonated 2-aza-diene, in the synthesis of Daphniphyllum alkaloids. We give one example. The diol 92 gives the unstable dialdehyde 93 by Swern oxidation and hence the protonated

818

34 Functionality and Pericyclic Reactions

azadiene 94 by reaction with ammonia. As one of the aldehydes in 93 is joined to a quaternary centre and cannot form an enamine, only one such azadiene can be formed. But now that you can see the side chain ‘R1’, what happens to 94? Several pericyclic reactions are possible.

R R

R

H (COCl)2 DMSO

OH OH

1. NH 3 CH2Cl2

OHC

Et3N CH2Cl2

R1

H

OHC

92

H

HN

?

2. HOAc

R1

93

94

The product is the cage amine 97. The azadiene must cycloadd to the nearer of the two unactivated alkenes in the conformation 95 (mechanism not shown for clarity). Though the regioselectivity is what we should expect from the HOMO of the dienophile and the LUMO of the protonated 2-azadiene, the short tether (two saturated carbon atoms) is probably responsible. It is certainly responsible for the stereochemistry as the tether delivers the dienophile to the top face of the diene.16 The product 96 cyclises by addition of the remaining alkene to the immonium ion to give 97. We shall see more of these tandem sequences later in the sections on sigmatropic rearrangements. R

H

H

R

R

aza-DielsAlder

N

H

95

N

HN 97

96

Intramolecular Diels-Alder Reactions with Imines Simple imines give good yields of intramolecular Diels-Alder reactions providing always that the second ring formed is stable and that usually means five- or six-membered. Grieco has developed a simple method whereby an aldehyde, also containing a remote diene, is combined with an amine in acidic solution. The immonium ion cannot usually be isolated as cycloaddition 99 is rapid. A simple example 98 gives the tricyclic amine17 101

Me CHO

MeNH2.HCl

N

H

Me

N

H

Me

N

H2O, EtOH

98

99

100

101

This reaction is obviously useful for the synthesis of hydroquinolines and compounds such as pumiliotoxin C 102. The aza-Diels-Alder disconnection 103 requires a preliminary FGA. The required aldehyde 105 was made from the known enantiomerically pure 106.

34 Intramolecular Diels-Alder Reactions with a Nitrogen Tether

Me

Me

H

azaDielsAlder

H

FGA

Me

N Pr H H H 103

Me

Me C–N

8a N H H H 102; pumilotoxin C

819

Pr

NH

CHO

imine

104

OR

Pr

105

OH

106

When the reaction was carried out by treating the aldehyde (⫺)-105 with ammonium chloride, the cycloaddition took place under quite mild conditions (75 ⬚C in aqueous ethanol) with complete regiochemical control but gave a mixture of two diastereoisomers 108 and 109 in a 70:30 ratio but none of the right isomer for the pumiliotoxin synthesis.17 The major product 108 has three of the chiral centres correct but is epimeric with pumiliotoxin C 102 at C-8a. Grieco has used this reaction successfully for other alkaloids.18 Me

Me

Me

Me

H

H

NH4Cl

8a

H2O, EtOH

CHO

75 ˚C

Pr

(–)-105

N Pr N Pr H H H H H H 108 2.2:1 108:109 109

Pr

NH2 107

This result shows both the strengths and the weaknesses of such intramolecular aza-Diels-Alder reactions. The stereochemical outcome is controlled by the tether and such considerations as endo/exo selectivity matter very little. The molecule must fold 110 so that the preferred transition state 111 has the newly formed carbocyclic ring (not the ring formed by the cycloaddition itself) in the best chair conformation 112. The reaction is useful only if you want the stereochemistry it naturally gives. H

H H

H

Me

Me NH2

Me

NH2

NH2

H

H

H

H 112

H 111

110

H

Intramolecular Diels-Alder Reactions with a Nitrogen Tether The last example shows how intramolecular Diels-Alder reactions can produce two rings simultaneously. A simple Diels-Alder reaction (i.e. not ‘aza’) with a nitrogen atom in the tether simply switches the two rings round: now two new bonds in the carbocyclic ring are made by the cycloaddition while one of them also completes the heterocyclic ring. In simple examples both the regio- and stereochemistry is controlled by the tether: the carbamate-stabilised enamine 113 gives only the cis ring junction 116 via an endo-like transition state 115. This time the relative stereochemistry is correct for pumiliotoxin C 102 though the yield is poor.19 H HH 190 ˚C

N CO2Me 113

N CO2Me 114

toluene 24 hours

N

CO2Me H

115

N

CO2Me 116; 25% yield

820

34 Functionality and Pericyclic Reactions

There are many examples of such reactions being used in synthesis. Often many new chiral centres are established in one step as in the lycorine synthesis of Boeckmann.20 The starting material 118 is prepared by elimination on the spirocyclic ammonium salt 117 and contains no 3D stereochemistry. OAc

CO2Me

MeO2C OAc O

O

DBU

N O

N

O 117

118

Cycloaddition gives a single diastereoisomer of adduct 120 with four new chiral centres. The molecule 118 must fold 119 so that the ester group is indeed endo to the diene but that, and the regioselectivity, are mainly due to the very short tether - just one CH2 group between two rigid rings. The product 120 is a late intermediate in a synthesis of lycorane. CO2Me

OAc

OAc 110 ˚C

O

4 hours toluene

N

O

O

118

MeO2C H

H N

O

OAc

CO2Me

H H H

H

O N

O

119

120

PART III – ‘ENE’ REACTIONS TO MAKE NITROGEN HETEROCYCLES Intramolecular Alder ‘Ene’ Reactions with a Nitrogen Tether The Alder ‘ene’ reaction is like a Diels Alder reaction in which one π-bond in the diene has been replaced by a C–H bond 121. It does not therefore form a ring and does not fit easily into any of the three classes of pericyclic reaction (cycloaddition, electrocyclic, and sigmatropic). Since a hydrogen atom is transferred from one component to the other it is best described as a ‘group transfer’ reaction.21 The regioselectivity is determined by the interaction 123 with the π-bond of the ‘ene’ (the HOMO) with the LUMO of the ‘enophile.’

HOMO ‘ene’

LUMO

‘enophile’ H 121

CO2Me

H 122

CO2Me

CO2Me

H 123

The ‘ene’ reaction forms a heterocyclic ring if the two components are tethered by a nitrogen atom. Simple examples described by Oppolzer and Snieckus show that there is some kinetic preference for the ‘endo’ product. After 24 hours at 150 ⬚C, the syn product 126 is the only one from the intramolecular ‘ene’ reaction22 125.

34 Intramolecular Alder ‘Ene’ Reactions with a Nitrogen Tether

SiMe3

SiMe3

SiMe3

N

O

N

N

O

150 ˚C

821

O

24 hours

CO2Me H 124

H CO2Me 125

CO2Me 126

If the reaction mixture, or compound 124 alone, is heated to ⬎250 ⬚C, the more stable anti isomer 129 is the only product. Though there are two carbonyl groups conjugated with the ‘enophile’ only the amide carbonyl can interact with the π-bond of the ‘ene’ so that the endo transition state 127 leads to the syn product 126. The exo transition state 128 leads to the anti product 129 and equilibration is presumably by reverse ‘ene’ reaction. Me3Si

SiMe3

SiMe3

N

SiMe3

N

N

N H H

O

H

H

O

H O

H H CO2Me

CO2Me

127 (endo)

MeO2C

H

O H CO2Me

H

128 (exo)

126

129

If the carbonyl group responsible for the endo secondary orbital overlap is not there, as in 130, the exo product 131 is formed at lower temperatures. This was just as well as the reaction was used to make allokainic acid 132 having the stereochemistry of 131. CF3

CF3

O

O N

N CO2Et CO2Et

H

CO2Et 130

80 ˚C 16 hours

H N CO2H

CO2Et CO2Et

CO2Et 131; 97% yield

CO2H 132; (+)-α-allokainic acid

Recognising an ‘ene’ disconnection of this sort is not easy but the relationship between the π-bond and the carbonyl group in the product (126, 129, or 131) is a help. Of course the reaction could also be used to make saturated compounds by subsequent hydrogenation and then even that clue is no longer there. What makes this chemistry all the more useful is that an asymmetric version has been devised and used to make a single enantiomer of allokainic acid.23 Initial results were not promising: the menthyl ester 133 gave anti 135 with less than 18% ee. However, the 9-phenylmenthyl ester 134 (see chapter 27) gave much better results. Regioselectivity and anti-stereoselectivity were unimpaired but asymmetric induction depended both on the geometry of the dienophile (E or Z-134) and the conditions. Thermal cycloaddition gave a 50:50 mixture of the two diastereoisomers anti-136 and anti-138 but with Lewis acid catalysis at lower temperatures a sensitivity to the geometry of the dienophile (E or Z-134) could be used to make either diastereoisomer and hence, after hydrolysis, either enantiomer of allokainic acid 132.

822

34 Functionality and Pericyclic Reactions

COCF3 CO2Et N CO2Et

COCF3

COCF3 N

N

CO2Et

CO2Et

CO2Et CO2R*

O R E- or Z-133; R = i-Pr E- or Z-134; R = CMe 2Ph O

Z-134 E-134

CO2Et CO2R*

anti-135; R = i-Pr anti-136; R = CMe 2Ph

anti-137; R = i-Pr anti-138; R = CMe 2Ph

70 ˚C, 80 hours

50:

Et2AlCl, –35 ˚C

95:

:5

Et2AlCl, –35 ˚C

15:

:85

:50

Natural (⫹)-α-allokainic acid 132 comes from anti-136 by two simple steps in 73% yield. The stereochemistry of the final centre is under thermodynamic control. COCF3 N

H N

CO2Et

H N

CO2H

NaOH

pH 3

CO2H

CO2Et

CO2H

H2S H2O

CO2H

CO2H anti-139

CO2R* anti-136

132

Intermolecular ‘Ene’-style Reactions with Oximes Nitrogen can be incorporated as an oxime into a different kind of ‘ene’ reaction that has been explored by Grigg and his group. The ‘ene’ component now bears no resemblance to a diene: one pair of electrons comes from the lone pair on nitrogen and the other from the OH bond of the oxime 140. The enophile is a more conventional enone and the initial product is a nitrone 141. No nitrogen heterocycle is formed in this step, but, if the enophile contains a second alkene, a 1,3-dipolar cycloaddition gives a bicyclic structure. The simplest reagent for this job is the rather unstable divinyl ketone (penta-1,4-dien-3-one, 143). Fortunately this can be released from the dichloroketone 142 with base and distilled with the solvent THF into the reaction mixture.24 R1

O

O

H

N

R

2

R1

H

H

H K2CO3

N

R

2 THF

O

140

Cl

O

141

O 142

O

Cl 143

The oximes are easily made but only the Z-oxime reacts. The temperature must be high enough for the isomers to equilibrate. With alkyl oximes 144 this is easy enough: equilibration occurs at 81 ⬚C in MeCN. The ‘ene’ reaction forms the nitrone 146 under the same conditions. RCHO +

O

H

81 ˚C

N R

R

O

O

H

N

R

N

MeCN

NH2OH

E-144

Z-144

O 145

O 146

34 The ‘Aza-Cope’ Rearrangement

823

The 1,3-dipolar cycloaddition follows with the regiochemistry shown 146a and the product is a bicyclic compound 147 with an N–O single bond that is easily reduced to give a seven-membered cyclic aminoketone 148 with stereochemistry.

R

O

O

O

R

O

N

O

N

R

OH

[H]

HN

N

R

O

O 146

146a

147

148

Oximes of aromatic aldehydes 149 do not equilibrate under these conditions but the Lewis acid catalyst HfCl4 both allows oxime equilibration and reverses the regioselectivity of the 1,3-dipolar cycloaddition 150 so that the final product is a piperidine 152. The regiochemical reversal is probably because the LUMO of the nitrone and the HOMO of the enone are used in 146a but, by lowering the LUMO energy of the enone, HfCl4 switches to the HOMO of the nitrone in 150. OH OH

HfCl4

Ar

OH

N

N

E-149

O

O

HfCl4

N

Ar

66 ˚C THF

Ar

O

O

145 HfCl4

Ar

150

Z-149

O

N

Ar

151

N H 152

PART IV – [3,3] SIGMATROPIC REARRANGEMENTS The ‘Aza-Cope’ Rearrangement The aliphatic Claisen (or oxy-Cope) rearrangement is a familiar [3,3] sigmatropic rearrangement 154, typically requiring temperatures of 200 ⬚C or so. The reaction is driven by the formation of a carbonyl group 155. R

R

[3,3]

heat

O

>200 ˚C

153

R O

O 154

155

The ‘aza-Cope’ rearrangement 157 has no special name and, at first sight, no special utility as it requires at least as high a temperature and the products are imines 158 very easily hydrolysed to the products 155 of the Claisen rearrangement. R

R

[3,3]

heat

R

N 156

>200 ˚C

R

N 157

R

R

H2O

R

N

O

158

155

The reaction is of value only if the products can be trapped with the nitrogen still in place. Stille achieved this by (i) forming the starting materials by forcing the imine to enamine conversion

824

34 Functionality and Pericyclic Reactions

by acylation of the imine 160 followed by reduction 162, (ii) accelerating the aza-Cope step by protonation of the amine 163, and (iii) trapping the imine products 164 by a second reduction 165. The enamine 162 can be isolated in 95% yield if the imine 160 is not isolated on the way.25 Other catalysts also effective in reducing the temperature for the sigmatropic rearrangement from >200 to about 100 ⬚C include TiCl4, Me3Al, and BF3.

CHO

COCl

N

NH2

LiAlH4

N

N

Et3N

O 161; 94% yield

160; 74% yield

159

162; 98% yield

LiAlH4

[3,3]

0.3-0.8 x HCl

N

reflux in dioxane [100-102 ˚C] 10 hours

N

N 0 ˚C

H

H

163

164

H 165; 83% yield

This sequence does not of course produce a nitrogen heterocycle. For that to happen we must have further reactions that make a second C–N bond. One of the earliest and most interesting is a pyridine synthesis26 using a variety of pericyclic steps starting with propargyl amine 167 and a 1,3-diketone 166. O

O

O

H3N

166

167

O

NaOH

195 ˚C

benzene reflux

PhNO2

N H 168; 74% yield

N 169; 50% yield

The stable enaminone 168 undergoes an aza-Cope reaction 170 at nearly 200 ⬚C in nitrobenzene. It may seem odd that a linear alkyne should take part in such a reaction but in fact this is not unusual. The product is an allene 171 that undergoes tautomerism to the more stable enaminone 172 then a [1,5] sigmatropic H shift before cyclising to a dihydropyridine 174. The cyclisation is formally a six electron disrotatory electrocyclic reaction 173. The solvent PhNO2 oxidises 174 to the pyridine 169. O

O

170

•

•

[3,3]

N H

O

O

O H

N 171

H

N H 172

[1,5]H

H

N H

N H 173

174

Though dramatic, this is an unusual example of such processes. Much more has been made of tandem aza-Cope and Mannich reactions, particularly by Overman. A typical sequence begins with the addition of a vinyl-lithium, e.g. 176, to a protected α-aminoketone 177 to give, in this case, the product 178 from Felkin-like attack (chapter 21) on the side of the ketone anti to the amino group.27

34 The ‘Aza-Cope’ Rearrangement

O

O O

825

OH

O

O

Ar 14:1 diastereoisomers CN 67% anti isolated

t-BuLi

+ N

CN H

Br

N

Ph

Li 175

Ph 178

177

176

The cyanide group can be eliminated with silver(I) catalysis to give the starting material 179 for an aza-Cope rearrangement. The product 182, however, has a new heterocyclic ring and has also rearranged. The old five-membered carbocyclic ring is now six-membered. The first step is indeed a [3,3] sigmatropic rearrangement on the immonium ion 179. The product 180 is another immonium salt and is also an enol. Mannich cyclisation 181 gives the product 182. This is the tandem aza-Cope (or azonia-Cope) rearrangement/Mannich ring closure sequence. OH

1.1 x AgNO3

Ar

OH

Ar

OH

O

Ar

[3,3]

=

178 EtOH 2 hours 50 ˚C

H

N

N

N

H

N H

Ph

Ph 182; 94% yield

Ph

Ph 179

Ar

Mannich

180

181

The stereochemistry can be predicted from the way the starting material 179 folds. Transition states for [3,3] sigmatropic rearrangements prefer a chair-like conformation, in this case 184. The dotted bonds in 184 show that minimal change is needed to form the new σ-bond and break the old. The anti relationship between H and OH in 179 should be clear in 184. The chair six-membered ring in 184 leads to the syn relationship between Ar and the same H in 185. H

H

Ar

N R

H

Ar

N OH

R

183

Ar

N OH

OH

R

184

185

Such efficient transmission of stereochemical information suggest that asymmetric versions of this reaction should be possible. Using a single enantiomer of 1-phenylethylamine, a single enantiomer of the aminoketone 186 was prepared and converted into the starting material 187 for the aza-Cope reaction.28 The tandem reaction sequence is the same as that for 178 to 182 but gives a single enantiomer of 188 in 94% yield and ⬎99% ee. O

O

O

OH

Ar

O

Ar

Ag(I)

N

CN Ph

186

+ H Li 176

N

N

CN

Ph 187; 91% yield

H Ph 188; 94% yield

826

34 Functionality and Pericyclic Reactions

There is more to this chemistry than meets the eye. A simpler route to the [3,3] precursor, avoiding the need for stoichiometric silver, is set out in the chart below. Ph

Ph

HO

O

O

Ph

Ph

Ph

HO

O

Li

NH2

Me

NHCO2Et

Me

189; alanine

excess

Me NHCO2Et 191; 80% yield 20:1 diastereoisomers

190

Ph

N

Me

Me 192

Treatment of 192 with acid releases the immonium salt 193 that undergoes the tandem aza-Cope/Mannich sequence. The final product is a single diastereoisomer of the substituted pyrrolidine 195. Ph

Ph

O N

Me

Me

Ph

camphor sulfonic acid

HO

toluene 60 ˚C 1.5 hours

Me

Ph

Ph [3,3]

N

192

HO

O Mannich

Me

Me 193

Ph

N

Ph

Ph N

Me

Me 195 79% yield

Me 194

However, though 195 is a single diastereoisomer, it is completely racemic.29 The intermediate 194 is achiral though it is formed in a chiral conformation 194a that can give only one enantiomer 195a of the product. The problem is that, if the intermediate 194a is long lived enough for bond rotation, it may flip to the enantiomeric conformation 194b that gives the enantiomeric product 195b. O

OH Me

Ph

Ph

Ph Me

N

N

Me 195a

Me

ring flip

Ph H 194a

OH Me N Ph H 194b

Ph

O Ph

Ph Me N Me

Me 195b

Racemisation occurs because the intermediate 194 is achiral and long lived enough for bond rotation. This does not apply to the corresponding intermediate in the previous example 185. This compound is a cyclononadiene containing two E-alkenes. Such compounds are chiral and bond rotation is not possible because of the medium ring. Even if the chiral auxiliary is absent, a single enantiomer of 185 should, and does, rearrange to a single enantiomer of 182.

