Flavours and Fragrances (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Flavours and Fragrances Flavours and Fragrances Edited by Karl A. D. Swift Quest International, Ashford, Kent, UK W...

Author: A.D. Swift

518 downloads 2190 Views 11MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Flavours and Fragrances

Flavours and Fragrances Edited by

Karl A. D. Swift Quest International, Ashford, Kent, UK

WOo D H E A D

P uB LI s H I N G L I MI TE D

Cambridge England

Published by Woodhead Publishing Limited, Abington Hall, Abington, Cambridge CB 1 6AH, England www.woodheadpublishing.com The proceedings of the 1997 RSUSCI International Conference on Flavours and Fragrances held 30 April-2 May in Warwick, UK First published by The Royal Society of Chemistry 1997 Reprinted by Woodhead Publishing Limited 2005

0 Woodhead Publishing Ltd, 2005 The authors have asserted their moral rights This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but thc authors and the publishcr cannot assume responsibility for the validity of all materials. Neither the authors nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from the publisher. The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN- 13: 978-1-85573-780-8 ISBN-10: 1-85573-780-9 Printed in the United Kingdom by Lightning Source U K Ltd

Preface

This book is a compilation of the majority of the twenty one papers presented at the 1997 RSCfSCI flavours and fragrances conference at Scarman House, University of Warwick. The aim of the meeting was to bring together scientists from both industry and the academic world, who have a common interest in the chemistry of flavours and fragrances. The subject matter was intentionally broad, covering areas such as chemoreception, analytical techniques, essential oils, the synthesis of flavour and fragrance materials in the laboratory, clean efficient syntheses on a manufacturing scale, and the formation of flavours both in the cooking process and using biotechnology. The book is divided into the same sections as the original meeting. The meeting was spaced over two and a half days and saw speakers and delegates from all comers of the world exchanging ideas and information. Special thanks go to the Conference Secretariat Elaine Wellingham for helping me organise the conference on behalf of the Biological & Medicinal Chemistry Sector of the RSC and the Fine Chemicals Group of the SCI. The staff at Scarman House, University of Warwick, deserve a mention for their very professional and smooth running of such a wonderfid conference facility. Thank you also to all of the speakers and co-authors who firstly, submitted their manuscripts so promptly, and secondly, for co-operating with the editor! A very big thank you goes out to my team of proof readers, Jamie Mankee, Dave Munro, Lucy Swift, and Kim Yarwood. Their work is invaluable, and they probably know the manuscripts ‘off by heart’ by now! Finally, thank you to the RSC for agreeing to publish this book.

Contents

Chemoreception and Structure-Activity Relationships Toward a Rational Structure-Function Analysis of Odour Molecules: The Olfactory Receptor TM4 Domain Michael S. Singer, Gordon. M.Shepherd

3

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry Andreas Muheirn, Alex Hausler, Boris Schilling, Konrad Lerch

11

The Design and Synthesis of Novel Muguet Materials Karen J. Rossiter

21

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrance from the Skin Braja D.Mookerjee, Suba M. Patel, Robert W. Trenkle, Richard A . Wilson

36

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals Mikhail. Y. Gorbachov

48

Essential Oils and Analytical Derivatized Cyclodextrins in Enantiomer GC Separation of Volatiles Carlo Bicchi, A . D 'Arnato, V. Manzin

57

Production, Chemistry and Sensory Properties of Natural Isolates Mans H. Boelens

70

An Odour Sensing System for Use in Measuring Volatiles in Flavours and Fragrances Using QCM Junichi Ide, Takarnichi Nakarnoto, Toyosaka Moriizurni

87

The Aromatic Resins: Their Chemistry and Uses David A . Moyler, Robin A. Clery

96

...


Vlll

Studies Towards Structure Determination of Substituted Pyrazines Michael Zviely, Alexander Kern, Igal Gozlan, Ron Frim

116

Flavours Generation of Taste through (Redox) Biocatalysis Corja Laane, Ivonne Rieljens, Huub Haaker, Willem van Berkel

137

The Maillard Reaction in Flavour Formation Hugo Weenen, J. Kerler, J.G.M. van der Ven

153

Relationships between Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles AndyJ Taylor, R. S. T. Linforth

171

Organic Chemistry Synthesis and Odour Properties of Chiral Fragrance Chemicals Tetsuro Yamasaki

185

In Search of Nascent Musks ..... Or Not! Walter C. Frank

196

Synthesis and Application of Thiocarbonyl Compounds Shuichi Hayashi, S. Hashimoto, H. Kameoka, K. Sugimoto

209

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis Ivan V. Kozhevnikov

222

Subject Index

237

Chemoreception and Structure-ActivityRelationships

Toward a Rational Structure-Function Analysis of Odour Molecules: The Olfactory Receptor TM4 Domain Michael S. Singer and Gordon M. Shepherd SECTION OF NEUROBIOLOGY AND INTERDEPARTMENTAL NEUROSCIENCE PROGRAM, YALE UNIVERSITY SCHOOL OF MEDICINE, 236 FMB, 333 CEDAR STREET. NEW HAVEN, CT 06510,USA

1

INTRODUCTION

The family of olfactory receptor proteins (ORs) is currently the best candidate site for odour transduction in the vertebrate olfactory system. The first evidence for this was the identification of ORs as members of the G protein-coupled receptor (GPCR) superfamily, as predicted by an odour-induced adenylate cyclase activity”’ and supported by blockade of odour responses by the GPCR antagonists propranolol and a t r ~ p i n e Further .~ support came from estimates that mammals possess 1000 different OR subtypes, consistent with the capacity to detect many thousands of odour~.~-’ Direct evidence followed from two expression systems, which showed selective responses by rat subtype OR5 to two aldehydes of the lily of the valley class, lyral and lilia1.8‘9 Finally, computational analysis of the OR family has pointed to 5-10 amino acid residues likely to interact with odour molecules. Variations in these residues were postulated to account for different odour preferences across OR subtypes.”

-

Knowledge of ORs and their interactions with odour molecules obviously has important implications not only for basic research on olfactory transduction but also for practical applications in the fragrance industry. Given this potential importance, we have initiated studies of the mechanisms of odour molecule-OR interactions at the molecular level, using new advances in computational analysis of gene families. This paper focuses on OR proteins at the amino-acid level, as the basis for developing a rational structurefunction analysis of odour molecules. It is oriented toward organic chemists and perfhers in order to illustrate how data on ORs may inform the practice of fragrance and flavor chemistry. The OR fourth transmembrane domain (TM4), which has attracted interest in several studies, serves as a model for this discussion.

2 THE STRUCTURAL MOTIF OF G PROTEIN-COUPLED RECEPTORS Figure 1 shows the predicted tertiary structure of OR proteins, which is likely to be conserved across different species and subtypes. Like other GPCRS, ORs are predicted to have seven a-helical transmembrane domains (TMs), identified as hydrophobic spans of

4

Flnvours and Fragrances

18-25 amino acids. The amino terminus is extracellular, and the carboxyl terminus is intracellular," Conserved cysteines in the first (El) and second (E2) extracellular loops are likely to form a disulfide bond. Molecular models, correlated mutation analysis, and positive selection moments have pointed to an odour-binding pocket (A) formed by TM3, TM4, TM5 and TM6 (reviewed in ref. 10). The third intracellular loop (B), between TM5 and TM6, is believed to be the principal site of G protein activation. The extracellular loops (C) may serve multiple functions: binding odour molecules; docking olfactory binding proteins, or guidin axons of olfactory receptor neurons to their glomerular 8-13 targets in the olfactory bulb.

3 SEQUENCE DIVERSITY: VARIATIONS ON THE GPCR MOTIF Prior to the identification of ORs, several workers reasoned that amino acids in the odour binding site would vary across subtypes, in order to accommodate diverse odour molecule^.^^'^ This prediction was fulfilled by TM3, TM4, and TM5 of the rat ORs isolated by Buck and Axel: and yielded the first clue that TM4 may interact with odour molecules. Ben-Arie et a[.'' noted diversity in TM4 and TM5 of human ORs, labelling these domains and the loop between them (E2) an OR hypervariable region.I6 Both observations compared favorably with the related P-adrenergic receptor (PAR), where an ._ TM5 and TM6." agonist binding

Figure 1 Schematic model of an olfactory receptor, with extracellular surface at top. Transmembrane a-helices are shown as numbered cylinders, interhelical loops as lines. Three functional regions are marked: (A) predicted odour-binding pocket,(B) predicted site for G protein coupling, and (C) extra-cellular loops. Asterisk (*) indicates predicted disulfide bond between extracellular loops.

Toward a Rational Structure-FunctionAnalysis of Odour Molecules

5

4 POSITIVE DARWINIAN SELECTION Sequence diversity, as cited above, is dificult to interpret due to random mutations, which accumulate in noncritical areas of proteins. However, the significance of this diversity can be tested more rigorously at the nucleic acid level. This is carried out by measure of positive Darwinian selection, an evolutionary process which favors amino acid diversification over conservation. Under positive Darwinian selection, the relative rate of nonsynonymous mutations, which encode different amino acids, out aces the ?'-I9 This relative rate of synonymous mutations, which encode the same amino acid. process occurs in at least two classes of immune system molecules, where it is notably restricted to the antigen recognition sites. Ngai et of.*' reasoned that by analogy with immune system molecules, ORs would show positive Darwinian selection restricted to the odour binding site. Their analysis of ORs from the catfish subfamily 32 pointed to TM3 and TM4 as domains likely to interact with odour molecules.

5 MOLECULAR MODELS Evidence that rat subtype OR5 is activated preferentially by lyral or lilial' prompted us to build three-dimensional models of OR5 and other subtypes.2' These were based on bacteriorhodopsin coordinates,22 which have served as a template for several GPCR model^.^^-^' The models were further constrained by methods which predict the limits and relative rotations of transmembrane cr-heli~es.~'-~~ The OR5 model enabled us to search for candidate odour-binding sites by calculating interaction energies for a panel of odour ligands at various locations and orientations in the protein structure, a process referred to as docking. Lyral showed optimal interactions in a pocket bounded by TM3, TM4, TM5, TM6 and TM7, which showed considerable similarities with the PAR binding pocket. Three critical residues were identified in TM4: histidine 155, alanine 156, and histidine 159. Notably, these three residues were situated on the sanie side of the TM4 helical structure, near the extracellular part of the receptor. The importance of the TM4 residues, particularly His 155 and His 159, has been supported by subsequent models. Bajgrowicz and Broger2' built another model of OR5 based on an earlier PAR model, employing different modelling and docking software. The binding pocket was similar to that of Singer and Shepherd2' and indicated important roles for both His 155 and His 159. 6 CORRELATED MUTATION ANALYSIS The limitations of computer models and well documented difficulties of GPCR structure prediction and determination have prompted efforts to correlate protein sequence and receptor function directly, without the intermediate step of structural informati~n.~''~~ These efforts include correlated mutation analysis (CMA),28"2 which has proven useful for GPCRs such as opsins and muscarinic acetylcholine re~eptors.~'CMA compares several subtypes from the same receptor family, scanning the sequences for pairs of residues which remain conserved or mutate in tandem. Residues which behave in this way are said to show correlated mutations. They are believed to be functionally or structurally interdependent, based on the theory that one mutation compensated for the


6

deleterious effects of the other (see review in ref. 10). When we applied the method to rat OR sequences, 13 out of 310 residues were significantly correlated.33s34Most of these residues were concentrated in the upper half of TM4, TM5 and TM6, consistent with the predictions of Buck and Axel4 and Ben-Arie et al.," as well as the PAR binding pocket." The correlated residues also matched positively selected residues in the catfish sequences of Ngai er aZ.20Furthermore, there was a striking correspondence between the CMA data and the previously published computer models. The three residues identified by CMA in TM4 were exactly those found in the OR5 model: His 155, Ala 156, and His 159. Thus, evidence for the importance of these residues converged from several directions: amino acid hypervariability, overlap with the PAR binding pocket, positive Darwinian selection, two independent structural models, and the structure independent method of CMA. 7 HELICAL WHEEL STRUCTURE The secondary structure of TM4 can be modelled as a canonical or-helix (see ref. 35). Two important features of TM4 can be seen in the helical wheel diagram in Figure 2, which shows the orientations of residue side chains viewed en face, from outside the cell. His 155 and His 159 point approximately in the same direction as Ser 148 and Thr 152. The side chain polarity of all four residues, as well as possible cationic states in the histidines, provides strong evidence that they are oriented toward the receptor cleft (cf. refs. 36, 37), and probably into an odour binding site (cf. ref. 38). In contrast, other residues in TM4 are decidedly hydrophobic, consistent with placement in the lipid or helical interface. Exceptions are Ser 143, whose near cytosolic location may relieve it from the constraints of membrane hydrophobicity, and Thr 153, an amino acid which is known to be stable at helical interface^.^^ Thus, canonical (a-helix modelling of TM4) supports the hypothesis that His 155

Figure 2 Helical wheel diagram of TM4, viewed from outside the cell. Amino acids are based on rat subrype OR5 (ref: 8). Circles denote polar amino acids: filled circles, histidines; open circles, serines and threonines. Dotted lines indicate nonpolar amino acids. Note similar orientations ofpolar residues.

and His 159 could contribute to an odour-binding pocket. Simulations of a-helical domain interactions have recently been carried out with reasonable accuracy, as shown for glycophorin A.40 This approach may render it feasible to model interactions between odour molecules and TM4 at the atomic level.

Toward a Rational Structure-Function Analysis of Odour Molecules

I

Crabtree4' cited evidence that mercaptans, characteristically detected at low thresholds, bind to a transition metal center such as Cu(1). The arrangement of the histidines suggests that they could help to coordinate such an ion. 8 DUAL-HISTIDINE SUBSITE AND ODOUR DETERMINANTS

The biophysical properties of histidine make it possible to predict how His 155 and His 159 might bind odour molecules. The side chain of histidine contains an imidazole ring, capable of forming hydrogen bonds, electrostatic interactions, and van der Waals interactions. The imidazole ring shifts readily between protonated and deprotonated states due to its pK of 6.5, very near physiological pH. These properties make histidine quite versatile in its range of possible non-covalent bonds. Two histidines, such as His 155 and His 159, may also act cooperatively. This model is consistent with the predictions of Kosower,I4 who listed several pairs of amino acids postulated to interact with particular functional groups on odour ligands. Two histidines, for instance, were predicted to interact best with aldehyde or nitrile groups. We have further postulated that the "dualhistidine subsite" on ORs can bind an array of different polar groups or nonpolar regions with graded affinities, as shown in Figure 3. Collectively, these polar group and nonpolar regions may be called determinants. In this view, most odour molecules can be defined as an assembly of 2-5 determinants in a specific geometric arrangement. 9 SUBSITE-DETERMINANT MODEL

The consensus model for the odour-binding pocket of OR5, which has arisen out of the computational approaches described above, is shown in Figure 4. The dual histidine subsite can be seen on TM4. At least 3 more potential subsites were identified: a dual serine subsite (Ser 246 and Ser 249) on TM6; Ile 199 and Phe 206 on TM5; and Phe 322 on TM3.I' We postulate that each of these subsites binds specific determinants over an affinity spectrum as shown for the dual histidine subsite in Figure 3. The dual serine subsite, for instance, contains two hydroxyl groups predicted to be 5A apart, and may thus bind hydroxyl groups and mines with high affinity, via two hydrogen bonds.38 The phenylalanine and isoleucine residues may form surface-specific van der Waals forces or n-n orbital interactions with alkyl or phenyl determinants. These considerations help to define the fundamental biophysical units that mediate OR activation, which appear to be the elementary interactions between the determinants in an odour molecule and the subsites in the OR pocket.' The overall interaction between an odour molecule and an OR would then represent the sum of complementary interactions between 2-5 subsites and 2-5 determinants. Variations in the amino acids at each subsite would modulate preferences for particular determinants; for example, dual histidines for aldehyde, dual serines for hydroxyl, phenylalanines for phenyl, and aliphatic amino acids for alkyl determinants. The relative positions of subsites in the pocket, constrained by the protein structure, would provide the basis for discrimination of isomers and enantiomers. Steric hindrance and electrostatic repulsion would also shape OR specificity.

a


alaenyde R-CHO

alkane R-CH3

W

I

deterrnmanl

Figure 3 Potential interactions for difSerent Jirnctional groups (determinants) at the dualhistidine subsite (His 155 and His 159). Above, the dipole moment of the aldehyde carbonyl bond would allow for interaction with histidines as shown. Lower affinities would apply for an 0-H bond (hydroxyl) and N-H bond (amine) based on smaller dipole moments. The C-H bond of the alkane would not interact due to its lack of dipole moment. Below, hypothetical summary of the subsite afjinities for several determinanfs. Orher subsires would have drfferent aflnity spectra. The overall odour specificity of the receptor would represent a function of ihese spectra and the relative locations ofthe determinants.

Figure 4 End face view of the predicted odour-binding pocket in the OR5 receptor, based on the studies reviewed in this paper.

Toward a Rational Structure-Function Analysis of Odour Molecules

9

We postulate that the possible permutations of subsites in -1000 receptors would be sufficient to encompass the vast odour space that mammals, including humans, can discriminate. An important feature of our model is that ORs show specificity for combinations of determinants, rather than individual determinants or particular odours. The model should provide new insights into the biophysical mechanisms underlying odour perception, and help to guide fragrance and flavor chemists to a new level of structure-function analysis.

Acknowledgments We are grateful to Charles Greer, Emmanouil Skoufos, Donald Engelman, Robert Crabtree, Laerte Oliveira, Bob Bywater, Wilma Kuipers, Chris Sander, and Heinz Breer for helpful discussions, and to Gerrit Vriend for discussions and WHATIF software. This work was supported by grants to GMS from NIDCD and NASA, NINM and NIDCD through the Human Brain Project. MSS is supported by the Yale University MSTR References 1. U. Pace, E. Hansky, Y Salomon and D. Lancet D, Nature, 1985,316,255-258. 2. R. Reed, Neuron, 1992,8,205-209. 3. S . Firestein and G. M. Shepherd, NeuroReport, 1992; 3,661-664. 4. L. BuckandR. Axel, Cell, 1991,65,175-187. 5. D. Lancet, Ann. Rev. Neurosci., 1986,9,329-355. 6. G. M. Shepherd and S . Firestein, J SteroidBiochem. Molec. Biol., 1991,39,583-592. 7. D. Lancet, E. Sadovsky and E. Seidelmann, Proc. Natl. Acad. Sci., 1993,90,37153719. 8. K. Raming, J. Krieger, J. Strotrnann, L. Boekhoff, S. Kubick, C. Baumstark and H. Breer, Nature, 1993,361,353-356. 9. H. Kiefer, J. Krieger, J. D. Olszewski, G. Von Heijne, G. D. Prestwich and H. Breer, Biochem., 1996, 35, 16077- 16084. 10. G. M. Shepherd, M. S. Singer and C. A. Greer, Neuroscientist, 1996, 1,262-271. 1 1. U. Gat, E. Nekrasova, D. Lancet and M. Natochin, Eur. J. Biochem., 1994,225,11571168. 12. A. L. Hughes and M. K. Hughes, J. Molec. Evol., 1993,36,249-254. 13. M. S. Singer, G. M. Shepherd and C. A. Greer CA, Nature, 1995,377, 19-20. 14. E. M. Kosower, 'Molecular Mechanisms for Sensory Signals: Recognition and transformation', Princeton UP, Princeton, NJ, 1991. 15. N. Ben-Arie, D. Lancet, C. Taylor, M. Khen, N. Walker, D. H. Ledbetter, R. C a n T o m , K. Patel, D. Sheer, H. Lehrach and M. A. North, Hum. Mol. Gen., 1993,3, 229-235. 16. D. Lancet and N. Ben-Arie, Curr. Biol., 1994, 3,668-674. 17. C. D. S trader, I. S . Sigal and R. A. F. Dixon., FASEB J., 1989; 3, 1825-1832. 18. A. L. Hughes and M. Nei, Proc. Nail. Acad. Sci. USA, 1989; 86,958-962. 19. A. L. Hughes and M. Nei, Nature, 1988,335, 167-170.

