Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life

Ideas in Chemistry and Molecular Sciences Where Chemistry Meets Life Edited by Bruno Pignataro Ideas in Chemistry and...

Author: Bruno Pignataro

94 downloads 1153 Views 10MB 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

Ideas in Chemistry and Molecular Sciences Where Chemistry Meets Life

Edited by Bruno Pignataro

Ideas in Chemistry and Molecular Sciences Edited by Bruno Pignataro

Related Titles Pignataro, Bruno (ed.)

Waldmann, H., Janning, P. (eds.)

Ideas in Chemistry and Molecular Sciences

Chemical Biology

Advances in Synthetic Chemistry 2010

Learning through Case Studies 2009 ISBN: 978-3-527-32330-2

ISBN: 978-3-527-32539-9

Jacob, Claus/Winyard, Paul G. (eds.) Pignataro, Bruno (ed.)

Ideas in Chemistry and Molecular Sciences Advances in Nanotechnology, Materials and Devices

Redox Signaling and Regulation in Biology and Medicine 2009 ISBN: 978-3-527-31925-1

2010 ISBN: 978-3-527-32543-6

Carta, Giorgio/Jungbauer, Alois

Gjerde, Douglas T./Hoang, Lee/ Hornby, David

RNA Purification and Analysis

Protein Chromatography

Sample Preparation, Extraction, Chromatography

Process Development and Scale-Up

2009

2010

ISBN: 978-3-527-32116-2

ISBN: 978-3-527-31819-3

Pignataro, Bruno (ed.)

Hartmann, Roland K./Bindereif, Albrecht/ ¨ Astrid/Westhof, Eric (eds.) Schon,

Tomorrow’s Chemistry Today

Handbook of RNA Biochemistry

Concepts in Nanoscience, Organic Materials and Environmental Chemistry 2nd edition

Student Edition

2009 ISBN: 978-3-527-32623-5

2008 ISBN: 978-3-527-32534-4

Ideas in Chemistry and Molecular Sciences Where Chemistry Meets Life

Edited by Bruno Pignataro

The Editor Prof. Bruno Pignataro University of Palermo Department of Physical Chemistry Viale delle Scienze 90128 Palermo Italy

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Cover

Library of Congress Card No.: applied for

We would like to thank Dr. Adriana Pietropaolo (ETH Zürich) for providing us with the graphic material used in the cover illustration.

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at .  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Cover Design Adam Design, Weinheim Typesetting Laserwords Private Limited, Chennai, India Printing and Binding betz-druck GmbH, Darmstadt Printed in the Federal Republic of Germany Printed on acid-free paper ISBN: 978-3-527-32541-2 Set ISBN: 978-3-527-32875-8

V

Contents

Preface XIII List of Contributors Part I 1

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.7.1 1.7.2 1.7.3 1.7.4 1.7.5 1.7.6 1.8 1.8.1 1.8.2 1.9

2 2.1 2.2

XIX

Biochemical Studies

1

The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders 3 Fabio Arnesano Introduction 3 Metal Ions in the Brain 4 Brain Copper Homeostasis 5 Brain Copper and Neurodegenerative Disorders 8 The Role of Ubiquitin in Protein Degradation 9 Failure of the Ubiquitin System in Neurodegenerative Disorders 13 Interaction of Ubiquitin with Metal Ions 15 Thermal Stability of Ubiquitin 15 Spectroscopic Characterization of CuII Binding 15 Possible Implications for the Polyubiquitination Process 17 CuII -Induced Self-Oligomerization of Ub 17 Cooperativity between CuII -Binding and Solvent Polarity 18 Comparison with Other Metal Ions 19 Biological Implications 21 The Redox State of Cellular Copper 21 Ubiquitin and Phospholipids 22 Conclusions and Perspectives 23 Acknowledgments 24 References 24 The Bioinorganic and Organometallic Chemistry of Copper(III) 31 Xavi Ribas and Alicia Casitas Introduction 31 Bioinorganic Implications of Copper(III) 33

Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

VI

Contents

2.2.1 2.2.2 2.2.3 2.2.4 2.3 2.3.1 2.3.1.1 2.3.1.2 2.3.1.3 2.3.2 2.3.3 2.3.3.1 2.3.3.2 2.4 2.5

3 3.1 3.2 3.3 3.4 3.4.1 3.4.1.1 3.4.1.2 3.4.1.3 3.4.1.4 3.4.2 3.4.3 3.4.3.1 3.4.3.2 3.4.3.3 3.5 3.5.1 3.5.2 3.5.2.1 3.5.2.2 3.5.2.3 3.5.2.4 3.5.3 3.5.3.1

Dinuclear Type-3 Copper Enzymes 33 Particulate Methano Monooxygenase (pMMO) 36 Mononuclear Monooxygenating Copper-based Enzymes 38 Trinuclear Copper Models for Laccase 40 Organometallic CuIII Species in Organic Transformations 41 C–C Bond Formation in Organocuprate(I) Catalysis 42 Conjugate Addition to α-Enones 42 Acetylene Carbocupration 43 SN 2 and SN 2 Alkylations 43 Aryl–Heteroatom Bond Formation in Cu-mediated Cross-coupling Processes 44 Aromatic and Aliphatic C–H Bond Organometallic Functionalizations 45 Catalytic Systems 45 Stoichiometric Systems 47 Miscellany: Cuprate Superconducting Materials 51 Overview and Future Targets 51 References 52 Chemical Protein Modification 59 Gon¸calo J. L. Bernardes, Justin M. Chalker, and Benjamin G. Davis Introducing Diversity by Posttranslational Modification 59 Chemistry: A Route to Modified Proteins 60 Challenges in Chemical Protein Modification 61 Traditional Methods for Protein Modification 61 Lysine Modification 62 Activated Esters 62 Isocyanates and Isothiocyanates 62 Reductive Alkylation 62 IME Reagents 63 Glutamic and Aspartic Acid Modification 64 Cysteine 64 Alkylation 65 Disulfides 66 Desulfurization at Cysteine 67 Recent Innovations in Site-Selective Protein Modification 70 Dehydroalanine: A Useful Chemical Handle for Protein Conjugation 71 Metal-Mediated Protein Modification 71 Modification at Natural Residues 72 Iridium-Catalyzed Reductive Alkylation of Lysine 74 Modification of Unnatural Residues 74 Olefin Metathesis at S-Allyl Cysteine 77 Metal-Free Methods for Modifying Unnatural Amino Acids 78 Oxime Ligation at Aldehydes and Ketones 78

Contents

3.5.3.2 3.5.3.3 3.5.3.4 3.5.4 3.6

Azide and Alkyne Modification 79 Selective Modification of Tetrazole-Containing Proteins 80 Tetrazine Ligation 81 Dual Modification 81 Conclusion and Outlook 81 References 82 Part II

4

4.1 4.2 4.3 4.3.1 4.3.2 4.4 4.4.1 4.4.2 4.4.3 4.4.4 4.4.4.1 4.4.4.2 4.5

5

5.1 5.1.1 5.1.2 5.1.2.1 5.1.2.2 5.1.2.3 5.1.2.4 5.2 5.2.1 5.2.1.1 5.2.2 5.2.2.1 5.2.2.2 5.2.3

Drug Delivery 93

Vitamin B12 : A Potential Targeting Molecule for Therapeutic Drug Delivery 95 Pilar Ruiz-Sánchez Introduction 95 Transport Mechanism 96 Metabolism of B12 97 Adenosylcobalamin-Dependent Reactions 97 Methylcobalamin-Dependent Reactions 99 Vitamin B12 Derivatives 99 Structure 99 b-, d-, e-Cobalamin Derivatives 99 Modifications on the Ribose Moiety 101 β-Axial Position 102 Cobalamin Alkylation 102 Heterodinuclear Concept 103 Outlook 108 Acknowledgments 108 References 109 Strategies for Microsphere-Mediated Cellular Delivery 117 Rosario M. Sanchez-Martin, Lois M. Alexander, Juan Manuel Cardenas-Maestre, and Mark Bradley Introduction 117 Cellular Delivery 117 Delivery Devices 117 Liposomes 117 Cell-penetrating Peptides 118 Dendrimers 118 Nanomaterials 118 Microspheres 119 Biodegradable Microspheres 119 Preparation 119 Biostable Microspheres 120 Applications of Biostable Microspheres 120 Preparation 120 Microspheres and Solid-phase Chemistry 121

VII

VIII

Contents

5.2.3.1 5.2.3.2 5.2.3.3 5.2.3.4 5.2.4 5.2.4.1 5.2.4.2 5.2.5 5.2.5.1 5.2.5.2 5.2.6 5.2.6.1 5.3

Preparation of Microspheres 121 Fmoc Chemistry on Microspheres 121 Dual Functionality of Microspheres 122 Coupling Agents 124 Noncleavable Link to Microspheres 126 Microsphere-based Intracellular Sensing 126 siRNA Delivery 127 Cleavable Linkers 130 Ester Bonds 130 Disulfide Bonds 132 Bioconjugation 133 Streptavidin–Biotin 134 Future Perspectives 135 Acknowledgments 135 References 135 Part III

6

6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.2.5 6.2.6 6.3

7 7.1 7.2 7.2.1 7.2.2

Research in Therapeutics 141

Fundamental Processes in Radiation Damage to DNA: How Low-Energy Electrons Damage Biomolecules 143 Ilko Bald Radiation Damage and the Role of Low-Energy Electrons 143 How Chemical Bonds are Broken by Low-energy Electrons 145 DEA Studies of Gas-Phase DNA Building Blocks: The Nucleobases 147 DEA Studies on Model Compounds for the DNA Backbone 148 Electron Attachment to d-Ribose 148 Cross-Ring Cleavage of d-Ribose Proceeds with Selective Charge Retention 150 The Nature of the Transient Negative d-Ribose Anions 154 One Step Further: Tetraacetyl-d-Ribose 155 The Use of Laser-Induced Acoustic Desorption (LIAD) to Study DEA to Larger Biomolecules 159 Sugar–Phosphate Cleavage Induced by 0 eV Electrons: DEA to d-Ribose-5 -Phosphate 159 Outlook and Future Prospects 161 Acknowledgments 162 References 163 Structure-Based Design on the Way to New Anti-infectives 167 Anna Katharina Herta Hirsch Introduction 167 Isoprenoids and the Nonmevalonate Pathway 169 4-Diphosphocytidyl-2C-methyl-d-erythritol Kinase (IspE) 170 Structure of IspE 170

Contents

7.2.3 7.3 7.3.1 7.3.1.1 7.3.1.2 7.3.2 7.3.3 7.3.4 7.3.4.1 7.3.4.2 7.3.5 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.5 7.5.1 7.5.2

8

8.1 8.1.1 8.1.2 8.1.3 8.2 8.2.1 8.2.2 8.2.2.1 8.2.2.2 8.2.2.3 8.2.2.4 8.3 8.3.1 8.3.2 8.4 8.4.1 8.4.2

Active Site of IspE 170 Targeting the CDP-Binding Pocket of IspE 174 Design 174 Possible Ribose Analogues 175 Design of the Vector 176 Optimization of the Ribose Analogue 176 Importance of the Vector 178 Optimization of the Filling of the Small, Hydrophobic Pocket 179 The ‘‘55% Rule’’ 179 Evaluation of Inhibitors Featuring Different Sulfone Substituents 180 Summary of the First-Generation Inhibitors 182 X-ray Cocrystal Structure Analysis 182 Design of Water-Soluble Inhibitors 182 Enzyme Assays of Inhibitors Designed to be Water Soluble 183 Structural Analysis 184 Lessons Learnt from the Cocrystal Structure 185 Conclusions and Outlook 185 Conclusions 185 Outlook 186 Acknowledgments 186 List of Abbreviations 186 References 187 Drug–Membrane Interactions: Molecular Mechanisms Underlying Therapeutic and Toxic Effects of Drugs 191 Marlene Lucio, ´ José L. F. C. Lima, and Salette Reis Biological Membranes 191 Role of Membranes in Life Maintenance 191 Structure and Composition of Membranes 191 Dynamic Molecular Organization of Membranes 193 Drug–Membrane Interactions 195 Possible Effects of Drugs on Membranes 195 Clinical Relevance of the Drug–Membrane Interaction Studies 195 Contribution for Drug Development 195 Understanding Therapeutic and Toxic Effect of Drugs 197 Understanding Mechanisms of Multidrug Resistance 198 Controlling Enzymatic Inhibition 198 Analysis and Quantification of Drug–Membrane Interactions 199 Membrane Model Systems 199 Experimental Techniques 200 Drug–Membrane Interactions Applied to the Study of Nonsteroidal Anti-inflammatory Drugs (NSAIDs) 201 Drug Fundamental Physical–Chemical Studies 202 Membrane Structural and Dynamic Studies 203

IX

X

Contents

8.4.3 8.5

Results 205 Conclusions and Future Research Directions Acknowledgments 206 References 207

9

Targeting Disease with Small Molecule Inhibitors of Protein–Protein Interactions 215 Fedor Forafonov, Elena Miranda, Ida Karin Nordgren, and Ali Tavassoli Introduction 215 High-Throughput Screening of Chemical Libraries 216 High-Throughput Screening of Biosynthesized Libraries 220 Cyclic Peptide Inhibitors of AICAR Transformylase Activity 222 Cyclic Peptide Inhibitors of HIV Budding 224 Future Direction 226 Small Molecule Inhibitors of Tumor Hypoxia Response Network 226 Targeting Protein–Protein Interactions in Asthma 228 Targeting the Protein Interaction Networks of Influenza Virus 230 Acknowledgments 232 References 232

9.1 9.2 9.3 9.3.1 9.3.2 9.4 9.4.1 9.4.2 9.4.3

10 10.1 10.1.1 10.1.2 10.1.3 10.1.3.1 10.1.3.2 10.1.3.3 10.1.4 10.1.4.1 10.1.4.2 10.1.4.3 10.1.4.4 10.2 10.2.1 10.2.2 10.2.3 10.3

206

Cracking the Glycocode: Recent Developments in Glycomics 239 Lars Hillringhaus and Jürgen Seibel State of the Art 239 Introduction 239 Carbohydrate-Based Drugs 239 Carbohydrate Synthesis 240 Chemical Synthesis 241 Enzymatic Synthesis 243 Glycoprotein Synthesis 244 Glycomics 245 Mass Spectrometry 245 Microarrays 246 Cell, Tissue, and Metabolic Labeling 246 Bioinformatics 247 Some New Insights in Glycomics 247 Microwave-Assisted Glycosylation for the Synthesis of Glycopeptides 247 Highly Efficient Chemoenzymatic Synthesis of Novel Branched Thiooligosaccharides by Substrate Direction with Glucansucrases 251 Identification of New Acceptor Specificities of Glycosyltransferase R with the Aid of Substrate Microarrays 257 Future Perspectives 259 Acknowledgments 259 References 260

Contents

Part IV 11

11.1 11.2 11.2.1 11.2.2 11.3 11.3.1 11.3.2 11.3.3 11.3.3.1 11.3.3.2 11.4

O2 Reactivity at Model Copper Systems: Mimicking Tyrosinase Activity 267 Anna Company General Introduction: O2 Activation and Model Systems 267 Copper Proteins Involved in O2 Activation 268 Hemocyanin 269 Tyrosinase 270 O2 Binding and Activation at Biomimetic Cu Complexes 272 Copper–Dioxygen Adducts 272 Ligand Architecture: Influence on Reactivity toward O2 275 Hydroxylation of Aromatic Rings: Mimicking Tyrosinase Activity 278 Intramolecular Aromatic Hydroxylation 278 Intermolecular ortho-Hydroxylation of Phenolic Compounds 280 Concluding Remarks 285 References 286 Part V

12

12.1 12.1.1 12.1.2 12.1.3 12.1.4 12.2 12.3 12.3.1 12.3.2 12.3.3 12.3.4 12.4

13

13.1 13.2 13.3

Enzyme Chemistry 265

Structure–Property Relationship and Biosensing 291

Chirality in Biochemistry: A Computational Approach for Investigating Biomolecule Conformations 293 Adriana Pietropaolo Introduction 293 Molecular Chirality in Living Systems 293 Protein Secondary Structures 295 Protein Secondary Structure Assignment 296 Intrinsic Chirality and Protein Secondary Structures 297 Computational Techniques for Studying Protein Dynamics 297 Employing Chirality to Analyze Protein Motions 298 The Chirality Index 298 Using Chirality to Understand Protein Structure 300 Chirality Index as a Tool for Monitoring Protein Dynamics 302 Chirality and Circular Dichroism 305 Perspectives 308 Acknowledgments 309 References 310 Collisional Mechanism–Based E-DNA Sensors: A General Platform for Label-Free Electrochemical Detection of Hybridization and DNA Binding Proteins 313 Francesco Ricci Introduction 313 E-DNA Signaling Mechanism 315 E-DNA Sensor for DNA Binding Proteins Detection 319

XI

XII

Contents

13.4 13.5

Conclusions and Future Perspectives 321 Acknowledgments 324 References 324 Index 327

XIII

Preface The idea of publishing books based on contributions given by emerging young chemists arose during the preparation of the first EuCheMs (European Association for Chemical and Molecular Sciences) Conference in Budapest. In this conference I cochaired the competition for the first European Young Chemist Award aimed at showcasing and recognizing the excellent research being carried out by young scientists working in the field of chemical sciences. I then proposed to collect in a book the best contributions from researchers competing for the Award. This was further encouraged by EuCheMs, SCI (Italian Chemical Society), RSC (Royal Society of Chemistry), GDCh (Gesellschaft Deutscher Chemiker), and Wiley-VCH and brought out in the book ‘‘Tomorrow’s Chemistry Today’’ edited by myself and published by Wiley-VCH. The motivation gained by the organization from the above initiatives was, to me, the trampoline for co-organizing the second edition of the award during the second EuCheMs Conference in Torino. Under the patronage of EuCheMs, SCI, RSC, GDCh, the Consiglio Nazionale dei Chimici (CNC), and the European Young Chemists Network (EYCN), the European Young Chemist Award 2008 was again funded by the Italian Chemical Society. In Torino, once again, I personally learned a lot and received important inputs from the participants about how this event can serve as a source of new ideas and innovations for the research work of many scientists. This is also related to the fact that the areas of interest for the applicants cover many of the frontier issues of chemistry and molecular sciences (see also Chem. Eur. J. 2008, 14, 11252–11256). But, more importantly, I was left with the increasing feeling that our future needs of new concepts and new technologies should be largely in the hands of the new scientific generation of chemists. In Torino, we received about 90 applications from scientists (22 to 35 years old) from 30 different countries all around the world (Chem. Eur. J. 2008, 14, 11252–11256). Most of the applicants were from Spain, Italy, and Germany (about 15 from each of these countries). United Kingdom, Japan, Australia, United States, Brazil, Morocco, Vietnam, as well as Macedonia, Rumania, Slovenia, Russia, Ukraine, and most of the other European countries were also represented. In terms of applicants, 63% were male and about 35% were PhD students; the number of Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

XIV

Preface

postdoctoral researchers was only a small percentage, and only a couple of them came from industry. Among the oldest participants, mainly born between 1974 and 1975, several were associate professors or researchers at Universities or Research Institutes and others are lecturers, assistant professors, or research assistants. The scientific standing of the applicants was undoubtedly very high and many of them made important contributions to the various symposia of the 2nd EuCheMs Congress. A few figures help to substantiate this point. The, let me say, ‘‘h index’’ of the competitors was 20, in the sense that more than 20 applicants coauthored more than 20 publications. Some patents were also presented. Five participants had more than 35 publications, and, h indexes, average number of citations per publication, and number of citations, were as high as 16, 35.6, and 549, respectively. Several of the papers achieved further recognition as they were quoted in the reference lists of the young chemists who where featured on the covers of top journals. The publication lists of most applicants proudly noted the appearance of their work in the leading general chemistry journals such as Science, Nature, Angewandte Chemie, Journal of the American Chemical Society, or the best niche journals of organic, inorganic, organometallic, physical, analytical, environmental, and medicinal chemistry. All of this supported the idea of publishing a second book with the contributions of these talented chemists. However, in order to have more homogeneous publications and in connection to the great number of interesting papers presented during the competition, we decided to publish three volumes. This volume represents indeed one of the three edited by inviting a selection of young researchers who participated in the European Young Chemist Award 2008. The other two volumes concern the two different areas of synthetic chemistry and nanotechnology/materials-science and are, respectively, entitled ‘‘Ideas in Chemistry and Molecular Science: Advances in Synthetic Chemistry’’ and ‘‘Ideas in Chemistry and Molecular Sciences: Advances in Nanotechnology, Materials, and Devices’’. It is important to mention that the contents of the books are a result of the work carried out in several topmost laboratories around the world both by researchers who already lead their own group and by researchers who worked under a supervisor. I would like to take this occasion to acknowledge all the supervisors of the invited young researchers for their implicit or explicit support to this initiative that I hope could also serve to highlight the important results of their research groups. The prospect of excellence of the authors was evident from the very effusive recommendation letters sent by top scientists supporting the applicants for the Award. A flavor of these letters is given by the extracts from some of the sentences below: ‘‘The candidate creativity in polymer design for therapeutic impact is extremely impressive. I believe that he is an excellent example of an outstanding young European chemist’’; ‘‘Is an energetic, enthusiastic, articulate, bright researcher’’; ‘‘Talented and enthusiastic scientist who has a lot of determination and persistence. I was impressed by the strong preparation, enthusiasm and motivation. First rate

Preface

researcher’’; ‘‘. . .outstanding, very broad and far reaching research achievements’’; ‘‘The personality and the scientific research achievements will certainly pave to the candidate the way as a future leader of an independent and creative research group. Due to his original and uncommon approach together with the high quality of results, he will become well known in the field of medicinal chemistry. All this exciting work relied on his ability as a chemist and on his bright intellect, which allows him to move across disciplines seamlessly’’; ‘‘The candidate is a very articulate, knowledgeable, driven, clear minded, and extremely personable young scientist’’; ‘‘Application in this work has been outstanding, strong, powerful and very dedicated. The candidate will be in a very good position to make tremendous impact in the area of therapeutics and to become a leader in the area of organic chemistry in the next generation’’; ‘‘As one of the most talented graduate students to have emerged for our group he has the vision, skill and drive to make a real difference to the national research community’’; ‘‘. . .outstanding young scientist and a charming person’’; ‘‘. . .has an astonishing level of productivity’’; ‘‘. . .work of very high caliber’’; ‘‘Such deep and insightful investigations are refreshing and invigorating in my view especially in times that tend to favor simple, phenomenological results’’; ‘‘The candidate has been the intellectual driver and forged the necessary collaborations to address this complex problem. He deft integration of the various experimental and theoretic disciplines needed for the work is impressive’’; ‘‘The candidate performed outstanding research projects’’; ‘‘He should be a compelling and deserving candidate of the European Young Chemist award’’; ‘‘The candidate is for sure one of the most outstanding and promising young researchers that I know in the field of molecular science. Despite being young, he has a lot of experience and high motivation. He is excellent in the lab and extremely smart person’’; ‘‘The candidate has performed work of top quality and originality and of great breadth’’; ‘‘. . .a special individual who ranks among the very best’’; ‘‘Exceptional in many respect. The candidate commands superior experimental and intellectual skills and has shown great chemical structure intuition in the project’’. ‘‘. . .outstanding scientific debater. The personality is impeccable’’. ‘‘Extremely gifted young researcher who has already developed an outstanding track record of achievement while working in one of the world’s premier laboratories in the USA’’; ‘‘Has the skills, expertise and leadership potential to discover new compounds that will revolutionize the therapy of some difficult diseases. The work is well-aligned to the University’s strategy of supporting cross disciplinary research of the highest quality’’; ‘‘. . .unusually broad dimension of experimental skills’’; ‘‘This is a rare combination of expertise’’; ‘‘enables to undertake a diverse scope of key problems. In my opinion these are the type of individuals who will drive the field forward’’; ‘‘. . .has an engaging, outgoing, attractive personality to which others instinctively respond’’; ‘‘Brilliant young chemist with a great potential’’. The chapters written by the various contributors cover several areas of life science and range from more or less typical biochemical studies, to research relevant in the therapeutic field, to the enzyme or drug activity, to drug delivery, to computational structure–activity relation or sensors.

XV

XVI

Preface

In the area of biochemical studies, the book collects a contribution in the field of inorganic chemistry of the brain with the chapter on the role of metal ions (Cu(II) in particular) and the ubiquitin–proteasome system on neurodegenerative disorder (Arnesano). In another chapter, the strategies for chemical modification of proteins as well as selective installation of biochemical probes and tethering of therapeutic cargo to proteins are highlighted (Bernardes et al.). Another chapter (Ribas et al.) is centered on Cu(III) ion and its bioinorganic and organometallic chemistry. In particular, this contribution turns out to be relevant for understanding copper-dependent metalloenzymes. In the drug delivery domain, a chapter is dedicated to vitamin B12 as potential targeting molecule for therapeutic drug delivery (Ruiz Sánchez). This potentiality is vividly illustrated by the author with the following definition: vitamin B12 is an attractive ‘‘Trojan horse’’ for therapeutic drug delivery. A second contribution starts from the consideration that the cellular membrane poses a formidable barrier to the intracellular flux of materials circulating extracellularly due to its selective permeability based predominantly on ionic charge, hydrophobicity, and size. As such, many drugs, nucleic acids, proteins, and other investigative constructs are not able to translocate the cellular membrane alone and often require a delivery vehicle for function and/or activity. After considering various vehicles this contribution focuses on a highly competitive cellular delivery technology based on polymeric microspheres-mediated cellular delivery (Sanchez-Martin). In particular, the author presents different strategies that can be applied for the attachment of biomolecules to the microspheres and their applications as a delivery system. Monodispersed populations of robust cross-linked microspheres of defined sizes (from 200 nm to 2 µm) have been synthesized and standard solid phase multistep protocols have been applied to them. Several contributions are primarily dedicated to research that has relevance to the therapeutic field. The chapter (Lucio et al.) on molecular mechanisms, underlying therapeutic and toxic effects of drugs, highlights the importance of membrane composition and the dynamic molecular organization of membranes in drug–membrane interaction. Also, the work aims to exemplify an application of drug–membrane interaction studies in the evaluation of the nonsteroidal anti-inflammatory drug effect on membranes. These studies may prove valuable in the design of novel drug formulations with increasing efficacy and reduced side effects. A second contribution in the area deals with targeting diseases with small molecule inhibitors of protein–protein interactions (Tavassoli et al.). The chapter starts from the consideration that there are several thousand protein–protein interactions that control the majority of biological processes. There is therefore great potential for small molecule therapeutics that affect disease through modulation of protein–protein interactions. In this chapter, there is then a discussion on recent approaches taken toward controlling protein interactions with small molecules, from logical design using structural data to high-throughput screening of chemical and biological libraries. The authors go on to outline their own efforts in this

Preface

field, and end with a detailed discussion of some of the important protein–protein interactions currently being targeted. Another chapter (Hirsh) is dedicated to the design and synthesis of inhibitors of the kinase IspE as potential antimalarians opening a number of research avenues summarized in the chapter itself, while another contribution (Hillringhaus) deals with recent developments in glicomics – a technology that can be anticipated that it will strongly influence tomorrow’s therapy in the areas of various diseases. A last contribution (Bald) in this area deals with the important studies on the electron-induced modification of biomolecules, which may have potential implications for the design of new chemotherapeutic and radiosensitizing drugs and for the development of more efficient protocols in cancer therapy. Understanding how nature works in the area of enzyme chemistry is the goal of a chapter (Company) dedicated to the tyrosinase reactivity. In this chapter, in particular, it is shown how the model chemistry approach has allowed to reproduce and to better understand the mechanisms by which O2 is activated in the dinuclear copper protein just called tyrosinase. The studies summarized in another chapter (Ricci) refer to electrochemical biosensors, termed E-DNA sensors, that are based on the target binding–induced folding of electrode-bound DNA probes and gives contribution on E-DNA signaling mechanisms. In addition, studies on E-DNA sensors for DNA binding protein detection are also reported. The E-DNA sensing platform has then demonstrated to be not only a promising and appealing approach for the sequence-specific detection of DNA and RNA but also to be flexible enough to be adaptable for the detection of DNA–protein interaction. Taking into account that chirality selection is a key issue in many important biochemical phenomena such as protein folding and enzyme recognition, another chapter discusses the role of chirality in the chemistry of the life sciences. In particular, a novel methodology for the study of local chirality is presented, which provides one with a deeper understanding of the connection between secondary structure and protein flexibility. This computational study for investigating biomolecule conformation reported in the chapter (Pietropaolo) suggests that chirality is a central organizing principle in life that confers order at every length scale up to the full cell. Probably, everyone who reads this book will have their own opinion on what is relevant for the future of life chemistry, and in this respect I would like to notify that, due to its peculiar genesis, the book reflects the opinions of a select group of young chemists and therefore does not pretend to cover the whole area. The main aim is just to offer a variety of individual views that will provoke thought possibly giving an attractive insight into the minds and research areas for the next generation of chemical and molecular sciences, especially those related with life sciences and technologies. Starting from this point, I hope that the many ideas that can be grasped sailing through the various contributions by the young authors of the book should be very useful for helping to make several step forward in the field of chemistry and molecular sciences to solve problems related to our health or more generally to our

XVII

XVIII

Preface

life as well as to discover new tools in medicinal chemistry needed for the present and the future society. I cannot end this preface without acknowledging all the authors and the persons who helped me in the book project together with all the societies (see the book cover) that motivated and sponsored the book. I’m personally grateful to Professors Giovanni Natile, Francesco De Angelis and Luigi Campanella for their motivation and support in this activity. Palermo, October 2009

Bruno Pignataro

XIX

List of Contributors Lois M. Alexander University of Edinburgh Chemical Biology Section School of Chemistry Joseph Black Building West Mains Road Edinburgh EH9 3JJ Scotland Fabio Arnesano University of Bari ‘‘A. Moro’’ Department of Farmaco-Chimico Via E. Orabona 4 70125 Bari Italy Ilko Bald Freie Universität Berlin Institut für Chemie und Biochemie–Physikalische und Theoretische Chemie Takustraße 3 D-14195 Berlin Germany

Gonçalo J. L. Bernardes University of Oxford Department of Chemistry Chemistry Research Laboratory 12 Mansfield Road Oxford OX1 3TA United Kingdom Mark Bradley University of Edinburgh Chemical Biology Section School of Chemistry Joseph Black Building West Mains Road Edinburgh EH9 3JJ Scotland Alicia Casitas Universitat de Girona Department de Qu´ımica Grup de Qu´ımica Inorgànica i Supramolecular (QBIS) Campus Montilivi 17071 Girona – Catalonia Spain

and Interdisciplinary Nanoscience Center (iNANO) Aarhus University Ny Munkegade 8000 Aarhus C Denmark

Justin M. Chalker University of Oxford Department of Chemistry Chemistry Research Laboratory 12 Mansfield Road Oxford OX1 3TA United Kingdom


XX

List of Contributors

Anna Company Technische Universität Berlin Institut für Chemie. Sekretariat C2 Straße des 17.Juni 135 D-10623 Berlin (Germany) Benjamin G. Davis University of Oxford Department of Chemistry Chemistry Research Laboratory 12 Mansfield Road Oxford OX1 3TA United Kingdom Fedor Forafonov University of Southampton School of Chemistry University Road, Highfield Southampton SO17 1BJ United Kingdom Lars Hillringhaus University of Oxford Department of Chemistry Chemistry Research Laboratory 12 Mansfield Road Oxford OX1 3TA United Kingdom Ida Karin Nordgren University of Southampton School of Chemistry University Road, Highfield Southampton SO17 1BJ United Kingdom Anna Katharina Herta Hirsch Université de Strasbourg Institut de Science et d’Ingénierie Supramoléculaires 8 Allée Gaspard Monge Strasbourg 67000 France

José L. F. C. Lima Universidade do Porto Faculdade de Farmácia Requimte, Serviço de Qu´ımica-F´ısica Rua An´ıbal Cunha, 164 Porto 4099-030 Portugal Marlene Lúcio Universidade do Porto Faculdade de Farmácia Requimte, Serviço de Qu´ımica-F´ısica Rua An´ıbal Cunha, 164 Porto 4099-030 Portugal Juan Manuel Cardenas-Maestre University of Edinburgh Chemical Biology Section School of Chemistry Joseph Black Building West Mains Road Edinburgh EH9 3JJ Scotland Elena Miranda University of Southampton School of Chemistry University Road, Highfield Southampton SO17 1BJ United Kingdom

List of Contributors

Adriana Pietropaolo Universitá di Catania Dipartimento di Scienze Chimiche Viale A. Doria 6 Catania I-95125 Italy

Francesco Ricci University of Rome Tor Vergata Dipartimento di Scienze e Tecnologie Chimiche Via della Ricerca Scientifica 00133 Rome Italy

and

Pilar Ruiz-Sánchez Universität Zürich Anorganisch-chemisches Institut Winterthurerstr. 190 8057 Zürich Switzerland

´ ETH Zurich Department of Chemistry and Applied Biosciences Institute of Computational Science USI Campus Via Giuseppe Buffi 13 CH-6900 Lugano Switzerland Salette Reis Universidade do Porto Faculdade de Farmácia, Requimte, Serviço de Qu´ımica-F´ısica Rua An´ıbal Cunha, 164 Porto 4099-030 Portugal Xavi Ribas Universitat de Girona Department de Qu´ımica Grup de Qu´ımica Inorgànica i Supramolecular (QBIS) Campus Montilivi 17071 Girona – Catalonia Spain

Rosario M. Sanchez-Martin University of Edinburgh Chemical Biology Section School of Chemistry Joseph Black Building West Mains Road Edinburgh EH9 3JJ Scotland J¨ urgen Seibel University of Würzburg Institute of Organic Chemistry Am Hubland D-97074 Würzburg Germany Ali Tavassoli University of Southampton School of Chemistry University Road, Highfield Southampton SO17 1BJ United Kingdom

XXI

1

Part I Biochemical Studies


3

1 The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders Fabio Arnesano

1.1 Introduction

Ubiquitin (Ub) plays a crucial role in intracellular protein degradation via the proteasome and the autophagy–lysosome pathways [1]. Failure to eliminate misfolded proteins can lead to the formation of toxic aggregates and cell death [2]. Insoluble protein aggregates enriched with Ub are a hallmark of most neurodegenerative disorders including Parkinson’s, Alzheimer’s, amyotrophic lateral sclerosis and prion diseases [3, 4]. All of these disorders have been linked to metal accumulation and disturbance of redox and metal homeostasis in the brain [5–7], and metal ions have been implicated in the aggregation of disease-related, amyloidogenic proteins [8–12]. The potential role of metal ions in the aggregation of Ub has recently been examined [13, 14]. CuII is different from ZnII , NiII , AlIII , or CdII in that it binds to the N-terminal end of Ub, destabilizes the protein, and promotes its oligomerization into spherical particles. By mimicking the condition of low dielectric constant experienced near a membrane surface, the assembly of spherical oligomers of Ub yields a series of intermediate species leading to an extended nonfibrillar filament network. Aggregate disassembly is triggered by CuII chelation or reduction [14]. Intermediate annular and porelike structures, stabilized by the interaction of CuII -induced Ub oligomers with lipid bilayers, resemble toxic protofibrillar species produced by amyloidogenic proteins, which cause membrane permeabilization and disruption of metal homeostasis [15–19]. Susceptibility to aggregation of Ub represents a potential risk factor for disease onset or progression while cells attempt to tag and process toxic substrates. CuII binding and proximity to biological membranes appear to dramatically increase the aggregation propensity of Ub and other disease-related proteins, thus emphasizing the importance of preserving cellular compartmentalization and metal homeostasis for the correct functioning of protein degradation systems. Recent findings reinforce the vision of metal ions as key factors and promising therapeutic targets in protein conformational disorders [20, 21]. New strategies are being developed that will help to investigate their functional and pathogenic interactions in vivo. Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

4

1 The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders

1.2 Metal Ions in the Brain

The brain is a specialized organ that controls cognitive and motor functions. To carry out its functions the brain requires the highest concentrations of metal ions in the body and the highest per-weight consumption of body oxygen [22]. Metal ions in the brain fulfill catalytic and structural roles, which include the stabilization of biomolecules (e.g., MgII in nucleic acids, ZnII in Zn-finger transcription factors) or dynamic processes (e.g., NaI and KI in ion channels, CaII in neuronal cell signaling) [23]. The dynamic partitioning of these metal ions is controlled by ion-specific channels that selectively allow passage of ions in and out of cells. In the brain, the uneven distribution of NaI and KI ions across a cell membrane creates a potential that enables transmission of nervous pulses. CaII is also a key modulator of molecular information transfer within and between cells during neurotransmission; most eukaryotic cells either export or store CaII within membrane-enclosed vesicles to maintain cytosolic- free CaII levels at 100–200 nM, roughly 10 000-fold less than in the extracellular space [23]. More recently, considerable attention has been directed to the role of transition metal ions in the brain [22, 24]. Zn, Fe, Cu, and related d-block metals are emerging as significant players in both neurophysiology and neuropathology, particularly with regard to aging and neurodegenerative diseases [25]. Relatively high concentrations of these d-block metals are present within the different cellular compartments, the values ranging from 100 to 1000 µM [22]. The metal concentrations in brain tissue are up to 10 000-fold higher than those in common neurotransmitters and neuropeptides. Not only do these metals serve as components of various proteins and enzymes essential for normal brain function, but, in the labile form, are also involved in specialized brain activities; therefore, if misregulation of their homeostasis occurs, toxicity, mediated also by oxidative stress in the case of Cu and Fe [26], could ensue. Oxidative stress has been identified in many neurodegenerative diseases, and is commonly associated with increased levels of at least one of these transition metal ions in specific brain regions [27]. Transporters for Cu, Zn, Fe, and Mn play an important part in the intracellular distribution of these metals [28], such that defects in their regulation, which could possibly occur with aging, may create an environment that could result in protein misfolding and aggregation, thereby accelerating degenerative conditions [2, 29, 30]. Notably, brain homeostasis of metals is intertwined with changes in one metal leading to changes in the levels of other metals [24]. This is well established for Cu and Fe, where decreased Cu bioavailability may result in altered Fe levels, and for Fe and Mn, where Fe deficiency leads to a significant increase in brain Mn. The following discussion will be focused on basic aspects of brain Cu homeostasis. The widespread distribution and mobility of Cu required for normal brain function, along with the numerous correlations between Cu misregulation and a variety of neurodegenerative diseases, have prompted interest in studying its roles in neurophysiology and neuropathology [26, 31].

1.3 Brain Copper Homeostasis


Copper is the third-most abundant transition metal in the brain, after Fe and Zn, with average neuronal Cu concentrations of ∼0.1 mM. This redox-active nutrient is distributed unevenly within brain tissue, as Cu levels in the gray matter are twoto threefold higher than those in the white matter. Cu is particularly abundant in the locus coeruleus (1.3 mM), the neural region responsible for physiological responses to stress and panic, as well as the substantia nigra (0.4 mM), the center for dopamine production in the brain [32]. The major oxidation states for Cu ions in biological systems are cuprous CuI and cupric CuII ; the former is more common in the reducing intracellular environment, and the latter is dominant in the more oxidizing extracellular environment [33]. Levels of extracellular CuII vary from 10−25 µM in blood serum, 0.5−2.5 µM in cerebrospinal fluid (CSF), and 30 µM in the synaptic cleft. Intracellular Cu levels within neurons can reach concentrations higher than 2–3 orders of magnitude [32]. Like Zn and Fe, brain Cu is partitioned into tightly bound and labile pools. Owing to its redox activity, Cu is an essential cofactor in numerous enzymes that handle the chemistry of oxygen or its metabolites, including cytochrome c oxidase (CcO), Cu, Zn superoxide dismutase (Cu,Zn-SOD1), ceruloplasmin (CP), dopamine β monooxygenase (DβM), peptidylglycine α-hydroxylating monooxygenase (PHM), and tyrosinase [31]. Because of its propensity to trigger aberrant redox chemistry and oxidative stress when unregulated, the brain maintains strict control over its Cu levels and distributions. An overview of homeostatic Cu pathways in the brain is given in Figure 1.1. Many of the fundamental concepts for neuronal Cu homeostasis are derived from studies in yeast, but the brain provides a more complex system with its own unique and largely unexplored inorganic physiology. There is little ‘‘free’’ Cu in the yeast cytoplasm, which is due to the tight regulation of metallochaperones [35, 36]; however, many open questions remain concerning the homeostasis of organelle Cu stores, particularly in higher organisms with specialized tissues. Some data suggest that yeast and mammals possess pools of labile Cu in the mitochondrial matrix [37]. Uptake of Cu by the blood–brain barrier (BBB) is considered to occur through the P-type ATPase ATP7A, which can pump Cu into the brain [38]. Mutations in the related gene lead to Menkes disease, an inherited neurodegenerative disorder that is globally characterized by brain Cu deficiency. This phenotype is mirrored by Wilson disease, which involves mutations in the ATP7B gene responsible for excretion of excess Cu from the liver into the bile. Loss of ATP7B function leads to abnormal increase of Cu in the liver [39]. The extracellular trafficking of brain Cu differs from that in the rest of the body. CSF, the extracellular medium of the brain and central nervous system, possesses a distinct Cu homeostasis from blood plasma, which carries Cu to organs in the rest of the body. Cp, a multicopper oxidase that is essential for Fe metabolism, is

5

6


Cu+

1

tr hC

Cu2+

Synapse

Reductase MT3 c

CcO

E Sco1/ Cox11

Atox1

S CC

PrP

ome ndos

Cox17

Metalloprotein

SOD1 AT

P7

?

?

A

Presynaptic cell

AP

P

Postsynaptic cell

Figure 1.1 A schematic model of neuronal copper homeostasis. Reprinted with permission from [34]. Copyright 2008 American Chemical Society.

the major carrier of CuII in the plasma, but houses less than 1% of Cu in CSF [32]. The primary protein or small-molecule ligands for Cu in CSF remain unidentified. Uptake of Cu into brain cells requires reduction of CuII to CuI . Steap proteins may fulfill this role like the yeast ferric and cupric reductases Fre1 and Fre2 [40]. Following reduction, CuI ions can be transported into cells through a variety of trafficking pathways [41, 42]. A major class of proteins involved in cellular Cu uptake is the copper transport (Ctr) protein family. Ctr1 is a representative member that is ubiquitously expressed. It resides predominantly in the plasma membrane and is essential for the survival of mammalian embryos and for Cu import into neurons and astrocytes [43]. Elevated Cu stimulates rapid endocytosis and degradation of Ctr [44]. Ctr1 contains three transmembrane helices, an N-terminal extracellular domain, and a C-terminal cytosolic domain. Electron crystallography revealed that Ctr1 is trimeric and possesses the type of radial symmetry associated with the structure of certain ion channels [45]. A region of low protein density at the center of the trimer is consistent with the existence of a Cu permeable pore. Mutagenesis studies have established that a methionine(Met)-rich motif in the N-terminal domain and a Met-rich motif at the extracellular end of the second transmembrane helix of Ctr1 play a pivotal


role in the mechanism of Cu uptake [46, 47]. The mechanisms of Cu translocation across cellular membranes, however, remain largely unknown. In addition to Ctr1, prion protein (PrP) and amyloid precursor protein (APP) are two other abundant Cu-binding proteins, found specifically at brain cell surfaces, implicated in Cu uptake/efflux [48, 49]. In particular, PrP is localized in synaptic membranes of presynaptic neurons. Mammalian PrP contains at least four octapeptide repeats in the N-terminal region that can bind CuII . Millimolar concentrations of CuII induce endocytosis of PrP, suggesting that PrP may act as a buffer for Cu in the synaptic cleft, maintaining presynaptic Cu concentrations while preventing CuII -related toxicity in the extracellular space [49, 50]. Upon its entry into brain cells, CuI can be funneled to its ultimate intracellular destinations through the use of Cu chaperone proteins or buffering by metallothioneins (MTs), such as MT1 (ubiquitously expressed) and MT3 (expressed in the brain) [51]. The metallochaperones function not only as intracellular Cu delivery proteins but also as protective agents against toxicity resulting from unbound and unregulated Cu ions [35, 36]. Three human Cu chaperones have been characterized so far: Atox1, CCS, and Cox17. Atox1 loads CuI into the Menkes and Wilson P-type ATPases, ATP7A and ATP7B, which mediate Cu delivery to the secretory pathway from the trans-Golgi network (TGN) to the plasma membrane [41]. Both Atox1 and ATP7A/B contain CXXC sequence motifs that are essential for CuI binding and exchange of CuI between the two partner proteins [52]. The combination of available structural and biochemical data suggests a docking model that involves CuI transfer through twoand three-coordinate intermediates [53–56]. ATP7A and ATP7B play multiple roles in neurons from the delivery of Cu to cuproenzymes involved in neurotransmitter synthesis and metabolism, such as DβM, to the removal of excess Cu via secretion or vesicular sequestration [39]. To carry out this function, ATP7A undergoes Cu-stimulated translocation from the Golgi to the plasma membrane [57]. Metabolic studies also revealed that translocation of ATP7A after N-methyl d-aspartate (NMDA) receptor activation is associated with rapid release of Cu from hippocampal neurons [58, 59], a finding that suggests a role for Cu in the modulation of synaptic activity [60]. The copper chaperone for superoxide dismutase (CCS) inserts Cu into SOD [61]. Cu,Zn-SOD1 is a ubiquitous component of the cellular antioxidant system, which catalyzes disproportionation of the superoxide anion to oxygen and hydrogen peroxide [62]. Active Cu,Zn-SOD1 is a dimer; each subunit binds one Cu and one Zn ion, and contains an intramolecular disulfide bond [63]. On the contrary, in the immature, apo form of SOD1, cysteines are in the reduced state and the protein is a monomer [64, 65]. CCS docks with and transfers the Cu ion to the latter form of SOD [66], and also catalyzes disulfide bond formation [67, 68]. CCS is made of three domains: the N-terminal domain I has a fold similar to Atox1 and contains a conserved CXXC motif, domain II has a fold similar to SOD1 and participates in target recognition, domain III is constituted by ∼30 residues and contains a CXC Cu-binding motif [69]. While domain III is essential for CCS activity in vivo, the requirement for domain I is only apparent under Cu-limiting conditions [69].

7

8


A third Cu chaperone, Cox17, is involved in Cu delivery to CcO in mitochondria [70, 71]. CcO is a 13-subunit complex embedded in the mitochondrial inner membrane and a key component of the respiratory chain that reduces oxygen to water [72]. Two Cu ions form a dicopper cluster, designated CuA , in a CcO subunit, while a third Cu ion, designated CuB , forms a dinuclear site with heme a in another CcO subunit [72]. Cox17 acts as CuI donor for Sco1 and Cox11 [73]. Sco1, in turn, transfers Cu to CuA [74], while Cox11 is involved in CuB assembly [75]. Cox17 contains two CX9 C motifs implicated in two intramolecular disulfide bonds [76] and a conserved CC motif essential for CuI binding [77]. Fully reduced Cox17 binds up to four CuI ions in a polycopper cluster and undergoes oligomerization [76, 78].

1.4 Brain Copper and Neurodegenerative Disorders

Disruption of Cu homeostasis is implicated in a number of neurodegenerative diseases, including Alzheimer’s disease (AD), prion diseases, Parkinson’s disease (PD), familial amyotrophic lateral sclerosis (ALS), Menkes disease, and Wilson disease [5]. In all these disorders, the deleterious effects of Cu stem from its dual abilities to bind ligands and trigger uncontrolled redox chemistry. A dominant risk factor associated with most neurodegenerative diseases is increasing age. A positive correlation with chronic occupational exposure to Cu and other metal ions in industrialized countries has also been recognized [79, 80]. Several studies have reported a rise in the levels of brain Cu from youth to adulthood [32]. However, biologically available Cu levels drop markedly with advanced age and in AD brain [20, 81]. The connection between Cu and AD pathology is due mainly to its reactions with APP and its β-amyloid cleavage product (Aβ), that result in imbalance of extracellular and intracellular brain Cu pools [20]. Aberrant binding of CuII to APP triggers its reduction to CuI with concomitant disulfide bond formation; this intermediate can then participate in reactive oxygen species (ROS) production [82]. Extracellular Aβ deposits from AD brains (amyloid plaques) are rich in Cu, in addition to Zn and Fe [83]. The MT3, released in the synaptic cleft by neighboring astrocytes, has the potential to ameliorate this adverse interaction, but is downregulated in AD [84]. Moreover, the β-secretase β-site of amyloid precursor protein cleaving enzyme (BACE1), involved in APP cleavage, possesses a CuI -binding site in its C-terminal cytosolic domain through which it interacts with domain I of CCS, indicating that intracellular Cu levels can have an impact on Aβ generation [85]. Altered brain Cu distribution in AD, with abnormal accumulation of Cu in amyloid plaques and Cu deficiency in neighboring cells, is accompanied by a loss of Cu-dependent enzymes (e.g., CcO, Cu,Zn-SOD1, CP). Therefore, administration of Cu chelators such as clioquinol, that can reverse Aβ aggregation and redistribute brain Cu pools acting as ionophores, can have dual beneficial effects [20]. It is also found that Cu-bound clioquinol and other Cu complexes can exhibit proteasome-inhibitory abilities [86, 87].

1.5 The Role of Ubiquitin in Protein Degradation

Prion diseases are also linked to brain Cu misregulation, where opposing CuII and MnII levels may influence the conversion of PrP into the toxic, protease-resistant form, PrPSc [88, 89]. PrP may act as a Cu-chelating agent, when extracellular Cu reaches high concentration peaks (15–300 µM) such as during synaptic transmission and depolarization [8]. Another hypothesis is that the binding of Cu to PrP could act directly to detoxify ROS, performing SOD-like activity [90]. In one proposal for prion toxicity, excess free Cu further exacerbates the disease by promoting oxidative stress [91]. Onset of PD is accompanied by death of dopaminergic neurons and intracellular accumulation of Lewy bodies [92], which are protein aggregates containing α-synuclein (α-syn), an abundant protein in the brain whose function is still unclear [93]. Monomeric α-syn, which has no persistent structure in aqueous solution, is known to bind anionic lipids [94] with a resulting increase in α-helix structure [95, 96]. Factors including oxidative stress and presence of various metal ions promote its fibrillation [97, 98]. CuII also promotes the self-oligomerization of α-syn [10] and its oxidation and aggregation in the presence of H2 O2 [99]. Although α-syn is widely expressed in the brain, inclusions of α-syn are commonly localized in the substantia nigra, locus coeruleus, and cerebral cortex, which are the regions where Cu is abundant [32]. In PD brain, increased levels of Cu are found in the CSF [100]. Familial ALS is an inherited neurodegenerative disorder stemming from mutations in Cu,Zn-SOD [101, 102]. Three main hypotheses exist regarding the molecular mechanisms of this disease: (i) the loss-of-function mechanism, which results in toxic accumulation of superoxide by lack of SOD1 protection, (ii) the gain-of-function mechanism, in which SOD1 exhibits enhanced peroxidase activity by aberrant redox chemistry, and (iii) the aggregation mechanism, where SOD1 aggregates are induced by decreased availability of Cu and Zn ions and are stabilized by intermolecular disulfide bonds [103, 104]. The role of Cu homeostasis in this disease remains uncertain, however molecular machineries controlling redox homeostasis in mitochondria [68] appear to be essential for intramolecular disulfide bond formation and correct metal incorporation into SOD1, two processes involving the CCS metallochaperone [105].


The unique morphology of neurons (with specialized zones for presynaptic neurotransmitter release and postsynaptic receptor activation) and the plasticity of synapses (which is tightly coupled to changes in the synaptic proteome) impose special challenges on the cellular machinery for both protein synthesis and degradation [106]. Protein degradation has important roles in both neuronal development and long-term synaptic plasticity. Moreover, many neurodegenerative diseases are associated with abnormal protein aggregates, implicating degradative dysfunction [4, 107].

9

10


Major proteases in eukaryotic cells are confined to specialized protein complexes (proteasomes) and organelles (lysosomes) to prevent nonspecific proteolysis. The ubiquitin-proteasome system (UPS) is responsible for degrading most intracellular, soluble proteins, but it can also degrade transmembrane proteins if they are extracted from the membrane into the cytosol (Figure 1.2). The lysosome degrades most membrane and endocytosed proteins, but it can also digest cytosolic proteins through autophagy (Figure 1.3) [1]. In both cases ubiquitin (Ub) plays a crucial role. Ub is a small protein of 76 aminoacids folded in a compact globular structure in which a mixed parallel/antiparallel β-sheet packs against an α-helix generating a hydrophobic core [108]. Not found in bacteria, this protein is ubiquitous in eukaryotes and has highly conserved sequences, the human and the yeast proteins differing by only three residues. The remarkable degree of sequence conservation underscores its important physiological role [109]. Ubiquitination is a posttranslational modification that forms an isopeptide bond between a lysine residue on the protein and the C-terminus of Ub. The ubiquitination requires four different classes of enzymes: E1–E4 [110, 111]. First, Ub is covalently conjugated to E1 (Ub-activating enzyme) in an ATP-dependent reaction, and then it is transferred to E2 (Ub-conjugating enzyme). E3 (Ub-protein ligase) then transfers Ub from E2 to the substrate protein and is largely responsible for target recognition through physical interactions with the substrate. After the first Ub has been attached (monoubiquitination), E3 can also elongate the Ub chain by creating Ub–Ub isopeptide bonds. Finally, E4 enzymes (chain elongation factors) are a subclass of E3-like enzymes that only catalyze chain extension (Figure 1.2) [110, 111]. Ub has seven lysine residues (K6, K11, K27, K29, K33, K48, and K63), all of which are available and indeed used in vivo for chain extension. The significance of complex ubiquitination patterns is only partially understood: K48 chains lead to degradation of the substrate by the 26S proteasome, whereas monoubiquitination and K63 chains are known to activate cell signals in several pathways including tolerance of DNA damage, inflammatory response, ribosomal protein synthesis, endocytosis, and protein trafficking [111]. The 26S proteasome is a large multisubunit complex of ∼2 MDa, localized in the cytosol and nucleus, and composed of a 20S proteolytic core and one or two 19S regulatory caps. After substrate–proteasome association, deubiquitinating enzymes (DUBs) and ATP-dependent unfoldase activities help the substrate to enter the proteolytic lumen of the 20S core by regenerating monomeric Ub [110, 111]. Notably, the cleavage of the isopeptide bond between the substrate and the most proximal Ub of the polyUb chain requires the metalloprotease activity of a 19S proteasome subunit, which contains a JAMM (JAB1 (Jun activation domain-binding protein-1)/MPN (Mpr1 Pad1 N-terminal domain)/Mov34 metalloenzyme) domain with a coordinated ZnII ion [112, 113]. Ub is also involved in the lysosomal degradation pathway. Lysosomes are organelles that contain acid hydrolases that break down biomolecules. The hydrolases in the lumen of lysosomes (pH 4–5) and late endosomes (pH 5–6) are highly active in acidic environments but loose their activities in the cytosol (pH ∼ 7.2) [114]. Many types of signals can regulate endocytosis and sorting of lysosomes, including


11

ATP ADP E2

E2

E1 Ubiquitin

E3 E3/E4

(K63)

DUB

DUB Target protein

DUB

DUB E3/E4 (K48) Proteasome translocation

Diffusion, chaperone, shuttling factor

26S proteasome

ATP ADP

DUB, unfoldase

Figure 1.2 The ubiquitin-proteasome system. Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Neuroscience [1], copyright 2008.

12


Micro autophagy

Exocytosis Endocytosis Recycling endosome

Macro autophagy Early endosome

Autophagosome

Lysosome Multivesicular body

Surface protein

Chaperone-mediated autophagy

Late endosome

Ubiquitin

Acid hydrolase

LAMP2 protein

Cytosolic protein

Chaperone

Figure 1.3 The autophagy-lysosome pathway. Reprinted by permission from Macmillan Publishers Ltd: Nature Reviews Neuroscience [1], copyright 2008.

monoubiquitination and K63-polyubiquitination [115, 116]. Intracellular proteins can enter lysosomes through several autophagic mechanisms. In macroautophagy, large amounts of cytosolic materials or even organelles are surrounded by a double-membrane structure (autophagosome) that fuses with lysosomes. In microautophagy, a small amount of the cytoplasm is internalized through lysosomal invagination (Figure 1.3) [117, 118]. Different classes of misfolded proteins partition between two separate intracellular compartments, one next to the nucleus (juxtanuclear) and the other near vacuoles (perivacuolar). Soluble ubiquitinated proteins go to the juxtanuclear compartment, whereas insoluble proteins accumulate in the perivacuolar compartment [119]. In addition to the spatial segregation, the fate of the proteins in these two compartments is divergent. Proteins in the juxtanuclear compartment are in close proximity to the 26S proteasome, whereas the perivacuolar compartment is targeted by proteins implicated in autophagy. Misfolded proteins may interact inefficiently with the normal quality control machinery and as a consequence are shunted to the perivacuolar pathways, as observed for PrP [119]. Enhancing the ubiquitination of a PrP enhances its targeting to the juxtanuclear compartment. Likewise, blocking the ubiquitination of proteins that normally go to the juxtanuclear compartment leads them to the perivacuolar compartment.

1.6 Failure of the Ubiquitin System in Neurodegenerative Disorders

1.6 Failure of the Ubiquitin System in Neurodegenerative Disorders

Protein misfolding and the subsequent assembly of protein molecules into aggregates of various morphologies represent common mechanisms that link a number of important human disorders, known as conformational diseases. These disorders include numerous neurodegenerative diseases and many systemic, localized, and familial amyloidoses [29]. A hallmark of these diseases is the accumulation of insoluble protein aggregates, termed amyloid, within affected cells. Amyloid fibrils have a characteristic cross-β structure with a ribbon-like β-sheet that extends over the length of the fibril and is comprised of β-strands that run approximately perpendicular to the long axis of the fibril. Backbone H-bonds that link the β-strands are nearly parallel to the fibril axis [30]. Numerous recent reports suggest that the toxicity of amyloidogenic proteins lies not in the insoluble fibrils that accumulate but rather in the soluble oligomeric intermediates [16]. Early oligomers have been identified as the primary toxic species for a number of amyloid diseases and shown to be cytotoxic to cells in culture and in tissue [15, 120]. These soluble oligomers include spherical particles and curvilinear structures, called protofibrils, that appear to represent strings of the spherical particles [30]. An antibody specifically recognizes soluble oligomers formed by several types of amyloidogenic proteins and peptides, which indicates that they have a common structure and may share a common pathogenic mechanism [17]. Such a mechanism appears to be related to the ability of these oligomers to interact with cell membranes and to form annular protofibrils that cause membrane permeabilization [19, 121, 122]. Disruption of intracellular CaII homeostasis and ROS production are among the earliest biochemical modifications following the interaction of protofibrillar assemblies with membranes [123]. In the case of extracellular aggregates, cell damage may be accomplished via a nonspecific pathway (pores) or through oligomer interaction with specific cell surface receptors, such as glutamatergic receptors involved in CaII influx [124]. Most, if not all, protein inclusions associated with neurodegenerative diseases are enriched with Ub [4]. Lewy bodies in PD, amyloid plaques and neurofibrillary tangles in AD, skeinlike inclusions and Lewy body-like inclusions in ALS, polyglutamine inclusions in Huntington’s disease (HD): all these inclusions show Ub immunoreactivity (Figure 1.4). Ub immunoreactivity is also present in vesicles in some areas of degeneration in AD [3, 125]. The frequent detection of proteasomes and lysosomes around Ub-enriched aggregates in postmortem brains implies that proteins within these inclusions are marked for degradation but not efficiently removed. Hence, it has been proposed that neurodegeneration might be linked to degradative dysfunction by several mechanisms, and protein aggregates might arise, at least in part, from impaired proteasomal and/or autophagic removal of the damaged proteins [4, 107]. One prevalent idea is that misfolded proteins resist degradation by, but not engagement with, the proteasome, resulting in proteasome inhibition. In this model, soluble misfolded proteins would be tagged with Ub and partly enter the

13

14


Ab

Ubiquitin

Merged

Figure 1.4 Ubiquitin-positive amyloid plaques from AD brain revealed by immunofluorescence. Reprinted by permission from Wiley-Blackwell: Journal of Cellular and Molecular Medicine, copyright 2008 [126].

proteasome, but because of their abnormal structures, would resist full entry and cause steric occlusion [127]. In particular, β-sheet structures, which have been implicated in aberrant conformations in several diseases, seem to be difficult for the proteasome to process when they occur at the site at which misfolding initiates. Moreover, overall proteasome activity in brain tissue decreases with age, and further loss is observed in degenerative conditions such as AD and PD [128]. Importantly, impairment of the UPS is an early response to soluble protein aggregates and not a consequence of inclusion formation [129]. Alternatively, the impairment of degradation might be responsible for the etiology of disease, and aggregate formation might be a secondary phenomenon [4]. The link between the Ub system and neurodegeneration has been strengthened by the identification of disease-causing mutations in genes coding for Ub [130] and several ubiquitination enzymes related to AD and PD [131]. The most compelling evidence for the involvement of the UPS in AD pathogenesis comes from a transcriptional misreading, which causes the deletion of two nucleotides in the mRNA coding for Ub [130]. The resulting frameshift mutant form of Ub, called UBB+1 , has a 19-aminoacid extension at the C-terminus and cannot bind to target proteins. However, it can be ubiquitinated by wild-type Ub. The polyubiquitinated UBB+1 cannot be degraded by the proteasome and probably inhibits its activity. Furthermore, it cannot be deubiquitinated [132]. One of the more common types of familial PD is caused by mutations in the gene that encodes the E3 Ub-ligase parkin [133]. Parkin is reported to possess monoubiquitination and K63-polyubiquitination activities [134]. In addition to parkin, an increasing number of E3 genes are now linked to neurodegenerative disorders [1]. Mutations in parkin cause early disease onset, with the loss of dopaminergic neurons in the substantia nigra in the general absence of Lewy bodies. On the other hand, the genetic association of UCHL1 (ubiquitin C-terminal hydrolase L1), a DUB, with familial and sporadic PD is controversial, but UCHL1 is found in Lewy bodies [131].

1.7 Interaction of Ubiquitin with Metal Ions

UCHL1 is a neuron-specific DUB and one of the most abundant proteins in the brain, which helps to maintain monomeric Ub levels. It possesses DUB activity in its monomeric form and ligase activity in its dimeric form. A mutation linked to familial PD promotes its dimerization and the K63-polyubiquitination of α-syn [135]. The cellular concentrations of the two forms of Ub, free Ub and polyUb chains, are closely interconnected and may change because of various cellular events; for instance heat stress induces an increase in polyUb chains at the expenses of free Ub [136]. The finding that modest reduction in the level of free Ub in brain is linked to synaptic dysfunction and neuronal degeneration associated with loss-of-function mutations in DUBs, suggests that adequate neuronal Ub supply should be maintained. Decreased Ub availability in neurons of mice is sufficient to cause neuronal dysfunction and death [137]. On the other hand, increased Ub concentrations in CSF of patients affected by neurodegenerative diseases and amyloidosis may have a neuroprotective effect [138]. The presence of Ub within inclusion bodies was noted in a large variety of neurodegenerative diseases [3, 4, 107, 125], suggesting that disruption of Ub homeostasis may be a common factor in the pathogenesis of these disorders [137].

1.7 Interaction of Ubiquitin with Metal Ions 1.7.1 Thermal Stability of Ubiquitin

Ub has been widely used as model for stability, folding, and structural studies, the protein remaining soluble and folded over a wide pH range and at high temperatures [139]. The effect of various cations on the thermal stability of Ub was assessed by differential scanning calorimetry (DSC) experiments carried out on Ub solutions of different metals (CuII , ZnII , NiII , AlIII , and CdII ). CuII , in contrast with other cations, has a specific negative effect on the thermal stability of Ub, both in terms of unfolding temperature (Tm ) and enthalpy (H) [13]. Far–UV circular dichroism (CD) spectra indicated very little overall change in Ub secondary structure upon CuII binding, however thermal denaturation curves revealed a decrease of Tm from 100 ◦ C for native Ub to 90 ◦ C for the CuII –Ub system. 1.7.2 Spectroscopic Characterization of CuII Binding

Electrospray ionization mass spectrometry (ESI-MS) indicated binding of two CuII ions to the protein with ∼fourfold different affinities. The first CuII site has an affinity constant of ∼107 M−1 , as determined from spectrophotometric measurements, and the EPR (electron paramagnetic resonance) parameters (g|| = 2.30 and A|| = 159 × 10−4 cm−1 ) are consistent with a tetragonal N1 O3 (type II)

15

16


CuII site [13]. The unpaired electron of a type II CuII site is characterized by relatively long electronic relaxation times (10−8 –10−9 seconds); therefore, its coupling with the nuclear spins has a dramatic effect on the nuclear relaxation and consequently severely affects the NMR linewidths [140]. For most residues in the proximity of the metal center that completely escaped signal identification in 1 H-detected heteronuclear single quantum coherence (HSQC) spectra, 13 C resonances could be recovered using tailored NMR experiments based on 13 C direct detection [141, 142]. These latter experiments are intrinsically less affected by paramagnetism-enhanced relaxation than conventional experiments. As the paramagnetic dipolar contributions to nuclear relaxation depend on the square of the gyromagnetic ratio of the observed nucleus, going from 1 H(γH = 2.67 × 108 ) to 13 C(γC = 6.73 × 107 ) detection produces a decrease in relaxation rate by a factor of ∼16[140, 143]. The ‘‘protonless’’ approach was originally applied to a CuII -binding protein involved in Cu trafficking and homeostasis in bacteria [144, 145], and subsequently extended to human Cu,Zn-SOD1 [146, 147]. In the case of Ub, conventional and 13 C-detected experiments allowed the identification of resonances for most aminoacid residues with the only exception of those directly coordinated to the Cu ion: the first CuII site is located at the N-terminus of the protein and involves the Met1 nitrogen and three oxygen donor ligands (residues 16–18) in a tetragonal arrangement; the second CuII -anchoring site involves the imidazole nitrogen of His68 [13] (Figure 1.5). CuII has excited electronic states of relatively high energy, which produce small orbital contributions to the electronic spin moment [140, 143]. As a consequence, the magnetic susceptibility anisotropy of CuII is generally small. Nonetheless some nuclei of Ub experienced small but measurable pseudocontact shifts (PCSs) arising from dipolar coupling with the unpaired electron [13]. The size and the sign of PCS allowed to define the orientation of the magnetic susceptibility tensor χ, whose main axis is nearly orthogonal to the tetragonal plane formed by the CuII donor atoms. Such a CuII coordination environment was also supported by the presence

Met1

His68 Lys63 Figure 1.5

The two CuII -binding sites of ubiquitin determined by paramagnetic NMR [13].


of a weak absorption band at 680 nm in the visible spectrum, assignable to the d–d electronic transition [13]. All NMR spectral changes induced by CuII on Ub are abolished upon addition of stoichiometric amounts of ascorbic acid and reduction of CuII to CuI [13, 148]. Therefore, under ‘‘normal’’ conditions, in which intracellular Cu is mainly CuI , Ub does not bind Cu. On the other hand, addition of the tridentate ligand iminodiacetic acid (IDA) mobilizes CuII from the first to the second affinity site, in which three coordination positions of CuII are firmly taken by the IDA ligand, and the protein only provides the fourth donor atom coincident with an easily accessible nitrogen of the His68 imidazole ring [13, 149] (Figure 1.5). Cu reduction or chelation can suppress the CuII -induced protein destabilization. 1.7.3 Possible Implications for the Polyubiquitination Process

Since residues close to the N-terminus are involved in the formation of a β-strand, CuII binding may destabilize the protein starting from its N-terminal region. Potential oxygen donor ligands for the first site are Met1 and Val17 carbonyl groups, and Glu16 and Glu18 carboxylate groups [13]. In native Ub, the first residue, Met1, is involved in two key H-bonds [108]: one occurring between the amino-terminal group and the CO of Val17 and the other between the side chain sulfur atom of Met1 and the amide NH of Lys63 (Figure 1.5). Therefore, CuII binding to Ub might hamper the proteins’ turnover and other in vivo signaling events regulated by Lys63-polyubiquitination [111] such as the autophagic clearance of protein inclusions [150]. In the polyubiquitination process, Lys63 of the acceptor Ub attacks Gly76 (the C-terminal residue) of the donor Ub forming a Lys63–Gly76 isopeptide bond. The reaction is catalyzed by the Ub-conjugating enzyme Ubc13 and a Ub-ligase. The structural elucidation of Ub covalently bound to Ubc13 has revealed the molecular determinants of Lys63-polyubiquitination [151]. In this structure, several H-bonds help positioning the Lys63 of the acceptor Ub in the active site of Ubc13 near the C-terminus of the donor Ub [151]. On this basis, binding of CuII to the acceptor protein could compromise the correct positioning of Lys63 by affecting the electrostatic interactions at the interface between the conjugating enzyme and the acceptor Ub. 1.7.4 CuII -Induced Self-Oligomerization of Ub

Since conditions that destabilize the native state of a protein render the macromolecular system more prone to aggregation, the interaction of Ub with CuII was hypothesized to be a factor promoting Ub aggregation. Furthermore, CuII ions were found to target the most aggregation-prone regions of Ub (the N-terminus and His68), as predicted by an algorithm based on the propensity of two residues to be found paired in neighboring strands within a β-sheet [152].

17

18


CuII

TFE

CuII + TFE

1 µm

1 µm

1 µm

Figure 1.6 AFM images of ubiquitin aggregate morphologies upon long-term incubation with CuII and/or TFE [14].

These predictions were corroborated by an in vitro study of Ub aggregation [14]. It was observed that micromolar concentrations of CuII ions were sufficient to induce self-oligomerization of Ub. Ub oligomers appear as spherical particles ranging from 5 to 25 nm (Figure 1.6), that can progress from dimers to large Ub conglomerates resistant to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). CuII reduction by ascorbic acid or CuII chelation by ethylenediaminetetraacetic acid (EDTA) or IDA, can trigger disruption of Ub oligomers. On the other hand, when the CuII -stabilized oligomers were added to a low-polarity medium, Ub aggregation increased dramatically. 1.7.5 Cooperativity between CuII -Binding and Solvent Polarity

An aqueous solution with a moderate amount (20%, v/v) of 2,2,2-trifluoroethanol (TFE) was used to mimic the local decrease of dielectric constant in the proximity of a membrane surface [153]. In the absence of CuII , 20% TFE does not affect Ub secondary structure; the protein forms short beaded chains composed of SDS-sensitive spherical subunits of 6–8 nm (Figure 1.6). In contrast, in the mixture containing both CuII and TFE, the Ub structure shows an increase of β-sheet content, as determined by far-UV CD and attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy [14]. Atomic force microscopy (AFM) indicates that the aggregation process proceeds through distinct steps characterized by clustering of spherical particles in annular species, formation of trigonal branched structures growing radially from the annular species, and interconnection of these branched structures in filament networks (Figure 1.6). Large aggregates can be disrupted by CuII reduction yielding a homogeneous population of spherical particles similar to those formed in TFE alone. The CuII -stabilized spherical oligomers of Ub form annular and porelike structures, in liposomes and planar phospholipid bilayers respectively [14] (Figure 1.7). The membrane-penetrating structures are not observed after pretreatment with EDTA, which was shown to destroy the Ub oligomers formed by incubation with CuII . The negativity to tests for amyloid confirmed that the aggregation process of Ub does not lead to fibril formation [14].


19

nm 5 4 3 2 1 0 250

nm 6 4 2 0 600

200 400

600

nm

nm

150 100

250 200

200 400

200 0

150

50 100

nm

nm

50

0

0

0

1 µm

1 µm

Figure 1.7 AFM images of annular and porelike structures formed by CuII -stabilized ubiquitin oligomers in phospholipid bilayers [14].

Kinetic stabilization of α-syn protofibrils, under conditions inhibiting their conversion to mature fibrils, was shown to enhance the harmful effects of aggregation and to accelerate disease progression [15, 154–156]. It was also shown that natural lipids destabilize and rapidly revert inert Aβ amyloid fibrils to soluble neurotoxic protofibrils [157]. Changes in the Aβ peptide structure promoted by metal ions appear to modify the effect of Aβ on lipid membranes. In particular, addition of CuII induces a greater peptide association with lipids and membrane insertion, and also results in an increased β-sheet content for the peptide [158]. 1.7.6 Comparison with Other Metal Ions

CuII appears to play a unique role in Ub destabilization and aggregation. Other metal ions (ZnII , NiII , AlIII , and CdII ) do not affect Ub stability, even at very high metal : protein ratios [13]. The different behavior of various metal ions could be ascribed to differences in their coordination properties. The N-terminus of Ub is the preferred CuII -binding site in pure water as well as in water with 20%

20


Zn

Figure 1.8

Packing of ubiquitin molecules in orthorhombic ZnII -ubiquitin crystals [159].