The Anionic ‘Aza-Cope’ Rearrangement Alternatively, the aza-Cope rearrangement can be greatly accelerated by an oxyanion rather in the style of the Cope rearrangement itself. The N/O acetal 196, easily prepared from cyclohexanone and an amino alcohol, undergoes a rearrangement in strong base to give the spirocyclic pyrrolidine30 198.

34 The Anionic ‘Aza-Cope’ Rearrangement

827

O

O

NH

KH

O

NH

N

18-crown-6

196

197

198

The strong base removes the proton from the nitrogen atom and opens the ring 199 to give the starting material 197 for an anion-accelerated aza-Cope rearrangement at 25 ⬚C in conformation 197a. The product is an enolate, rather than the enols we have been seeing up to now, and it cyclises 200 in Mannich style onto the imine to give the anion of the product 198. O

H2C

N

KH

196

N

[3,3]

N 198

O

O

200

197a

199

If the molecule is too rigid for the Mannich cyclisation to occur easily, it is possible to carry it out as a separate step. Overman’s approach to gelsemine, a very difficult target, involves an anionic aza-Cope rearrangement on 201 and the trapping of the enolate.31 The resulting ketone 202 is cyclised in acidic solution. Evidently the addition of the enol to the protonated imine goes better than the cyclisation of the enolate onto the neutral imine when the final product is as strained as 203. OH

MeO2C 1. KH

HN

R

CN

R

R

O

N

HCO2H

2. ClCO 2Me 3. KOH, MeOH, H 2O

reflux

H

201

N MeO2C

202

203

The first step is treatment of 201 with an excess of KH. Both the OH and the NH lose their protons and cyanide is eliminated 204 to give the starting material for the aza-Cope 205. The product is initially formed in a twisted conformation 206 that opens out to give the simple cis-fused imine enolate 206a. Trapping the enolate with ClCO2Me also converts the imine into the acylated enamine 207 and the enol ester is easily hydrolysed to give the ketone 202. O

O

O [3,3]

1. KH

201 N

R

CN

R CH2 205 R

204 R N

O

2. ClCO 2Me

MeO2C

N

N

N

R 206 OCO2Me

3. KOH

202 MeOH H2O

H 206a

H 207

828

34 Functionality and Pericyclic Reactions

In formic acid, the enamine is protonated on carbon to give the very electrophilic immonium ion 208 which cyclises in conformation onto the enol 208a in an intramolecular Mannich reaction to give the bicyclic skeleton 203 required for gelsemine.

HCO2H

MeO2C

R

R

OH

N

R

OH

=

202

N

N H

MeO2C

MeO2C

208

208a

203

PART V – OTHER REACTIONS This part contains one genuine pericyclic reaction and a collection of near relatives that all give nitrogen heterocycles by the combination of various unsaturated molecules.

Electrocyclic Reactions There are few useful examples of electrocyclic reactions giving nitrogen heterocycles and we shall give just one - a photochemical six electron electrocyclic reaction that closes a ring in a synthesis of lysergic acid 211. The first step is the acylation of the imine 209 and its regioselective conversion into the more conjugated enamine32 210. CO2H O NMe

O N

O COCl

NH Me

H

Et3N

BzN

BzN 210; 96% yield

209

HN 211; Lysergic acid

The electrocyclic step is a photochemical cyclisation of 210 under reducing conditions to give a mixture of diastereoisomers of 212 in about a 7:1 ratio. These were further reduced and the N-benzoyl group replaced to give a separable mixture of amines: the required diastereoisomer 213 was isolated in 61% yield. O N

BzN 210

hν NaBH4

Me

H

O

O

10:1 benzene: MeOH

H

H

N H

H BzN 212; 81% yield

1. LiAlH 4 2. PhCOCl

Me

H

O

O

3. crystallise

H

H

N H

Me

H BzN 213; 61% yield

The electrocyclic step can be drawn on the amide 210a but is perhaps more convincing when drawn on the delocalised version 210b. In either case this is a photochemical six electron electrocyclic reaction and therefore conrotatory. This determines the trans stereochemistry at two of the

34 Ring Closing Olefin Metathesis

829

new centres in the product 214: the ringed hydrogen atoms must both have rotated in the same sense (clockwise or anticlockwise). O

O

O N

N

Me

BzN

O

O hν

Me

O

H

H

BzN

Me

H

BzN

210a

N

six electron conrotatory electrocyclic

214

210b

Under the reaction conditions, the enolate is protonated by methanol 215 to give the more favourable cis ring junction 216 and borohydride reduces the immonium group stereoselectively to give the major diastereoisomer 217 of the final product. OMe H O H

H

N

H

H MeOH

Me

H

BzN

H

O

O

N

H

NaBH4

215

H

Me

N H

H

BzN

H

O

O

O

Me

H

BzN

216

217

Ring Closing Olefin Metathesis In principle ring closing metathesis (RCM, chapter 15) offers a simple route to unsaturated nitrogen heterocycles. The first disconnection is across an alkene inside the ring and normally two alkenes are created by the addition of a CH2 group at each end. Further disconnection might have simple alkylation at nitrogen in mind. A special problems in making nitrogen heterocycles this way might well be complexation of the metal by nitrogen so the nature of R might be important. X

Ring Closing

R

2 x C–N

N

R

N

R

NH2

Metathesis

X

Though any group could be added in theory, the addition of a CH2 group ensures that the other product will be gaseous ethylene so the thermodynamics are favourable. The Grubbs’ catalyst (Cy ⫽ cyclohexyl) is greatly to be preferred because of its stability and wide scope. This is not strictly a pericyclic reaction but in concept it is very like a cycloaddition followed by a reverse cycloaddition. Grubbs catalyst

R

N

R

N

+ CH2

CH2

PCy3 Grubbs Cl catalyst Ru Cl Ph Cy = cycloPCy3 hexyl

830

34 Functionality and Pericyclic Reactions

We shall discuss just the application of RCM to the asymmetric synthesis of pyrrolidine and piperidine derivatives. The asymmetry comes from an enantiomerically pure sulfinyl imine 218 that accepts a lithiated sulfone as nucleophile to give one diastereoisomer of adduct 219; R ⫽ SO2Ph. The chiral auxiliary is removed with acidic methanol and allylation on the nitrogen gave a potential RCM substrate 221. In the event this compound failed to cyclise with the Grubbs’ catalyst so the nitrogen was acetylated to make it a worse ligand. Now RCM was successful and the cyclic amide 223 formed in good yield. The amide can be hydrolysed quite easily to give the free amine 224 as a single enantiomer of the anti diastereoisomer.33 1. LDA 2.

O

O

p-Tol

SO2Ph = R

N

Ar

218 R

N Ar H 219; 85-100% yield

H2N

N

Ar

N

room temperature 2-3 hours

Ac 222; 73-83% yield

Ar

220; 62-88% yield

CH2Cl2

Et3N benzene

K2CO3 DMF

5 mol% Grubbs catalyst

R

MeCOCl

N Ar H 221; 72-79% yield

p-Tol

R

MeOH 0 ˚C

S

S

Br

TFA

R

SO2Ph

R

40% aqueous H2SO4

Ar

MeOH 100 ˚C

R

N

Ac

Ar

H

223; 80-89% yield

224; 60-74% yield

The stereoselectivity in the formation of 219 can be explained by a combination of the imine with the lithiated sulfone 225 in an intramolecular mechanism 225 with a chair transition state 226. The vinyl group goes axial to avoid a clash with the large groups around the sulfone. p-Tol S O

Li

Ar

Ar

N

O

O S

S Ph

O

SO2Ph

N

O

Li O S O p-Tol Ph

225

p-Tol

226

S

N H 219

Ar

The synthesis of pyrrolidines demands first of all the movement of one of the alkenes closer to the nitrogen atom. This can be done by equilibration of the allyl sulfone in 227 to the more stable conjugated vinyl sulfone 228 with catalytic base. Then allylation at nitrogen gave a possible metathesis substrate 229. In this series neither the amine 229, nor the corresponding benzamide would undergo metathesis.

SO2Ph

SO2Ph

cat. KOBu t THF, 0 ˚C

H2N

Ar 227

H2N

Ar

228; 73-85% yield

Br

K2CO3 MeCN

SO2Ph N Ar H 229; 72-83% yield

The problem is the sulfone. The trisubstituted alkene in 229 is too hindered. Reductive removal of the sulfone and acylation gave a good metathesis substrate 231 and hence the pyrrolidine

34 The Pauson-Khand Reaction

831

232. That the product was enantiomerically pure was demonstrated by chiral HPLC of the N-Boc compound and comparison of rotation with the known pyrrolidine 233. SmI2 THF/ HMPA

5 mole% Grubbs catalyst

PhCOCl

229

N H

–20 ˚C

Ar

N

Et3N CH2Cl2

Ar

COPh 231; 75-100% yield

230; 59-65% yield

Ar

N CH2Cl2 room temp.

Ph

N H

COPh

233; >99.9% ee

232 85-94% yield

The Pauson-Khand Reaction In chapter 6 we described the use of the remarkable Pauson-Khand reaction for the synthesis of cyclopentenones. If the components (CO, alkene and alkyne) are tethered by a nitrogen atom, a heterocycle will also be formed. The first stage in this process is to couple the cobalt carbonyl complex, e.g. 236, of a halo-alkyne with an amine containing the alkene in the side chain. The best way to do this is to react 234 with Co2 (CO) 9 to give 235 and then 236 and to capture this complex with the amine without isolation of intermediates.34 Ph

(CO)3 Co

Co(CO)3 Co2(CO)9

(CO)3Co

25 ˚C

(CO)3Co

Cl

0 ˚C

CO

R2NH

NR2

Ph

Ph 236

235

O

(CO)3Co

Cl

Ph

Cl 234

Co(CO)3

O

237

If the secondary amine is an allylic amine, the complex 239 rearranges with the incorporation of a molecule of carbon monoxide (see chapter 6) to give the cyclopentenone 240 with simultaneous formation of a pyrrolidone. This might be easier to see when you realise that 239 is the Co complex of 238. The yields are not wonderful (48–64%) but are reasonable for a reaction that does so much in one step. Co(CO)3

Ph

O

Ph

O

(CO)3Co

O

45 ˚C

N R 238

O

N R

Ph

N

R

240

239

A recent Organic Syntheses procedure35 describes a much higher yielding version using a cobalt complex formed from 241 by treatment with Et3SiH and cyclohexylamine. Only catalytic amounts (5 mol%) are needed to cyclise the sulfonamide 242 to the pyrrolidine 243. Co(CO)3 (CO)3Co

CO, 241

OH 241

Ts

N 242

Ts

N

67 ˚C

243; 86-93% yield

O

832

34 Functionality and Pericyclic Reactions

Metal-Catalysed Alkyne Trimerisation In the same vein but more in the style of Vollhardt’s co-trimerisation is a remarkable series of reactions that combine three alkynes. One is propargyl alcohol and the other two are tethered by a sulfonamide.36 In the simple case 245, trimerisation with the Grubbs’ catalyst (see above) gives one positional isomer 244 of a benzo-pyrroline while Wilkinson’s catalyst [(Ph3P)3RhCl] gives the other possible isomer 246. Me

OH

5 mole% Wilkinson's catalyst

TsN

OH TsN

TsN

25 ˚C toluene

H

244; 90% yield

Me

5 mole% Grubbs' catalyst

Me

OH

OH

40 ˚C CH2Cl2

246; 81% yield

245

Even more remarkably, moving the nitrogen atom in the tether so that it is conjugated with one of the alkenes 248, reduces the yield a little but doesn’t alter the regioselectivity at all. The CH2OH end of the propargyl alcohol reacts with the less hindered end of the amine when Grubbs’ catalyst is used 249 but the other way round with Wilkinson’s catalyst 247. With Grubbs’ catalyst the reaction is presumably a series of metatheses starting on the less hindered alkene. It is not yet understood how Wilkinson’s catalyst reverses this selectivity. Me

OH

N Ts 247; 67% yield

5 mole% Wilkinson's catalyst

OH

25 ˚C toluene

5 mole% Grubbs' catalyst

Me TsN

H 248

OH

40 ˚C CH2Cl2

Me

N Ts 249; 70% yield

OH

References General references are given on page 893 1. General Reference: D. L. Boger and S. M. Weinreb, Hetero Diels-Alder Methodology in Organic Synthesis, Academic Press, San Diego, 1987. 2. H. R. Snyder, R. B. Hasbrouck and J. F. Richardson, J. Am. Chem. Soc., 1939, 61, 3560; H. R. Snyder and J. C. Robinson, J. Am. Chem. Soc., 1941, 63, 3279. 3. A. Alberola, A. M. González, B. González, M. A. Laguna, and F. J. Pulido, Tetrahedron Lett., 1986, 27, 2927. 4. B. Serckx-Poncin, A.-M. Hesbain-Frisque and L. Ghosez, Tetrahedron Lett., 1982, 23, 3261. 5. K. T. Potts, E. B. Walsh, and D. Bhattacharjee, J. Org. Chem., 1987, 52, 2285. 6. D. L. Boger, W. L. Corbett, T. T. Curran, and A. M. Kaspar, J. Am. Chem. Soc., 1991, 113, 1713. 7. F. Sainte, B. Serckx-Poncin, A.-M. Hesbain-Frisque and L. Ghosez, J. Am. Chem. Soc., 1982, 104, 1428. 8. J. Barluenga, M. Thomás, A. Ballasteros and V. Gotor, J. Chem. Soc., Chem. Commun., 1987, 1195. 9. General Reviews: S. M. Weinreb and J. I. Levin, Heterocycles, 1979, 12, 949; S. M. Weinreb and R. R. Staib, Tetrahedron, 1982, 38, 3087; G. R. Heinzelman, I. R. Meigh, Y. R. Mahajan and S. M. Weinreb, Org. React., 2005, 65, 141. 10. G. A. Kraus and K. Neuenschwander, J. Org. Chem., 1981, 46, 4791. 11. S. J. Danishefsky and C. Vogel, J. Org. Chem., 1986, 51, 3915. 12. J. F. Kerwin and S. Danishefsky, Tetrahedron Lett., 1982, 23, 3739.

34 References

833

13. P. D. Bailey, D. J. Londesbrough, T. C. Hancox, J. D. Heffernan and A. B. Holmes, J. Chem. Soc., Chem. Commun., 1994, 2543. 14. General Review: E. Ciganek, Org. React., 1984, 32, 1. 15. Y. Ito, S. Miyata, M. Nakatsuka, and T. Saegusa, J. Am. Chem. Soc., 1981, 103, 5250. 16. R. B. Ruggieri, M. M. Hansen, and C. H. Heathcock, J. Am. Chem. Soc., 1988, 110, 8734. 17. P. A. Grieco and D. T. Parker, J. Org. Chem., 1988, 53, 3658. 18. P. A. Grieco and D. T. Parker, J. Org. Chem., 1988, 53, 3325; P. A. Grieco and A. Bahsas, Tetrahedron Lett., 1988, 29, 5855. 19. W. Oppolzer and W. Fröstl, Helv. Chim. Acta, 1975, 58, 590; W. Oppolzer W. Fröstl and H. P. Weber, Helv. Chim. Acta, 1975, 58, 593. 20. R. K. Boeckman, J. P. Sabatucci, S. W. Goldstein, D. M. Springer and P. F. Jackson, J. Org. Chem., 1986, 51, 3740. 21. I. Fleming, Pericyclic Reactions, Oxford University Press, 1999. 22. W. Oppolzer and V. Snieckus, Angew. Chem., Int. Ed., 1978, 17, 476. 23. W. Oppolzer, C. Robbiani, and K. Bättig, Helv. Chim. Acta, 1980, 63, 2015. 24. I. S. Saba, M. Frederickson, R. Grigg, P. J. Dunn and P. C. Levett, Tetrahedron Lett., 1997, 38, 6099; P. J. Dunn, A. B. Graham, R. Grigg, P. Higginson and I. S. Saba, Chem. Commun., 2000, 2033, P. J. Dunn, A. B. Graham, R. Grigg and P. Higginson, Chem. Commun., 2000, 2035. 25. G. R. Cook and J. R. Stille, J. Org. Chem., 1991, 56, 5578; G. R. Cook, N. S. Barta and J. R. Stille, J. Org. Chem., 1992, 57, 461. 26. K. Berg-Nielsen and L. Skattebøl, Acta Chem. Scand., 1978, B32, 553. 27. L. E. Overman, L. T. Mendelson and E. J. Jacobsen, J. Am. Chem. Soc., 1983, 105, 6629. 28. L. E. Overman and S. Sugai, Helv. Chim. Acta, 1985, 68, 745. 29. E. J. Jacobsen, J. Levin, and L. E. Overman, J. Am. Chem. Soc., 1988, 110, 4329. 30. M. Kakimoto and M. Okanawa, Chem. Lett., 1979, 1171. 31. W. G. Earley, E. J. Jacobsen, G. P. Meier, T. Oh and L. E. Overman, Tetrahedron Lett., 1988, 29, 3781; W. G. Earley, T. Oh and L. E. Overman, Tetrahedron Lett., 1988, 29, 3785. 32. T. Kiguchi, C. Hashimoto, T. Naito and I. Ninomiya, Heterocycles, 1982, 19, 2279. 33. R. Kumareswaran, T. Balasubramanian and A. Hassner, Tetrahedron Lett., 2000, 41, 8157. 34. J. Balsells, A. Moyano, A. Riera and M. A. Pericàs, Org. Lett., 1999, 1, 1981. 35. M. C. Patel, T. Livinghouse and B. L. Pagenkopf, Org. Synth., 2003, 80, 93. 36. B. Witulski, T. Stengel and J. M. Fernández-Hernández, Chem. Commun., 2000, 1965.

35

Synthesis and Chemistry of Azoles and other Heterocycles with Two or more Heteroatoms This chapter considers how aromatic heterocyclic compounds with two or more nitrogen (or other heteroatoms, chiefly O and S) atoms can be made. In particular it deals with the addition of new C-C and C-X bonds to previously prepared heterocycles of this kind.