10

Flavoitrs and Fragrances

20. J. Ngai, M. M. Dowling, L. Buck, R. Axel and A. Chess, Cell, 1993,72,657-666. 21. M. S. Singer and G. M. Shepherd, NeuroReport, 1994, 5, 1297-300. 22. R. Henderson, J. M. Baldwin, T. A. Ceska, F. Zemlin, E. Beckmann and K. H. Downing, Mol. Biol., 1990,213, 899-929. 23. S. Trumpp-Kallmeyer, J. Hoflack, A. Bruinvels and M. Hibert, J. Med. Chem., 1992, 35,3448-3462. 24. M. F. Hibert, S. Trumpp-Kallmeyer, A. Briunvels and J. Hoflack, Mol. Pharmacol., 1991,40,8-15. 25. D. M. Engelman, T. A. Steitz and A. Goldman, Ann. Rev. Biophys. Chem., 1986, 15, 321-353. 26. P. Cronet, C. Sander and G. Vriend, Prot. Eng., 1993,6,59-64. 27. J. Bajgrowiez and C. Broger, in 'Flavours, Fragrances and Essential Oils. Proceedings of the 13th International Congress of Flavours, Frangrances and Essential Oils', ed. K. H. C. Baser, AREP, Istanbul, 1995. 28. L. Oliveira, A. C. M. Paiva, G. Vriend, J. Comp. Aid. Molec. Des., 1993; 7,649658. 29. W. Kuipers, L. Oliveira, A. C. M. Paiva, F. Rippmann, C. Sander, G. Vriend, C. G. Krus, 1. van Wijngaarden, and A. R 1 Jzerman, in 'Membrane Protein Models: Experiment, theory, and speculation', ed. J. Findlay, BIOS, Oxford, 1 996. 30. L. M. Gregoret and R. T. S auer, Proc. Natl. A cad. Sci. USA, 1993,90,4246-4250. 31. U. Gdbel, C. Sander, R. Schneider and A. Valencia. Prot. Struct. Func. Gen., 1994, 18,309-3 17. 32. E. Neher, Proc. Natl. Acad Sci. USA, 91, 1994,98-102. 33. M. S. Singer, L. Oliveira, G. Vriend and G. M. Shepherd, Receptors and Channels, 1995,3,89-95. 34. G. Vriend, J. Mol. Graph., 1990, 8, 52-56. 35. C. Chothia, Ann. Rev. Biochem., 53, 1984,537-512. 36. D. Donnelly, J. P. Overington, S. V. Ruffle, J. H. A. Nugent and T. L. Blundell, Protein Sci., 1993,2,55-70. 37. D. Donnelly, J. B. C. Findlay and T. L. Blundell, Receptors and Channels, 1994, 2, 61-78. 38. M. S. Singer, Y. Weisinger-Lewin, D. Lancet and G. M. Shepherd, Receptors and Channels, 1996,4, 141-147. 39. K. R. MacKenzie, J. H. Prestegard and D. M. Engelman, Science, 1997,276,131-133. 40. P. D. Adams, D. M. Engelman and A. T. Briinger, Prot. Struct. Func. Gen., 1996, 26,257-261. 4 1 . R. Crabtree,J. Inorg. Nucl. Chem., 1977,40, 1453. 42. C. D. Strader, T. M. Fong, M. P. Grazianoand, M. R. Tota, FASEBJ., 1995,9,745754.

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry Andreas Muheim, Alex Hausler, Boris Schilling and Konrad Lerch* GIVAUDAN-ROW RESEARCH LTD., 8600 DUBENDORF, SWITZERLAND

1 INTRODUCTION Biotechnology, of which recombinant DNA technology is an important sub-discipline, has a long tradition in the production of food and flavours. Man first started to apply microbes around 3500 BC for the production of wine, beer, bread and many other food articles that became an indispensable part of our daily diet’. At the beginning such fermentations were carried out on a rather empirical level and only in the 19th century the scientific basis was laid by the discoveries of Louis Pasteur. Isolation and controlled cultivation of microbes became possible and about 20 years a o these techniques also found application in the K production of various flavour chemicals . In the early 1970’s, recombinant DNA technology emerged and soon started to become a significant part of today’s biotechnology. Immediate impacts of this new technology were observed in pharmaceutical research. Consequently, the first genetically engineered product, human insulin produced by bacteria, entered the market in 1982. Since then, more than 33 new drugs produced with recombinant DNA technology have been registered worldwide. In addition, 284 biotech drugs were in development in 1996, representing a three-fold increase since 1 9893. A similar change has been initiated in the food industry by the introduction of the FLAVRSAVR tomato in 1994 as the first genetically engineered whole food. Today, recombinant DNA technology has definitely found its way into the food industry, underlined by more than 3,600 transgenic field trials carried out by 1995. So far, 18 genetically engineered agricultural products have been approved for commercialization4. These include plants such as corn, cotton, soybeans and potatoes with improved pathogedpest resistance, herbicide tolerance and food quality’. The application of recombinant DNA technology in the flavour and fragrance industries is less advanced than in the pharmaceutical and food industries. Nevertheless, first products involving recombinant DNA technology in one or the other way have been commercialized. Today, recombinant DNA technology has also become an important part of the research activities of flavour and fragrance companies. The following chapters

12


illustrate areas of the flavour and fragrance industry that will be increasingly influenced by the use of recombinant DNA technology.

2 PRODUCTION OF AROMA CHEMICALS

2.1 Natural aroma chemicals Nature is a rich source of aroma chemicals of which several thousand have been identified and chemically synthesized. With the on-going trend towards natural flavours, aroma chemicals were increasingly required to be of natural origin. Separation techniques such as extraction and distillation of natural materials are successfully used in our industry. When these could not be achieved at economic costs, enzymatic or microbial conversions are used instead. Today, fermentative processes are employed to produce many aroma chemicals such as various aliphatic and aromatic acids (e.g. 2-methylbutyric acid and phenylacetic acid), different esters (e.g. ethyl-2-methybutyrate, methylanthranilate) and lactones (e.g. y- and &decalactones). For the production of high impact chemicals such as P-damascenone, methional or ynonalactone none of the above mentioned techniques could be reasonably applied so far. These chemicals are generally found in very small quantities in plant materials making their recovery an expensive endeavour. As no microbial or enzymatic conversions are evident, other approaches are needed. In nature, these aroma chemicals are formed by specific yet sometimes unknown pathways. However, some have been recently elucidated, as is the case for hraneol and yn~nalactone~.’.With the help of recombinant DNA technology the corresponding genetic information from the original source can be isolated and subsequently transferred into a suitable host strain. This allows an efficient microbial production of natural aroma chemicals. An application where we have explored this strategy is the production of cis-3hexenol*. Natural cis-3-hexeno1, also referred to as leaf alcohol, and its esters are of high demand as they are widely used in various fruit flavours. Traditionally, cis-3-hexenol is isolated from mint terpene fractions. In the plant, cis-3-hexenol is formed from linolenic acid via the hydroperoxide and cis-3-hexenal (Figure 1). As peppermint oil fractions could not satisfy the global need for natural cis-3-hexeno1, an enzymatic route starting out from linolenic acid was established’. The fatty acid is oxidized to the hydroperoxide using e.g. soya flour containing lipoxygenase. The conversion of the hydroperoxide to cis-3-hexenal is achieved by using a fruit source such as guava that was found to contain high activities of the hydroperoxide lyase. Reduction to the leaf alcohol is finally performed by yeast cells. The drawback of this enzymatic transformation, though independent of peppermint oil fractions, is the rather large amount of fruit that has to be processed. The lyase was shown to be the rate limiting factor in the enzymatic conversion of linolenic acid to cis-3-hexenol.

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry

noc-

_

-

13

Linolenic add

O2

1

Lipoxygenase

OOH

i

Hydroperoxide lyase

0

["'I -o n

j

Alcohol dehydrogenase

t

ur3-hexenol

Figure 1 Biosynthetic pathway of cis-3-hexenol in plants. Three enzymes are involved in the production of cis-3-hexenol starting from linolenic acid. Various plant materials serve as sources for these enzymes allowing a conversion of the acid to cis-3-hexenol at industrial scale. As in such a reconstituted production system the hydroperoxide lyase is the rate-limiting factor, the gene. coding for this enzyme was heterologously expressed in yeast cells to yield a highly active lyase material.

We have therefore purified the linolenic-acid-hydroperoxide-lyase from banana plants, allowing the determination of four independent, internal amino acid sequences. Degenerate oligonucleotides and resulting PCR-fragments helped to isolate the structural gene for the lyase from a banana cDNA library''. Interestingly, the DNA sequence shared 44% identity with the sequence of allene oxide synthase. The latter enzyme is important in the biosynthesis of methyl jasmonate, another important key flavour impact chemical. The heterologous expression of the lyase gene helped to overcome the drawbacks of the enzymatic production route supplying highly active lyase material. As can be seen in Figure 2, higher amounts of cis-3-hexenol have been produced in the presence of the recombinant yeast cells compared to homogenized bananas. To unify all three enzymes involved in the formation of cis-3-hexenol in yeast, we also cloned and coexpressed the lipoxygenase gene. This generated an even more efficient system to produce cis-3-hexenol. In addition, it should be pointed out that lipoxygenases and lyases with different specificities have been described that are involved in the degradation of fatty acids". Heterologous expression of such genes would allow the production of other important flavour chemicals such as 1-octene-3-01 or 2,6-nonadienal.

14


a :

b

.I

I:

Figure 2 GC analysis of the reaction mixtures in which the hydroperoxide of linolenic acid was incubated with a) homogenized banana material and reducing yeast cells and b) with recombinant yeast expressing the lyase gene and additional reducing yeast cells. The recombinant yeast process yielded higher amounts of cis-3-hexenol with no formation of trans-2-hexenol as side product.

Finally, applications based on our cloning of the lyase gene are not limited to microbial systems. The gene could also be transferred into plant hosts resulting in an increased formation of cis-3-hexenol/ cis-3-hexenal upon maceration of the plant. Ultimately, recombinant DNA technology could be used to enhance and to alter the flavour profile of fruits and plants by overexpressing key metabolic enzymes. As examples, strawberries high in furaneol or especially green smelling apples in which the lyase gene would be overexpressed can be imagined.

2.2 Tasty peptides Tasty peptides have been found in various food products such as meat, cheese, fish and y o g h ~ r t s l * ' ~In~ .order to improve or to boost flavours with such specific tasty peptides, they were so far synthesized either chemically or enzymatically. Both strategies, however, are not feasible for a large-scale and commercial production of tasty peptides due to high production costs14. We have therefore investigated the heterologous production of peptides using recombinant yeast strains. Several model peptides were chosen, among them an octapeptide known as beefy meaty peptide (BMP). This peptide was found in meat and was suggested to enhance the taste of beef gravy". Figure 3 shows the recombinant DNA technology approach used to produce the BMP in yeast cells. Linking the genetic information of the octapeptide to the yeast mating pheromone a-factor in a suitable expression vector allowed the secretion of the desired peptide into the culture medium. From the culture filtrate, the peptide can be easily recovered and used in a semi-

The Impact of Recombinant DNA-Technology on rhe Flavour and Fragrance Industry

15

purified form. Alternatively, intracellular accumulation of the peptide offers the possibility 16 to generate specially flavoured yeast extracts . This approach can be seen as a further improvement of yeast strains that have previously been engineered to contain high content of 5’-nucleotides (IMP, GMP).

Meat (flavor)

Peptides

L

/Yeast expression\

Purification

5 ’ -GAAGCTGAAGCTAAGGGTGACGAAGAATffTTGGCTTGA3

. . . . .a-factor..LysGlyAspGluGluSerLeuAla

Figure 3 Protease digestion of food proteins results in the formation of tasty peptides. Their organoleptic character and amino acid sequence can be determined after purification. As a new production strategy to obtain such peptides, yeast cells were transformed with an engineered yeast secretion vector. Upon induction, these cells started the synthesis of a fusion protein consisting of the desired peptide (in our case BMP) and the yeast a-factor. The desired peptide was then cleaved during the secretion process and could be easily recovered from the culture fluid in a semi-purified form.

3 PLANT ENGINEERING 3.1 Improvement of the flavour and fragrance profiles Plants are a major part of our daily diet and due to their smell and taste, .are established sources for raw materials used in the flavour and fragrance industry. More than 3,000 different essential oils have been analyzed and many of them are utilized in the creation of fine fragrances or serve as starting materials for the isolation and modification of chemicals”. As an illustration, 36,000 metric tons of d-limonene are extracted annually from citrus oils’*. Other commercially used examples include 1-carvone, geraniol and also menthol, the latter with an annual sales volume of about 2 billion USD. The way plants are nowadays industrially improved was dramatically changed by the new possibilities of using recombinant DNA technology. Presently, huge efforts are undertaken, for example, to increase the content and quality of fatty acids in oil crop

16


plants". Genetically engineered rapeseed underwent most field trials after potato, and in 1995 an engineered canola crop with high laurate was commercialized20*z'. Commercial examples of genetically engineered plants used in the flavour and fragrance industry are not yet known, but the example of a transgenic Pelargonium plant, commonly referred to as lemon geranium, can illustrate the potential of recombinant DNA technology22. In this example, the titre of geraniol was increased 4-fold and that of citronellol by 13-fold in the transgenic plant as compared to the wild-type. To optimally design such higher yielding plant species, an improved understanding of metabolic pathways as well as of the post-harvest biochemical reactions are required. DNA sequencing programs elucidating total plant genomes are expected to simplify the cloning of important gene sequences. 3.2 Safety aspects

The safety of genetically modified organisms has been assessed by the FDA23and the EC. Nevertheless, the public acceptance of products from transgenic plants is still rather low. This is especially the case in Western Europe where major concerns are expressed in Germany and Austria. On the other hand, little consumer reactions to genetically engineered food products have been observed in the USA. However, a recent European study showed that genetic engineering was ranked as a potential food risk similar to that of artificial food coloring and, interestingly, much safer than food irradiation or pesticide residuesz4.

4 NEW ENZYMES

The majority of industrial enzymes are used today in food preparations and in fabric care products. Both markets represent roughly a 160 million USD turnover annually2s. Therefore development of new enzymes is targeted mostly at these two segments. These enzymes also find limited applications in the flavour and fragrance industry. Examples include proteases for the peneration of food protein hydrolysates and lipases for the production of natural esters 6'27. In the past, many enzymes involved in the generation of flavours or flavour precursors have been characterized. It is a well known fact that during the post-mortem aging various hydrolyzing enzymes are released within the meat. This results in the formation of flavour precursors that are characteristic to the type of mea3'. The use of such enzymes in the flavour industry is limited as they are not available at reasonable costs. For the time being it does not appear that enzyme manufacturers will produce them due to the rather small market. The advent of recombinant DNA technology, however, has now great1 facilitated their large scale production rendering it feasible for flavour companies as welJ9. With the availability of such enzymes more authentic meat, cheese and other flavour mixtures could be generated. A first development in this direction can be seen in the area of enzyme-modified cheese. Treating milk proteins with commercially available proteases often results in bitter products. Though the occurrence of the bitter peptides has been extensively studied, screening of various commercially available proteases and mixtures thereof was needed to prepare pleasant, non-bitter flavours. As cheese flavours are the product of microbial


17

activities, extracellular enzymes from various starter cultures have been characterized. Interestingly, some of them were shown to have debittering activities3'. Such proteases removing off-flavours are of great interest to flavour industries which have started to clone and express the corresponding genes3'. This opens new avenues to making for example cheese flavours more characteristic and intensive. In recent years, new classes of enzymes such as oxidases (peroxidases and polyphenoloxidases) have been introduced for food and detergent applications. Oxidases are also important for the production of many different flavours. As an example, the oxidative degradation of amino acids yields various flavour aldehydes. L-amino acid oxidase (LAO) deaminates various amino acids, resulting in the formation of the corresponding keto acids which after decarboxylation yield flavour aldehydes (Figure 4). We have recently cloned the LAO gene from the filamentous fungus Neurospora crassa and overexpressed it homologously in the parent host3*. Alternatively, flavour aldehydes can also be produced by decarboxylation of amino acids and deamination by a rnonoamine oxidase. We have purified a novel monoamine oxidase (MAO) from Aspergillus niger and cloned the structural gene33. This FAD containing enzyme oxidizes various amines such as phenethylamine and methylthiopropylamine to the corresponding aldehydes. The gene coding for M A 0 was heterologously expressed in Escherichia coli. Incubating the above mentioned amines with protein extracts of such induced E. coli cells resulted in the formation of rnethional and phenylacetaldehyde. The broad substrate specificity makes this enzyme also attractive for the generation of various other flavour aldehydes. L- Amino Acid R NH2-CH-COOH

4

-7-

H202 + NH3

I

C-COOH

I1 I1 0

NH2-cH2

H20 + 0 2

LAO

R

I

coz

(L-Amino Acid Oxidase)

4

MA0 (Monoamine Oxidase

H202 + NH3

Decarboxylase

R

I -T b

Keto Acid

Figure 4

R

Decarboxylase

I

H20 + 0 2

Amine

Flavour aldehydes such enzymatically formed from for LAO and M A 0 have allowing the formation of precursor.

CHO

Flavour Aldehyde

as methional or phenylacetaldehyde were the corresponding amino acid. The genes coding been functionally expressed in microbial hosts flavour aldehydes when fed with the required

18


5 MOLECULAR OLFACTION AND TASTE A rather high impact of recombinant DNA technology can be expected in the field of molecular olfaction and taste. The first putative olfactory receptors were cloned in 199134. Since then the understanding of olfactory receptors and their signal transduction mechanisms has been drastically increased. It became widely accepted that olfactory receptors belong to the G-protein coupled seven transmembrane receptor family3’ that represent 60% of the targets for all drugs sold today. A possible interaction of odorants with a heterologously expressed mammalian receptor has been suggested in the case of lilialTMand lyralTM36. Such ligand-receptor models form a broad and scientific basis for the pharmaceutical industry to find new drugs targeting diseases such as AIDS, cancer or arthritis. Research in the field of molecular olfaction and taste has benefited a lot from the techniques and know-how developed for the discovery and screening of pharmaceutical drugs. Structureodor relationship studies, for example, have been widely applied in the search for novel fragrance molecules37338,but the lack of a three-dimensional structure of an olfactory receptor, so far, has hampered the efforts to model and study the odorant-receptor interactions. Thus, many recent efforts have been directed towards functional expression of olfactory receptors. A specific receptor can be expressed at the surface of a cell (Figure 5). These receptors are linked via ordinary signal transduction mechanisms to a reporter gene signaling receptor binding of a potential odorant. Such engineered screening systems are widely used in the pharmaceutical industry to test low molecular weight drugs.

0 0,

oaa o.

odorant

signal

enzyme

binding

transdudion

activity

Figure 5 Recombinant yeast cells expressing olfactory receptors at the cell surface can be used to screen a mixture of odorants. Upon specific odorant binding the intracellular signal transduction cascade is activated stimulating a reporter gene. The resulting enzyme activity correlates to the binding strength of the odorant and can be easily monitored. In contrast to olfaction, much less is known about taste receptors. Nevertheless, a simplified screening system for bitter and sweet compounds has been established. RuizAvila er 0 1 . ~isolated ~ receptors from the bovine tongue papillae, added recombinant Gprotein gustducin or transducin and incubated this reconstituted tongue receptor system in the presence of GTP-y-S and potential tastants. Trypsin digestion followed by Western blot analysis indicated if interaction between tastant and receptor occurred. This system which


19

so far is only applicable to bovine and not to human, not only allows the molecular screening of potent tastant, but more interestingly also of taste enhancing or blocking agents. In summary, screening system based on gustatory and olfactory receptors are feasible and will certainly become part of future investigations in flavour and fragrance industries. 6 OUTLOOK The flavour and fragrance industry is now at the point where the pharmaceutical industry was 20 years ago with respect to recombinant DNA technology. At that time, recombinant DNA technology entered the pharmaceutical research without much notice. The benefit of this technology was then clearly seen with the rather sudden emergence of first products and has thus developed to become an essential part of the research and production of new drugs. However, it can be foreseen that the flavour and fragrance industry will go through a similar phase until the commercial benefit of recombinant DNA technology is clearly recognized. It is already evident that recombinant DNA technology will become an important tool for the discovery and production of cheaper and novel flavour and fragrance chemicals. Furthermore, the technology is essential to ultimately advance our understanding of olfaction and taste. This will lead to the discovery of novel odor and taste modifying compounds, changing the way flavour and fragrance compositions will be formulated in the future. References

‘ ’ ’ * 10

11 12

13

14

‘5

16

Praeve P. e t a / . ,In: Fundamentals of biotechnology, Weinheim; Deerfield Beach, FL , 1987,1. Janssens L. eta/., Production of flavours by microorganisms, Proc. Biochem. 1992,27, 195. Facts & Figures, PhRMA facts April 1997. James C. and Krattiger A.F., Global review of the field testicg and commercialization of transgenic plants. ISAA Briefs No. 1 lsBBB Ithaca, NY, 1996,31. Beck C.I. and Ulrich T., Biotechnology in the food industry. Bioflechnology 1993,11, 895. Zabetakis 1. et a/.,The biosynthesis of 2,5-dimethyl-4-hydroxy-2H-furan-3-one and its derivatives in strawberry. In: Flavour science. 8th Weurman Symposium, 1996,90. Tress1 R. eta/., Formation of y- and Glactones by different biochemical pathways. In: Flavour science. 8th Weurman Symposium, 1996,141. Patent pending, Givaudan Roure. Gautier A. eta/., Firmenich patent 951215 UPC 940418. HBusler A. and Schilling B., Future impact of recombinant DNA technology on the production of natural aroma chemicals. Proceedings of the 5th Wartburg Symposium, Germany, 1997,in press. t!atanaka A., The fresh green odor emitted by plants, Food. Rev. Int. 1996,12,303. Aristoy M.C. and Toldra F., Isolation of flavour peptides from raw pork meat and dry-cured ham. In Food flavours: generation, analysis and process influence, Elsevier Science B.V., 1995. Mojarra-Guerra S.H. et a/., Isolation of low-molecular-weight taste peptides from Vacherin Mont d’Or cheese. J. Food Sci. 1991,56,4. Gill 1. et at., Biologically active peptides and enzymatic approaches to their production. Enz. Microb. Techno/.1996,18,162. Spanier A.M., BMP: a flavor enhancing peptide found naturally in beef. Its chemical synthesis, descriptive sensory analysis, and some factorsaffecting its usefulness. In: Food flavors: generation, analysis and process influence, Elsevier Science B.V., 1995,1365. Patent pending, Givaudan Roure.