TFE, as determined by NMR spectroscopy [14]. Despite the affinity of CuII for the imidazole nitrogen of histidine, the lack of a preorganized set of donor atoms near His68 renders the binding of CuII to this site less effective than its binding to the N-terminus, where the presence of carboxylate groups (Glu16 and Glu18) helps to accommodate the metal ion in a preorganized anchoring site. In contrast, His68 is the favorite binding site for ZnII , as deduced from the available X-ray structure of the ZnII –Ub adduct [159] (Figure 1.8). The ZnII ion completes its tetrahedral coordination by binding two water molecules and His68 from an adjacent Ub molecule. In the solid state, the adjacent Ub molecule plays a role similar to the IDA ligand in solution experiments, that is, to provide additional donor atoms to the metal ion. All trials to crystallize Ub in the presence of CuII were unsuccessful, even when using very small amounts of metal [159]. This could be a consequence of CuII destabilizing the protein structure and hampering crystallization. The analysis of molecular packing shows that in orthorombic crystals of the ZnII –Ub adduct, the N-terminus of one Ub molecule packs against helix α1 of another molecule. Each of the two Ub molecules, in turn, establishes complementary contacts with a third molecule, thus giving rise to a symmetric trimer stabilized by a series of intermolecular electrostatic interactions (H-bonds and salt bridges) (Figure 1.8). Helix α1 of native Ub contains a large number of unshielded backbone H-bonds, called dehydrons [160], which are generally correlated with membrane association [161] and aggregation propensity of a protein [162]. Mapping on the Zn–Ub structure reveals that dehydrons are clustered in the core of the trimer [159] (shown in black in Figure 1.8). Therefore, missing H-bond protection in the isolated Ub molecules is partially offset by crystal packing. Similarly, protein desolvation associated with a decrease in dielectric constant near a membrane surface is expected to foster intermolecular electrostatic interactions at the trimer interface. From this analysis it is also inferred that CuII binding at the N-terminus may enhance the aggregation propensity of these surface regions of Ub, thus shifting the equilibrium toward the formation of oligomeric species with bridging metal ions.

1.8 Biological Implications

1.8 Biological Implications 1.8.1 The Redox State of Cellular Copper

CuII reduction to CuI offsets Ub destabilization and aggregation [14]. As described above, Cu is generally believed to be transported in the cells by the plasma membrane permease Ctr1 in the +1 oxidation state; it is yet unproven, however, that CuI is the only permeant species [163]. There is unequivocal evidence that Ctr1 is not the only protein capable of mediating Cu entry into mammalian cells, and it is quite possible that CuII rather than CuI is transported by Ctr1-independent mechanisms, based on divalent metal ion inhibition [164, 165]. Consistent with the reducing environment of the cytosol, X-ray absorption spectroscopy indicates the presence of low-coordinate monovalent CuI in this compartment [33, 166, 167]; however, CuII appears to be abundant inside both normal and PD neurons of substantia nigra [168]. That CuII could gain access to special intracellular districts of normal or diseased cells is also supported by the recent identification of the Fre6 vacuolar cupric reductase [169]. In neurodegenerative diseases the intravesicular material undergoes ubiquitination [3, 170]. It has been also shown that a portion of α-syn is present in the lumen of intracellular vesicles and can be secreted to the extracellular space, thus possibly contributing to disease propagation [171]. Intravesicular α-syn is more prone to aggregation than its cytosolic counterpart [171] and this could be explained by the microenvironment of the vesicular lumen and increased metal ion concentrations. Preliminary findings indicate that α-syn and Ub, the two main components of Lewy bodies, form hybrid oligomers in the presence of CuII (Arnesano et al., unpublished observations). Intravesicular aggregates and Lewy bodies can then be propagated from diseased tissues to the cytosol of healthy neurons through prion-like transmission [172]. Importantly, it has been reported that extracellular aggregates including AD amyloids, can be internalized by mammalian cells, gain access to the cytosol, and colocalize with Ub and UPS components [173]. The mechanism of internalization is not clear; however, the recent discovery that PrP mediates Aβ-induced synaptic dysfunction [174], has led to the hypothesis that internalization of PrP may allow Aβ oligomers to reach intracellular compartments and interfere with proteasomal degradation [175] (Figure 1.9). Altogether, these findings strongly indicate that Ub may come in contact, in the cytosol, with extracellular aggregates enriched with oxidized CuII . Elemental mapping of amyloid deposits of AD brain revealed ‘‘hot spots’’ of accumulated metal ions, particularly Cu and Zn [83, 176]. Moreover, an increment of Cu density was observed in brain tissues that were positively stained for Ub [177], and AD amyloid plaques can be dissolved by CuII chelators [178]. Therefore, the interaction of Ub with CuII could be a pathological event taking place inside intracellular organelles or in the cytosolic compartment of cells which

21

22


Ab oligomer

Synapse

PrP coreceptor

PrP Lipid raft Neuron ?

?

Signal transduction

?

Transcription Proteasome

Nucleus

Figure 1.9 Prion and Aβ-induced synaptic toxicity. Reprinted by permission from Macmillan Publishers Ltd: Nature [175], copyright 2009.

attempt to tag and process toxic aggregates. Oxidative stress, membrane breaching, abnormal metal ion homeostasis, and metal miscompartmentalization can foster this process [6, 179]. 1.8.2 Ubiquitin and Phospholipids

Factors that favor protein and metal desolvation (e.g., lower dielectric constant near a membrane surface or in the proximity of inclusion bodies) [180] may significantly increase the CuII binding affinity and aggregation propensity of Ub. As an example, the CuII binding affinity of α-syn was shown to increase by 1 order of magnitude in the presence of lipids [181]. Lipids and vesicle membranes were found in Lewy bodies on autopsy [182], and lipid-bound soluble cytosolic oligomers were increased in brain extracts from patients with PD or dementia, with Lewy bodies [183]. The lipid composition of different subcellular compartment membranes can modulate protein folding and aggregation rates. Intracellular accumulation of Aβ was seen predominantly in multivesicular bodies and lysosomes and Aβ fibrillogenesis was found to be accelerated in the presence of endosomal and

1.9 Conclusions and Perspectives

lysosomal membranes. Furthermore, Aβ oligomerization was accelerated through the interaction with ganglioside clusters in lipid rafts [184]. Many physiologically relevant functions of Ub are carried out in the proximity of membrane surfaces. The ubiquitination process regulates endocytosis of membrane proteins, multivesicular body formation, and Golgi and endoplasmic reticulum functions [115]. A membrane-bound form of Ub, phosphatidyl–Ub, has been found in baculovirus particles [185] and in the virions of several other enveloped viruses [186]. The release of Ub from phosphatidyl–Ub, rather than the free cellular pool of Ub, facilitates protein ubiquitination events at membranes [187]. In-cell NMR spectroscopy allows one to determine how protein structures are influenced by their intracellular environment [188]. The application of this technique to the study of Ub in human cells has revealed an increase of protein dynamics and a decrease of folding stability in vivo as compared to those in vitro, as assessed by hydrogen exchange measurements [189]. This somewhat unexpected behavior has been attributed to nonspecific interactions of Ub with intracellular macromolecules, the cytoskeleton, and inner membranes [189], which may decrease the folding stability of Ub in cells. Ub exhibits large structural heterogeneity both in the free form and after binding to different partners [190]. Thus, binding events in cells might cause interconversion between different conformations of Ub, which may also destabilize folding and increase hydrogen exchanges rates. Alternatively, some of the substrates may preferentially bind to less folded states of Ub [190].

1.9 Conclusions and Perspectives

We found that Ub, the protein responsible for the cellular clearance of toxic aggregates linked to neurodegenerative disorders, is able to coordinate CuII and form in vitro assemblies similar to those it is supposed to break up [14]. The new findings underscore the importance of preserving cellular compartmentalization and Cu homeostasis for the correct functioning of protein degradation systems. Ub stability and function might be otherwise compromised [13]. The inorganic chemistry of the brain is inherently rich and remains an open frontier. The results linking Cu trafficking to neurodegenerative disorders are engaging, but much work remains to be done to fully elucidate the molecular roles of bound and labile metal ions, their physiological redox state and speciation, and their contributions to basic aspects of signaling within and between brain cells. Considerable attention is presently focused on the early stages of protein misfolding and aggregation (i.e., protein destabilization and oligomerization), with the aim of elucidating the molecular and biochemical basis of conformational and neurodegenerative disorders. Biophysical studies have been increasingly influential owing to their ability to provide a mechanistic rationale to better explain the effects of disease-causing mutations, oxidative stress, and the exacerbating role of Cu and other metal ions.

23

24


Of the features discussed above, NMR characterization of paramagnetic CuII proteins represents a challenge, however NMR experiments based on 13 C direct detection open new possibilities for solution and solid-state studies of disease-related CuII -binding proteins and their oligomeric assemblies [140, 143]. The obvious shortcoming of such studies is that they require experimental conditions that have little resemblance to the human brain. Moreover, the pathological processes are difficult to study in vitro, because it is difficult to reconstitute protein aggregation in an artificial environment. However, the advancements of in-cell NMR techniques [188], coupled to a variety of molecular and elemental imaging techniques including fluorescence and X-ray detection methods [34, 191], will allow detailed investigations at the atomic and microscopic level of globular, amyloid-forming, or intrinsically unstructured proteins in neuronal cells, and will help to characterize functional and pathogenic interactions of metal ions in vivo.

Acknowledgments

I am grateful to Prof. Giovanni Natile for precious suggestions and stimulating discussions. I also thank young collaborators and all the people involved in this research and, in particular, Prof. Ivano Bertini, Prof. Lucia Banci, and Prof. Enrico Rizzarelli. References 1. Tai, H.C. and Schuman, E.M. (2008) 2.

3.

4. 5.

6. 7.

8.

Nat. Rev. Neurosci., 9, 826–838. McClellan, A.J., Tam, S., Kaganovich, D., and Frydman, J. (2005) Nat. Cell Biol., 7, 736–741. Lowe, J., Blanchard, A., Morrell, K., Lennox, G., Reynolds, L., Billett, M., Landon, M., and Mayer, R.J. (1988) J. Pathol., 155, 9–15. Ciechanover, A. and Brundin, P. (2003) Neuron, 40, 427–446. Gaggelli, E., Kozlowski, H., Valensin, D., and Valensin, G. (2006) Chem. Rev., 106, 1995–2044. Barnham, K.J. and Bush, A.I. (2008) Curr. Opin. Chem. Biol., 12, 222–228. Davies, P., Fontaine, S.N., Moualla, D., Wang, X., Wright, J.A., and Brown, D.R. (2008) Biochem. Soc. Trans., 36, 1299–1303. Brown, D.R., Qin, K., Herms, J.W., Madlung, A., Manson, J., Strome, R., Fraser, P.E., Kruck, T.A., Von Bohlen, A., Schulz-Schaeffer, W., Giese, A., Westaway, D., and

9.

10.

11.

12.

13.

14.

Kretzschmar, H.A. (1997) Nature, 390, 684–687. Atwood, C.S., Moir, R.D., Huang, X., Scarpa, R.C., Bacarra, N.M., Romano, D.M., Hartshorn, M.A., Tanzi, R.E., and Bush, A.I. (1998) J. Biol. Chem., 273, 12817–12826. Paik, S.R., Shin, H.J., Lee, J.H., Chang, C.S., and Kim, J. (1999) Biochem. J., 340, 821–828. Morgan, C.J., Gelfand, M., Atreya, C., and Miranker, A.D. (2001) J. Mol. Biol., 309, 339–345. Drago, D., Bolognin, S., and Zatta, P. (2008) Curr. Alzheimer Res., 5, 500–507. Milardi, D., Arnesano, F., Grasso, G., Magri, A., Tabbi, G., Scintilla, S., Natile, G., and Rizzarelli, E. (2007) Angew. Chem. Int. Ed. Engl., 46, 7993–7995. ` V., Arnesano, F., Scintilla, S., Calo, Bonfrate, E., Ingrosso, C., Losacco, M., Pellegrino, T., Rizzarelli, E., and Natile, G. (2009) PLoS ONE, 4, e7052.

References 15. Lashuel, H.A., Hartley, D., Petre, B.M.,

16.

17.

18.

19. 20. 21.

22. 23.

24. 25.

26. 27.

28. 29. 30. 31. 32.

33.

Walz, T., and Lansbury, P.T. (2002) Nature, 418, 291. Bucciantini, M., Giannoni, E., Chiti, F., Baroni, F., Formigli, L., Zurdo, J., Taddei, N., Ramponi, G., Dobson, C.M., and Stefani, M. (2002) Nature, 416, 507–511. Kayed, R., Head, E., Thompson, J.L., McIntire, T.M., Milton, S.C., Cotman, C.W., and Glabe, C.G. (2003) Science, 300, 486–489. Kayed, R., Sokolov, Y., Edmonds, B., McIntire, T.M., Milton, S.C., Hall, J.E., and Glabe, C.G. (2004) J. Biol. Chem., 279, 46363–46366. Lashuel, H.A. (2005) Sci. Aging Knowledge Environ., 2005, e28. Bush, A.I. and Tanzi, R.E. (2008) Neurotherapeutics, 5, 421–432. Bolognin, S., Drago, D., Messori, L., and Zatta, P. (2009) Med. Res. Rev, 29, 547–570. Bush, A.I. (2000) Curr. Opin. Chem. Biol., 4, 184–191. Burdette, S.C. and Lippard, S.J. (2003) Proc. Natl. Acad. Sci. U.S.A, 100, 3605–3610. Thiele, D.J. and Gitlin, J.D. (2008) Nat. Chem. Biol., 4, 145–147. Sigel, A., Sigel, H., and Sigel, R. K. O. (eds) (2006) Neurodegenerative Diseases and Metal Ions: Metal Ions in Life Sciences, John Wiley & Sons, Ltd, Chichester, pp. 1–488. Madsen, E. and Gitlin, J.D. (2007) Annu. Rev. Neurosci., 30, 317–337. Barnham, K.J., Masters, C.L., and Bush, A.I. (2004) Nat. Rev. Drug Discov., 3, 205–214. Finney, L.A. and O’Halloran, T.V. (2003) Science, 300, 931–936. Soto, C. (2003) Nat. Rev. Neurosci., 4, 49–60. Chiti, F. and Dobson, C.M. (2006) Annu. Rev. Biochem., 75, 333–366. Kim, B.E., Nevitt, T., and Thiele, D.J. (2008) Nat. Chem. Biol., 4, 176–185. Smith, R.M. (1983) in Neurobiology of the Trace Elements, Vol. 1 (eds I.E.Dreosti and R.M. Smith), Humana Press, Clifton, pp. 1–40. Yang, L., McRae, R., Henary, M.M., Patel, R., Lai, B., Vogt, S., and Fahrni,

34.

35.

36.

37.

38. 39. 40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

C.J. (2005) Proc. Natl. Acad. Sci. U.S.A, 102, 11179–11184. Que, E.L., Domaille, D.W., and Chang, C.J. (2008) Chem. Rev., 108, 1517–1549. Rae, T., Schmidt, P.J., Pufahl, R.A., Culotta, V.C., and O’Halloran, T.V. (1999) Science, 284, 805–808. O’Halloran, T.V. and Culotta, V.C. (2000) J. Biol. Chem., 275, 25057–25060. Cobine, P.A., Pierrel, F., Bestwick, M.L., and Winge, D.R. (2006) J. Biol. Chem., 281, 36552–36559. Mercer, J.F. and Llanos, R.M. (2003) J. Nutr., 133, 1481S–1484S. Lutsenko, S. and Petris, M.J. (2003) J. Membr. Biol., 191, 1–12. Ohgami, R.S., Campagna, D.R., McDonald, A., and Fleming, M.D. (2006) Blood, 108, 1388–1394. Arnesano, F., Banci, L., Bertini, I., and Ciofi-Baffoni, S. (2004) Eur. J. Inorg. Chem., 2004, 1583–1593. Arnesano, F. and Banci, L. (2007) in Handbook of Metalloproteins (ed. A.Messerschmidt), John Wiley & Sons, Ltd. Lee, J., Prohaska, J.R., and Thiele, D.J. (2001) Proc. Natl. Acad. Sci. U.S.A., 98, 6842–6847. Petris, M.J., Smith, K., Lee, J., and Thiele, D.J. (2003) J. Biol. Chem., 278, 9639–9646. De Feo, C.J., Aller, S.G., Siluvai, G.S., Blackburn, N.J., and Unger, V.M. (2009) Proc. Natl. Acad. Sci. U.S.A, 106, 4237–4242. Puig, S., Lee, J., Lau, M., and Thiele, D.J. (2002) J. Biol. Chem., 277, 26021–26030. Guo, Y., Smith, K., Lee, J., Thiele, D.J., and Petris, M.J. (2004) J. Biol. Chem., 279, 17428–17433. Barnham, K.J., McKinstry, W.J., Multhaup, G., Galatis, D., Morton, C.J., Curtain, C.C., Williamson, N.A., White, A.R., Hinds, M.G., Norton, R.S., Beyreuther, K., Masters, C.L., Parker, M.W., and Cappai, R. (2003) J. Biol. Chem., 278, 17401–17407. Brown, D.R. and Kozlowski, H. (2004) Dalton Trans., 1907–1917.

25

26

1 The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders 50. Mendola, D.L., Pietropaolo, A.,

51.

52.

53.

54.

55.

56.

57.

58. 59.

60. 61.

62. 63.

64.

65.

Pappalardo, G., Zannoni, C., and Rizzarelli, E. (2008) Curr. Alzheimer Res., 5, 579–590. Hidalgo, J., Aschner, M., Zatta, P., and Vasak, M. (2001) Brain Res. Bull., 55, 133–145. Arnesano, F., Banci, L., Bertini, I., Ciofi-Baffoni, S., Molteni, E., Huffman, D.L., and O’Halloran, T.V. (2002) Genome Res., 12, 255–271. Wernimont, A.K., Huffman, D.L., Lamb, A.L., O’Halloran, T.V., and Rosenzweig, A.C. (2000) Nat. Struct. Biol., 7, 766–771. Arnesano, F., Banci, L., Bertini, I., Cantini, F., Ciofi-Baffoni, S., Huffman, D.L., and O’Halloran, T.V. (2001) J. Biol. Chem., 276, 41365–41376. Arnesano, F., Banci, L., Bertini, I., and Bonvin, A.M.J.J. (2004) Structure, 12, 669–676. Banci, L., Bertini, I., Cantini, F., Felli, I.C., Gonnelli, L., Hadjiliadis, N., Pierattelli, R., Rosato, A., and Voulgaris, P. (2006) Nat. Chem. Biol., 2, 367–368. Petris, M.J., Mercer, J.F., Culvenor, J.G., Lockhart, P., and Camakaris, J. (1996) EMBO J., 15, 6084–6095. Schlief, M.L., Craig, A.M., and Gitlin, J.D. (2005) J. Neurosci., 25, 239–246. Schlief, M.L., West, T., Craig, A.M., Holtzman, D.M., and Gitlin, J.D. (2006) Proc. Natl. Acad. Sci. U.S.A, 103, 14919–14924. Schlief, M.L. and Gitlin, J.D. (2006) Mol. Neurobiol., 33, 81–90. Sturtz, L.A., Diekert, K., Jensen, L.T., Lill, R., and Culotta, V.C. (2001) J. Biol. Chem., 276, 38084–38089. Fridovich, I. (1995) Annu. Rev. Biochem., 64, 97–112. Bordo, D., Djinovic, K., and Bolognesi, M. (1994) J. Mol. Biol., 238, 366–386. Arnesano, F., Banci, L., Bertini, I., Martinelli, M., Furukawa, Y., and O’Halloran, T.V. (2004) J. Biol. Chem., 279, 47998–48003. Doucette, P.A., Whitson, L.J., Cao, X., Schirf, V., Demeler, B., Valentine, J.S., Hansen, J.C., and Hart, P.J. (2004) J. Biol. Chem., 279, 54558–54566.

66. Rae, T.D., Torres, A.S., Pufahl, R.A.,

67.

68. 69.

70.

71.

72. 73.

74.

75.

76.

77.

78.

79. 80.

81.

and O’Halloran, T.V. (2001) J. Biol. Chem., 276, 5166–5176. Furukawa, Y., Torres, A.S., and O’Halloran, T.V. (2004) EMBO J., 23, 2872–2881. Kawamata, H. and Manfredi, G. (2008) Hum. Mol. Genet., 17, 3303–3317. Schmidt, P.J., Rae, T.D., Pufahl, R.A., Hamma, T., Strain, J., O’Halloran, T.V., and Culotta, V.C. (1999) J. Biol. Chem., 274, 23719–23725. Beers, J., Glerum, D.M., and Tzagoloff, A. (1997) J. Biol. Chem., 272, 33191–33196. Cobine, P.A., Pierrel, F., and Winge, D.R. (2006) Biochim. Biophys. Acta, 1763, 759–772. Yoshikawa, S. (2002) Adv. Protein Chem., 60, 341–395. Horng, Y.C., Cobine, P.A., Maxfield, A.B., Carr, H.S., and Winge, D.R. (2004) J. Biol. Chem., 279, 35334–35340. Lode, A., Kuschel, M., Paret, C., and Rodel, G. (2000) FEBS Lett, 485, 19–24. Hiser, L., Di Valentin, M., Hamer, A.G., and Hosler, J.P. (2000) J. Biol. Chem., 275, 619–623. Arnesano, F., Balatri, E., Banci, L., Bertini, I., and Winge, D.R. (2005) Structure, 13, 713–722. Banci, L., Bertini, I., Ciofi-Baffoni, S., Janicka, A., Martinelli, M., Kozlowski, H., and Palumaa, P. (2008) J. Biol. Chem., 283, 7912–7920. Heaton, D.N., George, G.N., Garrison, G., and Winge, D.R. (2001) Biochemistry, 40, 743–751. Tanner, C.M. (1989) Trends Neurosci., 12, 49–54. Gorell, J.M., Johnson, C.C., Rybicki, B.A., Peterson, E.L., Kortsha, G.X., Brown, G.G., and Richardson, R.J. (1999) Neurotoxicology, 20, 239–247. Zecca, L., Stroppolo, A., Gatti, A., Tampellini, D., Toscani, M., Gallorini, M., Giaveri, G., Arosio, P., Santambrogio, P., Fariello, R.G., Karatekin, E., Kleinman, M.H., Turro, N., Hornykiewicz, O., and Zucca, F.A. (2004) Proc. Natl. Acad. Sci. U.S.A, 101, 9843–9848.

References 82. Multhaup, G., Schlicksupp, A., Hesse,

83.

84.

85. 86. 87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

L., Beher, D., Ruppert, T., Masters, C.L., and Beyreuther, K. (1996) Science, 271, 1406–1409. Lovell, M.A., Robertson, J.D., Teesdale, W.J., Campbell, J.L., and Markesbery, W.R. (1998) J. Neurol. Sci., 158, 47–52. Meloni, G., Sonois, V., Delaine, T., Guilloreau, L., Gillet, A., Teissie, J., Faller, P., and Vasak, M. (2008) Nat. Chem. Biol., 4, 366–372. Dingwall, C. (2007) Biochem. Soc. Trans., 35, 571–573. Chen, D. and Dou, Q.P. (2008) Expert Opin. Ther. Targets, 12, 739–748. Cvek, B., Milacic, V., Taraba, J., and Dou, Q.P. (2008) J. Med. Chem., 51, 6256–6258. Brown, D.R., Hafiz, F., Glasssmith, L.L., Wong, B.S., Jones, I.M., Clive, C., and Haswell, S.J. (2000) EMBO J., 19, 1180–1186. Kralovicova, S., Fontaine, S.N., Alderton, A., Alderman, J., Ragnarsdottir, K.V., Collins, S.J., and Brown, D.R. (2009) Mol. Cell Neurosci, 41, 135–147. Brown, D.R., Wong, B.S., Hafiz, F., Clive, C., Haswell, S.J., and Jones, I.M. (1999) Biochem. J., 344, 1–5. Requena, J.R., Groth, D., Legname, G., Stadtman, E.R., Prusiner, S.B., and Levine, R.L. (2001) Proc. Natl. Acad. Sci. U.S.A, 98, 7170–7175. Wakabayashi, K., Tanji, K., Mori, F., and Takahashi, H. (2007) Neuropathology, 27, 494–506. Maroteaux, L., Campanelli, J.T., and Scheller, R.H. (1988) J. Neurosci., 8, 2804–2815. Sharon, R., Goldberg, M.S., Bar-Josef, I., Betensky, R.A., Shen, J., and Selkoe, D.J. (2001) Proc. Natl. Acad. Sci. U.S.A, 98, 9110–9115. Eliezer, D., Kutluay, E., Bussell, R. Jr., and Browne, G. (2001) J. Mol. Biol., 307, 1061–1073. Ulmer, T.S., Bax, A., Cole, N.B., and Nussbaum, R.L. (2004) J. Biol. Chem., 11, 9595–9603. Uversky, V.N., Li, J., and Fink, A.L. (2001) J. Biol. Chem., 276, 44284–44296.

98. Rasia, R.M., Bertoncini, C.W.,

99.

100.

101.

102.

103.

104.

105.

106. 107. 108. 109.

Marsh, D., Hoyer, W., Cherny, D., Zweckstetter, M., Griesinger, C., Jovin, T.M., and Fernández, C.O. (2005) Proc. Natl. Acad. Sci. U.S.A., 102, 4294–4299. Paik, S.R., Shin, H.J., and Lee, J.H. (2000) Arch. Biochem. Biophys., 378, 269–277. Pall, H.S., Williams, A.C., Blake, D.R., Lunec, J., Gutteridge, J.M., Hall, M., and Taylor, A. (1987) Lancet, 2, 238–241. Rosen, D.R., Siddique, T., Patterson, D., Figlewicz, D.A., Sapp, P., Hentati, A., Donaldson, D., Goto, J., O’Regan, J., Deng, H.X., Rahmani, Z., Krizus, A., McKenna-Yasek, D., Cayabyab, A., Gatson, S.M., Berger, R., Tanzi, R.E., Halperin, J.J., Herzfeldt, B., Van der Bergh, R., Hung, W.-Y., Bird, Deng, G., Mulder, D.W., Smyth, C., Laing, N.G., Soriano, E., Pericak-Vance, M.A., Haines, J., Rouleau, G.A., Gusella, J.S., Horvitz, H.R., Brown, R.H. Jr. (1993) Nature, 362, 59–62. Valentine, J.S., Doucette, P.A., and Potter, S.Z. (2005) Annu. Rev. Biochem., 74, 563–593. Deng, H.X., Shi, Y., Furukawa, Y., Zhai, H., Fu, R., Liu, E., Gorrie, G.H., Khan, M.S., Hung, W.Y., Bigio, E.H., Lukas, T., Dal Canto, M.C., O’Halloran, T.V., and Siddique, T. (2006) Proc. Natl. Acad. Sci. U.S.A, 103, 7142–7147. Banci, L., Bertini, I., Durazo, A., Girotto, S., Gralla, E.B., Martinelli, M., Valentine, J.S., Vieru, M., and Whitelegge, J.P. (2007) Proc. Natl. Acad. Sci. U.S.A, 104, 11263–11267. Furukawa, Y., Kaneko, K., Yamanaka, K., O’Halloran, T.V., and Nukina, N. (2008) J. Biol. Chem., 283, 24167–24176. DiAntonio, A. and Hicke, L. (2004) Annu. Rev. Neurosci., 27, 223–246. Ross, C.A. and Pickart, C.M. (2004) Trends Cell Biol., 14, 703–711. Vijay-Kumar, S., Bugg, C.E., and Cook, W.J. (1987) J. Mol. Biol., 194, 531–544. Pickart, C.M. and Eddins, M.J. (2004) Biochim. Biophys. Acta, 1695, 55–72.

27

28

1 The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders 110. Hershko, A. and Ciechanover, A.

111. 112.

113. 114.

115. 116. 117. 118.

119.

120. 121.

122. 123.

124.

125. 126. 127. 128.

(1998) Annu. Rev. Biochem., 67, 425–479. Pickart, C.M. and Fushman, D. (2004) Curr. Opin. Chem. Biol., 8, 610–616. Verma, R., Aravind, L., Oania, R., McDonald, W.H., Yates, J.R. III, Koonin, E.V., and Deshaies, R.J. (2002) Science, 298, 611–615. Ambroggio, X.I., Rees, D.C., and Deshaies, R.J. (2004) PLoS Biol., 2, E2. Luzio, J.P., Pryor, P.R., and Bright, N.A. (2007) Nat. Rev. Mol. Cell Biol., 8, 622–632. Hicke, L. and Dunn, R. (2003) Annu. Rev. Cell Dev. Biol., 19, 141–172. Piper, R.C. and Luzio, J.P. (2007) Curr. Opin. Cell Biol., 19, 459–465. Klionsky, D.J. and Emr, S.D. (2000) Science, 290, 1717–1721. Todde, V., Veenhuis, M., and van der Klei, I. (2009) Biochim. Biophys. Acta, 1792, 3–13. Kaganovich, D., Kopito, R., and Frydman, J. (2008) Nature, 454, 1088–1095. Haass, C. and Selkoe, D.J. (2007) Nat. Rev. Mol. Cell Biol., 8, 101–112. Quist, A., Doudevski, I., Lin, H., Azimova, R., Ng, D., Frangione, B., Kagan, B., Ghiso, J., and Lal, R. (2005) Proc. Natl. Acad. Sci. U.S.A., 102, 10427–10432. Lashuel, H.A. and Lansbury, P.T. Jr. (2006) Q. Rev. Biophys., 39, 167–201. Demuro, A., Mina, E., Kayed, R., Milton, S.C., Parker, I., and Glabe, C.G. (2005) J. Biol. Chem., 280, 17294–17300. Pellistri, F., Bucciantini, M., Relini, A., Nosi, D., Gliozzi, A., Robello, M., and Stefani, M. (2008) J. Biol. Chem., 283, 29950–29960. Lowe, J., Hand, N., and Mayer, R.J. (2005) Methods Enzymol., 399, 86–119. Oddo, S. (2008) J. Cell Mol. Med., 12, 363–373. Bence, N.F., Sampat, R.M., and Kopito, R.R. (2001) Science, 292, 1552–1555. Gray, D.A., Tsirigotis, M., and Woulfe, J. (2003) Sci. Aging Knowledge Environ., 2003, RE6.

129. Bennett, E.J., Bence, N.F.,

130.

131.

132.

133.

134. 135.

136. 137.

138. 139. 140. 141.

142.

Jayakumar, R., and Kopito, R.R. (2005) Mol. Cell, 17, 351–365. van Leeuwen, F.W., de Kleijn, D.P., van den Hurk, H.H., Neubauer, A., Sonnemans, M.A., Sluijs, J.A., Koycu, S., Ramdjielal, R.D., Salehi, A., Martens, G.J., Grosveld, F.G., Peter, J., Burbach, H., and Hol, E.M. (1998) Science, 279, 242–247. Leroy, E., Boyer, R., Auburger, G., Leube, B., Ulm, G., Mezey, E., Harta, G., Brownstein, M.J., Jonnalagada, S., Chernova, T., Dehejia, A., Lavedan, C., Gasser, T., Steinbach, P.J., Wilkinson, K.D., and Polymeropoulos, M.H. (1998) Nature, 395, 451–452. Lindsten, K., de Vrij, F.M., Verhoef, L.G., Fischer, D.F., van Leeuwen, F.W., Hol, E.M., Masucci, M.G., and Dantuma, N.P. (2002) J. Cell Biol., 157, 417–427. Shimura, H., Hattori, N., Kubo, S., Mizuno, Y., Asakawa, S., Minoshima, S., Shimizu, N., Iwai, K., Chiba, T., Tanaka, K., and Suzuki, T. (2000) Nat. Genet., 25, 302–305. Moore, D.J. (2006) Biochem. Soc. Trans., 34, 749–753. Liu, Y., Fallon, L., Lashuel, H.A., Liu, Z., and Lansbury, P.T. Jr. (2002) Cell, 111, 209–218. Parag, H.A., Raboy, B., and Kulka, R.G. (1987) EMBO J., 6, 55–61. Ryu, K.Y., Garza, J.C., Lu, X.Y., Barsh, G.S., and Kopito, R.R. (2008) Proc. Natl. Acad. Sci. U.S.A, 105, 4016–4021. Sixt, S.U. and Dahlmann, B. (2008) Biochim. Biophys. Acta, 1782, 817–823. Jackson, S.E. (2006) Org. Biomol. Chem., 4, 1845–1853. Arnesano, F., Banci, L., and Piccioli, M. (2005) Q. Rev. Biophys., 38, 167–219. Bermel, W., Bertini, I., Duma, L., Emsley, L., Felli, I.C., Pierattelli, R., and Vasos, P.R. (2005) Angew. Chem. Int. Ed., 44, 3089–3092. Bermel, W., Bertini, I., Felli, I.C., Piccioli, M., and Pierattelli, R. (2006) Prog. NMR Spectrosc., 48, 25–45.

References 143. Bertini, I., Luchinat, C., Parigi, G.,

144.

145.

146.

147.

148. 149.

150.

151.

152.

153.

154.

155.

156.

157.

and Pierattelli, R. (2008) Dalton Trans., 3782–3790. Arnesano, F., Banci, L., Bertini, I., Felli, I.C., Luchinat, C., and Thompsett, A.R. (2003) J. Am. Chem. Soc., 125, 7200–7208. Arnesano, F., Banci, L., Bertini, I., Mangani, S., and Thompsett, A.R. (2003) Proc. Natl. Acad. Sci. U.S.A., 100, 3814–3819. Bermel, W., Bertini, I., Felli, I.C., Kümmerle, R., and Pierattelli, R. (2003) J. Am. Chem. Soc., 125, 16423–16429. Bertini, I., Felli, I.C., Luchinat, C., Parigi, G., and Pierattelli, R. (2007) Chembiochem, 8, 1422–1429. Rice, M.E. (2000) Trends Neurosci., 23, 209–216. Nomura, M., Kobayashi, T., Kohno, T., Fujiwara, K., Tenno, T., Shirakawa, M., Ishizaki, I., Yamamoto, K., Matsuyama, T., Mishima, M., and Kojima, C. (2004) FEBS Lett., 566, 157–161. Tan, J.M., Wong, E.S., Kirkpatrick, D.S., Pletnikova, O., Ko, H.S., Tay, S.P., Ho, M.W., Troncoso, J., Gygi, S.P., Lee, M.K., Dawson, V.L., Dawson, T.M., and Lim, K.L. (2008) Hum. Mol. Genet., 17, 431–439. Eddins, M.J., Carlile, C.M., Gomez, K.M., Pickart, C.M., and Wolberger, C. (2006) Nat. Struct. Mol. Biol., 13, 915–920. Trovato, A., Chiti, F., Maritan, A., and Seno, F. (2006) PLoS. Comput. Biol., 2, e170 . Munishkina, L.A., Phelan, C., Uversky, V.N., and Fink, A.L. (2003) Biochemistry, 42, 2720–2730. Conway, K.A., Lee, S.J., Rochet, J.C., Ding, T.T., Williamson, R.E., and Lansbury, P.T.Jr. (2000) Proc. Natl. Acad. Sci. U.S.A., 97, 571–576. Conway, K.A., Rochet, J.C., Bieganski, R.M., and Lansbury, P.T.Jr. (2001) Science, 294, 1346–1349. Herrera, F.E., Chesi, A., Paleologou, K.E., Schmid, A., Munoz, A., Vendruscolo, M., Gustincich, S., Lashuel, H.A., and Carloni, P. (2008) PLoS ONE, 3, e3394 . Martins, I.C., Kuperstein, I., Wilkinson, H., Maes, E.,

158.

159.

160. 161.

162.

163.

164.

165.

166.

167.

168.

169. 170.

171.

Vanbrabant, M., Jonckheere, W., Van, G.P., Hartmann, D., D’Hooge, R., De, S.B., Schymkowitz, J., and Rousseau, F. (2008) EMBO J., 27, 224–233. Lau, T.L., Ambroggio, E.E., Tew, D.J., Cappai, R., Masters, C.L., Fidelio, G.D., Barnham, K.J., and Separovic, F. (2006) J. Mol. Biol., 356, 759–770. Falini, G., Fermani, S., Tosi, G., Arnesano, F., and Natile, G. (2008) Chem.Commun.(Camb.), 5960–5962. Fernandez, A., Chen, J., and Crespo, A. (2007) J. Chem. Phys., 126, 245103. Fernandez, A. and Berry, R.S. (2003) Proc. Natl. Acad. Sci. U.S.A, 100, 2391–2396. Fernandez, A., Kardos, J., Scott, L.R., Goto, Y., and Berry, R.S. (2003) Proc. Natl. Acad. Sci. U.S.A, 100, 6446–6451. Maryon, E.B., Molloy, S.A., Zimnicka, A.M., and Kaplan, J.H. (2007) Biometals, 20, 355–364. Lee, J., Petris, M.J., and Thiele, D.J. (2002) J. Biol. Chem., 277, 40253–40259. Moriya, M., Ho, Y.H., Grana, A., Nguyen, L., Alvarez, A., Jamil, R., Ackland, M.L., Michalczyk, A., Hamer, P., Ramos, D., Kim, S., Mercer, J.F., and Linder, M.C. (2008) Am. J. Physiol. Cell Physiol., 295, C708–C721. Pufahl, R.A., Singer, C.P., Peariso, K.L., Lin, S.-J., Schmidt, P.J., Fahrni, C.J., Cizewski Culotta, V., Penner-Hahn, J.E., and O’Halloran, T.V. (1997) Science, 278, 853–856. Xiao, Z., Loughlin, F., George, G.N., Howlett, G.J., and Wedd, A.G. (2004) J. Am. Chem. Soc., 126, 3081–3090. Chwiej, J., Adamek, D., Szczerbowska-Boruchowska, M., Krygowska-Wajs, A., Bohic, S., and Lankosz, M. (2008) J. Trace Elem. Med. Biol., 22, 183–188. Rees, E.M. and Thiele, D.J. (2007) J. Biol. Chem., 282, 21629–21638. Doherty, F.J., Osborn, N.U., Wassell, J.A., Heggie, P.E., Laszlo, L., and Mayer, R.J. (1989) Biochem. J., 263, 47–55. Lee, H.J., Patel, S., and Lee, S.J. (2005) J. Neurosci., 25, 6016–6024.

29

30

1 The Role of Copper Ion and the Ubiquitin System in Neurodegenerative Disorders 172. Li, J.Y., Englund, E., Holton, J.L.,

173.

174.

175. 176.

177. 178.

179.

180.

181.

Soulet, D., Hagell, P., Lees, A.J., Lashley, T., Quinn, N.P., Rehncrona, S., Bjorklund, A., Widner, H., Revesz, T., Lindvall, O., and Brundin, P. (2008) Nat. Med., 14, 501–503. Ren, P.H., Lauckner, J.E., Kachirskaia, I., Heuser, J.E., Melki, R., and Kopito, R.R. (2009) Nat. Cell Biol., 11, 219–225. Lauren, J., Gimbel, D.A., Nygaard, H.B., Gilbert, J.W., and Strittmatter, S.M. (2009) Nature, 457, 1128–1132. Cisse, M. and Mucke, L. (2009) Nature, 457, 1090–1091. Miller, L.M., Wang, Q., Telivala, T.P., Smith, R.J., Lanzirotti, A., and Miklossy, J. (2006) J. Struct. Biol., 155, 30–37. Ide-Ektessabi, A. and Rabionet, M. (2005) Anal. Sci., 21, 885–892. Cherny, R.A., Legg, J.T., McLean, C.A., Fairlie, D.P., Huang, X., Atwood, C.S., Beyreuther, K., Tanzi, R.E., Masters, C.L., and Bush, A.I. (1999) J. Biol. Chem., 274, 23223–23228. Donnelly, P.S., Xiao, Z., and Wedd, A.G. (2007) Curr. Opin. Chem. Biol., 11, 128–133. Liang, X., Campopiano, D.J., and Sadler, P.J. (2007) Chem. Soc. Rev., 36, 968–992. Lee, E.N., Lee, S.Y., Lee, D., Kim, J., and Paik, S.R. (2003) J. Neurochem., 84, 1128–1142.

182. Gai, W.P., Yuan, H.X., Li, X.Q.,

183.

184.

185. 186.

187.

188. 189.

190.

191.

Power, J.T., Blumbergs, P.C., and Jensen, P.H. (2000) Exp. Neurol., 166, 324–333. Sharon, R., Bar-Joseph, I., Frosch, M.P., Walsh, D.M., Hamilton, J.A., and Selkoe, D.J. (2003) Neuron, 37, 583–595. LaFerla, F.M., Green, K.N., and Oddo, S. (2007) Nat. Rev. Neurosci., 8, 499–509. Guarino, L.A., Smith, G., and Dong, W. (1995) Cell, 80, 301–309. Webb, J.H., Mayer, R.J., and Dixon, L.K. (1999) FEBS Lett., 444, 136–139. Welchman, R.L., Gordon, C., and Mayer, R.J. (2005) Nat. Rev. Mol. Cell Biol., 6, 599–609. Burz, D.S. and Shekhtman, A. (2009) Nature, 458, 37–38. Inomata, K., Ohno, A., Tochio, H., Isogai, S., Tenno, T., Nakase, I., Takeuchi, T., Futaki, S., Ito, Y., Hiroaki, H., and Shirakawa, M. (2009) Nature, 458, 106–109. Lange, O.F., Lakomek, N.A., Fares, C., Schroder, G.F., Walter, K.F., Becker, S., Meiler, J., Grubmuller, H., Griesinger, C., and de Groot, B.L. (2008) Science, 320, 1471–1475. Fahrni, C.J. (2007) Curr. Opin. Chem. Biol., 11, 121–127.

31

2 The Bioinorganic and Organometallic Chemistry of Copper(III) Xavi Ribas and Alicia Casitas

2.1 Introduction

Copper in oxidation state +3 is known in a limited number of compounds. In view of the electronic configuration of CuIII ion, 3d8 , the ligand-field theory predicts an enhanced stabilization in a square-planar type of geometry [1], and, indeed, this is the common geometry found in the vast majority of CuIII structures reported. Also, pairing of all d8 electrons affords diamagnetic species. Some of them are stable only in the solid state, such as NaCuIII O2 [2, 3], CuIII -bis-biuret [4], and CuIII -bis-oxamide [5], but decompose rapidly in solution; the generation of CuIII in water or acetonitrile has been achieved by means of pulse radiolytic [6, 7] or electrochemical experiments in the presence of macrocyclic amines [8, 9], but their reduction to CuII is usually very rapid. Diperiodato and ditellurato cuprates(III) have been generated in solution and studied as oxidizing agents of organic substrates [10]. However, it was not until the discovery by Margerum et al. in 1975 of CuIII –peptides complexes, stable in aqueous solutions and room temperature, that it boosted the interest in CuIII and its implications in biological systems. Examples of these are the tetrapeptide CuIII –glycine [11] and tri-α-aminoisobutyric acid tripeptide CuIII complex, [CuIII (H-3 Aib3 )] [12]. Since then, several other CuIII species have been stabilized at room temperature and in solution by using anionic ligands [13, 14]. For example, ligands containing two deprotonated oxime groups in combination with two imine N atoms stabilize CuIII complexes [15], as well as tetraanionic bis-oximato chelating ligands [16], allowing the characterization of their CuII /CuIII redox properties. Highly stable CuIII complexes with N-amido and alkoxo polyanionic chelating ligands were reported by Anson et al. [17] and studied their stability alteration by modifying the donor properties of ligands. More recently, Hanss et al. reported CuIII complexes bearing deprotonated hydroxyiminoamide ligands [18]. The comparison of CuIII /CuII redox potentials among several CuIII species in water and CH3 CN revealed the trend of enhanced stabilization of CuIII oxidation state by increasing the number of anionic donor atoms [19], and by introducing thiolate donor groups (Figure 2.1).


32

2 The Bioinorganic and Organometallic Chemistry of Copper(III) − O

O

N

N CuIII R1

R1

O O

R1=

O

O−

Me

E1/2 (CuIII/CuII) vs. SCE (mV)

N−

O

0.41

0.02

H3C H3C −0.51

O−

H3C H3C

S−

−0.78

N N H2

N CuIII O

0.10

O

O N O

N

CuIII N O OH 0.26

N

Figure 2.1 Redox potentials for well-established CuIII species bearing disubstituted o-phenylene-bis(amide) ligands (four on the left) [16], one tri-α-aminoisobutyric acid tripeptide ligand [12], and one bis(hydroxyiminoamide) ligand (on the right) [18].

Besides square-planar CuIII compounds, there are several examples reporting weak axial coordination yielding square-pyramidal geometry, and only two compounds have been reported exhibiting octahedral (Oh) geometries: the well-known paramagnetic K3 [CuIII F6 ] described in text books [1] and a more recent heterometallic compound, [LCoIII CuIII CoIII L](ClO4 )3 , with a CuIII S6 coordination environment [20]. In the latter case, the Oh CuIII species is no longer diamagnetic and the high spin paramagnetic (S = 1) species is characterized by EPR (electron paramagnetic resonance) spectroscopy. Additionally, Santo et al. have also shown a diamagnetic-to-paramagnetic conversion of a tris(2-pyridylthio)methylcopper(III) through a geometry change from trigonal bypiramidal to Oh upon addition of chloride [21]. The direct assignment of the oxidation state of the copper ion is not straightforward in many complexes. Transition metal complexes containing dithiolene ligands or bidentate ortho-disubstituted aromatic ligands bearing O, N, S, or Se as coordinating atoms [22] show a rich electrochemical behavior that opens the question whether the redox processes are ligand- or metal centered. Experimental and theoretical evidence has been reported on dearomatization of the benzene rings to give an open-shell o-semiquinonate type ligand, and the singlet diradical ground states of some neutral complexes have been characterized [23, 24]. Thus, the ligand role upon oxidation of the complex, usually referred as innocent or noninnocent character of the ligands, is still under discussion. A number of dithiolene- and diselenolene-type copper complexes in the formal oxidation state +3 are known [25–28], and several works discussing the real oxidation state of the metal center have been reported [26, 29]. In any case, the existence of an experimental technique to prove the oxidation state of copper was mandatory. DuBois et al. used synchrotron light source to set up the Cu K-edge X-ray absorption spectroscopy (XAS) technique in 2000, which provided a clear means of identifying CuIII [30]. They studied a varied number of CuIII species, from Margerum’s CuIII –peptide compounds to bis(µ-oxo)CuIII 2 species with peralkylated amine ligands; the distinct feature of CuIII is a 1 s → 3d transition centered on 8981 eV, approximately 2 eV

2.2 Bioinorganic Implications of Copper(III)

higher than the characteristic energy of this transition for CuII oxidation state. In mixtures with CuII , CuIII is identifiable if present in at least a 25%.


Copper versatility is reflected in the high diversity of biological functions that copper-dependent enzymes are capable to catalyze, and in the diverse number of Cu ions contained in each metalloenzyme active site [31]. Copper can be present as mononuclear, dinuclear, or trinuclear complexes, sharing the active site with other metals such as Fe and Zn or forming polymetallic aggregates. Among the different catalyzation processes, O2 activation and electron transfer processes are the most widely found. Enzymes eliciting monooxygenation (tyrosinase, Tyr; dopamine-β-monooxygenase, DβM; peptidylglycine α-hydroxylating monooxygenase, PHM; particulated metano monooxygenase, pMMO), dioxygenation (quercitinase and indole dioxygenase, IDO), and oxidation processes (galactose oxidase, GO; catechol oxidase, CO; laccase, Lc; cytochrome c oxidase, CcO) are widely spread in a large number of microorganisms, arthropods, and superior organisms including mammals. Electron transfer processes are catalyzed by mononuclear blue copper proteins (azurin and plastocyanin, Pc) found in plants. The participation of copper(III) species in processes catalyzed by mono-, di,and polynuclear copper metalloproteins has been often invoked [32]; despite this it has not been directly detected in natural systems so far. Research in bioinorganic chemistry and the development of low-molecular weight chemical models have, however, allowed the characterization of CuIII species that may be relevant to natural systems. It was actually after Margerum’s discovery of stable CuIII –peptide compounds in aqueous solutions that the scientific community began to think seriously in the possible biological implication of CuIII . Hereafter, we will briefly describe the most relevant biological processes where CuIII might have an important role. 2.2.1 Dinuclear Type-3 Copper Enzymes

Type-3 copper enzymes, namely hemocyanin (Hc), Tyr, and CO, are capable of binding molecular O2 to form a µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo species as the enzymatic oxidized form. Hc function is exclusively limited to O2 transport in arthropods and mollusks [33], whereas Tyr is usually found in bacteria, fungi, plants, and mammals, and is involved in the synthesis of melanins. Specifically, it catalyzes the hydroxylation of monophenols to obtain ortho-diphenols (catechols) in a first step, and reoxidizes the latter to o-quinones in a second step [34]. In dinuclear copper(I)-reduced deoxy forms, each metal center is coordinated to three histidinic-N residues, and reaction with O2 in a cooperative manner allows for the

33

34

2 The Bioinorganic and Organometallic Chemistry of Copper(III)

NHis NHis

CuII NHis

NHis O CuII N His O NHis

O2 transport (Hemocyanin)

NHis NHis

OH

CuI

CuI

NHis

NHis

NHis NHis

+ O2

R Phenolase reactivity (Tyrosinase)

R NHis NHis

O CuII NHis

O CuII O

NHis

OH OH

R

NHis

NHis

Figure 2.2 Distinct functionality of µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo species in hemocyanin and tyrosinase. An electrophilic attack of the peroxo core to the phenolic substrate is proposed to account for tyrosinase reactivity [35].

stabilization of the peroxo bridging group in a side-on geometry fashion for the oxidized form of both enzymes (Figure 2.2). Spectroscopic characterization of bioinorganic synthetic models of Hc and Tyr active sites anticipated the identical oxidized forms for both enzymes [34, 36–40], before X-ray structural data were recently available [41, 42]. The difference in reactivity is likely related to the accessibility of phenolic substrates to the active site, which is modulated by the tertiary structure of the polypeptidic chain. The latter has been also suggested by partial denaturalization of hemocyanin with urea, resulting in an active enzyme for phenol hydroxylation [43, 44]. The active species responsible for the hydroxylation of phenol in Tyr is thought to be the well-characterized µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo species, performing an electrophilic attack on the π-system of the phenol aromatic ring (Figure 2.2) [35, 45, 46]. Several synthetic compounds modeling the side-on peroxo-CuII 2 species support its capability to perform an electrophilic σ ∗ attack to a properly oriented aromatic ring [47], either in intramolecular ligand hydroxylations [48–53] or in intermolecular hydroxylation of exogenous phenolic substrates[54–58]. However, the real active species in Tyr was initially challenged in 1995 by Tolman and coworkers, who first described the existence of a bis(µ-oxo)CuIII 2 core in a synthetic model [59] and shortly after reported the coexistence of µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo species and bis(µ-oxo)CuIII 2 isomers in rapid equilibrium in solution at very low temperatures (Figure 2.3a) [60, 61]. The bis(µ-oxo)CuIII 2 core represented a step forward in the comprehension of the Cu2 –O2 interaction mechanism because the O–O bond is already broken. Since then, and considering the facile interconversion between the two isomeric species by changing solvent or counteranions [61–65], several groups have pushed the idea of the bis(µ-oxo)CuIII 2 core capability of performing aromatic hydroxylation, and thus proposing its implication in the reactivity of Tyr [66–68]. The most representative work in this line was published by Mirica et al. in 2005 [67], when


CuII

O

35

O CuII

CuIII

O

CuIII O

3.6 Å

2.8 Å

(a) +

2+ X N

N O

N

CuIII

CuIII N O

N

N

ONa N

X = Cl, F, CO2Me, CN Acetone −90 °C

OH

N

Warming up

O N

CuIII O N

(b)

N

CuIII O

N

X

Figure 2.3 (a) Equilibrium between isomeric µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo and bis(µ-oxo)CuIII 2 cores and (b) an example of exogenous phenolate hydroxylation by a bis(µ-oxo)CuIII 2 core [68].

they reported a dinuclear synthetic copper(I) complex bearing peralkylated diamine ligands that bind O2 to form a side-on peroxo species at −120 ◦ C, which evolutes to a bis(µ-oxo)CuIII 2 core upon coordination of an exogenous 2,4-di-tert-butylphenolate. Finally, the phenolic substrate is hydroxylated by the bis(µ-oxo)CuIII 2 core to afford quantitatively a mixture of the corresponding catechol and quinone. Theoretical analysis suggested an electrophilic σ ∗ attack of bis(µ-oxo)CuIII 2 core to the phenol ring. Company et al. have also made an important contribution substantiating the capability of a perfectly established bis(µ-oxo)dicopper(III) species to perform Tyr-like reactivity [68, 69]. In this synthetic model, the bis(µ-oxo)CuIII 2 core rapidly binds 4-chlorophenolate affording a new metastable intermediate species at −90 ◦ C, [CuIII 2 (µ-O)2 (p-Cl-C6 H4 O)(m-XYLMeAN )]+ , which decomposes toward the hydroxylation of the phenolate moiety to afford the corresponding 4-chlorocatechol (Figure 2.3b). Stopped-flow UV–Vis kinetic monitoring allowed to determine an electrophilic type of reactivity for a series of para-substituted phenolates, thus in line with Stack’s observations [67]. Our work contribution is centered in the fact that O–O bond cleavage can occur not only prior to C–O bond formation but also prior to phenol binding, further substantiating the idea that the bis(µ-oxo) core is competent for performing enzyme-like activity. Very recently, Herres-Pawlis et al. have reported another bis(µ-oxo)dicopper(III) species based on a permethylated-amine-guanidine ligand performing 2,4-di-tert-butylphenolate hydroxylation to 3,5-di-tert-butylcatecholate [70]. The fact that both isomeric Cu2 O2 cores are capable of performing the electrophilic aromatic hydroxylation of arene rings (either of phenolate-type or of intramolecular xylyl linkers), along with the known facile interconversion between both isomers depending on experimental conditions, suggests that either one may

X

OH

36


be the final core responsible for the aromatic hydroxylation. But, most importantly, what is really crucial is that the aromatic substrate can approach in intimate proximity and correct orientation to the Cu2 O2 core, and here seems to rely on the difference in reactivity between Tyr and hemocyanin. It has to be indicated, however, that a very recent report from Casella and coworkers rules out the existence of the bis(µ-oxo)CuIII 2 core in the natural Tyr [71], as previously suggested theoretically by Siegbahn et al. [72]. A complete discussion on reactivity of isomeric Cu2 O2 cores is found in Dr. Company’s chapter of this book is Chapter 11. 2.2.2 Particulate Methano Monooxygenase (pMMO)

Methane oxidation to methanol is performed by methanotrophic bacteria using molecular oxygen [73]. These bacteria reside at the boundary of aerobic and anaerobic environments, where CH4 and O2 are available. Two different enzymes with the same function are known to be expressed in methanotrophic bacteria such as Methylococcus capsulatus: a cytoplasmatic soluble methane monooxygenase (sMMO) and a membrane-bound known as particulate-methane monooxygenase (pMMO). The crystal structure of sMMO is known since 1993 [74], and establishes the existence of a dinuclear iron center that is responsible for the activation of O2 and reaction with CH4 to produce CH3 OH. On the other hand, it was known that pMMO contained Cu ions in its active site by means of spectroscopic studies [75], but unraveling its structure has been one of the most intriguing and exciting topics in bioinorganic chemistry. Very recently, Lieberman et al. overcame the difficulties of purifying and crystallizing a membrane-bound enzyme, and reported the crystal structure of pMMO from M. capsulatus at a resolution of 2.8 A˚ [76]. This structure showed the existence of three different subunits, with three metal centers: one mononuclear copper site, a dinuclear copper site 21 A˚ apart, and a mononuclear Zn site 19 A˚ apart from the dicopper site and 32 A˚ from the monocopper site (Figure 2.4a). The zinc site is considered to be occupied by Cu or Fe since its presence was not spectroscopically anticipated and pMMO was crystallized using a Zn buffer. Therefore, in sharp contrast to the iron mediated methane conversion to methanol in sMMO, the pMMO appears to be a copper-dependent enzyme that also hydroxylates CH4 . It has been intensively discussed in the literature whether monocopper or dicopper sites (or both) can perform O2 activation and subsequent methane oxidation to methanol [77–79]. Several highly oxidized copper-oxo species have been invoked to play a role in pMMO activity, and among them, mononuclear [CuIII =O]+ species [77] and dinuclear bis(µ-oxo)CuII CuIII core [77, 80] appear as the most reliable active species that can elicit the hydroxylation of methane (Figure 2.4b). Yoshisawa et al. evaluated computationally the formation of [CuIII =O]+ and bis (µ-oxo)CuII CuIII species by reproducing the biologic coordination environment certified by the X-ray structure [77]. According to their DFT (density functional


His 48

HN N

HN

N

Cu

Cu

N

NH

His 33

NH

C NH2 Gln 404

HN

N

HN His 137

NH

CuI

His 139

NH

O

HN

His 137 N

N

His 173

O N H

HN

His 33

His 33

HN His 137

NH N

CuIII O His 139 CuIII N NH

O

O His 139

CuII N NH

O C

Mononuclear metal site

CuII

O2

CuI His 139 NH

His 33

Glu 195

N

N

HN (b)

N H

O

Zn

N O

O

N

HN

Cu

Dinuclear copper site

N

N

O C

Glu 35

Monocopper site

HN His 137

Asp 156

NH N

O

(a)

His 160

His 137

His 72

37

N e−

CuIII O His 139 CuII N NH

O HN

NH N

His 33

Figure 2.4 (a) Three different copper sites identified in the pMMO crystal structure of M. capsulatus. (b) Proposal for O2 activation by the dinuclear copper site to obtain the most likely active bis(µ-oxo)CuII CuIII core.

theory) calculations, the mononuclear CuI center can activate O2 to form a CuII -superoxo species, followed by two consecutive H atom transfer from near tyrosine residues to generate a [CuIII =O]+ species and H2 O. This process was computed to be 17.8 kcal mol−1 endothermic, although was still considered possible under physiological conditions. On the other hand, the activation of O2 by the dinuclear CuI center affords a µ-η2 : η2 -(O2 )-CuII 2 side-on peroxo species that interconverts to a bis(µ-oxo)CuIII 2 isomer. One electron reduction of the latter leads to the formation of the dinuclear bis(µ-oxo)CuII CuIII species, an overall process computed to be 59.3 kcal mol−1 exothermic, thus much more favorable than the mononuclear [CuIII =O]+ . The latter result suggests that the dinuclear copper site might be the real executor of the pMMO catalytic activity. Indeed, the computed reaction path of methane hydroxylation by bis(µ-oxo)CuII CuIII (doublet ground state) shows a methane molecule weakly bound to the CuIII center (Figure 2.5), followed by H atom abstraction by a bridging oxo group (barrier 16.1 kcal mol−1 ) to form a nonradical intermediate, and final recombination of OH and CH3 ligands coordinated to the two distinct copper centers (rate determining barrier of 24.3 kcal mol−1 ). It is interesting to note that the computed mechanistic proposal of these authors shows the participation of organometallic CuIII –CH3 moieties in the reaction path of both [CuIII =O]+ and bis(µ-oxo)CuII CuIII species with methane. This mechanistic proposal is consistent with experimental hydroxylation of chiral ethane [81] and butane [82] by pMMO from M. capsulatus, since the alcohol products are obtained with 100% retention of configuration. The rate determining recombination step is, however, not consistent with the experimental H/D kinetic isotope effect (KIE = 5.2–5.5, 30 ◦ C for ethane hydroxylation) [81], although the authors

38


OH CH3 CH3

CuII

H O CuII

CuIII

O 16.1

TS2

TS1

1.9

−0.7

CH4 CuII

O O

CuIII O 24.3

OH

CuIII CuII

−43.6 CH3 CuIII

O

HO CuI

CH3 CuII

O

Figure 2.5 DFT energy diagram for the hydroxylation of methane by bis(µ-oxo)CuII CuIII species (energies in kilocalories per mole) [79].

believe that the protein environment can significantly stabilize the transition state for OH migrations. Full consideration of protein environment is computationally demanding, but necessary to address the viability of the mechanism proposed via bis(µ-oxo)CuII CuIII species from a computational point of view. Balasubramanian et al. also indicate that the coordination of the carboxylate moiety from Glu35 may not be correctly considered in these computational studies, since it is found 4 A˚ apart from the closer Cu center [78]. The calculated activation energies for the reaction of bis(µ-oxo)CuIII CuIII species with CH4 were found high enough to clearly discard this species as playing a role in pMMO activity [79]. In a recent report, Hakemian et al. have crystallized the pMMO enzyme from Methylosinus trichosporium OB3b [83], and found no copper in the mononuclear site, thus suggesting its nonparticipation in the reactivity, in line with theoretical proposals of bis(µ-oxo)CuII CuIII species as active species. 2.2.3 Mononuclear Monooxygenating Copper-based Enzymes

Peptidylglycine α-hydroxylating monooxygenase (PHM) and dopamine β-monooxygenase (DβM) are copper-dependent enzymes found in higher eukaryotes that catalyze the regio and stereospecific α-hydroxylation of a C-terminal glycine and the hydroxylation of dopamine to norepinephrine, respectively (Figure 2.6) [84]. The active sites of PHM and DβM are almost identical, and both consist of two inequivalent mononuclear Cu centers (CuA and CuB ) far apart in space (11 A˚ in PHM) [85] with no direct bridging ligands and no observable magnetic interactions between metals (Figure 2.7a). The crystal structure of PHM indicates that the copper atom is coordinated by two histidine and one methionine ligands at the CuB site (where O2 coordinates and substrate hydroxylation occurs), whereas copper is coordinated to three histidine ligands at

2.2 Bioinorganic Implications of Copper(III) O

O HO H

2e−, 2H+ COO−

N H

Peptide

H 2O

H OH

2e−, 2H+

NH2

NH2

DbM O2

(b)

Figure 2.6

COO−

N H

Peptide

PHM O2

(a)

39

H2O

Global aliphatic hydroxylation reactions catalyzed by (a) PHM and (b) DβM. NHis NHis

CuAI R H

SMet

O O CuBII NHis

(a)

(b)

S

CuBIII N

NH N

NH

R

NHis

CuAI

R NHis

H

NHis

OH SMet

NHis

NHis

H 3C H3C

H 2N

OH OH

H2N ∆E = 3.8 kcal mol−1

O CuBII NHis

H abstraction

O H3C

NHis

OH OH

H2N H3C

11 Å

R H

OH II S CuB N

OH OH

OH O rebinding

NH N NH

∆E = 6.5 kcal mol−1

II H3C S CuB H3C N

Figure 2.7 (a) Mononuclear CuII B -superoxo proposal as active species responsible for aliphatic H atom abstraction in both natural DβM and PHM enzymes (mononuclear CuA lies 11 A˚ apart from CuB site). (b) Theoretical QM/MM proposal of a CuIII B -oxo as the most active species in DβM [88b].

the CuA site (which provides an additional electron through long-range electron transfer (ET) to the CuB site) [85, 86]. The structural similarities and parallel H/D KIEs strongly suggest that both enzymes perform the C–H abstraction reaction via very similar molecular mechanisms. Based on the available experimental data, Klinman tested four mechanistic hypotheses involving different active species [84]: a one-electron reduced intermediate (copper-superoxo, CuB II (O−· 2 )), two-electron reduced species II − (copper-peroxo, CuB II (O2− 2 ) or a copper-hydroperoxo, CuB (OOH )), and, finally, III 2− a highly reduced copper-oxo CuB (O ) formed via the reductive cleavage of CuB II (OOH− ) by a proximal tyrosine. Based on the large amount of experimental data for PHM and DβM reactivity, the author concluded that a copper-superoxo intermediate is the most consistent alternative as active species toward H atom abstraction from the substrate (Figure 2.7a). A recent synthetic model supports the copper(II)-superoxo alternative [87].

NH N NH

40


There are, however, two recent computational works that challenge the CuB II -superoxo proposal as the active species [88, 89]. Yoshisawa et al. carried out calculations for a full enzyme Quantum Mechanics-Molecular Mechanics (QM–MM model) obtained by using the crystal structure information from DβM and PHM. They evaluated the reactivity of copper-superoxo, -hydroperoxo, and -oxo species in DβM. The activation energies for the H abstraction by the CuII -superoxo moiety were found to be 23.1 kcal mol, thus suggesting that the CuB II -superoxo species can promote the H abstraction in DβM. The CuB II -hydroperoxo species was ruled out as an active oxidant, because the activation barrier of that H abstraction was found to be more than 40 kcal mol−1 . Finally, they concluded that the CuB III -oxo species reasonably promotes the H abstraction because the activation energy was found to be only 5.4 kcal mol−1 , appearing as the most active species within the protein active site environment (Figure 2.7b) [88]. The hydrogen-bonding network between dopamine and three amino acid residues (Glu268, Glu369, and Tyr494) was found to play an essential role in substrate binding and the stereospecific hydroxylation of dopamine to norepinephrine. A second computational report on PHM reactivity also supports a CuIII -oxo as the active species for H atom abstraction [89]. Potential energy profiles computed for the H abstraction reaction performed by copper-superoxo species, CuB II (O−· 2 ), suggest that it is not the active species due to the high activation energy found. Inspection of the potential energy surfaces shows that the CuB II -hydroperoxo species is spontaneously formed by abstracting a proton from the surrounding solvent, but cannot abstract the hydrogen atom from the substrate. However, this species can spontaneously abstract another proton from the surrounding III +2 solvent to form [L·+ ([CuB III O]+ moiety with two unpaired electrons 3 CuB O] ferromagnetically coupled with L·+ 3 , in the quartet spin ground state), along with the release of a water molecule. Despite enormous efforts devoted to investigate DβM and PHM reactivity, there is no definitive consensus on the mechanism and/or the identity of the active species responsible for H abstraction. Bioinorganic models usually help in the understanding of mechanistic issues of natural systems. In the context of mononuclear monooxygenase enzymes, Hong et al. reported a mononuclear CuI -α-ketocarboxylate complex bearing a bulky bidentate ligand that upon activation of O2 undergoes an intramolecular hydroxylation of an arene moiety [90]. Although not experimentally directly detected, theoretical data pointed toward a CuIII -oxo species as the active species for this aromatic hydroxylation. Experimental isolation of these active intermediates to prove their oxidizing power toward exogenous substrates is a desired target for bioinorganic chemists, which may help in the interpretation of intimate mechanistic details in enzymes. 2.2.4 Trinuclear Copper Models for Laccase

Lc, ceruloplasmin (Cp), and ascorbate oxidase (AO) are the representative enzymes of type (2+3) trimer, and are involved in the lignin formation (plants)

2.3 Organometallic CuIII Species in Organic Transformations

3 [LCuI(CH3CN)]+

L=

NMe2

−80 °C, CH2Cl2 O2 3 CH CN 3

[L3Cu3O2]3+ ON

N II

III

Cu Cu Cu NMe2

N

O N

N II

N

Figure 2.8 Synthetic functional model for the O2 activation by a trinuclear copper center forming a (bis-µ3 -oxo)[CuII CuII CuIII ] species [93].

and degradation (fungi), among others [31c, 34]. They are responsible for the 4e− reduction of O2 to two H2 O molecules by using the electrons coming from one-electron oxidation of four substrate molecules. The active sites of these enzymes contain a trinuclear CuI active site (capable of O2 reduction) [91], with an additional mononuclear ‘‘blue’’ CuI site at 12 A˚ far apart (as a proton transfer site to inject one electron to the trinuclear site) [92]. A genuine functional model of these trinuclear centers is a monocopper(I) complex with N-permethylated (1R,2R)-cyclohexanediamine ligand, which reacts with O2 and generates a (bis-µ3 -oxo)[CuII CuII CuIII ] (Figure 2.8) [93]. The spectroscopic properties of this species, and especially, the bands at 290 and 355 nm assigned to oxo-to-CuIII ligand-to-metal charge transfer (LMCT), do not match the spectroscopic properties of the oxidized forms of the enzymes. The latter was interpreted in the sense that the fourth electron to reduce O2 to two O2− comes from the ‘‘blue’’ CuI site oxidation to CuII in the natural system, whereas from one CuII center to generate one CuIII in the synthetic model. Therefore, although CuIII cations probably do not play a role in the enzymatic reactivity, full characterization of the synthetic model has been crucial for the understanding of the natural systems.


The implication of organometallic CuIII –C species in different organic transformations catalyzed by copper reagents has been proposed for a long time since 1941, when Kharash discovered [94] the reactivity of organocuprates (R2 CuLi, for instance) [95] toward nucleophilic alkyl-, vinyl, and aryl anion delivery in regio and stereoselective C–C bond forming processes as conjugate additions [96], carbocuprations [97], SN 2 alkylations [98], and SN 2 allylations [99]. General consensus exists on the formation of R–CuIII –R’ species followed by subsequent reductive elimination to account for organocuprate reactivity [95, 100]. Indeed, the more than one century-old Ullmann chemistry [101–103], referred in general terms to all aromatic nucleophilic substitution reactions in

41

42


the presence of copper(I) salts to form new Caryl –Caryl , Caryl –O, Caryl –N, or Caryl –S bonds, is thought to proceed by participation of high-valent organometallic aryl-CuIII –Nucleophile, which undergoes reductive elimination to form the cross-coupling aryl–Nucleophile product [104, 105]. From a general perspective, the accessibility of four oxidation states for copper (from 0 to +3) is a distinctive characteristic that governs its chemistry and differentiates it from Pd reactivity, which basically functions with the Pd0 /PdII redox pair (although several systems working with PdII /PdIV redox pair have recently appeared) [106, 107]. In this section, the most intriguing implications of reported CuIII organometallic species in catalytic or stoichiometric organic transformations involving CuI and CuII reagents are described. 2.3.1 C–C Bond Formation in Organocuprate(I) Catalysis

Four basic types of organocopper are known: RCuI , R2 CuI M (M = Li, MgX), R3 CuI Li2 , and R3 CuIII . The polymeric nature of RCuI renders unreactive species, whether R2 CuI M and R3 CuI Li2 species generated in situ are the most useful carbon nucleophile reagents. The exact structural nature of the latter in solution remains a matter of debate [95]; however, it is clear that lithium cations play a role in the stabilization and reactivity of the intermediate species. On the other hand, R3 CuIII species are proposed in numerous mechanistic discussions, although there are only three structurally characterized examples of di- and tetraalkyl CuIII species: [CuIII (CF3 )2 (SC(S)NEt2 )] [108], [CuIII (CF3 )4 ]- [109], and [CuIII (CF2 H)4 ]- [110]. Their square-planar geometry can be regarded as a planar T-shape geometry with a fourth ligand to enhance stabilization. 2.3.1.1 Conjugate Addition to α-Enones Conjugated additions to α, β-unsaturated ketones and esters are among the most important reactions of organocuprates. Based on NMR spectroscopic studies and theoretical calculations, the reaction of R2 CuI Li with enones is directed by Li-carbonyl and copper-olefin interactions [111]. The key intermediate is a trialkyl-CuIII species that forms upon two-electron inner-sphere electron transfer (Figure 2.9). The organocopper R3 CuIII species then undergoes reductive elimination to afford the conjugate 1,4-addition product. Dorigo et al. already reported back in 1995 the theoretical stability of (CH3 )3 CuIII moieties further stabilized with one H2 O or NH3 molecule to complete a square-planar geometry [112]. Indeed, Snyder also predicted the possibility to isolate and even structurally characterize these CuIII species [113]. Very recently, Bertz et al. have used rapid injection nuclear magnetic resonance (RI-NMR) techniques at low temperatures (−100 ◦ C) to unequivocally characterize the largely sought and elusive trialkyl-CuIII intermediate upon reaction of 2-cyclohexenone and Me2 CuI Li


R1 CuI R1 Li Li R1 CuI R1

O R2

OLi R1 CuI R1 O Li Li R1 CuI R1

Reductive elimination

2

R

R1

OLi R2

CuIII

R12CuILi

43

R1

R1

+ R1CuI

Proposed mechanism for 1,4-conjugate addition of R2 CuI Li to α-enones.