PART I – INTRODUCTION Azoles: Heterocyclic Compounds with More than One Nitrogen Atom Names for a selection of these heterocycles: mainly 5- and 6-membered aromatic rings PART II – BUILDING THE RING a. The Simplest Disconnections The synthesis of allopurinol The synthesis of thiazoles b. Routes to Isoxazoles The synthesis of isoxazoles from dicarbonyl compounds The synthesis of isoxazoles by 1,3-dipolar cycloaddition c. Tetrazoles The synthesis of tetrazoles by 1,3-dipolar cycloaddition PART III – DISCONNECTIONS OUTSIDE THE RING a. N–C Disconnections: Azole Anions How to get reaction at nitrogen using azole anions The problem of tautomerism in azoles How to get reaction at nitrogen using silyl derivatives in acid solution b. C–X Disconnections Using electrophilic and nucleophilic aromatic substitution in five- and six-membered heterocycles. Chemo- and regioselectivity The synthesis of pyridazines c. C–C Disconnections The few reactions that work well with heterocycles. The Heck reaction, Friedel-Crafts acylation and a mention of lithiation d. Examples: An Anti-Ulcer Drug and Pentostatin An isocytosine-based drug with three heterocyclic rings Pentostatin: an example with fused five- and seven-membered rings A Designed Enzyme Inhibitor

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

836

35 Synthesis and Chemistry of Azoles

PART I – INTRODUCTION Azoles: Heterocyclic Compounds with More than One Nitrogen Atom Most important drug molecules are heterocycles and most of those heterocycles contain more than one heteroatom yet the chemistry of these systems is neglected and even despised by many organic chemists.1 A few examples should convince you that neglect of this area is quite unjustified for any serious organic chemist. The very first synthetic drug, antipyrine 1, is an aromatic pyrazolone with two adjacent nitrogens in a five-membered ring. Perhaps the most famous drug of all, penicillin 2, has a saturated five-membered ring with a nitrogen and a sulfur atom and of course the vital β-lactam used to disrupt bacteria cell wall synthesis. H H N

Me Ph

1 Antipyrine (1887)

O

N

N Me

H S 2 penicillin G (1940)

N

O O

CO2H

Ph

In more modern times, the first successful anti-ulcer drug, cimetidine 3, with its ability to block histamine synthesis in the stomach, is an imidazole and the classical anti-depressant valium 4 has a seven-membered ring heterocycle that is not (quite) aromatic.

3 Cimetidine (Tagamet) (1970s)

N

Me

H N

S

NHMe N

Me N H imidazole

O

N

CN

N

Cl

4 Diazepam (Valium)

Ph

Coming more up to date, Viagra 5, Pfizer’s impotence treatment that hit the headlines at the end of the last century has three heterocyclic rings - a saturated piperazine and two aromatic rings that separately make a pyrimidone and a pyrazole and together imitate a purine. The most successful drug to date, Omeprazole 6, is another anti-ulcer drug but with a unique action that inhibits the proton pump. It has a pyridine ring and a benzimidazole. pyrimidone O EtO

Me

HN

5 Viagra™ (Sildenafil)

OMe

N N pyrazole Me

Me

N Pr O S O

N pyridine

N

O S

N HN

6 Omeprazole (Losec) OMe

piperazine N

Me

benzimidazole

These rings are from four- to seven-membered, but five- and six-membered dominate. They may be saturated, partly saturated, or fully aromatic and they may contain one or more heteroatom. The heteroatoms may be nitrogen, oxygen, or sulfur but nitrogen dominates. In this chapter

35 Azoles: Heterocyclic Compounds with More than One Nitrogen Atom

837

we shall be looking at the synthesis of five- and six-membered rings, mostly aromatic, and mostly containing two or more nitrogen atoms. The chemistry will be very varied. It is worthwhile introducing these ring systems with a brief survey highlighting the main features of the various systems. All are based either on pyrrole 7 or pyridine 8. The distinction between the two types of nitrogen atoms present in these rings is an extremely important part of understanding aromatic heterocyclic chemistry. 7 pyrrole

N H

8 pyridine

N

N

H

H

N

N

sp2 lone pair

p lone pair

Pyrrole 7 has only two π-bonds so it needs the lone pair on the nitrogen atom to complete the aromatic sextet and the lone pair has to be in a p-orbital and delocalised round the ring. Pyridine 8 has three π-bonds so the lone pair on nitrogen is not needed. Besides, it is in an sp2 orbital in the plane of the ring and cannot be delocalised as it is orthogonal to the p-orbitals in the ring. Pyrroles are therefore nucleophilic and react with electrophiles at nitrogen or at carbon while pyridines react with electrophiles only at nitrogen and the ring itself is electrophilic (chapter 32). Adding an extra nitrogen atom (in thought only!) to the pyridine ring system creates three new heterocycles, pyridazine 9, pyrimidine 10, and pyrazine 11, and in all three the extra nitrogen is of the pyridine type with localised lone pairs on both nitrogen atoms. These three heterocycles are more electrophilic and even less nucleophilic than pyridine. 9 pyridazine

10 pyrimidine

N N

11 pyrazine

N

N N

N

Adding more nitrogen atoms to pyrrole gives a series of diazoles 12 and 13, triazoles 14 and 15 and tetrazole 16 but each has two π-bonds and so needs the lone pair of only one of the nitrogen atoms for aromaticity. Each extra nitrogen atom is of the pyridine sort with an sp2 lone pair. These heterocycles become less nucleophilic as more nitrogen atoms are added. In the diagrams below, the pyrrole-like lone pairs are marked in the ring and the pyridine-like lone pairs in sp2 orbitals outside the ring. 12; pyrazole

triazoles

14

13; imidazole N

N

N

15

16; tetrazole

N

N

N

N

N

N

N

N

N

N

N

H

H

H 1,2,3-

H 1,2,4-

H

When sulfur or oxygen atoms are present in a five-membered ring, they provide the pyrrole-like lone pair and any nitrogen atoms are necessarily pyridine-like. The four most important systems are oxazole 17 and thiazole 18 with 1,3-related heteroatoms and isoxazole 19 and isothiazole 20 with an N-O or N-S bond. 17 oxazole

N O

18 thiazole

N S

19 isoxazole

N O

20 isothiazole

N S

With fused heterocyclic rings, the situation can be quite complicated but it is really necessary only to decide whether each nitrogen atom is pyrrole- or pyridine-like. You can count electrons in

838

35 Synthesis and Chemistry of Azoles

each ring or in both rings together, whichever you prefer though in the case of the azaindolizine 21, it is more convincing to count the ten electrons in the periphery as the pyrrole-like nitrogen belongs to both rings. The three unmarked nitrogens in the purine 22 are all pyridine-like. 21 imidazo [1,2-a]pyridine (aza-indolizine)

N

pyridine -like pyrole -like

N N

N

22 purine N

pyrole -like

N H

The rest of the chapter is divided into two parts. We shall first consider how the rings themselves may be formed and then how new bonds (either C-X or C-C) may be built onto a previously formed ring. Both methods are important.

PART II – BUILDING THE RING a. The Simplest Disconnections The first disconnection to be considered is breaking all the C–X bonds in the ring. This is an obvious strategy for a pyrazole since one C–N bond is part of an enamine and the other is an imine 23. Both functional groups are made from amines and carbonyl compounds. Here hydrazine is the diamine that emerges from an easily made 1,3-dicarbonyl compound 24. R2 23 pyrazole

R3

R2 2 x C-N

imine R enamine

R1

N

1

R3 O

O

N H

+

NH2

H2N

hydrazine

24

An example is the pyrazole 25 needed for the synthesis of Viagra. Double C–N disconnection reveals a diketoester 26 with a convenient 1,3-diCO relationship that can be disconnected to a ketone 27 and unenolisable, reactive, and symmetrical diethyl oxalate. Pr 2

2 x C-N

N

EtO2C

EtO2C

N H 25

1

Pr 3

OEt

1,3-diCO

EtO2C

O Claisen ester

O 26

Pr +

O diethyl oxalate

O 27

The Claisen condensation is regioselective because the most stable enolate 28 of the product 26 is formed under the reaction conditions and it is more stable than the alternative 29. Reaction with hydrazine occurs at the two keto groups rather than at the less electrophilic ester. Pr CO2Et CO2Et diethyl oxalate

Pr H2N

27 EtO

EtO2C

O O 28

EtO2C

O O 26

NH2

25 H2O

EtO2C

O O 29

35 The Simplest Disconnections

839

The synthesis of allopurinol Many heterocycles were made by simple procedures of this kind and the syntheses were designed by breaking the molecule into reasonable pieces. Putting these aromatic molecules together generally relied on thermodynamics rather than kinetic selectivity. A good example is allopurinol 30, a treatment for gout that works by inhibiting xanthene 31 oxidase and thus preventing the deposition of painful crystalline uric acid 32 in the joints. Allopurinol imitates purine 22; the main change being the movement of a nitrogen atom in the right hand ring so that a pyrazole replaces the imidazole in purine. O

O

O

HN

H N

HN

N N N H 30; allopurinol

xanthine oxidase

H N

HN

O

N H2N N H 31; xanthine

O

N N H H 32; uric acid

O

Disconnection of hydrazine 30a from allopurinol leaves a diamide with an extra aldehyde group 33. Disconnection of that aldehyde leaves a simple and available symmetrical heterocycle that exists in the dihydroxy tautomer 34 and will require a one-carbon electrophile corresponding to the synthon 35. We shall meet reagents for this synthon soon. O

OH

O

HN

CHO

HN

2 x C-N

N N N H 30a

N

?

N

+ N

O 33

CHO

OH 34

35

No doubt such a synthesis could be carried out. We might also explore the alternative disconnection of the six-membered ring. Opening up the amidine 30b at first looks unpromising but tautomerism between 36 and 37 reveals an amidine and an excellent disconnection to a nitrile 38 from which hydrazine can be removed to give the much simpler 39. Now we must use the same disconnection as 33 to give the same synthon 35 and the even simpler 40. O

O

O

O

C=N

HN

N N 30b O

OHC

N H

H imine

N H

N H

N 36

NH2

OHC

amidine

N H

37 O

O C-N

CHO

+

X

NH2

N 39

NH

OHC N 35 + H

N H

N 38

N H HN

X

O NH

C–N

H

N H

H2N

C-N

O

±H

N 40

N 41

The synthesis is typical of old heterocyclic methods but has a delightful twist at the end. The one carbon electrophile corresponding to ⫹CHO is triethyl orthoformate CH(OEt)3. The other starting materials are very simple: formamide, cyanoacetic acid, and hydrazine and the synthesis is very short. Intermediate 40 is crystalline but need not be isolated as 34 g of cyanoacetic acid can

840

35 Synthesis and Chemistry of Azoles

be converted directly into 15 g of 42. The twist is that the last reaction is carried out by heating the dry yellow crystalline 38 at 150 ⬚C. No visible change occurs, but water is lost and the compound is entirely converted into allopurinol2 30. O

O HCONH2 formamide

O

HC(OEt)3 Ac2O

Ac2O, 70 ˚C

HO cyanoacetic acid

CN

N H

H

40

CN

NH2NH2

150 ˚C

30

38 AcOH

OEt

42

The synthesis of thiazoles The most obvious disconnection (1,1-diX) of a thiazole 43 reveals an acid derivative but the other starting material 44 is an impossible compound requiring NH2 and SH groups to be cis substituted on an alkene. Neither of these functional groups (primary enamine or thioenol) is stable and control of the geometry would also be impossible. The thiol 44 is also the thioenol of an α-amino thioketone 45: these are equally unknown. R2

N

O

1,1-diX

S

R3

R2

HS

R3

H2N

R2

S

R3

+

R1

R1

H2N

C-S and C=N

X

43; thiazole

44; does not exist

45; does not exist

The most productive disconnection is to keep the two heteroatoms in the same starting material, recognising that this can be redrawn as a thioamide. We start by disconnecting the C–N bond 43a of the enamine and then 46 the C–S bond to reveal a simple α-haloketone 48 and a compound 47 that may look strange at first but can be redrawn as a thioamide 49. R2

N

C-N

R1

NH

O

R2

R2

O

NH

C-S

NH2

+ S

R3

enamine

43a

R1

S 46

R1

R3

SH

47

47

R1

R3

X

S 49

48

The synthesis of thiazoles 43 adds an element of regioselectivity as the thioamide has soft (S; orbital controlled) and hard (N; charge controlled) nucleophilic sites and it is essential to differentiate equally clearly between the two electrophilic centres on the carbonyl part of the molecule. We need to know that a thioamide will react with an α-haloketone 48; X ⫽ Cl, Br so that the hard nitrogen atom attacks the ketone and the soft sulfur atom attacks the saturated carbon atom. Thus the simple non-steroidal anti-inflammatory drug Fentiazac 50 is made from thiobenzamide 51 and the simple bromoketone 52, made in turn by bromination of the ketoacid. The synthesis is just that.3

Fentiazac: Analysis Cl N

Cl NH2

C-N

Ph

CO2H

S 50

Ph C-S

S 51

O + Br

CO2H 52

35 Routes to Isoxazoles

841

b. Routes to Isoxazoles The synthesis of isoxazoles from dicarbonyl compounds The disconnection of isoxazoles is easier because there is obviously a molecule of hydroxylamine (NH2OH) waiting to be removed so that isoxazole 53 is ‘obviously’ made from 1,3-diketone 54 and hydroxylamine. That is all very well until we realise that we should also like to make isoxazole 55 from the same two starting materials! R2

R2 C–O enol ether

O

O N

R1

R2

C=N imine

1

N C–O enol ether

O

R

53

C=N imine

O

R1

54

55

Regioselectivity is not so straightforward here. Both N and O are good nucleophiles for carbonyl groups. There has been some success in controlling the reaction by conditions. Thus chemists at Wyeth making ‘glucose lowering agents for obese diabetic mice’ reacted hydroxylamine with the ketoesters 57 in acetic acid at room temperature overnight to give mainly isomer 58 while in concentrated HCl at lower temperatures for shorter times, isomer 56 predominated. It looks as though 56 is the kinetic and 58 the thermodynamic product but the yield of either isomer was poor4 (30–40% at best). O

OH 1. NH 2OH 2. conc HCl

Ar N

CO2Et

Ar

Ar O

room temperature overnight

cold, 1 hour

O

R

1. NH 2OH 2. AcOH

R

56

O

N H

R

57

58

It is better to differentiate the two electrophilic sites more sharply. The bromoenones 59 give some regioselectivity at different pHs. Hydroxylamine is usually added as the crystalline hydrochloride and base is needed to release the free amine. Carbonate does this and with K2CO3 in EtOH, isoxazole 60 is the product. It appears that conjugate addition of the nitrogen atom 61 occurs first.5 R

R

R NH2OH.HCl

Br Ph

O

Ph

O

NH

N

K2CO3/EtOH

59

R

Br

Ph

O

OH

Br Ph

OH

O

60

NH2

61

With the stronger base NaOEt in the same solvent, the isomeric isoxazole 62 is the main product. Presumably the hydroxylamine anion is the nucleophile and now the oxyanion does conjugate addition 64 in the first step. Attractive and rational though this method may be, the average yields of either isomer are only 25–50%. This type of approach relies too much on delicate choices between similar groups and a better way is needed. R Br

NH2OH.HCl

O 59

N

Ph 62

R

Br O

O

NaOEt/EtOH

Ph

R

R

Ph

O 63

NH2

Br Ph

O NH2

O 64

842

35 Synthesis and Chemistry of Azoles

The synthesis of isoxazoles by 1, 3-dipolar cycloaddition And this is it! Alkynes add to nitrile oxides in a concerted π4s ⫹ π2s cycloaddition 65 to give an isoxazole 66 in one step.6 The alkyne provides the two electrons and the nitrile oxide provides four (a lone pair on the oxyanion and one of the π-bonds in the alkyne). The nitrile oxide is an example of a ‘1,3-dipole’ being nucleophilic at one end and electrophilic at the other. It provides the four π-electrons from three atoms. The alkyne is a ‘dipolarophile’. You may guess that nitrile oxides are inherently unstable. In fact they do 1,3-dipolar cycloadditions on themselves to produce strange heterocycles 67 called furoxans. They must therefore be released in the presence of the dipolarophile and the dipolarophile must live up to its name and be better than the nitrile oxide at accepting the 1,3-dipole. R3

R2 R1

N

1,3-dipolar

R

R

1,3-dipolar

N

O

cycloaddition

O

R

R3 R2 N

R1

65

cycloaddition

N

O

R N O

N O 67

O

66

There are two good ways to do this. Dehydration of alkyl nitrocompounds 68, either with PhNCO or with Ph3P and DEAD (EtO2C-N⫽N-CO2Et) in a Mitsunobu elimination gives nitrile oxides 69, as does the 1,3-elimination of HCl from chloro-oximes 70. In the next section we shall show only the nitrile oxide but you should recall that it is generated in the reaction mixture by one of these reactions. H R

H N

O

Et3N

PhNCO

R

N

O

R

or Ph3P DEAD

O 68

Cl2

NOH

R

Cl 69

NH2OH

NOH

R

H

70

O H

71

The problem of regioselectivity remains. Monosubstituted alkynes usually react cleanly using the HOMO of the alkyne and the LUMO of the nitrile oxide. The product is exactly that type of isoxazole (72 or 73) that was so difficult to make from dicarbonyl compounds and hydroxylamine. Here regioselectivity is controlled because the two substituents (R1 and R2) are on different reagents. Conditions are very mild. R1

R1 R2

N

stand at room temperature

O

R2

O

O

N

R2

N

R1

72

R2 stand at room temperature

R1

N

O

73

The disconnection for this reaction is extremely simple 66a - just remove the nitrile oxide. However, if a fully substituted isoxazole is required, the disubstituted alkyne is no longer a satisfactory dipolarophile as it usually does not show enough regioselectivity. R3 R2

R2 R2

66 +

O R1

N 66a

R3

R3

R1

N

O

O N

R1 74

35 Routes to Isoxazoles

843

Better results are often obtained by using, as a latent alkyne, an alkene 77 with a substituent X that can be eliminated and that directs the regioselectivity of the cycloaddition. This means that a preliminary FGI on the isoxazole 75 (alkene to alkyl X) is required before the 1,3-dipolar disconnection 76. R

Analysis:

R

R

X

R

FGI

O N

R

R

1,3-dipolar

O R

75

R

+

R

N

O

X

cycloaddition

N 76

77

If X is electron-withdrawing, and the best example is X ⫽ Cl, the LUMO of the dipolarophile reacts with the HOMO of the nitrile oxide, i.e. O⫺ attacks the electrophilic end of the alkene, and the intermediate isoxazoline eliminates HCl 78 (not necessarily by an E2 reaction) to give the aromatic isoxazole 79.