20 l7 18

I’ 20

21

22

23

24

25

26

27 28

29

30

31

32 33

35

37

38

39


Cheetham P.S.J., The flavour and fragrance industry In: Biotechnology: the science and the business, Harwood Academic Publishers, 1991, 26, 481. Nonino E.A., Where is the citrus industry going? Perfumer 8 Navourist 1997, 22, 53. Murphy D.J., Engineering oil production in rapeseed and other oil crops. TlBTECH 1996, 14, 206. Goy P.A. and Duesing J.H., From pots to plots - genetically-modified plants on trial, Bioflechnology 1995, 13, 454. Liu K. and Brown E.A.. Enhancing vegetable oil quality through plant breeding and genetic engineering. f o o d Technology 1996, 11, 67. Pellegrineschi A. et a/., Improvement of ornamental characters and fragrance production in lemon-scented geranium through genetic transformation by Agrobacterium rhizogenes. Bioflechnology, 1994, 12 ,64. Hallagan J.R. and Hall R.L., Safety assessment of flavour ingredients produced by genetically modified organisms. ACS Symp., 1995, 605, 59. Hoban T.J., Consumer acceptance of biotechnology: an international perspective. Nature Biotechnology 1997, 15, 232. Wrotnowski C.. Unexpected niche applications for industrial enzymes drives market growth. Genetic engineering news 1997, 17, 14. Lieske B. and Konrad G., Protein hydrolysis the key to meat flavouring systems. Food Rev. Intern. 1994, 10, 287. Lecointe C. etal., Ester synthesis in aqueous-media in the presence of various lipases, Biotechn. Lett.1996, 18, 869. Spanier A.M. and Miller J.A., Role of proteins and peptides in meat flavor. ACS Symp. Ser. 1993, 528, 78. Headon D.R. and Walsh G.. The industrial production of enzymes. Biotechnology advances 1994, 12,635. Izawa-N et a/., Debittering of protein hydrolysates using Aeromonas caviae aminopeptidase, J. Agr. Food, 1997,45, 543. Quest Int. BV, Unilever patent EP 565172 A l . Niedermann D. and Lerch K., Molecular cloning of the L-amino-acid oxidase gene from Neurospora crassa, J. Biol. Chem. 1990, 265, 17246. Schilling B. and Lerch K., Cloning, sequencing and heterologous expression of the monoamine oxidase gene from Aspergillus niger, Mol. G. Genet. 1995, 247, 430. Buck L. and Axel R., A novel multigene family may encode odorant receptors - a molecular basis for odor recognition, Cell 1991, 65, 175. Shepherd G. eta/., Olfactory receptors - a large gene family with broad affinities and multiple functions. Neuroscientist 1996, 2, 262. Raming K. etal., Cloning and expression of odorant receptors. Nature, 1993, 361, 353. Bajgrowicz J. and Broger C., Molecular modelling in design of new odorants: scope and limitations. Proceedings of the 13th Congress of flavours, fragrances and essential oils,1995.3,1 Rossiter K.J., Structure-odor relationships. Chem. Rev. 1996, 96, 3201. Ruiz-Avila L eta/., Coupling of bitter receptor to phosphodiesterase through transducin in taste receptor cells. Nature 1995, 376, 80.

-

The Design and Synthesis of Novel Muguet Materials K. J. Rossiter QUEST INTERNATIONAL, ASHFORD, KENT TN24 OLT, UK

1 ABSTRACT

A qualitative structure activity approach, involving the hybridisation of two fragments found in two different known muguet odourants, led to the identification of several novel muguet smelling 3-alkoxypropan-1-01s. Several analogues of the best muguet lead, 2,2,7trimethyl-4-oxaoctan-1-ol,were synthesised and assessed in an attempt to better understand the structural requirements for a mupet odour. The effect of the methyl substitution pattern at C1, C2 and C3, and replacement of the ether linkage by an olefinic linkage, a triple bond and a methylene group were investigated. It was found that C1 and C3 should ideally be unsubstituted and that C2 should be dimethylated. Floral character was retained when the ether oxygen atom was replaced by a saturated or olefinic carbon. The muguet note appeared to be more pronounced in the 3-alkoxypropan-1-01 and the trans-alkenol, with the floral odour of the cis-alkenol and the fully saturated analogue shifting more in the direction of rose.

2 INTRODUCTION Since flowers from the muguet (lily of the valley) plant are very small and difficult to extract, it is impossible to produce a commercially viable blossom oil from this plant. Therefore, perfumers have to rely on synthetic substitutes, such as cyclamen aldehyde (l), Bourgeonal@(2), Lilial@(3), hydroxycitronellal (4), Lyral@( 5 ) and Dupical@(6) to create this odour type (Figure l), none of which have yet been identified in the rnuguet flower. In fact the odour of these materials is somewhat heavier than that of the living flower, which is soft and quite rose like, but because of their widespread use in perfiunery, the consumer and also the perfumer now use the term muguet to describe odours which are similar to these aroma chemicals rather than to the odour of the flower. The aldehydic muguet ingredients were among the first to be discovered and are still highly valued in perfumery today. Since all of the early muguet odourants were aldehydes the presence of the aldehyde functional group was originally believed to be a prerequisite for a muguet odour. However, today there are also a number of alcohols which can be used to create a muguet odour and these include Majantol@(7), Florosa Q@ (S), famesol (9),

22


(1)

(2)

Cyclamen aldehyde

Bourgeonal@ (Quest Int.)

Lilial@ (Givaudan)

LyraP ( I F F )

Dupical@ (Quest Int.)

(3)

I

OH (4)

Hydroxycitronellal

Figure 1 Synthetic aldehydic muguet odourants

2,6-dimethylheptan-2-01 (lo), and Mayol@ (1 1) (Figure 2). Farnes01'**~~ and dihydrofarne~ol~'~ have actually been found in the plant oil. The two classes of muguet odourants have quite different odour profiles. For example, Figure 3 shows the odour aspects which are significantly different for seven representative muguet materials. The three 3-(p-alkylphenyl)-propanals, cyclamen aldehyde (I), Bourgeonal@ (2), and Lilial@ (3), are more fruity with melon aspects; the two alcohols, Florosa Q@(8) and Mayol@(1 l), tend to have prominent herbal, lavender, pine and sandalwood notes; while the two materials which contain both an alcohol and aldehyde hnctional group, hydroxycitronellal (4) and Lyral@(9, are perceived as sweet. Pelzer et al.' studied the odour profiles of 181 substances possessing a typical muguet scent and found that the carbonyl compounds, in addition to the muguet aspects, exhibited lime blossom notes.

LO" q

OH

\

(7)

(8)

Majantol@ (Wacker)

(9)

Florosa Q@ (Quest lnt.)

Farnesol .OH

f (10)

Q

2,6-Dimethylheptan-2-01

Mayol@ (Firmenich)

Figure 2 Synthetic alcoholic muguet odourants

(11)

23

The Design and Synthesis of Novel Muguet Materials

Florosa Q H cyclamen aldehyde Lyral

1

hydroxycitronelkl

/

SWEET

1

H Lilial

Figure 3 Diflerences in odour profiles of muguet odourants

In more recent years, attention has been focused towards the discovery of nonaldehydic muguet materials. This is because of the instability associated with certain materials from this chemical class. Common side reactions include oxidation to the corresponding carboxylic acid and, in some cases, the corresponding lower homologue ketone; reaction with alcohols to form the corresponding acetals; aldol condensations; and the formation of aldehyde trimers (Figure 4). In general, the more hostile the product base the less stable the fragrance ingredient. For example, aldehydic ingredients are notoriously unstable in antiperspirant formulations which are strongly acidic due to partial hydrolysis of active antiperspirant agents such as aluminium chlorohydrate.

oxidation

/

m

Figure 4 Chemical instability of aldehyde group

acetal formation


24

The histogram in Figure 5 shows the relative organoleptic stability of five muguet ingredients in an aerosol antiperspirant. Two factors are considered when determining the organoleptic stability of an ingredient in a product base. The first is the degree to which the ingredient covers the base and the second is the persistence of performance over time. Thus in Figure 5 the taller the bar the more stable and better performing the ingredient is. The poor to moderate organoleptic performance of Lilial" (3), hydroxycitronellal (4),and Lyral" ( 5 ) is predominantly attributed to the chemical instability of the aldehyde group. For example, our studies have shown that the percentage of Lilial" remaining in an ethanol based roll-on antiperspirant after 4 weeks at 37°C is only 15%. Lilial" reacts with the ethanol in the base to form the corresponding diethyl acetal and undergoes oxidation. Bourgeonal" (2), on the other hand, performs well for a combination of reasons. Although Bourgeonal" reacts with the ethanol to form the corresponding diethyl acetal, it does not appear to undergo autoxidation. Consequently, the level of Bourgeonal" remaining after 4 weeks at 37°C is typically 65% as op osed to only 15% for Lilial". Another reason for the superior performance of Bourgeonal IS that it is a very potent material being 2 to 4 times stronger than Lilial" at the same concentration.' Florosa Q" (S), the only alcoholic muguet odourant in Figure 5, performs well because it is chemically very stable. Ether groups, apart from the possibility of peroxide formation, are relatively inert, and although possible side reactions of the alcohol group will depend upon the specific formulation of the antiperspirant, there is relatively little it could do other than acid catalysed dehydration, and this usually requires fairly high temperatures or lower pH. Consequently the level of Florosa Q" in, for example, an ethanol based roll on antiperspirant after 4 weeks storage at 37°C remains unchanged. The good as opposed to very good rating is a sensory effect.

8.

Organoleptic Stability

7

Hydroxycitmnellal Lilial

Lyml

Bourgmna Florosa

Figure 5 Organoleptic performance in an aerosol antiperspirant (3 7 ° C 12 weeks)

25

The Design and Synthesis of Novel Muguet Materials \

(7)

Majantol@ (Wacker)

/

(12) Reference 6

(13) Quest Research Material

Figure 6 Incorporation of the 2,2-dimethylpropan-l-oIJi.agment 3 DESIGN, SYNTHESIS AND ODOUR EVALUATION OF NOVEL MUGUET INGREDIENTS 3.1 Lead Identification The aim of the present study was to find a novel muguet odourant which is stable and which could potentially be prepared cost effectively on an industrial scale. It is clear from the above that the best chances of finding such a material lies within the alcohol chemical class. From a list of known muguet odourants it was observed that a few members (for example, 7, 12 and 13 in Figure 6) had one structural feature in common, the presence of a 2,2-dimethylpropan-l-ol fragment (14). One easy way of incorporating this fragment into a molecule is by the reductive cleavage of 5,5-dimethyl-ly3-dioxanes (IS), which in turn can be easily prepared from a carbonyl compound and the readily available and inexpensive 2,2-dimethylpropane-1,3-diol(Scheme 1). On a laboratory scale the most convenient way of reductively cleaving an acetal or ketal is by the use of reducing agents such as BH3 in tetrahydrofuran' or a combination of LiAIH4 and AIC13 in diethyl ether.' The latter reduces cyclic acetals in virtually quantitative yields and as such was the procedure adopted for this work. On an industrial scale, however, the preferred cheapest and safest rocess would be hydrogenolysis in the presence of an acid and a metal catalyst?*"*

P'

The target compounds 16 contain the 2,2-dimethylpropan-l-ol fragment connected via an ether linkage to the carbon skeleton of the starting carbonyl compound. The ether oxygen atom provides an area of high electron density at position four and as such may fulfil a similar role to the double bond and the phenyl ring often found at this position (see compounds 1,2,3,5,6,7 and 8).

26


a

Known muguet odourant Two desiredfragments

Starring materials used to incorporare HO above fragments in target hydroxy ether

target compound

Figure 7 Design of novel muguet odourants using the hybrid approach The criteria for selection of the carbonyl substrates was not only availability and price, but also, and more importantly, upon the similarity of their carbon skeleta to those of known muguet odourants. For example, the tricyclic fragment in Dupical@( 6 ) can be attached to the 2,2-dimethylpropan-I-o1 fragment (13) by the use of pketotricyclo[5.2.1 .02.6]decane (17). The resulting hydroxy ether (18), in addition to containing two fragments found in known muguet odourants, is also of a similar size to Dupical@(Figure 7). By using this qualitative structure activity approach it was anticipated that the chances of finding materials of the desired odour type would be increased. Ten 2,2-dimethyl-3-alkoxypropan1 -ols, which had been identified as target molecules using this hybrid approach, were prepared and their organoleptic properties assessed. A few of the bigger molecules, such as compound 18, had woody type notes, but the majority (70%) were floral in character and, although rose appeared to be the main floral aspect, three out of the seven floral odourants were also described as having muguet connotations (compounds 19-2 1, Figure 8).


27

(20) Novel 2,2-dimethyl-3-oxa-propan-l-ol

(12)

(21) Reference 12

Known muguet odourant

Figure 8 Novel 2,2-dimethyl-3-oxa-propan-l-ols possessing muguet odours

3.2 Lead Optimisation Pelzer et al.’ have suggested that the substitution pattern, particularly around the functional group, is an important criterion for alcohols and aldehydes to smell of muguet (Figure 9). Pelzer’s analysis of 73 alcohols produced, not only the distance constraints shown in Figure 9, but also the following rules: 1) C1 should be substituted by one to three alkyl groups, ideally three, provided that the alcohol group is not overshadowed too strongly by steric hindrance.

2) C2, C5, and to a lesser extent C6 and C7 should be substituted by a single alkyl group, ideally a methyl group. Dimethylation generally has a detrimental effect on the fragrance impression. 3) Where a double bond is present it should preferably be at C4 or C6. A double bond between C3 and C4 or, to a lesser extent between C2 and C3 generally has a negative influence on the odour.

Cl-C4 = 3.2 f Cl-Cg = 4.0 f Cl-Cg = 5.0 f Cl-Rl =4.7* Figure 9 Pelzer ’s muguet fiagment for alcohols

0.3A O.3A 0.3A 0.4A

28


3.2.1 Variation of the Methyl Substitution Pattern. In order to investigate the effect of introducing a methyl group at C1 for the 3-alkoxypropanol family, compound 19 was oxidised to the corresponding aldehyde (23) which was then subjected to a methyl Grignard reaction to yield 3,3,8-trimethyI-5-oxanonan-2-01(24) (Scheme 2).

(23)

Scheme 2

An easy way of changing the methyl substitution pattern at C1, C2 and C3 is to use a range of different 1,3-diols in the formation of the 1,3-dioxane precursors (Scheme 1). Therefore three analogues (compounds 25-27) of the best muguet lead (1 9) were prepared by reacting 3-methylbutanal with 2-methylpropane-l,3-diol, propane- 1,3-diol and 2methylpentan-2,4-dioI (hexylene glycol), all of which are cheap and readily available. In the case of hexylene glycol, where the diol is unsymmetrical, cleavage of the two different C - 0 bonds will give rise to two different products (27). The odour properties of these four hydroxyethers along with that of the parent compound 19 are listed in Figure 10. By comparing the odours of these five compounds, it was concluded that in this series of alkoxyalcohols dimethylation at position two is very important for a good muguet odour. Replacement of the methyl groups by hydrogen leads to a gradual decrease in the muguet character and the introduction of a fruity note. For example, compound 25, which has only one methyl group at position two, has a slightly more fruity and much less intense odour than 19. Removal of both methyl groups leads to complete destruction of the muguet character and a strengthening of the fruity note (26 and 27). The organoleptic properties of the individual isomers of 27 formed from the ring opening of 4,4,6-trimethyl2-(2-methylpropyl)-l,3-dioxanewere assessed by gc-sniffing and found to be very similar. Introduction of a methyl group at position one, whilst retaining dimethylation at C2, resulted in a very weak muguet odourant 24. In fact these findings contradict the aforementioned rules of Pelzer el ~ 1 They . ~ state that dimethylation at C2 generally has a detrimental effect on the fragrance impression, whereas in this series the dimethyl analogue 19 is the best with respect to both odour quality and intensity. They also suggest that C1 should be substituted by one to three alkyl groups, ideally three, provided that the polar group is not overshadowed too strongly by steric hindrance. From this statement one might expect 3 alkyl groups to be better than 2, which in turn are better than one. However, in this study the converse was found to be true with the secondary alcohol 24 being less intense than the primary alcohol 19. It is possible that the presence of both the methyl group at C1 and the dimethyl substitution at C2, in combination, lead to steric hindrance of the alcohol group and that this is the reason why 24 is much weaker than 19. Thus the rules of Pelzer and co-workers do not appear to apply to 3-alkoxypropanols. However, since they do not list all of the alcohols used in their study, it is impossible to tell whether or not they included ethers in their data set, or whether they restricted their set to compounds where the only functional group is hydroxyl.

The Design and Synthesis of Novel Miiguet Materials

29 ( I 9) Muguet, rose

(24) Very weak, muguet, rose, fruity

To+oH

( 2 5 ) Weak, fruity, rose, muguet

(26) Weak, fruity, woody T

O

A

o

H

TomoH+ Ton OH

(27) Fruity, green, herbal, woody

Figure 10 The effect of the methyl substitution pattern on the odour of 3-alkoxypropanols 3.2.2 Replacement of the Ether Linkage by a Double Bond. It is generally known within the perfiunery industry that the introduction of unsaturation in a molecule can lead to an increase in odour inten~ity.'~ Therefore the effect of replacing the ether linkage by an olefinic linkage was investigated (Figure 11). Since this structural change introduces rigidity into the molecule, one would expect, as is often the case with alkene geometric isomers, that cis- and trans-2,2,7-trimethyloct-4-en1-01 would have different odours. Molecular modelling was used to predict whether either of the isomers of 2,2,7trimethyloct-4-en- 1-01 (28) would have a muguet odour. 3.2.2.1 Modelling Work. Florosa Q@ was chosen as the muguet standard for the modelling work for two reasons. Firstly, it is relatively rigid conformationally and thus relatively easy to model, and secondly, the two isomers of Florosa Q@ are known to have different organoleptic properties. Sommer and GUntertl4 have reported that only the cisisomer, in which the hydroxyl and isobutyl groups are both equatorial, is the actual fragrance carrier. Our own findings suggest that both isomers have a muguet odour but that the cis-isomer is much more potent than the trans- (Figure 12). The two isomers of Florosa Q@ were separated by column chromatography and a series of dilutions in ethanol prepared. Forty nine people were asked to smell the dilutions from smelling strips, waiting a few minutes after the strips had been dipped for the ethanol to evaporate. They started with the lowest concentration and worked their way up to the most concentrated. They noted when they could first detect an odour, thus providing a measure of the relative odour threshold of the two isomers. Nine people could not smell the trans-isomer at the highest concentration (10% w/v) and two could not smell the cis-isomer. The results for those people who could detect an odour (40 and 47 for trans- and cis-Florosa Q@ respectively)

(19)

(284

Figure 11 Replacement of the ether group by a double bond

I

(2W

30


I

fl

x'""

, ;uI

trans-Isomer

cis-Isomer Strong muguet

Weak muguet / 8 % r i c

I6

I2 6 4

0.005 0.01

0.05

0.1

0.5

I

5

0.005 0.01

10

0.05

0.1

0.5

I

5

10

lconcl at which fin1 smelt

lconcl at which Rnt smelt

Figure 12 Dgerences in odour thresholds of cis- and trans-Florosa Q" are displayed as histograms in Figure 12. From the spread of data it can be seen that the trans-isomer is first detected at higher concentrations than the cis-isomer (i.e. it is less potent). The assumption was made that compounds will have a good muguet odour if they can adopt a conformation that closely resembles that of cis-Florosa Q". The alkoxypropanols are very flexible molecules and thus 2,2,7-trimethyl-4-oxaoctan-1-01 (19), which has a very similar carbon skeleton to that of Florosa Q@,can easily adopt relatively stable cyclic-type conformations (within 1-5 kcal/mol from the energy of the straight chain conformer, 19b) which resemble that of either trans- or cis-Florosa Q@(19a and 19c respectively) (Figure 13). In these conformations the distance between the two

A

A cis-Isomer, srrong muguer 0-0Distance = 4. I7A

trans-Isomer, weak-odourless 0-0 Distance = 3.49A

a

OH OH

( I 9a) 0-0 Distance = 3.53A

(19b)

( 19c)

0-0 Distance = 4.96A

0-0 Distance = 3.96A

Emrgy=-I 18.6(-125.2) kcal/mol

Energy = -120.8 (-129.9) kcallmol

Energy = - I 19.8 (-129.2) kcallmol

Figure 13 Comparison of the conformational arrangement of 19 with that of Florosa Q" (The cyclic-type conformations of 19 were tined to the cis and trans forms of Florosa Q" using the SYBYL MULTI-FIT tool. Heats of formation were calculated using the PM3 (and AMl) methods in MOPAC.)