Figure 2.9

Me Me Si Me

1. Inject TMSCI

Me2CuLi LiCN (−100 °C) 2. Inject O

O

H

Warming Li+ to −80 °C N− C Reductive CuIII Me elimination Me

OTMS + MeCuICNLi Me

NMR characterized Figure 2.10 Low-temperature RI-NMR characterization of an R3 CuIII intermediate species in an organocuprate conjugate addition reaction [100].

Li

X

Li I

R Cu R + H

H

Figure 2.11

“Trap” (oxidative addition) X = RCuR, halogen

Li

X

R Cu H

Li R “Bite” (reductive X Li CuI R R elimination)

Li III

H

H

Transmetalation

H

R Li

CuI

X Li H

Proposed trap-and-bite mechanism of acetylene carbocupration [95].

(Figure 2.10) [100], also supported by DFT calculations [114]. Warming of the CuIII intermediate up to −80 ◦ C afforded the conjugate 1,4-addition enolate. 2.3.1.2 Acetylene Carbocupration Synthesis of alkenyl derivatives can be achieved in a stereoselective manner by reaction of organocuprate(I) reagents with acetylene, a process that involves the intermediacy of a R2 CuIII (acetylide) species in a trap-and-bite mechanistic proposal (Figure 2.11) [95, 97]. These reactions are usually performed at room temperature or below. 2.3.1.3 SN 2 and SN 2 Alkylations The alkylation of alkyl halides (and epoxides) with organocuprate(I) reagents (Me2 CuI Li) occurs through a SN 2 mechanism in which the C–X (X = halide) bond is substituted by an incoming R–Cu σ bond to generate unstable Me2 CuIII R intermediates. The latter undergoes reductive elimination to afford the cross-coupled Me–R product. This mechanistic proposal is fully supported by the recent RI-NMR characterization of several CuIII tetracoordinate alkyl species upon reaction of ethyl iodide (EtI) with a variety of methyl Gilman reagents, Me2 CuLi·LiX (X = I, CN,

R H


44

CH3CH2I

Me2CuILi LiX (−100 °C)

X = I, CN, SCN, SPh

Me Et CuIII X Me

Li

Me2CuILi LiCN MeCuIXLi LiCN

Lil

Reductive Me elimination III CH3CH2CH3 Et Cu Me Li Me I Me2Cu Li

(a)

3 Me2CuLi (−100 °C) (b)

2

Cl

Me CuIII Me

MeCuIClLi + LiCI

+

Reductive Me elimination 2 CuIII Me Li Me Me2CuILi + MeCuI

CH3

Figure 2.12 Low-temperature RI-NMR characterization of (a) the first nonfluorinated R4 CuIII intermediate species in an SN 2 alkylation [115] and (b) η1 σ -allyl and η3 π -allyl-CuIII species [117].

SCN, SPh) [115, 116]. Among them, [CuIII Me3 Et]Li is the most stable and also the first example of a tetraalkylcopper(III) without fluorinated alkyl substituents (Figure 2.12a) [109, 110]. Similar reactivity is observed if allylic halides are used, but usually a mixture of SN 2 and SN 2 products is obtained. This observation suggested the existence of a regioisomeric σ -allyl-CuIII species, which has been certified very recently by means of RI-NMR with the identification of both η1 σ -allyl and η3 π-allyl-CuIII complexes upon reaction of allyl chloride and Me2 CuI Li·LiI (Figure 2.12b) [117, 118]. The σ complex is formed initially but transforms rapidly to the more stable π-allyl-CuIII , which undergoes reductive elimination to form 1-butene upon warming (the SN 2 product). Depending on the allylic halide nature and reaction conditions, different ratios of SN 2 and SN 2 products are obtained. 2.3.2 Aryl–Heteroatom Bond Formation in Cu-mediated Cross-coupling Processes

Impressive development of palladium-catalyzed methods for C–heteroatom and C–C bond formation [106] has obscured the older and classical Ullmann chemistry of copper-catalyzed C–N, C–O, and C–S bond formation reactions [101, 103, 104]. However, some of the deficiencies related to the use of Pd reagents such as their cost, high toxicity, use of sophisticated ligands, and low tolerance to reactive functional groups have prompted the revitalization of research on copper-based cross-coupling reactions between haloarenes and a wide range of N-nucleophiles (i.e., aromatic [119] and aliphatic [120] amines, imidazoles [121], and amides [122]), O-nucleophiles (phenols [123] and aliphatic alcohols [124]), and S-nucleophiles (aryl and alkyl thiols [125]). Equivalent methods for the synthesis of aryl-selenides [126] and aryl-phosphines [127] are also known. Examples of all these cross-coupling reactions exist with the mediation of copper(I) salts (CuI X, X = OAc, Br, I) (the use of a soft base and a ligand auxiliary additive is often required). The proposed mechanistic pathway consists in the oxidative addition of CuI to the haloarene to form an aryl-CuIII species, followed by coordination of the nucleophile source


45

Y X

CuIX

X

X = OAc, Br, I, Cl oxidative addition

L

Figure 2.13

Y = N, O, S CuIII X

X L

X−

CuIII Y


Cross-coupling of haloarenes and nucleophiles mediated by CuI salts.

and subsequent reductive elimination, to finally afford the arylation product (Figure 2.13) [104]. An impressive number of parallel methods have been reported by using organometallic-aryl compounds instead of haloarenes [104, 128], such as arylboronic acids [129], triaryl-bismuth(V)-diacetate (i.e., Ph3 Bi(OAc)2 ) [130], aryl-lead(IV)-triacetates [131], arylsiloxanes (i.e., PhSi(OMe)3 ) [132], and arylstannanes (i.e., PhSnMe3 ) [133], among others. These methods permit the obtention of the same O-, N-, and S-arylation products under significantly milder experimental conditions and shorter reaction times. Mechanistic proposals include initial transmetallation with the copper salt and suggest also the intermediacy of aryl-CuIII species prior to cross-coupling product formation for those methodologies using CuII salts (CuII (OAc)2 , CuII (OTf)2 , CuII Cl2 , CuII SO4 ·5H2 O) [104, 105]. In the latter, reactions require O2 atmosphere, what suggests the oxidation of aryl-CuII to aryl-CuIII in order to facilitate the subsequent reductive elimination step. 2.3.3 Aromatic and Aliphatic C–H Bond Organometallic Functionalizations

The transition metal activation of an inert alkane C–H bond is a challenging task [134], thus coupling a C–H bond activation with a subsequent functionalization step in a catalytic manner is even more complicated [107]. Enzymes can efficiently perform oxygenase processes through metal-oxo, -peroxo, or -superoxo species as active species [38–40, 84, 135], and it is in general believed that these catalyses do not occur through the intermediacy of M–C bonds. Hereafter, systems where a transition metal complex regioselectively cleaves a C–H bond to form organometallic M–C species that undergo functionalization via reductive elimination, provided a nucleophile is present, are discussed. 2.3.3.1 Catalytic Systems The well-known Shilov system is based on PtII /PtIV chloride complexes in aqueous solution capable of catalytic methane functionalization to methanol at 120 ◦ C [136]. Mechanistically, a PtII –CH3 species is initially formed, followed by oxidation to PtIV and final reductive elimination in the presence of H2 O to obtain CH3 OH and regenerate the PtII salt to restart the catalytic cycle. Some palladium-based systems that catalytically functionalize aromatic and aliphatic C–H bonds achieve their activation by means of the coordination-directed metallation strategy [134], where a functional group of the substrate coordinates

Y

+ CuIX

46


R1 (a)

But

( ) R2 1-2 FG H MLn

−H+

R1

( ) R2 1-2 MLn FG

( ) R2 1-2 FG X

X = C, O, N, Halide

O

Pd(OAc)2 (10 mol%)

But

Phl(OAc)2 (1 equiv.) I2 (1 equiv.) CH2Cl2 24 °C, 13 h

N

(b) R-X

L2PdII

Reductive elimination X L2PdIV X

X X

R-H

O N I

C-H bond cleavage HX

X L2PdII

R

R = alkyl, aryl X = OAc, OR, X, OH (c)

R1

X R

Oxidation X2

Figure 2.14 (a) Coordination-directed metallation strategy for C–H functionalizations; (b) example of arene halogenation mediated by PdII species; and (c) catalytic cycle proposed for PdII systems.

the metal in such a way that it is well positioned to regioselectively interact with a proximal C–H bond (Figure 2.14a). By using PdII reagents, PdII –C species are generated, followed by the subsequent functionalization step, which is facilitated by previously resorting to the oxidation of metallated PdII to PdIV species with the concomitant coordination of selected nucleophiles. This is an effective strategy to lower the barrier for the final reductive elimination step (Figure 2.14c) [107]. Following this methodology, broad scope halogenation of unactivated aromatic C–H bonds [137] (Figure 2.14b) and oxygenation of aliphatic sp3 C–H bonds [138] have been reported. A growing number of examples of C–H bond functionalization using CuII salts and involving a coordination-directed metallation have been reported recently [139–141], although the intimate mechanistic details remain quite obscure. For example, Chen et al. reported the ortho-acetoxylation and chlorination of aryl C–H bonds in phenyl-pyridine substrates mediated by CuII (OAc)2 and CuII Cl2 , respectively [139]. The coordination of CuII center to the pyridinic N accounts for the ortho-regioselectivity of the C–H activation step (Figure 2.15). A single electron transfer (SET) from the arene to reduce the metal to CuI , followed by the nucleophile insertion [140], proton loss, and a second SET mechanistic pathway,


N

X H

X = H, Me, F, Cl, CF3, OMe, COOMe, CHO, CH=CH2

Cu(OAc)2, O2 Nucleophiles (Nu−)

N

X

Solvent, 130 °C (40–90% yield)

Nu Nu = OH, OAc, p-CN-PhO, Cl, Br, I, PhS, MeS, TsNH, CN

Figure 2.15 Coordination-directed metallation strategy for the functionalization of arenes with a wide range of nucleophiles [139].

is proposed based on a nonobservable KIE; interestingly, the formation of a cyclometallated aryl-CuII species followed by reductive elimination to product and Cu0 is not discarded. These reactions are conducted under O2 , which is believed to reoxidate low-valent copper species to regenerate the CuII catalyst. Other recent CuII -mediated C–H bond catalytic functionalizations include the synthesis of benzimidazoles from amidines (C–N bond formation) [142] and of 2-arylbenzoxazoles from benzanilides via an intramolecular C–O bond formation [143]. In neither of the previous reports, a clear mechanistic picture is presented. 2.3.3.2 Stoichiometric Systems Although stoichiometric C–H bond functionalizations have minor synthetic impact compared to catalytic ones, they may allow for a more detailed mechanistic investigation of the overall process, and, in selected cases, important intermediates may be trapped. Ribas et al. have developed discrete systems based on triazamacrocyclic ligands (Ltriazamac ) that exhibit aromatic C–H bond activation upon complexation with CuII salts at exceedingly mild conditions [144]. The outcome of the reaction is the formation of equimolar amounts of stable, diamagnetic aryl-CuIII and CuI species (Figure 2.16). Full characterization of the metallated aryl-CuIII product by means of X-ray diffraction, NMR spectroscopy, cyclic voltammetry, and Cu K-edge XAS confirmed the high oxidation state of the metal center. Crystal structures showed ˚ and XAS spectra short CuIII –Caryl bond distances ranging from 1.85 to 1.90 A, exhibited a very intense transition at 8981 ± 0.5 eV, which has been demonstrated to be unequivocally indicative of the 1 s → 3d transition of CuIII compounds [30]. The singlet multiplicity for a d8 metal in a square-planar geometry agrees with the possibility of full NMR assignment of the diamagnetic compounds. The scope of this reaction was expanded to a whole family of triazamacrocyclic ligands [145], allowing us to understand the influence of steric and electronic factors in the properties of the aryl-CuIII complexes. It was found that the redox potential for the CuIII /CuII couple depends, as expected, on the electronic nature of the para-phenyl substituents as well as on the nature of the central macrocyclic amine (E1/2 (CuIII /CuII ) in the range between −201 and +6 mV versus Saturated Calomel Electrode(SCE)). Electron-withdrawing substituents in the phenyl ring destabilize the oxidation state +3, whereas electron donors produce the opposite

47


48

R1

HN

Ha NH CuII N

R1

2+

CH3CN 25 °C, < 1 min

0.5 HN CuIII

R

R2 [CuII(Ltriazamac)]2+ R1 = H, CH3, NO2 n = 0,1

2

[aryl-CuIII]2+ + 0.5 [CuI(CH3CN)4]+

R2 = H, CH3

C1

NH n

N

n

2+

1.901 N1

CuIII

N3

1.952

1.960

2.002 N2 (R1 = H; R2 = H, n = 1)

+ 0.5

(LtriazamacH)+

Figure 2.16 Aromatic C–H bond cleavage involving CuII disproportionation to afford aryl-CuIII and CuI species in a family of triazamacrocyclic ligands (crystal structure of one aryl-CuIII species included; H atoms omitted for clarity) [144, 145].

effect. Tertiary amines have a lower σ -donating capacity than secondary amines owing to a lower overlap capacity and also to solvation effects (N-methylation effect) [146], and this was manifested in a cathodic shift of the E1/2 redox potential. The isolation of stable aryl-CuIII species is important from three points of view. The first one is that no other example of structurally characterized aryl-CuIII species has been reported in the literature, and that these species can be considered as intermediate models of aryl–heteroatom cross-coupling reactions. The second one is that C–H activation is achieved very rapidly (seconds) and at room temperature, thus the mechanistic pathway by which this overall CuII disproportionation reaction takes place may provide some important clues to explain its fast reactivity [147], Finally, the third remarkable issue is the fact that we can make use of these aryl-CuIII species to evaluate their reactivity with nucleophiles as models of the aryl–heteroatom cross-coupling reactivity via a reductive elimination final stage. Huffman et al. have used three of these aryl-CuIII complexes to prove their ability for C–N coupling reactions with amide-type nucleophiles (Figure 2.17) [148]. The reactions occurred in a facile manner at room temperature to produce quantitatively the C–N coupled organic compound and [CuI (CH3 CN)4 ]+ , the expected products of a reductive elimination. The decay of aryl-CuIII features monitored by UV–Vis and 1 H-NMR indicate pseudo-first-order kinetics under excess of amide. They found that electron-withdrawing groups (−NO2 ) in the para position of the aromatic ring with respect to the ipso-carbon accelerate the reaction, whereas electron-donating groups disfavor it, thus clearly suggesting that the more electrophilic aryl-CuIII species react faster with a given N-nucleophilic substrate. On the other hand, the secondary amides exhibiting a higher pKa reacted faster, and a linear correlation between the logarithm of kobs and pKa (acidity) was found. Although the precise


R1

49

2+ H N O

HN CuIII NH

+

N

HN

CH3CN

O N R1

N

+ [CuI(CH3CN)4]+

25 °C, < 0.5 hours (And other amides)

N H

R1 = H, CH3, NO2 Figure 2.17 Facile C–N bond formation in aryl-amide coupling via reductive elimination of well-defined, isolated aryl-CuIII species [148].

mechanism of the reductive elimination step is not clear since no intermediate species are detected upon addition of the amide nucleophile, it seems clear that amide deprotonation is key for the reaction to occur. We are currently exploring the reactivity of aryl-CuIII species with O-nucleophiles to generate the aryl-O coupled products. Preliminary experiments confirm the possibility to obtain the products of reductive elimination upon reaction with aromatic and aliphatic alcohols [149]. Furthermore, the reactivity with hydroxide affords dinuclear bis-phenoxo-CuII 2 compounds in moderate yields [144, 150, 151]. We reported a few years ago the intramolecular aromatic hydroxylation of a 12-membered Ltriazamac type of ligand (named as H22m) upon reaction of its CuI complex with molecular O2 [152]. We have now evidence regarding the intermediacy of aryl-CuIII species in that process, although not isolable in this particular system due to the reduced size of the macrocyclic cavity. When H22m is reacted with CuII salts in the presence of water, the same bis-phenoxo-CuII 2 compound is obtained (Figure 2.18). Our reasoning to justify the isolation of final CuII species is that the thermodynamic stability of the final dinuclear bis-phenoxo species may be the driving force to oxidize the CuI species and recombine with the aryl-OH organic compounds, the original products of reductive elimination. The study of the reactivity of the corresponding CuI complexes, bearing the same Ltriazamac ligands, allowed us to report for the first time a family of CuI complexes that undergo regiospecific isotopic H/D exchange in D6-acetone, which represents an example of aromatic C–H activation by a CuI center (Figure 2.19) [153]. The reaction is reversible, quantitative (100% deuteration was found for the aromatic C–Ha bond, from minutes to hours depending on the ligand), and solvent dependent (occurs in D6-acetone, not in CD3 CN or CD2 Cl2 ). Our mechanistic proposal to explain the H/D exchange highlights the formation of a transient oxidative addition aryl-CuIII –hydride intermediate, followed by the H/D exchange assisted by the close proximity of a D6-acetone in its enolic form (Figure 2.19b). The observed reductive elimination CuI and Ltriazamac products by the reaction of aryl-CuIII species with hydride as nucleophile supports the H/D exchange mechanism. Noteworthy, the isolation of monodeuterated ligands at

50


+ CuI ; O2, CH3CN, 298 K

N1 II N3 Cu

2+

H2O NH

O1′

+ CuII ; CH3CN, H2O, 298 K HN CuIII NH N H

HN N H

N2

O1

N2′ N1′

Proposed aryl-CuIII intermediate

CuII N3′

Figure 2.18 Aromatic hydroxylation involving CuII disproportionation to afford a transient aryl-CuIII , which reacts with a nucleophilic water molecule to finally isolate bis-phenoxo-Cu2 II species (crystal structure ellipsoid diagram included; H atoms omitted for clarity) [152].

R1

R1 CD3CN

R1-H33m + CuI(CH3CN)4PF6

HN

Ha CuI NH

(a)

D

(b)

CD2

D3C

O

O H

CuI

N H

H, CH3, NO2

CD2

D3C

D CuI NH

HN CH3COCH3

N H R1:

CD3COCD3

R1

D

Oxidative addition

O

H

CuIII

CD3

D2C

1

H

R

H/D exchange

CD2

D 3C H

D R

CuIII

1

O D CuI

R1


Figure 2.19 (a) C–H activation in triazamacrocyclic ligands by CuI : isotopic H/D exchange at room temperature in D6-acetone. (b) Mechanistic proposal for H/D exchange reaction [153].

the C–Ha position will be an important tool to unravel important aspects of the reactivity of C–H bond activations by CuII within these systems. Up to this point, the work completes an overview on the different reactivity of these copper-containing systems depending on its oxidation state (CuII or CuI reactant species), but in either case the role of aryl-CuIII species has been demonstrated. The fundamental investigation on the mechanistic details of these intramolecular functionalization reactions within triazamacrocyclic systems will help in the development of catalytic versions of C–H bond activation processes, and may give important clues in the understanding of relevant CuII -catalyzed acetoxylation, halogenation, and arylation of aryl C–H bonds, among others [139].

2.5 Overview and Future Targets

Other reported stoichiometric C–H bond activation processes mediated by CuII species are the reaction of doubly-N-confused porphyrins (NCPs) with CuII (OAc)2 at room temperature to yield the NCP–CuIII species quantitatively in few hours [154], as well as a recent work from Ban et al. [155] that report the multiple C–H arylation of indoles and pyrroles mediated by stoichiometric amounts of CuII (O2 CCF3 )2 upon heating to 80 ◦ C for 24 hours. The latter reactivity is intriguing since it does not follow the coordination-directed functionalization strategy, thus further mechanistic investigation is required.

2.4 Miscellany: Cuprate Superconducting Materials

Owing to their relevance in potential technological applications, the inclusion of the CuIII role in high-temperature superconducting (HTSC) cuprates warrants their mention in this chapter, despite the still unclear participation of this oxidation state in the superconducting free electron flow. Cuprates as the well-known perovskite-type YBa2 Cu3 O7 (YBCO) or La0.85 Sr0.15 CuO4 present copper-oxide layers where the metal oxidation state is formally a mixture of CuII and CuIII [156]. Although it has been claimed that the holes may reside on the 2p-oxygen atom orbitals [157], many studies support the partial oxidation of CuII to CuIII , thus involving the unoccupied 3d Cu orbitals [158, 159]. Cu K-edge XAS and spectroscopic titration with p-anisidine have been reported to quantify the CuIII content in superconducting cuprate materials. In any case, the CuIII role in the origin of cuprate superconductivity remains unclear since their discovery 20 years ago [160]. The recently reported high-temperature iron-arsenic superconductors have rebursted the field [161] and may help in the understanding of the superconducting phenomena for HTSC, which falls beyond the Bardeen–Cooper–Schrieffer theory that can explain superconductivity only at low temperatures (100 keV) are well known and range from a simple sunburn to alterations of the genome, which might lead to cancer or long-term genetic diseases. Experimental evidence accumulates indicating that the damage caused by high-energy radiation can to a large extent be ascribed to a series of low-energy events [1–3]. In a first approximation, this can be rationalized when considering the typical dimensions of DNA that is assumed to be the most sensitive part of a cell, since an alteration of the genome might lead to the loss of genetic information. DNA is packed into the cell nucleus and surrounded by proteins to form a highly viscous polymer gel with high density [4]. The DNA double helix is wrapped around an octamer of histone proteins making up a nucleosome containing 146 bp [5]. The nucleosomes are further packed into a 30-nm fiber constituting the chromatin [6], which condenses into the chromosome during cell division. The mass of chromatine comprises twice as much protein than DNA, and in total the DNA covers only about 1% of the volume of a cell. Hence, the probability that high-energy radiation directly deposits energy into the DNA molecule is very low, and most of the damage is rather due to secondary particles generated along the radiation track, which are able to travel close enough to the DNA to induce damage. On a molecular level, radiation damage comprises a series of complex physical, chemical, and biological events that are only partially understood. Some of the processes eventually leading to DNA damage are depicted in Figure 6.1 ordered by the timescale of their appearance. After deposition of the initial radiation quantum, molecules are ionized and excited within femtoseconds or shorter thereby producing secondary electrons and ions. The thus generated electrons might possess sufficient energy to induce further ionization processes resulting in a distribution of the high energy of the primary quantum to a large number of secondary particles. Electrons are the most abundant secondary particles and are highly reactive before they are thermalized and solvated within 10−12 seconds. Molecules such as DNA can effectively be damaged by dissociative electron attachment (DEA) as Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

144

6 Role of Low-Energy Electrons in DNA Radiation Damage

hw

Chromosome

Primary radiation 100 eV to MeV

DNA

T C C A C A G G T A A T A T

Physical stage +

H2O+ → H+ + ·OH −

e + DNA → DNA-# → Fragments 10−12 s

Ionisation, excitation

e−aq

Final biological effects: mutations, cell death

A G C

Chemical stage Stand breaks Base loss Cross links

−

H2O + hw → H2O + e

10−15 s

T A T

10−9 s

10−3 s Diffusion

Figure 6.1 Outline of relevant processes in DNA radiation damage according to their timescale, starting from the initial events of ionization and excitation up to the biological effects of DNA damage. Shown is the type of damage at different time domains,

Biological stage Enzymatic response: repair/ misrepair

1s

weeks Enzymatic reactions

the respective effects on the DNA, and the involved pivotal physical, chemical, and biological processes. Dissociative electron attachment is a crucial process since it takes place very early in timescale.

is shown later in this chapter. Besides ions and electrons another highly reactive secondary particle is the hydroxyl (OH) radical that is produced by ionization and subsequent dissociation of water. In diffusion controlled reactions within 10−9 and 10−3 seconds the OH radical is able to induce severe chemical modifications in DNA such as strand breaks and base loss. The effects of the chemical stage are traditionally referred to as indirect damage; however, also damage caused by secondary electrons and ions taking place in the physical stage can be understood as indirect damage. Single strand breaks (SSBs) can be repaired by enzymes using the complementary strand as a template, but double strand breaks (DSBs) are barely repairable and usually result in cell death. Enzymatic reactions are triggered after some seconds and represent the biological stage of radiation damage. A misrepair or a lack of damage repair results in mutations or cell death. Apparently the final biological effects are determined by the initial ionization processes taking place very early after impact of the high-energy quantum and the majority of the DNA lesions is due to the indirect damage caused by secondary particles. According to very recent Monte–Carlo simulations [7], the most probable energy of secondary electrons that are produced in liquid water by high-energy proton or electron radiation (1–100 MeV) is about 9 eV. When all further ionization events that are caused by energetic secondary electrons are taken into account, almost one third of the electrons possess energy below 1 eV [7]. The importance of secondary

6.1 Radiation Damage and the Role of Low-Energy Electrons

electrons is obvious when considering that more than 80% of the primary energy of 1 MeV proton radiation is carried by low-energy electrons after the primary ionizations [7]. In the meantime it is well known that these low-energy electrons efficiently induce DSBs in plasmid DNA [2, 8, 9] at energies above 5 eV and SSBs even down to 0 eV with basically no threshold energy [10, 11]. A striking feature observed in the yield curves of SSB and DSB caused in plasmid DNA by low-energy electrons is a resonant behavior, that is, the presence of distinct maxima at energies around 10 (DSB and SSB) [2, 9], 2, and 0.8 eV (only SSB) [10]. A DNA strand break requires the cleavage of one or more chemical bonds within the DNA backbone, presumably a C–O or P–O bond between the sugar- and phosphate moieties. The characteristic resonant features of the DNA strand break yields indicate that bond cleavage is due to DEA, and a similar behavior was observed in the analysis of degradation products using high performance liquid chromatography (HPLC) [12] and very recently also by single photon ionization of neutral products [8] desorbed from the surface during electron irradiation of DNA. A direct observation of DNA damage caused by low-energy electrons by means of scanning tunneling microscopy (STM) was reported recently [13]. The observed structural changes of the physisorbed DNA upon irradiation with 8 eV electrons (extension of DNA islands) were ascribed to strand breaks or an unraveling of the double helix. A direct comparison of the damage induced by X-ray photons and secondary electrons reveals that the yield of strand breaks caused by secondary electrons is higher than that is caused directly by X rays [14]. 6.1.1 How Chemical Bonds are Broken by Low-energy Electrons

The dissociation energy of a typical chemical bond in an organic molecule is about 4–5 eV and superficially it is remarkable that electrons with considerably lower energy are able to induce dissociation of these bonds. In such a case the dissociation is driven by the electron affinity (EA) of the formed fragment anion, which might compensate the bond dissociation energy of a broken chemical bond. The initial step in DEA is the formation of a temporary negative ion [15–18] (TNI) by electron capture thereby occupying a formerly empty molecular orbital (MO) – usually an antibonding MO, which then leads to dissociation along a certain bond: e− + ABC −→ ABC#− −→ AB + C

(6.1)

The TNI (ABC#− ) may dissociate along different bonds or even involve complex rearrangements in polyatomic molecules: e− + ABC −→ ABC#− −→ A− + BC

(6.2)

e− + ABC −→ ABC#− −→ AC− + B

(6.3)

Depending on the specific energy of the incoming electron, different parent anionic states can be generated, which are commonly referred to as resonances. A TNI

145

146


ABC−

E

ABC

Γ

E

VAE = D(AB − C) − EA (AB) + E* AB + C EA (AB)

VAE

E* AB− + C D(AB − C)

0

0 EA (ABC)

QC Q −eq

Q(AB − C)

Figure 6.2 Schematic two-dimensional potential energy diagram illustrating the vertical transition from the neutral ground state of the molecule ABC to the anionic state ABC#− requiring the vertical attachment energy (VAE). As the anionic potential energy surface is repulsive in the Franck–Condon region, the generated molecular anion is referred to as transient negative ion (TNI) and dissociates into a stable

Ion yield

anionic and one or more neutral fragments. The energy dependence of the fragment ion formation reflects the initial Franck–Condon transition. The maximum of the ion yield corresponds to a Franck–Condon transition at longer bond distances where the dissociation probability is high due to the longer lifetime of the TNI with respect to AD and the shorter time it takes for dissociation resulting in a redshift of the DEA signal.

is formed in the geometry of the neutral molecule by a vertical Franck–Condon transition as is indicated in a simplified two-dimensional potential energy diagram in Figure 6.2. In a polyatomic molecule this represents a cut through the multidimensional potential energy surface. On the right-hand side of Figure 6.2, a typical DEA spectrum is displayed, which is obtained by scanning the energy of an electron beam within typically 0–15 eV while the yield of an anionic fragment is recorded by mass spectrometry as a function of the electron energy [17, 19]. The TNI in Eqs. (6.1–6.3) is labeled by ‘‘#’’ to indicate that it is intrinsically unstable toward loss of the extra electron (autodetachment (AD)). The AD lifetime τAD of the TNI is determined by the energy width of the resonance due to the Heisenberg uncertainty relation (τAD ≥ h/2π, with h being Planck’s constant (h = 6.6 · ×10−16 eV s) and the energy width of the resonance), and the probability for dissociation is determined by the ratio of the AD lifetime and the time it takes for dissociation: P = exp(−τDEA /τAD ). If the dissociation takes place along one single bond, the dissociation time can be as short as one vibrational period (≈10−14 seconds). In summary, the energetic position of a DEA signal in the ion yield curve is determined by the vertical attachment energy (VAE), convoluted with the (energy-dependent) probability for dissociation into a particular fragment ion.

6.1 Radiation Damage and the Role of Low-Energy Electrons

The intensity of the signal is governed by the AD rate and the Franck–Condon factor of the initial transition; that is, the square of the overlap integral of the neutral ground state and the anionic wave function. The intensity is expressed as (energy-dependent) cross section and is for biomolecules usually in the order of the geometrical cross section of the molecule (10−21 –10−22 m2 ; see [20] for a critical discussion of the absolute DEA cross section of the nucleobase thymine), but can be orders of magnitudes larger in halogen containing molecules. Dissociative electron attachment processes can be investigated experimentally by crossing an electron beam, having defined but variable energy, with a molecular beam. In the experiments described here the electron beam is generated by a trochoidal electron monochromator having an energy resolution of about 100 meV. Negative ions are detected by a quadrupole mass spectrometer as a function of the primary electron energy resulting in ion yield curves shown in the next section and schematically on the right-hand side of Figure 6.2. 6.1.2 DEA Studies of Gas-Phase DNA Building Blocks: The Nucleobases

Electron-induced reactions in biomolecules are discussed in detail in recent review articles [3, 17, 21, 22]. DEA to the DNA and RNA nucleobases thymine [20, 23–28], cytosine [23, 24], adenine [29, 30], guanine [29], and uracil [31, 32] has been studied extensively in the gas phase. All nucleobases (NBs) capture electrons at various energies below 10 eV and dissociate mainly by abstraction of a neutral hydrogen atom to form the closed shell anion [NB–H]− ,1) but also a decomposition of the pyrimidine or purine ring system, respectively, is observed. The abstraction of neutral H proceeds with remarkable site and energy selectivity [28, 33]; it was demonstrated that in thymine and uracil the hydrogen abstraction occurs within a sharp signal peaking at 1.0 eV exclusively by cleavage of the N1–H bond (which is connected to the sugar in nucleosides), whereas at an electron energy of 1.8 eV predominantly the N3–H bond dissociates. In the purine nucleobase adenine it was shown that at electron energies below 1.5 eV only the N9–H bond is broken [30]. These studies lead to the assumption that in DNA the NBs act as antennas for low-energy electrons, in that an excess electron gets initially localized on the nucleobase moiety and the extra charge then induces a bond cleavage within the DNA backbone resulting in a strand break. Different scenarios of DEA to single nucleotides have been calculated by different research groups using density functional theory (DFT) [34–40]. In general, it is confirmed that electrons of energy below 2 eV can attach to the NBs, which might be followed by sugar–phosphate bond cleavage due to coupling of the π ∗ orbital localized at the nucleobase with a σ ∗ orbital at the sugar–phosphate moiety. On the other hand, electron attachment to the phosphate group was predicted to proceed only at higher energies around 2 eV. An interesting alternative mechanism for low-energy electron–induced strand 1) The notation [M–H]− is commonly used

in mass spectrometry and the dash reads

‘‘minus’’and does not refer to a chemical bond.

147

148


breaks was suggested by Dabkowska et al. [41], who considered initial formation of a stable hydrogenated closed shell anion of cytidine followed by intramolecular proton transfer from the sugar moiety to the nucleobase eventually leading to C–O bond cleavage in the backbone. Owing to the low barriers of the reactions and the fact that only bound anions are involved, the rates of strand breakage are much higher than that in other proposed mechanisms. Experimental data on direct low-energy electron attachment to the DNA backbone is scarce, which is also due to the limited accessibility of suitable gas-phase model compounds for the DNA backbone. Nevertheless, in a recent study of low-energy electron–induced DNA strand breaks, it was demonstrated that the yield of neutral fragmentation products associated with a decomposition of the sugar moiety in the DNA backbone can be correlated very well with the resonant features in the strand break yield curves, whereas fragmentation products associated with the decomposition of the NBs give rise to different resonances [8]. High resolution electron energy loss (HREELS) and XPS (X-ray photoelectron spectroscopy) studies of self-assembled monolayers of DNA also suggest that electrons mainly interact with the DNA backbone [42]. In the following section, the experimental challenges that are encountered when investigating DNA backbone subunits are described, along with results demonstrating their ability to dissociatively capture low-energy electrons.

6.2 DEA Studies on Model Compounds for the DNA Backbone

The DNA backbone is composed of alternating 2-deoxy-d-ribose and phosphate moieties (Figure 6.3), but model compounds for the sugar unit and the phosphate group are not accessible in a straightforward way. In this section, DEA studies on the sugar molecule d-ribose are presented followed by an investigation of acetylated d-ribose, which is regarded as a considerably improved surrogate for the sugar unit in DNA or RNA due to its more similar chemical structure. Finally, an experimental approach that allows the study of DEA to more complex biomolecules in the gas phase is presented, and the discussion of DEA to the DNA backbone is completed by the presentation of results on d-ribose-5 -phosphate. The chemical structures of the investigated molecules are shown in Figure 6.3. 6.2.1 Electron Attachment to D-Ribose

The sugar moiety in DNA occupies a central position since it connects two phosphate groups of the backbone with a nucleobase, and a damage of the sugar would lead to either base release or a strand break. And indeed, the isolated (gas phase) RNA sugar d-ribose captures electrons at very low energies followed by numerous fragmentation reactions [43]. Examples of typical ion yields obtained from d-ribose are shown in Figures 6.4–6.6. All fragment ions are observed


O O P O O−

CH3 O C O

H O

NB

4

H

O

5 H O

CH3 C O O

O HO

OH

O

P

O

HO O O P O O− DNA backbone

H

H

3

2 OH 1

OH D-Ribose

R C5H10O5 150 amu

OH

O O O C C O CH3 CH3

H

Tetraacetyl-D-ribose TAR (CH3COO)4C5H6O 350 amu

149

OH H

H

H

OH

H OH

D-Ribose-5′-phosphate

RP H2PO3C5H9O5 230 amu

Figure 6.3 Part of the DNA backbone compared to the molecular structures of D-ribose in the pyranose form, tetraacetyl-D-ribose in the furanose form, which is more comparable to the sugar unit in DNA, and D-ribose-5 -phosphoric acid representing a model compound for the entire DNA backbone.

within a sharp signal close to 0 eV, and some fragment ions also appear from a resonant feature located at 6–8 eV (as shown in Figure 6.6). A typical fragmentation reaction of monosaccharide anions is the loss of neutral water molecules as was demonstrated with the pentoses 2-deoxy-d-ribose [44] and d-ribose [43], and the hexose d-fructose [45]: e− + M −→ M−# −→ [M–nH2 O]− + nH2 O

(6.4)

with n = 1–2 for d-ribose [43] and 2-deoxy-d-ribose [44], and 1–3 for d-fructose [45]. The thermodynamic threshold of a DEA reaction is given by the sum of the bond dissociation energies D that need to be broken subtracted by the sum of bond dissociation energies of newly formed bonds and the EA of the formed anionic fragment. For the abstraction of one H2 O molecule this is R H = D(C–OH) + D(C–H) − D(H–OH) − D(C=C) − EA(C5 H8 O4 ) (6.5) The EA of C5 H8 O4 is not known, but must be positive, and the sum of bond dissociation energies is zero when using typical values for the involved bonds [47] (D(C–OH) = 4.0 eV, D(C–H) = 4.2 eV, D(H–OH) = 5.2 eV, D(C = C) = 3 eV). Thus, the reaction is in total exothermic, but it is required that the activation barriers are very low so that the bond breaking and bond formation must proceed simultaneously. This reaction remarkably demonstrates how a complex reaction involving multiple bond breaking is solely triggered by the excess charge as it occurs close to 0 eV electron energy. The loss of different numbers of water molecules from the TNI most likely leaves the sugar ring intact. Nevertheless, several reactions that are associated with a fragmentation of the carbon–oxygen ring have been observed, that is, a cross-ring cleavage. The most dominant signals that arise from cross-ring cleavage are shown in Figures 6.4–6.6. The unambiguous assignment of the fragment masses to

150


5-13C-D-Ribose

D-Ribose

OH

C

OH

OH OH

Ion count rate (s−1)

101 amu: C4H5O3−

H C O OH H OH

1.0

1.5

SF6− OH

OH OH

OH OH

102 amu: 13CC H O − 3 5 3

CH2O, H2O

C4H6O3− (102 amu)

103 amu: 13CC H O − 3 6 3

102 amu: C4H6O3− 2.0

C

OH OH

102 amu: C4H6O3− 0.5

O 13OH

101 amu: C4H5O3−

e− (0 eV) +

0.0

O OH

13

O OH

1-13C-D-Ribose

0.0

0.5

1.0

1.5

2.0

0.0

0.5

1.0

1.5

2.0

Electron energy (eV) Figure 6.4 Ion yield curves for fragment anions at 101 and 102 amu, which are assigned to C4 H5 O3 − and C4 H6 O3 − , respectively. The DEA spectra of D-ribose labeled at C5 with 13 C (middle) are identical to the spectra of nonlabeled D-ribose (left) indicating that the loss of the carbon containing

neutral fragment occurs exclusively from C5 (as indicated in the reaction scheme). Conversely, the signals obtained from D-ribose labeled at C1 with 13 C (right) are shifted to 102 and 103 amu, respectively, confirming that the negative charge remains on the C1 site.The DEA spectra are adapted from [43].

specific formulas was only possible by the use of different isotope-labeled d-ribose molecules [43] containing either 13 C or deuterium (1-13 C-d-ribose, 5-13 C-d-ribose, and 1-D-d-ribose). Furthermore, detailed fragmentation studies of deprotonated d-ribose and d-fructose using matrix-assisted laser desorption and ionization (MALDI) [48] revealed a systematic fragmentation pattern, in that the negatively charged sugars loose multiples of neutral H2 O molecules and CH2 O units. Three examples of cross-ring cleavage reactions are described in detail in the next section. 6.2.2 Cross-Ring Cleavage of D-Ribose Proceeds with Selective Charge Retention

The dominant fragment anions associated with cross-ring cleavage of d-ribose subsequent to electron attachment were observed at 101, 102, 71, and 72 amu. The assignment of the product ions to a specific stoichiometry was supported by measurements with isotope-labeled molecules resulting in the following fragmentation scheme:


5-13C-D-Ribose 13

400

OH 200

C

O

151

1-13C-D-Ribose

OH

600

O OH

OH OH

300

OH C

13

OH OH 71 amu

71 amu C3H3O2−


0 600

400

300

200 72 amu C3H4O2−

72 amu − 2H3O2

13CC

0 600

400

200

73 amu

300 13

73 amu CC2H4O2−

0 0.0

0.5

1.0

1.5

2.0

0.0

0.5

1.0

1.5

Electron energy (eV) Figure 6.5 DEA spectra obtained at 71, 72, and 73 amu from 5-13 C-D-ribose (left) and 1-13 C-D-ribose (right), which are assigned to a cross-ring cleavage resulting in C3 H3 O2 − and C3 H4 O2 − , respectively. Two carbon atoms are excised as neutral fragments

exclusively from C5 and C4, since the spectra of 5-13 C-D-ribose are identical to nonlabeled D-ribose (not shown here) and in 1-13 C-D-ribose the signals are shifted to 72 and 73 amu, respectively.The DEA spectra are adapted from [43].

e− (ε ≈ 0 eV) −→ R−# −→ C4 H6 O− 3 (102 amu) + CH2 O + H2 O

(6.6)

−→ C4 H5 O− 3 (101 amu) + CH2 O + H2 O + H

(6.7)

−→ C3 H4 O− 2 (72 amu) + C2 H4 O2 + H2 O

(6.8)

−→ C3 H3 O− 2 (71 amu) + C2 H4 O2 + H2 O + H

(6.9)

Furthermore, the use of isotope-labeled molecules revealed another fascinating aspect of the DEA reactions to d-ribose and other monosaccharides, that is, the selective excision of neutral fragments containing C5 (see atom labeling in Figure 6.3). In Figure 6.4, the ion yield curves of the fragment anions at 102 and 101 amu from nonlabeled d-ribose are illustrated on the left-hand side. The spectra

2.0


1-13C-D-Ribose

D-Ribose – nonlabeled

59 amu O OH OH

Ion yield (a.u.)

152

O OH 13 C

OH OH

OH OH OH

59 amu C2H3O2−

0

1

2

3

4

5

6

7

8

9 10 11

60 amu 0

1

2

3

13

C

OH

5

6

7

8

O OH

O OH OH

OH OH

59 amu

OH OH

59 amu

D

60 amu 0

1

2

3

4

5

9 10 11

1-D-D-Ribose

5- C -D-Ribose 13

4

6

7

8

60 amu

0 1 2 3 9 10 11 Electron energy (eV)

Figure 6.6 Energy dependence of fragment ions at 59 and 60 amu obtained from DEA to D-ribose and its isotope-labeled analogs 5-13 C-D-ribose, 1-13 C-D-ribose, and 1-D-D-ribose. These anions are formed from two different resonances located close to 0 eV and at

4

5

6

7

8

9 10 11

6–8 eV. The selectivity of charge retention is less pronounced than for the heavier fragment anions; however, a clear preference for the C2 H3 O2 − anion to be formed from C1 is apparent.The DEA spectra are adapted from [46].

in the middle are taken from 5-13 C-d-ribose and show the same curves at the same masses as the nonlabeled d-ribose demonstrating that the neutral fragment in this reaction is exclusively generated from the C5 position. In contrast to this, both spectra are completely shifted by one mass unit to higher masses in 1-13 C-d-ribose (right-hand side) indicating that in this case the fragment anions contain 13 C, that is, the anomeric carbon atom. The situation is similar for the fragment anions at 72 and 71 amu, which corresponds to the loss of a neutral fragment containing two carbon atoms. In Figure 6.5 the spectra recorded at 71, 72, and 73 amu from 5-13 C-d-ribose (left) and 1-13 C-d-ribose (right) are compared. Apparently the signals obtained from 1-13 C-d-ribose are almost completely shifted to 72 and 73 amu, whereas the spectra of 5-13 C-d-ribose exhibit signals at 71 and 72 amu corresponding to the spectra of nonlabeled d-ribose (not shown here). Hence, these reactions proceed with selective charge retention on C1, similar to the prompt dissociation of deprotonated d-ribose and d-fructose recently observed in MALDI mass spectra [48]. Selective bond breaking involving the excision of C5 from d-ribose was also recently demonstrated in experiments using low-energy ion beams [49]. It should be noted that in DNA


the C5 atom of the sugar moiety is connected to the adjacent phosphate group and an excision of this atom directly results in a strand break. In addition to the cross-ring cleavage associated with loss of one or two carbon atoms also the abstraction of three neutral CH2 O units (or one C3 H6 O3 fragment) was observed resulting in the fragment anion at 59 amu [46]: e− (ε ≈ 0 eV) −→ R−# −→ C2 H3 O− 2 (59 amu) + C3 H6 O3 + H

(6.10)

The DEA spectra are shown in Figure 6.6 and the C2 H3 O2 − fragment ion is different compared to the ones discussed above in the sense that it is also generated in considerable amounts from a resonance located at higher energies between 6 and 8 eV. Similar to the selectivity of Reactions 6.6–6.9, the C2 H3 O2 − ion is mainly formed from the C1 site of the d-ribose molecule and the neutral fragment most likely consists of C5, C4, and C3 as can be seen on the very weak signal at 60 amu in 5-13 C-d-ribose. However, the selectivity is not complete and even slightly less pronounced on the higher energy resonance indicating that two different reaction pathways exist and a small fraction of the C2 H3 O2 − ions contains C5. The origin of the two resonances near 0 eV and at 6–8 eV is explained in the next subsection. The electron-induced selective bond breaking is a fascinating aspect of the DEA reactions, since selective bond breaking and bond formation has always been a major vision and challenge in chemistry and was pursued with enormous experimental effort especially in laser chemistry [50, 51]. Moreover, low-energy electron beams can be used as a comparatively convenient tool [52] to generate specific reaction products by site- and energy-selective bond breaking [53] at specific electron energies via the formation of defined negative ion states. According to the preceding discussion, a large number of fragment anions arise when a transient negative ion of d-ribose is formed near 0 eV. The DEA experiments reveal the mass-to-charge ratio of the anions and the energy at which they are formed. By means of isotope-labeled molecules, the stoichiometry of ionic and neutral fragments could be derived and the neutrals were found to be exclusively produced from the C5 site of the d-ribose molecule. The chemical structure of the fragments can be derived from the fact that the reactions leading to these fragments have to be exothermic as they occur close to 0 eV. DFT calculations at the B3LYP/6-31++G∗∗ level of theory using Gaussian 03 revealed the thermodynamic thresholds of Reactions 6.7 and 6.8 for different structures of the anionic and neutral fragments. In Figure 6.7 schematic energy diagrams are depicted starting from the lowest energy conformation of d-ribose, which is stabilized by three intramolecular hydrogen bonds between OH2, OH3, and OH4 [54]. According to the DFT calculations only one possible structure of the C4 H5 O3 − ion (at 101 amu) out of nine different linear and cyclic structures (not shown here) [55] is stable enough to render the reaction exothermic. The anion contains a carboxyl group and is stabilized by an intramolecular hydrogen bond as shown in Figure 6.7. The EA of the corresponding neutral molecule is with 3.9 eV remarkably high, and all other possible geometries for C4 H5 O3 − (not shown here) are at least 1.9 eV higher in energy [55]. Figure 6.7 shows two alternative compositions for the

153

154

6 Role of Low-Energy Electrons in DNA Radiation Damage D-Ribose:

72 amu: C3H4O2−

C5H10O5 101 amu: C4H5O3− CH2O H2O H

+ H2O CH(OH)CH(OH)

+0.003 eV

0

+ −0.098 eV −0.506 eV

+

+ CH3OH OH

H 2O

CH3COOH

−1.545 eV

Figure 6.7 Thermodynamic thresholds for the formation of C4 H5 O3 − (101 amu) and C3 H4 O2 − (72 amu), respectively, as calculated at the B3LYP/6-31++G** level of theory. For both fragment anions two neutral decay channels are indicated. The C4 H5 O3 − ion can either be associated with formation of neutral CH2 O + H2 O + H resulting in a threshold of −98 meV, or with CH3 OH

+ OH resulting in a threshold of −0.5 eV. For the C3 H4 O2 − (72 amu) anion thermodynamic thresholds of +3 meV for the neutral channel CH(OH)CH(OH) + H2 O or, alternatively, −1.5 eV for the the formation of neutral CH3 COOH + H2 O were calculated. For the anionic fragments, the most stable isomers or conformers out of many are displayed.

neutral fragments (CH2 O + H2 O + H, and CH3 OH + OH), which both result in exothermic reactions. However, the formation of water and formaldehyde might require less activation energy than the formation of methanol and the hydroxyl radical due to considerable rearrangements in the latter case. The most stable structure for C3 H4 O2 − (72 amu) found by DFT calculations is the propenoic acid anion as shown in Figure 6.7. Two alternative neutral channels accompanying the propenoic acid anion generation are shown on the left-hand side of Figure 6.7. The reaction is exothermic by 1.5 eV if acetic acid and water are formed, and slightly endothermic by 3 meV for the case that Z-dihydroxyethene and water are produced in addition to the propenoic acid anion. It is interesting to note that the spectra shown in Figure 6.5 exhibit rather broad signals, and the slightly better electron energy resolution employed in the measurements of 5-13 C-d-ribose even reveals two distinct features contributing to the formation of C3 H4 O2 − . Thus, two different reaction pathways having different thermodynamic thresholds might be operative from the same resonance, and the slightly endothermic reaction might appear only at the high-energy side of the resonance as a shoulder. 6.2.3 The Nature of the Transient Negative D-Ribose Anions

As is described in Section 6.1.1 electron attachment at energies below 4 eV proceeds without changing the electronic configuration of the neutral molecule,


what is referred to as a shape resonance [15, 17, 56] due to the explicit shape of the interaction potential between the incoming electron and the neutral molecule that is responsible for the trapping of the excess electron. As the d-ribose molecule is a saturated molecule the excess electron must occupy a σ ∗ orbital. However, a σ ∗ shape resonance is usually located at appreciably higher energies than 0 eV due to the high energy of the σ ∗ orbital in the neutral molecule that is not expected to be reduced in the negative sugar ion such that it appears close to 0 eV.2) For instance, in a recent systematic DEA study of different aliphatic and cyclic alcohols ∗ shape resonance is located around 2 eV. [57] it was demonstrated that a σO−H ∗ Furthermore, a σ shape resonance is expected to be comparatively localized on certain bonds that are cleaved to yield, for instance, [M–H]− in the case of alcohols ∗ arising from an O–H bond dissociation due to occupation of the σO−H antibonding MO. However, the manifold of different fragment anions observed in d-ribose arising from the 0 eV resonance through multiple bond breaking indicates that the initial negative ion state is rather delocalized. Such a ‘‘diffuse’’ negative ion state can be formed due to the large dipole moment of the sugar molecule. The extra electron is attracted by the positive site of the dipole moment and gets trapped into a very diffuse rydberg-like orbital to form a dipole-bound state (DBS) [58, 59]. It was shown previously by ab initio calculations [60] that d-fructose indeed can form a DBS, which may serve as a doorway to dissociation when the DBS couples to a valence state. The formation of DBS by electron attachment followed by dissociation was previously observed in the NBs uracil [61] and thymine [62] and also in other molecules, for example, ethylene carbonate [63]. The negative ion signals close to 0 eV are thus tentatively assigned to a DBS, but recent electron scattering calculations on d-ribose [46] predicted a series of shape resonances located between 7.98 and 9.49 eV. In this energy range, between 5.5 and 9.0 eV, the experiments indeed show weaker signals due to the formation of C3 H4 O2 − , C2 H3 O2 − (Figure 6.6), and OH− . This is consistent with the analysis of the spatial distribution of the calculations, which predict that the shape resonances lead to C–OH and C–C bond breaking [46]. 6.2.4 One Step Further: Tetraacetyl-D-Ribose

The DEA studies on d-ribose demonstrated that electron attachment to a sugar molecule is possible and, moreover, leads even at 0 eV to manifold fragmentation accompanied by a decomposition of the ring structure. Consequently, these reactions are highly relevant for the discussion of molecular mechanisms of DNA strand breaks since the fragmentation of the sugar moiety in DNA corresponds to both strand breaks and loss of a nucleobase and thus loss of genetic information. 2) According to Koopmans’ theorem, the

attachment energy (or electron affinity) corresponds to the negative of the virtual orbital energy (VOE) as obtained from a self-consistent field calculation. However,

orbital relaxation and electron correlation energies are neglected in this approximation, so that the VOE must be scaled by at least 0.5 eV. See [18] (J. Simons, J. Phys. Chem. A (2008), 112, 6401) for details.

155

156


However, apparently the structure of isolated gas-phase d-ribose is considerably different from the structure of sugar in DNA, since the latter adopts the furanose form in contrast to the pyranose structure of gas-phase d-ribose [54], which is shown in Figure 6.3. Additionally, the 2-deoxy-d-ribose unit in DNA does not possess any free hydroxyl groups, but it may be these hydroxyl groups that enable the electron attachment reactions described above in d-ribose. In order to elucidate whether the reactions described above for bare d-ribose also take place in a sugar that is bound to a more complex molecular environment, we investigated DEA to peracetylated d-ribose [64], which additionally can be obtained in pure furanosic form (Figure 6.3). Of course, the acetyl groups in tetraacetyl-d-ribose (TAR) are only a rough model for the nucleobase and the phosphate groups of the DNA molecule; nevertheless, the TAR molecule is a very impressing example to demonstrate how low-energy electrons react with a complex molecule containing several polar functional groups. Figure 6.8 shows a selection of DEA spectra obtained from TAR. The most dominant fragment anion was observed at 59 amu (top panel of Figure 6.8) and is due to the acetate ion generated through a single C–O bond dissociation: (CH3 COO)4 C5 H6 O + e− (0 eV; 1–3 eV; 7–11 eV) −→ CH3 COO− (59 amu) + (CH3 COO)3 C5 H6 O

(6.11)

It arises from three different resonant features located close to 0 eV, between 1 and 3 eV, and with smaller intensity at 7–11 eV. These distinct energy regimes are characteristic of the behavior of different parts of the molecule. The second spectrum was recorded at 70 amu and represents a group of anions [64] (100, 84, and 70 amu, but only the latter is shown here) that are only formed close to 0 eV and additionally at higher electron energies, but no intensity was observed at medium energies. The shape of the spectra is reminiscent to the DEA spectra of bare d-ribose, and so are the corresponding masses indicating that these fragment anions are associated with processes in the central sugar ring resulting in its decomposition. The 70 amu fragment is ascribed to the following reaction: (CH3 COO)4 C5 H6 O + e− (0 eV; 7–11 eV) −→ C3 H2 O− 2 (70 amu) + fragments

(6.12)

That is, the 0 eV signals are due to the initial formation of a DBS followed by a dissociation of the sugar ring and eventually leading to similar fragment anions as observed from bare d-ribose. In contrast to this the third spectrum of Figure 6.8 represents a group of heavier fragment anions (215, 161, 154, 119, and 113 amu; see [64] for details), that are exclusively generated from the resonance located at 1–3 eV and is due to the loss of the exocyclic acetate groups. The ion at 154 amu is ascribed to the following reaction: (CH3 COO)4 C5 H6 O + e− (1–3 eV) −→ C7 H6 O− 4 (154 amu) + 2CH3 COOH + CH3 CHO

(6.13)


40

59 amu CH3COO−

Ion yield (102 s−1)

20

0 6

0

1

2

3

4

5

6

7

8

4

9 10 11

70 amu C3H2O2−

2 0 12

0

1

2

8

3

4

5

6

7

8

9 10 11

154 amu C7H6O4−

4 0 0

1

2

3 4 5 6 7 8 Electron energy (eV)

Figure 6.8 Representative ion yield curves obtained from DEA to peracetylated D-ribose. The upper panel shows the DEA spectrum of the acetate anion, which is formed within three distinct resonant features. The signal close to 0 eV is tentatively assigned to the initial formation of a dipole-bound state, the peaks at 1–3 eV are ascribed to π ∗ shape resonances located on the acetate groups, and the signals at 7–11 eV are most likely due to core excited resonances

9 10 11

within the sugar ring. The middle spectrum is reminiscent to the DEA spectra of bare D-ribose and is associated with a decomposition of the sugar ring, thus no signal appears at 1–3 eV. The lower spectrum is representative for a group of heavier fragments associated with the abstraction of exocyclic groups exclusively through the π ∗ shape resonance at 1–3 eV.The DEA spectra are adapted from [64].

These spectra clearly indicate that the dissociation through DBS as observed in bare d-ribose is preserved in the appreciably more complex molecule TAR. More general, the experiments demonstrate that in a molecule composed of different moieties characteristic resonant features of the individual subunits are preserved. In TAR these are the polar exocyclic acetate groups leading to π ∗ shape resonances located at 1–3 eV and the central sugar ring resulting in ion formation close to 0 eV and at higher energies between 7 and 11 eV. In Figure 6.9 it is illustrated in the picture of a simplified two-dimensional potential energy diagram how different resonances contribute to the ion yield of, for example, the CH3 COO− ion at 59 amu. A TNI (indicated as TAR#− ) is formed by a

157

158


TAR−

TAR# −

TAR

E

E CH3COO + fragments

TAR− (DBS) 0

CH3COO− + fragments

0 Ion yield

Q Figure 6.9 Schematic potential energy diagram illustrating the formation of the fragment anion CH3 COO− through the initial formation of two different resonant states of tetraacetyl-D-ribose (TAR) resulting in two distinct signals in the ion yield curve. At 0 eV a dipole-bound state can be formed serving as a doorway to dissociation since the dissociative valence anionic state is occupied

via intramolecular charge transfer. The valence anionic state can also be accessed by a direct Franck–Condon transition resulting in a π ∗ shape resonance. An example for an MO representation of the excess electron in the π ∗ shape resonance is displayed, which was calculated at the B3LYP/6-31++G** level of theory using the stabilization method [64].

Franck–Condon transition requiring energy of about 2 eV. The MO representation of the TNI is a representative example resulting from ab initio calculations using the stabilization method [64]. The TNI formed at 2 eV is unstable with respect to dissociation and different fragment anions can result from this particular TNI, among them the acetate anion. The right-hand side of Figure 6.9 shows the resulting ion yield curve. The same fragment anion can also be formed at 0 eV when a DBS is initially formed. In the DBS the excess electron is still located relatively far away from the molecular framework, thus the geometry of the DBS corresponds to the geometry of the neutral and the binding energy of the excess electron is slightly positive. Consequently, the DBS can be formed at 0 eV, and an intramolecular charge transfer to the valence-bound state then leads to dissociation. The experiments with TAR yield important indications about the electron attachment reactions in DNA. The sugar unit is very active toward low-energy electrons also when it is bound within a more complex molecular environment; the only prerequisite seems to be the presence of polar side groups. In a DNA nucleotide, these are one nucleobase and two phosphate groups that in turn possess considerable electron affinities resulting in shape resonances located at the respective moiety.


In the next two sections, the investigation of a combined sugar–phosphate compound with respect to DEA is described along with the experimental challenges that are connected with the study of such molecules. 6.2.5 The Use of Laser-Induced Acoustic Desorption (LIAD) to Study DEA to Larger Biomolecules

The study of biomolecules in the gas phase, which are larger than single NBs or monosaccharides such as nucleosides or nucleotides, raises problems due to their thermal fragility. It was demonstrated previously that conventional sublimation of nucleosides such as thymidine in high vacuum by appropriate heating results in thermal decomposition [65]. As gas-phase experiments yield a wealth of information and due to their function as a reference for both theoretical calculations and experiments in more biologically relevant environments, more sophisticated methods are required to transfer biomolecules into the gas phase to make them accessible for DEA measurements. A very gentle method to desorb only neutral molecules without ionization is laser-induced acoustic desorption (LIAD) [66–68]. The sample molecules are deposited on a thin titanium foil, which is then irradiated from the backside with a pulsed Nd:YAG laser operated at 532 nm (15 Hz, 3 mJ per pulse, 2–6 ns). The laser pulse induces a shock wave that propagates through the metal foil and leads to a gentle desorption of sample molecules from the opposite surface, without resulting in fragmentation or ionization of the sample molecules. The desorbed neutral molecules then interact with an electron beam of defined but variable energy, and the thus formed anions are analyzed by a mass spectrometer. The proof-of-principle experiments were performed by using the nucleoside thymidine [69] and the sugar–phosphate compound d-ribose-5 -phosphate [69] (RP), which is shortly discussed in the next section. 6.2.6 Sugar–Phosphate Cleavage Induced by 0 eV Electrons: DEA to D-Ribose-5 -Phosphate

d-Ribose-5 -phosphate can be understood as a model compound for the entire DNA or RNA backbone. Figure 6.10a shows DEA spectra exhibiting formation of the H2 PO4 − anion and the ribose anion C5 H9 O5 − that are due to a C–O or a P–O bond cleavage, respectively: e− (≈ 0 eV) + (HO)2 OPO–C5 H9 O4 −→ H2 PO− 4 + C5 H9 O4

(6.14)

e− (≈ 0 eV) + (HO)2 OP–OC5 H9 O4 −→ H2 PO3 + C5 H9 O− 5

(6.15)

Both reactions occur close to 0 eV and represent a sugar–phosphate cleavage corresponding to a strand break in DNA or RNA. A more detailed analysis of the reactions by means of DFT calculations at the B3LYP/6-31G++∗∗ level of theory [69] reveals the energetics of the reaction pathways as shown in Figure 6.10b. The

159

0

2

4

6

8

0

2

4

6

8

10

12

14

−1 0

1

2

3 E (eV)

4

5

6 7

149 amu C5H9O5−

97 amu H2PO4−

8

9

(b)

HO

A

OH

P

O O

OH

O

OH

H OH

H

−0.19 eV

H

H

B

+0.43 eV

+0.78 eV

O

−

C

HO

H

OH

P

O

OH

H

O

+ H2PO3

−2.10 eV

O− + C H O 5 9 4

−0.20 eV

H OH

H

OH

Figure 6.10 (a) Ion yields of the phosphate anion H2 PO4 − and the D-ribose anion C5 H9 O5 − generated by DEA to D-ribose-5 -phosphate (RP) through a sugar–phosphate cleavage. Compared to the situation in DNA, both reactions correspond to a strand break. (b) Activation barriers for both reaction pathways. All the molecules were fully optimized at the B3LYP/6-31++G** level of theory and the given energies are in respect to the neutral ground state RP molecule, which is not shown here. A represents the substrate anion, B the transition states and C the sum of the isolated products. The DEA spectra and illustration are adapted from [69].

(a)


160


6.3 Outlook and Future Prospects

calculations indicate that ribose-5 -phosphate can form a stable molecular anion with an electron binding energy of 0.2 eV. The dissociation into the phosphate anion is driven by its thermodynamic stability reflected by the EA of the corresponding neutral (EA(H2 PO4 ) = 4.9 eV [70]). In contrast, the generation of the ribose fragment anion leads to products that are barely more stable than the RP− anion. Nevertheless, the activation energy for P–O bond breaking is appreciably lower than for the H2 PO4 − formation, making it a kinetically controlled reaction. The fact that two different bonds are broken and that the excess charge is subsequently localized on either the phosphate group or the sugar unit suggests that two different resonances are involved. The electron attachment mechanisms to d-ribose are discussed in detail above, but also the phosphate group is known to capture electrons at various energies down to 0 eV [71]. The low energy of the signals indicates that also in d-ribose-5-phosphate DBSs may be involved, which is supported by a recent theoretical and experimental study excluding low-lying π ∗ shape resonances in phosphate compounds [72].

6.3 Outlook and Future Prospects

The investigation of molecular mechanisms of DNA radiation damage and the role of low-energy electrons is beneficial for an improved understanding of the mode of action of radiosensitizers in radiation and chemoradiation therapy for tumor treatment [73, 74]. For instance, the chemotherapeutic agent cisplatin was shown to enhance the yield of electron-induced SSBs and DSBs in DNA by factors of 1.3–4.4 depending on the electron energy [74]. The enhancement is largest for 10 eV electrons, which is ascribed to an increase in the magnitude of DEA processes in the presence of cisplatin. A detailed study of the underlying DEA processes may have implications for the design of new chemotherapeutic and radiosensitizing drugs and for the development of more efficient protocols in cancer therapy. The understanding of molecular mechanisms of electron-induced DNA strand breaks has improved considerably over the past few years, but a number of controversial results still demonstrate the complexity of the system. For example, it still remains to be evaluated whether DNA strand breaks are mainly induced by initial localization of low-energy electrons at the NBs or in the backbone. The NBs are usually regarded as antennas for low-energy electrons, and the role of NBs is reflected, for instance, in the sequence dependence of electron capture by DNA [75]. However, at different electron energies different electron attachment mechanisms prevail and various experiments point out the active role of the sugar–phosphate backbone in electron attachment [8, 42, 69, 71, 76]. The experiments described in the present chapter indicate that electron attachment to the backbone subunits proceeds very effectively presumably via the initial formation of DBSs. It must be investigated if such DBSs are also relevant in a more biologically realistic environment [59].

161

162


In general, it is crucial to determine the effect of solvation by water molecules on electron-induced reactions in biomolecular systems. This question was tracked recently by experiments performed with films of short oligonucleotides covered by water [77] showing that fragment anion formation is increased by a factor of 1.6 in the presence of water. The recently initiated experiments combining DEA and LIAD offer a new scope of experiments that can be performed, for example, with oligonucleotides or peptides. The role of DNA binding proteins in electron-induced DNA damage is an important issue that was recently addressed by experiments using self-assembled monolayers [78] and must be further investigated. Apart from their relevance to radiation damage, the experiments on electroninduced reactions in isotope-labeled sugars (and previously in NBs) revealed that in these cases bond cleavage is highly energy- and site selective. As is indicated in Section 6.1.2 it has been shown in the case of pyrimidine nucleobases [33] and the purine nucleobase adenine [30] how the electron-induced hydrogen abstraction from the nucleobase can be controlled by simply changing the energy of the electron beam. More precisely, the energy of the incoming electron determines, on the one hand, the specific atom from which hydrogen is abstracted (either N1 or N3) and, on the other hand, if the negative charge remains on the nucleobase resulting in the dehydrogenated nucleobase anion or on the hydrogen atom leading to the hydride anion [26, 28]. In another DEA study about organic molecules containing different functional groups [79, 80], it was demonstrated that different TNIs are formed at electron energies specific for a certain functional group leading to fragmentation products at characteristic electron energies. In other words, a particular bond cleavage can be triggered just by tuning the energy of the electron beam. This can technically be applied to a controlled manipulation of materials at the nanoscale (nanolithography) [52]. Low-energy electron–induced polymerization of CH3 SSCH3 involving electron attachment, and subsequent dissociation was demonstrated very recently by means of STM [81]. The initial step is attachment of an electron emitted by the STM tip followed by dissociation along the S–S bond and subsequent reaction with an adjacent CH3 SSCH3 molecule. This example demonstrates the potential of DEA reactions as a route to synthesize novel materials using low-energy electron sources such as the STM, which has the advantage of possessing a remarkable spatial resolution.

Acknowledgments

The work presented here was conducted within my PhD project that was supervised by Prof. Eugen Illenberger who was an excellent mentor whom I hold in great respect and gratitude. Furthermore, I acknowledge very fruitful collaborations and discussions with Dr. Janina Kopyra and Dr. Iwona Dabkowska. Finally, a PhD fellowship by the Studienstiftung des deutschen Volkes is gratefully acknowledged.

References

References 1. Michael, B.D. and O’Neill, P. (2000) 2.

3. 4. 5.

6. 7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

Science, 287, 1603. Boudaiffa, B., Cloutier, P., Hunting, D., Huels, M.A., and Sanche, L. (2000) Science, 287, 1658. Sanche, L. (2005) Eur. Phys. J. D, 35, 367. Ball, P. (2003) Nature, 421, 421. Luger, K., Mader, A.W., Richmond, R.K., Sargent, D.F., and Richmond, T.J. (1997) Nature, 389, 251. Felsenfeld, G.and Groudine, M. (2003) Nature, 421, 448. Pimblott, S.M.and LaVerne, J.A. (2007) Radiat. Phys. Chem., 76, 1244. Chen, Y.F., Aleksandrov, A., and Orlando, T.M. (2008) Int. J. Mass Spectrom., 277, 314. Huels, M.A., Boudaiffa, B., Cloutier, P., Hunting, D., and Sanche, L. (2003) J. Am. Chem. Soc., 125, 4467. Martin, F., Burrow, P.D., Cai, Z., Cloutier, P., Hunting, D., and Sanche, L. (2004) Phys. Rev. Lett., 93, 068101. Panajotovic, R., Martin, F., Cloutier, P., Hunting, D., and Sanche, L. (2006) Radiat. Res., 165, 452. Zheng, Y., Wagner, J.R., and Sanche, L. (2006) Phys. Rev. Lett., 96, 208101. Mochiji, K., Hashimoto, H., Tanaka, Y., Ninomiya, N., and Takeo, M. (2007) Phys. Rev. B, 75, 094302. Cai, Z., Cloutier, P., Hunting, D., and Sanche, L. (2005) J. Phys. Chem. B, 109, 4796. Illenberger, E., and Momigny, J. (1992) Gaseous Molecular Ions. An Introduction to Elementary Processes Induced by Ionization, Springer, Steinkopff . Illenberger, E. (2000) Electron attachment processes to free and bound molecules, in Photoinonization and Photodetachment, Advanced Series in Physical-Chemistry, Vol. 10B (ed. C.-Y. Ng), World Scientific, Singapore Part II. Bald, I., Langer, J., Tegeder, P., and Ingolfsson, O. (2008) Int. J. Mass Spectrom., 277, 4.

18. Simons, J. (2008) J. Phys. Chem. A, 112,

6401. 19. Balog, R., Langer, J., Gohlke, S.,

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Stano, M., Abdoul-Carime, H., and Illenberger, E. (2004) Int. J. Mass Spectrom., 233, 267. Kopyra, J., Koenig-Lehmann, C., and Illenberger, E. (2009) Int. J. Mass Spectrom., 281, 89. McConkey, J.W., Malone, C.Y., Johnson, P.V., Winstead, C., McKoy, V., and Kanik, I. (2008) Phys. Rep.: Rev. Sec. Phys. Lett., 466, 1. Baccarelli, I., Bald, I., Illenberger, E., and Gianturco, F.A. (2009) Phys. Rep. Rev. Sec. Phys. Lett., in preparation. Huels, M.A., Hahndorf, I., Illenberger, E., and Sanche, L. (1998) J. Chem. Phys., 108, 1309. Abdoul-Carime, H., Gohlke, S., and Illenberger, E. (2004) Phys. Rev. Lett., 92, 168103. Abouaf, R., Pommier, J., and Dunet, H. (2003) Int. J. Mass Spectrom., 226, 397. Ptasinska, S., Denifl, S., Grill, V., Maerk, T.D., Scheier, P., Gohlke, S., Huels, M.A., and Illenberger, E. (2005) Angew. Chem. Int. Ed., 44, 1647. Ptasinska, S., Denifl, S., Mroz, B., Probst, M., Grill, V., Illenberger, E., Scheier, P., and Mark, T.D. (2005) J. Chem. Phys., 123, 124302. Ptasinska, S., Denifl, S., Scheier, P., Illenberger, E., and Maerk, T.D. (2005) Angew. Chem. Int. Ed., 44, 6941. Abdoul-Carime, H., Langer, J., Huels, M.A., and Illenberger, E. (2005) Eur. Phys. J. D, 35, 399. Denifl, S., Sulzer, P., Huber, D., Zappa, F., Probst, M., Mark, T.D., Scheier, P., Injan, N., Limtrakul, J., Abouaf, R., and Dunet, H. (2007) Angew. Chem. Int. Ed., 46, 5238. Hanel, G., Gstir, B., Denifl, S., Scheier, P., Probst, M., Farizon, B., Farizon, M., Illenberger, E., and Mark, T.D. (2003) Phys. Rev. Lett., 90, 4188104. Aflatooni, K., Scheer, A.M., and Burrow, P.D. (2005) Chem. Phys. Lett., 408, 426.

163

164

6 Role of Low-Energy Electrons in DNA Radiation Damage 33. Swiderek, P. (2006) Angew. Chem. Int. 34.

35. 36. 37. 38. 39. 40. 41.

42.

43. 44.

45.

46.

47. 48.

49.

50.

51.

Ed., 45, 4056. Berdys, J., Anusiewicz, I., Skurski, P., and Simons, J. (2004) J. Am. Chem. Soc., 126, 6441. Li, X., Sevilla, M.D., and Sanche, L. (2003) J. Am. Chem. Soc., 125, 13668. Simons, J. (2006) Acc. Chem. Res., 39, 772. Gu, J.D., Wang, J., and Leszczynski, J. (2006) J. Am. Chem. Soc., 128, 9322. Gu, J.D., Xie, Y.M., and Schaefer, H.F. (2006) J. Am. Chem. Soc., 128, 1250. Kumar, A.and Sevilla, M.D. (2007) J. Phys. Chem. B, 111, 5464. Kumar, A. and Sevilla, M.D. (2008) J. Am. Chem. Soc., 130, 2130. Dabkowska, I., Rak, J., and Gutowski, M. (2005) Eur. Phys. J. D, 35, 429. Vilar, M.R., do Rego, A.M.B., Ferraria, A.M., Jugnet, Y., Nogues, C., Peled, D., and Naaman, R. (2008) J. Phys. Chem. B, 112, 6957. Bald, I., Kopyra, J., and Illenberger, E. (2006) Angew. Chem. Int. Ed., 45, 4851. Ptasinska, S., Denifl, S., Scheier, P., and Mark, T.D. (2004) J. Chem. Phys., 120, 8505. Sulzer, P., Ptasinska, S., Zappa, F., Mielewska, B., Milosavljevic, A.R., Scheier, P., Mark, T.D., Bald, I., Gohlke, S., Huels, M.A., and Illenberger, E. (2006) J. Chem. Phys., 125, 044304. Baccarelli, I., Gianturco, F.A., Grandi, A., Sanna, N., Lucchese, R.R., Bald, I., Kopyra, J., and Illenberger, E. (2007) J. Am. Chem. Soc., 129, 6269. Lide D. (ed.) (2002) CRC Handbook of Chemistry and Physics. Bald, I., Flosadottir, H.D., Kopyra, J., Illenberger, E., and Ingolfsson, O. (2009) Int. J. Mass Spectrom., 280, 190. Deng, Z.W., Bald, I., Illenberger, E., and Huels, M.A. (2008) Angew. Chem. Int. Ed., 47, 9509. Levis, R.J., Menkir, G.M., and Rabitz, H. (2001) Science, 292, 709. Brixner, T. and Gerber, G. (2003) Chemphyschem, 4, 418.