Synthesis:

R

N

R

R

R

Cl

H

Cl

O

O

O

N

R

N

R 79

78

If on the other hand the substituent X is electron-donating, and various examples such as OAc, NR2, OSiMe3, or SR can be used, the regioselectivity is reversed with the HOMO of the dipolarophile attacking the LUMO of the nitrile oxide at the carbon end of the dipole. Elimination 80 gives the isomeric isoxazole 81. Synthesis: X = OAc, NR 2 OSiMe3, SR

R

X

X

R N

R H

O

R O

O

N 80

R

N

R 81

The three approaches can be compared with the same isoxazoles. 1,3-Dipolar cycloaddition with the alkyne 82 gives a mixture of regioisomers 83 and 84. If the chloroalkene 85 is used instead, only 84 is formed while the enamine 86 gives 83 only.7 O

Ph

N

O

H

82

83; 12% yield

Ph

O

O Ph

85

Ph

N

O

Ph

Cl O

86

Ph

NMe2

N

O

O Ph

Ph

84; 70% yield O Ph

O

83 only

N

Ph

N

84 only

844

35 Synthesis and Chemistry of Azoles

c. Tetrazoles The synthesis of tetrazoles by 1, 3-dipolar cycloaddition One group of heterocycles made almost exclusively by 1,3-dipolar cycloadditions is the tetrazoles 87. They have four nitrogen atoms in the five-membered ring and therefore only one carbon atom and, inevitably, no C–C bonds. Tetrazoles are as acidic as carboxylic acids (pKa about 5) and disconnection of the anion 88 into a nitrile and azide ion is the usual approach.

Tetrazoles: Analysis

R

R

R

N

N

FGI

87

N

N

N N

N

1,3-dipolar

N

N H

cycloaddition

N

N

N

88

There are many ways to make the reaction work well, one of the most popular using sodium azide buffered with ammonium chloride with LiCl in DMF. The product is the tetrazole anion 88 and the tetrazole 87 is released on acidification. The example 91 shows that the yields are excellent and that functionality can be included in the side chain.8 NaN3, LiCl

acid

RCN

88 NH4Cl, DMF, 100 ˚C

1. NaN 3, LiCl, NH 4Cl

MeO

CN

MeO

MeO

N N

DMF, 100 ˚C

89

87 work-up

N

2. acid work-up

N

N

N

N H 91; 100% yield

N

90

That the reaction occurs as expected between the HOMO of azide (which is, after all, an anion) and the LUMO of the electrophilic nitrile is confirmed by reactions with aryl nitriles ArCN with sodium azide. The yields are all close to quantitative but the reaction is faster when R is an electron-withdrawing group. Not everyone agrees that the reaction is concerted as the approach of the two linear molecules looks very hindered. However, a stepwise addition of azide ion 92 followed by cyclisation 93 looks, if anything, worse as two negatively charged atoms must attack each other in the cyclisation 93. However, this too is pericyclic (electrocyclic) and not ionic. R one step mechanism

N N

N N

86

two step mechanism R N N N N 92

R

N

N

N

N

86

93

This disconnection should of course be followed by others as the need arises. Thus the antihypertensive zolterine 94 containing a second heterocyclic ring - a piperazine - can be made from nitrile 95 and that comes from conjugate addition of the piperazine to acrylonitrile.

35 Tetrazoles

845

Zolterine (anti-hypertensive) Analysis: Ph

N

Ph

H N

N N 94; Zolterine

Ph

N

1,3-dipolar

N

N N

N

1,3-diX

NH

CN

95

+

CN

96

This synthesis was carried out using the method we have described and is short and unambiguous.9 Zolterine (anti-hypertensive) Synthesis: Ph

CN

N

Ph

N

NaN3, LiCl

NH

H N

N

95 NH4Cl, DMF, 100 ˚C

N N 94; Zolterine

96

N

If there is functionality directly attached to the carbon atom of the tetrazole, a problem may seem to arise. Thus many drugs are made from 5-aminotetrazole 97 and our disconnection requires the unlikely looking molecule H2N-CN. This is in fact ‘cyanamide’ available as a mixture with water and ‘stabilisers’ to prevent dimerisation to ‘dicyandiamide’ 98. The dimer is 50 times cheaper and can be used in the 1,3-dipolar cycloaddition under slightly different conditions.10 Many Tetrazole Drugs Made from 5-Aminotetrazole Analysis: Synthesis NH H NH2 NH2 N H2N CN N 1,3-dipolar H2N N N N H N N N N N 98; dicyancyanamide 97 diamide

HCl added slowly to NaN3 water

H N

H2N

N N 97; crystallises as stable hydrate N

A good example of a drug made from aminotetrazole is Merrell Dow’s anti-allergic agent MSD 427. You will notice that the drug 99 is actually the stable anion of the tetrazole. It makes sense to disconnect the aminotetrazole immediately to reveal the N-formyl derivative 100 of o-aminobenzoic acid (anthranilic acid) 101.11 Analysis: O

N N

N 99; MSD 427

N

O

O

N N

OR

2 x C–N

N H 100

C–N

OH

CHO

N

N

[ CHO] H2N

NH2 101

N

102

N H 97

Another fragment is a one carbon electrophile we have shown as ‘⫹CHO’. You will appreciate that the more heteroatoms there are in the ring, the more necessary it is to use small carbon fragments to make the rings. All that remains is to identify a reagent for 102 and to decide on the order of events. The reagent chosen was triethyl orthoformate HC(OEt)3 and the synthesis was simplicity itself.

846

35 Synthesis and Chemistry of Azoles

Synthesis

O O

HC(OEt)3

N

97 H

N H 103

O

N OMe

N + N H

OMe N N H

NH2 104

N

N

base

N

99

N H

105

PART III – DISCONNECTIONS OUTSIDE THE RING a. N–C Disconnections: Azole Anions Though tetrazole is the most dramatic example of a stable anion from an azole, all the azoles show enhanced acidity when compared with simple cyclic amines.12 Pyrrole (pKaH ⫺3.4) is a much weaker base than pyrrolidine (pKaH ⫹11.3) and it is protonated at carbon 107, not at nitrogen, to give the dearomatised ion 106 since the lone pair is delocalised round the ring. As an acid 108, it is much stronger than simple amines since the anion 109 is aromatic, like the cyclopentadienyl anion, but more stable because of the electronegative nitrogen atom. Pyrrole is about as strong an acid as a tertiary alcohol, so bases that you might use to make t-BuO⫺, such as NaH, are needed to make the anion. H

pKa as acid 16.5

pKa as base –3.4

H

N

H

N

N

H

H

H

106

107

108

N

B

109

Every additional nitrogen atom must be of the pyridine sort and this has the curious effect of increasing both the acidity and the basicity. There is another difference. Protonation now occurs on the pyridine-like nitrogen atom but both nitrogen atoms cooperate to deliver electrons in both pyrazole 111 and imidazole 112. In this process the pyridine-like nitrogen atom becomes pyrrole-like and vice versa but this is merely formal as the cations 110 and 113 are symmetrically delocalised and the two nitrogen atoms are equivalent. N N N

pKaH as base

N

H

H

pKaH as base

N

H 111; pyrazole

110

N

H

N

2.5

H

H N

7.0

H 112; imidazole

H 113

The action of pyrazole and imidazole as acids is simpler. In each case the anion (114 or 117) is aromatic, symmetrical, and more stable than the pyrrole anion because of the extra nitrogen atom. The two ‘diazoles’ pyrazole and imidazole are almost equal in acid strength. pKa as acid

N N

N

N N

14.2

N H

114a

114b

115

N B

H 116

N

pKa as acid

N

14.4

N

N N

B 117a

117b

As we add (this is a thought process and not a chemical reaction) more nitrogen atoms the trend to greater acidity continues. The two triazoles 119 and 120 are stronger acids and, as we have seen,

35 N–C Disconnections: Azole Anions

847

tetrazole 122 is about as strongly acidic as carboxylic acids. All this information seems rather dull but it is essential to have a grasp of the trends when planning to make bonds to the nitrogen atoms of the azoles. pKa as acid

N

1,2,3-triazole 1,2,4-triazole N

N

pKa as acid

N

N

N

N

H 119

H 120

9.4

118

N

N

N

tetrazole N

N

N

N

N

N

N

121

H 122

10.3

pKa as acid

N

N N

5.5

123

How to get reaction at nitrogen using azole anions In a nutshell: if you want to make a bond to a nitrogen atom, use the anion. The simplest examples are tetrazoles. The anti-inflammatory broperamole is the piperidine amide of the acid 124 that has a 1,3-relationship that can be disconnected first to reveal a simple tetrazole 125 available by 1,3-dipolar cycloaddition. N

N

N

N N

1,3-diX

CO2H

N N

+

+ NaN3

NH

CN

1,3-dipolar

CO2H Br 126

Br

Br

125

124

The tetrazole synthesis was done with azide and acetic acid and base-catalysed conjugate addition followed by treatment with SOCl2, to make the acid chloride, and piperidine completed the synthesis.13 N

N

N

NH

CN N

NaN3

N N N

CO2H

HOAc CO2H

Br

Br

Br 124

125

126

The indole-based drug 129 is clearly made by acylating an indole 127 and this was the last step in the synthesis, The problem here is chemoselectivity. The answer is the usual one of ‘last in, first out’. The anion of the indole will be needed to ensure reaction at nitrogen but the anion of the tetrazole will be formed first. Two molecules of NaH make the dianion 128 and the less stable anion, that of the indole, reacts first.14 N N HN

N

N

NaH

N N

DMF

N

N

ArCOCl

N N

HN

N O

H 127

128

Cl

129

N N

848

35 Synthesis and Chemistry of Azoles

In many cases where azole anions are used, problems of regioselectivity arise too. There are many fungicidal drugs based on triazoles, usually 1,2,4-triazoles, and the anions of these compounds may have two different sites to react. The ketone 131 is made by alkylation of the triazole anion with the primary alkyl chloride15 130 and is the basis for many anti-fungal acetals 132. OR N

O

N

O

N N H

Cl

N N

N

OR

O HO

O N N

OH

base

Cl

Cl

Cl

Cl

Cl

131

130

Cl 132; antifungal drugs

The anion of 1,2,4-triazole 133 might be alkylated elsewhere. Reaction at N-1 or N-2 leads to the same product 134 but reaction at N-4 gives symmetrical 135 instead. It is not always easy to predict where reaction will occur but reaction at N-1 or N-2 is usually wanted and there is at least a 2:1 statistical factor in favour of this regioselectivity. We shall explore this question in the next section. N 2N 1 N

R

N

4 RX

at N-1

N

at N-2

N

R N

R 134

133

N

N at N-4 N N

N

134

135

The problem of tautomerism in azoles Any azole is a tautomeric mixture. That is not always obvious as, say, the two tautomers of imidazole itself are the same - they merely exchange protons in the way that a carboxylic acid does and this has no effect on reactions or synthesis.

H

N

N

N

N

H H

O

O

O

O

H

R R tautomerism of a carboxylic acid

tautomerism of imidazole

But as soon as there is substitution (anywhere except at the carbon atom between the two nitrogens 136) the two tautomers become different and the sample contains both in equilibrium. Thus nitration of 136 gives a tautomeric mixture of 137 and 138. O2N

O2N

NO2

N

N R 136a

H

H

N R 136b

N

H

N R 137

N

N

N

H

R 138

There is no point in trying to separate these tautomers as each would rapidly equilibrate with the other. If the next reaction is intended to occur at one of the nitrogen atoms and so the anion is to be used, then all is well as both tautomers 137 and 138 give the same anion 139. To be precise, the anion 139 is delocalised (139a and 139b are just two different ways to draw the same anion) and protonation at either nitrogen atom gives one of the tautomers 137 or 138. With such an

35 N–C Disconnections: Azole Anions

849

unsymmetrical anion, reasonable regioselectivity is to be expected, and the product on reaction of the anion 139 with an electrophile is isomer 140 rather than the other 141. tautomers O2N base

N

N

H

delocalised forms O 2N N

N

O 2N

O2N R

137

R

139a

E

N

E O2N

N

N

N

N

N

H

138 R

E

R 141 not formed

R 140

base

N

N

O 2N

R

139b

Why should that isomer 140 be formed? The real reason is that the largest coefficient in the HOMO of the anion is at that nitrogen atom. This is not very helpful as it is not easy to work out quickly the distribution of coefficients in such molecules. Fortunately there is an easy way to find the answer. Look for the isomer of the product with the longest conjugated system between the pyrrole-like nitrogen and the functional group. Here the answer is clear-cut. The isomer formed 140 has conjugation from N lone pair to nitro group all round the ring as in 140a/b while the other isomer 141 has that conjugation only round part of the ring 141a/b. O O

O

N

E

O N

N

E

R 140a

O

N

O N

N

O N

N

R 140b

O

N

N N

N

E

R 141a

E

R 141b

This chemistry is used in the synthesis of metronidazole16 (Flagyl®) 143 a drug for the treatment of parasitic infections of the urinary tract. The substituent R is just methyl and the electrophile is ethylene oxide. The nitro group makes the imidazole NH in 137/8 even more acidic and NaOH is a strong enough base for the N-alkylation. The yield of 143 shows that regioselectivity is complete. O2N

O2N

O2N

O

NO2

N

HN

HN

N

N

NH NaOH

Me 142

Me 137; R = Me

Me 138; R = Me

HO

N

N

Me 143; 100% yield

How to get reaction at nitrogen using silyl derivatives in acid solution Not all reactions can be carried out in basic solution so we also need a method for adding electrophiles to heterocyclic nitrogen atoms in acidic solution. This is an important problem in nucleoside synthesis,

850

35 Synthesis and Chemistry of Azoles

especially when modified nucleosides are needed for anti-viral compounds. A nucleoside consists of a ‘base’, a purine or a pyrimidine, joined through a nitrogen atom to a sugar. If the sugar is ribose we have an RNA-style nucleoside 145 and if deoxyribose, DNA-style 144. O Me 144 a nucleoside from DNA

thymine– a pyrimidine

adenine– a purine N

NH N

NH2

O

N

N

O

145 a nucleoside from RNA

N

O

HO

HO deoxyribose

HO

HO

ribose OH

The obvious disconnection to make in any nucleoside is at the very strategic C–N bond between the two rings. Here is a proposed synthesis of cytidine 146. It is also obvious to make the nitrogen a nucleophile 147 and the sugar the electrophile 149 by the addition of a leaving group 150. The oxygen atom in the ring makes a cation at that position 149 rather favourable. Cytidine Analysis:

Synthons:

Reagents:

NH2

NH2

NH2 N

N

N N

1,1-diX

N

O 147

O

O

C–N

HO

O 148

O

X

O HO

HO

N Y HO

OH

OH

HO

146; cytidine

OH

HO

149

150

The leaving group on the sugar 150 could be a halide or acetate or even an alkoxide as we then would have an acetal. But these compounds will react in acidic rather than basic solution and so the problem is the structure of 148: what can be Y? The answer is an SiMe3 group, attached to the oxygen rather than the nitrogen and we have a famous silicon-directed nucleoside synthesis, the Vorbrüggen coupling.17 The silyl group can be added in a number of ways and it will be important to protect all the OH groups. Cytidine Synthesis:

NH2 N

O

OH

OH HO 153; ribose

NHSiMe3

N

N

(Me3Si)2NAc

N

N O H151 HO

NHSiMe3

OSiMe3

N

152 O BzO BzO 154

O OAc

OBz

BzO OBz BzO 155 protected cytidine

O

35 N–C Disconnections: Azole Anions

851

You may be surprised by the stereoselectivity of this coupling. The benzoate on the 2’ position actually participates 157 after the acetate is lost 156 so that the silylated pyrimidine has to attack from the top face 159. Silylated cytosine reacts rather like a silyl enol ether. This type of stereochemical control applies only to ribonucleosides and is more difficult to achieve without the 2’ OH group. O O

O

O

BzO

O

BzO O

BzO 156

O

O

BzO 157

Ph NHSiMe3

O

158 NHSiMe3 N

N SiMe3

N O

N

O

O

O

O O

BzO Ph

OAc O

BzO

Ph

O

O

NHSiMe3

N N

O

BzO

BzO

BzO Ph

BzO

BzO 160

159

OBz BzO 155 protected cytidine

OBz

One synthesis of the important anti-viral drug AZT 163 (Azidothymidine, used in the treatment of HIV) shows this. The doubly silylated thymine 161 combines with the azido-sugar 162 under Lewis acid catalysis to give AZT after work-up. Only 33% of the required diastereoisomer 163 can be isolated and about the same amount of the other is formed.18 OSiMe3 161 silylated thymidine

O

N N

NH OSiMe3

O TBDMSO

t-BuMe2SiOTf

N O

OMe

HO

O

163 AZT AZido Thymidine

162 N3

N3

Other synthetic anti-viral compounds, such as carbovir 169, lack the oxygen atom in the ring so that a route based on an SN1 style stereoselective displacement appears impossible. Carbovir has a double bond in the ring however so that an η3 palladium allyl cation complex allows regio- and stereo-selective displacement of an ester leaving group (chapter 19). The palladium attacks the alkene 164 from the opposite side to the leaving group to give 165. The nucleophile then comes in from the opposite side to the palladium and at the end of the allyl cation farthest from the other substituent to give 166 with overall retention of configuration. OCO2Et BnO 164

Pd(PPh3)4

Nu

NuH

BnO

BnO 165

PdL3

166

852

35 Synthesis and Chemistry of Azoles

The starting material 164 can be made as a single enantiomer by the use of an oxazolidinone chiral auxiliary (chapter 27). Treatment with the chloropurine 167 catalysed by Pd(0) gave only the correct stereoisomer 168 in 63% yield - not a high yield but better than 33% - and deprotection and hydrolysis gave the guanine-derived carbovir19 169. Cl

Cl N

N

N

Pd(PPh3)4

N

N

N

1. BCl 3

OBn N

165

N H

O

NH2

167

N

NH2

NH

OH N

2. NaOH

NH2

N

169 carbovir

168

b. C–X Disconnections Using electrophilic and nucleophilic aromatic substitution in five- and six-membered heterocycles. Chemo- and regioselectivity Five-membered rings with two or more heteroatoms are usually good at electrophilic substitution as one of the heteroatoms must be either O or S or a pyrrole-like nitrogen atom any of which supply lone pair electrons. Pyrazole 12, imidazole 13, oxazole 17, thiazole 18, isoxazole 19 and isothiazole 20 are examples. Multiple substitution can be a problem but is less so than for pyrrole because there are fewer carbon atoms available and the pyridine-like nitrogen atoms deactivate the ring. 12; pyrazole

13; imidazole

17; oxazole

18; thiazole

N

N

19; isoxazole

20; isothiazole

N N

N N H

O

N H

N

O

S

S

An example20 is the antibiotic 171. The nitro group on the furan ring of 170 was originally put in by nitration but now bromination is carried out in acetic acid. Furan is fundamentally more reactive towards bromine than the thiazole ring. However, the furan is deactivated by the nitro group while the thiazole is activated by the amino group. In fact bromination occurs exclusively at the one remaining position on the thiazole. Br Br2

O2N

S

O N 170

NH2

HOAc

O2N

S

O N 171; 86% yield

NH2

Six-membered rings with two or three nitrogens (they must all be pyridine-like) are very unreactive towards electrophilic attack but good at nucleophilic attack if there is any remotely feasible leaving group, even OR or better SR. The main ring systems are pyridazine, pyrimidine, pyrazine and triazine and the places where nucleophilic substitution occurs best are marked with an ‘X’. We shall see examples of these reactions in the next section.