31


w Figure 14 Top: superimposition of 19a and cis-Florosa Q". Bottom left: trans-2,2,7trimethyloct-4-en-1-01. Bottom right: cis-2,2,7-trimethyloct-4-en1-01. oxygen atoms are virtually identical and the isobutyl group is located in a similar position in space. It is postulated that the muguet aspects of 19 are associated with the cyclicconformation (19c). The similarity between 19c and cis-Florosa Q" is clearly seen in the top of Figure 14, where the two molecules have been superim osed (the green molecule is 19c, and the molecule coloured by atom type is cis-Florosa Q . Turning now to the alkene derivative, 2,2,7-trimethyloct-4-en-1-01 (28). Although both trans- (bottom left, Figure 14) and cis-2,2,7-trimethyloct-4-en-l-ol (bottom right, Figure 14) can adopt the postulated active cyclic-type conformation, only the trans-isomer has the isobutyl group located in a very similar position to that of 19c and cis-Florosa Q@ (top, Figure 14). It was thus predicted that trans-2,2,7-trimethyloct-4-en-I-ol would have a muguet, rose odour and that cis-2,2,7-trimethyloct-4-en-I-ol would not.

4

3.2.2.1 Synthesis of trans- and cis-2,2,7-trimethyloct-4-en-l-ol.Trans-2,2,7trimethyloct-4-en01 was successfully prepared using the route outlined in Scheme 3. This process uses Claisen chemistry to ensure trans-c~nfiguration'~ of the double bond. 2Methylpropanal was treated with vinyl Grignard reagent to yield 5-methylhex-1-en-3-ol' (29). The vinyl alcohol (29) was reacted with 3-methylpropanal to give predominantly the trans-isomer of 2,2,7-trimethyloct-4-enal(30). The aldehyde was reduced to the corresponding alcohol (28a) using sodium borohydride. The isomeric composition of the product was 95% trans and 5% cis (capillary gc, rpa).

32

Flavours and Fragrances 1-

Cis-2,2,7-trimethyloct-4-enolwas prepared in five steps (scheme 4). 4-Methylpent1-ene was brominated to give 1,2-dibrom0-4-methylpentane,which on subsequent dehydrobromination using the method of Dehmlow and ThieserI6 yielded 4-methylpent-I yne. The latter was coupled with the THP ether of 3-bromo-2,2-dimethylpropan-l-ol according to the procedure of Schwarz and Waters". Deprotection yielded 2,2,7trimethyloct-4-yn- 1-01 (3 l), which was hydrogenated under Lindlar conditions to give the desired cis-alcohol (28b). For comparative purposes the fully saturated analogue, 2,2,7-trimethyloctan-1-01 (32) was also prepared by catalytic hydrogenation of trans-2,2,7-trimethyloct-4-en-l-ol. The organoleptic properties of 2,2,7-trimethyl-4-0xaocten-l-ol(19), trans-2,2,7trimethyloct-4-en- 1-01 (28a), cis-2,2,7-trimethyIoct-4-en-1-01 (28b), 2,2,7-trimethyloct-4yn-1-01 (31) and 2,2,7-trimethyloctan-l-ol(32) were compared (Figure 15). The odours of these samples were assessed as 1% and 10% solutions in diethyl phthalate by a panel of perfumers and fragrance chemists. The organoleptic purity of each sample was checked by gc olfactometry.

(31)

Scheme 4


33

(19) Muguet, rose

(28a) Muguet, rose, woody

(28b) Rose, geraniol-like,rnuguet

(31) Weak, minty (minty note from allene impurity)

(32) Rose,geraniol-like, muguet

Figure 15 Replacement of the C-0 linkage and its effect on odour

All of the materials were floral in odour, except for 2,2,7-trimethyloct-4-yn-l-ol (31) which was described as weak and minty. This note was shown, by gc olfactometry, to be due to the presence of a trace amount of the corresponding allene. The odour descriptors given to the remaining four compounds varied significantly from subject to subject thus making it very difficult to draw definite conclusions about the effect of structure on the muguet character of these alcohols. The trans-isomer 28a and the alkoxyalcohol 19 were generally described as having a muguet odour with some rose notes, which was consistent with the modelling predictions. The trans-alkene also possessed some woody character. The two materials which were described by the widest range of descriptors were the fully saturated alcohol and the cis-alkenol. Some assessors perceived these as being very muguet like whilst others perceived them as being definitely rose and geraniol like. The fact that these two compounds presented the greatest problems with regard to subjectivity, could possibly be explained by the fact that their structures do not satisfactorily meet the requirements for a muguet odour. The saturated compound 32 is also more conformationally mobile, thus allowing it to adapt to fit a wider range of receptors. Flexible molecules tend to have complex odour profiles and as such are often described using a number of odour descriptors. These molecules can adopt a large number of energetically favourable conformations each of which may be responsible for triggering a different odour response. For example, Yoshii et investigated the stable conformations of (R)-ethylcitronellyl oxalate and found that the most stable compact conformations fitted their previously published benzenoid musk model. One of the stable conformations partially resembled one conformer of S-citronellol, a rose odourant. They concluded that these conformations could be responsible for ethyl citronellyl oxalates main odour quality (musk) and its secondary odour quality (rose) and that other notes, such as woody and fresh, might be explained by further conformational comparisons with other structure-odour models. Rigid compounds on the other hand usually have a well-

34


defined odour which can be described using only one or two words (odour descriptors). They are thus easy to classify and are also relatively easy to model. It is for these two reasons that structure-odour relationship studies have been concentrated on odour groups such as ambergris, bitter almond, musk and sandalwood rather than odour groups such as floral, fruity and green. A recent review on the field of structure-odour relationships is provided by Rossiter.”

4 CONCLUSIONS

1) Several novel muguet-smelling 3-alkoxypropan- 1-01s were identified using a hybrid qualitative structure activity relationship approach.

2) In the 3-(3-methylbutoxy)propan-1-01 series it was found that dimethylation at position two is very important for a good muguet character. Successive replacement of the methyl groups by hydrogen leads to a gradual decrease in the muguet character and the introduction of a fruity note. This seems to contradict earlier SAR models. 3) Floral character is retained when the ether oxygen atom is replaced by a saturated or olefinic carbon. The muguet note appears to be more pronounced in the 3alkoxypropan- 1-01 and the trans-alkenol, whereas the cis-alkenol and the fully saturated analogue appear to be more rose-like. 4) Common problems encountered with the development of structure-odour relationships

were exemplified by this work. These include the subjectivity of odour, the importance of organoleptic purity, and the dilemmas associated with the modelling of conformationally flexible molecules. Acknowledgements

I would like to thank the following colleagues from Quest International. Ian Payne for the organoleptical stability data of muguet ingredients, Steven Rowland for analysis of the breakdown products of fragrance ingredients in antiperspirant base, and Kerry McInerney and Anne Richardson for the sensory profiling work. I would also like to thank Joe Metcalfe (Oxford University) and Karin Rose (Portsmouth University) who, during undergraduate placements at Quest International, helped with the synthesis of the 3-alkoxypropan-1-01s and the determination of the relative odour thresholds of the ‘8 Florosa Q isomers respectively.

References

1. 2. 3.

M. Boelens, H. J. Wobben and J. Heydel, Perfum. Flavor., 1980,5 (6), 2. D. P. Anonis, Perfum. Flavor., 1987, 11 (6), 31. E. J. Brunke, F. J. Hammerschmidt, F. Rittler and G . Schmaus, SOFW J., 1996 122 (9), 593.


4.

5.

6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

35

H. Surburg, M. Guentert and H. Harder, “Recent Developments of Flavour and Fragrance Chemistry”: Proceedings of the 3rd International Haarmann & Reimer Symposium, VCH Publishers, Weinheim, Germany, 1993, p. 103. R. Pelzer, U. Harder, A. Krempel, H. Sommer, H. Surburg and P. Hoever, “Recent Developments of Flavour and Fragrance Chemistry”: Proceedings of the 3rd International Haarmann & Reimer Symposium, VCH Publishers, Weinheim, Germany, 1993, p. 29. C. G. Cardenas, H. M. Hoffmann and B. J. Kane, Perfum. Flavor., 1993,lS (l), 11. H. I. Bolker and B. I. Fleming, Can. J. Chem., 1974,52,888. E. L. Eliel, V. G. Badding, and M. N. Rerick, J. Am. Chem. Soc.,1962,84,2371. W. L. Howard, J. H. Brown Jr., J. Org. Chem., 1961,25, 1026. P. A. Gorin, J. Org. Chem,, 1959,24,49. E. F. M. Ghenassia and A. J. Lakodey, PC UK PRODUITS CHIMIQUES UGIINE KUHLMANN, European Patent 0092 463 Al, published 26/10/83 C. Anselmi, M. Centini, M. Mariani, A. Sega and P. Pelosi, J. Agric. Food Chem., 1992,40,853. P. A. Edwards and P. C. Jurs, Chem. Senses, 1989,14 (2), 281. H. Sommer and M. Giintert, Haannan & Reimer Contact, 1993,59, 9. P. Vittorelli, T. Winkler, H. J. Hansen and H. Schmid, Hefv. Chim. A m , 1968, 51, 1456. E. V. Dehmlow and R. Thieser, Tetrahedron, 1986,42 (13), 3568. M. Schwan. and R. M. Waters, Synthesis, 1972,2,567. F. Yoshii, S. Hirono and I. Moriguchi, Quant. Struct. Act. Relat., 1994, 13, 144. F. Yoshii, S. Hirono, Q. Liu and I. Moriguchi, Chem. Senses, 1992, 17,573. K. J. Rossiter, Chem. Rev., 1996,96 (8), 3201.

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrance from the Skin Braja D. Mookerjee”, Suba M. Patel, Robert W. Trenkle and Richard A. Wilson INTERNATIONAL FLAVORS & FRAGRANCES INC.. I5 15 HIGHWAY 36. UNION BEACH, NJ 07735, USA

1. INTRODUCTION It is a common belief that people perceive a fragrance by the relative volatility of its components. Thus, a fragrance is described as having a topnote, a middle note, and a bottom note. A new technology has now developed to prove that this is not so. A fragrance is actually perceived by the diffusion of molecules which is an inherent property of the compounds and is independent of their molecular weight, boiling point, and odour threshold or odour value. This technology is called Aura of Aroma@ and has been trademarked by IFF. By means of this technology we have shown, for the first time that, depending upon the fragrance, various skins may or may not have a significant effect on the emission of fragrance molecules. Both the technology and the results will be discussed in detail. 1.1 What is “Aura of Aroma@?”

Figure 1 shows an actual photograph of an eclipse of the sun taken from National Geographic magazine. When the sun is totally eclipsed by the moon, the surrounding glow is called the “Aura.” Similarly, if we consider a drop of fragrance, just as in the case of the sun in the photo, the molecules surrounding the drop form an Aura of that particular fragrance. It is a common belief that a fragrance smells layer by layer. For example, one smells first the most volatile components, called the “Top Note”, then one smells the “Middle Note” which contains the components with boiling points in the middle range, and finally one perceives the highest boiling molecules which constitute the “Bottom Note”. In reality, this is not the case. When a drop of fragrance is placed on the skin, several selected molecules from the lowest boiling to the highest boiling, irrespective of their molecular weights, boiling points, and vapour pressures, come into an “Aura,” eventually hit our nose, and give our first impression of that particular fragrance. The composition of this Aura is dependent on a characteristic property of each fragrance molecule known as its “diffusivity.”

*Correspondence to: Dr. Braja D. Mookerjee.

Aura of Aroma": A Novel Technology KO Study the Emission of Fragrance from the Skin

Figure 1. Eclipse of the sun. (Taken from National Geographic magazine)

Figure 2. Dendrobium superbum orchid

31

38


1.2. What is “Diffusivity?”

Diffusivity is the inherent property of a compound to emit its molecules into the air. One compound is said to be more diffusive than another if its molecules tend to pass into the air to a greater extent than those of the other compound. This is called “relative diffusivity.” Diffusivity is independent of boiling point, molecular weight, odour threshold, or odour value. This phenomenon is true not only for a fragrance placed on the skin but also for a living flower’s fragrance which forms an Aura. This Aura is actually due to the aroma molecules which are constantly being produced by the living flower and are coming off the surface of the petals.

2 AURA OF AROMA@THE EVOLUTION OF LIVING FLOWER@ TECHNOLOGY. Indeed, the Aura of Aroma@ technology evolved naturally from the Living Flower@ technology. In this connection, we have now developed a novel technology to capture the aroma molecules surrounding the flower petals without touching the flower or any part of the plant. This is called SPME, Solid-Phase Micro-Extraction. To our knowledge, no one has ever before applied this technique to the analysis of the Living Flower@aroma. Figure 2 shows the Dendrobium superbum orchid, now growing in our greenhouse, which comes from New Guinea and is unique in the world because of its raspberry-like fruity odour. The SPME needle, which is nothing more than a 2-3 mm glass fibre coated with a highboiling liquid adsorbent, is placed in close proximity to the flower without touching it and is kept there for a period of from 112-1 hour depending on the odour strength of the blossom. The aroma molecules around the petals are adsorbed onto the fibre which is then analysed by GUMS to give the aroma profile of that particular flower. The data in the first column of Table 1 shows the composition of the living orchid fragrance, which represents more than 98% of its aroma profile. When we reconstituted this fragrance we discovered that it lacked the diffusivity of the Dendrobium orchid. We then analysed the headspace volatiles of this recreated living orchid fragrance. This study revealed the very interesting phenomenon that the headspace composition of a living flower fragrance consists of several highly diffusive molecules the nature of which is independent of their molecular weight, boiling point, and odour strength. For example, benzyl acetone, with the lowest molecular weight did not increase at all in the headspace, whereas linalool which has a higher molecular weight than benzyl acetone, dramatically increased more than 15 times. Benzyl acetate increased 25 times, while 2-tridecanone, a so-called Middle Note component based on its molecular weight and boiling point, increased 275 times. Oxyphenylon, another Middle Note component and the compound responsible for the characteristic raspbeny odour decreased substantially in the Aura, while high boiling compounds like 2-pentadecanone and ethyl myristate were still

39

Aura of Aroma": A Novel Technology to Study the Emission of Fragrance from the Skin

Mol.

ComDound

Benzyl Acetone Benzyl Acetate Linalool Oxyphenolon 2-Tridecanone 2-Pentadecanone Ethyl Myristate

m

Topnote Topnote Topnote Middle Note Middle Note Bottom Note Bottom Note

148 150 154 164 198 226 256

Living Orchid Fragrance %! 0.02 0.20 2.20 1 1.90 0.02 69.00 14.80

Aura %! 0.03 5.20 34.10 1.70 5.50 33.50 8.50

Table 1. Aura of Aroma of Living Flower (Dendrobium superbum orchid).

substantially present in the Aura. When we recreated this Aura and added it back to the living orchid fragrance at a level of lo%, the fragrance now became very diffusive. 3 AURA OF AROMA@ON SKIN. Now, we want to know what happens when a true fragrance is applied to the skin. In other words, we want to study the Aura of the fragrance. When a fragrance is applied to the skin of a woman, a natural Aura of Aroma@surrounds her body. We want to study this Aura of Aroma@on skin. Now, I want to describe the technique we use to study this Aura. We have developed a very simple yet elegant technique in which 10 microliters of perfume is applied to a clean inner forearm of a woman. Immediately a small glass globe is placed over it and sealed by contact with the skin. The same SPME needle as used for Living Flower studies is inserted through a septum in the top of the vessel and positioned so that the fibre tip is approximately 1 cm above the sample area. The needle is kept in place for a period of 1/2-1 hour following which it is immediately analysed by G C M S to determine the composition of the fragrance Aura. In this connection, we would like to mention that, to our knowledge, nobody has ever used the SPME method to study the release of fragrance from skin. We have done this work at least three years ago and disclosed it in 1996. This novel technique was described in a recent trade journal called Spray Technology & Marketing' . The first fragrance we studied was the classical fragrance Shalimar created by Guerlain in 1925 (Table 2).

40


w

ComDonent Limonene Linalool Linalyl Acetate Ethyl Vanillin Coumarin Methyl Ionone Musk Xylol

Topnote Topnote Topnote Middle Note Middle Note Middle Note Bottom Note

30.0 1.7 9.9 0.2 1.7 1.1 Trace

Aura on Skin

w 20.4 17.9 21.6 1.6 7.8 2.1 0.3

Table 2 Comparison of Aura of Shalimar on Skin & Fragrance Oil. You can see that limonene, the most volatile component, which constitutes 30% of the fragrance oil, is only present at 20% of the Aura, whereas linalool, which is only 1.7% in the oil increases eight times in the Aura. Similarly, Linalyl acetate also doubled in the Aura. Ethyl vanillin, a high boiling chemical which was used for the first time in Shalimar to the extent of 0.2% in the oil, dramatically increased eight times in the Aura. Similarly, methyl ionone, another high boiling compound used for the first time in Shalimar, doubled in the Aura. Most interesting is the musk xylol, the highest boiling compound which is only in trace amounts in the oil, was 0.3% of the Aura. This shows that the Aura of a fragrance is composed not only of highly volatile molecules but also that both low-boiling and higher boiling compounds are highly diffusive. Next, we studied a very successful 1990’s fragrance, Amarige, introduced by Givenchy (Table 3). The Aura studies showed that Top Note constituents like linalool, benzyl acetate and styrallyl acetate increased in the Aura, but, at the same time, medium-boiling components like cashmeran, which could not even be detected in the fragrance oil, showed up in the Aura to the extent of 0.5%. Similarly, other high-boiling compounds like bacdanol, cedramber, and is0 E super doubled in the Aura. At the same time, Hedione, a very renowned and widely used fragrance chemical, is not diffusive at all, and, in reality,

ComDonent Linalool Benzyl Acetate Styralyl Acetate Cashmeran Bacdanol Hedione Cedramber is0 E super Ambrox Benzyl Salicylate Muskalactone

m Topnote Topnote Topnote Middle Note Middle Note Middle Note Middle Note Bottom Note Bottom Note Bottom Note Bottom Note

1.7 4.9 1.2 0.0 0.2 29.9 1.5 7.1 0.2 32.5 0.9

Aura on Skin

U.hu 17.9 22.7 9.7 0.5 0.5 4.9 4.9 12.1 0.1 1.1 0.4

Table 3 Comparison of Aura of Amarige on Skin & Fragrance Oil.

Aura of Aromam: A Novel Technology to Study the Emission of Fragrance from the Skin

ComDound Ethyl Linalool Linalyl Acetate Floralozone Cyclogalbaniff Dihydro Myrcenol Linalool Limonene beta Ionone Polysantol is0 E Super Ambrox Hedione Galaxolide Tonalid

Increase or Decrease

yoin Qid

!%dAua

inAm

0.7 10.4

2.8 36.0 0.3 0.7 10.6 11.5 1.4 6.2 0.4 4.8 0.4 5.0 0.6 0.1

4 times 3 times 3 times 3 times 2 times

0.1 0.2 5.8 7.7 4.5 2.5 0.2 4.8

0.7 25.9 5.5 3.3

41

1.5 times 0.3 times 3 times 2 times

same 0.5 times 0.2 times 0.1 times 0.03 times

Table 4 Comparison of Aura of Unisex Fragrance on Skin (Ihr) & Fragrance Oil can be considered to function primarily as a diluent. However, so-called Bottom Note compounds like Ambrox and musklactone, which is the common name for cyclopentadecanolide, both among the highest boiling of fragrance chemicals, showed their presence in the Aura in appreciable amounts indicating that they are very diffusive molecules which play important roles in the first impression of Amarige. Next, we studied a highly successful unisex commercial fragrance (Table 4). as we expected, ethyl linalool and linalyl acetate are enhanced in the Aura due to their high difisivity. Interestingly, middle-boiling range compounds like floralozone, cyclogalbaniff

ComDound Aldehyde AA Methyl Phenyl Acetate Ethyl Linalool Diphenyl Ether Cyclogalbaniff Methyl Ionones Ethyl Acetoacetate is0 E Super Ambrox Hedione Cyclopentadecanolide Galaxolide

w 0.05 0.02 5 .oo 0.01 0.20 2.50 1.80 2.60 0.20 18.00 4.80 14.00

Increase or Decrease 'YOin Aura 0.80

0.20 30.00 0.04 0.90 9.50 3.40 2.10 0.10 2.40 1.20 0.90

ir.uhl3

16 times 10 times 6 times 4 times 4 times 4 times 2 times

same 0.5 times

0.13 times 0.25 times 0.07 times

Table 5 Comparison of Aura of Feminine Fragrance on Skin (1 hr) & Fragrance Oil

42


and beta ionone all increased by a factor of three in the Aura. At the same time, highboiling compounds like polysantol, is0 E super, and Ambrox were present in the Aura in appreciable quantities due to their diffusivity. Once again, please note the poor diffusivity of Hedione as well as Galaxolide and Tonalid. The next perfume we studied was a very successful modem feminine fragrance also introduced in the 90’s (Table 5). In addition to two very volatile chemicals like Aldehyde AA and methyl phenyl acetate, highly diffusive middle-boiling range compounds like ethyl linalool, diphenyl ether, cyclogalbaniff, and methyl ionone, together with high-boiling yet very diffusive materials like iso E super, Ambrox, and Cyclopentadecanolide are responsible for this unique feminine note. Please observe that Hedione does not influence the Aura due to its nondiffusivity. Finally, we studied a very sensuous and highly successful woman’s fragrance, recently launched in both Europe and the US.(Table 6). It is very easy to see from the data in Table 6 why this fragrance is so diffusive. Three extremely diffusive chemicals which cannot even be detected in direct analysis of the perfume oil, Lolitol, Passionfruit Compound, and Methyl Octin Carbonate, are readily seen in the Aura. In addition, several other very diffusive molecules have been used including Givaudan’s Givescone, IFF’S Floralozone, akha and beta Damascone, cis Jasmone, and Undecavertol as well as the already discussed Ethyl Linalool, Linalyl Acetate, Methyl Ionones, and Cyclopentadecanolide.