52. Mason, N.J. (2008) Int. J. Mass Spec-

trom., 277, 31. 53. Bald, I., Dabkowska, I., Illenberger, E.,

54.

55.

56. 57.

58.

59. 60. 61.

62.

63.

64.

65.

66.

67.

68.

and Ingolfsson, O. (2007) Phys. Chem. Chem. Phys., 9, 2983. Guler, L.P., Yu Y.-Q., and Kenttaemaa, H.I. (2002) J. Phys. Chem. A, 106, 6754. Bald, I. (2007) Low Energy Electron Induced Reactions in Gas Phase Biomolecules – The Role of the Sugar and Phosphate Moiety in DNA Damage PhD thesis, Freie Univerität Berlin, Berlin. Schulz, G.J. (1973) Rev. Mod. Phys., 45, 423. Ibanescu B.C., May O., Monney, A., and Allan, M. (2007) Phys. Chem. Chem. Phys., 9, 3163. Hotop, H., Ruf, M.W., Allan, M., and Fabrikant, I.I. (2003) Adv. Atom. Mol. Opt. Phys., 49, 85. Jordan, K.D. and Wang, F. (2003) Ann. Rev. Phys. Chem., 54, 367. Sommerfeld, T. (2007) J. Chem. Phys. 126, 124301/1. Scheer, A.M., Aflatooni, K., Gallup, G.A., and Burrow, P.D. (2004) Phys. Rev. Lett., 92 068102. Burrow, P.D., Gallup, G.A., Scheer, A.M., Denifl, S., Ptasinska, S., Mark, T., and Scheier, P. (2006), J. Chem. Phys., 124, 124310. Stepanovic, M., Pariat, Y., and Allan, M. (1999) J. Chem. Phys., 110, 11376. Bald, I., Kopyra, J., Dabkowska, I., Antonsson, E., and Illenberger, E. (2007) J. Chem. Phys., 126, 074308. Ptasinska, S. Denifl, S., Gohlke, S., Scheier, P., Illenberger, E., and Maerk, T.D. (2006) Angew. Chem. Int. Ed., 45, 1893. Golovlev, V.V., Allman, S.L., Garrett, W.R., Taranenko, N.I., and Chen, C.H. (1997) Int. J. Mass Spectrom., 169/170, 69. Perez, J., Petzold, C.J., Watkins, M.A., Vaughn, W.E., and Kenttamaa, H.I. (1999) J. Am. Soc. Mass Spectrosc., 10, 1105. Shea, R.C., Petzold, C.J., Campbell, J.L., Li, S., Aaserud, D.J., and

References

69.

70.

71.

72.

73. 74.

Kenttaemaa, H.I. (2006) Anal. Chem., 78, 6133. Bald, I., Dabkowska, I., and Illenberger, E. (2008) Angew. Chem. Int. Ed., 47, 8518. Wang, X.-B., Vorpagel, E.R., Yang, X., and Wang, L.-S. (2001) J. Phys. Chem. A, 105, 10468. Konig, C., Kopyra, J., Bald, I., and Illenberger, E. (2006) Phys. Rev. Lett., 97, 018105. Burrow, P.D., Gallup, G.A., and Modelli, A. (2008) J. Phys. Chem. A, 112, 4106. Gohlke, S. and Illenberger, E. (2002) Europhys. News, 33, 207. Zheng, Y., Hunting, D.J., Ayotte, P., and Sanche, L. (2008) Phys. Rev. Lett., 100, 4198101.

75. Ray, S.G., Daube, S.S., and Naaman, R.

76. 77. 78. 79.

80.

81.

(2005) Proc. Natl. Acad. Sci. U.S.A., 102, 15. Pan, X. and Sanche, L. (2006) Chem. Phys. Lett., 421, 404. Ptasinska, S. and Sanche, L. (2007) Phys. Rev. E, 75, 031915. Solomun, T. and Skalicky, T. (2008) Chem. Phys. Lett., 453, 101. Prabhudesai, V.S., Kelkar, A.H., Nandi D., and Krishnakumar, E. (2005) Phys. Rev. Lett., 95, 143202. Prabhudesai, V.S., Nandi, D., Kelkar, A.H., and Krishnakumar, E. (2008) J. Chem. Phys., 128, 154309. Maksymovych, P., Sorescu, D.C., Jordan, K.D., and Yates, J.T. (2008) Science, 322, 1664.

165

167

7 Structure-Based Design on the Way to New Anti-Infectives Anna Katharina Herta Hirsch

7.1 Introduction

With an estimated 300–500 million new infections and three million deaths annually, malaria and tuberculosis undoubtedly still pose a major health concern [1]. The need for the development of novel therapeutic approaches is ever-growing in light of the emergence of multi-drug-resistant parasites [2]. How can a new drug be identified? First, a lead compound, that is, a molecule with a promising biological activity that does not yet fulfill all the requirements but represents the starting point on the way to a new drug, has to be identified. A number of different strategies exist to achieve this goal: • High-throughput screening (HTS) of a large library of small molecules is of particular interest in cases in which no structural information or characterization of the biological target is available. The majority of lead compounds still comes from hits identified by HTS [3]. • Virtual screening has established itself as an alternative or complementary approach to classical HTS. Potentially active and/or drug-like compounds are selected from a library of compounds, using elaborate docking and scoring functions [4]. • Combinatorial chemistry is useful for the formation of large, small-molecule libraries. However, this approach is less effective for generating a great deal of structural diversity. • Nature can be of help by providing a rich and diverse source of structural inspiration [5]. The scaffold of a natural product, displaying interesting biological properties, could be developed into a new drug. • Structure-based design, a comparatively new field, has established itself in pharmaceutical research as a valuable alternative to traditional screening; the X-ray crystal structure of a target enzyme is used as a basis for lead compound identification and optimization. The increasing number of leads, identified and/or optimized using this rational approach, used for the development of new drugs illustrates Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

168

7 Structure-Based Design on the Way to New Anti-Infectives

this fact [6]. Until now, this strategy has been mainly used in the later stages of lead optimization. The last mentioned strategy will be employed for the purposes of this project. The aim of this approach is to identify synthetically accessible target molecules, with optimal stereoelectronic properties that are complementary to the binding site of the target enzyme and show minimal or no repulsive interactions when complexed to the enzyme. The identification of promising leads is aided by the medicinal chemist’s understanding of molecular recognition. While hydrophobic interactions between a lead and an enzyme are the main driving force for complexation, H-bonding interactions account for selectivity (Figure 7.1). Later stages of drug development are not the subject of this chapter. To determine the biological activity of a potential lead compound, a new biological target [7], that is, an enzyme or receptor that upon interference by the ligand/drug has an impact on the disease causing pathogen in the desired way, has to be identified. In an ideal case, the target is essential for the pathogen and not present in humans, thereby precluding any selectivity issues. Isoprenoids are an essential class of natural products, requiring the essential precursors isopentenyl diphosphate (IPP, 1) and dimethylallyl diphosphate (DMAPP, 2) for their biosynthesis (Scheme 7.1). Until recently, only one route to the universal isoprenoid precursors was known, the so-called mevalonate pathway, using acetyl-coenzyme A as the only building block [8]. A completely distinct alternative to this well-established biosynthetic route, now known as the nonmevalonate pathway, was discovered in the early 1990s, starting from pyruvate (3) and glyceraldehyde 3-phosphate (4) [9]. Interestingly, this biosynthetic pathway is exclusively used by a number of pathogens such as the malarial parasite Plasmodium falciparum and the tuberculosis-causing Mycobacterium tuberculosis and not by higher eukaryotes (e.g., humans), which means that inhibition of the constituent enzymes of the nonmevalonate pathway affects and kills only the parasites, leaving the patient untouched (Scheme 7.1). Thus, this pathway has provided a rich source of new, highly attractive drug targets. To illustrate the use of a structure-based design cycle, the development of the first inhibitors of the kinase IspE of the nonmevalonate pathway is described below, constituting a novel approach toward anti-infectives. First, the biosynthetic pathway

Target enzyme

Inhibitor

Key requirements: Steric complementarity Electronic complementarity Noncovalent interactions

Figure 7.1

Schematic representation of structure-based inhibitor design.

7.2 Isoprenoids and the Nonmevalonate Pathway

O

O S

O CO2−

CoA

Acetyl CoA

+

H

3

Pathogens: nonmevalonate pathway

Humans: Mevalonate pathway

OPP

OP OH 4

+

OPP 2

1

Target: block the kinase IspE

Isoprenoids Scheme 7.1 Schematic representation of the biosynthesis of the isoprenoid precursors IPP (1) and DMAPP (2). P = phosphate group.

is briefly presented with an overview of the known inhibitors of the constituent enzymes. Subsequently, the chosen target, the kinase IspE, is introduced and its structure is discussed in some detail. This will provide the basis for the rationale of the design of the first-generation inhibitors. Second, the active site is explored by showing the possibility of applying concepts from supramolecular chemistry to an enzymatic context. Finally, the development of water-soluble inhibitors is described with the aim of obtaining an X-ray cocrystal structure to verify the proposed binding mode.

7.2 Isoprenoids and the Nonmevalonate Pathway

There are more than 35 000 known isoprenoids, which fulfill a myriad of important biological functions. Despite their striking structural diversity, all isoprenoids are biosynthesized from the two simple five-carbon building blocks 1 and 2 (Scheme 7.1). This concept is also known as the isoprene rule [10]. The nonmevalonate pathway starts with the head-to-tail condensation of the two- and three-carbon precursors 3 and 4. A total of seven enzymes catalyze the conversion of these starting materials into the essential isoprenoid precursors 1 and 2. Fosmidomycin, an inhibitor of the second enzyme of the pathway (IspC), has been shown to cure malaria in rodents, thereby validating the constituent enzymes as drug targets [11]. This finding triggered research efforts aimed at the elucidation of the structures

169

170


and mechanisms of the participating enzymes. As a result, detailed structural and mechanistic data exist for most enzymes [12], setting the stage for lead generation by structure-based design. A number of enzymes of the nonmevalonate pathway have been chosen as targets to achieve this goal. The rather hydrophilic nature of the active sites renders the development of low-molecular-weight inhibitors challenging. Thus, it comes as no surprise that the few reported inhibitors either bear phosphate or phosphonate groups or display rather modest inhibition (Table 7.1) [13]. 7.2.1 4-Diphosphocytidyl-2C-methyl-D-erythritol Kinase (IspE)

The absence of known inhibitors, the fact that the kinase IspE belongs to the nonmevalonate pathway, and the availability of an X-ray crystal structure make the fourth enzyme of the pathway an ideal target for rational design of potent, drug-like inhibitors without the use of the problematic phosphate or phosphonate moieties. IspE (EC 2.7.1.148) employs adenosine 5 -triphosphate (ATP) and Mg2+ cations for the phosphorylation of the C(2)-hydroxyl group of 4-diphosphocytidyl2C-methyl-d-erythritol (5) to afford 4-diphosphocytidyl-2C-methyl-d-erythritol 2-phosphate (6, Scheme 7.2) [18]. This central reaction is the only ATP-dependent step of the whole biosynthetic pathway. Sequence comparisons have shown that IspE belongs to the galactose/ homoserine/mevalonate/phosphomevalonate (GHMP) kinase superfamily [19]. 7.2.2 Structure of IspE

In 2003, the first crystal structures of IspE from Thermus thermophilus and Escherichia coli were solved, of the apoenzyme and of a ternary complex, respectively [20, 21]. Recently, the crystal structure of Aquifex aeolicus IspE was solved as a complex with a number of natural ligands [22], a synthetic substrate mimic [22], and synthetic cytidine [23] as well as cytosine (Section 7.4) [24] derivatives. To date, no crystal structure is available for IspE from a pathogen, for example, M. tuberculosis or P. falciparum. IspE generally crystallizes as a homodimer with each monomer displaying the characteristic two-domain fold of the GHMP kinase superfamily that consists of an ATP- and a substrate-binding domain. The dimer clasps around a solvent-filled channel, featuring two active sites at either end (Figure 7.2). 7.2.3 Active Site of IspE

The active site of E. coli IspE was used for modeling. It can be divided into three main pockets: the ATP-, the cytidine 5 -diphosphate (CDP)-, and the methylerythritol (ME)-binding pockets. Molecular modeling, using the program MOLOC

R2

N

O

N

R4 R3

Cl

O

OH

N

F

H

O

PO32−

CO2−

n

CDP 2 X+ O2 S N H

+ 2 CDP 2 X

Substrate and transition-state analogues

O

R

1

O

O

H N

2

R

OH

O

N

N

R3

N

N

R4

H N

NH

Mechanism-based inhibitors

NH

4

HN

HO

N NH2

IC50 = 0.45 mM [16] Ki = 0.011 µM [17]

Ki = 0.9 nM [15]

IC50 = 0.08 mM [14]

Inhibition

The inhibition of the best inhibitor is given. CDP = cytidine 5 -diphosphate; DXS and IspF: first and fifth enzyme of the nonmevalonate pathway, respectively; IC50 = concentration of inhibitor at which 50% maximum initial velocity is observed; Ki = inhibition constant.

NH2 O

No known inhibitors at the onset of the present design cycle NMe2

R1

O

O

O

Inhibitors

Known inhibitors of enzymes of the nonmevalonate pathway.

IspF IPP isomerase

IspC IspE

DXS

Enzyme

Table 7.1

7.2 Isoprenoids and the Nonmevalonate Pathway 171

172


NH2

DXS O

O

IspC

+

OP

H

CO2−

IspD

OH

3

1 OH

4

N

HO 2

3

O

4

O

O P O P O O−

OH

O

N

O

O− 5

HO

OH

ATP IspE ADP

OPP 1

NH2

IspF IPP isomerase

IspG IspH

N

PO

OH

O OH

O

O P O P O O− O−

OPP 6

2 Scheme 7.2

Simplified version of the nonmevalonate pathway.

AMP-PNP

5 Figure 7.2 Schematic representation of the ternary complex of E. coli IspE, 5 -adenyl-β, γ -amidotriphosphate (AMP-PNP) and the substrate 5, cocrystallized as a homodimer (Protein Data Bank (PDB) code: 1OJ4) [20].

O

HO

N

OH

O

7.2 Isoprenoids and the Nonmevalonate Pathway Small, hydrophobic pocket

ME-binding pocket

Leu15

Leu28

Lys10 H 2N

Asn12

Lys96 Asn10

O H2N

Phe185

+

His26 O NH

Val157 O

H3N+

N

−O

Pro182

NH

Thr181 OH N

O

O

S

NH3 O

Val156 Asp141 Met100

H2N H N

HN Tyr25

173

Leu66

OH CDP-binding pocket

ATP-binding pocket

Figure 7.3 Schematic representation of the four pockets of the active site of E. coli IspE. The residues of the glycine-rich loop were omitted for clarity.

[25], revealed that an additional small, hydrophobic pocket lies adjacent to the CDP-binding site (Figure 7.3). The ATP-binding pocket features a glycine-rich phosphate-binding loop, typically displaying an adjacent positively charged N terminus of an α helix [20]. The adenine moiety is accommodated in a hydrophobic cleft lined by Val57, Val60, Leu66, Ile67, Lys96, and Met100. Numerous hydrophobic contacts offered by this pocket certainly make a large contribution to the binding enthalpy, as predicted by MOLOC. Additional stabilization is derived from a network of H bonds to the nucleobase moiety. The ribose moiety of 5 -adenyl-β, γ -amidotriphosphate (AMP-PNP) is solvent-exposed and does not show any contacts with the protein. The CDP-binding pocket accommodates the cytosine moiety in a π sandwich, consisting of the side chains of Tyr25 and Phe18 held in place by – stacking interactions. The ribose moiety benefits from stabilization by a pseudo-π sandwich that is composed of the aromatic side chain of Tyr25 and the aliphatic Pro182. A pseudo-π sandwich can be defined as a π sandwich, in which one of the two aromatic rings is replaced by an aliphatic ring. His26 is a key residue for the recognition of cytosine, which involves a total of three H-bonding interactions. The ribose and phosphate groups are stabilized by a number of solvent-mediated interactions and a H bond from the side chain of Tyr25 (Figure 7.4). IspE was shown to have high substrate specificity [21]. Hence, the cytosine moiety must play a key role in substrate recognition and thus selectivity, providing the starting point for the structure-based design of the first-generation inhibitors (Section 7.3).

lle67 Asn65 Val160

174


Asn12

Leu15

Lys10

Leu28

Asp141

Phe185 His26

Val156 Thr181 Pro182 (a)

Tyr25

(b)

Figure 7.4 (a) CDP- and ME-binding and the small, hydrophobic pockets of E. coli IspE. (b) Active site of E. coli IspE (PDB code: 1OJ4) [20]. Color code: protein skeleton: C: gray; skeleton of 5: C: green; O: red; N: blue; P: orange.

7.3 Targeting the CDP-Binding Pocket of IspE 7.3.1 Design

Careful examination of the active site revealed the CDP-binding pocket to be more attractive as a target for inhibitor design than the other binding pockets for a number of reasons. A perfect setup for a double π sandwich and other recognition features should endow potential inhibitors with higher selectivity. Furthermore, given the myriad of proteins that use ATP as a cofactor, inhibitors designed to target the ATP-binding pocket bear a high risk of selectivity issues. Given that neither substrate nor cofactor bind to the adjacent, hydrophobic pocket, presumably, both affinity and selectivity could be gained by occupying it. As this small cavity lies adjacent to the CDP-binding pocket, potential inhibitors should be designed to occupy both the CDP-binding pocket and the newly discovered subpocket, leaving the hydrophilic ME- as well as the ATP-binding pockets unoccupied. In a first round of design, cytosine was chosen as a central scaffold to position potential inhibitors in the CDP-binding pocket. According to modeling, the nucleobase moiety was predicted to be sandwiched between Tyr25 and Phe185. By analogy to the natural substrate 5, the cytosine moiety was postulated to be able to form H bonds to His26. It was concluded that the central platform should be decorated with a suitable ribose analogue at the N(1) position, which should be held in place by the pseudo-π sandwich and a vector designed such that its

7.3 Targeting the CDP-Binding Pocket of IspE

175

substituent would be placed in the hydrophobic subpocket. If this were to be achieved, numerous hydrophobic interactions would result. By connecting the linker to the C(5) position, it should serve to address the catalytically essential residues Lys10 and Asp141. Because of the modular design, the different components, that is the ribose analogue, the vector, and the central scaffold, should be easy to vary and optimize (Figure 7.5a). 7.3.1.1 Possible Ribose Analogues As ribose analogues, both heteroalicyclic and aromatic rings can be envisaged to fill the space provided by the pseudo-π sandwich (Figure 7.5b, compounds of type 7 and 8). Introduction of a saturated ring featuring a sulfur atom, for example, a tetrahydrothiophenyl ring, presumably would enable an additional sulfur–aromatic interaction with the phenolic ring of Tyr25 [27]. Modeling predicted both enantiomers of a tetrahydrothiophenyl derivative to bind with similar strength due to the conformational flexibility of five-membered rings. Four questions need to be answered regarding the choice of the ribose analogue: (i) Which is the ideal ring size? (ii) Is an aromatic or an aliphatic system favored? (iii) Are heteroatoms beneficial? (iv) Is a connecting methylene linker beneficial or detrimental to affinity? NH2 NH2 4 N 5 2 N 6 O

CDP-binding pocket

R2

N N R1

O

Small, hydrophobic pocket

R1 = ribose analogue

Inhibitor R2 Ribose analogue

(a)

O NH2 N O

N H

(c)

clogP

(±)-9 Alicyclic amine 10 OH

O S

R

O NH2

2

O

(b)

Saturated cyclic Methylene bridged

S

Inhibitor R

clogPa 0.3 to 1.0 −0.4 to 1.5

Figure 7.5 (a) Modular design of first-generation inhibitors. (b) Ribose analogues, (c) different vectors, and (d) sulfone substituents envisaged for first-generation inhibitors. a clogP values (calculated

O S

N

R2 = Et, c -Pr, CH2CF3

7 8

N H

N

N R1

Inhibitor R1

1.1 to 2.3 1.2

(d)

(±)-11 Acyclic (±)-12 Cyclic

clogP 0.3 to 2.0 −0.03 to 2.6

partitioning coefficient) were calculated with the program ACD/LogP [26]. The tetrahydrothiophenyl and the tetrahydrofuranyl derivatives of type 7 and 8, respectively, are chiral.

R

176


7.3.1.2 Design of the Vector For the vector, a propargylic sulfonamide substituent at the C(5) position of cytosine was envisaged, which displays three attractive features. First, the alkyne ensures a certain rigidity and linearity. Second, N-substituted sulfonamides are known to prefer a conformation in which the lone pair of the nitrogen atom bisects the O–S–O angle, resulting in a staggered arrangement [28]. In its preferred conformation, the sulfonamide is expected to form ionic H bonds to the side chains of Lys10 and Asp141. Finally, small complementary sulfone substituents (R2 = Et, Figure 7.5a), should orient directly into the subpocket. However, before this part of the vector is optimized (Section 7.3.4), the importance of the sulfonamide’s contribution to affinity should be evaluated (Section 7.3.3). An exemplary inhibitor of type 7 features a tetrahydrothiophenyl ring as a ribose mimic (R1 ) and an ethyl group as a sulfone substituent (R2 ) (Figure 7.5b). According to modeling, the ligand should benefit from H bonds to His26, Lys10, and Asp141 as well as from the postulated sulfur–aromatic interaction (Figure 7.6a). An overlay of the natural substrate 5 and the potential inhibitor showed the sulfone moiety to be almost perfectly superimposed with the C(2)-hydroxyl group that is to be phosphorylated (Figure 7.6b). The synthesis of the first representative of the first-generation inhibitors was achieved using a convergent strategy based on the Sonogashira cross-coupling reaction [29, 30]. With the synthetic route in place, the optimization of both modules of the inhibitors could be initiated in parallel. 7.3.2 Optimization of the Ribose Analogue

It was envisaged to introduce a number of different ribose analogues (Figure 7.5b). Different ring types had to be tested: saturated, aromatic, or heteroaromatic rings, all of which should be directly attached to the N(1) of the nucleobase (ligands of type 7). Furthermore, the influence of a connecting methylene group between N(1) Asn12

Leu15 Leu28 Phe185

His26 2.9

2.9

Phe32 3.1 2.9

3.0

Lys10 Asp141

Pro182 3.4 (a)

Val156 Tyr25

(b)

Figure 7.6 (a) MOLOC-generated molecular model of the exemplary inhibitor in the active site of E. coli IspE (PDB code: 1OJ4) [20]. (b) Superposition of the potential inhibitor and the natural substrate 5. Color code: S: yellow; C skeleton of the inhibitor: cyan.


of cytosine and the ribose analogue needed to be evaluated both for aliphatic and (hetero-)aromatic substituents (compounds of type 8). In addition, it was deemed important to test an acyclic ribose mimic, featuring an ester group. Replacement of the sulfur atom of the tetrahydrothiophenyl ring of the exemplary inhibitor by a methylene moiety to afford a cyclopentyl ring as a ribose analogue was expected to yield information on the postulated sulfur–aromatic interaction. Synthesis of this set of potential inhibitors was achieved using similar synthetic routes [24, 29]. Using the established photometric assay to determine the activity of potential inhibitors against IspE, the IC50 and Ki values for all new ligands were determined (Table 7.2) [24, 29, 31]. A competitive mechanism with respect to substrate binding was assigned to most ligands. In a few cases, however, a mixed competitive (Kic )–uncompetitive (Kiu ) mode of inhibition was found. In agreement with the observation that small cytosine derivatives possess remarkably low water solubility [24], none of the inhibitors showed high water solubility, despite the low calculated partitioning coefficient (clogP) values (Figure 7.5b). The results obtained are highly satisfactory: First, as no inhibitors of IspE had been described at the onset of this design cycle, obtaining the first active compounds with Ki values in the upper-nanomolar range constitutes an important achievement. Second, inhibition was possible in the absence of a phosphate or phosphonate group. The ligands described constitute the first example of potent, drug-like inhibitors of an enzyme of the nonmevalonate pathway. Furthermore, variation of the ribose mimic has a clear effect on affinity, that is, structure–activity relationships (SARs) could be observed. A methylene-bridged tetrahydrofuranyl ring is the poorest ribose substitute with a double-digit micromolar Ki value. Introduction of a methylene-bridged aromatic ring (benzyl) or an open alkyl chain bearing an ester moiety as a ribose mimic clearly also has a negative effect on affinity. However, it seems that a methylene-bridged ring is tolerated as long as it is not too bulky: the methylene-bridged cyclobutyl and pyrazolyl derivatives feature similar affinities to that of the inhibitor bearing a cyclopentyl ring directly attached to N(1) of cytosine. On the basis of the affinities, there does not seem to be any Inhibitory activities (E. coli IspE) of compounds of type 7 and 8.

Table 7.2

Ribose analogue

Kic (µM)a

Tetrahydrothiophenyl Cyclopentyl Cyclobutylmethyl CH2 –3-pyrazolyl Benzyl CH2 CO2 Et CH2 –2-tetrahydrofuranyl

0.29 ± 0.1 1.5 ± 0.2 1.5 ± 0.2 1.6 ± 0.1 3.7 ± 0.5b 4.2 ± 0.6c 32.3 ± 2.8

a The

IC50 values can be found in 24, 29, 30. inhibition: Kiu = 23.5 ± 7.1 µM. c Mixed inhibition: K = 21.6 ± 6.2 µM. iu b Mixed

177

178


significant difference between a heteroaromatic ring such as pyrazolyl and the alicyclic cyclopentyl ring. However, if clogP values are taken into account, the latter inhibitor is significantly more lipophilic; hence, the resulting more favorable partitioning from the aqueous buffer into the less polar protein environment could partially account for its affinity. This implies that heteroatoms afford some affinity. Finally, the tetrahydrothiophenyl ring is the best ribose mimic, affording the lowest Ki value. Presumably, its improved affinity could be ascribed to the favorable sulfur–aromatic interaction. This interaction was quantified when comparing the affinities of the tetrahydrothiophenyl derivative with an inhibitor bearing the same sulfone substituent (cyclopropyl) and a cyclopentyl ring as a ribose analogue. The affinity is lowered by a factor of five, corresponding to a free-enthalpy increase of G300 K ≈ 1 kcal mol−1 , in agreement with published data [27]. Both rings benefit from stabilization by the pseudo-π sandwich. Thus, the higher affinity could be explained by favorable interactions of the phenolic side chain and the sulfur atom of the sandwiched ring. 7.3.3 Importance of the Vector

Before optimizing the propargylic sulfonamide vector, its significance has to be confirmed. According to modeling, the sulfonamide group forms two H bonds to Lys10 and Asp141. To verify this prediction experimentally, four derivatives featuring modified vectors and established ribose analogues were designed and synthesized (Figure 7.5c): (i) The N-methylated derivative of the most potent inhibitor, featuring a cyclopropyl ring as sulfone substituent and a tetrahydrothiophenyl ring as ribose analogue, served to quantify the importance of the ionic H bond between the sulfonamide NH and the side chain of Asp141. (ii) Different propargyl amine derivatives [compounds of type (±)-9] as well as propargyl alcohol derivative 10 were envisaged to evaluate the effect of substituting the sulfonamide altogether. The set of target molecules were synthesized following similar routes [24]. Using the established enzymatic assay, the IC50 and Ki values of the potential inhibitors were determined [24, 29, 31]. N-methylation of the sulfonamide nitrogen atom clearly affected affinity, illustrating the importance of the postulated ionic H bond. Modeling showed the inhibitor to be still accommodated in the active site without any repulsive interactions but with one less H bond. Comparison with the most potent inhibitor enabled the quantification of the postulated ionic H bond, given that both ribose analogue and sulfone substituent remained unchanged. The inhibitory potency was reduced by a factor of nearly 10 upon N-methylation (Ki = 2.5 µM); thus, a contribution of up to G300 K = 1.3 kcal mol−1 to the overall binding free enthalpy could be ascribed to this H bond alone (Figure 7.7a). As expected, substitution of the sulfonamide moiety by heterocyclic amines led to a decrease in inhibitory potency. A piperidyl-substituted derivative maintains the highest affinity (Ki = 4.7 µM). A possible explanation could be an ionic H bond between the side chain of Asp141 and the piperidinium residue (Figure 7.7b). According to modeling, the piperidinyl ring should be located at the entrance of the

7.3 Targeting the CDP-Binding Pocket of IspE Leu15 Asn12

Leu28 His26

Phe185

Leu28 3.2

Lys10

2.9

Phe32

His26 Phe185

Phe32 2.9

Asn12

Leu15

2.9

2.8

2.9

3.2

3.1 2.9

Lys10

Pro182 Pro182 (a)

Val156 Tyr25

Val156 Asp141

Asp141 (b)

Tyr25

Figure 7.7 MOLOC-generated molecular model of (a) the N-methylated and (b) the piperidinyl-substituted inhibitors in the active site of E. coli IspE (PDB code: 1OJ4) [20]. Reproduced with permission from the Royal Society of Chemistry.

small, hydrophobic pocket. In this way, a number of hydrophobic interactions with the residues lining this cavity are still possible. A pyrrolidinyl-substituted derivative was proposed to bind in a similar manner, displaying fewer hydrophobic contacts because of its smaller ring. As a result, a decrease in binding affinity was observed (Ki = 11.8 µM). The alcohol 10 shows very weak affinity for IspE; consequently, no Ki value could be determined. This observation could be explained by the lack of both a sulfonamide moiety and an alkyl substituent to benefit from potential interactions in the small, hydrophobic cavity. In summary, important contributions of the sulfonamide moiety to the observed binding affinity were clearly confirmed through this set of derivatives. With the sulfonamide moiety validated as a good vector, the next logical step was therefore the optimization of the sulfone substituent R (Figure 7.5d). 7.3.4 Optimization of the Filling of the Small, Hydrophobic Pocket

Inspired by the finding that optimal volume occupancy can be an important contributor to affinity [32], the small, hydrophobic pocket of IspE was carefully examined. For this purpose, a series of derivatives of type (±)-11 and (±)-12 were designed, differing only in the sulfone substituent as shown in Figure 7.5d. By keeping the scaffold, the vector and the ribose analogue constant, the effects of different substituents on affinity should be directly comparable. The target molecules were readily synthesized [24]. The choice of substituents was guided by the predictions made by molecular modeling and a concept from conventional supramolecular chemistry – the ‘‘55% rule’’ – that was recently applied to enzymes for the first time [33]. 7.3.4.1 The ‘‘55% Rule’’ Investigation of the optimal volume occupancy of the cavity space confined by capsular synthetic receptors by Mecozzi and Rebek led to the ‘‘55% rule,’’ stating

179

180


that the most stable inclusion complex forms if 55 ± 9% of the apolar space is occupied by the guest [34]. This concept holds true, from synthetic supramolecular chemistry [35]. Recently, Zürcher et al. applied it to the filling of a hydrophobic cavity in the active site of the antimalarial target plasmepsin II [33]. The van-der-Waals interactions in the cavity are not ideal at smaller packing coefficients (PCs). At higher PCs, however, large entropic losses, resulting from a decrease in the mobility of the binding partners, counteract enthalpic gains. The small, hydrophobic pocket of E. coli IspE was estimated to be around 100 A˚ 3 by filling the pocket with a hydrocarbon network [34, 36]. Thus, the PCs were calculated for different target molecules, which guided their design. 7.3.4.2 Evaluation of Inhibitors Featuring Different Sulfone Substituents A series of potential inhibitors were subjected to the enzymatic assay. The inhibitory activities as well as the calculated PCs for selected compounds are summarized in Table 7.3. A number of derivatives prepared feature Ki values in the nanomolar range. In general, small alkyl chains with a maximum chain length of two carbon atoms or cyclic alkyl substituents up to a ring size of five are well suited to fill the cavity; hence, it is presumably more flexible on the sides than at its bottom. The three-membered ring seems to be ideal to fill this lipophilic pocket. A 2,2,2-trifluoroethyl group almost affords the same affinity as the most potent inhibitor (featuring a cyclopropyl ring)

Inhibitory activities (E. coli IspE) and PCs of compounds of type (±)-11 and (±)-12.

Table 7.3

Sulfone substituent

µM) Kic (µ

PC (%)

Cyclopropyl 2,2,2-Trifluoroethyl Isopropyl Cyclobutyl Ethyl Cyclopentyl 1,1,1-Trifluoromethyl sec-Butyl n-Hexyl Cyclohexyl Methyl n-Butyl n-Propyl Phenyl

0.29 ± 0.1 0.36 ± 0.1 0.52 ± 0.1 0.56 ± 0 0.64 ± 0.1 0.89 ± 0.1a 1.2 ± 0.3 1.8 ± 0.3 2.0 ± 0.3 2.5 ± 0.4b 2.6 ± 0.1 8.0 ± 0.1 8.2 ± 1.7c 16.3 ± 1.0

56 n.d. 62 69 45 83 n.d. n.d. 107 97 28 76 61 n.d.

inhibition: Kiu = 19.0 ± 9.2 µM. inhibition: Kiu = 67.7 ± 35 µM. c Mixed inhibition: K = 27.3 ± 11 µM. iu n.d. = not determined. a Mixed

b Mixed


and is nearly twice as strong as the corresponding ethyl derivative (Figure 7.6a). The affinity of isopropyl- and cyclobutyl-substituted ligands is very similar to that of the ethyl-substituted derivative; however, the alkyl residues seem to be slightly too large for optimal volume occupancy, leading to a decrease in affinity. As the pocket is not properly filled by ligands with a smaller alkyl substituent (methyl), the binding affinity is reduced. Evaluation of the set of derivatives in terms of their PCs showed binding affinity to correlate with volume occupancy: a cyclopropyl ring has a PC of 56% and the lowest Ki value (Ki = 0.29 µM, Figure 7.8a). Lower (28% for methyl and 45% for ethyl substituents) or higher PCs (62% for isopropyl and 69% for cyclobutyl) are mirrored by weaker inhibition. A slight increase in size of the sulfone substituent – that is, extension of the ethyl substituent by one or two carbon atoms – led to a strong decrease in affinity. According to modeling, the n-propyl substituent might still fit into the cavity at the cost of adopting the energetically less favorable gauche conformation (PC 61%). The n-butyl substituent, however, cannot be accommodated by the pocket, even when contorted. Thus, the propargylic sulfonamide linker could equally well undergo a conformational change to direct this substituent out of the pocket into solvent-exposed space. Larger substituents (PC > 80%), such as cyclopentyl, cyclohexyl, or n-hexyl, were predicted to direct their alkyl substituents toward the opposite direction, that is, toward the solvent (Figure 7.8b). The increased lipophilicity (cf. clogP values) and the resulting more favorable partitioning to the less polar protein environment from the aqueous buffer could explain the increased binding affinity. In summary, by exploring the small, hydrophobic pocket, the affinity of the inhibitors was improved and another example for the application of the ‘‘55% rule’’ to an enzymatic context was provided.

Lys10 Phe185 Leu15 Asp141 (a)

Leu28

(b)

Figure 7.8 (a) van-der-Waals surfaces of the cyclopropyl ring of the most potent inhibitor of type (±)-11 and the protein in the small, hydrophobic pocket. (b) MOLOC-generated molecular model of the alkyl-substituted inhibitors in the active site of E. coli IspE

(PDB code: 1OJ4) [20]. Color code: C skeleton of the inhibitors: methyl, cyan; ethyl, magenta; n-propyl, green; n-hexyl, light pink.Reproduced with permission from the Royal Society of Chemistry.

181

182


7.3.5 Summary of the First-Generation Inhibitors

In conclusion, addressing the CDP-binding pocket was rewarded with the discovery of the first, potent, small-molecule inhibitors of IspE. The affinity could be improved through rational modifications of the lead compound, giving access to an extensive set of SARs; for the time being, they provide the only indication that the proposed binding mode is correct. X-ray-crystallographic studies had to be performed to validate this hypothesis (Section 7.4).

7.4 X-ray Cocrystal Structure Analysis

Efforts were undertaken to obtain a cocrystal structure of the first-generation inhibitors and IspE. The kinase is very sensitive to crystallization conditions; even traces of organic cosolvents are sufficient to preclude crystal growth. Presumably, because of the lack of water solubility of the first-generation inhibitors, no cocrystal structures could be obtained. Thus, a water-soluble derivative had to be designed (Section 7.4.1). Provided it resembled the structure of the inhibitors in hand, it could prove the binding mode of this class of compounds. 7.4.1 Design of Water-Soluble Inhibitors

Modeling as well as clogP values were used to guide the selection of promising target molecules (Figure 7.9). To improve the water solubility of the first-generation inhibitors without a concomitant decrease in affinity, the sulfone substituent or the sulfonamide vector (target molecules of type (±)-13) or the ribose analogue (ligands of type 14) could be modified. Given that modestly potent inhibitors could be cocrystallized (Section 7.2.2), the proposed series of target molecules looked promising as long as water solubility could be achieved. Two derivatives were obtained by varying NH2

O

Inhibitor (±)-13 14

R2

N R1

R1

R2

clogP

2-Tetrahydrothiophenyl Carboxylic-acid-based

Morpholinyl based Cyclopropyl

–0.49 to 0.9 –0.6 to 0.4

Figure 7.9 Target molecules to improve the water solubility of the first-generation inhibitors.

7.4 X-ray Cocrystal Structure Analysis

the vector, whereas the remaining three resulted from a modification of the ribose analogue. Introduction of a carboxylic acid or the corresponding ester functionality, a morpholinyl, or an oxetanyl substituent should afford the desired property. Oxetanes, in particular, were recently described to endow compounds with improved physicochemical properties, namely, affording enhanced solubility and decreased lipophilicity [37]. 7.4.2 Enzyme Assays of Inhibitors Designed to be Water Soluble

Among the six compounds specifically prepared to obtain water-soluble inhibitors, two (the morpholinyl-substituted sulfonamide (Ki = 13.1 µM) and the oxetanyl derivative (compound 15, Figure 7.10b, Ki = 28.7 µM)) do not require addition of dimethyl sulfoxide as a cosolvent to perform the enzyme assay [24]. While the former still requires ethanol as a cosolvent, the latter is water soluble, setting the stage for X-ray crystallographic studies (Section 7.4.3). Water solubility was achieved at the expense of potency: the affinity was decreased by nearly two orders of magnitude, when replacing the tetrahydrothiophenyl ring of the most potent inhibitor by the oxetanyl substituent of inhibitor 15. Because of their poor affinities, only the IC50 values of the carboxylic acid derivatives were determined (upper micromolar range). The weak inhibitory potency of the morpholinyl derivative lacking the sulfonamide moiety could be explained by the bad match between the electronics of the morpholinyl moiety and the small, hydrophobic subpocket. Thus, introduction of an oxygen atom to the piperidinyl derivative of type (±)-9 (Section 7.3.3), affording the corresponding morpholinyl-substituted ligand, resulted in a decrease in affinity by almost a factor of nine. A similar decrease in affinity was observed when changing the sulfone substituent from cyclohexyl (compound of type (±)-12) to morpholinyl. The poor affinity of the carboxylic acid derivatives could be attributed to the less than optimal ribose mimics.

Lys145

Tyr31 O

2.9 3.2

His25 3.1 3.2

3.0

3.0 2.6

NH2

Asn11

2.3 Asp130

O

N O

(a)

N H

N

(b)

Figure 7.10 (a) Binding mode of inhibitor 15 in the CDP-binding pocket of active site B (PDB code: 2VF3) [24]. (b) Structure of compound 15.

CO2Et 15

O S

183

184


7.4.3 Structural Analysis

The structure of the oxetanyl derivative 15 in complex with A. aeolicus IspE was determined to 2.2 A˚ resolution (PDB code: 2VF3) [24]. Two molecules are present in the asymmetric unit, featuring ‘‘active sites A and B.’’ Evidence for the proposed binding mode of inhibitor 15 was provided by the cocrystal structure. The water-soluble ligand indeed binds in the CDP-binding pocket. As a result of the differences in amino-acid sequences of A. aeolicus and E. coli IspE [30], the sulfonamide and nucleobase moieties form somewhat different H-bonding patterns compared to those predicted by modeling. In particular, two additional H bonds stabilize the cytosine moiety of ligand 15 (Figure 7.10a). The cytosine moiety is engaged in four H bonds to the side chain and backbone amide of His25 in active site B. The amino group forms an additional H bond to the backbone amide C=O of Lys145. The cytosine ring is sandwiched by the side chains of Tyr24 and Tyr175 (= Phe185 in E. coli IspE). A network of H bonds to the side chains of Asn11, Tyr31, and Asp130 stabilizes the sulfonamide moiety (Figure 7.10a). Although AMP was present in the crystallization conditions, the electron density observed in the glycine-rich loop of the ATP-binding pocket was incompatible with this compound. Modeling and refinement of the density as diphosphate proved successful. The latter unit forms numerous H bonds to the backbone-amide NH groups of the glycine-rich loop, in analogy to the diphosphate moiety of AMP-PNP in the crystal structure of the E. coli enzyme complex [20]. With one exception (Gly95), all the backbone-amide NH moieties of glycine residues of the loop are involved in H-bonding. The residues of the glycine-rich loop, which are not glycine, however, do not participate in H-bonding interactions with diphosphate. This underlines the importance of the presence of glycine residues in phosphate-binding pockets. Presumably, glycine is ideally suited for the binding of phosphate groups in such loops as it can adopt the required conformation to wrap the loop around the bound phosphate group [38]. For future efforts aimed at the design of inhibitors for IspE from pathogens such as P. falciparum or M. tuberculosis, the finding that the cyclopropyl substituent of ligand 15 is accommodated in the hydrophobic pocket is of particular interest (Figure 7.10b). Given that the change from Phe185 (in E. coli IspE) to Tyr175 (in A. Aeolicus IspE) increases the hydrophilicity of the small pocket, the cyclopropyl ring was not expected to be located in this subpocket as it precludes solvation of the hydroxyl group of Tyr175. This is an important finding, given that the enzymes from P. falciparum and M. tuberculosis possess a conserved tyrosine residue at this position, and that – according to homology modeling – the phenolic hydroxyl group is directed into the small pocket. The propargylic sulfonamide vector of this class of inhibitors is therefore highly suitable for precisely addressing the phenolic hydroxyl group in future design cycles.

7.5 Conclusions and Outlook

7.4.4 Lessons Learnt from the Cocrystal Structure

By designing water-soluble inhibitors and solving the cocrystal structure of one ligand with A. aeolicus IspE, the suggested binding mode was validated. Thus, the first design cycle was successfully completed. The inhibitor occupies the CDP-binding pocket and exhibits a Ki value in the lower micromolar activity range. This proof of concept opens the way for further modification and optimization of the inhibitors aimed at the development of ligands with activity against IspE from medically important organisms. Thus, the cocrystal structure solved represents an important step on the way to anti-infectives with a novel mode of action.

7.5 Conclusions and Outlook 7.5.1 Conclusions

The active site of the kinase IspE features both highly polar and hydrophobic subpockets. Because the highly polar subpockets do not lend themselves well to structure-based drug design, the inhibitors were targeted to the lipophilic regions of the active site. Using structure-based design, the first inhibitors of the enzyme were developed, which display drug-like properties and highly satisfactory potency. A number of conclusions can be drawn from the series of inhibitors: • Inhibition of IspE is possible whilst bypassing the highly polar phosphate-binding pocket in the absence of a phosphate moiety. Therefore, the ligands synthesized constitute the first potent and drug-like inhibitors of an enzyme of the nonmevalonate pathway. • The inhibitors target only the CDP-binding pocket, potentially affording better selectivity than inhibitors designed to bind to the ATP-binding pocket. • The majority of the binding free enthalpy is derived from the cytosine scaffold and the propargylic sulfonamide vector. • A systematic variation of the vector’s substituent provided access to SARs, confirming the proposed binding mode. In addition, these modifications gave rise to the second example of the ‘‘55% rule’’ for the optimal filling of cavities applied to an enzymatic context. • Preparation of a water-soluble derivative enabled crystallographic studies of an enzyme–inhibitor complex, validating the proposed binding mode. This cocrystal structure provided invaluable structural information, especially in view of fine-tuning the ligands to obtain activity against the pathogenic enzymes. In conclusion, the new target IspE revealed itself to be ideally suited to structure-based inhibitor design. The inhibitors synthesized in the context of this project represent a successful application of this design strategy.

185

186


7.5.2 Outlook

The design and synthesis of the first-generation inhibitors of IspE opened up a number of future research avenues. First, as a result of the observed difference in amino acid sequences between the model system E. coli and the pathogenic enzymes – marked by the replacement of the phenylalanine residue lining the hydrophobic pocket by a tyrosine – the structure of the inhibitors will have to be fine-tuned to address this structural difference. Second, the bisubstrate approach could be tried to obtain even more potent and selective inhibitors of IspE. In principle, it should be possible to decorate the established scaffold to occupy the different pockets of the active site. Third, it would be very rewarding to identify an attractive mimic in the quest for new cytosine analogues. As opposed to the other nucleobases, very few cytosine substitutes have been described to date. Finally, identification of a generally applicable way of rendering promising inhibitors water soluble without having to undertake a time-consuming design cycle and synthesis of the new derivatives would be of great interest. In this way, attractive structures could be designed, synthesized, and assayed without worrying about their physicochemical properties such as solubility.

Acknowledgments

This work was carried out under the supervision of Prof. F. Diederich and in collaboration with the groups of Prof. A. Bacher and Prof. W. N. Hunter. Financial support by the Roche Research Foundation and the ETH Research Council is gratefully acknowledged.

List of Abbreviations

AMP-PNP ATP clogP CDP DMAPP GHMP HTS IC50 IPP IspE Ki ME

5 -adenyl-β, γ -amidotriphosphate adenosine 5 -triphosphate calculated partitioning coefficient cytidine 5 -diphosphate dimethylallyl diphosphate galactose/homoserine/mevalonate/phosphomevalonate high-throughput screening concentration of inhibitor at which 50% maximum initial velocity is observed isopentenyl diphosphate 4-diphosphocytidyl-2C-methyl-d-erythritol kinase inhibition constant methyl-erythritol

References

PC PDB SAR

packing coefficient protein data bank structure–activity relationship

References 1. http://www.cdc.gov/malaria/impact/

2.

3.

4.

5.

6.

7.

8.

index.htm (accessed 23 February 2009); http://www.cdc.gov/tb/WorldTBDay/ resources global.htm (accessed 23 February 2009). (a) Baird, J.K. (2005) Effectiveness of antimalarial drugs. N. Engl. J. Med., 352, 1565–1577; (b) Kritski, A.L., de Jesus, L.S.R., Andrade, M.K., Werneck-Barroso, E., Vieira, M.A., Haffner, A., and Riley, L.W. (1997) Retreatment tuberculosis cases. Factors associated with drug resistance and adverse outcomes. Chest, 111, 1162–1167; (c) Ollé-Goig, J.E. (2006) The treatment of multi-drug resistant tuberculosis – a return to the pre-antibiotic area? Trop. Med. Int. Health, 11, 1625–1628. Haney, S.A., LaPan, P., Pan, J., and Zhang, J. (2006) High-content screening moves to the front of the line. Drug Discov. Today, 11, 889–894. Schneider, G. and Böhm, H.-J. (2002) Virtual screening and fast automated docking methods. Drug Discov. Today, 7, 64–70. Butler, M.S. (2004) The role of natural product chemistry in drug discovery. J. Nat. Prod., 67, 2141–2153. (a) Kubinyi, H. (1998) Structure-based drug design. Chim. Oggi, 16, 17–22; (b) Klebe, G. (2000) Recent developments in structure-based drug design. J. Mol. Med., 78, 269–281; (c) Van Dongen, M., Weigelt, J., Uppenberg, J., Schultz, J., and Wikström, M. (2002) Structure-based screening and design in drug discovery. Drug Discov. Today, 7, 471–478. Knowles, J. and Gromo, G. (2003) Target selection in drug discovery. Nat. Rev. Drug Discov., 2, 63–69. (a) Bloch, K. (1992) Sterol molecule: structure, biosynthesis, and function. Steroids, 57, 378–383; (b) Bochar, D.A., Friesen, J.A., Stauffacher, C.V.,

9.

10.

11.

12.

13.

and Rodwell V.W. (1999)Biosynthesis of mevalonic acid from Acetyl-CoA, in Comprehensive Natural Products Chemistry: Isoprenoids, Including Carotenoids and Steroids (ed. D. E.Cane), Elsevier Science Publishers, Amsterdam, pp. 15–44. (a) Rohmer, M., Knani, M., Simonin, P., Sutter, B., and Sahm, H. (1993) Isoprenoid biosynthesis in bacteria: a novel pathway for the early steps leading to isopentenyl diphosphate. Biochem. J., 295, 517–524; (b) Schwarz, M.K. (1994) Terpen-Biosynthese in Gingko biloba: ¨ Eine Uberraschende Geschichte. Dissertation, ETH Zürich; (c) Broers, ¨ S.T.J. (1994) Uber die Frühen Stufen der Biosynthese von Isoprenoiden in E. coli. Dissertation, ETH Zürich; (d) Eisenreich, W., Schwarz, M., Cartayrade, A., Arigoni, D., Zenk, M.H., and Bacher, A. (1998) The deoxyxylulose phosphate pathway of terpenoid biosynthesis in plants and microorganisms. Chem. Biol., 5, R221–R233. Ruzicka, L. (1953) The isoprene rule and the biogenesis of terpenic compounds. Experientia, 9, 357–367. Jomaa, H., Wiesner, J., Sanderbrand, S., Altincicek, B., Weidemeyer, C., Hintz, M., Türbachova, I., Eberl, M., Zeidler, J., Lichtenthaler, H.K., Soldati, D., and Beck, E. (1999) Inhibitors of the nonmevalonate pathway of isoprenoid biosynthesis as antimalarial drugs. Science, 285, 1573–1576. Hunter, W.N. (2007) The nonmevalonate pathway of isoprenoid precursor biosynthesis. J. Biol. Chem., 282, 21573–21577. Hirsch, A.K.H. and Diederich, F. (2008) The nonmevalonate pathway to isoprenoid biosynthesis: a potential

187

188

7 Structure-Based Design on the Way to New Anti-Infectives source of new drug targets. Chimia, 62, 226–230. 14. (a) Altincicek, B., Hintz, M., Sanderbrand, S., Wiesner, J., Beck, E., and Jomaa, H. (2000) Tools for discovery of inhibitors of the 1-Deoxy-d-xylulose 5-Phosphate (DXP) synthase and DXP reductoisomerase: an approach with enzymes from the pathogenic bacterium Pseudomonas aeruginosa. FEMS Microbiol. Lett., 190, 329–333; (b) Mueller, C., Schwender, J., Zeidler, J., and Lichtenthaler, H.K. (2000) Properties and inhibition of the first two enzymes of the nonmevalonate pathway of isoprenoid biosynthesis. Biochem. Soc. Trans., 28, 792–793. 15. (a) Kuzuyama, T., Shimizu, T., Takahashi, S., and Seto, H. (1998) Fosmidomycin, a specific inhibitor of 1-Deoxy-d-xylulose 5-phosphate reductoisomerase in the nonmevalonate pathway for terpenoid biosynthesis. Tetrahedron Lett., 39, 7913–7916; (b) Herforth, C., Wiesner, J., Heidler, P., Sanderbrand, S., Van Calenbergh, S., Jomaa, H., and Link, A. (2004) Antimalarial activity of N6-substituted adenosine derivatives. Part 3. Bioorg. Med. Chem., 12, 755–762; (c) Walker, J.R. and Poulter, C.D. (2005) Synthesis and evaluation of 1-Deoxy-d-xylulose 5-phosphate analogues as chelation-based inhibitors of methylerythritol phosphate synthase. J. Org. Chem., 70, 9955–9959; (d) Woo, Y.H., Fernandes, R.P.M., and Proteau, P.J. (2006) Evaluation of fosmidomycin analogs as inhibitors of the Synechocystis sp PCC6803 1-Deoxy-d-xylulose 5-phosphate reductoisomerase. Bioorg. Med. Chem., 14, 2375–2385; (e) Haemers, T., Wiesner, J., Van Poecke, S., Goeman, J., Henschker, D., Beck, E., Jomaa, H., and Van Calenbergh, S. (2006) Synthesis of α-substituted fosmidomycin analogues as highly potent Plasmodium falciparum growth inhibitors. Bioorg. Med. Chem. Lett., 16, 1888–1891; (f) Devreux, V., Wiesner, J., Goeman, J.L., Van der Eycken, J., Jomaa, H., and Van Calenbergh, S. (2006) Synthesis

and biological evaluation of cyclopropyl analogues of fosmidomycin as potent Plasmodium falciparum growth inhibitors. J. Med. Chem., 49, 2656–2660. 16. (a) Crane, C.M., Kaiser, J., Ramsden, N., Rohdich, F., Eisenreich, W., Hunter, W.N., Bacher, A., and Diederich, F. (2006) Fluorescent inhibitors for IspF, an enzyme in the nonmevalonate pathway for isoprenoid biosynthesis and a potential target for antimalarial therapy. Angew. Chem., 118, 1083–1087; Angew. Chem. Int. Ed., 45, 1069–1074; (b) Baumgartner, C., Eberle, C., Diederich, F., Lauw, S., Rohdich, F., Eisenreich, W., and Bacher, A. (2007) Structure-based design and synthesis of the first weak non-phosphate inhibitors for IspF, an enzyme in the nonmevalonate pathway of isoprenoid biosynthesis. Helv. Chim. Acta, 90, 1043–1068. 17. (a) Reardon, J.E. and Abeles, R.H. (1986) Mechanism of action of isopentenyl pyrophosphate isomerase: evidence for a carbonium-ion intermediate. Biochemistry, 25, 5609–5616; (b) Muehlbacher, M. and Poulter, C.D. (1988) Isopentenyl-diphosphate isomerase: inactivation of the enzyme with active-site-directed irreversible inhibitors and transition-state analogues. Biochemistry, 27, 7315–7328; (c) Poulter, C.D., Muehlbacher, M., and Davis, D.R. (1989) Isopentenyldiphosphate isomerase. Mechanism of active-site-directed irreversible inhibition by 3-(fluoromethyl)-3-butenyl diphosphate. J. Am. Chem. Soc., 111, 3740–3742; (d) Lu, X.J., Christensen, D.J., and Poulter, C.D. (1992) Isopentenyl-diphosphate isomerase: irreversible inhibition by 3-methyl-3,4-epoxybutyl diphosphate. Biochemistry, 31, 9955–9960. 18. (a) Kuzuyama, T., Takagi, M., Kaneda, K., Watanabe, H., Dairi, T., and Seto, H. (2000) Studies on the nonmevalonate pathway: conversion of 4-(cytidine 5 -diphospho)-2C-methyl-d-erythritol to its 2-phospho derivative by 4-(cytidine 5 -diphospho)-2C-methyl-d-erythritol Kinase. Tetrahedron Lett., 41, 2925–2928;

References

19.

20.

21.

22.

23.

(b) Lüttgen, H., Rohdich, F., Herz, S., Wungsintaweekul, J., Hecht, S., Schuhr, C.A., Fellermeier, M., Sagner, S., Zenk, M.H., Bacher, A., and Eisenreich, W. (2000) Biosynthesis of terpenoids: YchB protein of Escherichia coli phosphorylates the 2-hydroxy group of 4-diphosphocytidyl-2C-methyl-derythritol. Proc. Natl. Acad. Sci. U.S.A., 97, 1062–1067. Bork, P., Sander, C., and Valencia, A. (1993) Convergent evolution of similar enzymatic function on different protein folds: the hexokinase, ribokinase, and galactokinase families of sugar kinases. Protein Sci., 2, 31–40. Miallau, L., Alphey, M.S., Kemp, L.E., Leonard, G.A., McSweeney, S.M., Hecht, S., Bacher, A., Eisenreich, W., Rohdich, F., and Hunter, W.N. (2003) Biosynthesis of isoprenoids: crystal structure of 4-diphosphocytidyl-2C-methyl-derythritol kinase. Proc. Natl. Acad. Sci. U.S.A., 100, 9173–9178. Wada, T., Kuzuyama, T., Satoh, S., Kuramitsu, S., Yokoyama, S., Unzai, S., Tame, J.R.H., and Park, S.Y. (2003) Crystal structure of 4-(cytidine 5’-diphospho)-2C-methyl-d-erythritol kinase, an enzyme in the nonmevalonate pathway of isoprenoid synthesis. J. Biol. Chem., 278, 30022–30027. Sgraja, T., Alphey, M.S., Ghilagaber, S., Marquez, R., Robertson, M.N., Hemmings, J.L., Lauw, S., Rohdich, F., Bacher, A., Eisenreich, W., Illarionova, V., and Hunter, W.N. (2008) Characterization of Aquifex aeolicus 4-diphosphocytidyl-2C-methyl-derythritol kinase – ligand recognition in a template for antimicrobial drug discovery. FEBS J., 275, 2779–2794. Crane, C.M., Hirsch, A.K.H., Alphey, M.S., Sgraja, T., Lauw, S., Illarionova, V., Rohdich, F., Eisenreich, W., Hunter, W.N., Bacher, A., and Diederich, F. (2008) Synthesis and characterization of cytidine derivatives that inhibit the kinase IspE of the nonmevalonate pathway for isoprenoid biosynthesis. ChemMedChem, 3, 91–101.

24. Hirsch, A.K.H., Alphey, M.S., Lauw,

25.

26.

27.

28.

29.

30.

31.

32.

S., Seet, M., Barandun, L., Rohdich, F., Hunter, W.N., Bacher, A., and Diederich, F. (2008) Inhibitors of the kinase IspE: structure–activity relationships and co-crystal structure analysis. Org. Biomol. Chem., 6, 2719–2730. Gerber, P.R. and Müller, K. (1995) MAB, a generally applicable molecular-force field for structure modeling in medicinal chemistry. J. Comput.-Aided Mol. Des., 9, 251–268. ACD/LogP & ACD/LogD, Advanced Chemistry Development, Inc. (2005)Toronto. www.acdlabs.com. (a) Meyer, E.A., Castellano, R., and Diederich, F. (2003) Interactions with aromatic rings in chemical and biological recognition. Angew. Chem., 115, 1244–1287; Angew. Chem. Int. Ed., 42, 1210–1250. Brameld, K.A., Kuhn, B., Reuter, D.C., and Stahl, M. (2008) Small molecule conformational preferences derived from crystal structure data. A medicinal chemistry focused analysis. J. Chem. Inf. Model., 48, 1–24. Hirsch, A.K.H., Lauw, S., Gersbach, P., Schweizer, W.B., Rohdich, F., Eisenreich, W., Bacher, A., and Diederich, F. (2007) Nonphosphate inhibitors of IspE protein, a kinase in the nonmevalonate pathway for isoprenoid biosynthesis and a potential target for antimalarial therapy. ChemMedChem, 2, 806–810. Hirsch, A.K.H. (2008) A novel approach towards antimalarials: design and synthesis of inhibitors of the kinase IspE . Dissertation, ETH Zürich. Illarionova, V., Kaiser, J., Ostrozhenkova, E., Bacher, A., Fischer, M., Eisenreich, W., and Rohdich, F. (2006) Nonmevalonate terpene biosynthesis enzymes as antiinfective drug targets: substrate synthesis and high-throughput screening methods. J. Org. Chem., 71, 8824–8834. Zürcher, M. and Diederich, F. (2008) Structure-based drug design: exploring the proper filling of apolar pockets at enzyme active sites. J. Org. Chem., 73, 4345–4361.

189

190

7 Structure-Based Design on the Way to New Anti-Infectives 33. Z¨ urcher, M., Gottschalk, T., Meyer,

S., Bur, D., and Diederich, F. (2008) Exploring the flap pocket of the antimalarial target plasmepsin II: the ‘‘55% Rule’’ applied to enzymes. ChemMedChem, 3, 237–240. 34. Mecozzi, S. and Rebek, J.Jr. (1998) The 55% solution: a formula for molecular recognition in the liquid state. Chem. Eur. J., 4, 1016–1022. 35. (a) Purse, B.W. and Rebek, J.Jr. (2006) Self-fulfilling cavitands: packing alkyl chains into small spaces. Proc. Natl. Acad. Sci. U.S.A., 103, 2530–2534; (b) Hooley, R.J. and Rebek, J.Jr. (2007) Self-complexed deep cavitands: alkyl chains coil into a nearby cavity. Org. Lett., 9, 1179–1182; (c) Gottschalk, T., Jaun, B., and Diederich, F. (2007) Container molecules with portals:

reversibly switchable cycloalkane complexation. Angew. Chem., 119, 264–268; Angew. Chem. Int. Ed., 46, 260–264. 36. (2006) Spartan’06, Wavefunction Inc., Irvine . 37. (a) Wuitschik, G., Rogers-Evans, M., Müller, K., Fischer, H., Wagner, B., Schuler, F., Polonchuk, L., and Carreira, E.M. (2006) Oxetanes as promising modules in drug discovery. Angew. Chem., 118, 7900–7903; (b) Angew. Chem. Int. Ed., 45, 7736–7739. 38. (a) Hirsch, A.K.H., Fischer, F.R., and Diederich, F. (2007) Phosphate recognition in structural biology. Angew. Chem., 119, 342–357; Angew. Chem. Int. Ed., 46, 338–352.

191

8 Drug–Membrane Interactions: Molecular Mechanisms Underlying Therapeutic and Toxic Effects of Drugs Marlene Lúcio, José L. F. C. Lima, and Salette Reis

8.1 Biological Membranes 8.1.1 Role of Membranes in Life Maintenance

Membranes are of fundamental importance for biological systems [1]. They provide for cellular and subcellular organelles’ compartmentalization and maintain the essential differences between the cellular and extracellular environments. The external membrane (called the plasma membrane) protects the cell from the loss of important molecules and the penetration of undesired substances from outside. This control of the internal cell environment provides for a regulation of the concentration of electrolytes and the maintenance of the osmotic pressure, which are part of the cellular homeostasis. Internal membranes also perform important functions separating incompatible biochemical reactions, maintaining intracellular transport, and playing a major role in energy transfer and storage [2]. Continuous and reliable communication between compartments (intracellular) and between cells (extracellular) is therefore required, involving processes that occur at the surface of membranes or are mediated by membranes. Membranes also serve as a matrix for the spatial organization of proteins among which are the enzymes and their cofactors, holding otherwise scattered molecules in functional contact [3]. 8.1.2 Structure and Composition of Membranes

In the past two decades there has been considerable advance in the knowledge of membrane structure. According to the Fluid Mosaic model [4], biological membranes are based on a fluid lipid bilayer, in which the lipids (and many proteins) show a rapid lateral diffusion. Such rapid diffusion implies that lipids and proteins are more or less randomly distributed in the plane of the membrane with a nonspecific order. However, this rather simple picture of ‘‘proteins floating in a sea of lipids’’ Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

192

8 Molecular Mechanisms Underlying Therapeutic and Toxic Effects of Drugs

can hardly match either the heterogeneous composition of the membrane or the membrane molecular organization. Indeed, more recent analysis of the bilayer structure suggests that the composition of biological membranes is not uniform, but shows a heterogeneous distribution of components, both laterally and across the bilayer [5, 6]. Cell membranes, nowadays, are considered to consist of solid domains coexisting with fluid membrane lipids that may also contain proteins [7–10]. The most important difference encountered in the actual membrane model from the fluid mosaic model is that a high degree of spatiotemporal order also prevails in the fluid membrane and in membrane domains [11], and this order seems to be essential for the functioning of lipid-embedded and integrated proteins [12, 13]. Therefore, the earlier concept of a lipid matrix, which was merely a support for embedded proteins, has evolved, and evidence is accumulating that lipid bilayer structure and dynamics play a crucial role in membrane functionality. According to this, it is worthwhile to describe in more detail, the lipid composition of biomembranes. Most of the membrane common lipids belong to three classes: phospholipids, glycolipids (lipid–sugar conjugates), and sterols (e.g., cholesterol). On average, a biological membrane contains about 40–70% of phospholipids, 0–50% of cholesterol, and 0–26% glycolipids [2]. The major components of the membranes are thus the phospholipids, which are amphiphilic molecules constituted of four components: one or two fatty acids designated as the hydrophobic tails, a glycerol backbone (or sphingosine in the case of the sphingolipids), a phosphate, and an alcohol attached to the phosphate. The phosphate and alcohol together form a polar head group. The hydrophobic tails of the phospholipids can differ in their length or in the number of unsaturated bonds. Because of the fatty acids synthesis cycles in the metabolism, practically all of the naturally occurring saturated fatty acids in natural membranes are palmitic (C16), stearic (C18), and myristic (C14). From the unsaturated fatty acids, the oleic (C18:1), which is a stearic acid with one double bond in the middle of the chain, is the most important. Most natural lipids contain one or more double bonds in cis conformation (e.g., higher animals, plants) while a trans conformation is found in bacteria. The level of saturation is important in controlling the fluidity of the membranes by lowering the phase transition temperature, as well as the hydration and mobility of the bilayer. Unsaturated fatty acids are therefore important for the proper functioning of biomembranes, and actual nutritional recommendations mirror this knowledge. The heterogeneity of lipids is also reflected in their polar heads. The most frequently occurring alcohol moieties are the amino acids serine, ethanolamine, choline, glycerol, and inositol [2]. The polar headgroups can substantially alter the function of a membrane since this is the part of the lipid that is present on the membrane surface, and is often responsible for the interaction with the surrounding. The polar groups of phospholipids are usually neutral or negatively charged. A positive charge is rare (cationic lipids are, however, important in for example, gene delivery, in which they form complexes with nucleic acids; the so-called lipoplexes). The electric charges of membrane lipids usually fall into a range of +1 to −[2]. Most of the neutral lipids are zwitterionic, characterized by strong dipole moments.

8.1 Biological Membranes

The electrostatics of lipid membranes is a complex issue, but strongly affects membrane–protein interactions, domain formation, as well as other membrane functions. In conclusion, membranes are complex supramolecular liquid-crystalline assemblies mainly composed of phospholipids. This complexity also arises from the fact that in most membranes different regions and domains with defined lipid and protein compositions usually coexist and similar domains are not always entirely equal [14]. Finally, membrane proteins and lipids may both be subjected to regulatory processes in response to pathophysiological situations or nutritional pharmacological interventions, which in turn may alter the activity and functions of the membrane [14]. Hence, membranes gather lipids and proteins, both fulfilling prominent roles in certain cellular processes, equally relevant in many cases. Therefore, drug–membrane interactions represent a wide and complex field in medicinal chemistry and can involve studies of drugs that are capable of influencing the protein function both directly and indirectly via lipid modulation. The protein regulation achieved by drug–membrane interactions can finally induce changes in cell signaling and gene expression, which might serve to reverse the pathological state. The literature on drug–membrane interactions is vast and multifaceted, rendering it a difficult task to elaborate a comprehensive review on the subject. In view of this, the present chapter will be restricted to examples of applications of drug–membrane interaction studies focusing on the influence of drugs on the lipid components. 8.1.3 Dynamic Molecular Organization of Membranes

A characteristic feature of the amphiphilic molecules in general, and lipids in particular, are their ability to spontaneously self-aggregate into organized and functional structures [15, 16]. In an aqueous environment, the lipid molecules conceal their hydrophobic tails constituting an interior shielded by their polar headgroups, whose surface interacts favorably with water. The driving force for lipid self-assembly is the hydrophobic effect [17, 18]. The molecular aggregation process, is itself a complex phenomenon, which still lacks experimental studies in atomic detail. The entire process has been however demonstrated using molecular dynamic simulations [19, 20]. As a starting point, from a random solution of phospholipids in water, a perfect bilayer forms spontaneously. The bilayer, a lamellar phase, is one of the several different possible structures into which lipids can self-organize. The diversity of the most common aggregation forms is shown in Figure 8.1. The extraordinarily rich polymorphism of lipid aggregates results from a subtle balance of attractive and repulsive interactions between these molecules at the water/lipid interface [21]. This balance is strongly affected by external conditions and of the lipid themselves. The major external factors determining the aggregation state of lipids are the temperature and the amount and the nature of the solvent. But other factors, such as the pressure, pH, and ion concentration also

193

8 Molecular Mechanisms Underlying Therapeutic and Toxic Effects of Drugs Increasing temperature

Increasing temperature

Lc

Increasing hydration

194

Lb

Pb ′ QI

HI

HII

La

M

Single chain Short double-chain lipids

Figure 8.1 Schematic dependence of phase transitions of lipid–water mixtures on hydration, temperature, and also on length of the alkyl tail of a phospholipid. Lc corresponds to the lamellar crystalline phase; Lβ

Long double-chain lipids

to the lamellar gel phase; Lα to the lamellar liquid-crystalline phase; HI to the hexagonal phase; HII to the inverted hexagonal phase; M to the micellar phase; Pβ to the ripple phase; and QII to the inverted cubic phase.

have a significant influence. Geometrical properties that affect the phase behavior are the size of the lipid headgroups, the length and saturation of lipid tails. In Figure 8.1, a schematic representation of the phase transitions induced by temperature, hydration, and lipid type is presented. At physiological conditions, the most common and the most biologically relevant lamellar phase is the fluid phase also named as liquid-crystalline phase (Lα ). In this phase, the rotational freedom of a single C–C bond results in the vigorous thermal movement of fatty acid chains of phospholipids ensuring the fluidity of the entire bilayer. Upon lowering the temperature, the bilayer undergoes a number of phase transitions ending in the lamellar crystalline phase (Lc ). The main phase transition from the fluid to the ordered phase occurs at a melting temperature (Tm ). Here, the tails become almost fully ordered and pack into a hexagonal lattice, forming a gel phase (Lβ ). Some disorder is still present, distinguishing the gel phase from the crystalline phase. For some lipids (dimyristoylphosphatidylcholine, DMPC; dipalmitoylphosphatidylcholine, DPPC), upon cooling the liquid-crystalline phase (Lα ) state to a temperature below the Tm , the formation of an intermediate gel phase, called a ripple phase (Pβ ) is observed. Decreasing the temperature further, a tilted gel phase (Lβ ) is formed.

8.2 Drug–Membrane Interactions

As mentioned above, the transition from a lamellar to a nonlamellar phase can be triggered both by temperature changes (thermotropic phase transition) and by (de)hydration (lyotropic phase transition). Upon increased hydration, the lamellar structure ruptures and long wormlike aggregates such as in the hexagonal HI phase, or small micelle aggregates, M form. On the other hand, dehydration (or temperature rise) can trigger the formation of inverted aggregates such as the HII or cubic phases QII . The described phase transitions occurring within a membrane can be considered as integrative phenomena of the dynamic molecular organization of membranes, which can be divided in conformational movements (related with the intramolecular movements and the long range order of lipid packing) and the translational movements (which indicate the lateral position of the molecule in the plane of the bilayer and are related with the short-range order of lipid packing). The common used term membrane fluidity can thus be considered to contain both conformational (microviscosity) and translational (lateral diffusion) dynamics [22].