35 C–X Disconnections

853

X X

X X

N

N N X pyridazine (X = H)

N X pyrimidine (X = H)

X

N

N

N X X pyrazine (X = H)

N

X N X triazine (X = H)

When there is a nitrogen atom common to two rings, one ring may be good at electrophilic and the other at nucleophilic substitution. In this example known as imidazo[1,5-b]pyridazine 172, bromination occurs in the five-membered ring to give 173, while nucleophilic substitution occurs if there are leaving groups X in the six-membered ring 174. It is inevitable that the nitrogen common to both rings is of the pyrrole type and these systems are probably best considered as 10 electron aromatic compounds.

Br2

N

N

N

X

N

N

N

N X

Br 172

N

N

174; X = Cl, SMe, OMe

173

The synthesis of pyridazines The ‘most obvious’ disconnection depends on which Kekulé form of the pyridazine you happen to draw. Disconnection of both C–N bonds reveals either 175 an encouraging 1,4-diketone 176 and a most discouraging molecule of diimide 177 or, by imine disconnection, 178 an encouraging molecule of available hydrazine and a rather discouraging Z-enedione 179. R2 R1

N

R2

N

R1

175

O

O

176

R2

NH HN 177 diimide

R1

N

N

R2 R1

O

178

O

H2N

NH2

hydrazine

179

There is one useful case of the second disconnection. Maleic anhydride 180 has to have a Z alkene because of its cyclic structure. Reaction with hydrazine gives a diamide 181 that prefers to exist as 182 and is known as ‘maleic hydrazide’. OH

O O O

O

180 maleic anhydride

H2N

NH2

H2O

O

N H 181

NH

O

N H

N

182; 85% yield

This product is useful because it gives the dichloride 182 with POCl3. This compound reacts cleanly once with ammonia to give 184 and then again with any nucleophile, including amines under more vigorous conditions, to give 185; Nu ⫽ NHR, OR, SR etc.21

854

35 Synthesis and Chemistry of Azoles

OH

Cl

N N H 182

O

N

N

Cl

NH2

NH2

NH3

POCl3

Nu

EtOH

Cl

N

N

Nu

184; 70% yield

183

N

N 185

A more general approach to pyrazines is based on an FGA strategy. Removal of the alkene also removes all the problems. Disconnection 186 of hydrazine gives the same simple 1,4-diketone 177 we saw before and the most obvious ways to make that 187 are the d1 ⫹ a3 or d2 ⫹ a2 strategies, both requiring some umpolung. R2 R1

N 175

N

R2

FGA oxidation

R1

N 186

N

imines

a R2 R1

O

b

O

d1

R1

O 187

a3

O 176

O

NH2

H2N X

R2

a

b R1

R2

2 x C=N

R2

O

R1

O

O

d2

a2

A simple example is the synthesis of Cyanamid’s broad leaf herbicide 193 by the d1 ⫹ a3 strategy. m-Fluorobenzaldehyde combines with morpholine and cyanide to give the aminonitrile 188. The marked proton is acidic because of the cyanide group. The anion is good at conjugate addition (these d1 reagents were specifically designed by Stork for conjugate addition, chapter 9) and adds well to the methacrylate ester to give the 1,4-dicarbonyl compound 189. Cyclisation with hydrazine and aromatisation with bromine gives the pyridazolone 191 and nucleophilic aromatic substitution by the POCl3 and nucleophile (here methoxide) sequence gives the herbicide 193 in good yield.22 1.

O

HN

ArCHO

O NC

2. KCN, H 2O

N

Ar H N

CO2R

O

N

N

NH2NH2

H 188

NaOMe 2. HOAc

Ar

EtOH, H 2O

189

O

N POCl3

CF3

1. CO2R

N

O Br2

Ar

HOAc

190 Cl

N MeO

CF3

H N

N

OMe

CF3

MeOH

191

192

193

c. C–C Disconnections The few reactions that work well with heterocycles. The Heck reaction, Friedel-Crafts acylation and a mention of lithiation The ever-versatile Heck reaction (chapter 18) works well with most heterocycles providing one can put the halide or triflate in the right place on the heterocyclic ring. This example shows that an unprotected nucleoside 194, made by iodination of deoxyuridine, reacts cleanly with ethyl acrylate

35 Examples: an Anti-Ulcer Drug and Pentostatin

855

in a typical Heck reaction without any protection of the OH or NH groups in the nucleoside.23 The reactions of nucleosides we have used so far have all required complete protection. O I

O EtO2C

NH N

NH

EtO2C

O

N cat. Pd(OAc) 2 Ph3P, Et 3N dioxan 85 ˚C, 3 hours

HO

O

HO

HO

HO 195; 80% yield

194

The Heck reaction makes a C–C bond and adds a highly functionalised fragment that can be elaborated into many other functional groups. The same is true of the Friedel-Crafts reaction that generally works well with azoles or other heterocycles able to do electrophilic substitution. As usual, it is best to have only one free position so that no regioselectivity problems arise. In the Heck reaction, the site of attack is marked by the iodine atom, but the Friedel-Crafts can occur at any free position. Our example is a pyrazole that acylates cleanly with Lewis acid catalysts to give eventually the herbicide pyrazolate 196. Cl O Cl

Me

O

Friedel-Crafts

TsO

Me

Cl C–C

+ TsO

Cl

N

Cl

N

Me 196; pyrazolate (herbicide)

N N Me

197

198

It turns out that it is better to acylate the pyrazolone in a standard Friedel-Crafts fashion to give 201. Whether you draw the pyrazole 200 as a pyrazolone 199 or not, there are plenty of electrons from one of the nitrogen atoms to activate the remaining free position in the ring. Subsequent tosylation fixes the molecule in pyrazole form as pyrazolate.

Me

O

Me

Me

Ar 197

O

NH N Me 199

HO

TsCl

N N Me 200

AlCl3

HO

N N

196 pyrazolate (herbicide)

Me 201

d. Examples: an Anti-Ulcer Drug and Pentostatin We give two examples: an anti-ulcer drug with three heterocyclic rings and an anti-viral and antitumour compound, pentostatin.

856

35 Synthesis and Chemistry of Azoles

An isocytosine-based drug with three heterocyclic rings A series of anti-ulcer drugs such as cimetidine and ranitidine were the most successful drugs of all time when they first appeared. They work by inhibiting histamine action in the stomach. We shall look at a more modern version, a drug from what was then SmithKline Beecham, another H2 receptor histamine antagonist24 202. It is particularly appropriate for this chapter as it contains three heterocyclic rings with two heteroatoms each: an imidazole, a pyrimidone, and a thiazole and it allows us to revise material from other chapters as well as this.

O H N

202 former SKB histamine receptor H2 antagonist

Me S

N

S

N

imidazole N H

thiazole N

N H pyrimidone

Strategically it is best to disconnect somewhere in the middle of the molecule and the ease of nucleophilic substitution on pyrimidones suggests the C–N bond exo to that ring 202a. This gives us a relatively simple imidazole fragment 203 and a fragment with the other two rings 204. The leaving group on the pyrimidone 204 (X) could be Cl but need only be as good as OMe. O

Analysis H Me N N

S

N S

N H 202a

N

N H

C–N nucleophilic aromatic substitution

O

H N

Me S

N

NH2 +

S

N N

N H

X 203

204

The imidazole 203 can easily be disconnected 203a back to the simpler imidazole 205 and H2NCH2CH2SH. You may forecast selectivity problems in the combination of unactivated 205 and the unprotected aminothiol, but wait. Alcohol 205 can be made from the ester 206 and that can be made from formamidine 207 and a suitably activated acetoacetate 208 (X is again a leaving group). H N N

Me

NH2 S

203a

C–S

H N

Me

FGI

OH

N 205

H N

Me

2x C–N

+

H N

CO2Et 206

O

NH2 NH 207

X

CO2Et 208

The remaining fragment can be disconnected 204a at the six-membered ring by removal of some simple amidine derivative 209. The remainder 210 has a 1,3-diCO relationship so a Claisen ester disconnection gives yet again a one-carbon electrophile and a simple ester 211. This is best made from the unsaturated ester 212 and some sort of aldol reaction on 213.

35 Examples: an Anti-Ulcer Drug and Pentostatin

O

O S

N

NH

2 x C–N

N

N H

X

857

NH2

X

204a

S

+ RO

209

EtO2C

S

CHO +

210

EtO2C

FGI

N

H

O

1,3-diCO

S

N

OHC

aldol

N

211

S N 213

212

Now we must put the molecule together again. 2-Bromothiazole is available so lithiation and carbonylation with DMF gives 213 and an aldol (Knoevenagel) reaction with malonic acid gives 214 without a separate decarboxylation step. The best one-carbon electrophile is ethyl formate (HCO2Et) and thiourea makes a suitable derivative of 209 for displacement. Br

S

1. BuLi

N

CH2(CO2H)2

213 61% yield

2. Me 2N.CHO

HO2C

S

H2/Pd/C

EtOH

212

2-bromothiazole

H2SO4

N

pyridine piperidine

214 O

O

NH2

EtO2C

S

HCO2Et

S

Na, Et 2O

211 MeOH 45 ˚C

N

O

S

HN

NH2

S

EtOH reflux

215

N H

N

MeI NaOH EtOH H2O

S

N N H

MeS

216

N 217

The other half starts with formation of the imidazole by the Bredereck reaction. The leaving group in 208 can be an acetate but more interestingly the nucleophile can be an excess of formamide HCONH2. The alcohol 205 reacts selectively in strongly acidic solution with the aminothiol to give 218 without protection. After isolation as the HBr salt, 203 is liberated and reacted with 217 to give the complete molecule.

O

H N

excess HCONH2

CO2Et

208; X = OAc

HS

LiAlH4

206 AcO

NH2

Me OH

N 205

48% HBr

H N

Me 2Br

NH3

pH 11

N H

S

217

203

202

K2CO3 H2O

218; 98% yield as crystalline salt

Pentostatin: an example with fused five- and seven-membered rings Pentostatin 219 is clearly a modified deoxynucleoside with a seven-membered ring instead of the normal six-membered ring of a purine.25 It is simplest to remove the stereochemistry in the sevenmembered ring by FGI of the alcohol to a ketone 220. We can then disconnect the sugar, expecting to use the sort of chemistry earlier in the chapter, maybe a silyl derivative of the purine analogue 221 and a derivative of the sugar 222, maybe with X ⫽ OAc.

858

35 Synthesis and Chemistry of Azoles

OH HN

HN

N N

O

O

FGI

N

N

N N

HN 1,1-diO

N

O

O

HO

N H

+

C–N

HO

HO 219; pentostatin

N

O

X

HO

HO

HO

220

222

221

The sugar is available deoxyribose but the seven-membered ring heterocycle 221 must be made. Disconnection of the one atom electrophile from between the two nitrogen atoms reveals a diamine 223 that can be made by reduction of a dinitro compound 224. Removal of nitromethane from 224 gives a simple nitroimidazole carboxylic acid 225. Clearly the nitro group will be added by direct nitration but we can take advantage of its presence to reconnect an alkene 226 to the carboxylic acid with the idea of using the nitro group to activate the methyl group in 227. Nitration of 4-methylimidazole would be expected to go where we want it as the only alternative, C-3, is between the two nitrogen atoms and less reactive to electrophiles. O

O

O N

N

1,1-diO

N H

N

O2N

2 x C–N

N H

H2N

221a HO2C

N N H

O2N

224

R

reconnect

N H

O2N

223

1,3-diO

C–C

N

FGI

H2N

HN

N O 2N

225

Me

N

O2N

N H

C–C

N H

227

226

The synthesis starts with the routine nitration of available 228 and the base-catalysed reaction of 227 with benzaldehyde to give 229 (226; R⫽Ph). It is necessary to protect this and benzylation in base gives predominantly the isomer 230 with the longest conjugated system. Ozonolysis (ozonation) reveals the stable crystalline carboxylic acid 232. After deprotection, both isomers give 225.

Me

N

N H 228 (available) Ph

Me

N

HNO3 H2SO4

Ph base

N H

O2N

N H 229, >80% yield O2N

227

Ph

N Ph O N O 230

N

PhCHO

N

+

N

O2N

>95% yield, 3:1 isomers 231

N H

Cl

K2CO3 DMF

HO2C

N

1. O 3 2. HCO 3H

Ph

Ph

O2N

N

Ph 232, 75% yield crystalline acid

35 A Designed Enzyme Inhibitor

859

Now the seven-membered ring must be built up. Here an azole-based reagent, CDI, carbonyl di-imidazole proves invaluable. Reaction with 232 gives the acylimidazole 233 and this acylates the potassium ‘enolate’ of nitromethane in good yield to give 234 and hence, by reduction, the diamine 235. The N-benzyl group can now be removed and the synthesis of the seven-membered ring is completed with the one-carbon electrophile HC(OEt)3.

O N

HO2C

N

198 heat

N

+ N

N

232

H2N

H2N N H

H2N

Ph 234; 75% yield

HC(OEt)3

N HN

EtOH DMSO

N

Ph 235; crystalline 74% yield diHCl salt

N

O2N

O N

Pd/C pH10:1 syn:anti

The first step is the formation of an alkyne-Ru π-complex 249 that adds bromide ion, presumably transferred from Ru in view of the stereochemistry of 250. Now the enone 247 and the vinyl-Ru complex 250 couple with retention of the stereochemistry of 250 and transfer of the Ru to oxygen to make a Ru-enolate. This is known to be true as work-up in the absence of the aldehyde gives the ketone 252.

892

36 Tandem Organic Reactions

R1

RuCp

RuCp

Br

R1

Br 249

R2

R1

RuCp

R2

R1

247

250

Br

ORuCp 251

Br

O 252

If the aldehyde is present, a Ru-aldol reaction gives a single geometrical and diastereomeric aldol product 248 and the Ru is lost as CpRu⫹ so it can catalyse the next cycle. There are quite a few extra reagents needed to make the reaction go well: catalytic Sn(II)Br2 and the curious spirocyclic ammonium salt. This latter is essential.

References General references are given on page 893 1. T.-L. Ho, Tandem Organic Reactions, Wiley, New York, 1992. 2. M. J. Chapdelaine and M. Hulce, Org. Reacts., 1990, 38, 225. 3. D. P. Curran, Aldrichim. Acta, 2001, 33, 194; D. P. Curran and D. M. Rakiewicz, J. Am. Chem. Soc., 1985, 107, 1448. 4. G. H. Posner, J. J. Sterling, C. E. Whitten, C. M. Lenz and D. J. Brunelle, J. Am. Chem. Soc., 1975, 97, 107. 5. M. Suzuki, T. Kawagishi and R. Noyori, Tetrahedron Lett., 1982, 23, 5563; R. Noyori and M. Suzuki, Angew. Chem., Int. Ed., 1984, 23, 847. 6. M. Suzuki, A. Yanagisawa and R. Noyori, J. Am. Chem. Soc., 1985, 107, 3348. 7. K. K. Heng and R. A. J. Smith, Tetrahedron, 1979, 35, 425. 8. C. H. Heathcock and J. A. Stafford, J. Org. Chem., 1992, 57, 2566. 9. H. Nagaoka, K. Kobayashi, T. Okamura, and Y. Tamada, Tetrahedron Lett., 1987, 28, 6641. 10. G. H. Posner and S.-B. Lu, J. Am. Chem. Soc., 1985, 107, 1424. 11. G. H. Posner, J. P. Mallamo and A. Y. Black, Tetrahedron, 1981, 37, 3921. 12. G. H. Posner, S.-B. Lu, E. Asirvatham, E. F. Silversmith and E. M. Shulman, J. Am. Chem. Soc., 1986, 108, 511; G. H. Posner and E. M. Shulman-Roskes, J. Org. Chem., 1989, 54, 3514. 13. R. V. Stevens, P. M. Lesko, and R. Lapalme, J. Org. Chem., 1975, 40, 3495; C. P. Forbes, J. D. Michau, T. van Ree, A. Wiechers and N. Woudenberg, Tetrahedron Lett., 1976, 935. 14. T. Arai, H. Sasai, K. Aoe, K. Okamura, T. Date, and M. Shibasaki, Angew. Chem., Int. Ed., 1996, 35, 104. 15. S. G. Davies, N. M. Garrido, O. Ichihara and I. A. S. Walters, J. Chem. Soc., Chem. Commun., 1993, 1153; S. G. Davies and D. R. Fenwick, J. Chem. Soc., Chem. Commun., 1995, 1109; N. Asao, N. Tsukada and Y. Yamamoto, J. Chem. Soc., Chem. Commun., 1993, 1660; N. Asao, T. Uyehara and Y. Yamamoto, Tetrahedron, 1990, 46, 4563. 16. C. H. Heathcock, S. K. Davidsen and S. Mills, J. Am. Chem. Soc., 1986, 108, 5650. 17. Y. Tamura, H. Maeda, S. Akai and H. Ishibashi, Tetrahedron Lett., 1982, 23, 2209. 18. S. J. Hays, T. C. Malone and G. Johnson, J. Org. Chem., 1991, 56, 4084. 19. C. Agami, F. Couty, M. Poursolis, and J. Vaissermann, Tetrahedron, 1992, 48, 431. 20. L. E. Overman, K. L. Bell, and F. Ito, J. Am. Chem. Soc., 1984, 106, 4192. 21. T. Kametani, Y. Kato, T. Honda, and K. Fukumoto, J. Am. Chem. Soc., 1976, 98, 8185. 22. B. B. Snider and E. A. Deutsch, J. Org. Chem., 1983, 48, 1822. 23. S. Raucher, K.-W. Chi, K.-J. Hwang, and J. E. Burks, J. Org. Chem., 1986, 51, 5503. 24. W. H. Pearson and K.-C. Lin, Tetrahedron Lett., 1990, 31, 7571. 25. S. Blechert, R. Knier, H. Schroers and T. Wirth, Synthesis, 1995, 592. 26. M. Watanabe and V. Snieckus, J. Am. Chem. Soc., 1980, 102, 1457. 27. J. Tailor and D. G. Hall, Org. Lett., 2000, 2, 3715. 28. Clayden, Organic Chemistry, page 997. 29. M. E. Jung, P. A. Blair, and J. A. Lowe, Tetrahedron Lett., 1976, 1439. 30. D. Schinzer and Y. Bo, Angew. Chem., Int. Ed., 1991, 30, 687. 31. T. Ohshima, K. Kagechika, M. Adachi, M. Sodeoka, and M. Shibasaki, J. Am. Chem. Soc., 1996, 118, 7108. 32. W. J. Zuercher, M. Hashimoto, and R. H. Grubbs, J. Am. Chem. Soc., 1996, 118, 6634. 33. B. M. Trost and A. B. Pinkerton, J. Am. Chem. Soc., 2000, 122, 8081.