ComDound

Lolitol Passion Fruit Comp’d. Methyl Octin Carbonate Givescone Floralozone alpha Damascone beta Damascone Ethyl Linalool Undecavertol Linalyl Acetate cis Jasmone Methyl Ionone Cyclopentadecanolide Galaxolide

M

0.10 0.01 0.10 0.04 1.40 0.30 2.00 0.10 2.00 0.90 8.00

O ‘ A 0.10 0.0 1 0.10 1.20 0.10 0.50 0.20 6.00 1 .oo 7.80 0.20 4.60 0.20 1 .oo

Increase or Decrease iILAU3

- times - times - times 10 times 10 times 5 times 5 times 4 times 3 times 3 times 2 times 2 times 0.25 times 0.13 times

Table 6 Comparison of Aura of Woman’s Fragrance on Skin (lhr) & Fragrance Oil

Aura of Aroma": A Novel Technology to Study the Emission of Fragrance from the Skin

43

From these studies we have proved beyond a doubt that the first impression of a fragrance is not only due to the highly volatile so-called Top Note chemicals, as once believed. Actually, it consists of a combination of highly volatile, middle-boiling range molecules, as well as high molecular weight high-boiling compounds possessing a quality called high diffusivity. Table 7 summarises our finding with regard to the aroma molecules, from the lowestboiling to the highest boiling, which appear simultaneous1y in the Aura of any fragrance in which they are present. Relatively higher molecular weight sulphur compounds like 8puruMenthanethiol and Passionfruit Compound play key roles in many fragrances at very low levels not only due to their extremely low odour thresholds but also because of their very high diffusivities.

Extremely Diffusive Fragrance Molecules *(Molecular Weight) Passionfruit Compound (160)* Methyl Octin Carbonate (168) 8-puruMenthanethio1 ( I 70)

Diffusive Topnote Molecules Aldehyde AA ( I 38) Lolitol (144) Linalool ( I 54) Dihydro Myrcenol(lS6) Styralyl Acetate (1 64) cis-Jasmone ( 164) Ethyl Linalool(l68) Diphenyl Ether (170) Linalyl Acetate (196)

Diffusive Middle Note Molecules Coumarin (146) Ethyl Vanillin (166) Floralozone ( 190) a & PDamascones ( 192) lonones & Methyl lonones (192/206) Cyclogalbaniff (198) Cashmeran (206) Cedramber (236)

Diffusive Bottom Note Molecules Cedrene.Cedrol (204/222) a & PSantalol(220) Polysantol(222) Patchouli Alcohol (222) is0 E Super (234) Ambrox (236) Cyclopentadecanolide (240)

Highly Used NON-Diffusive Molecules Hedione Benzyl Salycilate Galaxolide Tonalid

Table 7

44


At the same time, various so-called Middle Note compounds ionones, Cashmeran, and Cedramber as well as Bottom sesquiterpenic materials such as Cedrol, Santalol, and very high compounds like Ambrox and Cyclopentadecanolide appear in which constitute the first real impression of any fragrance.

like Ethyl Vanillin, the Note compounds like boiling amber and musk the Aura of fragrances

Compounds like Hedione, Benzyl Salicylate, and Galaxolide, which are used throughout the fragrance industry in ton quantities, are relatively non-diffusive. In other words, they play little role in the Aura of a fragrance. Now, one could easily imagine that a creative perfumer, selecting from compounds mentioned in Table 7, could easily create a long lasting and, at the same time, highly diffusive characteristic fragrance as he or she desires. 4. EFFECT OF SKIN ON EMISSION OF FRAGRANCE It is commonly believed that the smell of a fragrance differs greatly from skin to skin. Therefore, we studied whether the fragrance composition changes depending upon the skin to which it is applied. At IFF we have set out to study the emission of fragrance from skin. We have utilised the Aura technology for the study of this interesting phenomenon. Now, the question comes: what kind of skin should we study? A number of investigators have performed these studies using the skin of various Caucasian people, and they stressed the importance of fat and moisture content of the skin. We think that these are important considerations, but, we have selected the skins from a global perspective. In other words, we have selected people from different parts of the world. We selected a professional Indian lady in her mid ~ O ’ Sa, Jamaican lady of the same age who works in a clerical position, a light-skinned Caucasian girl in her mid ~ O ’ S and , a managerial Caucasian lady in her mid 50’s who has never married. Each of these women were first placed on a bland diet and asked to maintain their normal working conditions without too much physical activity. They were also asked to clean their skin by washing with non-fragranced soap. The first perfume that we selected for our skin effect studies was the unisex fragrance the Aura of which we discussed earlier (see Table 4). Two drops of the fragrance were applied to the forearm skin of each lady and the Auras were collected for one hour starting immediately after application and then analysed. These experiments were repeated twice. Table 8 show the composition of fragrance from each skin. One could easily see that among the so-called highly volatile components there is essentially no change from skin to skin. The Limonene composition varies from 9-13%; the second major component, Linalool, varies from 14-15%; the major constituent, Linalyl

45

Aura of Aroma": A Novel Technology ro Study the Emission of Fragrance from the Skin

Woman % SM(45 yr)

ComDonent cis 3-Hexenol

Limonene Dihydro Myrcenol cis 3-Hexenyl Methyl Carbonate Linalool Benzyl Acetate 4-Terpinenol Ethyl Linalool (2-10 Aldehyde Carveol Citronellol Ally1 Amy1 Glycolate Citral Linalyl Acetate alpha Terpinyl Acetate Citronellyl Acetate

w 0.03 10.70 9.50 0.30 14.10 0.40 0.10 1.80 0.05 0.40 0.50 0.20 0.20 30.90 0.30 0.10

Woman % Woman % CP(47 yr) PAM(36 yr)

Woman % CB(53 yr)

Jamaican

Caucasian

Caucasian

0.04 9.30 10.30 0.30 15.40 0.40 0.10 2.00 0.05 0.40 1.oo 0.30 0.20 3 1.OO 0.30 0.10

0.03 13.30 9.30 0.30 13.90 0.40 0.10 1.80 0.05 0.40 0.30 0.20 0.10 32.0 0.30 0.10

0.03 10.10 9.20 0.30 13.80

0.40 0.10 1.80 0.04 0.40 0.50 0.20 0.20 3 1.30 0.30 0.10

Table 8 Emission of Unisex Fragrance from Different Skins (Applied two drops to skin and immediately collected for 1 hour.) Acetate, basically did not change at all. Similarly, none of the minor components changed appreciably. Table 9 shows the composition of the Auras for the Middle and Bottom Notes. Once again, there is basically no change from skin to skin. The slight changes observed in the case of beta Ionone, Coumarin, and Tonalid for one of the Caucasian ladies is considered within the range of experimental error. Since this fragrance #1 was introduced as a unisex fragrance we also studied the effect of skins of five different male lab workers with ages between 20 and 55. One was Jamaican, one was Indian, and the rest were Caucasian. These results were compared with those obtained for the four women previously described. Within the limits of experimental error, essentially no difference was observed among the composition of the so called lower-boiling components of this fragrance from men to women. This was also found to be the case for the higher-boiling components. It is obvious that different skins do not significantly affect the emission of this particular fragrance. The next perhme we studied for the effect of skin on the emission of fragrance was another women's fragrance which was first introduced into the U.S. in 1994. In this case, together with the four women we also included one young male.

46


Woman YO Woman YO Woman YO Woman Yo CP(47 yr) PAM(36 yr) CB(53 yr) SM(45 yr) Component Neryl Acetate Geranyl Acetate cis Jasmone Coumarin Cyclogalbaniff Floralozone Caryophyllene beta Ionone Pol ysantol Lilial Helional Kharismal is0 E Super Ambrox Galaxolide Tonalid

.h&j!j

Jamaican

1 .oo 1.60 0.10 0.10 0.20 0.20 0.60 4.10 0.20 0.60 1.60 1.20 2.00 0.10 0.20 0.10

1

.oo

Caucasian 1.10 1.70 0.10 0.05 0.10 0.10 0.70 1.90 0.10 0.10 0.10 0.50 1.10 0.10

1.60 0.20 0.10 0.30 0.20 0.60 3.60 0.20 0.30 0.20 1.20 1.70 0.10 0.20 0.10

Ca_ucasian 1 .oo

0.10 0.03

1.60 0.20 0.10 0.30 0.20 0.60 4.20 0.20 0.50 0.20 1.40 2.30 0.20 0.20 0.10

Table 9 Emission of Unisex Fragrance from Different Skins (Cont’d) (Applied two drops to skin and immediately collected for 1 hour.) The Auras were collected from the five subjects for one hour immediately after application of the fragrance. As can be seen in Table 10, there is no significant change in the composition of the initial Auras from all of these five skins. Woman YO

Woman YO

Woman YQ

Woman %

(45 Yr)

(47 yr)

(36 yr)

(53 yr)

Jamaican

Caucasian

Caucasian

L2xUskQ

2.80 7.40 20.30 1.80 6.60 ISO 0.80 2.10 I 1.40 13.10 15.00

0.80

3.30 4.20 20.30 I .80 5.20 0.90 1.10 2.50 14.10 15.80 16.50 0.60

2.10 4.60 18.10 I.60 5.70 1.10 0.90 2.30 12.70 14.60 18.10 1.oo

0.20 1 I.20

12.90

2.80 4.80 24.90 I.70 3.40 1.10 1.30 3.10 14.20 14.00 13.40 0.50 0.30 10.70

m

€Qmw.nm Dihydro Myrcenol Phenyl Ethyl Alcohol Benzyl Acetate Ethyl Linalool Citronellol Phenyl Ethyl Acetate Linalyl Acetate Dimethyl Octyl Acetate Citronellyl Acetate Geranyl Acetate gamma Methyl lonone Lilial Bacdanol iso E Super Cyclopentadecanolide Galaxolide

TN TN

TN TN TN TN TN TN TN TN

MN MN BN 6N BN BN

3.40 5.90 20.50 I .90 6.40 I.20 0.90 2.10 11.20 12.70 15.0 0.80 0.20 12.20

TN = Topnote, MN = Middle Note, BN = Bottom Note Table 10 Emission of Women’s Fragrance from Different Skins (Applied two drops to skin and immediately collected for 1 hour.)

Man % (20yr)

0.30 12.50

47

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrancefrom the Skin

(45 Yr)

Woman % (47 yr)

Woman % (36 yr)

Woman % (53 yr)

Ilu!ian

LimaiGm

Caucasian

Caucasian C-aucasian

Woman %

ComDonent

Dihydro Myrcenol Phenyl Ethyl Alcohol Benzyl Acetate Ethyl Linalool Citronellol Phenyl Ethyl Acetate Linalyl Acetate Dimethyl Octyl Acetate Citronellyl Acetate Geranyl Acetate gamma Methyl Ionone Lilial Bacdanol is0 E Super Cyclopentadecanolide Galaxolide

TN TN

TN TN

Man % (20Yr)

0.50

0.60 0.50

1.10

TN TN

TN

TN TN TN

m MN BN BN BN BN

0.30 4.00 9.30 22.70 2.80 1 .so

38.60

5.20 17.60 3.70 53.20 1.70

0.40 5.20 10.80 24.50 2.80 1.50 35.60 0.80 0.80

2.40 I1.50 3.40 2.20 56.20 1.90 2.40

1.60 3.90 15.40 2.90 2.30 48.10 I .so 1.90

TN = Topnote, MN = Middle Note, BN = Bottom Note Table 11 Emission of Women’s Fragrance from Different Skins (Applied two drops to skin, after 45 minutes, collected for 1 hour.) Since this particular fragrance lasts longer on the skin than the unisex fragrance, in another experiment we collected the Auras after waiting for a period of 45 minutes following the application of the fragrance. Table 1 1 shows that irrespective of the skin, after 45 minutes none of the skins retained any of the lower-boiling components, however, for the higherboilers, there are some significant differences. From these results, we can conclude that, for the two fragrances we studied which are of totally different composition, there were basically no changes observed in the initial Auras of the fragrances. However, after waiting approximately one hour, the feminine fragrance shows some differences in the composition of the higher-boiling constituents of the Auras. Therefore, we may assume that skin may have some effect depending on the composition of the fragrance. To really establish the effect of skin, obviously we must do more studies on various fragrances.

5. ACKNOWLEDGEMENT We would like to thank Mr. Gene Grisanti, Chairman and CEO of IFF, and Ms Lisa DelBeccaro, Director of Advertising and Public Relations, IFF Inc., for permitting us to present and publish this paper. References 1. “IFF Announces New Method for Living Flower@Analysis”. Spray Technology & Marketing, p26. Edited by Michael N. SanGiovanni, Published by Industry Publications, Inc., Fairfield, New Jersey, October, 1996.

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals M. Yu. Gorbachov INSTITUTE OF CHEMISTRY OF AS, MD 2028 KISHINEV, REPUBLIC OF MOLDOVA

1. INTRODUCTION

The musk odour is found within four very different chemical families. These are macrocyclic compounds, in particular lactones and ketones, aromatic nitrocompounds, aromatic benzenoids, such as acyl tetralins and indanes, and a few steroids’. The commercial importance of musk odourants, the fact that musk is an odour facet which is well defined, and also that the aromatic musks are fairly ridgid has resulted in a large number of studies into the relationship between structure and musk odour. Some, on the basis of the observation that there are subtle odour differences between the different families, have assumed that there is more than one musk receptor, and searched for correlations within a given group. Others have tried to speculate on molecular parameters common to all groups. One such study was that of Bersuker et al.* They found that two independent molecular fragments with special geometrical and electronic characteristics were required for a compound to smell of musk. The first fragment consists of a polar group (CO, NO, OH) whose electronegative heteroatom is situated symmetrically, and is at a distance of 6.7 k 0.5 A with respect to two methyl (or methylene) groups, the distance between the later being 2.5 f 0.5 A. The second fragment includes two other methyl (or methylene) groups situated at a distance of 5.5 f 0.5 8, from each other. Both of these conditions are true provided that there are no bulky substituents close to the functional group making the later sterically unaccessible. These two structural fragments are called hereafter the active fragments I and I1 respectively. Their presence in the investigated molecular structure results in its musk odour. In the absence of either or both fragments the musk odour disappears. These fragment rules allow a qualitative description of the influence of insignificant changes of molecular stucture on the presence (or absence) of musk odour. However, the problem of the dependence of the musk odour intensity remains unsolved.

2. RESULTS AND DISCUSSION This paper describes the first part of the present work on understanding the relationships between chemical structure and musk intensity. Twenty compounds including representatives from the three main musk families have been investigated (Figure 1). All

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals

0.00

49

* -0.03

0.0 1

o.oo*

0.00

-0.18

-0.08

H

0.02 0.0 1 0.00

(4)

(3)

0.00

*

0.00

0.02*

(6)

C 0.15*

0.00 oo.o*

-0.24

0.02

*o.oo (7) 0.00 oo .w,

H

(9)

b

C

0 -0.22

Figure 1 The charges on the activity fragments' atoms in the compounds under investigation

50


*o.o I

*-0.0 1

0.14i "\\

-0'39

1

"O2W 0.00

*o.oo

-0.25

(13) 0.02*

0.00

*-0.01

i

o,C),o \ c=@

P *-0.0 1

-0.23

0.00

0.00

(14)

-0.3 1

/

0.02 0 . 0 (15) *o.oo 2 0 C 0.02

o-02Qc=@

-0.27

0.02

*-0.05

Figure 1 Continued

-0.27

*-0.05


51

of these compounds possess a musk odour of varying intensities. The compounds 1-9 have a weak musk odour, and the compounds 10-20 have a moderate to very strong musk odour (see the work in ref.2). The active fragments I and I1 for each compound are shown in Figure 1 by means of circles and asterisks respectively. Some geometrical and electronic characteristics of the compounds are presented in Table 1. Table 1 Some geometrical and electronic characteristics of the organic molecular systems 1-20.

p-dipole moment, Q-complete charge on all carbon and heteroatoms of fragments I and 11, El-energy of HOMO, EZenergy of LUMO The geometrical parameters and the dipole moments (p) were calculated by means of the MM2 method3. In Table 1, R1 is the distance between the electronegative heteroatoms and the methyl (or methylene) groups of the active fragment I. R2 is the distance between the two methyl (or methylene) groups of the active fragment 11. The distance between the methyl (or rnethylene) groups of the active fragment I changes very little for the twenty compounds under investigation. The electronic parameters, which are also resented in Table 1 and in Figure 1 were calculated on the basis of the CND0/2 method (the atomic charges, the energy levels, and the energy gaps). The relative calculation error of this method is small for different sequences of organic compounds, belonging to different structural classes’.’.

B

52


A comparison of the dipole moments and the distances for the active fragments I and I1 (Figure 1 and Table 1) shows that these molecular descriptors alone can not be used to descriminate between the compounds 1-9 with the weak musk odour, and the compounds 10-20 with the moderate or strong musk odour. Both the atomic charges on the individual atoms of the two active fragments and the energy levels for the HOMO (El) and the LUMO (E2) are also incapable of classifying these two classes of musk compounds. However, the value of Q (Table I ) which is the complete charge on all the carbon and heteroatoms of the fragments I and I1 may be a useful discriminating feature. The compounds with a weak musk odour have negative Q values, which are less or equal to -0.21 (all the charges in Table I are given in the electronic charge units). For moderate and strong musks the negative Q values are more negative than -0.21.

Table 2 The dependence of Appel’s Intensity of the musk odour on the values of AE. Compound

Energy (e.V.)

El

Exaltolide Ethylene Brassylate -Ambrettolide Versalide Musk Ambrette Musk Ketone Musk Xylene

- 13.237 -13.071 -1 2.414 -1 1.232 -12.779 - 12.39 1 - 1 2.852

A1

AE

E2 4.763 4.417 4.643 3.205 0.503 0.0 19 -0.460

I

18.000 17.488 17.057 14.437 13.282 12.410 12.312

7

6 6

5 4 3 3

The cycloketones 16 and 17, and the nitro musks 19 and 20 possess the maximum negative Q values, and thus are the strongest musks in this data set. For compounds 1 and 3, their interaction with with an odourant bioreceptor is via their respective bromine and chlorine atoms. These atoms have relatively large van der Waals radii and hence relatively small negative atomic charges. Consequently these compounds have low Q values and thus a weak musk odour (together with the sulphur containing macrocyclic ether 2). However, the value of Q taken alone is not enough for a complete quantitative description of the musk odour intensity. Let us consider the compounds shown in Table 2. These are all strong musk odourants. The quantum chemical calculation shows they have much the same value of Q (Q = -0.31). Nevertheless the Appel’s intensity (AI) of their musk odour is different. It can be seen from Table 2 that for the compounds under consideration there is a clear cut dependence of the values of A1 on the values of the energy gap (AE) between the above mentioned energies E2 and El. An increase in the value of AE by e.V. leads to an increase in the corresponding value of A1 by one unit. Thus a conclusion may be drawn. The musk odour intensity appears to be dependent upon both the complete atomic charge (Q) on the carbon and heteroatoms of the two active structural fragments and the energy gap (AE) between the two frontier molecular orbitals (the HOMO and the LUMO). It seams reasonable to introduce the term 5, to describe the intensity of the musk odour for all compounds under investigation.