8.2 Drug–Membrane Interactions 8.2.1 Possible Effects of Drugs on Membranes

It was previously mentioned that cell membranes, composed of lipids and proteins, function as a permeability barrier, maintain ion gradients across the membrane and steady state of fluxes, and possess recognition sites for communication and interaction with other cells. Disturbance of such a complex system by the uptake of exogenous compounds, like drugs and lipids, can cause several effects in membrane structure and dynamics, which in turn, are responsible for severe changes in the performance of the cell, including the function of transmembrane receptor proteins and proteins responsible for signal transduction. The degree of membrane perturbation will depend on both the structure and physicochemical properties of the exogenous molecules and membrane lipids involved, as well as the way of supplying the molecules to the membrane system. The effects on membranes that can arise from drug–membrane interactions are summarized in Figure 8.2. 8.2.2 Clinical Relevance of the Drug–Membrane Interaction Studies 8.2.2.1 Contribution for Drug Development Understanding the effects of drugs on membranes is of great importance for current and future pharmaceutical research. Drug discovery and development has typically considered interactions between the ligand and the molecular target (e.g.,

195

196


Conformation Location

CD

Fluorescence XD

Membrane / water partition

Drug

Derivative spectroscopy Drug–membrane Mode of action toxicity Membrane

Phase transitions

Structure AFM, BAM, SAXS WAXS, FFEM, XD

Surface charge Zeta-potential

Figure 8.2 Schematic representation of the analytical tools available for the evaluation of possible effects of drugs on membranes (effects on membrane structure, surface charge, dynamic properties, and

Dynamic properties

DSC

NMR, Fluorescence

phase transition) and effects of the membrane on drug molecules (effects on drug partition between aqueous and lipid phase, drug location and orientation, and drug conformation).

protein receptor), with the assumption that the membrane bilayer acts as a passive surrounding environment. In recent years, however, it has become apparent that membrane constituents, particularly phospholipids, interact with drugs and other hydrophobic chemicals [1, 23–25]. Partitioning of drugs into lipid bilayers of biological membranes plays a significant role in their uptake, transport, bioavailability, and distribution [24–26]. As most drugs are administered orally, their ability to transport across the intestinal epithelium, a monolayer of cells that line the interior of the intestine, is an important issue. For efficient absorption from the gastrointestinal tract, a compound has to cross the plasma membrane of enterocytes lining the gut lumen. The majority of the compounds cross the plasma membrane by passive diffusion through the lipid bilayer [27]. The main determinants for absorption are therefore drug–lipid interactions. After the drug has reached the circulation it will redistribute in the body. The distribution can be limited due to some blood–tissue barriers, such as the blood–brain barrier [28]. The blood–brain barrier is formed by endothelial cells sealed by tight junctions in brain capillaries, and several proteins are involved in the formation of tight junctions [28]. The blood–brain barrier is also metabolically very active and contains large amount of transport proteins [28]. However, most of the compounds going through these barriers must

8.2 Drug–Membrane Interactions

cross the lipid bilayer of the endothelial cells [28]. Thus, drug–lipid interactions also play a crucial role in the distribution of compounds across blood–tissue barriers. Furthermore, if the drug target is intracellular, the permeation through the plasma membrane of the target cell is of outmost importance. As different cells can have different plasma membrane compositions and physicochemical properties, a thorough knowledge of drug–lipid interactions would help the development of optimal properties for drug candidates [29–32]. 8.2.2.2 Understanding Therapeutic and Toxic Effect of Drugs In order to understand the therapeutic action of any drug it is pivotal to know the dynamics of interaction with the lipid membrane and its crossing rate. This information is particularly useful when the lipid membrane represents the actual target for the drug and, in this case, the studies of the drug–membrane interaction are obviously fundamental to understand the drugs mode of action and consequently their therapeutic effect. In this context, earlier studies on neuroactive compounds such as anesthetics, tranquillizers, antidepressants, and so on, have suggested a strong involvement of a nonspecific impairing action of these compounds on the lipid bilayer in their neurological effects. The major cause of the sedative side effect of the antidepressant imipramine has been attributed to its ability to penetrate membrane structures [33, 34]. Studies on anesthetics like halothane and enflurane have proved that the preferred position of interaction of inhalation anesthetics with the lipid bilayer is the membrane interface [35]. Besides drugs, other neuroactive compounds include pollutants that should also be investigated for both their environmental and health toxicity. Popular insecticides like parathion, malathion, lindane, dichlorodiphenyl trichloroethane (DDT), and allethrin induce perturbations of membrane permeability and fluidize the membrane and the changes could be partially related to their primary insecticidal activity [36–40]. Also, a number of other drugs have been described to alter lipid order and domain formation. Examples include cyclosporine [41], steroids [42], trifluoroperazine [43], and antibiotics like aminoglycosides [44, 45] or macrolides (azithromycin) [46–53]. This latter compound has also shown to markedly inhibit endocytosis [52, 53], probably by interacting with lipids, modifying biophysical properties of membrane and affecting membrane’s dynamics in living cells [46, 47]. The membrane interactions, as a part of the mechanism of action, is further shown for amphiphilic drugs like cisplatin [54–59], amphotericin B [60, 61], and antimicrobial peptides such as magainin, cecropin, and defensin [62–68], for which no protein receptors have been identified. Knowledge about drug–lipid interactions might contribute to a better understanding of the drug activity but might also shed light on the severe side effects of the drug, of which, the effects on membrane biophysics are pointed out as one of the most important causes. Examples of the correlation between the effect of drugs on membrane biophysics and their toxicity are widely described in the literature [69–83].

197

198


8.2.2.3 Understanding Mechanisms of Multidrug Resistance Multidrug resistance (MDR) is a critical problem in cancer chemotherapy. Overexpression of P-glycoprotein (PgP) on the plasma membrane is frequently observed in drug-resistant cancer cells leading to the failure of chemotherapy [84–86]. PgP is an integral membrane protein that also modulates drug absorption, bioavailability, tissue deposition, and excretion. Therefore, not only anticancer drugs, but also several other drugs are known to be efficiently pumped out from the cell by PgP, which thus plays an important role in compromising AIDS chemotherapy and other drug-treatable diseases [87–89]. A common conclusion emerging from the studies performed on MDR modulation was that the interaction of drugs with the lipid bilayer was a primary determinant for substrate recognition of PgP. Consequently, substrate binding to PgP should be considered in two steps, the initial partitioning of drugs to the lipid bilayer followed by binding to transmembrane region of the enzyme [86, 90–93]. Despite this knowledge, conventionally, drug–lipid interactions are neglected, either as a possible explanation for increased intracellular concentration of the substrate in studies of PgP function, or as a way of finding inhibitors of PgP. Indeed, besides the interaction of MDR modulators with transporter proteins, their interaction with plasma membrane lipids may also contribute to reverse the molecular mechanisms of MDR. According to this, a drug that is able to induce alterations of bilayer properties, especially altering fluidity and permeability, should play an essential role in the processes of MDR modulation. Hence, most of the MDR reversing agents are preferentially soluble in lipids and they may also exert an influence on the physical properties of lipid bilayers [94–96]. For example, phenothiazines, considered as MDR reversing agents affect the properties of the lipid bilayer causing phase separation in membranes containing lipid microdomains of sphingolipids and cholesterol [95]. These induced changes in the lipid environment of membrane transporters like PgP can affect either drug export or import [95]. In conclusion, perturbation of lipid phase of the cell membrane by modulators seems to be important for potentiating the anticancer drug therapy and other therapies, whose efficacy is affected by MDR mechanisms [95]. 8.2.2.4 Controlling Enzymatic Inhibition A growing body of evidence has shown that lipid–bilayer structure and dynamics play a key role in membrane functionality [97]. Hence, the lipid disturbance by the uptake of drugs, can lead to changes in many membrane properties, which then can induce severe alterations in the performance of the enzymes. In fact, binding of drugs to membrane lipids can lead to alterations in the function of proteins, as shown for phospholipase A2 [98], 5-lipoxygenase [99], cytochrome c oxidase [100], lysosomal phospholipases, phosphatidylinositol-specific phospholipase C, sphingomyelinase [70], and protein kinase C [101]. For instance, it has been reported that the inhibitory effect of the mycotoxin Fumonisin B1 on ceramide synthase activity might not be a typical enzyme–inhibitor type direct interaction, but a consequence of mycotoxin-induced changes in the dynamic organization of the lipid phase [83].

8.3 Analysis and Quantification of Drug–Membrane Interactions

Consequently, in addition to specific interactions between drugs and enzymes, drugs that act as enzymatic inhibitors are also likely to modify the bulk physical properties of the membrane. The mechanism of these effects on membrane biophysics is an area of current interest, and therefore justifies the increasing interest to study drug–membrane interactions [102]. 8.3 Analysis and Quantification of Drug–Membrane Interactions 8.3.1 Membrane Model Systems

Membranes of cell surfaces have distinct lateral structures [10, 103] whose molecular organization and dynamics are essential for their interaction with specific ligands and proteins [104, 105]. Since plasma membranes are too complex to be characterized in simple physical–chemical terms, the quantitative study of such questions proceeds most conveniently by means of model systems [103]. Therefore, to understand the basic functions of membranes, or mechanisms by which they undergo specific transitions, simple in vitro models typically consisting of only one or two components are used. A lipid bilayer membrane can be artificially reconstituted in various forms from purified or synthesized lipid molecules. Indeed, as previously mentioned in Section 8.1.3, in an aqueous environment, phospholipids spontaneously self-assemble into bilayer membranes through hydrophobic interactions between their acyl chains [106]. On the basis of the aggregation properties of lipids in aqueous media, several membrane models were developed for studying drug–membrane interactions namely liposomes, planar bilayers, and monolayers [104]. Liposomes are a good model for biomembranes and have been widely used for studying drug delivery and interaction of drugs with the phospholipid bilayer [107]. Liposomes are spherical vesicles consisting either of a single lipid bilayer or multiple lipid bilayers that enclose an internal aqueous volume [108]. They can be obtained easily by hydrating dry thin films of phospholipids. These vesicles form spontaneously as a result of the tendency to minimize the energetically unfavorable interactions of bilayer edges with water [109]. Liposomes can vary widely in size, ranging from few nanometers referred to as small unilamellar vesicles (SUVs) to the micrometer size of real cells also referred to as giant unilamellar vesicles (GUVs). Spontaneously formed liposomes are usually multilamellar and inhomogeneous in shape. However, various techniques can be used to reduce the size of vesicles and the number of bilayers (e.g., by ultrasound, so-called sonication and extrusion through polycarbonate membranes [108]), in order to obtain large unilamellar vesicles (LUVs). Other important models used for studying the structural properties of membranes are the monolayers. In fact, the use of monolayers of phospholipids as a membrane model has regained interest recently, since it is half of a lipid bilayer and thus constitutes a very convenient first step to approach the two-dimensional

199

200


structure of biomembranes [110–117]. Phospholipidic monolayers dispersed on the top of an aqueous subphase provide a highly informative approach for studying drug–lipid interactions at the air–water interface, as there is no curvature effect (unlike vesicles) and the lateral packing can be precisely controlled [111, 113, 118]. Moreover, the control of the state of the hydrocarbon chains and the relative concentration of the drug (both on the monolayer and in the subphase) allows a molecular control of the parameters associated with the effects of drugs [113]. Supported phospholipid bilayers (SPBs) can also act as simple in vitro models for physiological membranes gathering the advantages of providing for a lipid bilayer with no curvature effect. These bilayers are placed in a solid substrate, and a hydration layer, typically with a thickness of around 1–2 nm, exists between the lower leaflet of the SPB and the solid substrate. The presence of this layer provides the bilayer with flexibility and fluidity, creating a dynamic system where lipid molecules can, laterally, diffuse freely [23]. More recently, alternative models made of lipid bilayer-modified microbeads and nanomaterials [119] have been extensively studied. The microbeads and nanomaterials-supported lipid bilayers (nanoSLBs) somewhat mimic the native environment of membrane-associated biomolecules. Thus, from a bionic point of view, they can be considered as curved lipid membranes with high structural integrity [119]. Compared with other lipid membranes, an attractive feature of nanoSLBs is that some states of mimetic biomembranes could be monitored by tracking the changes of some intrinsic properties of the nanomaterials. Thus, the nanoSLBs could be used as a biomembrane model studying some membrane properties and biological functions, such as the fluidity of membranes, and the interaction between proteins and membranes [119]. 8.3.2 Experimental Techniques

The molecular description of the interactions between drugs and membrane constituents is a powerful tool to understand their mechanism of activity or toxicity. In view of this, a large number of methods have been developed for analyzing and quantifying various aspects of drug–membrane interactions (Figure 8.2). Experimental techniques used to study model membranes can be grouped into the following categories [1]: microscopic (optical and electron microscopy (EM)); spectroscopic (nuclear magnetic resonance (NMR), infrared (IR), electron spin resonance (ESR), fluorescence, X-ray diffraction (XD), neutron diffraction); thermodynamic (calorimetry, electrochemical methods); hydrodynamic (high performance liquid chromatography (HPLC), dynamic light scattering (DLS), zeta potential, viscosity); chemical; and mechanical (pipette aspiration). Before analyzing the properties of lipids of interest, they have to be isolated from the source and then separated. HPLC is a very accurate technique that allows the separation of different fractions of lipids [1]. The size of the vesicles formed can be determined by a simple turbidity method or more accurately by DLS. A macroscopic view of the cell is provided by EM.

8.4 Drug−Membrane Interactions Applied to the Study of NSAIDs

Especially, the application of the freeze-fracture EM (FFEM) technique allows the direct visualization of the architecture of the membrane and the embedded proteins [120]. The surface charge of the membrane can be also measured by determining the zeta potential. To obtain structural properties (such as area per lipid molecule, volume, thickness, or order parameter), microscopic (atomic force microscopy (AFM) and Brewster angle microscopy (BAM)), spectroscopic, diffraction (X-ray; neutron), or scattering methods (small angle X-Ray scattering and wide angle X-Ray scattering (SAXS and WAXS)) are used [103, 117, 121, 122]. Dynamic properties such as diffusion, permeability, or fluidity of membranes are measured by NMR, X-ray, or fluorescence [80, 123, 124]. Phase transitions can be detected and characterized using calorimetric methods such as differential scanning calorimetry (DSC) [102, 107, 124–126]. Besides analyzing the effect on the physicochemical properties of membranes, it is also necessary to evaluate the influence of the membrane itself in the drug behavior and thus determine drug partition, localization, orientation, and conformation within the membrane. This information can be obtained from a vast number of techniques which include derivative spectroscopy, fluorescence quenching, XD, circular dichroism (CD), and so on [1, 127, 128].

8.4 Drug–Membrane Interactions Applied to the Study of Nonsteroidal Anti-inflammatory Drugs (NSAIDs)

In the previous sections of this chapter, the clinical relevance of studying the interactions of drugs with membranes was briefly presented. This section will cover the particular case of nonsteroidal anti-inflammatory drugs (NSAIDs) since these drugs can be presented as an iconic example of the clinical importance of studying drug–membrane interactions. NSAIDs are one of the world’s most prescribed groups of drugs for acute and chronic inflammatory diseases, and their continuous use is associated with severe gastrointestinal toxicity [127]. Principal targets for the NSAIDs in controlling pain and inflammation are the cyclooxygenases (COXs), which are membrane-associated enzymes. To bind with the targets these drugs have to pass through the biomembranes, and thus their interactions with lipid bilayers are expected to play a major role in guiding their COX inhibition [127]. However, the sequence of events resulting from COX inhibition does not totally explain the gastric toxicity of NSAIDs. Such mechanisms are complex and mucosal damage can also be related to a local acid effect of the dissolved drug [127]. Most NSAIDs are weakly acidic, lipid-soluble compounds. Since the cell membranes on the stomach wall contain lipids for protection against strong acids, they offer little resistance to the lipid-soluble NSAIDs.

201

202


The observation of the strong influence of NSAIDs on membranes either by their inhibition effect on the activity of the membrane-located enzyme COX, or by their topical effect on gastric mucosa, kept our investigation group interested in exploring the membrane changes upon the interaction with NSAIDs molecules. Therefore, the focus of our studies has been placed on the effects of NSAIDs on membranes, and specifically the ability of these drugs to bind to surface-active phospholipids, as well as their effect on the lipid dynamic properties of membrane models, once the results elucidate about the effects of these drugs compromising the integrity of gastric mucosal barrier. In order to identify and isolate the effect of different membrane parameters on the interaction of NSAIDs, the biomimicry of the bilayer lipid membrane has been achieved by the use of membrane models [127, 128] (liposomes and monolayers) once they show structural similarities to the lipid matrix of cell membranes. Occasionally, more complex systems were used like mouse splenocytes, mouse macrophages cell line – J774, human leukemia monocyte cell line – THP-1, human granulocytes, and mononuclear cells [129, 130]. The experimental techniques used in the interaction studies of NSAIDs with membrane model systems can constitute an example for those willing to apply the same approach to other drugs or compounds and will be grouped as techniques more focused on drugs (e.g., drugs physical–chemical studies) and techniques more focused on the effects of drugs on membrane (e.g., membrane structural and dynamic studies). 8.4.1 Drug Fundamental Physical–Chemical Studies

When studying the interaction of drugs with biomembranes, it is highly recommended to start with physical–chemical studies that include the evaluation of drug acid–base properties in aqueous or organized media; drug solubility in several media; drug liposome/water partition coefficient (Kp ); and drug location within membranes. Indeed, one major question is to determine the likely mechanism by which the drug interacts with the membrane, which is dependent of the drugs hydrophilic/lipophilic balance. Therefore, in the study of the interaction of any compound with model membrane systems, the determination of the partition coefficient should be the first step. After this information is obtained, structural and dynamic studies can then be carried out. The extent of interaction of a solute with a membrane model system is evaluated in a quantitative way from its partition coefficient. The Kp determinations can be successfully obtained using several techniques (for a review see reference [131]). Among all the techniques, we have chosen the spectroscopic (derivative spectroscopy and fluorescence quenching techniques), which permitted, in the case of NSAIDs, to evaluate and compare the extent of penetration and/or interaction of the drugs with membrane phospholipids [128, 132–134]. When spectroscopic techniques are used, the measured parameter is a combination of signals from the ‘‘aqueous’’ and ‘‘lipidic’’ solute subpopulations, their relative weight depending on the partition


coefficient. Spectroscopic techniques are thus usually based on the analysis of signals originating from both phases. Therefore, there is no demand for their physical separation, which is a significant advantage [131], since physical separation of phases may be laborious and may result in equilibria perturbation. Fluorescence techniques, on the other hand, are amongst the most sensitive; low solute concentrations can be used, leading to a small perturbation of the lipidic membranes. Understanding the role of the lipid or polar moiety of drugs in altering the partitioning between membrane and aqueous environments will also assist in predicting the subcellular localization of these therapeutic substances. Moreover, fluorescence quenching studies and zeta potential measurements can be used to preview the location of drugs within the membrane. Fluorescence quenching can provide information on the position and orientation of drugs within the bilayer, as the precise location in the membrane of the fluorescent probes used is known. Furthermore, for the same probe, the extent of quenching is inversely proportional to its distance to the drug that acts as fluorescence quencher. In this way it was possible to predict the membrane location of NSAIDs, by measuring the decrease in fluorescence of the probe inserted in the lipid membrane when increasing concentrations of drugs were added [128, 132–135]. Additional information on NSAIDs location has been obtained by assessing their influence on the membrane surface potential, as it provides an indication of the type of interactions between drug and lipid bilayer surfaces [132, 133, 135]. Measurement of acid–base properties of drugs (e.g., pKa ) is also essential since drugs can face different physiologic pH conditions, and consequently their chemical behavior may change. Briefly, this method consists of a titration of a strong basic (or acidic) solution of the drug with a strong acid (or base) added with an automatic piston burette, controlled by a personal computer, under a nitrogen stream. The absorption spectrum of the drug is recorded in a system where a peristaltic pump enables circulation of the solution from the reaction pot, through a flow-cell of the spectrophotometer and back to the reaction pot. The acid–base properties of NSAIDs have been assessed by this method [132, 133]. 8.4.2 Membrane Structural and Dynamic Studies

After drug physical–chemical studies have been conducted, which enabled a deeper knowledge about drugs lipophilicity and about the charged forms of drugs, structural and dynamic studies have been carried out to evaluate the interaction and penetration of NSAIDs in lipid bilayers and the resultant variations in the thermodynamic parameters and lipid phase transitions. In this case, DSC is the appropriate technique and can be used to determine the onset temperature of the lipid chain–melting and other phase transitions, and the area under the curve (i.e., the enthalpy), which is representative of the cooperativity. The first effect observed by DSC when NSAIDs were interacting with membranes of DPPC and DMPC was the disappearance of the pretransition peak, which represents the phase transition

203

204


between bilayers in an Lβ organization, where the lipid molecules are parallel and tilted with respect to the bilayer normal, to a Pβ phase where the bilayer displays periodic ripples. The second effect was a shift on the main transition temperature, Tm , to a lower temperature, which represents the melting of the acyl chains from the Pβ to the Lα phase. Finally, the third effect was a broadening of the peak, indicating a lower cooperativity during the melting process. The calorimetric enthalpy, H did not change or changed slightly. This behavior is typical of an intercalation of the drugs into the lipid bilayer and indicates that the drug strongly interacts with liposomes. More specifically, it indicates that the molecule acts as a spacer causing fluidization and disordering of the lipid molecules in the membrane [127, 135]. Modifications of membrane biophysical properties elicited by the drugs under study have also been evaluated by fluorescence measurements of anisotropy [128]. Measurements of fluorescence anisotropy of probes inserted in lipid bilayers are a powerful tool to report changes in membrane microviscosity. Upon excitation with polarized light, the emission from lipid membrane labeled with probes is also polarized. The extent of polarization of the emission is described in terms of the anisotropy (r). The emission can become depolarized by a number of processes. One cause of depolarization is changing the microviscosity of the lipid media where the probe is inserted, which provokes rotational diffusion changes of the probe. Therefore, a change in fluorescence anisotropy of the probe will report a change in the microviscosity of the membrane where the probe rotates. Thus, in the case of NSAIDs there was a decrease in the fluorescence anisotropy in a concentration-dependent manner and we could conclude that, in agreement with DSC studies, NSAIDs increased membrane fluidity. Besides the effects of drugs on membrane microviscosity, it is also important to evaluate the effects on the order and packing of the lipids. As a result, the effects in symmetry and in the long- and short-range organization of bilayers; in the molecular packing of the bilayers as well as in the chain conformation of the molecules upon the interaction of the bilayers with NSAIDs have been investigated by SAXS and WAXS [127]. Scattering methods, such as SAXS and WAXS, have been useful to measure structural features on length scales between 1 nm up to several hundreds of nanometers by analyzing the scattering pattern at very low or very wide angles from the direct X-ray beam. These techniques provide complementary information: with SAXS it is possible to measure the long distances that correspond to the bilayer thickness; with WAXS it is possible to measure the short distances that correspond to chain order. Another highly sophisticated X-ray technique that was used to study the influence of NSAIDs in lipid membranes was grazing-incidence X-ray diffraction (GIXD), which provides the most precise structural parameters of condensed monolayers [127]. In GIXD, a monochromatic X-ray beam is directed on the surface at an angle of incidence that is smaller than the critical angle for total external reflection. Thus, the beam penetrates the subphase by only about 5 nm, and GIXD becomes a surface-sensitive technique. Consequently, the beam is diffracted by the structural pattern of the monolayer. The scattered intensities are detected as functions of the horizontal and the vertical scattering angle.


The corresponding in-plane and out-of-plane components of the scattering vector allow the calculation of distinct lattice parameters, the tilt angle, and the direction of the aliphatic chains, as well as the correlation length, that is, the size of a homogeneously structured lipid domain. Therefore, along with the studies in bilayers systems, models consisting of phospholipids spread at the air–water interface as a Langmuir monolayer and NSAIDs used as a second component dissolved in the subphase allowed the variation of physical–chemical parameters of the interface as lateral pressure and surface providing the measurement of pressure-area isotherms, which are simple models for studying the drug’s penetration [127]. Simultaneously, BAM measurements have permitted to follow the drug adsorption to the lipid monolayer and detect possible drug–lipid fluidity changes. 8.4.3 Results

Results obtained with the aforementioned techniques provided evidence that the NSAIDs evaluated have destabilizing effects, especially in the lipid gel phase, with a possible increase of the bilayer’s wettability, which could be transposed to a similar effect in the stomach’s mucus phospholipid gel layer with the development Digestive acids

Prostaglandines

Figure 8.3 Brewster angle microscopy image of the effect of NSAIDs on lipid membranes. Lipid membranes in the absence of NSAIDs (upper image) shows fluid phase (dark area) and solid domains (clear spots). NSAIDs show a fluidizing effect

on membrane (lower image) reducing the solid–lipid domains and increasing the fluid content of the membrane. On the right side is an schematic interpretation of the fluidizing effect in vivo.

205

206


of gastric erosions (Figure 8.3). The NSAIDs fluidizing effects have proved to be parallel with their membrane preferential location and are also related with their physical–chemical properties. For instance, in the lipid gel phase (Lβ ), indomethacin and acemetacin showed to be located inside the hydrophobic core of the membrane with the ionized carboxyl group anchored at the headgroup region. This membrane location parallels the highest destabilizing effects presented by these drugs, which in turn can be related with their gastric toxicity. Indeed, the destabilizing effects of these drugs on membranes may be part of the mechanism by which these NSAIDs attenuate the hydrophobic barrier properties of the stomach’s mucus phospholipid gel layer with the consequent increase in the back diffusion of luminal acid into the mucosa and the development of erosions (Figure 8.3). In comparison, nimesulide, one NSAID studied, which has a less damaging gastrointestinal profile, showed minimal effects of the lipid gel phase [127]. Further studies will be made with several other NSAIDs to extend the conclusions achieved. 8.5 Conclusions and Future Research Directions

The aims of this chapter were to highlight the importance of membrane composition and the dynamic molecular organization of membranes in drug membrane interaction studies, to point out the effects of such interactions on membrane properties, and finally to exemplify an application of drug–membrane interaction studies in the evaluation of the NSAIDs effect on membranes in their normal fluid state (allowing the evaluation of the therapeutic effect of drugs) and in the gel state typical of the gastric protective barrier (allowing the evaluation of the toxic effects of drugs). The studies presented offer starting points for further exploration of the interaction of drugs with membranes. Drugs affect not only the phospholipids, but also other components of the cell membranes such as membrane proteins. Studying the interactions of drugs with different membrane proteins, membrane compositions, and levels of hydration and pH, as well as studying interactions of drugs with the lipid microdomains are fertile fields for future investigation. Results gathered with these studies can provide useful information to understand the therapeutic effects of drugs, as well as their different selectivity and toxicity. Ultimately, these studies may prove valuable in the design of novel drug formulations with increased efficacy and reduced side effects. Acknowledgments

Partial financial support for this work was provided by Fundaçaõ para a Ciência e Tecnologia (FCT – Lisbon), through the contract PTDC/SAU-FCF/67718/2006. We are grateful to Dr Gerald Brezesinski and his research group at the Max Planck Institute (Golm, Germany) for his support and cooperation in some of our studies. We thank HASYLAB at DESY, Hamburg, Germany, for beam time and support through the contract I-20080033 EC.

References

References 1. Seydel, J.K., Wiese, M., Mannhold,

2. 3.

4.

5.

6.

7.

8.

9.

10.

11.

R., Kubinyi, H., and Folkers, G. (eds) (2002) Drug-Membrane Interactions: Analysis, Drug Distribution, Modeling, Wiley-VCH Verlag GmbH, Weinheim. Small, D.M. (ed.) (1998) Handbook of Lipid Research, Plenum Press, London. Berg, J.M., Tymoczko, J.L., and Stryer, L. (eds) (2006) Biochemistry, W. H. Freeman and Company. Singer, S.J. and Nicolson, G.L. (1972) The fluid mosaic model of the structure of cell membranes. Science, 175, 720–731. Welti, R. and Glaser, M. (1994) Lipid domains in model and biological membranes. Chem. Phys. Lipids, 73, 121–137. Balasubramanian, K. and Schroit, A.J. (2003) Aminophospholipid asymmetry: a matter of life and death. Annu. Rev. Physiol., 65, 701–734. Lommerse, P.H., Spaink, H.P., and Schmidt, T. (2004) In vivo plasma membrane organization: results of biophysical approaches. Biochim. Biophys. Acta, 1664, 119–131. Brown, D.A. and London, E. (2000) Structure and function of sphingolipidand cholesterol-rich membrane rafts. J. Biol. Chem., 275, 17221–17224. Frankel, D.J., Pfeiffer, J.R., Surviladze, Z., Johnson, A.E., Oliver, J.M., Wilson, B.S., and Burns, A.R. (2006) Revealing the topography of cellular membrane domains by combined atomic force microscopy/fluorescence imaging. Biophys. J., 90, 2404–2413. Jacobson, K., Mouritsen, O.G., and Anderson, R.G. (2007) Lipid rafts: at a crossroad between cell biology and physics. Nat. Cell Biol., 9, 7–14. Vereb, G., Szollosi, J., Matko, J., Nagy, P., Farkas, T., Vigh, L., Matyus, L., Waldmann, T.A., and Damjanovich, S. (2003) Dynamic, yet structured: the cell membrane three decades after the Singer-Nicolson model. Proc. Natl. Acad. Sci. U.S.A., 100, 8053–8058.

12. Simons, K. and Ikonen, E. (1997) Func-

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

tional rafts in cell membranes. Nature, 387, 569–572. Mouritsen, O.G. and Jorgensen, K. (1992) Dynamic lipid-bilayer heterogeneity: a mesoscopic vehicle for membrane function? Bioessays, 14, 129–136. Escriba, P.V., Gonzalez-Ros, J.M., Goni, F.M., Kinnunen, P.K.J., Vigh, L., Sanchez-Magraner, L., Fernandez, A.M., Busquets, X., Horvath, I., and Barcelo-Coblijn, G. (2008) Membranes: a meeting point for lipids, proteins and therapies. J. Cell. Mol. Med., 12, 829–875. Israelachvili, J.N., Mitchell, D.J., and Ninham, B.W. (1977) Theory of self-assembly of lipid bilayers and vesicles. Biochim. Biophys. Acta, 470, 185–201. Whitesides, G.M. and Grzybowski, B. (2002) Self-assembly at all scales. Science, 295, 2418–2421. Tanford, C. (1979) Interfacial free energy and the hydrophobic effect. Proc. Natl. Acad. Sci. U.S.A., 76, 4175–4176. Tanford, C. (1978) The hydrophobic effect and the organization of living matter. Science, 200, 1012–1018. Marrink, S.J. and Tieleman, D.P. (2001) Molecular dynamics simulation of a lipid diamond cubic phase. J. Am. Chem. Soc., 123, 12383–12391. Marrink, S.J., Lindahl, E., Edholm, O., and Mark, A.E. (2001) Simulation of the spontaneous aggregation of phospholipids into bilayers. J. Am. Chem. Soc., 123, 8638–8639. Kocherbitov, V. (2005) Driving forces of phase transitions in surfactant and lipid systems. J. Phys. Chem. B, 109, 6430–6435. Mouritsen, O.G., Kinnunen, P.K.J., Merz, K.M.J., and Roux, B. (eds) (1996) Role of Lipid Organization and Dynamics for Membrane Functionality, Birkhäuser, Berlin. Nussio, M.R., Sykes, M.J., Miners, J.O., and Shapter, J.G. (2009) Kinetics membrane disruption due to drug

207

208


24.

25.

26.

27.

28.

29.

30.

interactions of chlorpromazine hydrochloride. Langmuir, 25, 1086–1090. Seydel, J.K., Coats, E.A., Cordes, H.P., and Wiese, M. (1994) Drug membrane interaction and the importance for drug transport, distribution, accumulation, efficacy and resistance. Arch. Pharmazie, 327, 601–610. Seydel, J.K., Velasco, M.A., Coats, E.A., Cordes, H.P., Kunz, B., and Wiese, M. (1992) The importance of drug-membrane interaction in drug research-and-development. Quant. Struct.-Act. Relat., 11, 205–210. Mouristen, O.G., Anderson, H.K., Anderson, J.S., Davidsen, J., Nielsen, L.K., Jørgensen, K., Testa, B., van De Waterbeemd, H. Folkers, G., and Guy, R. (eds) (2001) Structure of Liposomal Membranes in Relation to Permeation, Wiley-VCH & Verlag Helvetica Chimica Acta, Weinheim, Zürich. Kansy, M., Fischer, K., Kratzat, F., Senner, B., Wagner, B., Parilla, I., Testa, B., van De Waterbeemd, H., Folkers, G., and Guy, R. (eds) (2001) High-throughput Artificial Membrane Permeability Studies in Early Lead Discovery and Development, Wiley-VCH & Verlag Helvetica Chimica Acta, Weinheim, Zürich. Krämer, S.D., Abbott, N.J., Begley, D.J., Testa, B., van De Waterbeemd, H., Folkers, G., and Guy, R. (eds) (2001) Biological Models to Study Blood-brain Barrier Permeation, Wiley-VCH Verlag GmbH Helvetica Chimica Acta, Weinheim, Zürich. Neves, P., Berkane, E., Gameiro, P., Winterhalter, M., and de Castro, B. (2005) Interaction between quinolones antibiotics and bacterial outer membrane porin OmpF. Biophys. Chem., 113, 123–128. Grancelli, A., Morros, A., Cabanas, M.E., Domenech, O., Merino, S., Vazquez, J.L., Montero, M.T., Vinas, M., and Hernandez-Borrell, J. (2002) Interaction of 6-fluoroquinolones with dipalmitoylphosphatidylcholine monolayers and liposomes. Langmuir, 18, 9177–9182.

31. Giacometti, A., Cirioni, O., Ghiselli, R.,

32.

33.

34.

35.

36.

37.

38.

Mocchegiani, F., Orlando, F., Silvestri, C., Bozzi, A., Di Giulio, A., Luzi, C., Mangoni, M.L., Barra, D., Saba, V., Scalise, G., and Rinaldi, A.C. (2006) Interaction of antimicrobial peptide temporin L with lipopolysaccharide in vitro and in experimental rat models of septic shock caused by gram-negative bacteria. Antimicrob. Agents Chemother., 50, 2478–2486. Chaudhuri, S., Banerjee, A., Basu, K., Sengupta, B., and Sengupta, P.K. (2007) Interaction of flavonoids with red blood cell membrane lipids and proteins: antioxidant and antihemolytic effects. Int. J. Biol. Macromol., 41, 42–48. Bauer, M., Megret, C., Lamure, A., LaCabanne, C., and Fauran-Clavel, M.J. (1990) Differential scanning calorimetry study of the interaction of antidepressant drugs, noradrenaline, and 5-hydroxytryptamine with a membrane model. J. Pharm. Sci., 79, 897–901. Fisar, Z., Fuksova, K., and Velenovska, M. (2004) Binding of imipramine to phospholipid bilayers using radioligand binding assay. Gen. Physiol. Biophys., 23, 77–99. Yoshida, T., Taga, K., Okabayashi, H., Kamaya, H., and Ueda, I. (1989) Changes in surface capacitance and conductance parallel to phospholipid membranes associated with phase transition: effects of halothane. Biochim. Biophys. Acta, 984, 253–256. Moya-Quiles, M.R., Munoz-Delgado, E., and Vidal, C.J. (1994) Interactions of the pyrethroid insecticide allethrin with liposomes. Arch. Biochem. Biophys., 312, 95–100. Antunes-Madeira, M.C., Videira, R.A., and Madeira, V.M. (1994) Effects of parathion on membrane organization and its implications for the mechanisms of toxicity. Biochim. Biophys. Acta, 1190, 149–154. Antunes-Madeira, M.C., Almeida, L.M., and Madeira, V.M. (1993) Depth-dependent effects of DDT and lindane on the fluidity of native membranes and extracted lipids. Implications for mechanisms of toxicity.

References

39.

40.

41.

42.

43.

44.

45.

46.

47.

Bull. Environ. Contam. Toxicol., 51, 787–794. Jorgensen, K., Ipsen, J.H., Mouritsen, O.G., Bennett, D., and Zuckermann, M.J. (1991) The effects of density fluctuations on the partitioning of foreign molecules into lipid bilayers: application to anaesthetics and insecticides. Biochim. Biophys. Acta, 1067, 241–253. Sreeramulu, K., Liu, R., and Sharom, F.J. (2007) Interaction of insecticides with mammalian P-glycoprotein and their effect on its transport function. Biochim. Biophys. Acta-Biomembr., 1768, 1750–1757. Soderlund, T., Lehtonen, J.Y., and Kinnunen, P.K. (1999) Interactions of cyclosporin A with phospholipid membranes: effect of cholesterol. Mol. Pharmacol., 55, 32–38. Wenz, J.J. and Barrantes, F.J. (2003) Steroid structural requirements for stabilizing or disrupting lipid domains. Biochemistry, 42, 14267–14276. Hendrich, A.B., Wesolowska, O., and Michalak, K. (2001) Trifluoperazine induces domain formation in zwitterionic phosphatidylcholine but not in charged phosphatidylglycerol bilayers. Biochim. Biophys. Acta-Biomembr., 1510, 414–425. Schanck, A., Mingeot-Leclercq, M.P., Tulkens, P.M., Carrier, D., Smith, I.C., and Jarrell, H.C. (1992) Interactions of aminoglycoside antibiotics with phospholipids. A deuterium nuclear magnetic resonance study. Chem. Phys. Lipids, 62, 153–163. Colome, C., Alsina, M.A., Busquets, M.A., Haro, I., and Reig, F. (1993) Interaction of aminoglycosides and colistin with model membranes – liposomes and monolayers. Int. J. Pharm., 90, 59–71. Berquand, A., Fa, N., Dufrene, Y.F., and Mingeot-Leclercq, M.P. (2005) Interaction of the macrolide antibiotic azithromycin with lipid bilayers: effect on membrane organization, fluidity, and permeability. Pharm. Res., 22, 465–475. Tyteca, D., Schanck, A., Dufrene, Y.F., Deleu, M., Courtoy, P.J., Tulkens, P.M., and Mingeot-Leclercq, M.P. (2003)

48.

49.

50.

51.

52.

53.

The macrolide antibiotic azithromycin interacts with lipids and affects membrane organization and fluidity: studies on Langmuir-Blodgett monolayers, liposomes and J774 macrophages. J. Membr. Biol., 192, 203–215. Berquand, A., Mingeot-Leclercq, M.P., and Dufrene, Y.F. (2004) Real-time imaging of drug-membrane interactions by atomic force microscopy. Biochim. Biophys. Acta-Biomembr., 1664, 198–205. Fa, N., Lins, L., Courtoy, P.J., Dufrene, Y., Van Der Smissen, P., Brasseur, R., Tyteca, D., and Mingeot-Leclercq, M.P. (2007) Decrease of elastic moduli of DOPC bilayers induced by a macrolide antibiotic, azithromycin. Biochim. Biophys. Acta-Biomembr., 1768, 1830–1838. Fa, N., Ronkart, S., Schanck, A., Deleu, M., Gaigneaux, A., Goonnaghtigh, E., and Mingeot-Leclercq, M.P. (2006) Effect of the antibiotic azithromycin on thermotropic behavior of DOPC or DPPC bilayers. Chem. Phys. Lipids, 144, 108–116. Montenez, J.P., VanBambeke, F., Piret, J., Schanck, A., Brasseur, R., Tulkens, P.M., and MingeotLeclercq, M.P. (1996) Interaction of the macrolide azithromycin with phospholipids. 2. Biophysical and computer-aided conformational studies. Eur. J. Pharmacol., 314, 215–227. Tyteca, D., Van Der Smissen, P., Mettlen, M., Van Bambeke, F., Tulkens, P.M., Mingeot-Leclercq, M.P., and Courtoy, P.J. (2002) Azithromycin, a lysosomotropic antibiotic, has distinct effects on fluid-phase and receptor-mediated endocytosis, but does not impair phagocytosis in J774 macrophages. Exp. Cell Res., 281, 86–100. Tyteca, D., Van Der Smissen, P., Van Bambeke, F., Leys, K., Tulkens, P.M., Courtoy, P.J., and Mingeot-Leclercq, M.P. (2001) Azithromycin, a lysosomotropic antibiotic, impairs fluid-phase pinocytosis in cultured fibroblasts. Eur. J. Cell Biol., 80, 466–478.

209

210

8 Molecular Mechanisms Underlying Therapeutic and Toxic Effects of Drugs 54. Maheswari, K.U., Ramachandran, T.,

55.

56.

57.

58.

59.

60.

61.

62.

and Rajaji, D. (2000) Interaction of cisplatin with planar model membranes – dose dependent change in electrical characteristics. Biochim. Biophys. Acta-Biomembr., 1463, 230–240. Speelmans, G., Staffhorst, R.W.H.M., Versluis, K., Reedijk, J., and de Kruijff, B. (1997) Cisplatin complexes with phosphatidylserine in membranes. Biochemistry, 36, 10545–10550. Burger, K.N.J., Staffhorst, R.W.H.M. and De Kruijff, B. (1999) Interaction of the anti-cancer drug cisplatin with phosphatidylserine in intact and semi-intact cells. Biochim. Biophys. Acta-Biomembr., 1419, 43–54. Speelmans, G., Sips, W.H., Grisel, R.J., Staffhorst, R.W., Fichtinger-Schepman, A.M., Reedijk, J., and de Kruijff, B. (1996) The interaction of the anti-cancer drug cisplatin with phospholipids is specific for negatively charged phospholipids and takes place at low chloride ion concentration. Biochim. Biophys. Acta, 1283, 60–66. Suwalsky, M., Hernandez, P., Villena, F., and Sotomayor, C.P. (2000) The anticancer drug cisplatin interacts with the human erythrocyte membrane. Z. Naturforsch., C: J. Biosci., 55, 461–466. Wang, K., Lu, J.F., and Li, R.C. (1996) The events that occur when cisplatin encounters cells. Coord. Chem. Rev., 151, 53–88. Baginski, M., Sternal, K., Czub, J., and Borowski, E. (2005) Molecular modelling of membrane activity of amphotericin B, a polyene macrolide antifungal antibiotic. Acta Biochim. Pol., 52, 655–658. Matsuoka, S., Ikeuchi, H., Umegawa, Y., Matsumori, N., and Murata, M. (2006) Membrane interaction of amphotericin B as single-length assembly examined by solid state NMR for uniformly C-13-enriched agent. Bioorg. Med. Chem., 14, 6608–6614. Bechinger, B. (1997) Structure and functions of channel-forming peptides: magainins, cecropins, melittin

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

and alamethicin. J. Membr. Biol., 156, 197–211. Matsuzaki, K. (1998) Magainins as paradigm for the mode of action of pore forming polypeptides. Biochim. Biophys. Acta, 1376, 391–400. Oren, Z. and Shai, Y. (1998) Mode of action of linear amphipathic alpha-helical antimicrobial peptides. Biopolymers, 47, 451–463. Sitaram, N. and Nagaraj, R. (1999) Interaction of antimicrobial peptides with biological and model membranes: structural and charge requirements for activity. Biochim. Biophys. Acta, 1462, 29–54. Shai, Y. (1999) Mechanism of the binding, insertion and destabilization of phospholipid bilayer membranes by alpha-helical antimicrobial and cell non-selective membrane-lytic peptides. Biochim. Biophys. Acta, 1462, 55–70. Foged, C., Franzyk, H., Bahrami, S., Frokjaer, S., Jaroszewski, J.W., Nielsen, H.M., and Olsen, C.A. (2008) Cellular uptake and membrane-destabilising properties of alpha-peptide/beta-peptoid chimeras: lessons for the design of new cell-penetrating peptides. Biochim. Biophys. Acta-Biomembr., 1778, 2487–2495. Hitz, T., Iten, R., Gardiner, J., Namoto, K., Walde, P., and Seebach, D. (2006) Interaction of alpha- and beta-oligoarginine-acids and amides with anionic lipid vesicles: a mechanistic and thermodynamic study. Biochemistry, 45, 5817–5829. Molitoris, B.A., Meyer, C., Dahl, R., and Geerdes, A. (1993) Mechanism of ischemia-enhanced aminoglycoside binding and uptake by proximal tubule cells. Am. J. Physiol., 264, F907–F916. Mingeot-Leclercq, M.P. and Tulkens, P.M. (1999) Aminoglycosides: nephrotoxicity. Antimicrob. Agents Chemother., 43, 1003–1012. Gallis, H.A., Drew, R.H., and Pickard, W.W. (1990) Amphotericin B: 30 years of clinical experience. Rev. Infect. Dis., 12, 308–329. Barwicz, J. and Tancrede, P. (1997) The effect of aggregation state of amphotericin-B on its interactions with

References

73.

74.

75.

76.

77.

78.

79.

80.

cholesterol- or ergosterol-containing phosphatidylcholine monolayers. Chem. Phys. Lipids, 85, 145–155. Andreoli, T.E. (1974) The structure and function of amphotericin B-cholesterol pores in lipid bilayer membranes. Ann. N. Y. Acad. Sci., 235, 448–468. Cohen, B.E. (1992) A sequential mechanism for the formation of aqueous channels by amphotericin B in liposomes. The effect of sterols and phospholipid composition. Biochim. Biophys. Acta, 1108, 49–58. de Kruijff, B. and Demel, R.A. (1974) Polyene antibiotic-sterol interactions in membranes of Acholeplasma laidlawii cells and lecithin liposomes. 3. Molecular structure of the polyene antibiotic-cholesterol complexes. Biochim. Biophys. Acta, 339, 57–70. de Paula, E. and Schreier, S. (1996) Molecular and physicochemical aspects of local anesthetic-membrane interaction. Braz. J. Med. Biol. Res., 29, 877–894. Malheiros, S.V.P., Pinto, L.M.A., Gottardo, L., Yokaichiya, D.K., Fraceto, L.F., Meirelles, N.C., and de Paula, E. (2004) A new look at the hemolytic effect of local anesthetics, considering their real membrane/water partitioning at pH 7.4. Biophys. Chem., 110, 213–221. Yun, I., Cho, E.S., Jang, H.O., Kim, U.K., Choi, C.H., Chung, I.K., Kim, I.S., and Wood, W.G. (2002) Amphiphilic effects of local anesthetics on rotational mobility in neuronal and model membranes. Biochim. Biophys. Acta, 1564, 123–132. Colell, A., Garcia-Ruiz, C., Lluis, J.M., Coll, O., Mari, M., and Fernandez-Checa, J.C. (2003) Cholesterol impairs the adenine nucleotide translocator-mediated mitochondrial permeability transition through altered membrane fluidity. J. Biol. Chem., 278, 33928–33935. Cronier, F., Patenaude, A., Gaudreault, R.C., and Auger, M. (2007) Membrane composition modulates the interaction between a new class of antineoplastic agents deriving from aromatic

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

2-chloroethylureas and lipid bilayers: a solid-state NMR study. Chem. Phys. Lipids, 146, 125–135. Grunicke, H., Grunewald, K., Helliger, W., Scheidl, F., Wolff-Schreiner, E., and Puschendorf, B. (1983) Inhibition of tumor growth by an alkylation of the plasma membrane. Adv. Enzyme Regul., 21, 21–30. Ihlenfeldt, M., Gantner, G., Harrer, M., Puschendorf, B., Putzer, H., and Grunicke, H. (1981) Interaction of the alkylating antitumor agent 2,3,5-tris(ethyleneimino)benzoquinone with the plasma membrane of Ehrlich ascites tumor cells. Cancer Res., 41, 289–293. Theumer, M.G., Clop, E.M., Rubinstein, H.R., and Perillo, M.A. (2008) The lipid-mediated hypothesis of fumonisin B1 toxicodynamics tested in model membranes. Colloids Surf. B: Biointerfaces, 64, 22–33. Bosch, I. and Croop, J. (1996) P-glycoprotein multidrug resistance and cancer. Biochim. Biophys. Acta, 1288, F37–F54. Gottesman, M.M. and Pastan, I. (1993) Biochemistry of multidrug resistance mediated by the multidrug transporter. Annu. Rev. Biochem., 62, 385–427. Omote, H. and Al-Shawi, M.K. (2006) Interaction of transported drugs with the lipid bilayer and beta-glycoprotein through a solvation exchange mechanism. Biophys. J., 90, 4046–4059. Ambudkar, S.V., Kimchi-Sarfaty, C., Sauna, Z.E., and Gottesman, M.M. (2003) P-glycoprotein: from genomics to mechanism. Oncogene, 22, 7468–7485. Borst, P. and Elferink, R.O. (2002) Mammalian ABC transporters in health and disease. Annu. Rev. Biochem., 71, 537–592. Leslie, E.M., Deeley, R.G., and Cole, S.P. (2005) Multidrug resistance proteins: role of P-glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue defense. Toxicol. Appl. Pharmacol., 204, 216–237. Lu, P., Liu, R., and Sharom, F.J. (2001) Drug transport by reconstituted P-glycoprotein in proteoliposomes.

211

212


91.

92.

93.

94.

95.

96.

97.

98.

99.

Effect of substrates and modulators, and dependence on bilayer phase state. Eur. J. Biochem., 268, 1687–1697. Omote, H. and Al-Shawi, M.K. (2002) A novel electron paramagnetic resonance approach to determine the mechanism of drug transport by P-glycoprotein. J. Biol. Chem., 277, 45688–45694. Romsicki, Y. and Sharom, F.J. (1997) Interaction of P-glycoprotein with defined phospholipid bilayers: a differential scanning calorimetric study. Biochemistry, 36, 9807–9815. Romsicki, Y. and Sharom, F.J. (1999) The membrane lipid environment modulates drug interactions with the P-glycoprotein multidrug transporter. Biochemistry, 38, 6887–6896. Drori, S., Eytan, G.D., and Assaraf, Y.G. (1995) Potentiation of anticancer-drug cytotoxicity by multidrug-resistance chemosensitizers involves alterations in membrane fluidity leading to increased membrane-permeability. Eur. J. Biochem., 228, 1020–1029. Michalak, K., Wesolowska, O., Motohashi, N., Molnar, J., and Hendrich, A.B. (2006) Interactions of phenothiazines with lipid bilayer and their role in multidrug resistance reversal. Curr. Drug Targets, 7, 1095–1105. Pajeva, I.K., Wiese, M., Cordes, H.P., and Seydel, J.K. (1996) Membrane interactions of some catamphiphilic drugs and relation to their multidrug-resistance-reversing ability. J. Cancer Res. Clin. Oncol., 122, 27–40. Bergethon, P.R. and Simons, E.R. (eds) (1990) Biophysical Chemistry – Molecules to Membranes, Springer-Verlag, New York. Mustonen, P. and Kinnunen, P.K. (1991) Activation of phospholipase A2 by adriamycin in vitro. Role of drug-lipid interactions. J. Biol. Chem., 266, 6302–6307. Pande, A.H., Qin, S., and Tatulian, S.A. (2005) Membrane fluidity is a key modulator of membrane binding,

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

insertion, and activity of 5-lipoxygenase. Biophys. J., 88, 4084–4094. Nicolay, K. and de Kruijff, B. (1987) Effects of adriamycin on respiratory chain activities in mitochondria from rat liver, rat heart and bovine heart. Evidence for a preferential inhibition of complex III and IV. Biochim. Biophys. Acta, 892, 320–330. Kumar, R., Holian, O., Cook, B., and Roshani, P. (1997) Inhibition of rat brain protein kinase C by lipid soluble psychotropics. Neurochem. Res., 22, 1–10. Epand, R.M. (1997) Biophysical studies of lipopeptide-membrane interactions. Biopolymers, 43, 15–24. Thormann, E., Simonsen, A.C., Nielsen, L.K., and Mouritsen, O.G. (2007) Ligand-receptor interactions and membrane structure investigated by AFM and time-resolved fluorescence microscopy. J. Mol. Recognit., 20, 554–560. Banerjee, S., Bennouna, M., Ferreira-Marques, J., Ruysschaert, J.M., and Caspers, J. (1999) Lipid-drug interaction and colligative properties in phospholipid vesicles. J. Colloid Interface Sci., 219, 168–177. Ohki, S. (1981) Membrane-potential, surface-potential, and ionic permeabilities. Physiol. Chem. Phys., 13, 195–210. Suzuki, H. and Takeuchi, S. (2008) Microtechnologies for membrane protein studies. Anal. Bioanal. Chem., 391, 2695–2702. Liu, Q., Qu, Y., Van Antwerpen, R., and Farrell, N. (2006) Mechanism of the membrane interaction of polynuclear platinum anticancer agents. Implications for cellular uptake. Biochemistry, 45, 4248–4256. Lasic, D.D. (ed.) (1993) Liposomes – from Physics to Applications, Elsevier, New York. Lasic, D.D. (1988) The mechanism of vesicle formation. Biochem. J., 256, 1–11. Brezesinski, G. and Mohwald, H. (2003) Langmuir monolayers to study interactions at model membrane

References

111.

112.

113.

114.

115.

116.

117.

118.

119.

surfaces. Adv. Colloid Interface Sci., 100-102, 563–584. Brockman, H. (1999) Lipid monolayers: why use half a membrane to characterize protein-membrane interactions? Curr. Opin. Struct. Biol., 9, 438–443. Caetano, W., Ferreira, M., Tabak, M., Sanchez, M.I.M., Oliveira, O.N., Kruger, P., Schalke, M., and Losche, M. (2001) Cooperativity of phospholipid reorganization upon interaction of dipyridamole with surface monolayers on water. Biophys. Chem., 91, 21–35. Hidalgo, A.A., Caetano, W., Tabak, M., and Oliveira, O.N. (2004) Interaction of two phenothiazine derivatives with phospholipid monolayers. Biophys. Chem., 109, 85–104. Jutila, A., Soderlund, T., Pakkanen, A.L., Huttunen, M., and Kinnunen, P.K. (2001) Comparison of the effects of clozapine, chlorpromazine, and haloperidol on membrane lateral heterogeneity. Chem. Phys. Lipids, 112, 151–163. Agasosler, A.V., Tungodden, L.M., Cejka, D., Bakstad, E., Sydnes, L.K., and Holmsen, H. (2001) Chlorpromazine-induced increase in dipalmitoylphosphatidylserine surface area in monolayers at room temperature. Biochem. Pharmacol., 61, 817–825. Varnier Agasoster, A. and Holmsen, H. (2001) Chlorpromazine associates with phosphatidylserines to cause an increase in the lipid’s own interfacial molecular area–role of the fatty acyl composition. Biophys. Chem., 91, 37–47. da Silva, A.M.G. and Romao, R.I.S. (2005) Mixed monolayers involving DPPC, DODAB and oleic acid and their interaction with nicotinic acid at the air-water interface. Chem. Phys. Lipids, 137, 62–76. Nagy, I.B., Haro, I., Alsina, M.A., Reig, F., and Hudecz, F. (1998) Interaction of branched chain polymeric polypeptides with phospholipid model membranes. Biopolymers, 46, 169–179. Zhang, L.X., Liu, J.Y., and Wang, E.K. (2008) A new method for studying

120.

121.

122.

123.

124.

125.

126.

127.

the interaction between chlorpromazine and phospholipid bilayer. Biochem. Biophys. Res. Commun., 373, 202–205. Ladaviere, C., Toustou, M., Gulik-Krzywicki, T., and Tribet, C. (2001) Slow reorganization of small phosphatidylcholine vesicles upon adsorption of amphiphilic polymers. J. Colloid Interface Sci., 241, 178–187. Lewis, R.N., Tristram-Nagle, S., Nagle, J.F., and McElhaney, R.N. (2001) The thermotropic phase behavior of cationic lipids: calorimetric, infrared spectroscopic and X-ray diffraction studies of lipid bilayer membranes composed of 1,2-di-O-myristoyl-3-N,N,N-trimethylaminopropane (DM-TAP). Biochim. Biophys. Acta, 1510, 70–82. Tristram-Nagle, S. and Nagle, J.F. (2004) Lipid bilayers: thermodynamics, structure, fluctuations, and interactions. Chem. Phys. Lipids, 127, 3–14. Panicker, L. and Mishra, K.P. (2006) Nuclear magnetic resonance and thermal studies on the interaction between salicylic acid and model membranes. Biophys. Chem., 120, 15–23. Tsogas, I., Tsiourvas, D., Nounesis, G., and Paleos, C.M. (2006) Modeling cell membrane transport: interaction of guanidinylated poly(propylene imine) dendrimers with a liposomal membrane consisting of phosphate-based lipids. Langmuir, 22, 11322–11328. Castelli, F., Raudino, A., and Fresta, M. (2005) A mechanistic study of the permeation kinetics through biomembrane models: Gemcitabine-phospholipid bilayer interaction. J. Colloid Interface Sci., 285, 110–117. Gicquaud, C., Auger, M., Wong, P.T.T., Poyet, P., Boudreau, N., and CGaudreault, R. (1996) Interaction of 4-tert-butyl-[3-(2-chloroethyl)ureido] benzene with phosphatidylcholine bilayers: a differential scanning calorimetry and infrared spectroscopy study. Arch. Biochem. Biophys., 334, 193–199. Lucio, M., Bringezu, F., Reis, S., Lima, J.L.F.C., and Brezesinski, G.

213

214


128.

129.

130.

131.

(2008) Binding of nonsteroidal anti-inflammatory drugs to DPPC: structure and thermodynamic aspects. Langmuir, 24, 4132–4139. Lucio, M., Ferreira, H., Lima, J.L.F.C., Matos, C., de Castro, B., and Reis, S. (2004) Influence of some anti-inflammatory drugs in membrane fluidity studied by fluorescence anisotropy measurements. Phys. Chem. Chem. Phys., 6, 1493–1498. Sousa, C., Nunes, C., Lucio, M., Ferreira, H., Lima, J.L.F.C., Tavares, J., Cordeiro-Da-Silva, A., and Reis, S. (2008) Effect of nonsteroidal anti-inflammatory drugs on the cellular membrane fluidity. J. Pharm. Sci., 97, 3195–3206. Ferreira, H., Lucio, M., Lima, J.L.F.C., Cordeiro-da-Silva, A., Tavares, J., and Reis, S. (2005) Effect of anti-inflammatory drugs on splenocyte membrane fluidity. Anal. Biochem., 339, 144–149. Santos, N.C., Prieto, M., and Castanho, M.A.R.B. (2003) Quantifying molecular partition into model systems of biomembranes: an emphasis on optical

132.

133.

134.

135.

spectroscopic methods. Biochim. Biophys. Acta-Biomembr., 1612, 123–135. Ferreira, H., Lucio, M., Lima, J.L.F.C., Matos, C., and Reis, S. (2005) Effects of diclofenac on EPC liposome membrane properties. Anal. Bioanal. Chem., 382, 1256–1264. Ferreira, H., Lucio, M., Lima, J.L.F.C., Matos, C., and Reis, S. (2005) Interaction of clonixin with EPC liposomes used as membrane models. J. Pharm. Sci., 94, 1277–1287. Ferreira, H., Lucio, M., de Castro, B., Gameiro, P., Lima, J.L.F.C., and Reis, S. (2003) Partition and location of nimesulide in EPC liposomes: a spectrophotometric and fluorescence study. Anal. Bioanal. Chem., 377, 293–298. Lucio, M., Nunes, C., Gaspar, D., Golebska, K., Wisniewski, M., Lima, J.L.F.C., Brezesinski, G., and Reis, S. (2009) Effect of anti-inflammatory drugs in phosphatidylcholine membranes: a fluorescence and calorimetric study. Chem. Phys. Lett., 471, 300–309.

215

9 Targeting Disease with Small Molecule Inhibitors of Protein–Protein Interactions Fedor Forafonov, Elena Miranda, Ida Karin Nordgren, and Ali Tavassoli

9.1 Introduction

There exists in every cell an intricate network of thousands of protein interactions, which control and regulate the majority of cellular processes and functions, from intercellular communication, proliferation, growth, and differentiation to programmed cell death. Alterations in protein interaction networks are the key regulatory components of almost all diseases, from viral infection to heart disease or cancer. There is therefore tremendous potential for small molecule modulators of protein–protein interactions to play a key role in the next generation of healthcare. Such compounds would treat the cause, not the symptoms of disease, at an earlier stage and on a more fundamental level; they would also serve as tools to enhance our current understanding of biological processes. As a result of this potential for novel therapeutic agents, there has been a great deal of recent effort to develop methodologies that allow the facile and rapid identification of small molecules capable of selectively disrupting specific protein–protein interactions [1, 2]. There are however, several factors that make identifying such compounds extremely challenging. Unlike enzymes, which have substrates that serve as an ideal starting point for the design of inhibitors, no such small molecule starting points exist for protein–protein interactions. The often large and uncharted protein surface combined with a lack of cavities and small molecule binding sites, results in the discovery of small molecule inhibitors being quite challenging. The size of the interacting protein interfaces can vary from 750 to 5000 A˚ 2 , and are typically stabilized via a series of buried hydrophobic patches [3]. Such interactions and their subsequent stabilization is thought to occur through van der Waals contacts between nonpolar amino acid residues, driven by the gain in free energy resulting from elimination of the polar aqueous environment. Electrostatic forces and hydrogen bonds are also believed to play an important part in partner recognition and subsequent stabilization of the protein dimer [4]. Such stabilizing regions, and more specifically the residues responsible for stabilizing the interaction are termed hot spots [4]. These hot spot residues have traditionally been identified using a combination of X-ray crystallography and site-directed Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

216

9 Protein Interactions with Small Molecules

mutagenesis, replacing individual amino acids with alanine (alanine scanning), and measuring the effect on the total binding energy of the complex. This approach tends to be quite laborious, and consequently available data on hot spot residues is limited, but the recent advances in computational power have resulted in several in silico approaches identifying hot spots on protein surfaces. Given the structural diversity and large number of protein interactions, it seems reasonable to assume that each interacting interface is also specific, unique, and diverse. It is also thought however, that proteins with multiple interactions tend to use the same hot spots, with structural changes varying the hot spot, adapting it to present the same residues in different structural contexts. Small molecules that target hot spots may also inadvertently affect other protein–protein interactions, with subsequent physiological side effects. The presence of structural and biophysical data for two interacting proteins, allows the possibility of rationally designed small molecule inhibitors, or in silico screening of virtual libraries. Although the availability of such structural data for interacting proteins is relatively rare, there have been several examples of protein–protein interaction inhibitors identified by these methods [5–7].

9.2 High-Throughput Screening of Chemical Libraries

The recent advances in chemical biology and functional genomics have given rise to new methods and opportunities for drug discovery, and in turn new methods for the generation of small molecule libraries [8–12]. Combinatorial chemistry, the method of choice for assembling chemically synthesized libraries has also undergone rapid development over the last few years [13–15]. Yet, despite providing almost unlimited functional group diversity, chemically synthesized libraries of small molecules still lack a suitable, straightforward method for their decoding. Split and pool protocols allow for the chemical synthesis of libraries with up to ∼105 members [16, 17]; however, the paths to identifying the active members of such libraries are often laborious and complicated [18]. Another approach to identifying inhibitors of protein–protein interactions is to use high-throughput screens. The composition of the libraries used in such screens can vary from small molecules to peptides, with library size varying from a few hundred to a hundred million. The challenge with this method of drug discovery lies in identifying the binding site and mode of action of the active compounds. Also the compound/protein ratio in such screens may be quite high and not reflect realistic physiological levels. The advantage of high-throughput approaches is that unlike rational design, they tend to not bias the compounds to one binding site, instead allowing the most effective binding site (even previously unknown allosteric sites) to be targeted. An early example is inhibition of Myc, a basic helix-loop-helix leucine zipper (bHLHZip) transcription factor that has been identified as an oncogenic effector, involved in inducing lymphoid tumors [19]. Through dimerization with Max, another


bHLHZip protein, its transcriptional activity drives cell proliferation, stimulates angiogenesis, and represses differentiation. The Myc/Max protein–protein interaction has therefore become a potent target for the development of small molecule inhibitors. Using a fluorescence-based high-throughput screen, a library of 7000 peptidomimetic compounds was assayed for inhibitors of Myc/Max dimerization [19]. Four compounds were identified, the most potent of which, produced 38% dissociation between the protein partners at 25 µM. However, it was found that another leucine zipper, Jun, was also inhibited by the identified compounds and with an IC50 of only 75 µM. Although these compounds did not show potency high enough to warrant further investigation, the methodology illustrated the potential use of high-throughput screens for the identification of small molecule inhibitors of protein–protein interactions. Another well-explored protein–protein interaction with relevance to cancer is the one between the human protein double minute 2 (HDM2) and the tumor suppressor protein p53. The tumor suppressor p53 is a potent transcription factor that controls and regulates cellular response to damage and stress [20]. Intracellular p53 levels are elevated in response to a variety of stress signals, leading to cell cycle arrest or apoptosis, and as such, it is frequently inactivated in a variety of cancers through its dimerization with HDM2. The binding of HDM2 to p53, upregulates p53 degradation, resulting in the suppression of its transcriptional activity (and inhibition of cell cycle arrest and apoptosis). This interaction has been identified as a potential target for new, broad-based anticancer therapies. There have been two recent reports of HDM2/p53 inhibitors from high-throughput screens. Hoffmann-La Roche, in New Jersey, United States, recently reported a series of tetra-substituted imidazoles (termed Nutlins) that bind HDM2 and inhibit its interaction with p53 (Figure 9.1) [21]. The most potent compound, Nutlin-3, had an IC50 of 90 nM. The mode of inhibition and the binding location of these compounds were identified by determining the crystal structure of the HDM2/Nutlin-2 complex. It had been postulated that an inhibitor of HDM2/p53 interaction should lead to stabilization and accumulation of p53 protein, resulting in activation of p53-regulated pro-apoptotic genes. Small molecule inhibitors of this interaction were therefore expected to cause cell cycle O

OH Br

O

N

N

Cl

N

O

N

N O

O

N

N O

Br

O Cl

Nutlin-2 Figure 9.1

NH

The structure of Nutlin-2 and Nutlin-3.

Nutlin-3

217

218


arrest and apoptosis. This was shown to be the case in a series of in vitro and in vivo experiments and in multiple cellular models [21]. The HDM2/p53 interaction was also targeted by Johnson & Johnson, where a library of 338 000 compounds designed by Directed Diversity software were screened in parallel for binding to HDM2, by monitoring changes in the thermostability of the protein complex using a specially designed instrument [22]. This methodology utilizes fluorescent dyes to monitor protein unfolding as a function of temperature, allowing the detection of compounds binding to target proteins by measuring the resultant increase in thermal stability. The natural agonist from p53 fills a relatively deep hydrophobic pocket of HMD2 through three side chains, Phe19, Trp23, and Leu26 (Figure 9.2) [23]. A novel series of benzodiazepinediones were identified as potent inhibitors of the HDM2/p53 interaction, exhibiting a Kd of 67 nM and an IC50 of 420 nM [22]. As with the nutlin series, crystal structures were obtained and used to observe the nature of the inhibitor–protein interaction and utilized to improve agonists. The compounds identified as inhibitors bind to the same region as this α-helical portion, and insert aromatic or aliphatic moieties into the same hotspot pockets of HDM2. In the case of Nutlin-2, the bromophenyl moiety is observed sitting in

Phe19

Leu26

Trp23

Figure 9.2 Structure of the MDM2 oncoprotein bound to the p53 tumor suppressor transactivation domain. Three key residues on p53 are highlighted.


the Trp pocket, while the remaining bromophenyl is occupying the Leu pocket, and the ether side chain is directed toward the Phe pocket. As with the bound peptide, the interaction is largely through nonspecific Van der Waals contacts with no hydrogen bonds involved. The interaction of HDM2 with p53 produces a more open conformation of HDM2, whereas it appears to close over the small molecule inhibitors resulting in a more concave contact. These structures also clearly demonstrate the remarkable efficiency of the inhibitors as compared to the p53 peptide. The inhibitors bind 25% less of the exposed HMD2 surface (804 vs. 1073 A˚ 2 ) than the minimal peptide binding epitope with less than a third of the molecular weight, but manage to achieve a 10-fold higher binding affinity. Protein–protein interaction targets need not be limited to anticancer therapeutics. In fact, host proteins are regularly usurped by viruses and other pathogens as part of the infection process and hence constitute an enticing target for combating infections. This is demonstrated by HIV, with protein–protein interactions heavily involved in all steps of the viral life cycle (Figure 9.3). Maraviroc (UK-427 857) is a recent anti-HIV compound, marketed by Pfizer, and discovered by screening their in-house compound library against the protein–protein interaction between the human chemokine receptor CCR5 and

Entry

Replication

Budding Figure 9.3 The lifecycle of HIV: viral entry involves the interaction of the HIV gp120 protein with the human CD4 and CCR5 or CXCR4 proteins. HIV uses the host cell’s machinery to replicate its genetic material

and core proteins, which have been integrated into the host’s DNA. Immature virus particles bud from the host cell, by utilizing host proteins, and go on to infect other cells.

219

220


F F

O

NH

N

Figure 9.4

N

N N

The structure of the HIV entry inhibitor, maraviroc.

the HIV envelope protein gp 120 [24]. This interaction is a key mediator of cell–cell fusion and HIV invasion. In order to identify a lead compound, Pfizer screened their compound file against CCR5, using a chemokine radioligand-binding assay. This screen monitors the level of dimerization through the measurement of an 125 I-labeled protein partner, in this case gp120, and its subsequent dissociation on protein–protein interaction disruption. One of the most potent and ligand-efficient (defined as binding affinity per heavy atom) lead compound was optimized for binding potency, antiviral activity, absorption, and pharmacokinetics as well as selectivity against key human targets such as the hERG channel by screening a further 1000 analogs. This resulted in maraviroc (Figure 9.4) with a low nanomolar IC90 potency and broad-spectrum antiviral activity. Through this development process, Pfizer developed and optimized two bespoke assays, one measuring envelope binding to cell surface receptors and a second modeling subsequent membrane fusion events. Importantly there was no unrelated activity observed, especially with CCR2, which has closest sequence identity to CCR5.