General References Full details of important books referred to by abbreviated titles in the chapters to avoid repetition. Clayden, Lithium: J. Clayden, Organolithiums: Selectivity for Synthesis, Pergamon, 2002. Clayden, Organic Chemistry: J. Clayden, N. Greeves, S. Warren and P. Wothers, Organic Chemistry, Oxford University Press, Oxford, 2000. Collins, Chirality in Industry: eds. A. N. Collins, G. N. Sheldrake and J. Crosby, Chirality in Industry: The Commercial Manufacture and Applications of Optically Active Compounds, Vol I, 1992, Vol II, 1997, Wiley, Chichester. Comp. Org. Synth.: eds. Ian Fleming and B. M. Trost, Comprehensive Organic Synthesis, Pergamon, Oxford, 1991, six volumes. Corey, Logic: E. J. Corey and X.-M. Cheng, The Logic of Chemical Synthesis, Wiley, New York, 1989. Designing Syntheses: S. Warren, Designing Organic Syntheses, Wiley, Chichester, 1978. Disconnection Textbook: S. Warren, Organic Synthesis, The Disconnection Approach, Wiley, Chichester, 1982. Disconnection Workbook: S. Warren, Workbook for Organic Synthesis, The Disconnection Approach, Wiley, Chichester, 1982. Drauz and Waldmann: K. Drauz and H. Waldmann, Enzyme Catalysis in Organic Synthesis, VCH, Weinheim, Two Volumes, 1995, ISBN 3-527-28479-6. Eliel: E. L. Eliel and S. H. Wilen, Stereochemistry of Organic Compounds, Wiley, New York, 1994. Fieser, Reagents: L. Fieser and M. Fieser, Reagents for Organic Synthesis, Wiley, New York, 20 volumes, 1967–2000, later volumes by T.-L. Ho. Fleming, Orbitals: Ian Fleming, Frontier Orbitals and Organic Chemical Reactions, Wiley, London, 1976. Fleming, Syntheses: Ian Fleming, Selected Organic Syntheses, Wiley, London, 1973. Houben-Weyl; Methoden der Organischen Chemie, ed. E. Müller, and Methods of Organic Chemistry, ed. H.-G. Padeken, Thieme, Stuttgart, many volumes 1909–2004. House: H. O. House, Modern Synthetic Reactions, Benjamin, Menlo Park, Second Edition, 1972. Morrison: Asymmetric Synthesis, ed. J. D. Morrison, Academic Press, Orlando, 5 Volumes, 1983–5. Nicolaou and Sorensen: K. C. Nicolaou and E. Sorensen, Classics in Total Synthesis: Targets, Strategies, Methods. VCH, Weinheim, 1996. Second volume now published. Ojima, Asymmetric: I. Ojima (ed.) Catalytic Asymmetric Synthesis, Second Edition, Wiley, New York, 2000. Saunders, Top Drugs: J. Saunders, Top Drugs: Top Synthetic Routes, Oxford University Press, Oxford, 2000. Vogel: B. S. Furniss, A. J. Hannaford, P. W. G. Smith, and A. R. Tatchell, Vogel’s Textbook of Practical Organic Chemistry, Fifth Edition, Longman, Harlow, 1989.

Index Acetal, 195 Selective formation under thermodynamic control, 59, 61 Acifran, 207 Acorene, 67 Acyl anion equivalents, 67, 72, and see also chapter 14 See also synthons, d1 Aldol reaction of, 169, 695 Derivatives of cyclic vinyl ethers, 203, 210 Diketones synthesis, 216 Dithians, 205 Functionalised Wittig reagents, 203, 214 Intramolecular, 242, 266 Lithium derivatives of allenyl ethers, 203, 212 Modified acetals, 203–205 Nitroalkanes, 203, 214 Ester d1 synthon, 203, 209 Oxidative cleavage of allenes, 212 Vinyl ethers, 212 Yamamoto’s ROCH(CN) 2, 209 Acyloin reaction, 73, 178 Silicon modification, 73 Agelastatin, 465, 493 AL-4862 (Azopt), 510, 511 Aldol reaction, 201 See also specific substrates Acid catalysed, 214, 297, 314 anti-Selective of boron enolates, 611 of ester enolates, 18, 85 of lithium enolates, 778, 791 of silyl enol ethers, 795–796 Diastereoselective, 52, 332 E and Z nomenclature, 355, 405 Enone synthesis by, 58, 67, 621 Enzymatic, 436, 459 Equilibrating reaction, 19, 51, 73 Homoaldol reaction, 189, 194 1,5-Induction, 708

1,3-anti Induction, 702 1,4-syn Induction, 706 Mukaiyama aldol reaction, 32 Of lithium enolates, see chapters 3–6, plus 778, 791, 800 Of enamines, 30 Of extended enolates, 155, 159 Of silyl enol ethers, 20, 163, 191 Of zirconium enolates, 49 Open transition state of silyl enol ethers, 48 Proline-catalysed, 579 Stereochemical control, 690 Stereoselective aldol reaction, 44, 51, 404, 670, 737, 739, 868 syn-Selective of boron enolates, 47 of phenylthio esters, 47 of ketone enolates, 48 of silyl enol ethers, 48 of zirconium enolates, 49 of thioester boron enolates, 611 Synthesis of (⫹)-discodermolide, 709 Tandem Michael and aldol reaction, 869 Alkaloids, 475, 501 Ligands for the Sharpless asymmetric dihydroxylation, 538 Alkenes, 277, 512, 543 See also Wittig and Horner-Wadsworth-Emmons reactions, 155, 157 Amine catalysed isomerisation, 228 Control of geometry, see chapter 15 Epoxidation of, 224, 273 Markovnikov addition, 283 anti-Markovnikov addition, 284 Isomerisation, 353, 362 Intramolecular trapping of by iminium ions, 668, 877 Photochemical isomerisation, 250 Polyene cyclisations, 302 Radical isomerisation, 251 cis-Selective synthesis, 145

Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren Copyright © 2007 John Wiley & Sons, Ltd

896

Alkenes (Continued) Stereospecific Z-alkene synthesis, 712 Stereospecific E-alkene synthesis, 252 Synthesis by selenoxide elimination, 213 trans-Selective synthesis, 232 Reaction with electrophiles, 256, 433 regioselective, 520, 531 chemoselective, 571, 604 Alkynes, 248, 262 Bromination of, 789 Hydroboration of, 263 Michael addition to, 130, 621 Reduction to alkenes, 248 E-Selective LiAlH4 reduction of, 248 1,3-Allylic shift, 174 (⫹)-α-Allo-kainic acid, 821–822 Allopurinol, 839 Allyl anions equivalents, 173 See chapter 12, 315 Allylboration, asymmetric, 718, 738 Allyl dianion, 183 Allylic alcohols, 340, 343, 359 See chapter 19, 735 Chromium(VI) mediated rearrangement of, 7, 85 Directed Simmons-Smith cyclopropanation, 585 Heck reaction of, 854–855 Palladium-catalysed reaction of, 6, 267 Rearrangement of, 83 Reaction with nucleophiles, 109, 112, 115 Stereochemically controlled epoxidations of, 350 Synthesis from aldehydes and ketones, 342 by enone reduction, 341 from iodolactonisation products, 341 by Wittig reaction, 342–343 Allylic alkylation, asymmetric, 685 Allylic phosphates, 343 Ambruticin, 563 Amino acids, see chapters 22 and 23 Diazotisation of, 467 Nucleophilic substitution of α-lactones, 513, 517 Conversion to, 321, 666 α-chloro acids, 467 epoxides, 350, 363 (R)-amino acids, 382, 449 Reduction of, 468 Synthesis from 2-acylamino acrylates, 568 Amino alcohols, 499 Synthesis of quinolone antibiotics, 484 oxazolidinone antibiotics, 485 anti-viral nucleoside analogues, 486 Aminohydroxylation, 528, 552 Amino-nitriles, 199 Michael addition of, 228, 621

Index

Hydrolysis of, 624, 744 Amlodipine, 439, 676, 677 (R) and (S)-Amino-1-phenylethanol, 718 Amopyroquine, 12–13 Amphotericin, 118 (⫹)-Anatoxin-A, 772 Anomeric effect, 46–47 Antheridic acid, 60 (⫺)-Aplysistatin, 689 AspartameTM, 467, 570 Asperazine, 329 (⫹)-Aspicilin, 549 (⫹)-Astericanolide, 690 Asymmetric deprotonation, 522 Of tropinone, 522 With sparteine, 522 Asymmetric induction, 528, 533–534 Reagent based control, see chapter 24 Atpenin B, 783 Aza-enolates, 142–143, 145 See enolates Azatyrosine, 193, 605 Aziridine, 696, 778 Reaction with azide, 844, 847 AZT (Azidothymidine), 851 Baeyer-Villiger rearrangement, 36 Of cyclohexanones, 35, 147, 471 Enzymatic, 487, 575 Regioselectivity, 593–594, 614 Reverse selectivity, 664 With engineered baker’s yeast, 664 (⫺)-Balanol, 533–534 Baker’s yeast, 652, 664 Reduction of ketones, 510 Mnemonic for ketone reduction, 530–531 Bakke nitration, 773 Katritzky’s improvement of, 773 BAY 43-9006, 772 Baylis-Hillman reaction, 166 Asymmetric, organic catalysis of, 161, 166 Beckmann rearrangement, 889 Tandem Beckmann rearrangement and allyl silane cyclisation, 889 Benzyne, 866, 880 Formation using ortho-lithiation, 96–100 Bestatin, 480 Birch reduction, 622 Asymmetric, 622 Combined with iodolactonisation, 617, 682 Asymmetric of heterocycles, 751 Schultz’s asymmetric, 682 Synthesis of extended enolates, 161, 165 Blastmycin, 427 BMS 181100, 574

Index

Bombykol, 232 Boranes, 256, 264, 299 See also hydroboration, 284 Asymmetric borane reduction of unsymmetric ketones, 507 Alpine-Borane®, 509 Ipc2BCl (DIP-Chloride®) reduction with, 510 ketones, 507–510 β-keto-acids, 219 Reaction of chiral allylic boranes with imines, 513 Boronic acids, 756, 772 Synthesis from alkynes, 266 Vinyl boronic acids, synthesis, 330, 334 Boronic esters, 329 Vinyl boronic esters, synthesis, 330, 334 Bredereck reaction, 857 Brefeldin, 270 (⫹)-Brevicomin, 312, 667 Bromination, 333, 495 Of alkenes, 511–512 Of benzene, 622 Bromolactonisation, 293, 419 Brown’s crotyl borane, 739 (⫺)-Brunsvigine, 64 Buchwald-Hartwig coupling, 756 Buproprion, 778, 806 Bürgi-Dunitz angle, 695, 702 Cadinene, 66 (S)-Camptothecin, 763 Cannabispirenones, 777, 792 Captopril, 465, 476 Carboalumination, 267 Carbocupration, 268 Carbohydrates, 473 Synthesis of D-glyceraldehyde, 474 Carbo-metallation, 267 Carbonylation, 299 Of organo-zirconium reagents, 121 Of palladium σ-complexes, 122 Carboxylic Acids, 37, 63 Lithium enolate formation, 289 Reaction with vinyl-lithium reagents, 269 β-Carotene, 329 CBS reduction of ketones, 575 Synthesis of cetirizine, 576 Cecropia juvenile hormone, 279 Cedrene, 67 (⫹)-(R)-Cetrizine, 448 (⫺)-Chanoclavine, 595 Chelation control, 683, 696, 802 Of Grignard reagents, 13, 51, 113, 115–117 Of Zn(BH4) 2 reduction, 427–428 Chiral auxiliaries, 466, 469, 613 See chapter 27, 763 See also specific auxiliary

Asymmetric aldol with, 327–328 Conjugate (Michael) addition with, 613 Diels-Alder reaction with, 613–614 Chiral Pool, 676, 683 See chapter 23, 244, 514 New chiral pool, 588 amino-indanols, 491 C2 symmetric diamines, 555 epichlorhydrin, 637 glycidol, 487 phenylglycine, 488–489 Tables of most common substrates, 497–501 Chloroenones, 261 Chloroquine, 446 Chrysanthemic acid, 460 trans-Chrysanthemic acid, 460 Cilastatin, 585 (⫺)-Cinatrin B, 358 (S)-(⫺)-Citronellol, 472, 571 Claisen ester condensation, 209, 224 Claisen(-Cope) Rearrangement, 224, 246 Aza-Claisen rearrangement, 823 Claisen-Ireland rearrangement, 355 Chirality transfer, 420, 432 Diastereoselectivity, 419–422 Palladium(II) catalysed, 283, 318 Regioselectivity, 320–322 Stereochemistry in, 354, 380 Claisen ester condensation, 209, 224 Conformational control, 410, 412 Concanamycin, 268 Conjugate addition, see chapter 9, plus 268–271 See Michael addition, 287, 309 Conjugate substitution, 310, 312 See Michael addition, 287, 309 Cope rearrangement, 352–353 Oxy-Cope rearrangement, 823 Aza-Cope, 823–825 anionic aza-Cope, 827 tandem aza-Cope and Mannich reactions, 824 Copper, 64, 119 Aryl-amine cross coupling, 124–125 Catalysed SNAr reaction, 484 Organo-copper reagents, 129, 137 Michael addition of, 148, 228 stereoselective, 228, 233 structure, 248 Organo-cuprates, 131, 316, 619, 708 acylation of, 237–238, 463 alkynyl cuprates, 269–270 lithium cuprates, 132, 147 Michael addition of, 148, 228 reaction with alkyl halides, 116 structure, 248

897

898

Copper (Continued) vinyl cuprates, 261, 270, 327, 486 Michael addition of, 148, 228 Cyanocuprates, 131 Michael addition of, 148, 228 Corey-Fuchs reaction, 322 Coriolin, 74, 284 CP-060S, 455 Crixivan (Indinavir), 491, 493 Crocacin C, 327 Crocacin D, 327 Cuprates, 314, 316, 327 See organo-copper reagents, 129 α-Curcumene, 18–19 Curtius rearrangement, 481, 586 Cyanohydrin, 655, 665 Enzymatic formation, 567, 670 Enzymatic reaction of, 632 Cycloadditions, 747, 809 See also Diels-Alder reaction and 1,3-dipolar cycloaddition, 822, 885–886 2⫹2 Photochemical, 134 2⫹2 Ketene cycloadditions, asymmetric, 587 Synthesis of enalapril, 589 Cycloheptanones Synthesis using oxyallyl cations, 71, 78 Cyclopentannelation, 195, 284 Cyclopentanones, 79, 664 Synthesis, 81 by oxidative rearrangement of tertiary allylic alcohols, 83 by Pauson-Khand reaction, 71, 79 using chromium carbenes, 86 using oxyallyl cations, 71, 78–79 using ynoate(s), 85 Cyclopropanes, 192, 193, 240 See also Simmons-Smith reaction, 349, 351 Synthesis, 350 asymmetric, 355, 382 using Michael addition, 426, 577 using sulfoxonium ylids, 128 Danishefsky’s diene, 291–292, 561 DARVON, 454, 507 (S,S)-iso-ddA, 486 Decarboxylation, 140, 148, 250, 360 Dendrobine, 195 11-Deoxytetrodotoxin, 352 Deracemisation, 505, 517–518 Of arylpropionic acids, 654 Of α-halo acids, 515, 518 Desymmetrisation Reactions, 528, 558 Comparison with kinetic resolution, 560 Opening of anhydrides, 528

Index

Of bis-allylic esters, 594 Opening epoxides, 399 Of immobilised enzymes, 675 Of a symmetrical Diels-Alder adduct, 677 Of a dihydropyridine, 677 Then kinetic resolution, 646 With lipases, 656 Mono esterification of diols, 561 Diastereoselection of 1,3-diols, 424 Diazines, 761 Diazoketones, rearrangement to carboxylic acids, 844, 847 Diazotisation, 467 In the electrophilic substitution of benzene, 622 DIBAL, 666 Ester reduction, 653 Nitrile reduction, 23 1,3-Dicarbonyls, 785 Synthesis, 785 Alkylation of, 848 1,4-Dicarbonyls, 785 Synthesis by Stetter reaction, 220 1,5-Dicarbonyls, 785 Synthesis, 785 Diels-Alder reaction, 291, 315, 345 See also hetero-Diels-Alder reaction, 561–562 Asymmetric, 296 organic catalysis, 577 C2-symmetric Lewis acid catalysis, 583 with box and pybox catalysts, 584 with Oppolzer’s chiral sultam, 615 with chiral auxiliaries attached to the diene, 617 Aza-Diels Alder reaction, 819, 884, 888 Tandem aza-Diels-Alder and allyl boronate reactions, 864, 884 Tandem Aza-Diels-Alder and aza-ene reactions, 864, 888 Danishefsky’s diene, 561–562, 617 Enantioselectivity in catalysed, 584 Endo-selectivity, 401–403, 425, 506 Hetero-Diels-Alder reaction, 345, 561 Asymmetric, 342, 355 auxiliary controlled, 614–618 copper bis-oxazoline catalysed, 584 chromium Salen catalysed, 562 Synthesis of ambruticin, 563, 564 Lewis acid catalysed, 169, 214 Of azadienes, 812 1-azadienes, 812 2-azadienes, 814 Of β-bromo alkynes, 315 Of β-sulfonyl alkynes, 315 Of imines, 687 Intramolecular, 686, 700, 739