C=

IQIAE


I I

53

I

Where Q is the absolute value (module) of Q. 6 is the product of Q I and AE. In the last column of Table 1 are given the values of E, for the compounds 1-20. The moderate to strong musks (10-20) have 5>3.00; for the weak musks (1-9) E, furanone (HDF) 4-hydroxy-5-methyl-3(2H)-furanone (HMF) 2-ethyl-4-hydroxy-5-methyl-3(2H)furanone (EHMF) 3-hydroxy-4,5-dimethyl-2(5H)furanone (sotolon) 3-hydroxy-6-methyl-2(2H)-pyranone Compounds containing nitrogen 2,3-diethyl-5-methylpyrazine 2-acetyl- 1-pyrroline 6-acetyltetrahydropyridine 2-acetylpyridine cysteine/glucose (71), cysteinelrhamnose (71) prolinelglucose ( I 6) prolinelglucose (16) proline/glucose (16)

1.0 (31), 0.09 (41) 0.1 (62) 1.0 (63), 1.6 (64) I9 (63)

earthy, roasty roasty,popcorn-like roasty, burnt caramel-like roasty, caramel-like

seasoning-like

seasoning-like

(74) 20 (43), 1. I5 (52)

caramel-like, sweet

cysteinelribose (70). cysteinelglucose (7 I), cysteine/rhamnose (7 I ) cysteine/rhamnose (7 1)

cysteine/rhamnose (71)

8500 (17), 23000

0.3 (53),0.001 (54) 15.0 (17)

cysteine/ribose (70), cysteine/glucose (7I), cysteinel rharnnose (7I), proline/glucose (16) cysteine/ribose (70)

10 (52), 100 (53)

cysteine/glucose (7I )

Detected inb

caramel-like,strawberrylike caramel-like, burnt chicory

24.0 (I 7)

[pgkg] in water

Odor threshold

caramel-like, sweet, meaty

Odour description

Flavour impact components ofprocessflavouringsa(continued)

Compound

e2

3

s.

9

2

3

3

-_

f:5

F

a

162

Flavoirrs and Fragrances

Scheme 1 : Bicyclic structures of 3-deoxyglucosone

0

o+oH

’

‘CH,OH

OH

HO & ~ ~ o H ’ z 3

Ho*H

5

OH

4

O H

HO

&cH20H

HO

7

8

The Maillard Reaction in Flavour Formation

163

In addition they found differences in the composition of volatiles, depending on whether APR’s were heated dry or in aqueous solution. Heinzler and Eichner” investigated volatiles formation from fructosealanine, under roasting conditions. The main aroma components formed were 2-acetylpyrrole, 5-methylfurfural and 2-acetylfuran. Other flavour substances included pyrazines, pyrroles, pyridines and h a n s . Unfortunately neither van de Ouweland et al. nor Heinzler and Eichner compared their results with the breakdown of the corresponding unreacted amino acid and sugar mixtures. Huyghues-Despointes” el al. and Keyahani et ~ 1investigated . ~ ~ the formation of volatiles from proline-glucose ARP and phenylalanine-glucose ARP using pyrolysidGCNS analysis. They compared the ARP’s with unreacted glucose, phenylalanine and a glucose + phenylalanine mixture. Clear differences were observed between ARP’s and the corresponding amino acidglucose mixtures. Unfortunately the results are difficult to compare with solution Maillard chemistry, as many of the products are quite different from what is obtained in an aqueous process flavouring. 4.2 Thiazolidine Carboxylic Acids

When cysteine is heated in the presence of a reducing sugar, ARP’s are not formed but instead thiazolidine carboxylic acids are formed as intermediate^^^'^^. These thiazolidine carboxylic acids are relatively stable, but decompose at their melting pointsg4.When heated to their melting points, 4-alkylthiazolidinecarboxylic acids were found to give 2,4-dialk 1-2,3-dihydrothiophenes and 2,4-dialkylthiophenes in yields varying from 0.16-18% L . The anionic form of these thiazolidinecarboxylic acids is apparently relatively stable, and has been proposed to inhibit browning and flavour developments5. The reactivity of the thiazolidines can be enhanced by the use of reactive sugars and buffer salts (low pH). The inhibiting effect of the thiazolidine pathway can apparently be avoided by using A m ’ s + cysteineg4. 4.3 Deoxyglycosones

The formation of deoxyglucosones from ARP’s was demonstrated by Beck et a1.86 using diaminobenzene as a trapping agent. ARP’s from glucose and maltose were found to give mainly 1- and 3-deoxyglucosones, but also 4-deoxyglucosone. The role of the deoxyglucosones in the Maillard reaction in foods and the human body has been very well reviewed by Led1 and Schleicher”. The structure of 3-deoxyglucosone was studied by Weenen and TjanSgvg9, who showed that 3-deoxyglucosone in aqueous solution exists as a mixture of five major relatively stable bicyclic isomers, with no free carbonyl moiety present (Scheme I). When heated under acidic conditions, intramolecular condensation takes place, resulting in relatively high yields of 4(hydroxymethyl)fuhralg8. When heated with asparagine methylated pyrazines are formedg8.Surprisingly, the yields of these methylated pyrazines were lower from 3deoxyglucosone, than from glucose or fructose, suggesting that carbohydrate fragmentation in the Maillard reaction is more efficient from other intermediates (see below).


164

Scheme 2: Proposed formation of a-dicarbonylhgments

-

1. Formation of glyoxal

sugar, HRP

RA

glycolaldehyde b -.

[Ol

glyoxal

2. Formation of pyruvaldehyde

AR P

-b

RA glyceraldehyde-aa adduct

\

pyruvaldehyde

1-Done. 3-Done

3. Formation of diacetyl and 2J-pentanedione

1-Done

1. isom. A 2. RA

RA : retroaldolisation isom: isomerisation aa: amino acid

HCHO diacetyl

-b

2,3-pentanedione

165

The Maillard Reaction in Flavour Formafion

Figure 1: Formation of alpha-dicarbonyl compounds f?om carbohydrates and alanine (reprinted with permission h m ref 92. Copyright 1997 Elsevier Science Ltd)

I2O

2

80

-P

60

e 8 0

e .BU

40

I

m

.c

-a 20 m 0

glyoxal

pyruvaldehyde

w diacetyl

Ipentanedione

I66

5


CARBOHYDRATE FRAGMENTATION

Products formed in the Maillard reaction from carbohydrate fragments include pyrazines, thiazoles, carbocylic compounds and other heterocyclic compounds, and are always formed in relatively low yields”. Carbohydrate fragmentation seems to be the yield determing step in these reactions, and therefore deserves special attention. When using cysteamine as a trapping agent for carbonyl containing intermediates, Weenen and Tjan” found that hydroxyacetone was formed from glucose and fructose, but not from 3-deoxyglucosone. They explained this by postulation that hydroxyacetone is formed by P-cleavage from isomerised 1 -deoxyglucosone. More recently Weenen and Apeldoom used 1,2-diaminobenzene as a trapping agent for a-dicarbonyl compounds, and concluded that carbohydrate fragmentation in the Maillard reaction takes place by retro-aldolisation of deoxyglycosones, the starting sugars (or their amino acid adducts) and possibly also of ARP’s and HRP’s9’. These results are in contrast with earlier conclusions based on pyrazine formation studies, but upon closer inspection are in agreement with the results of these experiments”. Four a-dicarbonyl containing fragments were detected: glyoxal, pyruvaldehyde, diacetyl and 2,3-pentanedione (Figure 1). As expected 3-deoxyglucosone was a particularly efficient precursor of pyruvaldehyde, but to our surprise xylose was also a relatively good precursor of pyruvaldehyde. This may be related to the expected lower stability of deoxypentosones. The formation of glyoxal can best be explained as resulting from retroaldolisation of the starting sugar or the corresponding amino acid adduct. The resulting glycolaldehyde should then be oxidised to give glyoxal, for example via a Strecker-like mechanism (Scheme 2). If HRP’s cleave via retroaldolisation, C-3 fragments are formed with only one carbonyl, i.e. glyceraldehyde or hydroxyacetone (+ amino acid adducts). Hydroxyacetone can isomerize to glyceraldehyde, and glyceraldehyde can generate pyruvaldehyde, upon p-elimination of H20. A summary of the proposed mechanisms for the formation of glyoxal, pyruvaldehyde, diacetyl and 2,3-pentanedione from normal sugars, ARP’s, HRP’s, 1- / 3-deoxyosones is given in Scheme 2. 6

DISCUSSION

Aroma character impact components can be determined by either Charm Analysis’’, l9,2 1.22 , or Headspace dilution analysis2’. These Aroma Extract Dilution Analysis methods have been applied to thermally treated foods and to process flavourings. Results so far indicate a large degree of similarity between thermally treated foods and process flavourings, as expected, with as notable exceptions, S-acetyl-2,3-dihydro- 1,4thiazine and 2-( 1 -mercaptoethyl)furan, which have only been detected in process flavourings so far. Research on Amadori rearrangement products, Heyns rearrangement products, thiazolidines and deoxyglycosones is fragmentary, but shows the importance of these reactive intermediates for the course of the Maillard reaction. Results so far suggest that fragmentation occurs from the dexoyglycosones, ARP’s and HRP’s, but also from the starting sugars, or their amino acid addition products. The implications for the overall scheme of flavour formation in the Maillard reaction are summarized in Scheme 3.

Flavour Substances

+

Scheme 3: The Maillard reaction in flavour formation (Reproduced with permission from ref. 92. Copyright 1997 Elsevier Science Ltd)

Ffavotir Substances

168


References

I. 2.

L.C. Maillard, C.R. .4cad. Sci. Ser., 1912.2, 154. M. Amadori, Atti. R. Accad. Naz.. Lincei. Mem. CI. Sci. Fis. Mat. Nat., I93 I , 13, 72. 3. V.A. Yaylayan and A. Huyghues-Despointes, Crit. Rev. Food Sci. Nutr., 1994, 34(4), 32 1. 4. K. Heyns and H. Noack, Chem. Ber., 1962,720. 5. A. Strecker, Ann. 1862, 123,363. 6. A. Schonberg and R. Moubacher, Chem. Rev., 1952,50,261. 7. W.Z. Ruckdeschel, Z. ges. Brauw. 1914,37,430 + 437. 8. W.J. Herz and R.S. Schallenberger, Food. Res., 1960,25,491. 9. I.D. Morton, P. Akroyd and C.G. May, G.B. Patent 836,694, 1960. 10. G. MacLeod and M. Seyyedain-Ardebik, Crir. Rev. Food Sci. Nutr., 1981, 14,309. 11. J.E. Hodge, J. Agric. Food Chem., 1953, 1, 928. 12. H. Maarse, C.A. Visscher, L.G. Willemsens, L.M.Nijssen and M.H. Boelens, ‘Volatile Compounds in Food. Qualitative and Quantitative Data.’ Supplement 5 to the Sixth Edition. TNO Nutrition and Food Research, Zeist, The Netherlands, 1994. 13. R. Tressl, B. Helak, N. Martin, and E. Kersten, ‘Thermal Generation of Aromas’, American Chemical Society, Washington, DC, 1989, 156. 14. M. Guntert, J. Bruning, R. Emberger, M. Kopsel, W. Kuhn, T. Thielmann and P. Werkhoff, J. Agric Food Chem., 1990,38,2027. 15. P. Schieberle, Z. Lebensm. Unters. Forsch, 1990, 191,206. 16. D.D. Roberts and T.E. Acree, ‘Thermally generated flavors’, ACS Symposium Series 543, 1994, 7 1. 17. T. Hofmann, PhD Thesis, Technical University of Munich, 1995. 18. T.E. Acree, J. Barnard and D.G. Cunningham. Food Chem., 1984, 14,273. 19. F. Ullrich and W. Grosch, Z. Lebensm. Unters. Forsch., 1987, 184,277. 20. H. Guth and W. Grosch, Flavour Fragrance J., 1993,8, 173. 21. W. Grosch, Flavour Fragrance J., 1994,9, 147. 22. W. Grosch, Trends Foods Sci. Technol., 1993,4, 68. 23. T.E. Acree, ‘Flavor Science - Sensible Principles and Techniques’, American Chemical Society, Washington, D.C., 1993. 24. P. Schieberle, ‘Characterization of Food-Emerging Methods’, Elsevier, Amsterdam, 1995. 25. M. Rothe and B. Thomas, Z. Lebensm. Unters. Forsch., 1963, 119,302. 26. H. Guth and W. Grosch, J. Agric. Food Chem., 1994,42,2862. 27. E.L. Pippen and E.P. Mecchi, J. FoodSci., 1969,34,443. 28. R.G. Buttery, W.F. fladdon, R.M. Seifert and J.G. Turnbaugh, J. Agric. Food Chem., 1984,32,674. 29. P. Schieberle, personal communication. 30. D.G. Guadagni, R.G. Buttery, and J.G. Turnbaugh, J. Sci. Food Agric.. 1972, 23. 1435. 3 1. C. Cerny and W. Grosch, Z. Lebensm. Unters. Forsch., 1993, 196, 123.

The Maillard Reaction in Flavour Formation

32. 33. 34. 35. 36. 37. 38. 39. 40. 4 1. 42. 43. 44. 45. 46. 47. 48. 49. 50. 5 1. 52. 53.

54.

55. 56. 57. 58. 59. 60. 6 1. 62. 63. 64. 65. 66. 67. 68.

169

U. Gasser and W. Grosch, Z. Lebensm. Unters. Forsch., 1988, 186,489. R. Kerscher and W. Grosch, Z. Lebensm. Unters. Forsch., 1997,204, 3. U. Gasser and W. Grosch, Z. Lebensm. Unters. Forsch., 1990, 190,3. J. Kerler and W. Grosch, Z. Lebensm. Unters. Forsch., 1997, submitted for publication. U. Gasser and W. Grosch, Lebensmittelchemie, 1991,45, 15. I. Blank, A. Sen and W. Grosch, Z. Lebensm. Unters. Forsch., 1992, 195,239. P. Semmelroch and W. Grosch, Lebensm. Wiss. Technol., 1995,28,3 10. P. Schieberle, ‘Progress in Flavour Precursor Studies’, Allured Publishing Corporation, Carol Stream, USA, 1993, 343. P. Schieberle, J. Agric. Food Chem., 1991,39, 1 141. J. Kerler, PhD Thesis, Technical University of Munich, 1996. J. Kerler and W. Grosch, J. FoodSci., 1996,61 (6), 1271. P. Semmelroch and W. Grosch, J. Agric. Food Chem., 1996,44, 537. C. Milo and W. Grosch, J. Agric. Food Chem., 1995,43,459. H. Guth and W. Grosch, Lebensm. Wiss. Technol., 1993,26, 171. C. Cerny and W. Grosch, Z. Lebensm. Unters. Forsch., 1992, 194,322. R. Wagner and W. Grosch, Lebensm. Wiss. Technol., 1997,30, 164. C. Milo and W. Grosch, J. Agric. Food Chem., 1993,41,2076. C. Milo and W. Grosch, J. Agric. Food Chem., 1996,44,2366. M. Rychlik and W. Grosch, Lebensm. Wiss. Technol., 1996,29,515. P. Schnermann and P. Schieberle, J. Agric. Food Chem., 1997,45, 867. P. Semmelroch, G Laskawy, I. Blank and W. Grosch, Flavour Fragrance J., 1995, 10, 1. W. Grosch, G. Zeiler-Hilgart, C. Cerny and H. Guth, ‘Progress in Flavour Precursor Studies’, Allured Publishing Corporation, Carol Stream, USA, 1993, 329. A. Kobayashi, ‘Flavor Chemistry - Trends and Developments’, ACS Symposium Series 338, Washington DC, 1989. S . Fors, ‘The Maillard reaction in Foods and Nutrition’, ACS Symposium Series 215, Washington DC, 1988. J.C. Leffingwell and D. Leffingwell, Perfum Flavor, 1991, 16,2. I . Blank, A. Sen and W. Grosch, ‘ASIC, 14th Colloque, San Francisco, 1991. P. Schieberle, Lebensmittelchemie, 1993,47, 15. P. Schieberle, Food Chem., 1996,55 (2), 145. P. Schieberle and W. Grosch, 2. Lebensm. Unters. Forsch., 1987, 185, 11 1 . H. Holscher, PhD Thesis, University of Hamburg, 1991. R.G. Buttery, L.C. Ling, B.O. Juliano and J.G. Turnbaugh, J. Agric. Food Chem., 1983,31, 823. R. Teranishi, R.G. Buttery, and D.G. Guadagni, ‘Geruchs- und Geschmacksstoffe’, Verlag Hans Carl, Nilmberg, 1975, 178. R.G. Buttery and L.C. Ling, J. Agric. Food Chem., 1995,43, 1878. R.G. Buttery, L.C. Ling, B.O. Juliano, Chem. Ind., 1982,958. P. Schieberle, J. Agric. Food Chem., 1995,43,2442. P. Schieberle and W. Grosch, Z. Lebensm. Unters. Forsch., 1994, 198,292. M. Czemy, R. Wagner and W. Grosch, J Agric. Food Chem., 1996,44,3268.

I70


69. P. Schieberle, Z.Lehensm. Un1er.s. Forsch.. 1991, 193, 558.

70. 7 1. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

89. 90.

91.

T. Hofmann and P. Schieberle, cJ. Agric. Food Chem., 1995,43,2 187. T. Hofmann and P. Schieberle. -J. Agric. Food Chem., 1997, 45, 898. U. Gasser. PhD Thesis. Technical University of Munich, 1990. R. Tress1 and R. Silvar, J. Agric. Food Chem.. 1981,29, 1078. R.G. Buttery, G.R. Takeoka, G.E. Kramnier and L.C. Ling, Lehensm. Wiss. Technol.. 1994,27, 592. T. Hofmann and P. Schieberle. J. Agric. Food Chem., 1996,44, 25 1. W. Grosch and P. Schieberle, FoodRev. In[., 1997, submitted for publication. K.O. Herz and S.S. Chang, Advcin. Food Res., 1970, 18, 1. I.D. Morton, P. Akroyd, and C.G. May, U K Patent 836694, 1960. G.A.M. Ouweland, H.G. Peer and S.B. Tjan. ‘Flavor of foods and beverages’, G. Charalambous (ed.), Academic Press, 1978, 13 1. M. Heinzler and K. Echner, Z. Lehensm. Unlers. Forsch., 1991, 192,445. R. Wittman and K. Echner, Z. Lehensm. Unters. Frosch. 1989,188,212. A. Huyhues-Despointes, V.A. Yaylayan and A. Keyhani, J. Argric. Food Chem., 1994,42,25 19. A. Keyhani and V.A. Yaylayan, J. Agric. Food Chem., 1996,44.223. K.B. de Roos. In: Flavor precursors, R. Teranishi, G.R. Takeoka and M. Guntert (eds.), American Chemical Society, 1992. 203. G.P. Rizzi. A.R. Steinle and D.R. Patton. In: Food Science and Human Nutrition, G. Charalambous (ed.), Elsevier, 1992, 73 1. J. Beck, I;. Led1 and T. Severin, Lehensm. Un/ers. Forsch.. 1989, 188, 118. F. Ledl, E. Schleicher. Angew. C‘hemie In[. Ed. Engl., 1990, 29, 565. H. Weenen and S.B. Tjan, In: Flavor Precursors, R. Teranishi, G.R. Takeoka and M. Guntert (eds.), ACS Symposium Series 490, American Chemical Society, Washington. 1992, 2 17. H. Weenen and S.B. Tjan, In: Trends in flavor research, H. Maarse and D.G. van der Heij (eds.), Elsevier Science B.V., Amsterdam, 1994, 327. H. Weenen and W. Apeldoorn, In: Flavour Science: Recent developments, A.J. Taylor and D.S. Mottram (eds.). Royal Society of Chemistry, Cambridge, 1996,211. H. Weenen, S.B. Tjan, P.J. de Valois. N. Bouter, A. Pos and H. Vonk, In: Thermally generated flavors, Maillard, microwave and extrusion processes, T1i.H. Parliment, M.J. Morello and R.J. McGorrin (eds.), ACS Symposium Series 543, America1 Chemical Society, Washington DC, 1992, 142.