9.3 High-Throughput Screening of Biosynthesized Libraries

In contrast to chemical libraries, biologically synthesized libraries of small molecules are often several-fold larger in size and allow for very straightforward identification of the active members [25–27]. When combined with an in vivo screen, biological libraries become part of a powerful, rapid, and facile method for the screening of a large number of compounds against a chosen target [11, 27–29]. Biological libraries are principally polypeptides, typically embedded within, or fused to larger molecules in order to overcome the host cell’s catabolic machinery. Alternatively, intracellular stability is achieved by constraining the ends of the peptide with noncovalent and covalent interactions, with varying results [30, 31]. We have pursued intracellular backbone cyclization to generate biostable peptide libraries [32]. This procedure, termed SICLOPPS [33, 34] utilizes the Synechocystis sp. PCC6803 DnaE split intein [35] By rearranging the order of the elements of the


O Active intein

O

O HZ N H

O Thioester

Figure 9.5 SICLOPPS mechanism: the expressed fusion protein folds to form an active intein. An N- to S-acyl shift at the target N-terminal intein junction produces a thioester, which undergoes transesterification with a side chain nucleophile (serine or

IN IC

IC

IN

H2N

H N

T

H2N

H N

Z O N H

O Lariat

O

T

O

HN HZ

T

et peptide arg

N H

S NH2

NH2 O

et peptide arg

O

O HZ

T

et peptide arg

H2N

H N

N H

SH

O

O

O et peptide arg

O

O

IC

IN

HS

O Cyclic peptide

cysteine; X = O or S) at the C-terminal intein junction to form a lariat intermediate. An asparagine side chain liberates the cyclic product as a lactone, and an X- to N-acyl shift generates the thermodynamically favored lactam product in vivo.

intein, an active cis-intein (IC : target peptide : IN ) is yielded, which upon splicing results in cyclization of the target protein/peptide sequence (Figure 9.5). In order to utilize SICLOPPS for the biosynthesis of cyclic peptide libraries [33], degenerate oligonucleotides encoding those peptides are introduced between the IC and IN genes, while making sure that the correct reading frame is maintained throughout, using a PCR-based method [32]. The variable segment is encoded in the form NNS where N represents any of the four DNA bases (A, C, G, or T) and S represents C or G. The NNS sequence generates 32 codons and encodes all 20 amino acids while eliminating the ocher (UAA) and opal (UGA) stop codons from the library. The intein chemistry requires the first amino acid to be a nucleophilic cysteine or serine. There are no limits on the number of amino acids in the target peptide, allowing cyclic peptides of various sizes. We typically use five variable amino acids because the theoretical number of library members at DNA levels (34 million) is within the number of transformants that can be readily achieved. We have combined SICLOPPS with homodimeric and heterodimeric bacterial reverse two-hybrid systems (RTHSs) [28, 29]. Our RTHS is based on the bacteriophage regulatory system [36], using chimeric repressor fusions and promoter sequences to link the disruption of targeted fusion protein heterodimers to the expression of three reporter genes (Figure 9.6). The homodimeric RTHS uses the bacteriophage 434 DNA-binding protein, and the heterodimeric RTHS uses a mutant P22.434 and a wild-type 434 DNA-binding protein. HIS3 (imidazole glycerol phosphate dehydratase) [37] and KanR (aminoglycoside 3 -phosphotransferase for kanamycin resistance) are two chemically tunable, conditionally selective reporter genes. The third reporter gene, LacZ (β-galactosidase) is used to quantify the protein–protein interaction through β-galactosidase assays. Combining the aforementioned RTHS with SICLOPPS yields a powerful method in which each bacterial cell is utilized for the intracellular synthesis, and screening of a small molecule against the targeted protein–protein interaction; reporting the affinity and selectivity of potential inhibitors. This approach differs from conventional approaches to drug discovery, as traditionally, compound libraries are screened for in vitro activity with the aim of demonstrating in vivo selectivity

221

222


Y

X

Y

X

No transcription 434

P22

434

P22

434

P22

434

P22

HIS3

Kan

LacZ

Kan

LacZ

(a)

434

X

Y

434

P22

P22

434

Transcription

P22

HIS3

(b)

Figure 9.6 Heterodimeric RTHS. (a) The target proteins are expressed as fusions with the 434 or P22.434 bacteriophage DNA-binding proteins, which associate to form a functional repressor that prevents expression of the reporter genes, inhibiting growth on minimal media. (b) A small

molecule inhibits the protein–protein interaction of the target proteins (p6 and UEV), allowing transcription and translation of the reporter genes that rescue growth by induction of HIS3 and KanR . LacZ is used to quantify repression by β-galactosidase assays.

at a later stage. In our genetic selection system, inhibitors were selected on the basis of in vivo activity and screened for in vivo target selectivity intracellularly. The mode of action and site of binding are determined for the most active compounds afterward. We have used this approach to uncover inhibitors of several protein–protein interactions, two of which are outlined below. 9.3.1 Cyclic Peptide Inhibitors of AICAR Transformylase Activity

The de novo purine biosynthetic pathway [29], used by virtually all organisms to produce purine nucleotides consists of 10 enzymatic reactions that convert 5-phosphoribosyl-1-pyrophosphate to inosine monophosphate (IMP) [38]. The final two steps of this pathway (Figure 9.7) are catalyzed by the product of the purH gene 5-amino-4-imidazolecarboxamide ribonucleotide transformylase/inosine 5 -monophosphate cyclohydrolase (ATIC), a 64-kDa bifunctional protein possessing two distinct domains [39]. The C-terminal aminoimidazole-4-carboxamide ribonucleotide (AICAR) transformylase domain (residues 200–593) catalyzes the transfer of a formyl group from N10 -formyl-tetrahydrofolate (10-f-THF) to AICAR. The final step of the pathway (intramolecular cyclization of 5-formyl-AICAR) is catalyzed by the N-terminal IMP cyclohydrolase domain (residues 1–199). The individual AICAR transformylase and IMP cyclohydrolase domains have been

9.3 High-Throughput Screening of Biosynthesized Libraries O N 2-

O4P O

N

HO OH AICAR

Figure 9.7

223

O N

NH2 10-f-THF

THF

NH2

2-

O4P O

AICAR transformylase

N

HO OH

O NH2

N H

N H2O

O

2-

O4P O

IMP cyclohydrolase

N

HO OH

FAICAR

IMP

Final two steps of the de novo purine biosynthesis pathway catalyzed by ATIC.

shown to be active when expressed separately [40–42]. ATIC is highly conserved from Escherichia coli to humans. Of the two pathways available for purine synthesis, normal cells favor the salvage pathway whereas cancer cells rely on the de novo pathway [43]. Inhibiting enzymes of the de novo pathway therefore represents an attractive approach in the search for anticancer agents. Current approaches rely on folate analogs such as methotrexate, an inhibitor of dihydrofolate reductase that has been in clinical use for over 50 years [44]. Owing to its low sequence homology with other folate-dependent enzymes, specific antifolate inhibitors of ATIC have been relatively scarce [45]. As well as potential uses in the treatment of malignant diseases, ATIC inhibitors may have uses in the treatment of inflammatory diseases such as rheumatoid arthritis. Inhibition of ATIC and the resultant increase in intracellular AICAR levels also leads to the inhibition of adenosine deaminase and adenosine kinase. The subsequent build up of adenosine is in turn, thought to be responsible for the anti-inflammatory response of several nonsteroid ATIC inhibitors [46–49]. The crystal structure of ATIC [41] shows the dimer to have an interacting interface of ∼5000 A˚ 2 , demonstrating the magnitude of the challenge involved in disrupting this interaction with a small molecule. We used our genetic selection methodology outlined above, to screen two SICLOPPS libraries of around a hundred million members for inhibitors of ATIC homodimerization. The screen identified around a hundred sequences, which were ranked for their activity. The two most active cyclic peptides were chemically synthesized for testing in vitro (AICAR transformylase assay). The most potent cyclic peptide was found to have a Ki of 17 µM and we determined its mode of action by progress curve analysis [50–52]. The cyclic peptide showed competitive inhibition of enzyme dimerization with respect to 10-f-THF and noncompetitive inhibition with respect to AICAR, which was consistent with both the ordered binding observed for the enzyme and stabilization of the catalytic dimer by 10-f-THF [53, 54]. Furthermore, the Ki of c-1a obtained by this method (18 µM) closely matches that obtained assuming competitive inhibition using a linear plot. This work demonstrated the potential of our genetic selection methodology, by uncovering a six-membered cyclic peptide, capable of disrupting a protein–protein interaction with an interface of 5000 A˚ 2 . Whether the selected inhibitors prevent dimerization by binding to a hot spot [55] in the ATIC protein interface or elsewhere remains unclear. Studies are currently underway in our laboratory to develop the

NH N

224


next generation of inhibitors, and to determine the exact binding site of the inhibitor. 9.3.2 Cyclic Peptide Inhibitors of HIV Budding

Another area of interest for our group is studying, and disrupting the protein–protein interactions involved in viral life cycles [56]. In the case of HIV, the poor fidelity of genomic replication is advantageous to the virus, allowing the evolution of a number of drug-resistant mutants. Around 40% of infected individuals are estimated to harbor drug-resistant strains of HIV, with a further 5% exhibiting complete resistance to reverse transcriptase and protease inhibitor drugs. Recent advances in molecular research have allowed better understanding of the intricate interplay between viral and host proteins during the HIV life cycle, uncovering new points of possible intervention less susceptible to acquired resistance. A number of retroviruses including HIV, Ebola, and human T-cell leukemia virus type 1 contain a highly conserved P(T/S)AP tetrapeptide motif in their late budding domain, thought to be essential for viral budding. In the case of HIV the PTAP motif is found in the p6 region of Gag, the protein that drives virus assembly and release. P6 has been shown to specifically bind the UEV (Ubiquitin E2 variant) domain of TSG101 [57], a human protein residing on the late endosome, involved in the formation of multivesicular bodies. TSG101 activity is normally mediated by binding to the endosomal protein HRS (hepatocyte growth factor-regulated tyrosine kinase substrate), through a PSAP motif [58]. The HIV p6–TSG101 interaction occurs through a competing PTAP motif. Upon binding to p6, TSG101 is diverted from its normal site of action to the plasma membrane, a process that is vital to the initiation of viral budding [59, 60]. Current HIV therapies target the entry and replication steps (Figure 9.3) in the viral life cycle, with no drugs currently on the market that target HIV budding. We saw great potential in using our genetic selection methodology to uncover a cyclic peptide that disrupts the interaction of the HIV p6, with the human TSG101 protein. We therefore constructed a heterodimeric RTHS (Figure 9.6) to assay SICLOPPS libraries for inhibitors of the p6–UEV protein–protein interaction. We used a plasmid library coding for SGWXXXXX (X = any amino acid) cyclic peptides. The invariable motif of the peptide was designed to contain serine (required nucleophile for intein processing), glycine (avoids racemization during chemical synthesis), and tryptophan, which functions as a chromophore for HPLC purification. The library size recovered by electroporation into the selection strain was 1 × 108 transformants. After several rounds of secondary screening, three cyclic peptides remained, with SGWIYWNV being the most active in the p6–UEV RTHS (Figure 9.8). We also constructed and screened a SGWXXPXXPXX library, to target the PTAP binding motif on TSG101. Selectants from this library were not carried forward as they were less active than those from the fully randomized library; this highlights the power and advantage of genetic selection over conventional methodologies.


NH O

O N H O HO

HN

NH

HN

NH O

NH

O

OH

H N O

O H2N O

Figure 9.8

O HN

N H

Cyclic peptide SGWIYWNV.

The conventional approach would be to target the p6–UEV interaction with PTAP mimics, artificially limiting the available inhibitor binding sites. Genetic selection however, allows the whole surface area of the two proteins to be assayed, uncovering the most effective inhibitors from the library, regardless of binding site. The key step for us was to demonstrate that the inhibitor of the p6–UEV interaction uncovered using our genetic selection methodology, also inhibited HIV budding from host cells. The assay used monitors viral budding from human 293T cells transfected with a GFP-tagged Gag construct (Fugene transfection), which causes viruslike particle release [61]. This is based on the fact that the HIV Gag protein is itself sufficient to bud from the cell surface and give rise to particles that are morphologically similar to genuine HIV viral particles [60, 61]. The intracellular delivery of the cyclic peptides was enhanced by utilizing a cell penetrating peptide sequence derived from the HIV Tat internalization domain [62, 63]. The arginine-rich Tat sequence was synthetically bound to the cyclic peptides via a disulfide bond, formed by modifying the nonrandom region of the cyclic peptide from SGW to CGW and by adding a cysteine to the beginning of the Tat sequence. The Tat-tagged cyclic peptide (Figure 9.8) was found to inhibit Gag VLP release by more than 60% when administered to the 293T cells at 10 µM. In comparison, the Tat peptide alone (negative control) had no effect at the same concentration. The inhibitory effect of the cyclic peptide on Gag budding was found to be dose-dependent, and specific to VLP release at all concentrations, with an IC50 of 7 µM. We also assayed the cyclic peptide for its effect on the physiological function of TSG101. Epidermal growth factor receptor (EGFR) degradation is a process that is regulated by, and dependant on the TSG101–HRS interaction. We measured the effect of Tat-tagged cyclic peptide 11 on EFGR degradation in HeLa cells [58], and found that it does not affect this process. The collective data suggests that the cyclic peptide inhibitor acts by inhibiting dimerization of TSG101 and HIV p6, thus inhibiting HIV budding from 293T cells. This compound promises the possibility of targeting the viral budding of HIV,

225

226


via the p6–TSG101 interaction. As this process relies on host proteins, it is less likely to be susceptible to acquired resistance than current drugs that only target the viral proteome. We are currently attempting to evolve more potent inhibitors using second-generation SICLOPPS libraries (based on the selected sequences) and peptidomimetics [64]. The isolation of the active cyclic peptide from a library of ∼108 members demonstrates the power of genetic selection and the premise of controlling cellular processes through modulation of specific protein–protein interactions. The methodology used allows rapid identification of small molecule inhibitors of other protein–protein interactions, yielding a powerful and novel approach to drug discovery.

9.4 Future Direction

The majority of protein interaction networks contain weak spots: protein–protein interactions that are critical to a cellular process and yet are not ‘‘backed up’’ by other protein–protein interactions that would take over in case of malfunction (genetic mutations, etc.). Thus, small molecules that target such weak points have great potential to form the next generation of therapeutic agents. In the following sections we outline three areas in which inhibitors of the given protein–protein interactions would represent a major step forward from current therapies. 9.4.1 Small Molecule Inhibitors of Tumor Hypoxia Response Network

As tumors grow, they are constantly outstripping the supply of oxygen and nutrients that are delivered by the vascular network. Growth of primary and metastatic tumors is therefore dependant on angiogenesis, the development of new blood vessels from an existing vascular network, delivering oxygen and nutrients necessary for growth. The cellular hypoxia (deprivation of oxygen) response network is therefore a major driver of tumor progression. Hypoxic areas form when the growth of a tumor outstrips local neovascularization, thereby creating areas of inadequate perfusion. HIF-1 is the most potent inducer of the expression of genes coding for glycolytic enzymes, vascular endothelial growth factor (VGEF) and erythropoietin [65]. HIF-1 is a heterodimer that consists of HIF-1α whose cellular concentration is tightly regulated by oxygen, and HIF-1β, which is constitutively expressed in the nucleus. In the absence of oxygen, HIF-1α forms an active complex with HIF-1β that binds to hypoxia response element, thereby activating the expression of numerous hypoxia response genes (Figure 9.9). The potency of HIF-1 activity in tumors and its correlation to angiogenesis, tumor growth, and metastasis is well established [66, 67]. Immunohistochemical analysis of human tumor biopsies has revealed dramatic overexpression of HIF-1α in common cancers [68], which is associated with the regulation of genes involved


Normoxia

227

Hypoxia

O2

PHD2

OH OH

PHD2

OH

HIF-1a

HIF-1a

Ub VHL

Ub

VHL

HIF-1a Proteasome degradation

HIF-1a

HIF-1a Stabilization

HIF-1a

HIF-1a

HIF-1b

HIF-1a mRNA Expression of target genes Figure 9.9 The hypoxia response network. In normoxia, HIF-1α is hydroxylated by prolyl hydroxylases (PHDs), allowing the von Hipple–Lindau (VHL) protein to bind, leading to its ubiquitination and degradation. Under hypoxic conditions, the oxygen substrate for PHDs is no longer present,

resulting in stabilization and build up of HIF-1α, which translocates to the nucleus and heterodimerizes with HIF-1β, to form a transcription activator of pro-angiogenic genes, which stimulate new blood vessel formation.

in angiogenesis such as VEGF. Clinically, HIF-1α overexpression has been shown to be a marker of highly aggressive disease and is associated with poor prognosis and treatment failure in a number of cancers [69, 70]. Hypoxic cells convert to a glycolytic metabolism, become resistant to apoptosis, and are more likely to migrate to less hypoxic areas of the body (metastasis). Furthermore, the loss or gain of HIF-1 activity is negatively or positively correlated with tumor growth and angiogenesis [71, 72].

228


A small molecule inhibitor of HIF-1 heterodimerization, could potentially act as a very potent anticancer agent, inhibiting angiogenesis in tumors and preventing their growth. We are currently in the process of uncovering such agents using our genetic selection methodology, outlined previously. 9.4.2 Targeting Protein–Protein Interactions in Asthma

Asthma is a chronic disease of the lung where there is an inappropriate response to an innocuous substance. These allergens can be harmless, everyday molecules that when inhaled by an asthmatic subject, trigger an immediate and acute response by their immune system, characterized by inflammation and contraction of the muscles surrounding the airways. This intense inflammation damages the lining of the lung and over time leads to profound structural changes known as remodeling. At the root of this process (as with most other cellular processes) lie a series of protein–protein interactions that play a key role in disease development and progress. Such interactions are upstream of the majority of current drug targets, and are potentially ideal points for combating disease. It is becoming increasingly evident that many of these interactions cannot be targeted with conventional druglike molecules [73]. The promise of this approach is demonstrated by the current excitement surrounding therapeutic antibodies that target extracellular protein interactions (e.g., infliximab) [74]. Although highly specific for their molecular targets and stable in human serum, antibodies are difficult and expensive to manufacture and not cell permeable. Small molecule inhibitors that bind protein surfaces with antibody-like affinity and selectivity offer the potential to overcome such issues [75]. An important aspect of allergic inflammation in asthma is the migration of Th2 cells and eosinophils into the airway. Th2 cells support the influx of eosinophils via cytokine secretion, and are sources of interleukin-4 (IL-4) and IL-13 that have been shown to induce chemokine production in various airway cells [76–78]. IL-4 and IL-13 are two key cytokines associated with airway inflammation and airflow obstruction in asthma [79, 80]. The predominant biological effects of both IL-4 and IL-13 are driven through a heterodimeric receptor complex composed of IL-4 receptor α (IL-4Rα) and IL-13Rα1 subunits, or the IL-4Rα and γ C receptor complex (Figure 9.10). IL-4/IL-13 binding to either receptor complex instigates an intracellular signaling cascade resulting in the homodimerization of signal transducer and activator of rranscription factor 6 (STAT-6) (Figure 9.10) [81]. This is thought to be critical for Th2 differentiation of T helper cells and essential for isotype class switching to IgE. STAT-6 expression has furthermore shown to be required for the development of experimental allergic asthma in mice [81, 82]. IL-4 and IL-13 are proposed to play key roles in the pathogenesis of severe asthma, through modulating airway hyperresponsiveness, subepithelial fibrosis, and mucus production [83]. There has been considerable effort directed toward therapies based on the inhibition of either IL-4 or IL-13 by soluble receptor proteins or antibodies


TYK2

JAK1

P

P

P Homodimerization

P

STAT-6

P STAT-6

STAT-6

STAT-6

P

STAT-6

STAT-6

P

IL-13

IL-4Ra

P

IL-13Ra1

JAK1 IL-4Ra

JAK3

YC

P

or

IL-4

IL-4

P Inflammation and fibrosis

Figure 9.10 The IL-4 and IL-13 signaling pathways in asthma. Binding of the interleukins to the receptor complex activates phosphorylation of janus kinase (Jak) 1 and 3, and tyrosine kinase (Tyk) 2. These phosphorylated sites recruit STAT-6, which

leads to its phosphorylation and subsequent dimerization. STAT-6 dimers translocate to the nucleus, where they activate transcription of a number of target genes that lead to inflammation and fibrosis.

[84]. The shared receptor complex of IL-4Rα and IL-13Rα1, however, leads to many overlapping functions and thus renders single-target therapies largely ineffective. Previous work has shown that the expression of STAT-6 in the bronchial epithelium is higher in subjects with severe asthma than in subjects with mild asthma or normal controls [85]. Thus, this presents an ideal therapeutic target for the treatment of asthma. The ability to develop small molecule antagonists has significant advantages over the current therapies, which are expensive to manufacture and require either repeated hospital visits or daily injections. As both cytokines are thought to predominantly instigate their intracellular signaling cascade via activation of STAT-6, this seems an ideal point for intervention. Our laboratory is currently studying the role of these protein–protein interactions in asthma, and developing small molecule inhibitors.

229

230


9.4.3 Targeting the Protein Interaction Networks of Influenza Virus

In humans, influenza viruses are common and highly contagious pathogens of the upper respiratory tract, and seasonal epidemics affect 10–20% of the general population. However, the virus can also be deadly, with pandemic epidemics that killed up to a million people worldwide. All human influenza viruses belong to the family Orthomyxoviridae. There are three genera, corresponding to influenza types A, B, and C. Influenza A viruses, uniquely, are divided into subtypes based on major antigenic specificities of their hemagglutinin (HA) and neuraminidase (NA) proteins. Type A is the most important of the influenza viruses, causing alternate annual outbreaks and epidemics during the winter seasons. Pandemics of influenza A have occurred about three times per century since 1700 and were manifested by the worldwide spread of the disease, typically with high morbidity and mortality. The most extreme pandemic in the twentieth century was that of Spanish influenza of 1918–1919, which killed an estimated 20–40 million people throughout the world [86]. Emergence of new influenza strains in the human population occurs via transmission from other animal species, especially poultry. Typically, human and avian influenza viruses are different with little cross-species infection. Occasionally, direct avian–human transmission can occur, often with enhanced pathogenicity, as demonstrated by the emergence of the recent H5N1 avian influenza in many countries throughout Southeast Asia; human-to-human spread of H5N1 avian influenza has not been conclusively documented [87]. The genome of influenza A viruses consists of eight separate segments of single-stranded, negative-sense RNA encoding for viral RNA polymerase subunits and structural and functional proteins. Mutations resulting from the error-prone replication of the single-stranded RNA, and an ability to produce genetic hybrid viruses by the mixing of gene segments from two viruses (when dual infection occurs), contribute to the evolutionary success of these viruses and the difficulty associated with the development of anti-influenza therapies. Influenza viruses, like other enveloped viruses, use a membrane fusion strategy to deliver their genomes to the cytosol of target cells. Receptor-binding and fusion are mediated by the viral HA. HA binds to sialic acid (SA) residues on glycoproteins or glycolipids on the host cell surface. The specificity of the HA determines the tropism of the virus with respect to species and target cells [88]. One potential approach to antiviral therapy involves the inhibition of virus/host cell binding via blocking of the viral receptor. Ideal preventative agents must be nontoxic to mammalian cells, broadly specific across viral subtypes, and cost effective. Previous reports have demonstrated that millimolar concentrations of solubilized monomeric SA prevent influenza A agglutination of chicken erythrocytes (cRBC), presumably by binding to the HA molecule on the virus. However, each influenza virion uses about 300–600 envelope HA spikes to develop strong polyvalent binding to host cell receptors, while the binding of soluble monomeric SA derivative to HA is very weak. Thus, nonphysiological and potentially toxic concentrations of soluble monomeric SA are


required to competitively disrupt this polyvalent interaction and prevent viral internalization into the cells. To circumvent these problems, multiple sialyl moieties were conjugated to a synthetic polymer backbone to provide a polyvalent inhibitor. Even if ‘‘polymeric inhibitors’’ based on polyacrylamide backbones are more effective than monomeric SA at inhibiting influenza-induced agglutination of red blood cells, they were toxic to mammalian cells [89]. Other works have focused on the acidification of the endosome compartment: it is generally accepted that the M2 proton channel is inhibited by amantadine and its derivate rimantadine [90, 91]. Both compounds appear to be of prophylactic activity against influenza A, and may be important in the treatment of immunocompromised and elderly patients. They show, however, the following disadvantages: induction of viral resistance and exhibition of adverse effects on the central nervous and gastrointestinal system [92]. Fusion of influenza virus from within endosomes is inhibited by agents that raise the intraendosomal pH, such as NH4 Cl, monensin, or chloroquine or those that could bind selectively to a pocket of the HA2 chain, thereby stabilizing its nonfusogenic conformation and inhibiting the exposure of the fusion peptide [92, 93]. Influenza viruses replicate and transcribe their segmented negative-sense single-stranded RNA genome in the nucleus of the infected host cells, processes that are performed by the virally encoded RNA-dependent RNA polymerase, which is a trimeric complex composed by PB1, PB2, and PA [94]. PB1 is the central protein, containing to different domains interacting with PB2 and PA. In a recent work, Ghanem et al. [95] have shown that peptides that specifically bind to the protein–protein interaction domains in these subunits interfere with polymerase complex assembly and viral replication. In the host nucleus, the viral primary transcription produces proteins necessary for replication. Viral transcription is however, critically dependent on on-going cellular transcription, in particular, on the activities associated with the cellular DNA-dependent RNA polymerase II (Pol II) [96]. Thus if cellular Pol II is blocked by specific inhibitors, no viral mRNA can be synthesized [97]. The viral polymerase complex is not the only viral component shown to interact with the host Pol II transcription machinery. Many viruses encode proteins that modify the host transcription or translation apparatus to favor the synthesis of viral proteins over those of the host cell. As a result, the synthetic capability of the host cell is directed principally to the production of new virus particles [96]. The viral NS1, a multifunctional 26-kDa protein, was reported to inhibit host immune responses [98]. NS1 binds and sequesters dsRNA, interfering with the host mRNA processing and facilitating the preferential viral mRNA translation. The combination of all these functions makes the NS1 protein a very potent inhibitor of immunity and allows the influenza virus to efficiently escape the immune surveillance and to establish infection in the host. It is however clear that NS1 also acts directly to modulate other important aspects of the virus replication cycle, including viral RNA replication, viral protein synthesis, and general host cell physiology. Two important domains have been described in NS1 accomplishing its multiple functions: the N-terminal RNA-binding domain (residues 1–73) and the C-terminal effector domain (residues 74–230), which predominantly mediates interactions

231

232


with host cell proteins, and also stabilizes the RNA-binding domain [99]. The presence of different domains, the resolution of its crystal structure, and the discovery of its interaction with host components provide several potential targets for the development of antiviral compounds. Given the numerous roles of NS1 during virus replication, one potential target for anti-influenza drug design may be to disrupt conserved interactions of NS1 with RNA or its protein–protein interaction with cellular and viral factors. In this regard, peptide-mediated inhibition of the NS1–CPSF30 interaction has recently been described as a ‘‘proof-of-principle’’ approach to limit virus replication in tissue culture. Unfortunately, such a virus-specific strategy allows for virus mutation and the development of drug resistance [100]. Other works have been focused on the inhibition of NS1 binding to vRNA [101]. A recent report outlined the use of a yeast-based assay to identify compounds that phenotypically suppress NS1 function [102]. Interestingly, in this study, cells lacking an interferon response were drug resistant, suggesting that the compounds block interactions between NS1 and the interferon system. The effects of the compounds were specific to NS1, because they had no effect on the ability of the severe acute respiratory syndrome coronavirus papain-like protease protein to block β-interferon promoter activation. These recent findings suggest that function of NS1 can be modulated by chemical inhibitors and that such inhibitors will be useful probes of biological function, and starting points for development of anti-influenza therapies [102]. A number of other host factors have also been shown to interact with influenza viral components, and for some of these interactions, there is evidence that they have functional significance during the viral life cycle. For others, however, much needs to be learned before we can fully appreciate their role in the infected cell for either the viral RNA synthesis machinery or for the virus’ attempts at thwarting normal cellular functions. The development of proteomic tools in recent years will certainly aid our attempts to identify some of these physical interactions. The greatest challenge, however, remains in establishing what, if any, role a newly discovered interacting partner plays in the influenza life cycle and to selective and specific inhibitors of their function. Acknowledgments

A.T. would like to thank Prof. Stephen Benkovic for his guidance and support. We would also like to thank Prof. Stanley Cohen, Prof. Alex Horswill, Prof. Sergey Savinov, Dr Todd Naumann, and Prof. Quan Lu. References 1. Arkin, M.R. and Wells, J.A.

(2004) Small-molecule inhibitors of protein-protein interactions: progressing towards the dream. Nat. Rev. Drug Discov., 3, 301–317.

2. Wells, J.A. and McClendon, C.L.

(2007) Reaching for high-hanging fruit in drug discovery at proteinprotein interfaces. Nature, 450, 1001–1009.

References 3. Young, L., Jernigan, R.L., and Covell,

4.

5.

6.

7.

8.

9.

10.

11.

12.

D.G. (1994) A role for surface hydrophobicity in protein-protein recognition. Protein Sci., 3, 717–729. Moreira, I.S., Fernandes, P.A., and Ramos, M.J. (2007) Hot spots–a review of the protein-protein interface determinant amino-acid residues. Proteins, 68, 803–812. Gao, Y., Dickerson, J.B., Guo, F., Zheng, J., and Zheng, Y. (2004) Rational design and characterization of a Rac GTPase-specific small molecule inhibitor. Proc. Natl. Acad. Sci. U.S.A., 101, 7618–7623. Trosset, J.Y., Dalvit, C., Knapp, S., Fasolini, M., Veronesi, M., Mantegani, S., Gianellini, L.M., Catana, C., Sundstrom, M., Stouten, P.F., and Moll, J.K. (2006) Inhibition of protein-protein interactions: the discovery of druglike beta-catenin inhibitors by combining virtual and biophysical screening. Proteins, 64, 60–67. Enyedy, I.J., Ling, Y., Nacro, K., Tomita, Y., Wu, X., Cao, Y., Guo, R., Li, B., Zhu, X., Huang, Y., Long, Y.Q., Roller, P.P., Yang, D., and Wang, S. (2001) Discovery of small-molecule inhibitors of Bcl-2 through structure-based computer screening. J. Med. Chem., 44, 4313–4324. Boldt, G.E., Dickerson, T.J., and Janda, K.D. (2006) Emerging chemical and biological approaches for the preparation of discovery libraries. Drug Discov. Today, 11, 143–148. Borghouts, C., Kunz, C., and Groner, B. (2005) Peptide aptamers: recent developments for cancer therapy. Expert Opin. Biol. Ther., 5, 783–797. Chang, K., Elledge, S.J., and Hannon, G.J. (2006) Lessons from nature: microRNA-based shRNA libraries. Nat. Methods, 3, 707–714. Matsumoto, S., Akashi, H., and Taira, K. (2006) Screening and determination of gene function using randomized ribozyme and siRNA. libraries Handb. Exp. Pharmacol., 173, 197–221. Rothe, A., Hosse, R.J., and Power, B.E. (2006) Ribosome display for improved

13.

14.

15. 16.

17.

18.

19.

20.

21.

22.

biotherapeutic molecules. Expert Opin. Biol. Ther., 6, 177–187. Baldwin, J.J. (1996) Design, synthesis and use of binary encoded synthetic chemical libraries. Mol. Divers., 2, 81–88. Edwards, P.J. and Morrell, A.I. (2002) Solid-phase compound library synthesis in drug design and development. Curr. Opin. Drug Discov. Dev., 5, 594–605. Huwe, C.M. (2006) Synthetic library design. Drug Discov. Today, 11, 763–767. Tan, D.S. (2005) Diversity-oriented synthesis: exploring the intersections between chemistry and biology. Nat. Chem. Biol., 1, 74–84. Webb, T.R. (2005) Current directions in the evolution of compound libraries. Curr. Opin. Drug Discov. Dev., 8, 303–308. Irwin, J.J. (2006) How good is your screening library? Curr. Opin. Chem. Biol., 10, 352–356. Berg, T., Cohen, S.B., Desharnais, J., Sonderegger, C., Maslyar, D.J., Goldberg, J., Boger, D.L., and Vogt, P.K. (2002) Small-molecule antagonists of Myc/Max dimerization inhibit Myc-induced transformation of chicken embryo fibroblasts. Proc. Natl. Acad. Sci. U.S.A., 99, 3830–3835. Vogelstein, B., Lane, D., and Levine, A.J. (2000) Surfing the p53 network. Nature, 408, 307–310. Vassilev, L.T., Vu, B.T., Graves, B., Carvajal, D., Podlaski, F., Filipovic, Z., Kong, N., Kammlott, U., Lukacs, C., Klein, C., Fotouhi, N., and Liu, E.A. (2004) In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science, 303, 844–848. Grasberger, B.L., Lu, T., Schubert, C., Parks, D.J., Carver, T.E., Koblish, H.K., Cummings, M.D., LaFrance, L.V., Milkiewicz, K.L., Calvo, R.R., Maguire, D., Lattanze, J., Franks, C.F., Zhao, S., Ramachandren, K., Bylebyl, G.R., Zhang, M., Manthey, C.L., Petrella, E.C., Pantoliano, M.W., Deckman, I.C., Spurlino, J.C., Maroney, A.C., Tomczuk, B.E., Molloy, C.J., and Bone, R.F. (2005) Discovery and cocrystal structure of benzodiazepinedione

233

234


23.

24.

25.

26.

27.

28.

29.

30.

31.

HDM2 antagonists that activate p53 in cells. J. Med. Chem., 48, 909–912. Kussie, P.H., Gorina, S., Marechal, V., Elenbaas, B., Moreau, J., Levine, A.J., and Pavletich, N.P. (1996) Structure of the MDM2 oncoprotein bound to the p53 tumour suppressor transactivation domain. Science, 274, 948–953. Dorr, P., Westby, M., Dobbs, S., Griffin, P., Irvine, B., Macartney, M., Mori, J., Rickett, G., Smith-Burchnell, C., Napier, C., Webster, R., Armour, D., Price, D., Stammen, B., Wood, A., and Perros, M. (2005) Maraviroc (UK-427,857), a potent, orally bioavailable, and selective small-molecule inhibitor of chemokine receptor CCR5 with broad-spectrum anti-human immunodeficiency virus type 1 activity. Antimicrob. Agents Chemother., 49, 4721–4732. Falciani, C., Lozzi, L., Pini, A., and Bracci, L. (2005) Bioactive peptides from libraries. Chem. Biol., 12, 417–426. Ulrich, H. (2005) DNA and RNA aptamers as modulators of protein function. Med. Chem., 1, 199–208. Uttamchandani, M., Wang, J., and Yao, S.Q. (2006) Protein and small molecule microarrays: powerful tools for high-throughput proteomics. Mol. Biosyst., 2, 58–68. Horswill, A.R., Savinov, S.N., and Benkovic, S.J. (2004) A systematic method for identifying small-molecule modulators of protein-protein interactions. Proc. Natl. Acad. Sci. U.S.A., 101, 15591–15596. Tavassoli, A. and Benkovic, S.J. (2005) Genetically selected cyclic-peptide inhibitors of AICAR transformylase homodimerization. Angew. Chem. Int. Ed. Engl., 44, 2760–2763. Gururaja, T.L., Narasimhamurthy, S., Payan, D.G., and Anderson, D.C. (2000) A novel artificial loop scaffold for the noncovalent constraint of peptides. Chem. Biol., 7, 515–527. Jermutus, L., Honegger, A., Schwesinger, F., Hanes, J., and Pluckthun, A. (2001) Tailoring in vitro evolution for protein affinity or

32.

33.

34.

35.

36.

37.

38.

39.

40.

stability. Proc. Natl. Acad. Sci. U.S.A., 98, 75–80. Tavassoli, A. and Benkovic, S.J. (2007) Split-intein mediated circular ligation used in the synthesis of cyclic peptide libraries in E. coli. Nat. Protoc., 2, 1126–1133. Scott, C.P., Abel-Santos, E., Jones, A.D., and Benkovic, S.J. (2001) Structural requirements for the biosynthesis of backbone cyclic peptide libraries. Chem. Biol., 8, 801–815. Scott, C.P., Abel-Santos, E., Wall, M., Wahnon, D.C., and Benkovic, S.J. (1999) Production of cyclic peptides and proteins in vivo. Proc. Natl. Acad. Sci. U.S.A., 96, 13638–13643. Wu, H., Hu, Z., and Liu, X.Q. (1998) Protein trans-splicing by a split intein encoded in a split DnaE gene of Synechocystis sp. PCC6803. Proc. Natl. Acad. Sci. U.S.A., 95, 9226–9231. Hu, J.C., O’Shea, E.K., Kim, P.S., and Sauer, R.T. (1990) Sequence requirements for coiled-coils: analysis with lambda repressor-GCN4 leucine zipper fusions. Science, 250, 1400–1403. Joung, J.K., Ramm, E.I., and Pabo, C.O. (2000) A bacterial two-hybrid selection system for studying protein-DNA and protein-protein interactions. Proc. Natl. Acad. Sci. U.S.A., 97, 7382–7387. Zalkin, H. and Dixon, J.E. (1992) De novo purine nucleotide biosynthesis. Prog. Nucleic Acid Res. Mol. Biol., 42, 259–287. Ni, L., Guan, K., Zalkin, H., and Dixon, J.E. (1991) De novo purine nucleotide biosynthesis: cloning, sequencing and expression of a chicken PurH cDNA encoding 5-aminoimidazole-4-carboxamideribonucleotide transformylase-IMP cyclohydrolase. Gene, 106, 197–205. Rayl, E.A., Moroson, B.A., and Beardsley, G.P. (1996) The human purH gene product, 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase/IMP cyclohydrolase. Cloning, sequencing, expression, purification, kinetic analysis, and domain mapping. J. Biol. Chem., 271, 2225–2233.

References 41. Greasley, S.E., Horton, P., Ramcharan,

42.

43.

44.

45.

46.

47.

48.

49.

J., Beardsley, G.P., Benkovic, S.J., and Wilson, I.A. (2001) Crystal structure of a bifunctional transformylase and cyclohydrolase enzyme in purine biosynthesis. Nat. Struct. Biol., 8, 402–406. Wolan, D.W., Greasley, S.E., Beardsley, G.P., and Wilson, I.A. (2002) Structural insights into the avian AICAR transformylase mechanism. Biochemistry, 41, 15505–15513. Jackson, R.C. and Harkrader, R.J. (1981) in Nucleosides and Cancer Treatment (eds M.H.N. Tattersall and R.M. Fox), Academic Press, Sydney, pp. 18–31. McGuire, J.J. (2003) Anticancer antifolates: current status and future directions. Curr. Pharm. Des., 9, 2593–2613. Cheong, C.G., Wolan, D.W., Greasley, S.E., Horton, P.A., Beardsley, G.P., and Wilson, I.A. (2004) Crystal structures of human bifunctional enzyme aminoimidazole-4-carboxamide ribonucleotide transformylase/IMP cyclohydrolase in complex with potent sulfonyl-containing antifolates. J. Biol. Chem., 279, 18034–18045. Baggott, J.E., Morgan, S.L., Ha, T., Vaughn, W.H., and Hine, R.J. (1992) Inhibition of folate-dependent enzymes by non-steroidal anti-inflammatory drugs. Biochem. J., 282 (Pt 1), 197–202. Allegra, C.J., Drake, J.C., Jolivet, J., and Chabner, B.A. (1985) Inhibition of phosphoribosylaminoimidazolecarboxamide transformylase by methotrexate and dihydrofolic acid polyglutamates. Proc. Natl. Acad. Sci. U.S.A., 82, 4881–4885. Cronstein, B.N., Naime, D., and Ostad, E. (1993) The antiinflammatory mechanism of methotrexate. Increased adenosine release at inflamed sites diminishes leukocyte accumulation in an in vivo model of inflammation. J. Clin. Invest., 92, 2675–2682. Gadangi, P., Longaker, M., Naime, D., Levin, R.I., Recht, P.A., Montesinos, M.C., Buckley, M.T., Carlin, G., and Cronstein, B.N. (1996) The

50.

51.

52.

53.

54.

55.

56.

57.

anti-inflammatory mechanism of sulfasalazine is related to adenosine release at inflamed sites. J. Immunol., 156, 1937–1941. Kuzmic, P. (1996) Program DYNAFIT for the analysis of enzyme kinetic data: application to HIV proteinase. Anal. Biochem., 237, 260–273. Stone, S.R. and Fromm, H.J. (1980) Studies on the mechanism of adenosine 5 -monophosphate inhibition of bovine liver fructose 1,6-bisphosphatase. Biochemistry, 19, 620–625. Bauer, M., Griengl, H., and Steiner, W. (1999) Kinetic studies on the enzyme (S)-hydroxynitrile lyase from hevea brasiliensis using initial rate methods and progress curve analysis. Biotechnol. Bioeng., 62, 20–29. Vergis, J.M., Bulock, K.G., Fleming, K.G., and Beardsley, G.P. (2001) Human 5-aminoimidazole-4-carboxamide ribonucleotide transformylase/inosine 5 -monophosphate cyclohydrolase. A bifunctional protein requiring dimerization for transformylase activity but not for cyclohydrolase activity. J. Biol. Chem., 276, 7727–7733. Bulock, K.G., Beardsley, G.P., and Anderson, K.S. (2002) The kinetic mechanism of the human bifunctional enzyme ATIC (5-amino-4-imidazolecarboxamide ribonucleotide transformylase/inosine 5 -monophosphate cyclohydrolase). A surprising lack of substrate channeling. J. Biol. Chem., 277, 22168–22174. DeLano, W.L. (2002) Unraveling hot spots in binding interfaces: progress and challenges. Curr. Opin. Struct. Biol., 12, 14–20. Tavassoli, A., Lu, Q., Gam, J., Pan, H., Benkovic, S.J., and Cohen, S.N. (2008) Inhibition of HIV budding by a genetically selected cyclic peptide targeting the Gag-TSG101 interaction. ACS Chem. Biol., 3, 757–764. Pornillos, O., Alam, S.L., Davis, D.R., and Sundquist, W.I. (2002) Structure of the Tsg101 UEV domain in complex with the PTAP motif of the HIV-1 p6 protein. Nat. Struct. Biol., 9, 812–817.

235

236

9 Protein Interactions with Small Molecules 58. Lu, Q., Hope, L.W., Brasch, M.,

59.

60.

61.

62.

63.

64.

65.

66.

67.

Reinhard, C., and Cohen, S.N. (2003) TSG101 interaction with HRS mediates endosomal trafficking and receptor down-regulation. Proc. Natl. Acad. Sci. U.S.A., 100, 7626–7631. Demirov, D.G., Ono, A., Orenstein, J.M., and Freed, E.O. (2002) Overexpression of the N-terminal domain of TSG101 inhibits HIV-1 budding by blocking late domain function. Proc. Natl. Acad. Sci. U.S.A., 99, 955–960. Garrus, J.E., von Schwedler, U.K., Pornillos, O.W., Morham, S.G., Zavitz, K.H., Wang, H.E., Wettstein, D.A., Stray, K.M., Cote, M., Rich, R.L., Myszka, D.G., and Sundquist, W.I. (2001) Tsg101 and the vacuolar protein sorting pathway are essential for HIV-1 budding. Cell, 107, 55–65. Hermida-Matsumoto, L. and Resh, M.D. (2000) Localization of human immunodeficiency virus type 1 Gag and Env at the plasma membrane by confocal imaging. J. Virol., 74, 8670–8679. Mae, M. and Langel, U. (2006) Cell-penetrating peptides as vectors for peptide, protein and oligonucleotide delivery. Curr. Opin. Pharmacol., 6, 509–514. Mie, M., Takahashi, F., Funabashi, H., Yanagida, Y., Aizawa, M., and Kobatake, E. (2003) Intracellular delivery of antibodies using TAT fusion protein A. Biochem. Biophys. Res. Commun., 310, 730–734. Andronati, S.A., Karaseva, T.L., and Krysko, A.A. (2004) Peptidomimetics – antagonists of the fibrinogen receptors: molecular design, structures, properties and therapeutic applications. Curr. Med. Chem., 11, 1183–1211. Dery, M.A., Michaud, M.D., and Richard, D.E. (2005) Hypoxia-inducible factor 1: regulation by hypoxic and non-hypoxic activators. Int. J. Biochem. Cell Biol., 37, 535–540. Hewitson, K.S. and Schofield, C.J. (2004) The HIF pathway as a therapeutic target. Drug Discov. Today, 9, 704–711. Shi, Y.H. and Fang, W.G. (2004) Hypoxia-inducible factor-1 in tumour

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

angiogenesis. World J. Gastroenterol., 10, 1082–1087. Melillo, G. (2006) Inhibiting hypoxia-inducible factor 1 for cancer therapy. Mol. Cancer Res., 4, 601–605. Jensen, R.L. (2006) Hypoxia in the tumourigenesis of gliomas and as a potential target for therapeutic measures. Neurosurg. Focus, 20, E24 . Lauzier, M.C., Michaud, M.D., Dery, M.A., and Richard, D.E. (2006) [HIF-1 activation during tumour progression: implications and consequences]. Bull. Cancer, 93, 349–356. Bardos, J.I. and Ashcroft, M. (2005) Negative and positive regulation of HIF-1: a complex network. Biochim. Biophys. Acta, 1755, 107–120. Escuin, D., Simons, J.W., and Giannakakou, P. (2004) Exploitation of the HIF axis for cancer therapy. Cancer Biol. Ther., 3, 608–611. Hopkins, A.L. and Groom, C.R. (2002) The druggable genome. Nat. Rev. Drug Discov., 1, 727–730. Shealy, D.J. and Visvanathan, S. (2008) Anti-TNF antibodies: lessons from the past, roadmap for the future. Handb. Exp. Pharmacol., 181, 101–129. Millward, S.W., Fiacco, S., Austin, R.J., and Roberts, R.W. (2007) Design of cyclic peptides that bind protein surfaces with antibody-like affinity. ACS Chem. Biol., 2, 625–634. Komiya, A., Nagase, H., Yamada, H., Sekiya, T., Yamaguchi, M., Sano, Y., Hanai, N., Furuya, A., Ohta, K., Matsushima, K., Yoshie, O., Yamamoto, K., and Hirai, K. (2003) Concerted expression of eotaxin-1, eotaxin-2, and eotaxin-3 in human bronchial epithelial cells. Cell. Immunol., 225, 91–100. Yuan, Q., Campanella, G.S., Colvin, R.A., Hamilos, D.L., Jones, K.J., Mathew, A., Means, T.K., and Luster, A.D. (2006) Membrane-bound eotaxin-3 mediates eosinophil transepithelial migration in IL-4-stimulated epithelial cells. Eur. J. Immunol., 36, 2700–2714. Ip, W.K., Wong, C.K., and Lam, C.W. (2006) Interleukin (IL)-4 and IL-13 up-regulate monocyte chemoattractant protein-1 expression in human

References

79.

80.

81.

82.

83.

84.

85.

86.

bronchial epithelial cells: involvement of p38 mitogen-activated protein kinase, extracellular signal-regulated kinase 1/2 and Janus kinase-2 but not c-Jun NH2-terminal kinase 1/2 signalling pathways. Clin. Exp. Immunol., 145, 162–172. Dabbagh, K., Takeyama, K., Lee, H.M., Ueki, I.F., Lausier, J.A., and Nadel, J.A. (1999) IL-4 induces mucin gene expression and goblet cell metaplasia in vitro and in vivo. J. Immunol., 162, 6233–6237. Doucet, C., Brouty-Boye, D., Pottin-Clemenceau, C., Canonica, G.W., Jasmin, C., and Azzarone, B. (1998) Interleukin (IL) 4 and IL-13 act on human lung fibroblasts. Implication in asthma. J. Clin. Invest., 101, 2129–2139. Kuperman, D.A. and Schleimer, R.P. (2008) Interleukin-4, interleukin-13, signal transducer and activator of transcription factor 6, and allergic asthma. Curr. Mol. Med., 8, 384–392. Kuperman, D., Schofield, B., Wills-Karp, M., and Grusby, M.J. (1998) Signal transducer and activator of transcription factor 6 (Stat6)-deficient mice are protected from antigen-induced airway hyperresponsiveness and mucus production. J. Exp. Med., 187, 939–948. Wills-Karp, M. (2004) Interleukin-13 in asthma pathogenesis. Immunol. Rev., 202, 175–190. Kasaian, M.T. and Miller, D.K. (2008) IL-13 as a therapeutic target for respiratory disease. Biochem. Pharmacol., 76, 147–155. Mullings, R.E., Wilson, S.J., Puddicombe, S.M., Lordan, J.L., Bucchieri, F., Djukanovic, R., Howarth, P.H., Harper, S., Holgate, S.T., and Davies, D.E. (2001) Signal transducer and activator of transcription 6 (STAT-6) expression and function in asthmatic bronchial epithelium. J. Allergy Clin. Immunol., 108, 832–838. Hilleman, M.R. (2002) Realities and enigmas of human viral influenza: pathogenesis, epidemiology and control. Vaccine, 20, 3068–3087.

87. Hampson, A.W. and Mackenzie, J.S.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

(2006) The influenza viruses. Med. J. Aust., 185, S39–S43. Sharon, N. and Lis, H. (2004) History of lectins: from hemagglutinins to biological recognition molecules. Glycobiology, 14, 53R–562. Reuter, J.D., Myc, A., Hayes, M.M., Gan, Z., Roy, R., Qin, D., Yin, R., Piehler, L.T., Esfand, R., Tomalia, D.A., and Baker, J.R. (1999) Inhibition of viral adhesion and infection by sialic-acid-conjugated dendritic polymers. Bioconjugate Chem., 10, 271–278. Pinto, L.H. and Lamb, R.A. (2007) Controlling influenza virus replication by inhibiting its proton channel. Mol. Biosyst., 3, 18–23. Samuel, C.E. (2006) Virus-host interaction minireview series: human immunodeficiency virus, hepatitis C virus, and influenza virus. J. Biol. Chem., 281, 8305–8307. Luscher-Mattli, M. (2000) Influenza chemotherapy: a review of the present state of art and of new drugs in development. Arch. Virol., 145, 2233–2248. Huckriede, A., Bungener, L., Stegmann, T., Daemen, T., Medema, J., Palache, A.M., and Wilschut, J. (2005) The virosome concept for influenza vaccines. Vaccine, 23 (Suppl 1), S26–S38. Engelhardt, O.G. and Fodor, E. (2006) Functional association between viral and cellular transcription during influenza virus infection. Rev. Med. Virol., 16, 329–345. Ghanem, A., Mayer, D., Chase, G., Tegge, W., Frank, R., Kochs, G., Garcia-Sastre, A., and Schwemmle, M. (2007) Peptide-mediated interference with influenza A virus polymerase. J. Virol., 81, 7801–7804. Amorim, M.J. and Digard, P. (2006) Influenza A virus and the cell nucleus. Vaccine, 24, 6651–6655. Chan, A.Y., Vreede, F.T., Smith, M., Engelhardt, O.G., and Fodor, E. (2006) Influenza virus inhibits RNA polymerase II elongation. Virology, 351, 210–217.

237

238

9 Protein Interactions with Small Molecules 98. Kochs, G., Garcia-Sastre, A. and

Martinez-Sobrido, L. (2007) Multiple anti-interferon actions of the influenza A virus NS1 protein. J. Virol., 81, 7011–7021. 99. Lin, D., Lan, J., and Zhang, Z. (2007) Structure and function of the NS1 protein of influenza A virus. Acta Biochim. Biophys. Sin., 39, 155–162. 100. De Clercq, E. (2006) Antiviral agents active against influenza A viruses. Nat. Rev. Drug Discov., 5, 1015–1025.

101. Maroto, M., Fernandez, Y., Ortin, J.,

Pelaez, F., and Cabello, M.A. (2008) Development of an HTS assay for the search of anti-influenza agents targeting the interaction of viral RNA with the NS1 protein. J. Biomol. Screen., 13, 581–590. 102. Basu, D., Walkiewicz, M.P., Frieman, M., Baric, R.S., Auble, D.T., and Engel, D.A. (2009) Novel influenza virus NS1 antagonists block replication and restore innate immune function. J. Virol., 83, 1881–1891.

239

10 Cracking the Glycocode: Recent Developments in Glycomics Lars Hillringhaus and Jürgen Seibel

10.1 State of the Art 10.1.1 Introduction

Cell surfaces are covered by various glycans, which characterize the status of the cell and participate in molecular recognition processes involved in embryonic development, cancer cell metastasis, bacterial and viral infections, and the initiation of immune responses [1]. Changes in the presented glycans are associated with certain disease states like cancer and inflammation. In vertebrates there exist different types of glycoconjugates like glycoproteins, glycolipids, proteoglycans, and glycosylphophatidylinositol anchors. Unlike nucleic acids and proteins, which are linear polymers, glycans exist as complex branched structures. Additionally, they are not directly encoded by the genome but synthesized by the combined interaction of carbohydrate synthesizing and modifying enzymes. Owing to their enormous structural diversity they are predestinated to encode biological information. To decipher this so-called glycocode and determine the correlation between specific glycans and biological events is the goal of glycomics. Several tools have been developed to analyze the glycome ranging from high-throughput analysis on microarrays to in vivo imaging of dynamic changes in the glycome. Cracking the glycocode should enhance the development of carbohydrate-based drugs against infectious diseases and cancer as well as tools for early-stage diagnosis. 10.1.2 Carbohydrate-Based Drugs

As glycans are involved in numerous biological processes they have an enormous potential as therapeutic agents. Nevertheless, the number of carbohydrate-based therapeutics in the market is relatively low, which is mainly due to the difficulties in the analysis and synthesis of glycans. Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

240

10 Cracking the Glycocode: Recent Developments in Glycomics

Several small molecules containing carbohydrate moieties are used as therapeutics against infectious diseases and cancer. The most famous ones are zanamivir (Relenza) and oseltamivir (Tamiflu), which are used in the treatment of influenza. Both are structural analogs of neuraminic acid and inhibit influenza neuraminidase, which is essential for the release of replicated viruses from the host cell. Aminoglycosides belong to a diverse class of aminosugar-containing glycans. They are used to treat bacterial infections as they inhibit the biosynthesis of bacterial proteins by binding bacterial ribosomes. Acarbose is used for the treatment of type II diabetes. It functions as an α-glucosidase and an α-amylase inhibitor regulating intestinal carbohydrate digestion, thereby reducing glucose absorption. It is a pseudotetrasaccharide from microbial origin and is produced by fermentation. The glycosaminoglycan heparin has been used as an antithrombic agent since the 1940s. It activates the serine protease inhibitor antithrombin III, which blocks thrombin and factor Xa in the coagulation cascade [2]. Heparin, which is isolated from animal organs is a heterogeneous mixture of polysaccharides and severe side effects have been observed [3]. By chemical and enzymatic fragmentation low-molecular-weight heparins (LMWHs) can be obtained, which have a longer half-life and fewer side effects, but are still heterogeneous and less active than heparin. In the early 1980s, a pentasaccharide was discovered as the critical sequence that is responsible for the anticoagulant activity of heparin [4]. A synthetic analog of the pentasaccharide is in the market since 2002, and it was shown that its use in major orthopedic surgery led to a decrease in the risk of thrombosis by more than 50% compared to LMWH [5]. Erythropoietin (EPO) is the most widely produced recombinant protein drug in the world and is used in the treatment of anemia. Its in vivo activity and half-life could be increased by the introduction of new glycosylation sites [6]. The surfaces of viruses, parasites, and bacteria are covered by glycans, which usually differ from those of the host cells. Several vaccines containing these glycan motifs have been developed against bacterial and infectious diseases [7]. As tumor cells have altered glycans on their surfaces, vaccines containing these structures were shown to elicit immune responses and are actually used in clinical trials [8, 9]. 10.1.3 Carbohydrate Synthesis

Glycans are difficult to isolate as they are only displayed in small quantities on cell surfaces. Additionally there exist numerous glycoforms of each glycoprotein which differ in their glycosylation pattern and sites. Defined glycan structures are essential to study the roles of glycans in biological processes in detail. Whereas nucleic acids can be synthesized by polymerase chain reaction and proteins by recombinant expression, adequate methods for the synthesis of carbohydrates are still not available. The existing chemical and enzymatic glycosylation techniques are often combined for the synthesis of pure and defined glycans.

10.1 State of the Art

10.1.3.1 Chemical Synthesis Carbohydrates are polyfunctional molecules that carry hydroxyl groups of similar reactivity. Protecting group strategies have to be applied in carbohydrate synthesis to avoid side reactions. Selective protection of hydroxyl groups can be achieved by using orthogonal protecting groups, which can be cleaved independently under different reaction conditions [10]. Another important factor is the control of the stereoselectivity of the glycosylation. In most glycosylation methods anomeric mixtures are obtained. The stereochemistry can be controlled by an accurate choice of neighboring groups, reaction temperature, and solvent [11]. A substituent at C-2, which can participate in the forming of the transition state (e.g., an ester, phenylsulfanyl, or phenylselanyl group) will conduct the synthesis toward the 1,2-trans configured product, whereas with alkyl or benzyl groups at C-2 or 2-deoxy sugars, the 1,2-cis configured product is predominantly formed due to the anomeric effect [12]. Nevertheless, the synthesis of the 1,2-cis configured glycans is a difficult task and anomeric mixtures are often obtained, which have to be separated subsequently by column chromatography. For the glycosylation reaction, a glycosyl donor carrying a leaving group at the anomeric center is first activated to an oxocarbenium ion, which is then attacked by a nucleophile that can be a sugar-bearing a free hydroxyl group or an amino acid–like serine or threonine. There exist several glycosylation methods, which differ in the kind of leaving group and the conditions that are necessary for its activation. Many complex glycans including natural products have been synthesized using different glycosylation methods like the Koenings–Knorr, the trichloracetimidate, or the glycal method among many others. Trichloracetimidate glycosyl donors were used in the syntheses of several natural products such as calicheamicin [13, 14], macroviracin D [15], or complex gangliosides [16, 17]. Biantennary, tribranched, and tetrabranched PSA glycopeptide fragments were synthesized using the glycal assembly technique [18]. Chemical synthesis is often the only way to get pure glycans, which are necessary for use in medical applications. Nevertheless, most syntheses require multiple steps and purification procedures, reaction times up to several months, and the products are obtained only in low yields. Automated Synthesis The investigation of the glycome requires a large library of structurally defined glycans. Compared to genomics and proteomics where nucleic acids and peptides are synthesized with automated synthesizers, adequate tools for the synthesis of glycans are still in their infancy. However, promising techniques for the automated synthesis of glycans have been developed and applied in various complex glycan syntheses. One-Pot Synthesis The one-pot synthesis is based on the different reactivities of carbohydrate building blocks [19]. The reactivity is determined by the substituents that consist of finely tuned electron-pushing and -pulling groups. The glycans are synthesized by successive addition of the building blocks to a single reaction vessel, normally starting from the most reactive to the less reactive one. The

241

242

10 Cracking the Glycocode: Recent Developments in Glycomics BzO

BnO BnO

OBn O

(BzN)O

OBz O

HO 2

O(NBz) O STol O(ClBn)

HO 3

STol NHTroc

BnO BnO

STol OLev

NIS, TfOH, CH2Cl2, −20 °C 67%

1

OBn BzO O O OLev

OBz (BzN)O O O NHTroc

O(NBz) O STol O(ClBn) 4: R = Lev

H2NNH2, HOAc, THF, 0 °C 95% 5: R = H HO

OBn O

O OBn BnO

BzO O

STol OBn

BnO OBn

5

7

NIS, TfOH, CH2Cl2, −40 °C 62%

6

OBn O

OBz OBn BzO (BzN)O O(NBz) O O O BnO O O (BnCl)O TrocHN O OBn O OBn O BnO O 8 BnO BnO OBn OBn BnO

OPMP

OBn rt

OBn O

OPMP

OBn

(a) Zn, AcOH (b) Ac2O, pyridine

8

(c) NaOMe, MeOH (d) H2, Pd /C 45%

HO

OH O

HO

HO

OH O

O

O O HO OH

HO

NHAc

OH O HO O

OH

9 HO

OH O OH

O HO

OH O

OPMP

OH

Figure 10.1 One-pot synthesis of Globo H (Tol, toluyl; Lev, levulinyl; Troc, 2,2,2-trichloroethoxycarbonyl; Bn, benzyl; Bz, benzoyl; PMP, 4-methoxyphenyl; NIS, N-iodosuccinimide).

glycan is thereby elongated from the nonreducing to the reducing end. The one-pot synthesis was automated by implementing a computer program named Optimer, which reverts to a database where the reactivities of more than 400 building blocks are saved. The synthesis of the tumor-associated antigen Globo H, a complex hexasaccharide, by automated one-pot synthesis demonstrates the potential of this method (Figure 10.1) [20]. Similarly, Takashi and coworkers synthesized a library of 54 linear and 18 branched trisaccharides by a one-pot, two-step glycosylation using a commercial synthesizer in 63–99% yields [21]. However, the one-pot synthesis is limited to finding suitable building blocks of adapted reactivity, which are not yet commercially available and often difficult to synthesize. Since the reaction is done in a single reaction vessel no purification of the intermediates is necessary. After addition of all building blocks the final product has to be purified properly because by-products and unreacted educts are not removed during the synthesis. Solid-Phase Synthesis An automator for carbohydrate synthesis was developed by Seeberger and coworkers [22]. It is based on a modified peptide synthesizer. As building blocks, glycosyl phosphates and trichloracetimidates with temporary protecting groups are used. The reducing end of the first building block is coupled to a polystyrene resin by an easily cleavable linker. For the glycosylation, an excess


of the next building block is added followed by washing and filtration to remove by-products and remaining reagents. After selective removal of the temporary protecting group the next building block is added. After several cycles of coupling and deprotection steps the fully protected product is cleaved from the resin, deprotected, and thoroughly purified. Many biologically important oligosaccharides have been prepared by automated solid-phase synthesis. The dimeric combination of the Lewis blood group antigen Ley -Lex , which is found on tumor cells was prepared using only five building blocks [23]. The nonasaccharide was synthesized in 6% yield within 23 hours, which is much faster than the solution-phase synthesis, which requires several months. Although only about 36 building blocks are needed to assemble 75% of known mammalian oligosaccharides, the building blocks are laborious and time-consuming to synthesize and not yet commercially available. The advantage is that complex oligosaccharides can be assembled within one day or less and there is only one final purification step necessary. 10.1.3.2 Enzymatic Synthesis With chemical methods, almost any complex oligosaccharide can be synthesized but the syntheses require an enormous expense. In nature, glycans are synthesized with the help of specific enzymes termed glycosyltransferases. Isolated enzymes or even whole cells can be used for the synthesis of glycans in vitro [24]. Enzymes are highly regio- and stereoselective so that no protecting groups are necessary. For the enzymatic synthesis of glycans, mainly, glycosidases, glycosynthases, and glycosyltransferases have been used. Glycosidases and Glycosynthases In nature there exist two classes of glycosidases termed endoglycosidases and exoglycosidases, which are responsible for the trimming of oligosaccharides. In vitro they can be used for oligosaccharide synthesis if the acceptor is a better nucleophile than water. Although they are seldom regioselective and give lower yields than glycosyltransferases, they are generally more stable and can be used in some organic solvents [25–28]. Furthermore, glycosidases and the required glycosyl donors are well accessible from bacteria and fungi. Glycosynthases are engineered glycosidases, where a catalytic nucleophile is substituted through mutation by a neutral amino acid like glycine, alanine, or serine. The enzymes thereby lose their hydrolytic activity and are able to form glycosidic bonds using mainly glycosyl fluorides as substrates. Glycosyltransferases Glycosyltransferases synthesize glycans in cells, especially in mammalian cells employing nucleotide-activated sugars as glycosyl donors [29]. Glycosyltransferases catalyze the transfer of a glycosyl donor to a hydroxyl group of another sugar or to a serine, threonine, or asparagine residue of a protein with excellent regio- and anomeric stereoselectivity [30]. Usually they are highly substrate specific, but tolerate small modifications of the donor and acceptor [31, 32]. Glycosyltransferases can be further classified into retaining and inverting ones. Retaining glycosyltransferases do not change the stereochemistry at the anomeric position, whereas inverting glycosyltransferases change the anomeric

243

244


stereochemistry. Their in vitro application for glycan synthesis is hampered by limited availability as they exist only in low concentrations in tissue. In addition they are often membrane-bound or -associated, difficult to isolate and often unstable. Recombinantly, they can only be expressed in low yields, which results in a limited large-scale producibility. Though the required nucleotide-activated glycosyl donors are commercially available they are unstable and expensive. Since many glycosyltransferases are strongly inhibited by the released nucleotides, during glycosyltransfer, the nucleotide-activated sugar has to be used in excess. Alternatively, the nucleotide can be degraded or recycled to the nucleotide-activated sugar. A system for the in situ regeneration of nucleotide-activated sugars by which the concentration of the activated sugar and thus, its inhibitory effect on the glycosyltransferase is kept low, was developed by Wong and Whitesides [33]. 10.1.3.3 Glycoprotein Synthesis The structural and functional diversity of proteins is remarkably increased by posttranslational modifications such as methylation, phosphorylation, sulfation, and glycosylation. The glycosylation of proteins is the most complex posttranslational modification and essential for several biological processes like protein folding, cell targeting and adhesion, signal transduction, and many others [34]. Glycoproteins are an emerging class of therapeutics and it has been shown that the glycan moieties are necessary to improve their stability and pharmacokinetic properties [35]. However, glycoproteins exist in nature in numerous glycoforms, which are not appropriate for systematic structure–function studies and therapeutic applications. Thus, different methods have been developed for the preparation of homogeneous glycoproteins. Peptide Ligation Glycopeptides can be synthesized by solid-phase peptide synthesis (SPPS) incorporating glycosylamino acids, by chemoenzymatic manipulation of the glycan moiety of a glycopeptide or by direct glycosylation of a peptide. For SPPS approaches extensive carbohydrate protection is necessary, which is accompanied by the acid and base lability of the glycosylamino acids [36]. SPPS approaches are limited to a maximum peptide size of 50 amino acids. For the synthesis of larger peptides, ligation techniques have to be used where a synthetic glycopeptide is ligated with a peptide or protein based on chemical or biochemical approaches. In the native chemical ligation (NCL) a glycopeptide bearing a C-terminal thioester is reacted with an N-terminal cysteine residue of another peptide whereby a natural peptide bond is formed [37–40]. Several glycopeptides like a HIV viral protein gp 120 fragment [41, 42], complex-type N-linked glycopeptides [43], and a MUC2 motif carrying several O-GalNAc moieties [44] have been synthesized using NCL approaches. The peptide thioester or the cysteine-containing peptide, which are required for the NCL reaction can be generated by recombinant protein expression techniques termed expressed protein ligation (EPL) [38, 45, 46]. Using EPL, Bertozzi et al. synthesized the muzinelike glycoprotein GlyCAM-1, which serves as a ligand for the leukocyte adhesion molecule l-selectin [47]. Peptide fragments can also be ligated chemically by traceless Staudinger ligation where an azide is reacted with


a phosphinothioester resulting in a native peptide bond [48]. Enzymatic ligation is possible by the use of proteases, which are highly stereo- and chemoselective and do not require a cysteine residue at the ligation site [49]. Enzyme Remodeling Homogenous glycoproteins can be obtained by enzymatic remodeling of their glycan moieties. In a first step the glycans are trimmed by glycosidases and in a second step extended by glycosidase- or glycosyltransferase-catalyzed transglycosylation. Hydrolysis as a competing side reaction can be reduced by using organic solvents [50, 51] or oxazoline derivates [49, 52, 53]. Oxazoline donors were used by Wang et al. for the endoglycosidase-catalyzed synthesis of homogenous glycoforms of the protein RNase [54]. Molecular and Cell Biological Techniques Several techniques have been developed to modify the natural glycosylation machinery of a cell. The biosynthetic pathways of cells were augmented by the introduction of new glycosyltransferase genes or regulated by using small molecules that heterodimerize proteins. The natural product rapomyzin, for example, was used for chemically induced dimerization to regulate the production of sLex in living cells [55]. By in vivo suppressor tRNA technology it is possible to introduce unnatural amino acids to the genetic code resulting in a kind of ‘‘pretranslational glycosylation’’ [35]. These unnatural amino acids are either glycosylamino acids [56–58] or carry tags [59] for subsequent site-selective glycosylation. Schultz et al. used this approach to incorporate the glycosylamino acids β-GlcNAc serine and α-GalNAc threonine into myoglobin in vivo [60, 61]. 10.1.4 Glycomics

The glycome is defined as the entity of glycans in a particular cell, tissue, or organism. Glycomics is the structural and functional study of the glycome. The glycome depends on several factors like the genome, which encodes carbohydrate synthesizing, modeling, and transporting enzymes; the transcriptome; and the proteome. In addition, the glycome is also influenced by nutrient levels, which is not the case for the genome and proteome [62]. Thus the glycome is highly complex and dynamic and its analysis requires a combination of several tools as it cannot be predicted by sole analysis of the genome and proteome. 10.1.4.1 Mass Spectrometry Mass spectrometry (MS) is the primary technique for the analysis of small glycan quantities. Information about the glycan structures as well as the glycan attachment sites of proteins is obtained by various experiments. With modern MS methods, glycomes of whole organisms have been profiled and serum samples have been screened to discover disease-related biomarkers [63]. To use MS for high-throughput analysis of, for example, serum samples, the inevitable procedures for glycan sample purification had to be accelerated. Methods have been developed by which carbohydrates are captured onto a solid support by an oxime linkage. The solid

245

246


support is subsequently washed and the carbohydrates eluted by cleavage of the linker [64] or oxime exchange [65]. However, progress in MS glycan analysis is hampered by extensive sample preparation before analysis. Glycoconjugates have to be separated depending on their attached aglycones (N-linked, O-linked, and glycolipids) and modifications like O-acetyl, methyl, and sulfate groups are often removed during sample preparation [63]. 10.1.4.2 Microarrays Microarrays are ideally suited for the high-throughput analysis of glycans. Lectins and antibodies are highly specific for certain glycans and can be used for glycan profiling. For this purpose, lectins and antibodies are immobilized on the array surface and afterward, fluorescence labeled glycoproteins from cell lysate or fluid samples are added. This technique is excellently suited for a fast comparison of glycomes, but it reveals only limited structural details. Since 2002, carbohydrate microarrays are used where isolated or synthesized glycans are noncovalently or covalently immobilized on the array surface. The glycans have a defined orientation on the surface, which is similar to their presentation on cell surfaces. The immobilized glycans can be probed with MS [66, 67], fluorescence labeled antibodies or lectins [68], or cell imaging [69]. Glycan arrays have been used to profile carbohydrate specificities of lectins, growth factors, cytokines, antibodies, microbial toxins, viruses, and whole cells, to determine the specificity of human, avian, and mutated influenza hemagglutinin, to develop vaccines against malaria, and to find cellular markers of infectious diseases [70]. Furthermore, the activity and selectivity of various glycosyltransferases [70, 71] as well as new substrate specificities of bacterial glycosyltransferases [72] and inhibitors [70] have been identified. Thus, glycan microarrays are emerging tools for biomarker and drug discovery. 10.1.4.3 Cell, Tissue, and Metabolic Labeling Molecular imaging techniques have been developed to analyze changes in the glycome during pathogenic or normal biological processes. The glycans in cells in tissues can be imaged by using fluorescent-labeled lectins and antibodies. Lectins have been used in cultured cell lines [73, 74] and tissue sections [75–80], but their in vivo applicability is limited because they possess only low affinity to carbohydrates and are furthermore tissue-impermeable and often toxic. Antibodies have been used for fixed cells and one in vivo application has also been reported [81]. Nevertheless, their use for in vivo applications is limited by their tissue impermeability and low affinity. As well as lectins, they can only be used to study a present state of the glycome, whereas dynamic studies are difficult to conduct. A novel approach for a dynamic study of the glycome in vivo encompasses feeding cells with modified glycans, which the cell subsequently incorporates in the biosynthesis of cell surface glycans. The modified glycans carry a chemical reporter group, which can be labeled by click-chemistry with imaging probes and then visualized by fluorescence microscopy. Most glycan subtypes have been

10.2 Some New Insights in Glycomics

imaged by metabolic labeling using azido or alkynyl derivates of sialic acid, N-acetylgalactosamine, N-acetylglucosamine, mannose, and fucose [82]. Glycans could be imaged in many cultured cell lines, and also in living organisms such as zebrafish [83], where changes in the glycome during embryonical development were mapped by metabolic labeling. 10.1.4.4 Bioinformatics A major challenge for bioinformatics is to correlate the data obtained from the various experiments since they reveal information of different kinds and hierarchical levels. Some techniques, for instance, provide information about the glycan structures, whereas others reveal which glycans are presented on specific cells or tissues. Bioinformatic tools from genomics and proteomics research are seldom adaptable for glycomics since the glycome is not directly encoded by the genome [84]. The correlation of data from glycomics with genomic, transcriptomic, and proteomic data is essential to see, for example, how the expression of enzymes involved in the synthesis or transport of glycans determines the glycome of a certain cell or tissue. In recent years, comprehensive glycan databases have been created and international consortia have been formed to collaborate in a systematic study of the glycome.

10.2 Some New Insights in Glycomics 10.2.1 Microwave-Assisted Glycosylation for the Synthesis of Glycopeptides

Malignant transformations of cells are often associated with the presentation of altered glycoproteins on their surfaces [85]. Synthetic vaccines containing the displayed carbohydrate antigens like Ley or Globo H conjugated to an immunogenic carrier protein like keyhole limpet hemocyanin have been shown to elicit antitumor cell antibody responses in mice and patients [86]. The carbohydrate antigens appear only in small amounts on the surface of tumor cells, and thus cannot be isolated, but rather have to be synthesized. For the solid-phase synthesis of glycopeptides Fmoc-protected glycosylamino acids are generally required. So, there is an urgent need for easy and efficient methods for the synthesis of glycosylamino acids. Carvalho et al. reported about the glycosylation of Fmoc-Ser-OBn with 2-aceta mido-2-deoxy-3,4,6-tri-O-acetyl-α-d-glucopyranosylchloride, which is activated by mercuric bromide [87]. The glycosylation of Fmoc-Ser-OH with β-glucose was reported to proceed under BF3 ·OEt2 promotion in yields from 30 to 37% and reaction times from 2 to 18 hours [88–91]. The described glycosylations require multistep reactions for the synthesis of the glycosyl donors and heavy metals like Hg(CN)2 , HgBr2 , or AgOTf as promoters. Long reaction times are necessary and the products are obtained in low yields.

247

248


OH FmocHN 10 O OAc RO AcO

O AcO

FeCl3, 4 Å MS toluene OAc

OBn OAc RO AcO

Microwave

O

O

AcO

OBn

FmocHN O Figure 10.2 FeCl3 -promoted synthesis of glycosylamino acids. Reprinted with permission from [93].