Index

with azadienes, 811–814 with imines, 812, 815 with a nitrogen tether, 819, 820 Stereoselectivity of, 224, 304, 851 Differential crystallisation (entrainment), 449 With racemisation, 451–453 Typical procedure, 449 4, 5-Dihydrostreptazolin, 244 Dihydroerythramine, 134 Dihydroxylation reaction, 282, 304 See also Sharpless asymmetric dihydroxylation reaction, 538 Of alkenes, 538, 542 Racemic dihydroxylation reaction, 737 (⫹)-cis-Diltiazem, 621 Synthesis using, 812 Sharpless asymmetric dihydroxylation reaction, 587 Jacobsen epoxidation, 663 Evans’ oxazolidinones, 621 1.2-Diols, 15 Conversion to epoxides, 38 1,3-Dipolar cycloaddition, 248, 658 Isoxazole synthesis, 841–843 Tetrazole synthesis, 847 Tandem reactions involving, 885 (⫹)-Discodermolide, 709 Large scale synthesis, 712, 742 (⫹)-Disparlure, 531–532 Dithians, 207 Acylation of, 237 Alkylation of, 247 Dithioacetal monoxides, 204, 209 Exchange, 209 Synthesis of, 257 Conversion to carbonyl compounds, 267 Dithioacetal monoxides, 204, 209 E1 reaction, 228 E1cB reaction, 228 E2 reaction, 232 Efavirenz, 515, 591 Electrocyclic reactions, 828 Electrophilic substitution, 807 Of benzene, 803 By oxygen on benzene, 778 Of pyridine, 812, 824 α-Eleostearic acid, 129 Enalapril, 478 Enamines, 19 See specific reactions also, 181 A1,3 strain Acylation of, 238, 621 Alkylation of, 30, 143, 159 Aldol reaction of, 158–159

Extended, 158–159 Reaction with, 181 alkyl halides, 196–197 α-carbonyl halides, 142 α-halo carbonyls, 38 oxyallyl cations, 71, 78 Michael addition of, 136 Regioselective formation of, 142, 149 Robinson annulation of, 226 [3,3] Sigmatropic rearrangement of, 246 Synthesis of 1,4-dicarbonyls, 785 1,5-dicarbonyls, 785 Ene (Alder ene) reaction, 884 Oxo-ene reaction, 297, 299 Asymmetric, 304, 327 Asymmetric carbonyl ene, 590 Aza-ene reaction, 864, 884 Intramolecular, 890 with a nitrogen tether, 819, 820 with oximes, 822 Tandem ene reactions, 882 Enolates, 16–20 See also homo-and extended enolates, 158, 159 Acylation, 181 Alkylation of, 186 Aldol reaction of, see chapters 3–6, plus 695, 736 Aza-enolates, 143, 145 alkylation of, 186 alkylation of aldehydes, 145 alkylation and Michael, 146 extended, 155 metallated hydrazones, 139, 145 reaction with alkyl halide, 159, 162 reaction with epoxides, 180 Synthesis, 180 Alkylation of, 186 Alkylation of aldehydes, 145 Alkylation and Michael, 146 Capture by electrophiles, 863, 866 Chiral, 867–868 RAMP and SAMP, 469, 599, 601 Schöllkopf’s bislactim ethers, 603 Seebach’s relay chiral units, 606 from hydroxyl acids, 607 from Evans’ oxazolidinones, 609 William’s chiral glycine enolate equivalent, 605 cis-Boron enolate generation, 408–410 Conjugate substitution reaction of, 866–869 Electrophilic attack by oxygen, 602, 694 MoOPH hydroxylation of, 778, 800 N-Sulfonyl oxaziridine hydroxylation of, 802 Nitrosation with nitroso compounds, 786 Reaction of endocyclic enolates, 413

899

900

Enolates (Continued) Reaction of exocyclic enolates, equatorial vs axial attack, 399 Reaction with epoxides, 38, 116 Of 1,3-Dicarbonyls, 785 “Naked” from silyl enol ethers, 146 Reaction with epoxides, 180 Regioselective formation of, 178, 260 trans-Boron enolate generation, 425 O-Acylation of, 425 Enols, 16, 29, 784 See also enolates and enolisation, 784 Specific enol equivalents, 17 Bromination of, 789 Selenium oxide oxidation of, 785–786 Nitrosation with nitrites, 786 Enol Ethers, 139, 796, 799 Synthesis from enols, 778 Asymmetric hydroboration of, 512 Enolisation, 189 Regioselectivity, 196 Under kinetic control, 606, 670 Under thermodynamic control, 668, 670 Under equilibrating conditions, 73, 75, 140 Enones, 163 Oxidation of, 281, 301 Synthesis, see chapter 5, also 286 by acid catalysed aldol reaction, 579 by acylation of vinyl metal reagents, 63 by aldol reaction, 71, 74 by base catalysed aldol reaction, 579 by bromination of enols, 789 by palladium(II) oxidation of silyl enol ethers, 777 using sulfur and selenium compounds, 790 E-Selective synthesis by Wittig style ‘aldol’ reaction, 208, 232, 236 Regioselective asymmetric hydrogenation of, 572 Enzymes, 636 Acylation of alcohols, 631 Desymmetrisation, 646, 627 with immobilised enzymes, 657 with lipases, 656 of a symmetrical Diels-Alder adduct, 677, 687 of a dihydropyridine, 812, 824 Effects of amines on lipases and esterases, 651, 659 Engineered aldolase, 669 Ester formation and hydrolysis, 651, 653 Kinetic resolution with, 655 as resolving agents, 456 by ester hydrolysis, 457 of secondary alcohols with lipases, 458 of α-acetoxysulfides, 639 of amines, 659 of ibuprofen with esterases, 517, 654

Index

of hemiacetals using lipase, 193, 794 with proteolytic enzymes, 459 with racemisation, 460 Nucleophilic addition to carbonyl groups, 665 Addition of cyanide to aldehydes, 665 Catalysed aldol reactions, 667 Acylating enzymes, 660 Oxidation with, 701, 739 asymmetric dihydroxylation of benzenes, 725, 742 Baeyer-Villiger rearrangement, 36, 794–795 Baeyer-Villiger rearrangement with engineered baker’s yeast, 36, 118, 238 epoxidation, 224, 261 Polymer-supported reagents and enzymes, 658 Reduction of ketones using baker’s yeast, 652 Synthesis of, 658 prostaglandins, 661 syringolide, 667 Versus whole organisms, 656 (1R,2S)-(⫺)-Ephedrine, 470–471, 515 Epibatidine, 315, 749, 772, 779 Epichlorhydrin, 487–488, 499 As a chiral pool reagent, 465, 487 Nucleophilic attack on, 490, 505, 513 Reaction with α-naphthol, 530 Epothilone A, 669–670 Epoxidation, 670, 673 See also epoxides, Sharpless asymmetric and Jacobsen epoxidations, 559, 663 Asymmetric epoxidation with chiral sulfur ylids, 516 With sulfur ylids, 128, 516 With sulfonium ylids, 333–334, 339 With mCPBA, 795–797 Epoxidation with VO(acac) 2 /t-BuOOH, 281 Epoxidation with H2O2 /base, 284, 289 Enzymatic, 436, 459 Epoxides, 467, 499 See also epoxidations, 350, 559 Chemoselective synthesis, 667, 741 Rearrangement to aldehydes, 62, 132 Reaction with azide, 280, 309 enolates, 313–314 Grignard reagents, 591, 666 Me3SiBr, 418 Regioselective reduction of, 277, 286 Regioselective synthesis, 30 Stereoselective synthesis, 48, 443 using allylic alcohols, 55, 63 Synthesis from amino acids, 193 1,2-diols, 528, 538 trans-Di-axial product, 414, 416 Eschenmoser fragmentation, 793

Index

Erythrina alkaloid, 134, 876 Esomeprazole, 524, 759 Ester, 18, 147 Lithium enolate formation, 146, 804 Alkylation of, 30, 143 Aldol reaction of, 225, 158–159 Reduction of, 223 Erythronolides, 292 Eucarvone, 128 Evans-Mislow rearrangement, 694 Evans’ oxazolidinones, 599, 609 Asymmetric alkylation with, 717 Asymmetric Michael addition with, 11, 13, 870 syn-Aldol reaction with boron enolates, 43, 47 anti-Aldol reaction by Lewis acid catalysis, 149, 160 Evans-Tishchenko reduction, 424 Extended enolates, see chapter 11 Alkylation, 154, 156 Generation of kinetic enols, 798 Ketones, 651, 652 Kinetic product, 155 Reaction at α-position, 157–158 Reaction at γ-position, 158–159 Synthesis using Birch reduction, 165 Thermodynamic generation, 155–157, 159 Thermodynamic product, 155 Felkin-Anh Selectivity, 237, 429–431, 666 Conformation, 681, 683 Control, 683 Fenarimol, 116 Fenfluramine, 435, 451 Fentiazac, 840 Flexibilene, 5–7, 267 Fluconazole, 447 (S)-Fluoxetine (Prozac), 510 Flupirtine, 752 Flutriafol, 3, 447 Formylation, 564, 761 By ortho-lithiation, 763–764 Fosinopril, 476 (⫹)-FR900482, 696 FR-900848, 392 Friedel-Crafts reaction, 34, 38, 55 Acylation, 37 Alkylation, 34, 35 Aliphatic Friedel-Crafts reaction, 55, 64 intramolecular, 74 Regioselectivity, 77, 80 With heterocycles, 165 Fries rearrangement, 92 Anionic Fries rearrangement, 107 Regioselectivity, 107, 131

901

(⫺)-Fumagillol, 64 Fusilade (Fluazifop butyl), 456 Geiparvirin, 19 Gibberellin, 257 (⫹)-Grandisol, 260–261, 723 Grignard reagents, 13, 51 Acylation of, 52 Alkylation of, 57 Allyl Grignard reagents, 173 Cerium-catalysed addition of, 313 Chelation-controlled addition, 247 Chiral Grignard reagents, 732 Desymmetrisation of a symmetrical anhydride using a chiral Grignard, 732 Reaction with aldehydes, 793, 815 CO2, 827, 831 epoxides, 38 ketones, 43–44 nitriles, 117, 199 Reduction of aldehydes and ketones, 793 Structure, 200 Synthesis, 193 Vinyl Grignard, 256–257 addition to ketones, 258 synthesis, 260 Synthesis by exchange, 264 Gingerol, 32 Ginkgolide, 139, 148 Grandisol, 260–261 Half-chair, 247, 347 Halicholactone, 349 Halogen dance, 761 Halolactonisation, 289 See also bromo- and iodo-lactonisation, 290, 293 Hantszch pyridine synthesis, 439 Heck reaction, 317–323, 755, 762 Asymmetric, 763 Mechanism of, 317, 765 Intramolecular, 773, 797 Reaction of allylic alcohols, 346 Synthesis of strychnine, 320 asperazine, 329 With heterocycles, 334, 587 Tandem asymmetric Heck and Palladium-allyl cation reaction, 864, 890 Henry reaction, 225 Hetero-Diels-Alder reaction, 345 See Diels-Alder reaction, 393 Hexaconazole, 447 Hirsutene, 24–25

902

Homoenolate, see chapter 13 Acetal derived, 193 Allyl carbamates, 201 Aryl coupling of, 193 Cyclopropyl, 194 Enantioselective generation, 193 From 3-membered rings, 467 Generation from 2-haloacids, 191 Generation from 2-chloroesters, 191 Heteroatom-substituted allyl anions, 189, 196 Michael addition of zinc homoenolates, 192–193 Of allyl amines, 198 Of allyl silanes, 197 Of allyl sulfides, 197 Of carboxylic acid, addition to aldehydes and ketones, 190 Of sulfones, 186 acylation of, 183 synthesis of, 237 Reaction with ketones and aldehyde, 180 Reaction with electrophiles, 180 Synthesis of enantionmerically pure β-oxo-amino acids, 193 Zirconium, 192 Hoppe’s ‘defensive strategy’, 190 Houk conformation, 351, 400 Houk control, 697 Horeau principle, 391 Horner-Wadsworth-Emmons reaction, 62, 136, 233, 733 Still and Gennari modification, 236 E-selective, 241 Z-selective, 342 β-Hydride elimination, 121–123 Hydantoin synthesis, 488 Hydroalumination, 263–265 Of alkynes, 263 Of propargyl alcohols, 265 Hydroboration, 263 Of alkenes, 263 Of alkynes, 263 Asymmetric hydroboration of alkenes, 512 Asymmetric hydroboration of enol ethers, 513 Carbonylation of alkyl boranes, 299 Ketone synthesis using cyanide, 301 Hydrogenation, 51, 122 Asymmetric hydrogenation of alkenes, 493 Asymmetric hydrogenation of carbonyl groups, 572 Catalytic asymmetric homogeneous hydrogenation, 568 With C2 symmetrical phosphine Rh complexes, 572 With C2 symmetrical BINAP Rh and Ru complexes, 572 With Raney nickel, 621 Reduction of 2-acylamino acrylates, 569 Regioselective asymmetric hydrogenation of enones, 572

Index

Of alkynes, 80 Transfer hydrogenation, 320, 479, 640 Noyori’s synthesis of menthol, 574 Using Wilkinson’s catalyst, 568 Hydrometallation, 120 Of alkynes, 80 Hydrosilylation, 120 Intramolecular, 124, 128, 133 Metal catalysed, 120 Hydrostannylation, 262 Of terminal alkynes, 262 Hydroxylation, 781 Hydroxy acids Synthesis of, 783 TADDOL, 469 propylene oxide, 470 streptazolin from tartaric acid, 475, 482 (2S,4R)-Hydroxypipecolic acid, 720 Hydrozirconation, 266 Of alkenes, 270 Imines, 601 Enamine tautomerism of, 848 Hydrogenation of, 478, 572 Tandem iminium ion formation and vinyl silane cyclisation, 880 Imidazoles, 859 Nitration of, 858 Synthesis by Bredereck reaction, 857 Indicine, 543 Organo-indium reagents, 184 Allyl indiums, 184 synthesis of all-trans dienes, 184 Indolizomycin, 239 Iodo-etherification, 674 Iodolactonisation, 682–683 Combined with Birch reduction, 165 Stereoselectivity, 168, 185 Ircinal A, 244 Iridomyrmecin, 287, 289 Isoxazoles, 841 Synthesis by 1,3 dipolar cycloaddition, 842 Itraconazole (Sporanox), 686 Jacobsen Epoxidation, 491, 528 Catalyst, 51 Synthesis, 51 of crixivan, 491 of ditiazem, 557 cis-Jasmone, 71–72, 211 Jones oxidation, 332 Julia reaction, 494–495, 241 Kocienski modification, 241 Juvabione, 51

Index

Kalihinene X (⫺)-α-Kainic acid, 357 Kedarcidin, 755 Kendomycin, 135 Ketorolac, 460, 655, 656 Ketones, 11, 15, 18 Asymmetric borane reduction of unsymmetrical, 507 Asymmetric addition of carbon nucleophiles, 515 Baker’s yeast reduction of, 688 Corey’s CBS reduction of, 575 Ipc2BCl (DIP-Chloride®) reduction of β-keto-acids, 510 Reduction of using; Ipc2BCl (DIP-Chloride®), 510 K-Selectride®, 721 REDAL, 731 Kinetic control, 28, 33 Of enolisation, 666 Kinetic resolution reactions, see chapter 28 See also enzymes, 635 Of diastereomeric mixtures, 384, 447 Double methods, 635 Dynamic kinetic resolutions, 636 reaction of epichlorhydrin, 637 of α-acetoxysulfides, 639 of enzymes and metals together, 640 by hydrogenation, 640–641 Parallel kinetic resolution, 641 With racemisation, 655 Regiodivergent resoltuions, 644 S Values, 630 By Sharpless asymmetric epoxidation, 647 Synthesis of 2-methyl-tetrahydropyran-4-one, 730 With chiral DMAP, 631 Unwanted kinetic resolutions, 635 Knoevenagel reaction, 141, 857 (⫹)-Lactacystin, 718, 734 Lactonisation, 224, 289 See also halo- and sulfenyl-lactonisations, 292–293 Enantioselective lactonisation of achiral hydroxy diacids, 521 LAF389, 495 Lamivudine, 638–639 (⫹)-Laurencin, 481 Lindlar’s catalyst, 248 Lipase, 358, 457–460 Kinetic resolution of hemiacetals, 193 Kinetic resolution of secondary alcohols, 458 Lipstatin, 22 Lipoic acid, 35 alpha-Lithiation, 108–110 Lateral lithiation, 111 Chemoselectivity, 111

903

ortho-Lithiation, 91, 96 Anionic Fries rearrangement, 107, 112 Benzyne formation by, 110 Dilithiations, 106 Directing groups, 98, 99 containing oxygen, 98 containing nitrogen, 99 fluorine, 103, 108 Multiple lithiations, 101 Multiple directed lithiations, 101 Of fluoroanisoles, 103 Of pyridine, 750–751 Of quinolones, 484 Regioselectivity, 508, 513 Lithium, 17–19, 115 Organo-lithium reagents, 269, 514 acylation of, 628, 631 allyl lithium reagents, 175–176, 181 carbonylation of, 779, 857 conjugate substitution reaction of, 316 reaction with alkyl halides, 115 carboxylic acids, 117 epoxides, 128 structure, 128 vinyl-lithium reagents, 269 synthesis, 270 Lithium amide addition, 17, 112, 522 Asymmetric, 522 Lithium-halogen exchange, 99 Longifolene, 15 Lotrafiban, 676 Luche reaction, 260, 891 Lycorine, 683, 685 Lysergic acid, 828 (⫹)-Macbecin I, 699–700 Macrolactonisation, 150 Yamaguchi lactonisation, 670 Mannich reaction, 824, 828, 872, 875 Reverse Mannich reaction within chiral resolutions, 454 Tandem aza-Cope and Mannich reactions, 824 Tandem Michael and Mannich reactions, 869 Mannicone, 21 Mappicine, 762 Markovnikov addition, 283–284 Marimastat, 726 McMurry reaction, 242 Intramolecular, 242 Meldrum’s acid, 24, 141 Aldol reaction of, 161 (⫺)-Menthol, 499, 573 L-methyl DOPA, 450 Methyl shikimate, 418

904

Mercuration, of alkenes, 283, 294, 440 Mercuration-reduction, 283 Metathesis, 243–244, 494 Asymmetric, 241 Cross, 247 Grubbs’ Catalysts, 243 Hoveyda-Grubbs catalyst, 494 Mechanism, 243 Molybdenum-catalysed, 800–801 Schrock’s molybdenum catalyst, 586 Ring closing, 829 tandem ring-closing and ring-opening, 864, 891 Methoxatin, 251 Methylation, 140, 249 Methylenation, 135 Methylenomycin, 81 O-Methyljoubertiamine, 204, 216 Metronidazole (Flagyl®), 849 Mevinolin, 788 MGS002836, 685 Michael (conjugate) addition, see chapter 9, also 39, 619 Asymmetric, 619 addition of lithium nucleophiles, 619 with chiral auxiliaries, 620 with chiral sulfoxides, 621 with Evans’ oxazolidinones, 600 with 8-phenylmenthyl esters, 619 induction by a chiral auxiliary, 868 Copper catalysed, 684 Conjugate substitution reaction, 316 Followed by reaction with electrophiles, 316, 319, 323 Fuctionalised Michael donors, 136 Intramolecular, 180 To alkynes, 261 Of alkynyl cuprates, 269–270 Of amino-nitriles, 199 Of anilines, 812 Of cuprates, 133 Of cyanocuprates, 131 Of 1,3-dicarbonyl compounds, 37, 141 Of organo-copper reagents, 129–131 Of enamines, 30, 35, 38 Of enolates, 35, 46 Of extended enolates, 159 Of heteroatom nucleophiles, 135 Of Grignard reagents, 178, 193 Of nitroalkanes, 204, 206, 218 Of nitroketones, 216 Of silyl cuprate, 217 Of silyl enol ethers, 777–778, 791, 796 Of sulfur ylids, 128, 516 Of sulfonium ylids, 128, 343 Of sulfoxonium ylids, 128 Of tetrazoles, 844