Relationship between Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles A. J. Taylor and R. S . T. Linforth DEPARTMENTAL OF APPLIED BIOCHEMISTRY AND FOOD SCIENCE, UNIVERSITY OF NO’ITINGHAM, SUITON BONINGTON CAMPUS, LOUGHBOROUGH LEI 2 SRD, UK

1. INTRODUCTION

For many years, flavour scientists have studied the relationship between the flavour perceived by consumers and the qualitative and quantitative aspects of the mixture of flavour chemicals that cause perception in humans. Broadly, research has fallen into two distinct areas. The first area has studied single flavour chemicals, particularly those that produce a sensation of sweetness. The sweetening power of individual compounds has been determined by sensory analysis and then related to the amount of compound administered, to derive a quantitative relationship. These mathematical relationships are often referred to as the Psychophysical Laws.’,’ The second area has studied foods (or model systems) which contain complex flavour mixtures and attempted to derive relationshipsbetween the flavour composition and the perceived flavour of the sample. To implif) the problem, some workers have tried to quantifjl individual flavour compounds and relate the concentration of these individual compounds to flavour characteristics e.g. the gassy note of a flavour might be related to the amount of hexanal or the total amount of the Ca aldehydes and alcohols. While this approach reduces the complexity of the ?roblem, it has been criticised as being too simplistic (see for example Bootp ). Since some flavour molecules can interact with more than one sensor in the nose (or in the mouth and nose in some instances), there is some evidence to suggest that the relationship between perception and the amount of a compound administered, depends also on the concentrations of other compounds present in the flavour. Recent developments in analytical methodology have allowed the concentration of flavour volatiles to be measured in the nose during eating4 at concentrations around 10ppbv which brings the sensitivity into a range where many compounds have their odour rhreshold values and which is therefore useful to flavour scientists. Our laboratory has developed an interface for an Atmospheric Pressure Ionisation Mass Spe~trometer~’~ and this provides a measure of volatile concentration with the time of eating (typically 20 to 60 seconds). The data produced by this technique, now include a temporal dimension as concentration of the volatile changes with time. The purpose of this paper is to consider how these data might be correlated with measurements of sensory perception using simple food systems and simultaneous Time-Intensity sensory measurement combined with real time in-nose measurements of volatile concentration. Some background to the problem is

172


presented initially, followed by preliminaty experimental data obtained t?om gelatin gels and then from starchlsucroseconfectionery. L.1 Psychophysical Laws

The relationship between sweetening power and concentration of the sweetening agent was first investigated last century with the results laying the foundation for psychophysics, a discipline that has been recently reviewed by Hoppe.’.’ This work has led to a number of “Laws” which relate the stimulus (S) applied (i.e. the concentration of the flavour compound) to the response (R) or perceived flavour intensity by assuming there is a mathematical relationship between R and S. The relationship proposed by Stevens (also known as the Power Law) has been widely used. Stevens Law

R=aSb

(1)

‘The exponent b can be obtained by subjecting people to a series of stimuli of different concentrations,determining the response and then plotting Ln R against Ln S. The slope of the line is b and the intercept is a. The value of b varies with experimental conditions and with the range of concentrations used according to Hoppe.’ Stevens obtained a value of 1.3 for b in the case of sweetness associated with sucrose.’ 1.2 Correlation of composition with sensory properties

The correlation of sensory properties with the composition of complex flavours is subject to the limitations explained in the Introduction. Nevertheless, several studies have derived empirical relationships by collecting data and subjecting them to various mathematical treatments. This approach ignores any possible mechanisms for olfaction, nor does it consider interaction between volatiles in a direct way although the maths may indicate such interactions. This chapter is not intended as a complete review of the literature but mention should be made of the work on Cheddar cheese flavour by Bacremont & Vickers’ as well as work on tea by Togari and coworkers.’ This latter paper is a good example of what can be achieved by correlation of volatile composition with sensory properties, Togari and co-workers analysed many samples of tea for total volatile composition imd then correlated these data with sensory attributes for the samples. They attempted to describe sensory attributes like “flowery” in terms of the mixture of compounds that correlated (positively and negatively) with that attribute. They described the rationale to their work clearly in their paper and stressed that a sufficiently wide range of samples must be analysed to obtain good quality data. An example of the results they obtained is shown below. Sweet floral = -0.059l[pentanal] -0.671[2-heptanone] + 0.562[linalooI] +0.693[2-phenylethanol]+0.713[jasminelactone] - 0.134 It is not clear how widely applicable this method is although, presumably, it would need to be validated for each food product. In particular, is the method capable of discriminating subtle differences in flavour (and their relationship to compositional changes) or is it just a crude indicator of quality? Further studies will no doubt answer these questions.

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

I73

Delahunty" and co-workers studied the volatile compounds present in the buccal lieadspace during eating of cheeses and attempted to correlate the amounts of volatiles present over the time of eating with the perceived quality of the cheeses. Samples were obtained over four time ranges (0 to 15; 0 to 30;0 to 45; 0 to 60 sec) which produced a cumulative trace of volatiles with time due to the experimental method employed. Delahunty reported that straight buccal headspace analysis allowed prediction of sensory attributes but time course studies improved predictive ability. This finding suggests that the time course of release does play a role in flavour perception although the relationship is not clear. 1.3 Mechanisms of olfactory perception

Recently, understanding of the structure of olfactory receptors and their arrangement both in the nose and within structures of the brain (reviewed by Shepherd at the Warwick meeting) has given the potential to model the relationship between the interaction of a volatile with the receptor and the subsequent sensory signal. This helps in the design of appropriate neural network models which could show the type of neural signal that might be produced by any given mixture of volatiles. However, this information is still some distance away from the verbal description of flavour perception on which we rely for consumer assessment of a food product. During the first session of the meeting, the problems that humans encounter when trying to describe odours by use of appropriate words were discussed. Van Toller had demonstrated in his presentation that people were more able to describe verbally, shapes or colours rather than flavours and it was suggested that human recognition of flavour might be more akin to recognition of human faces i.e. it is easy to recognise people by their faces but extremely difficult to describe verbally the features that distinguish one face from mother. This assertion suggests that the pattern of volatiles sensed by the nose (and other sensors) may be the key to recognition and, therefore, that attempts to relate sensory and compositional data should follow this route. This is especially significant when it is considered that most flavours are due to a mixture of chemicals and, while there are certain character impact compounds, they rely frequently on the presence of other compounds to produce a good flavour sensation. Booth and co-workers3 have described an interesting approach to explain the recognition of strawberry aroma. Their data suggests that recognition depends on the combination of two separate olfactory stimulation patterns and their work presents a unique approach to the whole subject of sensory recognition. 1.4 Rationale for measuring in-nose concentration

The rationale for measuring the concentration of volatiles in the expired air of humans during the consumption of food has been discussed in detail e l ~ e w h e r e ~and ~'~'~ only a brief overview is given here. Basically, there are three types of volatile extraction und analysis, each of which has its merits. Determination of the fotal volatile profile of a food is often achieved by steam distillation or solvent extraction of the food followed by GC-MS. Similarly, the volatiles above a food, the hedpace, can be determined under a range of conditions of temperature, water content etc. either at equilibrium or under dynamic conditions. However, neither of these analyses takes into account the changes that take place on eating and it has been shown that, in many foods, the volatile profile measured in expired air from the nose (nosespace or b r e d by breath) is very different

174


tom the headspace profile in both qualitative and quantitative terms

.~.. 'J.'*

Current thinking

is that the volatile profile measured by nosespace analysis must be very close to the profile

experienced by the olfactory epithelium and thus, if the relationship between volatiles and perception is to be studied, then this seems the logical type of instrumental analysis to perform. 1.5 Measurement of in-nose or in-mouth concentration

Previous used trapping of volatiles on adsorbents like Tenax, followed by GC-MS analysis to record the volatiles released over periods of time during eating. Using MI-MS, with the interface developed at Nottingham, it is possible to monitor 12 to 20 ,volatiles simultaneously in real time and provide a record of the way that the concentrations of volatiles in the nose or mouth change during eating and after swallowing. In our nosespace method, the subject inserts a soft plastic tube into one nostril and air is sampled by a second capillary tube that fits into the soft plastic tube at right angles. This arrangement allows subjects to breathe normally with no great hindrance. Since the MI-MS samples tidal air from the subjects, the inspired air contains only those compounds present in the laboratory air while the expired air contains products of metabolism from the lungs (e.g. acetone) as well as volatiles released in the mouth and rransported retronasally to the nose. The traces obtained show breath-by-breath release md Figure 1 indicates the behaviour of two different volatiles (ethyl butyrate and ethanol) released from a gelatidsucrose gel as well as the presence of a liver metabolite (acetone) in the expired air of a healthy male volunteer

Ethyl Butyrate 14 01

loo

* 0

Ethanol

,100

I

*

~

1I

I_

0 1

3

~

~ 1340

1380

1380

1400

l

S

R-C-OR

II

S

0 R-C-SR'

'c= RJ

-I

R-C-SR'

II

S

Figure 1 Lawesson's Reagent and Typical Thiation ofCarbonyls

21 1

Synthesis and Application of Thiocarbonyl Compounds

2.1 Synthesis of Thiolactones The delta and gamma lactones have been identified in various h i t s , milk, and butter, and the lactones are often key components of the flavour. In this chapter the Authors describe the synthesis of thiolactones. A range of gamma lactones from butyro to dodecalactonewere converted to their corresponding thiolactones. The gamma lactones were mixed with half a mole equivalent of Lawesson's reagent in toluene and then refluxed for three hours with stimng. After the reaction was complete the Lawesson's reagent was removed by filtration and the thiolactones were isolated by distillation and column chromatography(silica gel, mixed solvent of hexane/ether). The corresponding delta lactones were refluxed in xylene for five hours, and then purified as for the gamma lactones. In result, all lactones were converted to the correspondingthiolactones (Figure 2). The yield of the delta lactones was lower than that for the corresponding gamma lactones. The reason behind this is currently being researched. As the electric environment around the carbonyl group is similar, the Authors assumed that the conformation of the side chain is directly connected to the reactivity. The most stable conformation of the delta lactone ring is the chair type. When a side chain is in an axial position, the carbonyl and side chain would be in close proximity to each other, thus this environment may disturb the reaction.

Lawesson's Reagent in Toluene I Reflux I 3 h

H

:a92%

C4H9 : e 91%

.

C5H11 f 75% C6Hi3 : g 8 1 % OH15 : h 79% C8Hi7 : i 88%

Lawesson's Reagent in Xylene I Reflux I 5 h

C4H9 : a47% C5Hii : b 63% c6H13 : C 5 5 % C7Hi5 : d 6 9 %

Figure 2 Thiation of y and &lactones by Lawesson's reagent

212


The products of the reaction were identified by instrumental analysis. The Infra red spectrum shows the absorption of a typical lactone carbonyl to be 1772 cm”. In the thiolactone products the absorption has shifted to 1238 cm”. The mass spectrum showed that the molecular ion peaks had increased by 16 after the reaction. These results were obtained for each delta and gamma lactone reaction product. (Table 1)

Table 1 IR and MS Analysis Data of Thiolactones

IR Spectral Data (vmax cm-I) 1772 (C=O)

1238 (C=S)

MS Spectral Data ~~~~

R

Lactone

H CH3 C2H5 C3H7 C4H9 CSHl 1 C6H 13 C7H 15 C8H17

m/z 86(M+) m/z 100(M+)

m/z 114(M+)

m/z 128(M+) m/z m/z m/z m/z m/z

142(M+) 156(M+) 170(M+) 184(M+) 198(M+)

Thiolactone m/z 102(M+) m/z 116(M+) m/z 130(M+) m/z 144(M+) m/z 158(M+) m/z 172(M+) m/z 186(M+) m/z 200(M+) m/z 2 14(M+)

IR Spectral Data (vmax cm-I) 1772 (C=O)

1238 (C=S)

MS Spectral Data R C4H9 C5H11 C6H 13 C7H15

Lactone m/z 156(M+) m/z 170(M+) m/z 184(M+) m/z 198(M+)

Thiolactone m/z 172(M+) m/z 186(M+) m/z 200(M+) m/z214(M+)

Synthesis and Applicarion of Thiocarbonyl Compounds

213

The 13C-NMR data is shown in Table 2. The chemical sift of the carbon at No.2 position has obviously changed from 172 ppm to 222 ppm during the reaction. Every delta and gamma lactone commonly showed these results. Table 2 IH and 13C NMR Data of Thiobctones I I

I

R= C6H13 (y-Decathiolactone)

1H N M R 6 (PPm) 4.87 (H) 3.04 (2H) 1.79 1.98 (2H) I .61 1.75(2H) 1.21 1.58 (8H) 0.90 (3H)

---

I

13C N M R

6 @pm) 222.48 C2 O172.32 (C=O) 44.85 C3 34.05 C4 91.05 CS 31.53 29.56 28.90 25.21 22.40 13.95

R=CSH 1 1 (GDecathiolactone) IHNMR

13C N M R

6 @pm)

6 (PP4

4.35 (H) 2.92 (2H) 1.92 2.00 (2H) 1.56 1.72 (2H) 1.24- 1.38(8H) (3H) 0.90

223.60 40.65 35.19 26.89 83.58 31.55 24.73 22.46 18.09 13.97

--

c2 *174.50 (C=O) C3 C4 C5 C6

2.2 Application of thiolactones to flavour and fragrance. The thiolactones obtained from the reactions showed unique and strong odours. The odour ofgamma butyrothiolactone was strong, green, oily, and fatty. The thiocarbonyl group

214


seemed to especially impart the odour character of green. Gamma valero and gamma hexathiolactones had similar green odours. Gamma heptathiolactone showed a mellow, green, and fatty odour. From gamma octa to gamma undecathiolactone, the odour was also strong, wonderfully mellow, green, and fatty. The corresponding delta thiolactones had also changed their odour clearly from their lactone analogues. Delta nonathiolactone had a mild sulphury green and ether like note. Delta decathiolactone had a mellow, green, and fatty odour, delta undecathiolactone was sulphur. Both the gamma and delta thiolactones show unique strong green notes, and because of this they can be used for the creation of flavours and fragrances. The Authors applied the thiolactones to flavour and fiagrances and examined the results. (Table 3). The butter flavour with thiolactones included had a mellow, natural, and long lasting odour and taste. The white peach flavour with thiolactones included showed a mellow, natural, and gave the impression of freshly squeezed juice from a well ripened peach. The lily of the valley fragrance including gamma undecathiolactone had a mellow and long lasting scent. These results suggest that the thiolactones give volume, richness, and natural freshness to fruits, daily products, cooking flavours, and that they give body and naturality for fragrances.

Table 3 Application of thiolactones to Flavour and Fragrance

Butter Flavour Ingredient Diacetyl Butyl Acetate Acetoin Butyric Acid Hexanoic Acid Octanoic Acid 2-Heptanone 2-Nonanone GNonalactone GNonathiolactone &Decalactone GDecathiolactone GUndecalactone BUndecathiolactone GDodecalactone BDodecathiolactone Propylene glycol Ethanol

Formula A

Formula B

2.0 2.0 4.0 6.0 8.0 4.0 I .o 2.0 5.0

2.0

40.0 1 .o

25.0

200.0 700.0

2.0 4.0 6.0 8.0

4.0 I .o 2.0 5.0 0.1 40.0 0.1 I .o 0.1 25.0 0.1

200.0 700.0


215

Table 3 Continued. White Peach Flavour Ingredient Ethyl maltol Vanillin Ethyl Acetate Ethyl Myristate Benzaldehyde Benzyl alcohol Linalool 2-Methyl butyric acid y-Octalactone y-Oetathiolactone y-Decalactone y-Decathiolactone GDecalactone &Decathiolactone y-Undecalactone y-Undecathiolactone GDodecalactone &Dodecathiolactone NPC Peach Base Propylene glycol Ethanol

Lily of the Valley fragrance Ingredient Hydroxycitronellal Linalool Ylang ylang oil Jasrnin absolute Rhodiol Terpineol Benzaldehyde Heliotropine Caldamon oil Benzyl propionate Benzyl salicylate y-Undecalactone (1 0%) y-Undecathiolactone (10%) Vanillin Anisyl formate Cyclamen aldehyde Undecenal Civetone (3%) Citronellyl acetate

Formula A

Formula B

0.1 0. I 5.0 8.0 1.o 30.0 7.0 3.0 1.o

0.1 0.1

3.0 5.0

20.0 3.8 10.0

200.0 700.0

5.0 8.0 1 .o

30.0 7.0 3.0 1.o 0.1 3.0 0.1 5.0 0.1 20.0 0.1 3.8 0.1 10.0 200.0 700.0

Formula A

Formula B

300 150 20 30

300 150

20 30

150

I50

I50

150

1 25

I 25 I

I 15

15

60

60 0.9

1

2

30 3 2 20 40

0.1 2 30 3 2 20 40

216


References 1. J.W.Scheeren, P.H.J.Ooms and R.J.Nivard, Synthesis, 1973, 149. 2. H.Hoffmann and G.Schumacher, Tetrahedron Lett., 1967,31,2963. 3. S.Oae, A.Nakanishi and N.Tsujimoto, Chem. and Ind., 1972,575. 4. P.Salama, M.Poirier and M.Caissie, Heterocycles, 1995,41, 1 1,2481. 5. B.S.Pedersen, SScheibye, N.H.Nilsson and S.O.Lawesson, Bull. SOC.Belg. 1978,87,3, 223. 6. BSPedersen, S.Scheibye, K.Clausen and S.O.Lawesson, Bull. SOC.Belg. 1978,87,4, 293. 7. S.Scheibye, J.Knstensen and S.O.Lawesson, Tetrahedron, 1979,35, 1339.

3 OTHER FUNCTIONALITY OF THIOLACTONES

It is well known in Japan that some plants that contain organosulphur compounds have special bioactivity. For example wasabi (Japanese horseradish) has antimicrobial activity, and garlic is known to activate the human body"'. In this chapter the Authors describe the antimicrobial activity and Superoxide dismutase like activity of the thiolactones. 3.1 Antimicrobial Activity The antimicrobial activity was anaylzed by the paperdisk diffusion method. The procedure is as follows, firstly the organisms diluted in a suitable solvent were coated onto the culture medium. Secondly the paper disk is set onto the medium, and finally the solution being tested is added to the disk, and everything is then incubated for 24 or 48 hours. If the sample has antimicrobial activity an inhibitory circle will form around the paper disk. The diameter of this inhibitory circle is measured as the intensity of antimicrobial activity. The organisms tested are shown in table 4, and the results of the tests are shown in table 5. Activity was shown against only E. coli and S.aureus. The highest activity was shown by the low molecular weight gamma thiolactones. The high molecular weight derivatives showed a much weaker effect. The delta thiolactones showed the same trend. Gamma butyro, gamma valero, and gamma hexathiolactones may be used as antimicrobial additives. Table 4 Examined Organisms

Staphylococcus aureus Bacillus subtilis Escherichia coli Saccharomyces cerevisiae Penicillium chrysogenum Aspergillus niger

I F 0 3060 IF0 1213 IF0 12734 sw-33 I F 0 6352 IF0 6342

217


Table 5 Intensity of Antimicrobial Activity

Sample

dose

Intensity (mm) E. coli S. aureus

13 7 11 6

y-Heptathiolactone y-Octathiolactone y-Nonathiolactone y-Decathiolactone y-Undecathiolactone y-Dodecathiolactone

1000 1000 1000 1000 1000 1000

8 6 4 3

11 5 10 4 5 3 3 2 2 2

1 0

0

6-Nonathiolactone 6-Decathiolactone GUndecathiolactone bDodecathiolactone

1000 1000 1000 1000

3 1 1 0

2 2 1 0

y-Butyrothiolactone

1000

500 y-Valerothiolactone

1000

y-Hexathiolactone

1000

500 500

10 4

1

not active against other organisms

3.2 Superoxide Dismutase Like Activity

As the second functional property superoxide dismutase like activity was examined. The superoxide radical is well known and has been studied greatly'". These radicals are generated inside our bodies by autoxidation of haemoglobin or other systems. The radicals damage DNA or denature proteins, thus can initiate cancers or cause ageing. Our bodies have a natural inhibitor of superoxide anion radical, it is the enzyme called super oxide dismutase (SOD). The SOD reduces the superoxide anion to hydrogen peroxide, which is then converted to water by the catalase enzyme. It has been reported that some extracts of herbs reduce the superoxide anion radical just like SOD. Some reports describe some organosulphur compounds as having SOD like activity'. The Authors examined the activity by the NBT method'". The outline of the NBT method is shown in figure 2. This method uses xanthine and xanthine oxidase as the generator of superoxide anion radicals. Initially xanthine is oxidized to uric acid by xanthineoxidase, it is during this oxidation that the superoxide radicals are generated. The sample and nitroblue tetrazolium (NBT) are then mixed into this system. When a thiolactone has SOD like activity, a proportion of the superoxide radicals are reduced to hydrogen peroxide, the

218


rest react with NBT causing a change to the formazan form. At this time the colour of NBT changes from yellow to blue. The intensity of activity is found by measuring the absorbance of the formazan form. If a sample has a very strong SOD like activity, the superoxide radical anions will be consumed, thus leaving too few to cause a colour change. The intensity of SOD like activity was assessed by the following expression.

EXPRESSION (B-(A-C))IBxlOO (Yo) A; is absorbance of 560 nm of sample with XOD system. B; is absorbance of 560 nm of only XOD system. Without sample system. C; is absorbance of sample without XOD. When sample is coloured, the absorbance should be modified.

Nltroblue tetrazolium (yellow)

/“IJric acid

f Formazan (blue)

Diformazan (blue)

Figure 3 Outline of NBT Method (Xanthine-xanthinoxidase, Nitroblue tetrazolium method) The results of the examination of SOD like activity are shown in table 6 . Only gamma thiolactones were examined, and the strongest activity was seen from the higher molecular weight derivatives. Gamma deca, and gamma undecathiolactones showed especially high inhibitory ratios. These results show the opposite in terms of molecular weight from that found for antimicrobial activity. Gamma deca and gamma undecathiolactones can be used as functional flavour or fragrance ingredients.