It has been reported that FeCl3 promotes the β-glycosylation of aliphatic alcohols with β-peracetylated glycosides [92]. So, we carried out the reaction with β-peracetylated carbohydrates and Fmoc-Ser-OBn 10 under reflux conditions in toluene using FeCl3 as a promoter. Although the β-linked glycosylamino acids were formed, we observed many by-products. The optimal temperature for a minimum of site products and a maximum yield was 45 ◦ C. TLC studies revealed that not the activation but rather the addition of the acceptor was problematic. Neither other promoters nor an excess of the glycosyl donor or an increase of the temperature could improve the yield. This showed that steric reasons obviously hindered the addition. It was reported that steric hindrance could be overcome through microwave radiation. So, we carried out the same reactions under microwave radiation by which the reaction time could be reduced from several hours to just 4 minutes (Figure 10.2). It was not possible to activate the α-anomers of the glycosyl donors. Using a mixture of peracetylated α- and β-glucose 13, only the β-anomer reacted and the α-anomer could be reisolated. We were able to glycosylate Fmoc-Ser-OBn with various peracetylated mono- and disaccharides (Table 10.1). For the synthesis of the β-glycoside 18 from peracetylated lactose 17, a literature method requires three steps with an overall yield of 9%, whereas our method requires only one step with a reaction time of 8 minutes and a yield of 54% [94]. The glycosylation is assumed to proceed under participation of the neighboring group at C-2 to yield selectively the β-glycosides, as reported by Lemieux [95]. The oxocarbenium ion reacts with the more stable acyloxonium ion, which is formed by the attack of the acyl group at C-2. Solvents with low polarity like toluene probably promote this mechanism by inefficient solvation of the oxocarbenium ion [96]. The formation of orthoesters, which is often described in literature was not observed. We also performed the reaction with peracetylated N-acetylglucosamine 19. We were able to isolate the respective oxazoline 20, which can be reacted with various glycosyl acceptors [97]. Nevertheless, it was not possible to react the oxazoline directly with Fmoc-Ser-OBn under FeCl3 promotion and microwave conditions. In conclusion, we could demonstrate that through microwave radiation the reaction time for the synthesis of various glycosylamino acids could be reduced

13

AcO AcO

11

AcO

AcO

Donor

O

OAc

AcO

O

OAc OAc

OAc

14

AcO AcO

12

AcO

AcO

Product

O

O

FmocHN

AcO

O

OAc

FmocHN

AcO

O

OAc

O

O

OBn

OBn

FeCl3 promoted glycosylation of Fmoc-Ser-OBn 10.

AcO

Table 10.1

4

2×4

4

4

Time (min)

Microwave

0 : 0 (0)

0 : 1 (61)

0 : 0 (0)

0 : 1 (52)

α/ß (yield %)

300

720

300

720

Time (min)

0 : 0 (0)

0 : 1 (22)

0 : 0 (0)

0 : 1 (31)

α/ß (yield %)

(continued overleaf )

Conventional

10.2 Some New Insights in Glycomics 249

AcO

AcO AcO

17

AcO

O

OAc

O AcO

19

OAc NHAc

O

OAc

AcO

O

OAc

AcO

O

OAc

AcO

O

OAc

(continued)

AcO O AcO 15

AcO AcO

Donor

Table 10.1

OAc

OAc

AcO AcO

18

AcO

AcO

AcO O AcO

O

OAc

20

N

O O

OAc

O AcO AcO

O

OAc

16

AcO AcO

Product

O

O

AcO FmocHN

O

OAc

FmocHN

AcO

O

OAc

O

O

OBn

OBn

4

2×4

4

2×4

4

4

Microwave

0 : 0 (0)

0 : 1 (85)

0 : 0 (0)

0 : 1 (54)

0 : 0 (0)

0 : 1 (52)

300

–

300

720

300

720

0 : 0 (0)

–

0 : 0 (0)

0 : 1 (16)

0 : 0 (0)

0 : 1 (10)

Conventional

250



from 5–10 hours to 4–8 minutes. This method is thereby providing easy access for building blocks needed for solid-phase glycopeptide synthesis. Metals such as silver and mercury, which are necessary in most glycosylations for the activation of the glycosyl donor, could be substituted by environmentally friendly FeCl3 . The ß-anomeric selectivity of the glycosyl donor allows separation and discrimination of donor substrates and could be used for multicomponent and multistep reactions. 10.2.2 Highly Efficient Chemoenzymatic Synthesis of Novel Branched Thiooligosaccharides by Substrate Direction with Glucansucrases

Carbohydrates exist as glycolipids or proteins on the cell surfaces. As they influence signaling processes like viral or bacterial infections, cancer metastasis, and inflammatory responses [98], large amounts of defined structure are needed to explore the details of these processes. Glycans exist in nature in extremely wide diversity, whereby the isolation of homogenous forms from natural sources is hardly possible. Even for a single protein, there generally exist several glycoforms, which differ in their glycosylation pattern and sites. Chemical synthesis is often the only way to obtain pure and defined structures, but it also requires complex protecting group strategies resulting in low yields and thorough purification steps. In nature, carbohydrates are synthesized by more than 200 glycosyltransferases, which exhibit a high regio- and stereoselectivity, but are often quite substrate specific. Using glycosyltransferases for the in vitro synthesis of carbohydrates is restricted by expensive substrates and limited availability of the membrane-bound glycosyltransferases. One class of carbohydrate synthesizing enzymes are bacterial glucansucrases also termed non-Leloir glycosyltransferases, which are produced by lactic acid bacteria such as Streptococcus, Lactobacillus, and Leuconstoc spp [99, 100]. Although they have high structural homology with glycosidases, they catalyze the synthesis of glycans, whereas glycosidases catalyze their hydrolysis. It is not clear until now, which structural details are responsible for these contrary functions. Bacterial glucansucrases synthesize glucose polymers of high molecular weight using sucrose as a substrate [24, 101]. They have a highly conserved catalytic domain, which is responsible for the binding of sucrose and a C-terminal domain that is responsible for glucan binding [99]. Depending on the respective glucansucrase, differently linked polymers like dextran (α-1,6), mutan (α-1,3), reuteran (α-1,4), and alternan (α-1,3 and α-1,6) are formed [102]. Besides polymers, oligosaccharides can also be synthesized if small molecules are used as acceptors [103–105]. Nevertheless, the product spectrum is limited, as the products are mostly α-1,6-linked and only glucose can be transferred by using sucrose as a substrate. It was demonstrated that the regioselectivity of enzymatic condensations could be altered by modifying the acceptor substrate as well as the enzyme [106, 107]. The regioselectivity of a glycosidase [108] as well as a Leloir glycosyltransferase [109] was changed by blocking the hydroxyl groups of the acceptor. We were

251


252

able to direct the selectivity of glucansucrases by substrate engineering. We used the glucansucrase R (GTFR) from Streptococcus oralis, which forms an α-1,6 glucose polymer with minor α-1,3 branches from sucrose as a substrate. Previous experiments showed that the GTFR can glucosylate various alcohols and amino acids indicating the diversity of this enzyme. The acceptor reaction with maltose as a substrate yielded panose as a product, in which glucose was transferred from α,1-6 to maltose. Using lactose as substrate, glucose was transferred α-1,2. The 6 -hydroxyl group was obviously not sufficiently accessible for the enzyme due to steric hindrance caused by the axial 4 -hydroxyl group of lactose. Thus the chemoselectivity of the enzyme could be controlled by the choice of the acceptor. In the acceptor reaction of GTFR with glucose 21 as a substrate, isomaltose 22, a disaccharide with α-1,6 linkage is formed as a product (Figure 10.3). We were able to change the selectivity of GTFR by modifying the substrate glucose at C-6. Glucose was blocked by a tosyl group at C-6 yielding the new acceptor

OH

OTs

O

HO HO

OR

OTs HO HO

OR

25: R = H 27: R = Me (a) 29: R = allyl (a)

21: R = H 23: R = Me (a)

GTFR or GTFA, sucrose

27: R = Me (a) 29: R = allyl (a)

GTFA, sucrose

GTFR, sucrose

OH

OH HO HO

O

O

HO HO

OH

O

HO HO

OTs

HO HO O

HO O

O

OR HO 6-Ts-glucose

HO 6-Ts-glucose

HO Glucose

HO HO

O

HO HO

O

OR

HO

O HO O

OTs O HO

OR

OR

HO

HO 22: R = H 24: R = Me (a) a-1,6

26: R = H 28: R = Me (a) 30: R = allyl (a) a -1,3

Figure 10.3 Acceptor-substrate-directed synthesis by GTFR and GTFA glycosyltransferases. Reprinted with permission from [112].

31: R = Me (a) 32: R = allyl (a) a-1,4


substrate 6-tosylglucose 25, which was reacted with GTFR and sucrose. The acceptor product 26 was glucosylated at C-3 in a yield of 48%, which is in correlation with the branching activity of GTFR. The yield could be increased up to 95% by using tosylated methyl-27 and allylglucose 29. The structural diversity was further expanded by using the glucansucrase A (GTFA) from Lactobacillus reuteri, which forms the α-1,4-linked glucose polymer reuteran [110]. Nevertheless, the acceptor products of GTFA are mainly α-1,6-linked [111]. We used the 6-tosylglucose derivates 27 and 29 as acceptor substrates for the GTFA and the products 31 and 32 were α-1,4-linked. So, it was possible to direct the main activity of the glucansucrases GTFR and GTFA to a side activity by substrate engineering. The tosyl as well as the allyl group provide access to further elaboration to build up branched glycoconjugates. Driguez et al. reported the synthesis of β-1,6-thio-linked oligosaccharides by anomeric S-alkylation through substitution of 6-halogenides of S-acetyl protected thiosugars [113]. We applied commercially available 1-thiosugars to minimize the necessary synthetic steps. The α-1,3-linked acceptor product 26 from the reaction with GTFR was first acetylated and then substituted with peracetylated thiosugars to yield trisaccharides in yields of around 30% (Table 10.2). The synthesis was further optimized so that no protecting groups were needed. Investigations showed that no substitution took place using monosaccharides, which is due to their lower nucleophilicity. So, the unprotected tosyl sugars could be substituted in DMF (dimethylformamide) at 80 ◦ C by thiosugars, which were previously deprotected and deprotonated with sodium methanolate. The yields were in the range of 90%. With this chemoenzymatic strategy, various branched thiooligosaccharides could be synthesized in a two-step synthesis (Table 10.2). Thiosugars, which carry a sulfur instead of an oxygen atom in their glycosidic bond, possess unique physiochemical properties differing from those of conventional O-glycosides. They have an enhanced water solubility and are tolerated by most biological systems. Besides, the S-glycosidic bond causes an enhanced stability against enzymatic or acid/base-catalyzed hydrolysis [114]. These characteristics qualify them as promising compounds for therapeutic applications. It has been reported that certain thio-linked disaccharides serve as fucosidase inhibitors [114, 115]. Furthermore it was shown, that thio-linked disaccharides decreased tumor cell line viability and hence, are potential antitumor agents [116]. Among the synthesized glycans, is the trisaccharide 45 with a terminal α-2,6-linked sialic acid. Sialic acids are involved in important biological processes like viral infection, immune cell activation, cancer metastasis, and neuronal development among many others [117]. An enhanced expression of α-2,6-linked sialic acids on tumor cells is often associated with poor prognosis for therapy [118]. The synthesized thiooligosaccharides, which are allyl-functionalized at the reducing end provide the opportunity for conjugation to proteins to yield neoglycoproteins or for immobilization to solid supports like microarrays, to investigate their potential as vaccines and their affinity to proteins of biological relevance.

253

AcO AcO

AcO AcO

36

O

OAc

O AcO

O AcO

OAc

AcO

36

O

OAc

OAc

O

OAc

O

OTs

OAc

O

OTs

OAc

OAc 37

34

AcO

AcO

AcO AcO

34

AcO

AcO

SH

SH OAc

O

OAc

OAc

O

OAc

SH OAc

O

OAc

Acceptor

Chemical synthesis of branched thiooligosaccharides.

OTs

AcO

33

AcO AcO

Donor

Table 10.2

AcO

O

OAc

AcO AcO

AcO AcO

39

38

O

AcO

O

OAc

O

AcO

O

OAc

35 AcO AcO

AcO

AcO

Product

O

AcO

AcO

AcO AcO

AcO

S

AcO

AcO

O

OAc

AcO

AcO

O

OAc

OAc

O

O AcO

S

AcO

S

OAc

OAc

34

31

75

Yield (%)

254


O

29

HO HO

29

HO HO O HO

O HO

HO

O

OH

HO

O

OH

HO 33 OCH3

HO HO

OTs

OCH3

HO

O OCH3

OTs

HO

O

OTs

OH

O

OH

SH

SH

SH

OH

O

OH

40

O

OH

40

33

HO HO

HO

HO

HO

HO

OH

HO HO

HO HO

41

HO

HO

43

S

O

O

HO

O

OH

42

HO

O

OH

HO HO

HO

O

OH

HO HO

HO

HO

HO

HO

O

S

HO

HO

O

OH

OH

OCH3

O HO

HO

O HO

87

91

(continued overleaf )

OCH3

OCH3

O

S

90

10.2 Some New Insights in Glycomics 255

HO HO

29

HO HO

Donor

(continued)

31

O HO

HO

O

OH

O HO

HO

O

OH

Table 10.2

OCH3

HO

O O

OTs

HO

O

OTs OH

33

OH

O

OH

O

SH

44

AcHN HO HO

HO HO

HO

Acceptor

SH

CO2H

HO HO

HO HO

O

O

HO

O

OH

HO O

46

45

OH O

HO

HO

HO

AcHN HO HO OH

HO

Product

S

O

O HO

OCH3

HO

HO

O

OH

HO

O

S

CO2H

90

47

Yield (%)

256



257

10.2.3 Identification of New Acceptor Specificities of Glycosyltransferase R with the Aid of Substrate Microarrays

Leloir glycosyltransferases are excellent tools for the regio- and stereoselective synthesis of glycans. Nevertheless, the transferases are not always accessible and as well as the required nucleotide-activated glycosyl donors, expensive and not fitted for large-scale synthesis. Non-Leloir glycosyltransferases, on the other hand, use sucrose as a substrate and it was shown that their regioselectivity can be controlled by substrate engineering [112]. Carbohydrate microarrays are excellent tools to analyze the structure and functions of glycans [70, 119, 120]. In most cases, isolated or readily synthesized glycans are first immobilized on the array and proteins added in a second step. There are some reports where glycoarrays are used to study the activity of glycosyltransferases [70, 71]. Our goal was to OH OH O HO HO

OH O

HO HO

HO O O HO

HO 47

HO HO

OH OH

OH

HO

O

O

HO OH HO

OH O HO

O

GTFR

O HO

OH

HO HO

+

OH

48

HO HO

O

HO

O

O OH HO

O

O HO

O

HO

HO O

OH N N N

N N N O

O

H3C N

H3C N O

O H3C N

H3C N

A

B

Figure 10.4 Enzymatic glycosylation of different acceptors on arrays. The immobilized alcohol on slide A and immobilized maltose were successfully glycosylated by the bacterial glycosyltransferase R and could afterward be detected by the binding of fluorescence labeled Concanavalin A. Reprinted with permission from [72].

A

B


identify new glycosylation pathways of bacterial glucansucrases on an array by variation of enzyme, acceptor, and substrate. As acceptors, azido-functionalized maltose and lactose were first coupled to a linker by 1,3-dipolar cycloaddition, and afterward, a primary aliphatic alcohol was also immobilized on a microtiter plate by NHS/EDC (N-hydroxy succinimide/1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) coupling in order to find out which of these substrates is glycosylated by the bacterial glycosyltransferase GTFR (Figure 10.4). Hence, a solution of glucosyltransferase GTFR and sucrose 47 was added to each microtiter well. After blocking the array with bovine serum albumin, fluorescence labeled Concanavalin A (ConA), which is a lectin specific for mannose and glucose but not for galactose, was added to detect the glycosylation of the immobilized acceptors. In literature, maltose was described to be a good acceptor for glycosyltransferases of the glucansucrase type such as dextransucrase (DSRS) [121], whereas for alcohols, no glycosylation was observed. After scanning the microtiter plate with a fluorescence scanner, we observed a stronger signal for maltose with added enzyme than for maltose without added enzyme (Figure 10.5). Goldstein et al. reported about a twofold increase of the ConA binding affinity for α-(1,6)-glucosylated maltose relative to mannose [122]. So, we interpreted the increased signal for enzymatically treated maltose as its successful glycosylation. The signal for enzymatically treated lactose was lower than for the microtiter wells without enzyme indicating that no glycosylation took place. The decreased signal is presumably due to less nonspecific binding of ConA. With the DSRS B-512F from L. mesenteroides, lactose is glucosylated at C-2[123]; so the immobilized lactose was probably not glucosylated because of steric hindrance resulting from the attached linker at the reducing end (Figure 10.5). The strongest binding was observed for 30 000 Aliphatic alcohol, sigmoidal Maltose, sigmoidal Lactose, sigmoidal Empty, sigmoidal Maltose Lactose Aliphatic alcohol Empty

25 000 Relative fluorescence

258

20 000 15 000 10 000 5 000 0 0

2

4

6

8

Time [h]

Figure 10.5 The arrays were monitored with the glucose-specific lectin ConA, with novel acceptor products derived from sucrose being identified with the glycosyltransferase, GTFR. Reprinted with permission from [72].

10

12

10.3 Future Perspectives

the immobilized alcohol, which shows its successful glycosylation. In summary, we could use the array platform as an excellent tool for the identification of new enzyme specificities as well as new acceptor products for the GTFR. After approval of the concept, the chip format should in future be reduced to small carbohydrate microarrays that offer the possibility to screen several hundreds acceptors in parallel fashion, which is not possible with conventional methods. This makes them important tools to enhance the scope of enzymes, which can be used to synthesize various complex glycans in solution and on-chip.

10.3 Future Perspectives

In 2003, the MIT’s magazine Technology Review had identified glycomics as one of the top 10 technologies that will change the future. During the past years, many efforts have been devoted to establish tools for the synthesis and investigation of glycans. Promising methods are now available for the synthesis of complex glycans. Nevertheless, each method has its strengths and weaknesses, which are often complementary to each other. Therefore, combined chemical and enzymatic approaches have to be applied for the synthesis of diverse glycan libraries. Although techniques to analyze the glycome are lagging behind those in genomics and proteomics, important advancements have been achieved. Metabolically labeled glycans have already been visualized in vivo and should in future be extended to mammalian disease models and human clinical settings [82]. Presenting the whole glycome of a certain cell on a single microarray would be a great advance in functional glycomics, which is until now only limited by the difficulties in the isolation and synthesis of glycans. Modification or even synthesis of the glycans on the chip by using engineered enzymes and substrates will enhance the diversity of existing glycan libraries. Microarrays have already proven to be excellent tools to study interactions between carbohydrates and proteins, viruses, and whole cells; sera of patients have also been screened [70]. High-throughput array analysis will in future lead to novel therapeutics and the discovery of highly specific disease biomarkers. Synthetic carbohydrate-based vaccines are already in clinical trials and are hopeful candidates against infectious diseases and cancer. New tools, algorithms, and data collections for glycobiology have to be developed and cross-linked with existing genomic and proteomic data collections [84]. So, most likely, many future insights in life sciences will be by achieved with the aid of glycomics.

Acknowledgments

J.S. and L.H. wish to express their gratitude to all past and present group members of J.S. who contributed to the described research. L.H. would also like to thank the German Environmental Foundation (DBU) for a PhD scholarship.

259

260


References 1. Varki, A. (2009) Essentials of Glycobio-

2. 3. 4. 5.

6. 7. 8. 9. 10. 11.

12. 13.

14.

15.

16.

17.

18.

19.

20.

logy, 2nd edn, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Capila, I. and Linhardt, R.J. (2002) Angew. Chem. Int. Ed., 41, 390. Tabeur, C. et al. (1999) Bioorg. Med. Chem., 7, 2003. Petitou, M., Casu, B., and Lindahl, U. (2003) Biochimie, 85, 83. Bauer, K.A., Hawkins, D.W., Peters, P.C., Petitou, M., Herbert, J.-M., Boeckel, C.A.A., and Meuleman, D.G. (2002) Cardiovasc. Drug Rev., 20, 37. Macdougall, L.C. (2000) Semin. Nephrol., 20, 375. Seeberger, P.H. and Werz, D.B. (2007) Nature, 446, 1046. Livingston, P.O. and Ragupathi, G. (2006) Hum. Vaccines, 2, 137. Galonić, D.P. and Gin, D.Y. (2007) Nature, 446, 1000. Wong, C.H., Ye, X.S., and Zhang, Z. (1998) J. Am. Chem. Soc., 120, 7137. Levy, D.E., Fügedi, P., and Levy, F. (2006) The Organic Chemistry of Sugars, Taylor & Francis, Boca Raton. Nicolaou, K.C. and Mitchell, H.J. (2001) Angew. Chem., 113, 1624. Halcomb, R.L., Boyer, S.H., and Danishefsky, S.J. (1992) Angew. Chem. Int. Ed. Engl., 31, 338. Nicolaou, K.C., Schreiner, E.P., Iwabuchi, Y., and Suzuki, T. (1992) Angew. Chem. Int. Ed. Engl., 31, 340. Mlynarski, J., Ruiz-Caro, J., and Fürstner, A. (2004) Chemistry, 10, 2214. Sugimoto, M., Fujikura, K., Nunomura, S., Horisaki, T., Ito, Y., and Ogawa, T. (1990) Tetrahedron Lett., 31, 385. Sugimoto, M., Fujikura, K., Nunomura, S., Ito, Y., and Ogawa, T. (1990) Tetrahedron Lett., 31, 1435. Dudkin, V.Y., Miller, J.S., and Danishefsky, S.J. (2004) J. Am. Chem. Soc., 126, 736. Zhang, Z., Ollmann, I.R., Ye, X.S., Wischnat, R., Baasov, T., and Wong, C.H. (1999) J. Am. Chem. Soc., 121, 734. Burkhart, F., Zhang, Z., Wacowich-Sgarbi, S., and Wong, C.H.

21.

22.

23.

24.

25. 26. 27.

28.

29.

30. 31.

32.

33. 34. 35.

36. 37. 38.

(2001) Angew. Chem. Int. Ed. Engl., 40, 1274. Takahashi, T., Adachi, M., Matsuda, A., and Doi, T. (2000) Tetrahedron Lett., 41, 2599. Plante, O.J., Palmacci, E.R., and Seeberger, P.H. (2001) Science, 291, 1523. Routenberg Love, K. and Seeberger, P.H. (2004) Angew. Chem. Int. Ed. Engl., 43, 602. Seibel, J., Jördening, H.-J., and Buchholz, K. (2006) Biocatal. Biotransform., 24, 311. Chinnathambi, S. and Lachke, A. (1997) Biotechnol. Tech., 11, 589. Laroute, V. and Willemot, R.M. (1992) Enzyme Microb. Technol., 14, 528. Vulfson, E.N., Patel, R., Beecher, J.E., Andrews, A.T., and Law, B.A. (1990) Enzyme Microb. Technol., 12, 950. Beecher, J.E., Andrews, A.T., and Vulfson, E.N. (1990) Enzyme Microb. Technol., 12, 955. Murrell, M.P., Yarema, K.J., and Levchenko, A. (2004) Chembiochem, 5, 1334. Roseman, S. (2001) J. Biol. Chem., 276, 41527. Qian, X., Sujino, K., and Palcic, M.M. (2000) Enzymatic Glycosylations with Non-natural Donors and Acceptors Wiley-VCH Verlag GmbH, vol. 2. Sujino, K., Uchiyama, T., Hindsgaul, O., Seto, N.O.L., Wakarchuk, W.W., and Palcic, M.M. (2000) J. Am. Chem. Soc., 122, 1261. Wong, C.H. and Whitesides, G.M. (1982) J. Org. Chem., 47, 2816. Helenius, A. and Aebi, M. (2001) Science, 291, 2364. Gamblin, D.P., Scanlan, E.M., and Davis, B.G. (2009) Chem. Rev., 109, 131. Kunz, H. (1987) Angew. Chem. Int. Ed. Engl., 26, 294. Dawson, P.E. and Kent, S.B.H. (2000) Annu. Rev. Biochem., 69, 923. Muir, T.W. (2003) Annu. Rev. Biochem., 72, 249.

References 39. Nilsson, B.L., Soellner, M.B., and

40. 41.

42.

43.

44.

45.

46.

47. 48.

49.

50.

51.

52.

53.

54. 55. 56.

Raines, R.T. (2005) Annu. Rev. Biophys. Biomol. Struct., 34, 91. Derek, M. (2006) Angew. Chem. Int. Ed., 45, 7668. Mandal, M., Dudkin, V.Y., Geng, X., and Danishefsky, S.J. (2004) Angew. Chem. Int. Ed., 43, 2557. Geng, X., Dudkin, V.Y., Mandal, M., and Danishefsky, S.J. (2004) Angew. Chem. Int. Ed., 43, 2562. Mezzato, S., Schaffrath, M., and Unverzagt, C. (2005) Angew. Chem. Int. Ed., 44, 1650. Hojo, H., Matsumoto, Y., Nakahara, Y., Ito, E., Suzuki, Y., Suzuki, M., Suzuki, A., and Nakahara, Y. (2005) J. Am. Chem. Soc., 127, 13720. David, R., Richter, M.P.O., and Beck-Sickinger, A.G. (2004) Eur. J. Biochem., 271, 663. Muir, T.W., Sondhi, D., and Cole, P.A. (1998) Proc. Natl. Acad. Sci. U.S.A., 95, 6705. Macmillan, D. and Bertozzi, C.R. (2004) Angew. Chem. Int. Ed., 43, 1355. Saxon, E., Armstrong, J.I., and Bertozzi, C.R. (2000) Org. Lett., 2, 2141. Li, B., Zeng, Y., Hauser, S., Song, H., and Wang, L.X. (2005) J. Am. Chem. Soc., 127, 9692. Fan, J.Q., Takegawa, K., Iwahara, S., Kondo, A., Kato, I., Abeygunawardana, C., and Lee, Y.C. (1995) J. Biol. Chem., 270, 17723. Akaike, E., Tsutsumida, M., Osumi, K., Fujita, M., Yamanoi, T., Yamamoto, K., and Fujita, K. (2004) Carbohydr. Res., 339, 719. Li, H., Li, B., Song, H., Breydo, L., Baskakov, I.V., and Wang, L.X. (2005) J. Org. Chem., 70, 9990. Rising, T.W., Claridge, T.D., Moir, J.W., and Fairbanks, A.J. (2006) Chembiochem, 7, 1177. Li, B., Song, H., Hauser, S., and Wang, L.-X. (2006) Org. Lett., 8, 3081. Kohler, J.J. and Bertozzi, C.R. (2003) Chem. Biol., 10, 1303. Fahmi, N.E., Dedkova, L., Wang, B., Golovine, S., and Hecht, S.M. (2007) J. Am. Chem. Soc., 129, 3586.

57. R¨ ohrig, C.H., Retz, O.A., Hareng, L.,

58.

59.

60.

61.

62. 63. 64.

65.

66.

67.

68.

69.

70.

Hartung, T., and Schmidt, R.R. (2005) Chembiochem, 6, 1805. Zhang, Z., Gildersleeve, J., Yang, Y.Y., Xu, R., Loo, J.A., Uryu, S., Wong, C.H., and Schultz, P.G. (2004) Science, 303, 371. Liu, H., Wang, L., Brock, A., Wong, C.-H., and Schultz, P.G. (2003) J. Am. Chem. Soc., 125, 1702. Xu, R., Hanson, S.R., Zhang, Z., Yang, Y.-Y., Schultz, P.G., and Wong, C.-H. (2004) J. Am. Chem. Soc., 126, 15654. Zhang, Z., Gildersleeve, J., Yang, Y.-Y., Xu, R., Loo, J.A., Uryu, S., Wong, C.-H., and Schultz, P.G. (2004) Science, 303, 371. Freeze, H.H. (2006) Nat. Rev. Genet., 7, 537. Pilobello, K.T. and Mahal, L.K. (2007) Curr. Opin. Chem. Biol., 11, 300. Lohse, A., Martins, R., Jorgensen, M., and Hindsgaul, O. (2006) Angew. Chem. Int. Ed. Engl., 45, 4167. Shimaoka, H., Kuramoto, H., Furukawa, J., Miura, Y., Kurogochi, M., Kita, Y., Hinou, H., Shinohara, Y., and Nishimura, S. (2007) Chem. Eur. J., 13, 1664. Fazio, F., Bryan, M.C., Blixt, O., Paulson, J.C., and Wong, C.H. (2002) J. Am. Chem. Soc., 124, 14397. Laurent, N., Voglmeir, J., Wright, A., Blackburn, J., Pham, N.T., Wong, S.C., Gaskell, S.J., and Flitsch, S.L. (2008) Chembiochem, 9, 883. Blixt, O., Head, S., Mondala, T., Scanlan, C., Huflejt, M.E., Alvarez, R., Bryan, M.C., Fazio, F., Calarese, D., Stevens, J., Razi, N., Stevens, D.J., Skehel, J.J., van Die, I., Burton, D.R., Wilson, I.A., Cummings, R., Bovin, N., Wong, C.H., and Paulson, J.C. (2004) Proc. Natl. Acad. Sci. U.S.A., 101, 17033. Nimrichter, L., Gargir, A., Gortler, M., Altstock, R.T., Shtevi, A., Weisshaus, O., Fire, E., Dotan, N., and Schnaar, R.L. (2004) Glycobiology, 14, 197. Liang, P.H., Wu, C.Y., Greenberg, W.A., and Wong, C.H. (2008) Curr. Opin. Chem. Biol., 12, 86.

261

262

10 Cracking the Glycocode: Recent Developments in Glycomics 71. Laurent, N., Voglmeir, J., and Flitsch,

72.

73. 74.

75. 76. 77.

78.

79. 80.

81.

82. 83.

84.

85. 86.

87.

88.

S.L. (2008) Chem. Commun., (37), 4400. Seibel, J., Hellmuth, H., Hofer, B., Kicinska, A.M., and Schmalbruch, B. (2006) Chembiochem, 7, 310. Carlsson, S., Carlsson, M.C., and Leffler, H. (2007) Glycobiology, 17, 906. Lizzi, A.R., D’Alessandro, A.M., Bozzi, A., Cinque, B., Oratore, A., and D’Andrea, G. (2007) Mol. Cell. Biochem., 300, 29. Griffith, C.M. and Sanders, E.J. (1991) Anat. Rec., 231, 238. Griffith, C.M. and Wiley, M.J. (1989) Anatomy Embryol., 180, 567. Paessens, L.C., Garc´ıa-Vallejo, J.J., Fernandes, R.J., and van Kooyk, Y. (2007) Immunol. Lett., 110, 65. Thurston, G., Baluk, P., Hirata, A., and McDonald, D.M. (1996) Am. J. Physiol.: Heart Circ. Physiol., 271, Tian, E. and Ten Hagen, K.G. (2007) Glycobiology, 17, 820. Truong, L.D., Phung, V.T., Yoshikawa, Y., and Mattioli, C.A. (1988) Histochemistry, 90, 51. Licha, K., Debus, N., Emig-Vollmer, S., Hofmann, B., Hasbach, M., Stibenz, D., Sydow, S., Schirner, M., Ebert, B., Petzelt, D., Bührer, C., Semmler, W., and Tauber, R. (2005) J. Biomed. Opt., 10, 041205. Laughlin, S.T. and Bertozzi, C.R. (2009) Proc. Natl. Acad. Sci. U.S.A., 106, 12. Laughlin, S.T., Baskin, J.M., Amacher, S.L., and Bertozzi, C.R. (2008) Science, 320, 664. von der Lieth, C.W., Bohne-Lang, A., Lohmann, K.K., and Frank, M. (2004) Brief. Bioinform., 5, 164. Hakomori, S.I. (1985) Cancer Res., 45, 2405. Ouerfelli, O., Warren, J.D., Wilson, R.M., and Danishefsky, S.J. (2005) Expert Rev. Vaccines, 4, 677. Carvalho, I., Scheuerl, S.L., Kartha, K.P.R., and Field, R.A. (2003) Carbohydr. Res., 338, 1039. Fahmi, N.E., Golovine, S., Wang, B., and Hecht, S.M. (2001) Carbohydr. Res., 330, 149.

89. Salvador, L.A., Elofsson, M., and

90.

91.

92.

93. 94. 95. 96. 97. 98. 99.

100.

101.

102.

103. 104. 105.

106.

Kihlberg, J. (1995) Tetrahedron, 51, 5643. Sol, V., Blais, J.C., Carré, V., Granet, R., Guilloton, M., Spiro, M., and Krausz, P. (1999) J. Org. Chem., 64, 4431. Steffan, W., Schutkowski, M., and Fischer, G. (1996) Chem. Commun., 313. Katsuraya, K., Ikushima, N., Takahashi, N., Shoji, T., Nakashima, H., Yamamoto, N., Yoshida, T., and Uryu, T. (1994) Carbohydr. Res., 260, 51. Seibel, J., Hillringhaus, L., and Moraru, R. (2005) Carbohydr. Res., 340, 507. Rajca, A. and Wiessler, M. (1995) Carbohydr. Res., 274, 123. Lemieux, R.U. (1954) Adv. Carbohydr. Chem., 9, 1. Wulff, G. and Rohle, G. (1974) Angew. Chem., 86, 173. Wittmann, V. and Lennartz, D. (2002) Eur. J. Org. Chem., (8), 1363. Varki, A. (1993) Glycobiology, 3, 97. Monchois, V., Willemot, R.M., and Monsan, P. (1999) FEMS Microbiol. Rev., 23, 131. Monsan, P., Bozonnet, S., Albenne, C., Joucla, G., Willemot, R.M., and Remaud-Simeon, M. (2001) Int. Dairy J., 11, 675. van Hijum, S.A., Kralj, S., Ozimek, L.K., Dijkhuizen, L., and van Geel-Schutten, I.G. (2006) Microbiol. Mol. Biol. Rev., 70, 157. Magali Remaud-Simeon, C.A., Joucia, G., Fabre, E., Bozonnet, S., Pizzut, S., Escalier, P., Potocki-Veronese, G., and Monsan, P. (2003)in Glucansucrases: Structural Basis, Mechanistic Aspects, and New Perspectives for Engineering, vol. 849 (eds G.Eggleston. and G.L. Cote), American Chemical Society, p. 90. Robyt, J.F. and Eklund, S.H. (1982) Bioorg. Chem., 11, 115. Robyt, J.F. and Walseth, T.F. (1978) Carbohydr. Res., 61, 433. Dols, M., Simeon, M.R., Willemot, R.M., Vignon, M.R., and Monsan, P.F. (1997) Carbohydr. Res., 305, 549. Montero, E., Alonso, J., Canada, F.J., Fernandez-Mayoralas, A., and

References

107.

108. 109.

110.

111.

112.

113.

Martin-Lomas, M. (1997) Carbohydr. Res., 305, 383. Nilsson, K.G.I. and FernándezMayoralas, A. (1991) Biotechnol. Lett., 13, 715. MacManus, D.A. and Vulfson, E.N. (2000) Biotechnol. Bioeng., 69, 585. Lairson, L.L., Watts, A.G., Wakarchuk, W.W., and Withers, S.G. (2006) Nat. Chem. Biol., 2, 724. Kralj, S., van Geel-Schutten, G.H., van der Maarel, M.J., and Dijkhuizen, L. (2004) Microbiology, 150, 2099. Kralj, S., Stripling, E., Sanders, P., van Geel-Schutten, G.H., and Dijkhuizen, L. (2005) Appl. Environ. Microbiol., 71, 3942. Hellmuth, H., Hillringhaus, L., Höbbel, S., Kralj, S., Dijkhuizen, L., and Seibel, J. (2007) Chembiochem, 8, 273. Greffe, L., Jensen, M.T., Chang-Pi-Hin, F., Fruchard, S., O’Donohue, M.J., Svensson, B., and Driguez, H. (2002) Chemistry, 8, 5447.

114. Witczak, Z.J., Chhabra, R., Chen, H.,

115.

116. 117. 118.

119. 120. 121.

122. 123.

and Xie, X.-Q. (1997) Carbohydr. Res., 301, 167. Witczak, Z.J., Sun, J., and Mielguj, R. (1995) Bioorg. Med. Chem. Lett., 5, 2169. Witczak, Z.J., Kaplon, P., and Dey, P.M. (2003) Carbohydr. Res., 338, 11. Varki, N.M. and Varki, A. (2007) Lab. Invest., 87, 851. Miles, D.W., Happerfield, L.C., Smith, P., Gillibrand, R., Bobrow, L.G., Gregory, W.M., and Rubens, R.D. (1994) Br. J. Cancer, 70, 1272. Paulson, J.C., Blixt, O., and Collins, B.E. (2006) Nat. Chem. Biol., 2, 238. Park, S., Lee, M.R., and Shin, I. (2008) Chem. Commun., 4389. Buchholz, K. and Monsan, P. (2003) Handbook of Food Enzymology, p. Marcel Dekker, 589. Goldstein, I.J., Hollerman, C.E., and Smith, E.E. (1965) Biochemistry, 4, 876. Robyt, J.F. and Eklund, S.H. (1983) Carbohydr. Res., 121, 279.

263

265

Part IV Enzyme Chemistry


267

11 O2 Reactivity at Model Copper Systems: Mimicking Tyrosinase Activity Anna Company

11.1 General Introduction: O2 Activation and Model Systems

The use of O2 as the oxidant in chemical transformations has become of great industrial interest, not only because it constitutes an environmentally friendly alternative to the highly toxic oxidants used nowadays in several chemical transformations but also due to its relatively low cost [1]. However, directly using molecular oxygen as an oxidant is not straightforward. The dioxygen molecule has a high potential reactivity that is kept in control by its particular electronic structure: molecular oxygen is a diradical having two unpaired electrons [2]. The two oxygen atoms share six electrons in the σ2pz , π2px , and π2py molecular orbitals and two unpaired electrons reside in the two degenerate antibonding π2px ∗ and π2py ∗ orbitals, leaving O2 with a formal bond order of two. The triplet ground state of O2 (S = 1) direct reaction with singlet molecules (S = 0) – the spin-paired state of most organic molecules representative of biological substrates – a spin-forbidden process [3, 4]. In order to overcome the high kinetic barrier inherent to the reactions of triplet O2 , in nature dioxygen is reduced from its abundant triplet ground state to a reactive singlet or doublet state (O2 − , O2 2− ) by oxidase and oxygenase enzymes. In many cases, these enzymes generate species of even greater reactivity by cleaving the O–O bond. Generally, nature’s strategy involves a transition metal (mainly Cu and Fe but also Mn, Ni, or Co), which in the appropriate oxidation state is able to react directly with triplet O2 to form dioxygen adducts that can participate in reaction pathways leading either to the selective incorporation of oxygen atoms into organic substrates or to their oxidation. Chemists have long sought to unravel the mechanisms of biological oxygen activation [5–9], with the ultimate objectives of inventing new reagents for organic synthesis and industrial catalysis, exerting control over reactivity and selectivity, and getting fundamental information about enzymatic reactions. A key issue concerns the nature of the oxidizing species that is directly responsible for effecting the reaction. In particular, the characterization of metal-based intermediates that act as the active species is critical to obtain detailed mechanistic insight. However, due to their nature as reactive oxidizing molecules, isolation or even direct observation Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

268

11 O2 Reactivity at Model Copper Systems: Mimicking Tyrosinase Activity

is, in general, difficult. Model chemistry is a useful tool to achieve this goal [10, 11]. It gains inspiration from biological systems to develop low-molecular-weight complexes that reproduce structural characteristics of an enzyme and hopefully its activity. These complexes are useful to understand how nature works, and in selected cases they constitute excellent catalysts that can elicit unprecedented and highly specific chemical transformations analogous to enzymes. In model chemistry, the accomplishment of these objectives often begins with the preparation of a suitable ligand (organic synthesis) and continues through several steps toward the ultimate goal of reproducing catalytic reactivity. In this review chapter, we present how the model chemistry approach has allowed to reproduce and to better understand the mechanisms by which O2 is activated in a dinuclear copper protein called tyrosinase is discussed.

11.2 Copper Proteins Involved in O2 Activation

Copper is found in most living organisms, and despite being present only in trace amounts, it is essential for life [12]. Generally, it is considered that biological copper serves exclusively as 1e− shuttle, alternating between CuI and CuII . The CuIII state is generally considered to be inaccessible because of the highly positive CuIII /CuII redox potentials [13]. However, several reaction mechanisms postulate the existence of intermediates in which the copper center reaches such an oxidation state that has so far not been directly detected in biological systems. The versatility of copper is clearly reflected by the wide range of biological reactions that it catalyzes. Nevertheless, in most cases, the role of this metal ion is directly related to two basic functions: electron transfer and O2 activation and transport [3, 14]. While the electron transfer activity is performed by the so-called ‘‘blue copper proteins,’’ transport and activation of O2 is carried out by a wide range of enzymes. Indeed, in biological systems, O2 -activating copper proteins emerge as a structurally diverse family of enzymes. Their active center may be constituted by one copper center (mononuclear copper proteins: amine oxidase and galactose oxidase) [15], by two metal sites (dinuclear copper proteins: dopamine β-hydroxylase, hemocyanin, tyrosinase, and catechol oxidase) [15, 16], or even by multinuclear configurations in which three copper centers are involved (trinuclear sites: laccase, ascorbate oxidase, ceruloplasmin, and FET 3) [16]. This variety of structural motifs allows the performance of a wide range of oxidative transformations including O2 transport, substrate hydroxylation, amine oxidation, or hydrogen atom abstraction. All these reactions take place through the activation of O2 by copper sites leading either to incorporation of oxygen atoms into an organic substrate (oxygenase activity) or to its oxidation accompanied by O2 reduction usually to water (oxidase activity) [9, 17, 18]. In the context of this chapter, those synthetic models capable of reproducing the activity exhibited by tyrosinase (the ortho-hydroxylation of phenols) are reviewed. From a structural point of view, this enzyme possesses a dinuclear coupled center also referred as type 3 copper site. However, before attempting to


HN

HN N

H N

N

HN N

4.6 Å

CuI

CuI

N

N NH

N

NH

NH

+

2

−

2

1.4 Å

H N

N N N

N H

HN

CuII

CuII

N 3.6 Å NH

269

N N N H

Oxy form

Deoxy form

rRaman: n(O – O) = 750 cm−1 UV– vis: l = 345 nm (20 000 M−1 cm−1), 570 nm (1000 M−1 cm−1) EPR silent Scheme 11.1 Structural changes of the dicopper site of hemocyanin upon oxygenation (• denotes oxygen atom from molecular O2 ).

reproduce the structure or reactivity of such an enzyme, the basic properties of type 3 copper proteins must be understood. In Sections 11.2.1 and 11.2.2, the main features of the two prototypical proteins (hemocyanin and tyrosinase) within this group are described. 11.2.1 Hemocyanin

Hemocyanins are large, multisubunit proteins that act as oxygen carriers (reversible O2 binding) in many molluscs and arthropods, analogous to hemoglobin in vertebrates [19]. Their active site is constituted by two copper atoms, each of them coordinated to three histidine residues. In the reduced (deoxy) form, the copper(I) ions show a trigonally distorted coordination geometry and they are separated by 4.6 A˚ (Scheme 11.1) [20]. However, in the oxygenated (oxy) form, the copper centers are oxidized by one electron to the +2 oxidation state and an oxygen molecule is reduced by two electrons to form a peroxide unit (O2− 2 ), which ends up ligated between the two metal sites in a µ − η2 : η2 coordination mode. Each of the five-coordinated CuII centers possesses a distorted square-pyramidal geometry; the Cu · · · Cu separation of 3.6 A˚ is significantly contracted relative to that in the reduced form (Scheme 11.1). The oxy form of hemocyanin has very distinctive spectroscopical signatures [21]. Its resonance Raman (rRaman) spectrum displays the characteristic O–O stretching frequency at ν = 750 cm−1 and UV–vis spectroscopy shows bands at λmax = 345 nm (ε ∼ 20 000 M−1 cm−1 ) and 570 nm (ε ∼ 1000 M−1 cm−1 ) responsible for the intense blue color of the oxidized form of the protein. Finally, no EPR (electron paramagnetic resonance) signal is detected due to strong antiferromagnetic coupling between the two CuII centers (S = 1/2). Several structural studies [22, 23] indicate that in hemocyanin there is a protein domain that shields the access of substrates to the dicopper active center. Because

270


OH

OH + O2

R Phenol Scheme 11.2

O OH

Tyrosinase (Phenolase)

O

Tyrosinase

Melanins

(Catecholase) R Catechol

R

ortho -Quinone

Biological function of tyrosinase toward the final synthesis of melanins.

this domain interferes with substrate binding, hemocyanin can play only one role, namely, that of an oxygen transporter by the reversible coordination of dioxygen to the active site. 11.2.2 Tyrosinase

Tyrosinase is an essential copper-containing enzyme that is found in all organisms. It is involved in browning processes of skin, hair, and fruit, and in wound healing or the immune response [16]. Tyrosinase initiates the synthesis of brown melanin pigments by catalyzing the ortho-hydroxylation of phenols to catechols (ortho-diphenols) and subsequent two-electron oxidation to ortho-quinones using molecular O2 as the oxidant (Scheme 11.2). The latter reaction is also catalyzed by the related enzyme catechol oxidase, which, however, is unable to mediate the phenol hydroxylation step. The crystal structure of catechol oxidase is available since 1998 [24], and despite the fact that the overall folding of this enzyme is completely different from the one found in hemocyanin, both proteins present an almost identical type 3 copper site. Remarkably, the X-ray structure of tyrosinase was not solved until very recently [25], but the configuration of its active site had been inferred by the similar spectroscopic features with hemocyanin [26]. Thus, for ages it has been well established that tyrosinase possesses an active site almost identical to the one found in hemocyanin and catechol oxidase both in the reduced and in the oxidized form (Scheme 11.1). This structural assignment was further corroborated by the X-ray studies by Matoba et al. on bacterial tyrosinase, a work that also allowed the characterization of the most relevant intermediates of the catalytic cycle of this enzyme [25]. It was found that the active site of tyrosinase and hemocyanin can be superposed showing a high degree of structural similarity [20, 25]. The different reactivity of both enzymes arises from the particular folding of the polypeptidic chain: while the accessibility to the active site in hemocyanin is blocked, the active center of tyrosinase is situated close to the molecular surface, a useful location to ensure the access of the substrate. Moreover, analysis of the crystal structure of tyrosinase indicates that the substrate-binding pocket has a large vacant space above the dicopper center, which allows direct interaction between the substrate (phenol) and the active site [25]. Remarkably, Itoh and coworkers demonstrated that hemocyanin can exhibit efficient monooxygenase activity (phenolase activity analogous to tyrosinase) if the


HN

(His) N

= (His)

N

(His) N CuI

CuI

N (His)

(His) N

2

N (His)

N (His)

Deoxy form (A)

N (His) CuII CuII N (His) (His) N N (His) Oxy form (B)

(His) N

O H2

OH

+ H+

R

H+

R

R

R

O (His) N

O II

Cu (His) N N H (His) (D)

CuII N (His)

N (His) N (His)

N (His) CuII N (His) N (His)

(His) N

CuII (His) N N (His) (C)

Scheme 11.3 Catalytic cycle proposed for the hydroxylation of a monophenolic substrate by tyrosinase (• denotes oxygen atom from molecular O2 ).

polypeptidic chain is altered by treatment with urea. This observation corroborates the identical active center in both proteins and the critical role that plays the folding of the protein in determining the ultimate function of a specific enzyme [27]. The proposed mechanism for phenol hydroxylation performed by tyrosinase is depicted in Scheme 11.3 [28]. In the first step of the catalytic cycle, the reduced form of tyrosinase (deoxy form A) interacts with molecular O2 to give the oxygenated intermediate (oxy form B) in which a peroxo ligand is coordinated in a µ − η2 : η2 mode to the two CuII centers. At this point, the substrate approaches the active center and it is preoriented by hydrophobic and hydrogen bonding interactions with nearby protein residues. The substrate binds to one of the metal centers (C) and hydroxylation of the aromatic ring occurs. To this end, the O–O axis of the peroxo ligand rotates and it points toward the phenolic ring. The proximity of the ortho position of the phenolic ring to the side-on coordinated peroxo group enables an electrophilic attack of the Cu2 O2 moiety on the aromatic ring, by which concomitant cleavage of the O–O bond occurs, so that the resulting diphenolic unit binds in a bidentate fashion (D). Finally, release of the ortho-quinone product regenerates the deoxy form (A) and a new catalytic cycle occurs. Overall, the active center of tyrosinase accommodates important changes during the catalytic cycle, which are possible only with a highly flexible dicopper site [25].

271

272


11.3 O2 Binding and Activation at Biomimetic Cu Complexes

The biological importance of the reactions involved in copper-regulated O2 metabolism and the potential technological utility of the oxidation reactions catalyzed by copper-dependent enzymes have fueled great effort in understanding the reaction mechanisms regulating O2 activation by these metalloproteins. A particular strategy to achieve these goals is the study of the reactivity exhibited by bioinspired synthetic models with O2 [9, 29, 30]. The study of mono- and dinuclear copper complexes with simple ligands reproducing the coordination environment of the metal center in natural proteins has allowed the characterization of a wide range of novel chemical species resulting from the interaction of O2 with copper(I). This Cun O2 type of compounds (generally formed by the oxygenation of CuI complexes) tends to be highly unstable and, in fact, the advance in stabilizing and characterizing them can be attributed partly to a greater accessibility of appropriate spectroscopic tools and to a better appreciation of the appropriate reaction conditions. Low temperatures (∼200 K), aprotic solvents (CH2 Cl2 , THF, acetone), and weak coordination anions are now standard conditions. Owing to their high instability, Cun O2 compounds are called the intermediate species formed upon CuI oxidation to CuII or CuIII accompanied by total or partial reduction of O2 . The identification and characterization of these synthetic intermediates are crucial to postulate possible reaction mechanisms in natural enzymes such as tyrosinase [31]. 11.3.1 Copper–Dioxygen Adducts

In the last couple of decades, several CuI complexes have been synthesized and their reactivity toward O2 has been thoroughly studied, which has allowed the determination of different coordination modes in Cun O2 type of systems. The nuclearity (n, number of copper centers activating a single O2 molecule) can range from 1 to 3 depending on the stoichiometry of the reaction [9, 29]. In tyrosinase there is a cooperative interaction between two copper sites to activate O2 (Scheme 11.3), which makes Cu2 O2 specially relevant for the purpose of mimicking such an enzyme. The most outstanding configurations exhibiting a Cu2 O2 core are trans-µ-1,2-peroxodicopper(II) (T P), µ-η2 : η2 -peroxodicopper(II) (s P), and bis(µ-oxo)dicopper(III) (O) (Figure 11.1). The biological relevance of trans-µ-1,2-peroxodicopper(II) (T P) intermediates has not been proved yet, but they are synthetically accessible through the use of mononuclear copper(I) complexes with tetradentate ligands that self-assemble upon O2 reaction [32, 37–41]. Their structural assignment is based primarily on distinctive intense UV–vis features that are responsible for their purple colors: λmax ∼ 530 nm (ε ∼ 10 000M−1 cm−1 ), and 600 nm (ε ∼ 7000M−1 cm−1 ). rRaman spectroscopy is also an adequate tool to identify this type of species that are characterized by an O–O stretching vibration at ν ∼ 830cm−1 ([18 O2 ] ∼ –45 cm−1 ) and a Cu–O stretching vibration at ν ∼ 555 cm−1 ([18 O2 ] ∼ –24 cm−1 ) [9].


H B N

N

N N

N

−

D5

N tpa

N B

Cu

O

O

CuIIL

trans -µ-1,2-Peroxo (TP) (a)

LCuII

Cu

O

CuIIL

Figure 11.1 (a) X-ray structure of the T P species [CuII 2 (O2 )(tpa)2 ]2+ formed upon reaction of [CuI (tpa)(EtCN)]+ with O2 in EtCN at −90 ◦ C [32]. (b) X-ray structure of the S P species [CuII 2 (O2 )(TpiPr2 )2 ] formed upon reaction of [CuI (TpiPr2 )]

O

CuIIIL O Bis(µ-oxo) (O)

LCuIII

O µ-h2:h2-Peroxo (sP) (b)

N D D

N O

O

Cu

N

LCuII

N

D D D21-Bz3tacn

TpiP r2

O

D5

D D N

N N

N N

273

(c) with O2 in acetone at −80 ◦ C [33, 34]. (c) X-ray structure of the O species [CuIII 2 (O)2 (D21 -Bz3 tacn)2 ]2+ formed upon reaction of [CuI (D21 -Bz3 tacn)(CH3 CN)]+ with O2 in CH2 Cl2 at −80 ◦ C [35, 36].

[CuII 2 (O2 )(tpa)2 ]2+ (Figure 11.1a) constitutes the prototypical example of these compounds and it is synthesized by reversible interaction of two [CuI (tpa) (EtCN)]+ units with a single O2 molecule at cryogenic temperatures [32, 37]. µ-η2 : η2 -Peroxodicopper(II) complexes (S P) were first discovered by Kitajima and coworkers with the resolution of the X-ray crystal structure of [CuII 2 (O2 )(TpiPr2 )2 ] [33, 34] (Figure 11.1b). Interestingly, the similarity between the spectroscopic features of this compound and those of the oxy form of hemocyanin (Scheme 11.3) strongly supported the notion that O2 binding in the oxidized form of this protein occurred with the same mode, as was later demonstrated by the resolution of the crystal structure of the natural protein [20]. This was a great success of synthetic model chemistry because through the characterization of a simple coordination complex, the correct and unprecedented coordination mode of O2 in a natural system could be established. S P species have very characteristic spectroscopic features. On the one hand, their UV–vis spectra exhibit a high energy charge transfer band at λmax = 340–380 nm (ε ∼ 18 000–25 000 M−1 cm−1 ) and a weaker lower energy band at 510–550 nm (ε ∼ 1000M−1 cm−1 ) conferring their characteristic violet color. On the other hand, their rRaman spectra display a characteristic low-energy O–O stretching vibration at ν = 730–760 cm−1 ([18 O2 ] ∼ –40 cm−1 ) [9].

D5

274


The first CuIII (µ-O)2 -CuIII core (O) was reported by Tolman in [35] and it presents completely different spectroscopic features compared to the rest of Cu2 O2 intermediates. The first structurally characterized O species was generated with the macrocyclic fac triamine ligand Bz3 tacn [35, 36], and despite its instability above −80 ◦ C, it could be fully characterized by several techniques including X-ray analysis of its perdeutero analog (Figure 11.1c). Since Tolman’s discovery, several O type of intermediates have been spectroscopically and, in selected cases, structurally characterized [42–45]. The structural differences between O and S P isomers portend different UV–vis and rRaman spectroscopic characteristics. O complexes generally exhibit two intense ligand-to-metal charge-transfer absorptions at λmax ∼ 300 nm (ε ∼ 20 000 M−1 cm−1 ) and 400 nm (ε ∼ 24 000 M−1 cm−1 ), whilst rRaman experiments reveal a characteristic and intense vibration at ν ∼ 600 cm−1 ([18 O2 ] ∼ –25 cm−1 ) [9]. Despite the fact that O species have never been directly detected in biological systems, it cannot be excluded that they may be involved in biological processes through the O–O bond breakage of the S P core. Indeed, one of the major contributions from Tolman and coworkers was the demonstration of the existence of a rapid equilibrium between S P and O species, thus indicating reversible scission and formation of the dioxygen O–O bond (Scheme 11.4) [46]. This equilibrium was clearly evidenced by the observation of reversible dramatic spectroscopic changes (UV–vis and rRaman) in a particular S P species upon changing the solvent. The new spectroscopic features could be unambiguously assigned to the new O compound; thanks to the resolution of its X-ray structure. Tolman’s work constitutes the first precedent of a transition metal complex where the O–O bond is formed and broken reversibly, and because of its significance, this equilibrium has been studied in depth [47–49]. It is influenced primarily by the steric demands of the ligand [47, 50, 51], although solvent [47, 51–53] or electronic features of the ligand [48, 54] can also affect its position.

LCuIII

O

CuIIIL

LCuII

O

CuIIL

O

O

Bis(µ-oxo) (O)

µ-h2:h2-peroxo (sP)

rRaman: n ~ 600 cm−1 (∆[18O2] ~ −25 cm−1)

rRaman: n ~ 750 cm−1 (∆[18O2] ~ −40 cm−1)

UV– vis: lmax ~ 300 nm (e ~ 20 000 M−1 cm−1) 400 nm (e ~ 24 000 M−1 cm−1)

UV– vis: lmax ~ 360 nm (e ~ 22 000 M−1 cm−1) 530 nm (e ~ 1000 M−1 cm−1)

Scheme 11.4 Equilibrium between O and S P complexes along with their spectroscopic features.


Mononuclear models

Dinuclear models

Cu+

Cu+

+

Cu+

Cu+

2

2

Cun +1

−2n 2

Cun +1

Cun +1

−2n 2

Cun +1

Scheme 11.5 Schematic representation of the O2 activation by mononuclear and dinuclear synthetic models to form Cu2 O2 species (• denotes oxygen atom from molecular O2 ).

11.3.2 Ligand Architecture: Influence on Reactivity toward O2

Our understanding on how oxygen can be activated by CuI centers has been advanced enormously thanks to the development of suitable ligand scaffolds. In general, biological models of dinuclear copper proteins that activate oxygen consist in mononuclear CuI complexes that self-assemble upon reaction with molecular O2 (Scheme 11.5, top). However, as described in Section 11.2.2, the active center of tyrosinase is constituted by a dinuclear copper site in which the two metallic ions are disposed in a particular configuration favoring a cooperative interaction with molecular O2 (Scheme 11.3) [25]. For this reason, an important step toward the mimicking of the natural active site of tyrosinase, and in general of dicopper proteins, is the synthesis of dinucleating ligands capable of establishing some control over the spatial arrangement of the two metal centers in the synthetic system [55]. These ligands are constituted by two coordination sites that hold two copper centers close enough to give rise to intramolecular interaction with O2 (Scheme 11.5, bottom). Traditionally, the connection between the two binding units of a dinucleating ligand is done with an alkyl group [56, 57] or an aromatic ring. A particular case of the latter is constituted by meta-xylyl linked dinuclear CuI complexes [58, 59] (dinucleating ligands, Scheme 11.6). Apart from the ligand nuclearity, another important point to design a good mimic of dinuclear copper proteins is the coordination environment: in order to reproduce the histidine-rich environment present in this type of proteins, the copper center should be surrounded by nitrogen atoms. In this sense, the use of ligands containing secondary or tertiary amines, pyridine groups, or even benzimidazole units has succeeded in the development of synthetic models [9, 29] suitable to reproduce the activity of dinuclear copper proteins in general, and tyrosinase in particular. The most relevant works in this field are discussed in Section 11.3.3. Another decisive factor that determines the final outcome of the interaction between copper(I) centers and molecular oxygen is the ligand flexibility because change in the oxidation state from +1 to +2 or even to +3 entails important modifications in the disposition of the coordination atoms around the copper center. This phenomenon is even more important in those cases where the copper–dioxygen adduct oxidizes a specific substrate. In these situations, direct

275

276


Mononucleating ligands

N

N N

NH N N

N N

N

N

D D

N

N

NH

N

N

MeAN L6Ph

N

N

N

N

N

N

N

N

N N

N

N

LAA

R R-PhPyNEt2 R = H, NO2, OMe

LPy2

DBED

LGG

LAG

R

Dinucleating ligands R N N

N

N N N

N

N

N

N

N N

N N

N

N

N

N N

N

N

N

N

N

N

N N

N

m -XYLMeAN

Me3m

R-XYL R = H,t Bu, OMe, F, CN, NO2

L66 (R = H) MeL66 (R = Me)

Scheme 11.6 Ligands used for the preparation of biomimetic copper complexes of tyrosinase.

coordination of exogenous substrates to the metal site is generally required causing even greater modifications in the environment around the Cun O2 core. The effect of ligand topology on oxygen activation by low-molecular-weight copper complexes has been studied in depth by different research groups [55]. Indeed, it has been evidenced that the structure of the ligand is important not only to achieve the desired O2 activation but also to perform tyrosinase-like reactivity (hydroxylation of aromatic rings). Among these studies, a recent work reported by our group clearly demonstrates the decisive role played by the flexibility of the ligand as far as oxygen activation by dicopper centers is concerned [60]. Two different dinucleating nitrogen-based ligands were used: Me3m and m-XYLMeAN (Scheme 11.6). In Me3m, the two binding triamine sites are connected by two meta-xylyl linkers giving a hexaaza macrocycle, while m-XYLMeAN is constituted by two triamine units connected by a single meta-xylyl group. The corresponding dicopper(I) complexes present important similarities (Scheme 11.7): in both cases, the copper sites are coordinated to three aliphatic nitrogen atoms and they possess a distorted trigonal planar geometry. The coordination sphere of the metal centers is structurally related to the one found in tyrosinase, in which each copper ion is coordinated to three histidine residues in the reduced form [25]. Finally, [CuI (MeAN)]+ (Schemes 11.6 and 11.7) constitutes a mononuclear analog of the above-mentioned dinuclear complexes.


N

N I

N Cu N

Cu

I

N N

I

N

N

N I

N

Cu Cu N

N N CuI N

2

N [CuI(MeAN)]+

I

MeAN 2+

[Cu 2(m -XYL [CuI2(Me3m)]2+ 2

)]

Acetone −80 °C

2

N

Slow decomposition to CuII species

N N

CuIII N

N CuIII

N

N

N CuII N

CH2Cl2 −80 °C

N CuII N N

[CuII2(µ-h2:h2-O2)(MeAN)2]2+

[CuIII2(µ-O)2(m -XYLMeAN)]2+ Scheme 11.7 Reactivity toward O2 of three structurally related dinuclear and mononuclear CuI complexes (• denotes oxygen atom from molecular O2 ) [60].

Its X-ray structure [61] shows a copper(I) center with a distorted trigonal planar geometry structurally equivalent to the one found in its dinuclear counterparts. Despite the almost identical structure of the copper(I) centers in the three complexes, their reactivity toward O2 was found to be completely dependent on the specific ligand. [CuI 2 (Me3m)]2+ reacted slowly with O2 even at room temperature and no metal–oxygen adduct was accumulated or detected in the course of the reaction. In contrast, reaction of [CuI 2 (m-XYLMeAN )]2+ in acetone at−80 ◦ C was much faster and a bis(µ-oxo) core (O) was trapped and spectroscopically characterized by UV–vis (λmax = 308 nm, ε = 20 000 M−1 cm−1 ; λmax = 413 nm, ε = 28 000 M−1 cm−1 ) and rRaman spectroscopy (ν = 600 cm−1 , [18 O2 ] = −23 cm−1 ). Finally, for the mononuclear analog [CuI (MeAN)]+ , a S P species, formed upon reaction of two complex units with a single O2 molecule, was exclusively detected [61]. Given the comparable coordination sphere and electronic properties of the CuI ions in the three complexes, it is rather unlikely that the initial O2 binding to a single CuI depends on the particular complex, and therefore the different O2 reactivity highlights the important role played by the second metal ion. In [CuI 2 (m-XYLMeAN )]2+ , the ligand is flexible, allowing copper sites to approach close enough to promote their synergistic actuation in O2 binding reduction, which is also true for its mononuclear analog [CuI (MeAN)]2+ . Instead, the rather rigid nature of the macrocyclic ligand (Me3m) imposes a higher barrier to this process,

277

278


shutting down the reaction. Thus, this rather simple model highlights the importance of cooperative effects between two copper centers to activate O2 . Further support for this concept comes from the recent observation that the active site of the natural enzyme tyrosinase accommodates remarkable changes occurring along its catalytic cycle, which require the presence of a rather flexible active center [25]. 11.3.3 Hydroxylation of Aromatic Rings: Mimicking Tyrosinase Activity

The reactivity of several Cu2 O2 species has been studied and analyzed in detail during the last two decades. The biological relevance of these species and the potential technological interest of the chemical transformations they perform have fueled great effort toward the understanding of basic aspects of their chemistry, such as their electrophilic/nucleophilic character or their acid/base behavior [29]. However, the main objective on synthetic models is reproducing the reactivity performed by enzymes. In particular, in this section, the focus is on those synthetic systems that reproduce the reactivity exhibited by tyrosinase: the hydroxylation of aromatic rings. These model systems contain a Cu2 O2 type of species (analogous to the dinuclear copper site present in the natural system) and they may be divided in two categories: (i) compounds that perform an intramolecular ligand aromatic hydroxylation and (ii) complexes that carry out the ortho-hydroxylation of exogenous phenolic compounds (natural substrates). In the following sections, the most outstanding examples of both classes have been reviewed. 11.3.3.1 Intramolecular Aromatic Hydroxylation Generally, copper–dioxygen species are stable only at low temperatures (−80 ◦ C) and warming up usually causes decomposition, typically to CuII complexes, via processes involving intramolecular oxidation of a supporting ligand. These processes can be considered as destructive and thus they may be undesired from the viewpoint of catalyst development, but they provide fundamental mechanistic information relevant to the reactivity of Cun O2 . As far as Cu2 O2 species is concerned, N-dealkylation [42, 62] or aliphatic hydroxylation [63] has been observed to occur especially upon formation of O and S P isomers. Another intramolecular oxidation occurring upon decomposition of selected Cu2 O2 adducts is the hydroxylation of an aromatic moiety included in the ligand structure. This reaction has attracted much attention of the scientific community, not only for the inherent interest of this transformation itself but also due to the reminiscence of this process to the activity exhibited by tyrosinase. Probably, the most deeply studied systems exhibiting this type of reactivity were described by Karlin and coworkers using dinucleating ligands of the R-XYL family (Scheme 11.6) [58, 59, 64]. Reaction of complexes [CuI 2 (R-XYL)]2+ with oxygen at room temperature resulted in the formation of [CuII 2 (R-XYL-O)(OH)]2+ , with the two oxygen atoms incorporated into the oxidized product coming from O2 (Scheme 11.8). Monitoring the reaction by UV–vis spectroscopy at low temperatures afforded the detection of a transient copper–dioxygen adduct when electron-withdrawing R groups (such

11.3 O2 Binding and Activation at Biomimetic Cu Complexes R

R

279 R

R +

N N

CuI N

N CuI N

N CuII

2

N

[CuI2(R-XYL)]2+

py

N CuII py

py

N CuII py

[CuII2(O2)(R-XYL)]2+

py

py

H

py

N CuII

N CuII

N CuII py

py

py

H

py

py

[CuII2(R-XYL-O)(OH)]2+

(sP)

Scheme 11.8 Mechanism of intramolecular aromatic ligand hydroxylation performed by [CuI 2 (R-XYL)]2+ system (• denotes oxygen atom from molecular O2 ).

as NO2 ) were used, which was unambiguously assigned as a S P species on the basis of its spectroscopic features (UV–vis: λmax = 358 nm, ε = 20 000 M−1 cm−1 ; λmax = 435 nm, ε = 5000 M−1 cm−1 ; λmax = 530 nm, ε = 1200 M−1 cm−1 . rRaman: ν = 747 cm−1 , [18 O2 ] = −40 cm−1 ); further support for this assignation comes from density functional theory (DFT) calculations. An electrophilic attack of the S P species over the m-xylyl ring followed by an intramolecular proton transfer was the proposed mechanism to explain the intramolecular ligand aromatic hydroxylation reaction, and it was supported by several experimental observations: (i) the lack of hydrogen/deuterium kinetic isotope effect indicated that the rate-determining step was the initial attack of the S P unit on the aromatic ring rather than a hydrogen atom abstraction event and (ii) the decrease in the hydroxylation rate upon increasing the electron-withdrawing character of the R group (measured Hammett parameter ρ = –2.1). The reactivity of several other dinuclear copper(I) systems with O2 also causes the intramolecular hydroxylation of the aromatic moiety of the ligand. However, in most cases, little is known about the exact mechanism of this process as no Cun O2 adducts are experimentally detected [65–70]. By analogy to Karlin’s system and on the basis of theoretical analyses [71], it is generally proposed that a S P type of species is responsible for the observed reactivity. Interestingly, a recent theoretical study shows that the mechanism of intramolecular aromatic hydroxylation by a dinuclear copper(I) complex upon reaction with O2 favors the participation of a S P species. However, the small energetic differences with the corresponding O isomer do not allow to definitively rule out the involvement of the bis(µ-oxo) core in this hydroxylation process [72]. It is interesting to notice here that there is a reported example in which a Cu2 O2 species different from the S P structure can perform the intramolecular aromatic hydroxylation [73]. In this work, it was observed that decomposition of the O compound supported by R-PhPyNEt2 ligands (R = NO2 ; UV–vis: λmax = 404 nm, ε = 13 000 M – 1 cm−1 ; rRaman: ν = 606 cm−1 , [18 O2 ] = –27 cm−1 ) resulted in partial ligand phenyl ring hydroxylation (Scheme 11.9). The decomposition of the Cu2 O2 species clearly involved the reaction with the aryl group because the nature of the aromatic substituents significantly influenced the decay rate. Similar to

280


R

Et2 N CuIII N

R

N CuIII N Et2

1. Warm 2. Aq. NH4OH (- Cu ions)

R

N H

NEt2

+

R

N NEt2

[CuIII2(µ-O)2(R-PhPyNEt2)2]2+ (O)

Scheme 11.9 Thermal decomposition of [CuIII 2 (µ-O)2 (R-PhPyNEt2 )2 ]2+ leading to partial hydroxylation of the aromatic phenyl ring (• denotes oxygen atom from molecular O2 ) [73].

what was observed in Karlin’s system, there was almost no kinetic isotope effect when the phenyl ring was fully deuterated and the rate of decay increased upon increasing the electron-donating properties of the R group. Thus, from this work it seems clear that an electrophilic aromatic substitution pathway is also viable with O type of species. Nevertheless, even though no indication of the presence of a S P isomer was found, hydroxylation by a small, undetectable amount of such an isomer in rapid equilibrium with the experimentally detected O core could not be ruled out in this case. 11.3.3.2 Intermolecular ortho-Hydroxylation of Phenolic Compounds The reactivity of Cu2 O2 species toward exogenous substrates is specially relevant to catalysis and it has been thoroughly studied. These reactions include oxygen atom transfer, hydrogen atom abstraction, oxidation of alcohols to aldehydes or ketones [74–76], and conversion of amines to nitriles [77, 78], among others. However, a specially relevant reaction performed by some Cu2 O2 adducts is the ortho-hydroxylation of phenolic compounds to catechols or quinones [79], due to its direct relevance to the catalytic transformation performed by tyrosinase. Casella and coworkers reported the first synthetically prepared Cu2 O2 adduct capable of performing the ortho-hydroxylation of exogenous phenolates to give the corresponding ortho-diphenols (catechols) [80]. It consisted in a dinuclear copper complex bearing the dinucleating ligand L66 that combines an aliphatic amine and two benziimdazole units as the coordinating groups (Scheme 11.6). Reaction of [CuI 2 (L66)]2+ with O2 at low temperature (−78 ◦ C) afforded a species with a UV–vis spectrum (λmax = 362 nm, ε = 15 000 M−1 cm−1 ; λmax = 455 nm, ε = 2000 M−1 cm−1 ) and a rRaman pattern (ν = 760 cm−1 , [18 O2 ] = −41 cm−1 ) fully consistent with a S P compound (Scheme 11.10). Reaction of this transient Cu2 O2 species with p-carbomethoxyphenolate at −60 ◦ C gave the corresponding catechol product in about 40% yield (with respect to the dioxygen adduct formed). The yield of catechol product increased up to 90% when the related S P complex [CuII 2 (O2 )2 (MeL66)]2+ reacted with the phenolic substrate [81]. Interestingly,


281

R CuII N

N

II

D

N

N Cu

Cu N

D

D N

N N

II

Cu

D N

N

N

N II

Cu

II

N

N

CuII O−

X

N

X

N O CuII

[CuII2(O2)(LPy2)2 ] 2+ (sP)

II

[Cu 2(O2)(L66)]

2+ s

II

( P) or [Cu 2(O2)(MeL66)]

CuII

2+ s

( P) X

H X

O−

X

OH O

p - Substituted phenolate X = Me, F, Cl, COOMe, CN

Catechol product

CuII

CuII H

Scheme 11.10 Reported synthetic S P species that perform the ortho-hydroxylation of exogenous phenolates to form catechols (left) along with their proposed mechanism of action (right) [81, 83] (• denotes oxygen atom from molecular O2 ).

reaction of [CuI (L6Ph)]+ , the mononuclear analog of [CuI 2 (L66)]2+ , with dioxygen at low temperatures also allowed the detection of a dinuclear S P species capable of oxidizing exogenous phenolates to catechols (oxidation of sodium p-carbomethoxyphenolate was obtained in 88% yield with respect to the theoretical amount of Cu2 O2 species) [82]. One year after the publication of the work by Casella and coworkers, another synthetic system capable of performing a similar reaction was reported by Itoh et al. [83]. In this case, the tridentate mononucleating LPy2 ligand (Scheme 11.6) was used and a S P compound was unambiguously formed upon self-assembling of two mononuclear [CuI (LPy2 )]+ complexes by reaction with O2 at low temperatures. Reaction of a series of lithium salts of para-substituted phenolates with the newly formed S P species resulted in the formation of the corresponding catechol derivates in good yields (ranging from 60 to 90% depending on the specific substrate) (Scheme 11.10). Both in Itoh (LPy2 ) and Casella’s (L66, MeL66, or L6Ph) systems, it was well established that the oxygen atom incorporated into the catechol product derived from O2 (as ascertained by experiments using 18 O2 ), and that the rate of hydroxylation increased upon increasing the electron-donating abilities of the para-substituents in the phenolic ring (Hammett analysis afforded ρ values of −1.8 for LPy2 and −1.84 for MeL66). Finally, rate constants of the reaction between S P and phenolates were dependent on substrate concentration following Michaelis–Menten type saturation curves. Such a kinetic behavior is indicative of the formation of a complex constituted by an association between the substrate and the peroxo intermediate in the course of reaction. These results are consistent with an electrophilic attack of the S P intermediate to the phenolate ring posterior to the formation of a binary complex between the peroxo core and the phenolic substrate

282


(Scheme 11.10, right). The similarities between this mechanism of action and the one proposed for tyrosinase suggest that these synthetic systems can be thought as true models for the natural enzyme. Remarkably, the reaction of [CuII 2 (O2 )(LPy2 )2 ]2+ with neutral phenols (rather than phenolate substrates) resulted in the exclusive formation of C–C coupling dimer products (no catechol was formed) [84]. Presumably, this reaction occurs through a proton-coupled electron transfer mechanism (PCET) by which phenoxyl radical intermediates are generated and they spontaneously collapse. In the biological system (Section 11.2.2), it is proposed that the neutral phenol acting as the natural substrate becomes readily deprotonated upon interaction with the active center, and thus it is the corresponding phenolate that directly coordinates to the copper site. Instead, phenol deprotonation in synthetic systems is not possible, thus precluding direct use of neutral phenols to model tyrosinase reactivity and enforcing reaction with previously deprotonated phenolate salts. Overall, the results arising from the works by Casella and Itoh on synthetic models of tyrosinase indicate that the presence of a S P type of species is necessary to perform the ortho-hydroxylation of exogenous phenolic substrates. This idea is further supported by the fact that this type of peroxo unit is also directly detected in the natural system where the proposed mechanistic pathway entails an O–O bond cleavage occurring concomitantly with hydroxylation (Scheme 11.3). Nevertheless, the aromatic hydroxylation mechanism continues being subject of intense discussion because the existence of a rapid equilibrium between S P and O species [46] (Scheme 11.4) hinders the assignment of the true hydroxylation species in natural systems. Moreover, some recent examples reported in the last three years clearly indicate that the O unit is also competent to perform tyrosinase-like reactivity. The first example was reported by Stack and coworkers (Scheme 11.11) [77, 85, 86]. In this case, reaction of the mononuclear [CuI (DBED)]+ with O2 at −80 ◦ C resulted in the formation of [Cu2 (O2 )(DBED)2 ]2+ , which was unequivocally assigned as a S P species on the basis of its characteristic spectroscopic features (UV–vis: λmax = 350 nm, ε = 36 000 M−1 cm−1 ; 485 nm, ε = 1200M−1 cm−1 ; 605 nm, ε = 900 M−1 cm−1 . rRaman: ν = 721 cm−1 , [18 O2 ] = −40 cm−1 ) [85]. Reaction of this Cu2 O2 species with sodium 2,4-di-tert-butylphenolate at −80 ◦ C afforded a 1 : 1 mixture of 3,5-di-tert-butylcatechol and 3,5-di-tert-butyl-1,2-benzoquinone products (∼80% yield based on the initial amount of copper complex). More interestingly, when the reaction between S P species and the phenolic substrate was monitored by UV–vis spectroscopy at extremely low temperatures (−120 ◦ C), the formation of a transient intermediate assigned as a bis(µ-oxo) core (O) with a coordinated phenolic substrate was observed [86]. The transformation of the initially detected S P species into a O compound upon substrate coordination was clearly evidenced by rRaman spectroscopy, in which a characteristic bis(µ-oxo) stretching frequency at 590 cm−1 ([18 O2 ] = −20 cm−1 ) appeared once the substrate was added. Detailed mechanistic studies on the hydroxylation reactions indicated that electron-deficient phenolic substrates caused a decrease in the decay rate (Hammett parameter ρ = –2.2), which supports a reaction path involving an electrophilic aromatic substitution.