Index

Organic catalysis of, 577, 582 Stereoselective, 563 copper catalysed, 562 Synthesis, 19 of 1,5-dicarbonyl compounds, 19, 34 of diketones, 38 Tandem, 863–864 Michael-Mannich reactions, 132 Michael-Michael additions, 869 stereochemical control in, 867 TiCl4 catalysed, 48 Versus 1,2-addition, 39 Microscopic reversibility, 415 Milbemycin β3, 330, 332 ‘MIMIRC’ sequence, 863, 871 Tuning with different Michael acceptors, 872 Mitsunobu reaction, 22, 208, 241 MK-0966, 84 Modhephene, 77 Monensin, 357, 419 Monomorine, 687, 688 MoOPH, 778, 800 Hydroxylation of enolates, 802, 804 Hydroxylation of steroids and amino acids, 801 Stereoselectivity, 801–802, 804 MSD 427, 845 Mukaiyama aldol reaction, 32 See also silyl enol ethers, 33 Multistriatin, 139, 143 Myxalamide A, 264 Nafuredin, 241 Nazarov reaction, 71, 77 Lewis acid catalysed, 169, 214 Regioselectivity, 245, 264 Effect of fluorinated substituents, 595, 686 Effect of silyl substituents, 78 Negishi coupling, 755, 761, 764, 772 (⫺)-Neoplanocin, 136 Nickel,160 Catalysed butadiene dimerisation, 164 Organo-nickel reagents allyl nickel complexes, 177 alkylation of, 186 cyclopentenone synthesis, 173, 178 synthesis of, 178 Nifluminic acid, 749, 757 Nitration, 758, 764, 766 Of benzene, 759, 766 Of pyridine, 758 Nitroalkanes, 204, 218 Conversion to ketones, 208, 218 Synthesis of, 219 Mono-protected ketones, 219 Strigol, 204, 219

Index

Nootkatone, 50 Norvir, 479 NOVRAD, 454, 507 N-Sulfonyl oxaziridines, 778, 802, 803 Asymmetric hydroxylation with camphor sultam derivatives, 804 Hydroxylation of enolates, 804 Stereoselectivity, 778, 814 Nuciferal, 195, 144 Nucleophilic addition, 590 Asymmetric nucleophilic attack by chiral alcohols, 517 Asymmetric addition to carbonyls, 515 Enzymatic, to carbonyl groups, 521 Felkin-Anh model, 696 To cyclic ketones, 399, 410 Ofloxacin, 484–485 Omeprazole, 524, 758 Organic catalysis, 577 Proline-catalysed aldol reaction, 579 Baylis-Hillman reaction, 581 Conjugate addition, 581 Diels-Alder reaction, 613–614 Robinson annelation, 790, 865 Reactions with hydroxy acetone, 579 Organo-copper reagents, 129, 137 See copper, 129 Organo-lithium reagents, 117 See lithium, 117 Organo-nickel reagents See nickel Organo-zinc reagents, 568, 591–592 See zinc, 592 Organo-zirconium reagents, 121 See zirconium, 121–122 Oxidation, 114, 118–119 Of aromatic compounds, 98 Asymmetric oxidation of sulfides, 523 Of enolates, 39, 139, 802, 804, 806 Oxidation state, 114, 119 Oxidative insertion, 123, 177 Oxyallyl cations, 78–79 Synthesis of cycloheptanones, 886 Synthesis of cyclopentanones, 79 Oxy-mercuration, of alkynes, 283 oxypalladation, 360 Ozonolysis, 428, 697 Palladium, 6–7, 83, and see chapter 18 see Heck, Stille and Suzuki Allylic substitution reaction, 340 asymmetric, 342 Aryl-amine cross coupling, 124–125 Carbonylation of σ-complexes, 97

Catalysed, 124 cross coupling of vinyl aluminiums, 271 SNAr reaction, 484 tin hydride reduction, 322 Diene synthesis, 325 β-hydride elimination, 121–123 Lactone synthesis by carbonylation, 124 Oxidation of sily enol ethers, 725 Oxypalladation, 283, 360 Reaction with monoepoxides of dienes, 339, 363 β-Panasinene, 134 Papain, kinetic resolution with, 459 Pauson-Khand reaction, 71, 79 Asymmetric, 135 Intramolecular, 74–76 Regioselectivity, 77, 80 Rhodium(I) catalysed, 83 Payne rearrangement, 532 Pederin, 204, 213 Pentostatin, 855, 857, 859 Peptide coupling, 467, 477, 479–480 With DCC, 544, 739 With EDC, 480 (⫺)-Petasinecine, 691–692 (R)-(⫹)-Perilla alcohol, 348 Periodate diol cleavage, 614, 726 (1R,3R)-Permethrinic acid, 460–461 Peterson reaction, 726, 878 (⫺)-PF1163B, 704 8-Phenylmenthol, 684, 817 Asymmetric Michael additions with 8-phenylmethyl esters, 619 2 ⫹ 2 Photochemical cycloaddition, 81 Phyllanthoside, 212 Pleraplysilin-1, 326 Pinacol rearrangement, 14–15 Chemoselectivity, 14 Lewis acid mediated, 15, 20 Selectivity of secondary versus tertiary alcohols, 14 Synthesis of longifolone, 15 Pipoxide, 350–351 PNU-142721, 753–754 (⫺)-podorhizon, 621 Prins Reaction, 295, 721, 876 Double Prins reaction, 298 Lewis acid catalysed, 217, 297, 169 Oxo-ene mechanism, 297 Stereoselectivity, 298 Tetrahydropyran syntheisis, 296 Propranolol, 487–488, 528, 530 Synthesis using Sharpless asymmetric epoxidation, 647 Sharpless asymmetric dihydroxylation, 537 Prostacyclin (PGI2), 365 Isocarbocyclin, 365

905

906

Prostaglandins, 84 PGA2, 84 Pseudo-asymmetric centres, 394 Pseudomeconin, 124 (⫹)-Pumiliotoxin B, 864, 878 Pumiliotoxin C, 402, 818 Pumiliotoxin 251D, 191 Pummerer oxidation, 314 Purinol, 839 Pyrazolate, 855 Pyrazole, 29, 315, 751 Synthesis from 1,3-dicarbonyls, 29, 37, 141 Synthesis of allopurinol, 839 Pyrenophorin, 204, 207–208, 211 Pyridazines, 853 Synthesis, 853 Pyridine, 12, 13, 85, 770–773 Bakke nitration of, 773 Katritzky’s improvement of, 773 Chlorination of, 786 Electrophilic substitution of, 750 direct electrophilic substitution, 752 Hydroxylation, by ortho-lithiation, 781 Iodination of, 754 N-oxides, 555 electrophilic substitution of, 750, 752, 807 tandem lithiation and nucleophilic substitution, 764 ortho-Lithiation of, 759 Nitration of, 751 Regioselectivity in electrophilic substitution of, 753 Sulfonation of, 768 Synthesis of imidazo[4,5-c]pyridines, 769 Vicarious nucleophilic substitution of, 769 Quercus lactones, 201 Radical cyclisation, 478, 683 RAMP, 469, 601 Raney nickel, 621 Reduction of alkenes, 670 Ramipril, 477 Reduction, 165, 248, 341, 342, 422, 507, 652 See also specific reagents and substrates, 713 Asymmetric using aluminium hydrides, 507 using baker’s yeast, 652 using BINAL-H, 508 using boranes, 508–510 using DARVON-H, 507 of unsymmetrical ketones, 507, 652 anti-Selective, 533 of β-hydroxy ketones, 422 Evans’ reduction of hydroxy ketones, 422, 427

Index

Of azide with PPh3, 122, 235, 318–319 Of imines, 687 Borane, 299 Using DIBAL, 331, 589 Ester, 355 Evans 1,3-reduction, 670 Evans-Tishchenko reduction, 424 Using K-Selectride®, 721 Using L-Selectride®, 220, 411 Using LiAlH4, 429, 488, 507 Using NaBH4, 684, 688, 695 Of nitriles, 799 Using samarium(II) iodide, 82 syn-Selective, 120, 408 Of β-hydroxy ketones, 422, 424 Reductive amination, 402, 437 Reformatsky reaction, 427 Regiodivergent resolutions, 644 Remote induction, 704 1,4-Control by sigmatropic shifts, 707 1,4-syn Induction in the aldol reaction, 706 1,5-Induction in the aldol reaction, 708 Resolution, 435–446 Chiral chromatography, 446 Of a hydroxy-acid, 437 Of a hydroxy-amine, 438 Of an amino-acid, 438 Via covalent compounds, 439 Of diastereomers, 447 Synthesis of Jacobsen’s Mn(III) epoxidation catalyst, 444 With racemisation, 451 (R)-Reticuline, 545 Retinal, 160 Retinol, 173, 185 Ritter reaction, 492 Robinson annelation, 577 Organic catalysis of, 579 Roseophilin, 75 Rubrynolide, 9, 23 Rubottom oxidation, 777, 796 Ruthenium, 7, 243 Catalysed four component coupling, 864, 891 Catalysed hydrogenation, 572, 574 Catalysed metathesis, 243–245 Samarium(II) iodide, 82 See also Evans-Tishchenko reduction, 424 N-O bond cleavage, 837 Reductive removal of a sulfone, 321, 602 SAMP, 602, 623 Santalene, 177 Sceletium A-4, 872 Scopadulcin, 69

Index

Seebach, 55, 56, 600 Relay chiral units, 606 Hydroxy ester alkylation, 602–603 Synthon nomenclature, 55–56 Selenenyl-lactonisation, 294 Selenium dioxide, 785 Enol oxidation, 785 Enone synthesis, 256, 621 Oxidation of enols, 784 Senepoxide, 164 Serricornine, 409–410 Sertraline, 443 Shapiro reaction, 255, 258–261 Sharpless asymmetric aminohydroxylation, 552 Sharpless asymmetric dihydroxylation, 537 Advanced dihydroxylation strategy, 551–552 Diastereoselectivity, 547, 551 Catalytic cycle, 540 Ligands for, 543 recommended ligands, 527 Methanesulfonamide, 542 Mnemonic device, 542 Regioselectivity, 547 Solvent dependence, 540 Synthesis of aspicillin, 549 (⫺)-propranolol, 554 reticuline, 545 indicine, 543 Sharpless asymmetric epoxidation, 527, 647, 725, 730, 738 Catalyst structure, 529 Kinetic resolution using, 638 Ligands for, 543, 584 Mnemonic device, 530, 542 Propranolol synthesis, 530 Modification after, 531 Synthesis of (⫺)-propranolol, 530 (⫹)-disparlure, 532 Summary, 537 Shikimic acid, 350, 616 [1,5] Sigmatropic shift, 767 [1,5]H shift, 767 [1,5]NO2 shift, 767–768 [2,3] Sigmatropic rearrangements, 247, 810 See also Wittig rearrangement Of allylic phosphates, 192 Of allylic sulfoxides, 339, 343 In allylation of ketones, 345 Synthesis of allylic alcohol, 129 E-Alkene synthesis, 225, 227 [3,3] Sigmatropic rearrangement, 245, 823 See also Claisen and Cope rearrangements, 352 Of allylic sulfones, 186, 239

907

Of chromate esters, 341 Stereochemical transmission, 691 Tandem [3,3] sigmatropic rearrangements, 823 Silanes, 178 See silicon, 178 Silicon, 183 See also silanes and siyl enol ethers, 182, 184 Cation stabilisation of, 182 Silanes, 182 allyl silanes, 182 alkylations, 284 asymmetric addition to adehydes, 515 asymmetric reaction with electrophiles, 30, 39 cobalt-stabilised cations, 182 deprotonation of, 521–522 electrophilic attack, 180, 256 heterocyclic synthesis, 181 Michael addition, 197 reaction with acetals, 195 epoxides, 180 aldehydes, 180 SE2’ reaction of, 433 Synthesis by Wittig reaction, 482 tandem Beckmann rearrangement and allyl silane cyclisation, 889 Vinyl silanes, 197 ‘ate’ complexes from, 274 conversion to carbonyl compounds, 197 reaction with electrophiles, 212 synthesis, 180 tandem iminium ion formation and vinyl silane cyclisation, 880 E-vinyl silanes, 259 synthesis from alkynes, 261 Silyl enol ethers, 17, 20 O-Acylation of, 484 Aldehydes of, 16, 18–20 Aldol reaction of, 19–20 syn-selective, 47 anti-selective, 46 Addition to aldehydes, 19–20 Alkylation of, 16–17, 30 tertiary halides, 36 Conjugate addition of, 39 Diels-Alder reaction of, 52, 164 Lewis acid catalysed aldol reaction of, 169, 159 Mukaiyama aldol reaction, 32 Palladium(II) oxidation of, 283 Rubottom oxidation, 796 Thermodynamic formation of, 15, 16, 20 Synthesis under equilibrating conditions, 73, 75, 140 Simmons-Smith cyclopropanation reaction, 348

908

Sonagashira coupling, 754, 860 Stannanes, 265, 327 See Tin and Stille coupling, 325 Stereochemical analysis, 385 Stereochemical descriptors, 381 Stereogenic centre, definition, 43 Stereoselectivity, definition, 43 Stereospecificity, definition, 43 Sterpuric acid, 797 Stetter reaction, 204, 220 Stille coupling, 325 Double stille couplings, 328 Variations of, 604 Synthesis of β-carotene, 328 Synthesis of asperazine, 329 Strecker reaction, 199, 450 Asymmetric, 452 Streptazolin, 482 Strigol, 219 Strychnine, 320 Substitution reaction, 316, 778 SE2’ reaction, 433 SEAr reaction, see chapters 7, 32, 33 SNAr reaction, 484 SN1 reaction, 308–309, 376 SN2 reaction, 34, 38, 43 SN2’ reaction, 340 Sulfenyl-lactonisations, 294 SuperQuat, 618 Suzuki coupling, 329–335, 662, 669–670 Synthesis of milbemycin β3, 330–332 Synthesis of brevicomin, 332, 333 Swainsonine, 885 Swern oxidation, 63, 321, 473 Sydowic acid, 93 (⫺)-Syringolide, 668 Synthons, 56 a1, see chapters 3, 4, 5 a3, see chapter 9 d1, see chapter 14 d2, see chapter 10 d3, see chapter 11 Seebach nomenclature, 56 Tandem organic reactions, see chapter 36 Involving 1,3-dipolar cycloadditions, 248 Involving Heck reaction, 317 Involving Michael (conjugate) addition, 619 and aldol reaction, 667 of chiral amines and aldol reactions, 863 heterocycle synthesis, 887 ‘MIMIRC’ sequence, 871 use of sulfur, 875 Involving pericyclic reactions, 880

Index

[3,3] sigmatropic rearrangements, 34, 69 aza-Diels-Alder reaction, 819, 884, 888 Diels-Alder reaction, 24, 52, 157 aza-ene reaction, 885 ene reaction, 297, 882 Involving metallation, 887 Involving metathesis, 891 Iminium ion formation and vinyl silane cyclisation, 880 Polymerisation terminated by cyclisation, 870 Asymmetric synthesis of pipecolic acids, 876 (⫹)-pumiliotoxin, 878 Synthesis of tricyclic amine, 875 2-vinyl indoles, 886 Tautomerism, 848 In azoles, 848 Taxodione, 303 Terpenes, 472 Chiral auxiliary synthesis, 472 8-phenylmenthol synthesis, 619 Tetrazoles, 844 Michael addition of, 870 Synthesis by 1,3 dipolar cycloaddition, 842 Synthesis of zolterine, 844 Thermodynamic control In formation of cyclic compounds, 686 Of double bond geometry, 227, 703 Of enolisation, 30, 785, 799 Thiazole, 840 Carbonylation of, 857 Synthesis, 840 Thienamycin, 520 Thyroxine, 476 Tin, 183 Organo-tin (stannane) reagents vinyl stannanes, 265, 336 synthesis by Shapiro reaction, 259, 342 Z-vinyl stannanes, 265 synthesis from alkynes by hydroalumination, 271 trans-Di-axial ring opening, 414, 416 Transmetallation, 114, 323 1,2,4-triazoles, 848 Alkylation of, 848 U100766, 485–486 Umpolung, 56, 854 Vernolepin, 158, 291 Vertinolide, 168 ViagraTM (Sildenafil), 779 (⫹)-Vincamine, 454

Index

Vinyl alanes, 271 Vinyl anion equivalents, 255, 264 See chapter 16 Vinyl cation equivalents, 309–310 See chapter 18 Vinyl coupling, 313 Vinyl halides, 313 E-vinyl bromide, 332 from vinyl silane, 309 vinyl iodide, 260–261, 264 synthesis by Shapiro reaction, 258 E-vinyl iodides, 264, 318 synthesis from boronic acids, 330, 333 Vorbrüggen coupling, 850

E-selective reaction of stabilised phosphonate ylids, 232–233 Horner-Wittig reaction, 236–237 Intramolecular reaction of phosphonium ylids, 579 ‘MIMIRC’ sequence, 871 Schlosser modification, 234–235 Unsaturated carboxylic acid synthesis, 76 Z-selective, of phosphonium salts, 157, 185, 196 Wittig rearrangement, 247 See also [2,3] Sigmatropic rearrangement, 247 Wittig Still rearrangement, 247

Wacker oxidation, 283 Cyclopentannelation using, 284 Weinreb amides, 118 Reaction with organometallic reagents, 115, 195 Synthesis from esters, 196 Wittig reaction, 230 Alkene synthesis, 230 Diene synthesis, 232 E-selective enone synthesis, 232

Zimmermann-Traxler transition state, 45 Zinc, 568 Organo-zinc reagents, 568, 592 asymmetric addition to aldehydes, 592 Zirconium, 49 enolates, syn-selective aldol reaction, 49 Organo-zirconium reagents, 121 reaction with acid chlorides, 121, 181, 192 carbonylation of, 122

Yomogi alcohol, 344

909