219

Table 6 Intensity of SOD like Activity Inhibitory ratio Sample y-Butyrothiolactone y-Valerothiolactone y-Hexathiolactone y-Heptathiolactone y-Octathiolactone y-Nonathiolactone y-Decathiolactone y-Undecathiolactone y-Dodecathiolactone

(YO) 7 15

6 16 29 50 75 92 47

References 1. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.95. 2. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.99. 3. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.199. 4. A.Battistoni, S.Folcarelli, R.Gabbianelli, C.Capo and G.Rotilio, Bi0chem.J.,1996,320, 713. 5 . G.Cao, E.Sofic and R.L.Prior, J.Agric. Food Chem.,1996,44,3426. 6. A.Rotman, Mol. Pharmacol, 1976,12,887. 7. T.Kuramoto, Shokuhin to kaihatsu, 1992, 27,3,22. 8. T.Miyazawa, Food Chem., 1995,119,27.

4 REDUCTION OF THE THIOCARBONYLS TO THIOLS To obtain more aroma chemicals pocessing a unique odour, the reduction of the thiocarbonyl group to the corresponding thiol was examined. Ethyl cinnamate, cis-jasmon, menthyl acetate, camphor, and menthone were converted to their corresponding thiocarbonyl compounds using Lawesson‘s reagent (Figure 4). The thiocamphor and thiomethone have unique green, and bumt sugar like odours. Both of these two compounds were then reduced with lithium aluminium hydride to their corresponding thiols (thiobomeol, and thiomenthol)(Figure 5). The thiobomeol had a cool, green, and tropical fruits like odour, whilst the thiomenthol had a strong green, mint like odour

220


Ethyl cinnamate

P

0

S

sweet. fnrity, balsamic

I

long lasting. honey like and masted

fruity and jasmin like

caramalic, honey like, burnt. masted

Menthyl acetate

-

-

----›

fruity. bergamot and lavendar like

Q.-Isweet, cooked vegetables

Figure 4 Thiation of Other Aroma Chemicals

Lawesson's Reagent Camphor

Thiocamphor

Thioborneol

camphoric

nutty. burned sugar

cool, tropical fruits

I

Menthone Menthol like

Thiomenthone Caramalic. burned sugar

Thiomenibol green, tropical fruits, grapefruit

~~

Figure 5 Reduction of Thiocarbonyl Compounds


22 1

5 CONCLUSION Various unique flavour and fragrance chemicals were obtained by the conversion of carbonyl compounds to thiocarbonyl compounds. Especially the thiolactones have a high potential to be key compounds for use in flavours and fragrances. Some thiolactones showed bioactivity thus they may be used in the future as functional additives. The development of new flavour chemicals is expected from the reduction of thiocarbonyls to thiols.

Heteropoly Acids as Catalysts for Fine Chemicals Synthesis Ivan V. Kozhevnikov LEVERHULME CENTRE FOR INNOVATIVE CATALYSIS, DEPARTMENT OF CHEMISTRY, THE UNIVERSITY OF LIVERPOOL, LIVERPOOL L69 3BX, UK

1 INTRODUCTION

Catalysis by heteropoly acids (HPAs) and related polyoxometalate compounds is a field of growing importance.'-" Many new and exciting developments are taking place in the field in both research and technology. This paper concentrates mainly on the use of HPAs as acid and redox catalysts for low-temperature, liquid-phase organic reactions relevant to the synthesis of fine chemicals. A series of reviews, discussing properties of HPAs and various aspects of HPA catalysis have been published.'-'' 1.1 Structure

HPAs are complex proton acids incorporating polyoxometalate anions (heteropolyanions) having metal - oxygen octahedra as the basic structural units.'' The first characterized and the best known is the Keggin heteropolyanion XMIz04,"-* where X is the central atom (Si4+, Pst, etc.), x is its oxidation state, and M is the metal ion (Mo6+,W6+, V5+,etc.). This anion is composed of a central tetrahedron XOa surrounded by 12 edge- and comer-sharing metal - oxygen octahedra M 0 6 (Figure 1).l2

Figure 1 The Keggin structure of a-XM,204t-s Among a wide variety of HPAs, the Keggins are the most stable and more easily available; these are the most important for catalysis. In this paper HPAs are understood as the Keggin acids, unless otherwise stated. Generally, their formulas will be

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis

223

abbreviated to XM, e.g., PW and SiW for H3PW12040 and H4SiWIZ040, respectively. 1.2 Properties

Why are HPAs interesting for catalysis? HPAs have several advantages as catalysts which make them environmentally and economically a t t r a c t i ~ e . ~They . ~ . ~have: (1) Discrete and mobile ionic structure. The structure is composed of discrete heteropolyanions loosely linked by countercations, unlike the network structure of e.g. zeolites and metal oxides. The structural mobility of solid HPAs is important for their use in heterogeneous catalysi~.~.~ (2) Very strong Bronsted acidity. HPAs are stronger than conventional mineral acids and solid acids such as H,SO, Si0,-A120,, zeolites, etc. Some solid HPAs, e.g. PW and its acid Cs salt CS,,~H,,PW, become superacids after thermal re treatment.^ (3) Redox properties. Particularly important is the ability of some polyoxometalates to reversible multielectron redox reactions under mild conditions as well as the ability to activate such common oxidants as 0, and H,O,.””” Moreover, acidbase and redox properties can be varied by changing the chemical composition of HPAs. (4) Very high solubility in water and oxygenated organic solvents such as lower alcohols, ethers, etc.; but HPAs are insoluble in hydrocarbons. ( 5 ) Fairly high thermal stability in the solid state which allows to use HPAs in heterogeneous catalysis at moderately high temperatures. (6) Such a unique property as “pseudoliquid phase”,’.’ etc. Not surprisingly, HPAs make efficient acid, redox and bifunctional catalysts both in homogeneous and heterogeneous systems. HPAs are used as model systems for fundamental research, providing unique opportunities for mechanistic studies at the molecular level. At the same time, they become increasingly important for applied catalysis. 1.3 Industrial application

In the last two decades, the broad utility of HPA catalysis has been demonstrated in a wide variety of synthetically useful selective transformations of organic substances.’-’ Several new industrial processes based on HPA catalysis have been developed and commercialized, for example oxidation of methacrolein, hydration of olefins (propene, n-butene and isobutene) and polymerization of tetrahydrofuran.’ There are also a few small-scale industrial processes relevant to tine chemicals synthesis to be discussed later.

’

2 ACID CATALYSIS 2.1 Advantages

Generally, HPAs have the following advantages over conventional acid ( I ) HPAs offer broad operational choice - they can be used in homogeneous, heterogeneous or biphasic systems. (2) Being stronger acids, and, therefore, more efficient proton donors, HPAs are generally more active catalysts than mineral acids and conventional solid-acid catalysts.

224


In particular in organic media, the molar catalytic activity of HPAs is 100 to 1000 times higher than that of H,S0,.‘6 (3) Efficient operating under milder conditions which is prerequisite of higher selectivity. (4)Lack of side reactions like sulfonation, chlorination, etc. due to the inertness of heteropolyanions towards organic molecules. ( 5 ) Heteropolyanions can play important role in reaction promoting by stabilizing organic intermediates, e.g. carbenium ions.’ 2.2 Organic reactions

Below are given some selected examples of liquid-phase acid-catalyzed reactions, demonstrating that HPAs in many cases offer strong options for more efficient and cleaner processing. 2.2.1 Homogeneously catalyzed reactions. The HPA-catalyzed hydration of C3-C4 olefins (Eq. ( I ) ) is an industrially important reaction, the hydration of propene being the first commercial process based on HPA catalysis.’ RCH=CH2 + H2O + RCH(OH)CH3

(1)

The hydration of isobutene is used for the separation of isobutene from the C4 hydrocarbon stream produced by cracking. As the catalyst, a concentrated aqueous solution of HPA is used. Compared to mineral acids, such as H2S04, HNO,, and HC104, HPA is 2-4 times more active per equal H’ concentration and shows a higher selectivity, minimizing side reactions such as isobutene ~ligomerization.””~ Mechanistic studies” showed that heteropolyanions play important role in promoting the reaction by stabilizing intermediate carbenium ions. HPAs are reportedly more efficient catalysts than H2S04 and HClO4 in the hydration of phenylacetylene’ (Eq. (2)). PhCECH + H2O + PhCOCH3

(2)

Cycloalkenes are hydrated to cycloalkanols with 99 YOselectivity in the presence of a catalyst consisting of a concentrated aqueous solution of an arylsulfonic acid and tungsten HPA.” PW and SiW are efficient catalysts for the homogeneous hydration of camphene to isobomeol, which is an intermediate in the synthesis of camphor.16 Synthesis of glycosides catalyzed by HPA is of industrial importance.’ Glycosides are used as new effective and biodegradable surfactants. HPA is several times more active than the conventional catalysts such as toluene sulfonic acid and ZnC12. Thus acetylated monosaccharides interact readily with alcohols in a homogeneous phase in the presence of 2% of HPA with respect to the sugar derivative at 7O-13O0C, yielding 7090% of glycosides” (Eq. (3)).

Heteropolyacids and Related Compounds as Catal.ystsfor Fine Chemicals Synthesis

AcO S Ac

O

A OAc

c

+

RoH

-

225

OAc P Ac.

c

O

OAc ~ O +R AcOH

The synthesis of vitamin E with HPA proceeds via the condensation of trimethylhydroquinone with isophytol (Eq. (4)). HPA provides the same yield and quality of vitamin E as ZnCl,, which is the best industrial catalyst. But in contrast to ZnCI,, the HPA catalyst is used in much smaller amounts and can be recycled.'*

2.2.2 Biphasic reactions. Separation of products and recovery and recycling of a catalyst often becomes much easier if a homogeneously catalyzed reaction can be performed in a biphasic system. HPAs due to their special solubility properties, i.e., high solubility in a variety of polar solvents and insolubility in nonpolar solvents, are suitable catalysts for operating under phase-transfer conditions. Polymerization of THF is used for the preparation of polyoxytetramethyleneglycol (PTMG), which is employed for manufacturing Spandex fibers and polyurethanes. PTMG is commercially produced by a two-step process, including ring-opening polymerization of THF with acetic anhydride catalyzed by HC104, followed by hydrolysis of the terminal acetate groups in the prepolymer. Aoshima et have developed a one-step process for the THF polymerization to directly yield PTMG (Eq. (5)) in the presence of PW as catalyst. nTHF + HzO + HO-[-(CHZ)~-O-],-H

(5)

The reaction proceeds in a two-phase system. The upper (product) phase is a PTMG solution in THF. The lower (catalyst) phase is a concentrated solution of HPA in THF. The polymer is formed in the HPA phase and transferred to the product phase. The process is performed continuously. The PTMG with a molecular weight of 500-2000 and a narrow molecular weight distribution is obtained from the product phase. PMo, PW, and SiW catalyze the biphasic cyclotrimerization of aldehydes, such as ethanal, propanal, butanal, 2-methylpropanal, etc., to produce 2,4,6,-trialkyl-1,3,5trioxanes in high yields (Eq. (6)), as reported by Sat0 et a/.''

R

Catalyst turnover number is more than 10000 for the propanal cyclotrimerization. At

226

Flavours a n d Fragrances

high conversions of aldehyde, the reaction mixture spontaneously separates into two phases, a product phase and a catalyst phase, which, depending on aldehyde can be solid or liquid. For the propanal cyclotrimerization, the reaction mixture separates into two liquid phases, and the recovered catalyst phase can be easily reused. Similarly, the formation of acetals between alcohols or diols and a range of aldehydes or ketones is catalyzed by HPAs in two-phase system.2’ HPAs catalyze acetoxylation and hydration of dihydromyrcene to yield dihydromyrcenol and dihydromyrcenyl acetate (Eq. (7)) which are useful as perfume ingredients.” These reactions occur with a 90% selectivity at room temperature in a twophase system. The heterogeneously catalyzed conversion with Si0,-supported HPA is also feasible.

OAc

OH

The alkylation of hydroquinone with isobutene to yield 2-t-butylhydroquinone and 2,5-di-t-butylhydroquinone catalyzed by H3PW12040,H6PzW18062and H6PzWz1071 under phase-transfer conditions in a biphasic system has been reported.’’ 2.2.3 Heterogeneottsly catalyzed reactions. The advantage of heterogeneous systems over homogeneous is easy separation of catalyst from reaction products. HPAs are very efficient and versatile catalysts for alkylation, dealkylation and transalkylation of These reactions are widely used for the preparation of antioxidants, bioactive substances, positional protection, etc. Bulk and 90,-supported PW and SiW are much more active than H,SO,, ion-exchange resins, aluminosilicates, etc. Even Nafion’, a polymeric perfluororesinsulfonic acid comparable in its strength to 100% H,SO,, is less active than PW.’4 Alkylation of p-cresol by isobutene with the use of HPA has been commercialized in Russia. It is a step in the synthesis of antioxidants. The use of HPA instead of H2SO4 provides the gain in selectivity of 7- 10% and almost completely eliminates toxic water pollution.’ HPAs and their salts are promising solid-acid catalysts for the Friedel-Crafts reactions, to replace the conventional homogeneous catalysts such as AICI, and H,SO,, which bring about serious environmental and operational problems.’ Insoluble acidic salt Cs, ’H,,PW shows high efficiency in acylation of activated arenes, such as p-xylene, anisole, mesitylene, etc., by acetic and benzoic anhydrides and acyl chlorides. This catalyst provides higher yields of acylated arenes than the parent PW, the latter being partly soluble in the reaction mixture.2b Esterification of dipicolinic acid with butanol is a step in the synthesis of pharmaceuticals (Eq. (8)). PW as a homogeneous catalyst is almost as efficient as sulfuric acid, yielding 100% of diester. An acidic salt, Ce,,,H,,PW, practically insoluble in butanol, was found to be fairly active as a heterogeneous catalyst. Although less

221

Heteropolyacids and Related Compounds as Catalystsfor Fine Chemicals Synthesis

active than homogeneous PW, it can be easily recovered and reused.” + 2BuOH

H02C

+ 2H20

(8)

BuO$ f i 0 2 B u

The insoluble salt C S ~ , ~ H ~ .and ~ P even W PW itself, which is highly soluble in water, can be included in the silica matrix by means of a sol-gel technique to be waterinsoluble and easily separable microporous solid-acid catalysts. The catalysts thus obtained have large surface areas (400-800 m2 g-’), strong acidity and are thermally more stable than Amberlyst- 15. They effectively catalyze the hydrolysis of ethyl acetate in aqueous phase, showing higher activities than Amberlyst-15 and H-ZSM-5. Remarkably, the immobilization of PW into the silica matrix effectively suppresses the HPA leaching to as low as 0.3% during the hydrolysis reaction (6OoC, 3 h).’* Silica-supported PW has been found to be an active and recyclable catalyst for the Diels-Alder reaction (Eq. (9)) in toluene medium, providing a 70-80% yield.29The Ce3’ salt, Ce0.87H0.4PW,showed a fairly high activity too, while C S ~ , ~ H ~ was . ~ P practically W inactive. Bulk PW exhibited a very low activity probably due to the small surface area and site blocking by diene polymers. 0

0

2.2.4 Shape-selective catalysis. Supported HPAs are important for applications because bulk HPAs are nonporous materials with low surface area. As supports, especially interesting, are materials with regular pore structure, e.g. zeolites, providing a shape selectivity. However, conventional zeolites are not suitable for this because their pores are too small to adsorb large HPA (12 A) molecules. Recently HPA (PW) supported on a novel mesoporous pure-silica molecular sieve MCM-4 1, having uniform pores 32 A in size, was prepared and chara~terized.’~~’’ The PW/MCM-41 compositions with PW loadings from 10 to 50 wt% have -30 8, uniformly sized mesopores. As shown by transmission electron microscopy, the PW species are mainly located inside the MCM-41 pores rather than on the outer surface.” PW/MCM-41 exhibits a higher catalytic activity than H,SO, or bulk PW and shows a shape selectivity in liquid-phase phenol alkylation.” This catalyst may be promising for shape-selective reactions of large organic molecules, particularly for the synthesis of fine chemicals.

3 LIQUID-PHASE OXIDATION The liquid-phase oxidation of organic substances catalyzed by polyoxometalates (POM) proceeds in homogeneous or biphasic systems, oxygen, hydrogen peroxide,

228


alkylperoxides, etc., being applied as oxidants. In contrast to acid catalysis, where easily available classical Keggin-type HPAs dominate, in the liquid-phase oxidation, a wide variety of transition-metal-substituted heteropolyanions are used. Multicomponent POM redox catalysts are often considered as robust inorganic metalloporphyrin 9.1 1.32 analogues. In this part, we will discuss oxidations with the use of the most important and environmentally benign oxidants - dioxygen and hydrogen peroxide. Several reviews on catalysis by POM in the liquid-phase oxidation have been pub~ished.2~3~5~9~l 1.33 3.1 Oxidation with dioxygen This section mainly reviews recent applications of Keggin-type mixed-addenda heteropolyanions PMo,~-,,V,,O~~(~+")~ (HPA-n) as catalysts for aerobic liquid-phase oxidation. Here HPA-n can mean either anion or acid as well as a partially protonated anion. The HPA-n catalytic system, discovered by Matveev et is the most efficient and versatile one in the POM series for oxidation by 02.3935*36 3.1.1 General principles. In liquid-phase oxidation, HPA-n with n = 2-6 are used. These compounds are remarkable because they are the reversibly acting oxidants under mild condition^.^"'^^ Generally reactions catalyzed by HPA-n proceed via a stepwise redox mechanism represented by Eqs. (10) and (1 1): HPA-n + Red + m H' + H,(HPA-n) + Ox H,(HPA-n) + m/4 O2 + HPA-n + m/2 HzO This mechanism includes stoichiometric oxidation of the substrate (Red) by HPA-n followed by reoxidation of the reduced form of the oxidant, H,(HPA-n), with dioxygen, where H,(HPA-n) = H , ( P M O ~ ~ - , , ~ + V , , - ~ + Vm~

Flavours and Fragrances (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Maillard Reactions in Chemistry, Food and Health (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Yeasts in Food: Beneficial and Detrimental Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food and Beverage Stability and Shelf Life (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Processed Meats: Improving Safety, Nutrition and Quality (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Colloids: Proteins, Lipids and Polysaccharides (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Emulsions and Foams (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food, diet and obesity (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional Foods: Principles and Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional and Speciality Beverage Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Endocrine Disrupting Chemicals in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Determination of Veterinary Residues in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Tracing Pathogens in The Food Chain (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Cereals Processing Technology (Woodhead Publishing in Food Science and Technology)

Environmentally-Compatible Food Packaging (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Handbook of Food Proteins (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Meat Products Handbook: Practical Science and Technology (Woodhead Publishing in Food Science, Technology and Nutrition)

The Maillard Reaction in Foods and Medicine (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Developing Children's Food Products (Woodhead Publishing Series in Food Science, Technology and Nutrition - Volume 204)

Encapsulation and Controlled Release (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dietary Fibre: Chemical and Biological Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dairy-Derived Ingredients: Food and Nutraceutical Uses (Woodhead Publishing in Food Science, Technology and Nutrition)

Reducing Saturated Fats in Foods (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Phytochemical Functional Foods (Woodhead Publishing in Food Science and Technology)

Frying - Improving Quality (Woodhead Publishing in Food Science and Technology)

Baking Problems Solved (Woodhead Publishing in Food Science and Technology)

Benders' dictionary of nutrition and food technology, Eighth Edition (Woodhead Publishing in Food Science, Technology and Nutrition)

Functional Foods: Concept to Product (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Flavours and Fragrances (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Maillard Reactions in Chemistry, Food and Health (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Yeasts in Food: Beneficial and Detrimental Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food and Beverage Stability and Shelf Life (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Processed Meats: Improving Safety, Nutrition and Quality (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Colloids: Proteins, Lipids and Polysaccharides (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Emulsions and Foams (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food, diet and obesity (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional Foods: Principles and Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Functional and Speciality Beverage Technology (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Endocrine Disrupting Chemicals in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Determination of Veterinary Residues in Food (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Tracing Pathogens in The Food Chain (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Food Preservation Techniques (Woodhead Publishing in Food Science and Technology)

Cereals Processing Technology (Woodhead Publishing in Food Science and Technology)

Environmentally-Compatible Food Packaging (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Handbook of Food Proteins (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Meat Products Handbook: Practical Science and Technology (Woodhead Publishing in Food Science, Technology and Nutrition)

The Maillard Reaction in Foods and Medicine (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Developing Children's Food Products (Woodhead Publishing Series in Food Science, Technology and Nutrition - Volume 204)

Encapsulation and Controlled Release (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dietary Fibre: Chemical and Biological Aspects (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Dairy-Derived Ingredients: Food and Nutraceutical Uses (Woodhead Publishing in Food Science, Technology and Nutrition)

Reducing Saturated Fats in Foods (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Phytochemical Functional Foods (Woodhead Publishing in Food Science and Technology)

Frying - Improving Quality (Woodhead Publishing in Food Science and Technology)

Baking Problems Solved (Woodhead Publishing in Food Science and Technology)

Benders' dictionary of nutrition and food technology, Eighth Edition (Woodhead Publishing in Food Science, Technology and Nutrition)

Functional Foods: Concept to Product (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Recommend Documents