+2

N H

N H O2 CuI NCMe MeTHF N −120 °C H [CuI(DBED)]+

t

O−

Bu

H N

N H N O CuIII CuIII O O H N

N

t

Bu

O CuII Cu O N N H H II

283

MeTHF −120 °C

t

Bu

OH t

O OH tBu +

Bu t

t

Bu

t

Bu 30% [CuIII2(µ-O)2(phenolate)(DBED)2]+ (O)

[CuII2(O2)(DBED)2]2+ (sP)

O Bu

30%

Scheme 11.11 Schematic representation of the O–O bond breakage prior to substrate hydroxylation in [CuI (DBED)]+ system [86] (• denotes oxygen atom from molecular O2 ).

This mechanism was further favored by the observed inverse secondary C–H/C–D isotope effect and by DFT calculations modeling the hydroxylation process. Therefore, in this work, it was spectroscopically observed not only that the O–O bond cleavage occurs prior to the hydroxylation but also that the O isomer is the real executor of the chemistry (Scheme 11.11). This observation contrasts with the above mentioned idea that formulates the true hydroxylation species as a side-on peroxo (S P) type of intermediate. More recently, our group reported that O species [CuIII 2 (O)2 (m-XYLMeAN )]2+ formed upon reaction of the dinuclear copper complex [CuI 2 (m-XYLMeAN )]2+ with O2 is also capable of performing the ortho-hydroxylation of phenolates to the corresponding catechols (Scheme 11.12) [87]. In fact, sodium para-chlorophenolate was oxidized into 4-chlorocatechol in 67% yield with respect to the initial dicopper complex. Moreover, UV–vis monitoring of the hydroxylation reaction indicated the formation of a new brown intermediate (UV–vis: λmax = 390 nm, ε = 8000 M−1 cm−1 ; 485 nm, ε = 4500 M−1 cm−1 ) showing a rRaman spectrum with the characteristic bis(µ-oxo) core vibration frequency at ν = 597cm−1 ([18 O2 ] = −26 cm−1 ) together with vibrations corresponding to the phenolic

O−Na+

N N N

CuIII

CuIII

N

X

N

X = Cl, F, CN, CO2Me

N

N

Acetone, −80 °C

N

N

[CuIII2(µ-O)2(m -XYLMeAN)]2+ (O)

CuIII

N CuIII N N O

X [CuIII2(µ-O)2(phenolate)(m -XYLMeAN)]+ (O)

Scheme 11.12 ortho-Hydroxylation of phenolates by the O species [CuIII 2 (µ-O)2 (m-XYLMeAN )]2+ [87] (• denotes oxygen atom from molecular O2 ).

OH X

H

284


substrate. No isotope-sensitive features that could be assigned to a S P species were observed. This rRaman data together with the similarity with the UV–vis spectrum of [CuIII 2 (µ-O)2 (phenolate)(DBED)2 ]+ [86] provided direct evidence for phenolate binding to the Cu2 O2 in this newly formed transient intermediate that could be formulated as [CuIII 2 (µ-O)2 (phenolate)(m-XYLMeAN )]+ . In analogy to the previously reported systems, the rate of hydroxylation increased upon increasing the electron-donating properties of the phenolate, which is indicative of an electrophilic oxidizing species that attacks the aromatic ring. Thus, the system reported by our group is similar to the previous one described by Stack and coworkers in the sense that the O isomer is the real executor of the hydroxylation. However, a significant difference can be found between these two systems: while in [CuIII 2 (µ-O)2 (m-XYLMeAN )]2+ the O–O bond breakage occurs prior to phenolate coordination, in [CuII 2 (O2 )(DBED)]2+ the O–O breakage is concomitant with phenolate binding. Substrate accessibility to Cu2 O2 cores is a very important factor in determining the nature of the oxidative process. Indeed, the active sites of hemocyanin and tyrosinase are almost identical, but the different substrate accessibility to the copper center determines their different final activity. Taking this concept into account, it is interesting to refer to a recent work reported by Stack and coworkers [88] in which a series of bidentate ligands were prepared: they contain either two nonbulky peralkylated diamine units (LAA ), or two sterically crowded guanidine coordinating groups (LGG ), or a hybrid of both constituted by one guanidine and one amine group (LAG ) (Scheme 11.6). The corresponding mononuclear copper(I) complexes self-assemble upon reaction with O2 to form [CuIII 2 (µ-O)2 (L)2 ]2+ assigned as O species on the basis of UV–vis (λmax ∼ 297 nm, ε ∼ 19 000 M−1 cm−1 ; λmax ∼ 390 nm, ε ∼ 23 000 M−1 cm−1 ) and EXAFS spectroscopy. Despite the similar bis(µ-oxo) core, their reactivity toward exogenous phenolates was found to be completely different (Scheme 11.13). The O species coordinated to LGG ligands did not react with phenolates presumably due to the hindered access of the phenolic substrates caused by the steric demands of the four guanidine units surrounding the copper sites. Instead, the O species formed with the peralkylated diamine ligand (LAA ) (with minimal steric demands) reacted with phenolates to form radical C–C coupling dimers by an electron transfer mechanism involving initial electron transfer from the phenolate to the CuIII center. Finally, when hybrid ligands LAG were present, the resulting O core exhibited tyrosinase-like reactivity and it performed the ortho-hydroxylation of phenolates to give the corresponding catechols. Presumably, the substitution of one of the amine groups by a guanidine results in a less oxidant copper center (at parity of oxidation state, a guanidine is much stronger nitrogen σ -donor than a peralkylated amine) and thus outer-sphere electron transfer mechanisms are diminished. Overall, this work indicates that ligand LAG forms an O species that balances both Cu2 O2 accessibility and moderate oxidizing power, which allows efficient hydroxylation of phenols.

11.4 Concluding Remarks

N

N CuIII

N

N

CuIII

N

CuIII

N

N

N

N

N N

CuIII N

N

N

N

N

N CuIII

CuIII

N

N

N

285

N N

N

N

[CuIII2(µ-O)2(LAA)2]2+ (O) O

−

[CuIII2(µ-O)2(LAG)2]2+ (O)

[CuIII2(µ-O)2(LGG)2]2+ (O)

O−

OH

O−

OH

No reaction H

HO (a)

C – C coupling

(b) ortho -Hydroxylation

(c)

Scheme 11.13 Reactivity toward exogenous phenolates of three related O compounds with LAA (a), LAG (b), and LGG (c) ligands [88] (• denotes oxygen atom from molecular O2 ).

11.4 Concluding Remarks

In the past decades enormous effort has been devoted to the development of suitable biomimetic systems that can structurally and even functionally reproduce the active site of a specific protein. These studies have decisively contributed to a better understanding on how nature works and they are considered as a powerful tool to get fundamental insight into the mechanistic and structural details of biological systems. In this review chapter, we have focused on the description of those synthetic model systems capable of reproducing the activity of tyrosinase, a dinuclear copper protein that catalyzes the ortho-hydroxylation of phenols to catechols. Despite the fact that the fundamental structure of this enzyme is nowadays firmly established by X-ray diffraction studies [25], some key aspects of its mechanism of action remain unclear, such as the characterization of the true hydroxylation species. On the one hand, the spectroscopic detection of a µ-η2 : η2 -peroxodicopper(II) species (S P) in the natural system together with the ability of some S P model systems to perform the ortho-hydroxylation of phenols [81, 83] suggests that this type of configuration is responsible for the hydroxylation event. On the other hand, the recent development of selected functional models of tyrosinase containing a bis(µ-oxo)-dicopper(III) core (O) together with theoretical studies postulating the electrophilic character of such a structure [89] indicates that O is also competent to perform the hydroxylation of phenols. Nevertheless, the

286


existence of a well-known equilibrium between S P and O isomers (reversible O–O bond breakage) [46] precludes the clear identification of the real oxidant. Indeed, the two isomers may be competent for such a transformation and it might be just a matter of finding dinuclear copper centers that balance substrate accessibility and suitable oxidizing power [88]. Future work in mimicking tyrosinase activity will be directed toward the transformation of the stoichiometric processes achieved by the reported functional models into catalytic and synthetically useful reactions to perform the selective oxidation of phenols. Most likely, these studies will require the use of adequate exogenous reducing agents in order to complete the catalytic cycle. In spite of the need for further work in this topic, all the successful accomplishments in mimicking tyrosinase’s activity and its active site must be considered as a landmark for the performance of equivalent studies with other O2 -activating copper proteins. Indeed, biomimetic studies on mononuclear and multinuclear copper oxidases and oxygenases are still in their initial stages of development, especially compared to dinuclear copper proteins. All knowledge acquired during the past decades to model such proteins will be undoubtedly essential for an efficient and successful mimicking of other proteins involved in O2 activation.

References 1. Bäckvall, J.-E. (2004) Modern Oxidation

2. 3.

4. 5.

6.

7. 8. 9.

10.

Methods, Wiley-VCH Verlag GmbH, Weinheim. Sawyer, D.T. (1991) Oxygen Chemistry, Oxford University Press, New York. Bertini, I., Gray, H.B., Stiefel, E.I., and Valentine, S.J. (2007) Biological Inorganic Chemistry: Structure and Reactivity, University Science Books, Sausalito. Kovaleva, E.G. and Lipscomb, J.D. (2008) Nat. Chem. Biol., 4, 186. Holm, R.H., Kennepohl, P., and Solomon, E.I. (1996) Chem. Rev., 96, 2239. Costas, M., Mehn, M.P., Jensen, M.P., and Que, L. Jr. (2004) Chem. Rev., 104, 939. Seigbahn, P.E.M. and Borowski, T. (2006) Acc. Chem. Res., 39, 729. Nam, W. (2007) Acc. Chem. Res., 40, 465. Mirica, L.M., Ottenwaelder, X., and Stack, T.D.P. (2004) Chem. Rev., 104, 1013. Kraatz, H.-B. and Metzler-Nolte, N. (2006) Concepts and Models in Bioinorganic Chemistry, Wiley-VCH Verlag GmbH, Weinheim.

11. Que, L. Jr. and Tolman, W.B. (2008) Na-

ture, 455, 333. 12. Kaim, W. and Schwederski, B. (1998)

13.

14.

15. 16.

17.

18.

19.

Bioinorganic Chemistry: Inorganic Elements in the Chemistry of Life, John Wiley & Sons, Ltd, Chichester. Cotton, F.A., Wilkinson, G., Murillo, C.A., and Bochmann, M. (1999) Advanced Inorganic Chemistry, 6th edn, John Wiley & Sons, Inc., New York. Lippard, S.J. and Berg, J.M. (1994) Principles of Bioinorganic Chemistry, University Science Books, Mill Valley. Klinman, J.P. (1996) Chem. Rev., 96, 2541. Solomon, E.I., Sundaram, U.M., and Machonkin, T.E. (1996) Chem. Rev., 96, 2563. Bento, I., Carrondo, M.A., and Lindley, P.F. (2006) J. Biol. Inorg. Chem., 11, 539. Solomon, E.I., Chen, P., Metz, M., Lee, S.-K., and Palmer, A.E. (2001) Angew. Chem. Int. Ed., 40, 4570. van Holde, K.E., Miller, K.I., and Decker, H. (2001) J. Biol. Chem., 276, 15563.

References 20. Cuff, M.E., Miller, K.I., van Holde, K.E.,

21.

22. 23.

24.

25.

26.

27.

28.

29. 30. 31. 32.

33.

34.

35.

36.

37.

and Hendrickson, W.N. (1998) J. Mol. Biol., 278, 855. Eickman, N.C., Himmelwright, R.S., and Solomon, E.I. (1979) Proc. Natl. Acad. Sci. U.S.A., 76, 2094. Decker, H. and Rimke, T. (1998) J. Biol. Chem., 273, 25889. Himmelwright, R.S., Eickman, N.C., LuBien, C.D., and Solomon, E.I. (1980) J. Am. Chem. Soc., 102, 5378. Klabunde, T., Eicken, C., Sacchettini, J.C., and Krebs, B. (1998) Nat. Struct. Biol., 5, 1084. Matoba, Y., Kumagai, T., Yamamoto, A., and Sugiyama, M.J. (2006) J. Biol. Chem., 281, 8981. Himmelwright, R.S., Eickman, N.C., LuBien, C.D., Lerch, K., and Solomon, E.I. (1980) J. Am. Chem. Soc., 102, 7339. Morioka, C., Tachi, Y., Suzuki, S., and Itoh, S. (2006) J. Am. Chem. Soc., 128, 6788. Decker, H., Schweikardt, T., and Tuczek, F. (2006) Angew. Chem. Int. Ed., 45, 4546. Lewis, E.A. and Tolman, W.B. (2004) Chem. Rev., 104, 1047. Schindler, S. (2000) Eur. J. Inorg. Chem., 2311. Decker, H., Dillinger, R., and Tuczek, F. (2000) Angew. Chem. Int. Ed., 39, 1591. Jacobson, R.R., Tyeklár, Z., Farooq, A., Karlin, K.D., Liu, S., and Zubieta, J. (1988) J. Am. Chem. Soc., 110, 3690. Kitajima, N., Fujisawa, K., and Moro-oka, Y. (1989) J. Am. Chem. Soc., 111, 8975. Kitajima, N., Fujisawa, L., Fujimoto, C., Moro-oka, Y., Hashimoto, S., Kitagawa, T., Toriumi, K., Tatsumi,K., and Nakamura, A. (1992) J. Am. Chem. Soc., 114, 1277. Mahapatra, S., Halfen, J.A., Wilkinson, E.C., Pan, G., Wang, X., Young, V.G. Jr., Cramer, C.J., Que, L. Jr., and Tolman, W.B. (1996) J. Am. Chem. Soc., 118, 11555. Halfen, J.A., Mahapatra, S., Wilkinson, E.C., Kaderli, S., Young, V.G. Jr., Que, L.Jr., Zuberbühler, A.D., and Tolman, W.B. (1996) Science, 271, 1397. Tyeklár, Z., Jacobson, R.R., Wei, N., Murthy, N.N., Zubieta, J., and Karlin,

38.

39.

40.

41.

42.

43.

44.

45.

46. 47.

48.

49.

50.

51. 52.

K.D. (1993) J. Am. Chem. Soc., 115, 2677. Becker, M., Heinemann, F.W., and Schindler, S. (1999) Chem. Eur. J., 5, 3124. Weitzer, M., Schatz, M., Hampel, F., Heinemann, F.W., and Schindler, S. (2002) J. Chem. Soc., Dalton Trans., 5, 686. Weitzer, M., Schindler, S., Brehm, G., Schneider, S., Hörmann, E., Jung, B., Kaderli, S., and Zuberbühler, A.D. (2003) Inorg. Chem., 42, 1800. Börzel, H., Comba, P., Katsichtis, C., Kiefer, W., Lienke, A., Nagel, V., and Pritzkow, H. (1999) Chem. Eur. J., 5, 1716. Mahadevan, V., Hou, Z., Cole, A.P., Root, D.E., Lal, T.K., Solomon, E.I., and Stack, T.D.P. (1997) J. Am. Chem. Soc., 119, 11996. Cole, A.P., Mahadevan, V., Mirica, L.M., Ottenwaelder, X., and Stack, T.D.P. (2005) Inorg. Chem., 44, 7345. Spencer, D.J.E., Aboelella, N.W., Reynolds, A.M., Holland, P.L., and Tolman, W.B. (2002) J. Am. Chem. Soc., 124, 2108. Aboelella, N.W., Lewis, E.A., Reynolds, A.M., Brennessel, W.W., Cramer, C.J., and Tolman, W.B. (2002) J. Am. Chem. Soc., 124, 10660. Tolman, W.B. (1997) Acc. Chem. Res., 30, 227. Mahadevan, V., Henson, M.J., Solomon, E.I., and Stack, T.D.P. (2000) J. Am. Chem. Soc., 122, 10249. Hatcher, L.Q., Vance, M.A., Sarjeant, A.A.N., Solomon, E.I., and Karlin, K.D. (2006) Inorg. Chem., 45, 3004. Maiti, D., Woertink, J.S., Sarjeant, A.A.N., Solomon, E.I., and Karlin, K.D. (2008) Inorg. Chem., 47, 3787. Lam, B.M.T., Halfen, J.A., Young, V.G. Jr., Hagadorn, J.R., Holland, P.L., Lledós, A., Cucurull-Sánchez, L., Novoa, J.J., Alvarez, S., and Tolman, W.B. (2000) Inorg. Chem., 39, 4059. Stack, T.D.P. (2003) Dalton Trans., 1881. Liang, H.-C., Henson, M.J., Hatcher, L.Q., Vance, M.A., Zhang, C.X., Lahti, D., Kaderli, S., Sommer, R.D., Rheingold, A.L., Zuberbühler, A.D.,

287

288


53. 54.

55. 56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

Solomon, E.I., and Karlin, K.D. (2004) Inorg. Chem., 43, 4115. Cahoy, J., Holland, P.L., and Tolman, W.B. (1999) Inorg. Chem., 38, 2161. Henson, M.J., Vance, M.A., Zhang, C.X., Liang, H.-C., Karlin, K.D., and Solomon, E.I. (2003) J. Am. Chem. Soc., 125, 5186. De, A., Mandal, S., and Mukherjee, R. (2008) J. Inorg. Biochem., 102, 1170. Mahapatra, S., Young, V.G.Jr., Kaderli, S., Zuberbühler, A.D., and Tolman, W.B. (1997) Angew. Chem. Int. Ed. Engl., 36, 130. Kodera, M., Katayama, K., Tachi, Y., Kano, K., Hirota, S., Fujinami, S., and Suzuki, M. (1999) J. Am. Chem. Soc., 121, 11006. Karlin, K.D., Nasir, M.S., Cohen, B.I., Cruse, R.W., Kaderli, S., and Zuberbühler, A.D. (1994) J. Am. Chem. Soc., 116, 1324. Pidcock, E., Obias, H.V., Zhang, C.X., Karlin, K.D., and Solomon, E.I. (1998) J. Am Chem. Soc., 120, 7841. Company, A., Lamata, D., Poater, A., Solà, M., Rybak-Akimova, E.V., Que, L. Jr., Fontrodona, X., Parella, T., Llobet, A., and Costas, M. (2006) Inorg. Chem., 45, 5239. Liang, H.-C., Zhang, C.X., Henson, M.J., Sommer, R.D., Hatwell, K.R., Kaderli, S., Zuberbühler, A.D., Rheingold, A.L., Solomon, E.I., and Karlin, K.D. (2002) J. Am. Chem. Soc., 124, 4170. Mahapatra, S., Halfen, J.A., and Tolman, W.B. (1996) J. Am. Chem. Soc., 118, 11575. Itoh, S., Nakao, H., Berreau, L.M., Kondo, T., Komatsu, M., and Fukuzumi, S. (1998) J. Am. Chem. Soc., 120, 2890. Becker, M., Schindler, S., Karlin, K.D., Kaden, T.A., Kaderli, S., Palanché, T., and Zuberbühler, A.D. (1999) Inorg. Chem., 38, 1989. Costas, M., Ribas, X., Poater, A., López Valbuena, J.M., Xifra, R., Company, A., Duran, M., Solà, M., Llobet, A., Corbella, M., Usón, M.A., Mah´ıa, J., Solans, X., Shan, X., and Benet-Buchholz, J. (2006) Inorg. Chem., 45, 3569. Fusi, V., Llobet, A., Mah´ıa, J., Micheloni, M., Paoli, P., Ribas, X., and Rossi, P. (2002) Eur. J. Inorg. Chem., 2002, 987.

67. Ghosh, D. and Mukherjee, R. (1998)

Inorg. Chem., 37, 6597. 68. Menif, R., Martell, A.E., Squattrito, P.J.,

69. 70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

and Clearfield, A. (1990) Inorg. Chem., 29, 4723. Menif, R. and Martell, A.E. (1989) J. Chem. Soc., Chem. Commun., 20, 1521. Casella, L., Gullotti, M., Pallanza, G., and Rigoni, L. (1988) J. Am. Chem. Soc., 110, 4221. Sander, O., Henss, A., Näther, C., Würtele, C., Holthausen, M.C., Schindler, S., and Tuczek, F. (2008) Chem. Eur. J., 14, 9714. Poater, A., Ribas, X., Llobet, A., Cavallo, L., and Solà, M. (2008) J. Am. Chem. Soc., 130, 17710. Holland, P.L., Rodgers, K.R., and Tolman, W.B. (1999) Angew. Chem. Int. Ed. Engl., 38, 1139. Obias, H.V., Lin, Y., Murthy, N.N., Pidcock, E., Solomon, E.I., Ralle, M., Blackburn, N.J., Neuhold, Y.-M., Zuberbühler, A.D. and Karlin, K.D. (1998) J. Am. Chem. Soc., 120, 12960. Zhang, C.X., Liang, H.-C., Kim, E.-I., Shearer, J., Helton, M.E., Kim, E., Kaderli, S., Incarvito, C.D., Zuberbühler, A.D., Rheingold, A.L., and Karlin, K.D. (2003) J. Am. Chem. Soc., 125, 634. Shearer, J., Zhang, C.X., Zakharov, L.N., Rheingold, A.L., and Karlin, K.D. (2005) J. Am. Chem. Soc., 127, 5469. Mirica, L.M., Rudd, D.J., Vance, M.A., Solomon, E.I., Hodgson, K.O., Hedman, B., and Stack, T.D.P. (2006) J. Am. Chem. Soc., 128, 2654. Mahadevan, V., DuBois, J.L., Hedman, B., Hodgson, K.O., and Stack, T.D.P. (1999) J. Am. Chem. Soc., 121, 5583. Battaini, G., Granata, A., Monzani, E., Gullotti, M., and Casella, L. (2006) Adv. Inorg. Chem., 58, 185. Santagostini, L., Gullotti, M., Monzani, E., Casella, L., Dillinger, R., and Tuczek, F. (2000) Chem. Eur. J., 6, 519. Palavicini, S., Granata, A., Monzani, E., and Casella, L. (2005) J. Am. Chem. Soc., 127, 18031. Battaini, G., De Carolis, E., Monzani, E., Tuczek, F., and Casella, L. (2003) Chem. Commun., 726.

References 83. Itoh, S., Kumei, H., Taki, M., Nagatomo, 87. Company, A., Palavicini, S.,

S., Kitagawa, T., and Fukuzumi, S. (2001) J. Am. Chem. Soc., 123, 6708. 84. Osako, T., Ohkubo, K., Taki, M., Tachi, Y., Fukuzumi, S., and Itoh, S. (2003) J. Am. Chem. Soc., 125, 11027. 85. Mirica, L.M., Vance, M., Rudd, D.J., Hedman, B., Hodgson, K.O., Solomon, E.I., and Stack, T.D.P. (2002) J. Am. Chem. Soc., 124, 9332. 86. Mirica, L.M., Vance, M., Rudd, D.J., Hedman, B., Hodgson, K.O., Solomon, E.I., and Stack, T.D.P. (2005) Science, 308, 1890.

Garcia-Bosch, I., Mas-Ballesté, R., Que, J.L., Rybak-Akimova, E.V., Casella, L., Ribas, X., and Costas, M. (2008) Chem. Eur. J., 14, 5335. 88. Herres-Pawlis, S., Verma, P., Haase, R., Kang, P., Lyons, C.T., Wasinger, E.C., Flörke, U., and Stack, T.D.P. (2009) J. Am. Chem. Soc., 131, 1154. 89. Spuhler, P. and Holthausen, M.C. (2003) Angew. Chem. Int. Ed., 42, 5961.

289

291

Part V Structure–Property Relationship and Biosensing


293

12 Chirality in Biochemistry: A Computational Approach for Investigating Biomolecule Conformations Adriana Pietropaolo

12.1 Introduction 12.1.1 Molecular Chirality in Living Systems

Since its discovery by Louis Pasteur in 1848, molecular chirality has captured the collective imagination of the chemical community. Pasteur’s realization that tartaric acid could be observed in number of chemically inequivalent forms (R, S, racemic and meso forms) led to a rich new vein of chemical discovery based on the study of what he christened chirality from the greek, cheir, meaning hand. Any molecule that has no improper rotational symmetry is by definition chiral because it cannot be superimposed on its mirror image. In other words molecules with no improper rotational symmetry can exist in two nonequivalent, enantiomeric forms, named R, S, following the Cahn–Ingold–Prelog priority rule or l, d in the Fisher notation for glyceraldehyde stereocenters (Figure 12.1), being these single notations not correlated each other. In nature many biomolecules do not have improper rotational symmetry and are thus chiral. All the natural amino acids, with the exception of glycine, are chiral and therefore have two enantiomeric forms commonly termed l and d. Amino acids are almost exclusively observed in the l form in biomolecules, while in natural carbohydrates the individual monomers are found in the d form. This selectivity in biomolecules, which is known as homochirality or using the Pasteur notation dissimilarity, has aroused the curiosity of many scientists; many of whom have expended a great deal of effort in attempts to understand why nature is so strict about chirality. Proteins, the building blocks of the cell, are typically composed of a chain of homochiral l-amino acids. Only in the cell wall of some microorganisms [1] and various aged human tissues (tooth, bone, brain, etc.) [2] d-amino acids are observed. Nanda et al. [3] recently investigated the role of homochirality in the folding free energy landscape of proteins. They showed that syndiotactic chains, in which there is an alternation in the chirality of adjacent monomers, have a large ensemble of accessible conformations (Figure 12.2). Hence, syndiotactic chains have a much Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

294

12 A Computational Approach for Investigating Biomolecule Conformations

Energy

Figure 12.1 Graphical representation of enantiomers, due to stereocenter chirality generated by the reflection through a vertical plane.

Achiral Homochiral Heterochiral

Unfolded Homo + heterochiral

Conformational states Figure 12.2 Energy landscape representation as described in [3]. Width of each cone represents the number of accessible conformational states, while their depth underlines the energy linked to structural stability. Unfolded conformations correspond to the highest energy level, here represented by the cone bases, whose perspective is given by the square.

greater degree of flexibility in their backbone, than isotactic polypeptides, in which the chiralities of all stereocenters are identical. The fact that glycine residues, the only amino acid that lacks the Cα stereocenter, can provide flexibility to proteins and can destabilize folded states [4, 5] strengthens the suggestion that chirality plays an important role in protein flexibility. This dependence of protein folding on chirality occurs because the chirality of individual amino acids can affect local secondary structures which in turn can affect the tertiary structure of the protein. For example, exchanging a number of l-amino acids for d-amino acids in Sso7d ribonuclease changes the conformation of the active site and prevents the protein function [6]. Chirality has therefore a deep role in complex systems like

12.1 Introduction

proteins in promoting the formation of inter-residue contacts that stabilize folded configurations. 12.1.2 Protein Secondary Structures

Protein structures are typically described using a system that separates the various aspects of their structures. The so-called primary structure describes the amino acid sequence of which the protein is composed, while tertiary structure refers to the global structure adopted by a single protein. This tertiary structure can be thought of in terms of a number of highly, regular substructural motifs, or secondary structure elements, the structures of which depend on the underlying primary structure. Proteins can adopt a number of different secondary structures (Figure 12.3), the most common being the right-handed helices (which include the α, the 310 and the π helices), the β sheet, the β turn and the left-handed helices (an example of which is the poly-l-proline II (PPII)). The following are the characteristics of each of these structures:

Figure 12.3 The most common protein secondary structures: α helix is shown by tubes, β sheets by thick ribbons, turn by narrow ribbons. H-bonds are shown by dotted lines.

295

296


• α helix: This structure, which was discovered by Linus Pauling and which typically involves more than four residues, has a pitch of 5.4 A˚ and is stabilized by hydrogen bonds between the carbonyl oxygen of residue i and the amide hydrogen of residue i + 4. • β sheets: This structure is flat and the hydrogen bonds it contains are either parallel or antiparallel depending on the orientation of the chain. Often an isolated β sheet is referred to as a bulge. • β turns: In these structures hydrogen bonds are formed between the ith and i + 3th residues. Right-handed 310 helices are β turns in which more than three residues are involved. ˚ • Poly-L-proline II: It has an elongated structure that involves a pitch of 9.3 A. • Coil: A Coil is any segment of an amino acid chain that has no secondary structure. Any given protein segment is said to have no secondary structure if every pair of (φ, ψ) angles is independent of all the other pairs of (φ, ψ) angles within the segment [7]. The secondary structure of a protein is very important as it is presumed that the global folding event is driven by local folding. Furthermore, as already mentioned, the tertiary structure, which controls the protein function, can be considered to be a collection of secondary elements. The incorrectly folded (misfolded) secondary structural elements can give rise to a different tertiary structure or even a protein with altered properties and can prevent the protein function. What is more, introduction of induced fit concepts surpassed rigid lock and key ideas in the theories of protein docking. This means that, now more than ever, it is necessary to understand the dynamical structures of proteins as well their static structures, since there is a direct link between protein conformational change and protein function. 12.1.3 Protein Secondary Structure Assignment

A number of methods for secondary structure assignment have been proposed (see, e.g., [8–11] and there is about 80% agreement between structural assignments made with different methods. The first key contribution was probably by Ramachandran et al. [12], who correlated the native distribution of the –N–Cα – and –Cα –C– protein dihedral angles (φ, ψ) with the secondary structure [13]. However, this approach fails when there is a great deal of conformational flexibility, which is common for peptides, because the flexibility gives rise to nonstandard backbone angles [14–16]. One of the most commonly used programs for secondary structure determination is the ‘‘dictionary of protein secondary structures’’ (DSSP) [17]. The DSSP is reliant on an algorithm that recognizes hydrogen bond patterns involving the C=O and N–H backbone atoms and neglects the φ and ψ dihedral angles. Structural classification is thus based on eight qualitatively, different classes, and small deviations in the backbone dihedral angles from the ideality, which are often important when it comes to proteins’ biological function, can be neglected.

12.2 Computational Techniques for Studying Protein Dynamics

Moreover, DSSP analysis is known to be error-prone when it comes to the exact detection of the edges of a given motif [18]. There are various other methods that improve on the original DSSP method. The first of these is the so-called STRIDE algorithm [19], in which a consideration of the contributions from the dihedral angles is included in addition to the descriptions of the hydrogen bond patterns in the identifiers for various classes. A recently reported algorithm for classifying turn structures [20] works by screening and clustering a large data set taken from a protein data bank. From this analysis the three normal, four open, and five reverse turn families emerged along with a number of new turn types. Despite all the algorithmic sophistication, the assignment of secondary structure to conformations that, like PPII structures [21, 22], depart strongly from any ideal backbone structure still remains a challenge. What makes the structural assignment of protein structure in this ‘‘twilight zone’’ [23] so difficult is the high degree of flexibility, which makes the coil state become a kind of catch-all structural assignment unless high resolution data are obtained. 12.1.4 Intrinsic Chirality and Protein Secondary Structures

In Section 12.1, only chirality related to stereocenters has been discussed. Such stereocenters have a chirality due to the asymmetry arising from a noncoplanar arrangement of atoms in which one atom is connected to ‘‘at least’’ three further nonequivalent atoms. The most ubiquitous example of this is a carbon bonded to four atoms that lie on the vertices of a tetrahedron centered on the carbon (Figure 12.1). This, however, is not the only form of chirality that can exist as there are countless other three-dimensional structures that have no stereocenter and are also chiral. An example of an intrinsically chiral structure that does not necessarily need to have any stereocenters is a helicoidal structure. Proteins tend to adopt more chiral (α helix, 310 helix, β turn, PPII) than achiral (β sheets) local, secondary structures.

12.2 Computational Techniques for Studying Protein Dynamics

Several computational techniques are available to study the motions of biological macromolecules, some commonly used examples of which are molecular dynamics (MD), Monte Carlo, and normal mode analysis (NMA) approaches. All these methods generally use the equations of classical mechanics together with a forcefield that is used to calculate the potential energies or forces. These forcefields are based on either terms that describe the interactions between atoms (atomistic models) or terms that describe the interactions between groups of atoms (mesoscale models). If the goal is to analysze the fine details of local arrangements, atomistic scale simulations are preferred. On the other hand, if the purpose is to obtain an

297

298


understanding of the spatial arrangements of the macromolecule, coarse-grained (mesoscale) approaches are preferred as the longer timescales accessible make it possible to examine huge rearrangements of the system. The simplest approach to study protein folding within a coarse-grained model is provided by the Go− potential [24]. The basic idea of which is to devise a forcefield consistent with the experimentally established structure of the native state, in which only the interactions between Cα –Cα are taken into account. These pairs are within some small cutoff distance (usually 7.5 A˚ [25]) in the native state structure. As only Cα atoms are included, the model provides only a very coarse determination of the protein motions. Another commonly used approach to describe large motions in proteins is NMA, which is based on the harmonic approximation of the potential energy. Here the dynamics of the molecule is then described in terms of a collection of independent harmonic oscillators on a basis of normal modes eigenvectors (an orthonormal set of directional vectors that represent the uncoupled motions of the system). Low-frequency oscillations represent large-amplitude, collective motions that often correlate well with the experimentally observed conformational changes associated with protein function. To perform an NMA, first, the potential energy of the system must be minimized, which is calculated using either an atomistic or coarse-grained approach, and then a diagonalization of the matrix of second derivatives of the potential energy [26] (the Hessian) should be performed. The method that can, in theory, provide the greatest amount of detail about protein motion is MD, which is made possible by the fact that forces are expressed as the gradient of a potential energy function. MD is a computational technique in which the time evolution of a set of interacting particles (generally atoms or molecules) is followed step by step by integrating their equations of motion. In contrast to Monte Carlo simulations, which employ stochastic dynamics to explore phase space, MD employs deterministic dynamics, which means that given an initial set of positions and velocities, the subsequent time evolution is completely determined and in principle reversible. The major disadvantage of this technique is that it is currently only possible to run at most a few microseconds of simulation, while most processes of interest in nature take place on the order of seconds. However, more advanced variants of the MD techniques have been developed [27–30] – in which longer timescales can be investigated – and used to obtain intriguing results [31].

12.3 Employing Chirality to Analyze Protein Motions 12.3.1 The Chirality Index

So far, we have discussed the main concepts used in the study of protein structure. However, studying the structural data that can be gleaned from molecular

12.3 Employing Chirality to Analyze Protein Motions

299

simulation is usually far from simple, because proteins are not static entities and instead exist as dynamical entities. Typically, the dynamical motions of proteins involve local rearrangements around the average structure; namely, that which can be probed experimentally using techniques such as NMR. A natural way to describe such local fluctuations in protein structures would involve an examination of changes in the local secondary structure of the protein. However, a description of these changes is beyond the scope of the standard tools such as DSSP [32] or STRIDE [33] that have been described. The exception to this is local chirality which is highly sensitive to local conformational motions and can thus provide insight into local conformational changes and the flexibility of particular protein segments [34]. The chirality method works as it is possible to calculate individual indices for fragments of the protein backbone. For each of these fragments, a chirality index can be computed using a method that was proposed by Marti et al. and Chou [35, 36] for low mass molecules and which succeeded in analyzing facial diastereoselectivity [37]. Here, however, instead of calculating the chirality index for the entire molecule, it is calculated for each individual fragment of the molecule separately (as shown in Figure 12.4). The value of the index for each protein fragment is expressed as follows: Ga,Na =

all p 4 gijkl 3N4a p

(12.1)

i,j,k,l

gijkl =

    

[(rij ×rkl )·ril ](rij ·rij )(rjk ·rkl )

0

if rij , rkl , ril , rjk ≤ re , a ≤ i, j, k, l ≤ Na + a − 1 otherwise

(12.2)

where a is the first (closest to the N terminal) atom of a given sequence of Na = 15 consecutive backbone atoms with coordinates r and rc = 12 A˚ is a cutoff radius, added to avoid the computation of negligible long-range terms. The efficacy of this method is that the instantaneous value of the chirality indices can be calculated only from a knowledge of the atomic coordinates. What NR

Ca N a

N C

Ca

C Ca

N

Ca

C Na

Figure 12.4 A backbone composed of 15 atoms. The G index is calculated for every permutation of four atoms inside this 15 atoms fragment. Then, second G index is calculated ahead of one atom, here underlined by a line starting from the Cα .

Ca

C

N

N

Na+a−1 C

N

C Ca

12 A Computational Approach for Investigating Biomolecule Conformations Table 12.1 Average G values and relative standard deviations of G for ideal secondary structures, involving at least NR residues. Each structure was built by sampling φ and ψ angles from a Gaussian distribution, centered at the ideal φ and ψ values with σ = 15◦ (see [30]).

Structure

G

σG

NR

α helix 310 helix β turn I β sheets PPII π helix

−0.04 −0.07 −0.07 +0.00 +0.10 −0.01

0.02 0.01 0.01 0.01 0.03 0.02

>3 >3 2,3 ≥2 >3 >3

0.10 PPII Sheets a helix Type I b turn 310 helix p helix

0.05

Ga

300

0.00

−0.05

−0.10 3

4

5

6

7

8

9

Residue number

Figure 12.5 Chirality index G along the backbone for different ideal secondary structures. The cutoff used is rc = 12 A˚ and Na = 15 atoms is considered.

is more, because it is possible to calculate the chirality index in idealized secondary structures (here reported in Table 12.1, Figure 12.5), we can determine which secondary structure a local region of the protein has and how much the structure deviates from that structure during its dynamical motions. 12.3.2 Using Chirality to Understand Protein Structure

Any function that purports to be able to determine secondary structure in proteins must provide a range of values for each of the possible secondary structures. To


assign these ranges, ideal secondary structures were analyzed. To do this, the periodicity in the backbone angles (φ, ψ) was used to generate ideal segments of protein structures for which chirality indices were computed. The periodicities employed were (−67◦ , −41◦ ) for α helix [34], (−49◦ , −26◦ ) for 310 helix [35], (−67◦ , −59◦ ) for π helix [34], (−60◦ , −30◦ , −90◦ , 0◦ ) for type I β turns [36], (−75◦ , 147◦ ) for PPII helix [37], and (−130◦ , 130◦ ) for sheet regions [38]. In all cases, the ω angles were kept fixed to the trans value of 180◦ . The type I β turn conformation is not periodic in proteins, as it generally involves only four consecutive residues; however, for the sake of simplicity, it was considered periodic in [30] so that comparison with the other structural motifs is possible. Figure 12.5 shows the behavior of the G index for the ideal structures along the ˚ The opposite backbone. Here, the index was calculated using Na = 15 and rc = 12 A. chirality indices of the right-handed α helix (negative) and the left-handed PPII helix (positive) is worth noting together with the observation that β sheets have no chirality because of their flat shape. Figure 12.6 shows the chirality index calculated along the backbone of a typical protein taken from the protein data bank of structures. It is possible to note from this figure that in most of this protein the chirality index is in one of the ranges corresponding to the idealized structures. In addition to its ability to recognize secondary structure types, the chirality index is very sensitive to poly-l-proline dihedrals. A positive peak in the G value indicates that at least one amino acid in a given protein region has the PPII structural motif. Figure 12.7 shows the G value of a model poly-l-proline peptide fragment (1JMQ, residues 51–60), which

Hemoglobin G0s a 310 Turn Bend Coil

0.05 0.00

−0.05

DSSP

Ga

0.10

6 5 4 3 2 1 0

−0.10 10

30

50

70 90 110 Residue number

Figure 12.6 Chirality index G along the backbone for hemoglobin protein, from X-ray structure. Typical secondary structures, with negative periodicity concerning the α helices and with typical G values for the

130

other secondary structures (cf. Figure 12.5), are easily identified. The DSSP assignment is also plotted as the numeric code: 310 = 0, turn = 1, bend = 2, bridge = 2.5, α = 3, sheets = 4, and coil = 5.

301


6 5 4 3 2 1 0

0.12 0.10 0.08 0.06 0.04

G G ideal ppII DSSP–coil

0.02 0.00

DSSP

0.14

Ga

302

3

4

5 6 Residue number

7

Figure 12.7 Chirality index G along the backbone for poly-L-proline II model peptide. The DSSP assignment is also plotted as the numeric code: 310 = 0, turn = 1, bend = 2, bridge = 2.5, α = 3, sheets = 4, and coil = 5.

is known to have poly-l-proline dihedral angles. A good overlap between the PPII ideal structure and the PPII model peptide exists in the residues labeled 3–5 in the figure; however, after residue 5, the G value of 1JMQ drops as it must take residues 7 and 8 into account that are not in the PPII conformation. Detection of the PPII structure using DSSP-like algorithms is hampered because prolines do not form hydrogen bonds. Therefore, despite the fact that this structure is observed in other polypeptides, its extended conformation (9.3 A˚ pitch) is not conducive to the assignment of a well-defined hydrogen bond pattern, which means they are commonly misclassified as loops or coils [22]. G’s sensitivity to these PPII structures makes the G function an important tool for the better identification of this particular structure class. 12.3.3 Chirality Index as a Tool for Monitoring Protein Dynamics

Because the chirality index can be calculated using the atomic coordinates, it can prove useful in examining protein dynamics, for example, the dynamics of a transition from an α helix to a 310 helix or from a coil to a helixes (Figure 12.8a and b, respectively). It is particularly useful for the examination of folding transitions, as it contains information on both the φ and ψ dihedral angles. Moreover, G index is stable during dynamics and in response to random thermal fluctuations [30], which proves that it can be used to follow dynamical processes. Furthermore, the average value of the chirality index can be used to detect protein motions, and


Ga

0.01 −0.01 −0.03

DSSP

Residue 5 G7 a 310 Turn Bend

6 5 4 3 2 1 0

−0.05 −0.07 −0.09 −0.11

0

20

40

60

80

100

120

Time (ns)

0.10 Residue 6

Ga

0.05

G10 DSSP a Turn 310 Bend Coil Sheets b bridge

0.00 −0.05 −0.10 −0.15

(b)

0

10

20

30 40 Time (ns)

50

6 5 4 3 2 1 0 DSSP

(a)

60

Figure 12.8 (a) α to 310 helix transition, followed by chirality. (b) Coil to helix transition, underlined from the positive-to-negative switch of Ga .

therefore useful for studying protein flexibility. This is an important issue in the study of intrinsically disordered proteins. Figure 12.9 reports the averages and standard deviations of the chirality indices calculated along the backbone of rigid (a) and flexible structures (b), respectively, during the course of an MD trajectory. As previously discussed, the average chirality provides information on the average local structure adopted, which, together with the standard deviations of G, provides information on the flexibility of local regions in the protein. The typical method used to measure this quantity is the root mean

303

−0.10

−0.05

0.00

3

4

5

6

7

8 9 10 11 12 13 14 15 Residue number

Hemoglobin 1–18 110 ns Hemoglobin 1–18 PDB

(b)

−0.10

−0.05

0.00

0.05

0.10

5

10

15 20 Residue number

25

1 REI 1–30 50 ns 1 REI 1–30 PDB

30

Figure 12.9 Standard deviations of G among (a) hemoglobin 1–18 and (b) 1REI 1–30 configurations. It is worth to note the persistence of the chirality index for a rigid peptide, as the first helix of hemoglobin, and the high values of the standard deviations for a flexible peptide, as the trajectories of the first 30 residues of 1REI immunoglobulin antigen. As comparison the G from PDB and from the trajectories is shown.

(a)

Ga

0.05

Ga

304



square deviations (RMSDs), which, in fact, measures only the difference in the structure from a given reference configuration and is not particularly sensitive to the changes in dihedral angles that local chirality measures. In particular high RMSD indicates that the structure highly differs from the reference structure. However, the reference structure may not be the true average structure and, even if it is, the RMSD may not represent significant motions particularly if the single configurations examined are far from the average structure. The standard deviation of the chirality index makes sense, however, as it is a deviation about an average in a quantity that measures secondary structures. This average quantity is zero if no secondary structure is present, whereas if a secondary structure is present then a constant value for the average chirality index is returned and this depends on the average secondary structure. Meanwhile, the standard deviation measures directly the flexibility of the secondary structure in a local region of the protein. 12.3.4 Chirality and Circular Dichroism

Circular dichroism (CD) experiments are connected in depth with chirality because the l and d enantiomers absorb polarized light differently. These experiments thus provide a useful probe for chiral molecules but not for the achiral ones. The CD spectra of proteins are strongly related to the backbone dihedral angles φ and ψ, and can be used to obtain insight into the average structure the proteins adopt in solution. What is more, combining information derived from CD and MD may help in unraveling the conformational ensemble adopted in solution. As an example, we consider the CD spectra of the avian prion hexarepeat region as a function of pH [16]. Prion proteins (PrPC ) are glycoproteins that in mammals, but not in avians, can cause prion diseases, which involve the incorrect folding (misfolding) of proteins. It is commonly believed that this misfolding involves a conversion from a structure rich in α helices to a βsheet–enriched pathogenic isoform (PrPSc ). Until now, NMR structure determination of the prion N-terminal portion has been hampered because of its flexibility. Mammal and avian proteins show different N-terminal tandem repeats, PHGGGWGQ and PHNPGY, respectively. Both these contain histidine, however; only the avian proteins involve tyrosine in their primary sequence. Both these residues are of particular interest because they are highly sensitive to pH variations, having average pKa values of 6.1 and 9.9, respectively [16]. The far UV–CD spectra of the avian tetra-hexarepeat (PHNPGY)4 , reported in Figure 12.10, show a signal varying with pH. At pH 4, a pH at which one would expect all histidyl residues to be protonated (average pKa 6.1), the spectrum is broad and has a minimum at 203 nm, a weak shoulder around 216 nm, and a maximum at 230 nm. In general, this shape indicates an equilibrium between different conformations, suggesting the presence of other secondary structure elements besides the random coil. This spectrum shape is similar to that found for shorter peptide fragments as reported in [15, 39]. For these smaller fragments, the presence of both random coil and β-turn structures was suggested for those primary

305

G

−0.09 −0.12

−0.12

0.00

0.03

0.06

0.09

0.12

3

250

−0.06

18

240

−0.09 21

220 230 λ (nm)

LH84+


210

−0.06

6

200

pH 4 pH 7 pH 8 pH 10

−0.03

3

−50 190

−40

−30

−20

−10

−0.03

0.00

0.03

0.06

0.09

0.12

∆e (M−1 cm−1)

0

G

10

6


18

LH4

21

LH41

G

−0.12

−0.09

−0.06

−0.03

0.00

0.03

0.06

0.09

3

6

4H+

−4H+

0.12

4H+

−4H+

L4−

LH84+


LH42

18

L4−

21

306



sequences that encompass a PXXP motif, a sequence that normally favors β-turn and/or PPII structures [40]. When the pH is increased the spectrum of the avian tetra-hexarepeat (PHNPGY)4 changes in a different way as shown in Figure 12.10. Up to pH 7, there is no change to the band centered at 230 nm, but there is a shift in the position of the minimum, which takes it toward 200 nm, and a decrease in the bands’ intensity. Further increases in the pH cause a general broadening of the spectra and a significant decrease of the signal at 230 nm. Finally, at pH 10, a new maximum at 250 nm is observed. It is well known that aromatic side chains give rise to a contribution to the far UV–CD spectra of peptides and that in phenols this far UV signal is redshifted in the deprotonated phenolate ion [41, 42]. Hence, the new maximum, observed at 250 nm, is attributable to the deprotonation of tyrosine residues. Furthermore, the decrease and subsequent disappearance of the positive maximum in the signal at 230 nm is also related to the protonation state of the tyrosine residue as it is caused by the phenolic group on it. Alongside the evidence the UV–CD spectrum provides about the protonation state of the tyrosine residue, changes in the peptide secondary structure, which occur as a function of pH, can also be detected. For instance, the strong positive band at 190 nm (Figure 12.10) and the shoulder found at 216 nm are features typical of β-turn-like conformations [43–45], which appear to be predominant at neutral and basic pH. The weakening of these signals at lower pH suggests that the equilibrium configuration is shifted, thus it would seem that the conformation adopted is strongly dependent on the protonation state of the histidine and tyrosine residues. A chirality analysis of the conformational ensemble obtained from MD simulations reveals that there is an enhancement in the number of turn conformations as pH increases (this is indicated by the negative sharp peaks shown in the bottom side of Figure 12.10). In particular, at acidic pH (LH4+ 8 ), the broad negative peak centered at residue 7 (Pro), which includes residues 6–8 (Tyr, Pro, and His, respectively), confirms the presence of a 310 helix structure [30], while the negative sharp peak centered at residue 11 (Gly) suggests the presence of a turn that includes residues 9–12 (Asn, Pro, Gly, and Tyr, respectively). At neutral pH (LH4 ), chirality analysis indicates the presence of a turn region that includes residues 4–6 (Pro, Gly, and Tyr, respectively). This turn is highly flexible, however, as indicated by the wide standard deviation and the average value for this negative peak is close to the upper bound of values observed for this particular structural motif [–0.1 : −0.06]. Two further turn regions that are signaled by much stronger negative peaks are observed at neutral pH and are centered at residues 11 and 17 (Gly). At basic

Figure 12.10 Up on the left, CD spectra of avian tetra-hexarepeat fragment as a function of pH. At basic pH, the shoulder approximately at 216 nm reveals the formation of type I β turn. Up, on the right, typical conformations of avian tetra-hexarepeat fragment at acidic (LH8 4+ ), neutral (LH4 ), and basic (L4− ) pH. Bottom, chirality index along the backbone of the avian tetra-hexarepeat trajectories in the three protonation states. The shift toward type I β turn is underlined from the negative sharp peaks, adopted mainly in the basic state, L4− , as shown in the CD spectra. A circle underlines the side chains hydrogen bonds. N and C termini are shown, respectively, from left to right.

307

308


pH (L4− ), those negative peaks that were observed at neutral pH show smaller values for their standard deviations. This suggests that the stability of turn regions, inside the peptide, is enhanced by the increasing pH. Moreover, by examining the chirality pattern along the backbone, we can observe that on changing the pH the chirality index varies most strongly for residues 4–8 (Pro, Gly, Tyr, Pro, and His, respectively) and for the C-terminal region. These are the regions in which His8 and Tyr24 are involved and this indicates the pivotal role that these residues play in any conformational change. The turn structures observed in the MD simulations are of course stabilized by the large number of proline residues. However, in this system it would appear that histidine, particularly histidine 8, also plays a key role in stabilizing turn regions. This is evidenced by the fact that at physiological pH, when this residue is expected to be deprotonated, turn regions become more stable. The above example has demonstrated how a combination of experimental CD and chirality analysis of simulation data can be used to explain how the conformational states accessible to this molecular system depend on pH. Furthermore, Figure 12.10 shows that at physiological pH the chirality pattern of the tetra-hexarepeat region reflects the periodicity in the primary structure. This finding reinforces the suggestion that local folding events mainly drive the global protein folding toward the tertiary structure.

12.4 Perspectives

The connections between chirality and biosystems, particularly proteins, have been discussed here at length. It has been shown that protein structures are particularly amenable to descriptions based on chirality concept since proteins have intrinsically asymmetric structures. In particular, a chirality index has been introduced that can be used to analyze the extent to which proteins are folded and the local flexibility of the protein. This index has well-defined values for the typical secondary structure elements and is particularly effective in detecting PPII motifs, often misclassified as coils when other techniques are employed. One word of caution regarding coil states must be mentioned here: Coils are often considered to be unstructured regions in the protein and thus on the face of it would be expected to be achiral. However, because proteins are isotactic, regions of proteins that adopt coiled states should present some degree of left-handed chirality and thus should not have a chirality index of zero [30]. A particularly interesting class of protein structures is the so-called coiled coils, examples of which include keratine and the muscle protein tropomyosin. These structures consist of α helices wound together in a manner that is stabilized by hydrophobic interactions. What makes them intriguing from a chirality point of view is that while the α helices themselves are right handed the structure is often overall left handed. Recently, a correlation between the chirality of the coiled coils and their mechanical properties has been found [46]. This correlation links the

12.4 Perspectives

plasticity of coiled coil structures with the partner intrinsic chirality in the selection of biochemical partners. A particularly intriguing question related to these discussions is the chirality of amyloid suprastructures. Amyloids, like A β 1−40 and hen lysozyme, are fibrillar aggregates of proteins with a characteristics cross-β conformation that until now were generally accepted to have a structure composed of left-handed helices. Recently, however, a surprising discovery was made concerning the peptide of serum amyloid A protein (SAA1−12 ). It was found that structures composed of amino acids with S stereocenters form right-handed helices, while those composed of R stereocenters form the left-handed suprastructures [47]. Hence, it is clear that supramolecular chirality depends on the structure organization and it need not be necessarily the same for all fibrils. Fascinating proteins with high symmetries are found in viruses. A virus consists of a fragment of deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) enclosed in a highly regular arrangement of capsomer proteins. In spherical viruses, this external, capsid shell has its individual capsomer molecules located on the vertex positions of an equilateral triangulation of a sphere [48] with an icosahedral symmetry [49, 50]. Viruses can also adopt nonspherical geometries in which there is a skew in the capsid, which can cause it to become chiral. These chiral-shaped capsids are found to require a higher rupture force than the perfect icosahedral capsids [51]. Proteins play a central role in the life sciences as both the molecular workhorses and building blocks of all living things. Simulation and experiment are beginning to provide genuine insight into how these complicated molecules function in isolation, but how they interact or organize themselves to form a cell is an area where a lot of work is still required. Recently, however, in a discovery that surprised the community, cells were found to have a chiral organizing principle, which works in terms of asymmetrical compartmentalization and probably originates in the centrosome [52]. Xu et al. [53] argue that intrinsic functional chirality is a property of eukaryotic cells that probably confers on them a selective advantage during the course of evolution. This finding suggests that chirality is a central organizing principle in life that confers order at every length scale up to the full cell. This organization on the basis of chirality must have emerged at some point in the past as an evolutionary response that protected the growth of life.

Acknowledgments

A number of people have contributed for the realization of the chirality index methodology; Dr. Luca Muccioli, Dr. Roberto Berardi, and Prof. Claudio Zannoni in thedevelopment of chirality method applied to protein structures. Dr. Diego La Mendola and Prof. Enrico Rizzarelli have made possible the combined experimental–computational approaches for studying the pH dependence

309

310


of the N-terminal region of avian prion protein, by using circular dichroism and chirality from MD simulations. I gratefully acknowledge Prof. Claudio Zannoni, Prof. Enrico Rizzarelli, and Prof. Michele Parrinello for valuable suggestions and discussions. A special acknowledgement goes to Dr. Gareth Tribello, Dr. Meher Prakash Ayalasomayajula, and Dr. Francesco Mauriello for several useful suggestions in writing this manuscript.

References 1. Corrigan, J. (1969) D-amino acids in 2.

3.

4.

5.

6.

7.

8.

9.

10.

animals. Science, 164, 142–149. Fujii, N. (2002) D-amino acids in living higher organisms. Orig. Life Evol. Biosph., 32, 103–127. Nanda, V., Andrianarijaona, A., and Narayanan, C. (2007) The role of protein homochirality in shaping the energy landscape of folding. Protein Sci., 16, 1667–1675. O’Neil, K.T. and Grado, W.F.D. (1990) Thermodynamic scale for the helix-forming tendencies of the commonly occurring amino-acids. Science, 249, 774–778. Maity, H., Lim, W.K., Rumbley, J.N., and Englader, S.W. (2003) Protein hydrogen exchange mechanism: local fluctuations. Protein Sci., 12, 153–160. Ladik, J.J. and Szekeres, Z.Z. (2006) A study on chirality in biomolecules: the effect of the exchange of L amino acids to D ones in Sso7d ribonuclease. J. Mol. Model., 12, 462–467. Smith, L.J., Fiebig, K.M., Schwalbe, H., and Dobson, C.M. (1996) The concept of a random coil – residual structure in peptides and denatured proteins. Fold Des., 1, R95–R106. Dupuis, F., Sadoc, J.F., and Mornon, J.P. (2004) Protein secondary structure assignment through Voronoi tesselation. Proteins, 55, 519–528. Richards, F.M. and Kundrot, C.M. (1988) Identification of structural motifs from protein coordinate data: secondary structure and first level supersecondary structure. Proteins, 3, 71–84. Majumdar, I., Krishna, S.S., and Grishin, N.V. (2005) PALSSE: a program to delineate linear secondary

11.

12. 13.

14.

15.

16.

17.

18.

structural elements from protein structures. BMC Bioinformatics, 6, 202. Ramachandran, G.N., Ramakrishnan, C., and Sasisekharan, V. (1963) Stereochemistry of polypeptide chain configurations. J. Mol. Biol., 7, 95–99. Schlick, T. (2002) Molecular Modeling and Simulation, Springer, New York. Zagrovic, B., Lipfert, J., Sorin, E.J., Millett, I.S., van Gunsteren, W., Doniach, S., and Pande, V.S. (2005) Unusual compactness of a polyproline type II structure. Proc. Natl. Acad. Sci. U.S.A., 102, 11698–11703. Pietropaolo, A., Raiola, L., Muccioli, L., Tiberio, G., Zannoni, C., Fattorusso, R., Isernia, C., La Mendola, D., Pappalardo, G., and Rizzarelli, E. (2007) An NMR and molecular dynamics investigation of the avian prion hexarepeat conformational features in solution. Chem. Phys. Lett., 442, 110–118. Pietropaolo, A., Muccioli, L., Zannoni, C., La Mendola, D., Maccarrone, G., Pappalardo, G., and Rizzarelli, E. (2008) Unveiling the role of histidine and tyrosine residues on the avian prion hexarepeat domain. J. Phys. Chem. B, 112, 5182–5188. Kabsch, W. and Sander, C. (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers, 22, 2577–2637. Mizuguchi, K. and Go, N. (1995) Comparison of spatial arrangements of secondary structural elements in proteins. Protein Eng., 8, 353–362. Frishman, D. and Argos, P. (1995) Knowledge-based protein secondary structure assignment. Proteins, 23, 566–579.

References 19. Meissner, M., Koch, O., Klebe, G.,

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

and Schneider, G. (2008) Prediction of turn types in protein structure by machine-learning classifiers. Proteins, 74, 344–352. Stapley, B.J. and Creamer, T.R. (1999) A survey of left-handed polyproline II helices. Protein Sci., 8, 587–595. Cubellis, M.V., Caillez, F., Blundell, T.L., and Lovell, S.C. (2005) Properties of polyproline II, a secondary structure element implicated in protein-protein interactions. Proteins, 58, 880–892. Martin, J., Letellier, G., Marin, A., Taly, J.F., de Brevern, A.G., and Gibrat, J.F. (2005) Protein secondary structure assignment revisited: a detailed analysis of different assignment methods. BMC Struct. Biol., 5, 17. Hayward, S. and Go, N. (1995) Collective variable description of native protein dynamics. Annu. Rev. Biophys. Biomol. Struct., 46, 223–250. Clementi, C., Vendruscolo, M., Maritan, A., and Domany, E. (1999) Folding Lennard-Jones proteins by a contact potential. Proteins Struct. Funct. Gen., 290, 253–266. Tama, F. and Brooks, C.L.III (2006) Symmetry, form and shape: guiding principles for robustness in macromolecular machines. Annu. Rev. Biophys. Biomol. Struct., 35, 115–133. Laio, A. and Parrinello, M. (2002) Escaping free-energy minima. Proc. Natl. Acad. Sci. U.S.A., 99, 12562–12566. Barducci, A., Bussi, G., and Parrinello, M. (2008) Well-tempered metadynamics: a smoothly converging and tunable free-energy method. Phys. Rev. Lett., 100, 020603-1–4. Branduardi, D., Gervasio, F.L., and Parrinello, M. (2007) From A to B in free energy space. J. Chem. Phys., 126, 10, Article No. 015403. Carter, E.A. (2008) Challenges in modeling materials properties without experimental input. Science, 321, 800–803. Laio, A. and Gervasio, F.L. (2008) Metadynamics: a method to simulate rare events and reconstruct the free energy in biophysics, chemistry and material science. Rep. Progr. Phys., 71, 126601.

31. Pietropaolo, A., Muccioli, L., Berardi,

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

R., and Zannoni, C. (2007) A chirality index for investigating secondary structure and their time evolution.Proteins, 70, 667–677. Solymosi, M., Low, R.J., Grayson, M., and Neal, M.P. (2002) A generalized scaling of a chiral index for molecules. J. Chem. Phys., 116, 9875–9881. Osipov, M.A., Pickup, B.T., and Dunmur, D.A. (1995) A new twist to molecular chirality: intrinsic chirality indices. Mol. Phys., 84, 1193–1206. Berardi, R., Cainelli, G., Galletti, P., Giacomini, D., Gualandi, A., Muccioli, L., and Zannoni, C. (2005) Can the pi-facial selectivity of solvation be predicted by atomistic simulation? J. Am. Chem. Soc., 127, 10699–10706. Pauling, L., Corey, R.B., and Branson, H. (1951) The structure of proteins; two hydrogen-bonded helical configurations of the polypeptide chain. Proc. Natl. Acad. Sci. U.S.A., 37, 205–211. Marti, D.N., Schaller, J., and Llinás, M. (1999) Solution structure and dynamics of the plasminogen kringle 2-AMCHA complex: 3(1)-helix in homologous domains. Biochemistry, 38, 15741–15755. Chou, K.C. (2000) Prediction of tight turns and their types in proteins. Anal. Biochem., 286, 1–16. Adzhubei, A.A. and Stemnberg, M.J.E. (1993) Left-handed polyproline-II helices commonly occur in globularproteins. J. Mol. Biol., 229, 472–493. Branden, C. and Tooze, J. (1999) Introduction to Protein Structure, Garland Publishing, New York. La Mendola, D., Bonomo, R.P., Impellizzeri, G., Maccarrone, G., Rizzarelli, E., Pietropaolo, A., and Zito, V. (2005) Copper(II) complexes with chicken prion repeats: influence of proline and tyrosine residues on the coordination features. J. Biol. Inorg. Chem., 10, 463–475. Feller, S., Ren, R., Hanafusa, H., and Baltimore, D. (1994) SH2 and SH3 domains as molecular adhesives: the interactions of Crk and Abl. Trends Biochem. Sci., 19, 453–458.

311

312

12 A Computational Approach for Investigating Biomolecule Conformations 42. Krittanai, C. and Johnson, W.C. (1997)

43.

44.

45.

46.

47.

Correcting the circular dichroism spectra of peptides for contributions of absorbing side chains. Anal. Biochem., 253, 57–64. Chakrabartty, A., Kortemme, T., Padmanabhan, S., and Baldwin, R.L. (1993) Helix propensities of the amino acids measured in alanine-based peptides without helix-stabilizing side-chain interactions. Biochemistry, 32, 5560–5565. Che, Y. and Marshall, G.R. (2006) Impact of cis-proline analogs on peptide conformation. Biopolymers, 81, 392–406. Rose, G.D., Gierasch, L.M., and Smith, J.A. (1985) Turns in peptides and proteins. Adv. Protein Chem., 37, 1–109. Bazan, J.F., Fletterick, R.J., McKinley, M.P., and Prusiner, S.B. (1987) Predicted secondary structure and membrane topology of the scrapie prion protein.Protein Eng., 1, 125–135. Neukirch, S., Goriely, A., and Hausrath, A.C. (2008) Chirality of

48.

49.

50.

51.

52.

53.

coiled coils: elasticity matters. Phys. Rev. Lett., 100, 038105-1–4. Rubin, N., Perugia, E., Goldschmidt, M., Fridkin, M., and Addadi, L. (2008) Chirality of amyloid suprastructures. J. Am. Chem. Soc., 130, 4602–4607. Caspar, D. and Klug, A. (1962) Physical principles in the construction of regular viruses. Cold Spring Harb. Symp. Quant. Biol., 27, 1–24. Caspar, D. and Klug, A. (1960) The structure of small viruses. Adv. Virus Res., 7, 225–325. Horne, R.W. and Wildy, P. (1961) Symmetry in virus architecture. Virology, 15, 348–373. Buenemann, M. and Lenz, P. (2008) Elastic properties and mechanical stability of chiral and filled viral capsids. Phys. Rev. E, 78, 051924-1–21. Xu, J., Keymeulen, A.V., Wakida, N.M., Carlton, P., and Berns, M.W. (2007) Polarity reveals intrinsic cell chirality. Proc. Natl. Acad. Sci. U.S.A., 104, 9296–9300.

313

13 Collisional Mechanism–Based E-DNA Sensors: A General Platform for Label-Free Electrochemical Detection of Hybridization and DNA Binding Proteins Francesco Ricci

13.1 Introduction

Pathogen detection or diagnosis of genetic diseases based on the sequence-specific identification of DNA has attracted significant attention [1, 2] due to the fact that these methods are usually more accurate and much faster than culture-based methods. Optical [1–4], electronic [5], acoustic [6], and gravimetric [7, 8] techniques have been applied to this goal. Although each of these approaches has its own advantages, the electrochemical detection of DNA hybridization appears promising due to its rapid response time, low cost, and suitability for mass production [9–12]. For this reason, many electrochemical DNA detection schemes have been described to date [11–13], the best of which achieve limits of detection ranging from picomolar to femtomolar [14–18]. Recently, Plaxco, Heeger, and coworkers, concomitantly with other groups, have introduced a number of single-step, label-free electrochemical biosensors, termed E-DNA sensors, that are based on the target binding–induced folding of electrode-bound DNA probes [19–22]. These sensors, in their original format, were designed as the electrochemical equivalent of optical molecular beacons [19–22]. The original E-DNA sensor is in fact comprised of a redox-modified ‘‘stem-loop’’ probe that is immobilized on the surface of a gold electrode via self-assembled monolayer chemistry (Figure 13.1). In the absence of target the stem-loop holds the redox moiety in proximity to the electrode, producing a large Faradaic current. Upon target hybridization, the stem is broken and the redox moiety moves away from the electrode surface. This, in turn, reduces the efficiency with which electrons are transferred to the electrode, thus suppressing the Faradaic current in proportion to the fraction of probe DNAs that have hybridized (Figure 13.1). This approach seemed to be particularly promising for oligonucleotide detection because of its rapidity, reagentless nature, and operational convenience [20, 22] and demonstrated to be a potentially promising method for the detection of both Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life. Edited by Bruno Pignataro Copyright  2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-32541-2

13 Collisional Mechanism–Based E-DNA Sensors MB

Target

MB

AC current (A)

314

−0.5

−0.4

−0.3

−0.2

−0.1

Potential (mV)

Figure 13.1 E-DNA sensors [13–16], in their original format, were designed as the electrochemical equivalent of optical molecular beacons comprised of a redox-modified ‘‘stem-loop’’ probe immobilized on the surface of a gold electrode. In the absence of a target, the stem-loop structure holds the electrochemical tag (here methylene

blue) into close proximity with the electrode surface, thus ensuring efficient redox of the electrochemical label (lower panel). Upon hybridization with the target sequence collisions between the redox tag and electrode are inhibited, decreasing the signaling current. Reproduced with permission from [19].

DNA and RNA sequences (see, e.g., commentary by Palececk et al. [23] and Thorp et al. [24]). E-DNA platform, as its optical counterpart, proved to be highly specific. E-DNA signaling is in fact driven by hybridization, and therefore, the sensor response is sequence-specific, thereby allowing to distinguish between a fully complementary 17-base target and a 50 000-fold excess of genomic DNA [20]. Unlike other similar optical approaches, the E-DNA platform is very selective and was often demonstrated several times to be effective in measuring DNA sequences in real complex matrices such as blood serum, soil extracts, and foodstuff (e.g., [25–27]). Also, an additional feature that differentiates this approach from the optical one, is that the E-DNA sensor is reusable. The redox-modified probe DNA is in fact strongly bonded to the interrogating electrode via a thiol-gold linkage and after hybridization, a brief, low ionic strength wash (30 seconds in room temperature distilled water) is usually sufficient to recover >99% of the original sensor signal even for sensors that have been challenged directly in complex matrices [25, 28, 29].

13.2 E-DNA Signaling Mechanism

Fabrication and interrogation procedures for E-DNA sensors are also straightforward [30] and suitable for in situ analysis with portable instrumentation and disposable sensors.


Even if E-DNA sensors [20–22] have been extensively studied in recent years, key issues in their fabrication were only recently disclosed, shedding light on the possible signal mechanism of the E-DNA platform. We have recently studied the effects of probe surface density, target length, and other aspects of molecular crowding on the signaling properties, selectivity, and response time of the E-DNA sensor trying to elucidate the possible mechanistic principles of this platform [29]. As a test bed for this study, we have employed a signal-off E-DNA sensor directed against 17-bases of the gyrB gene of Salmonella typhimurium [20, 31–33]. As with all E-DNA signal-off sensors, in the absence of the target DNA, the sensor gives a sharp, well-defined ACV peak at ∼–0.25 V (vs Ag/AgCl) consistent with the formal potential of the methylene blue redox moiety used (Figure 13.1). We found that the signal suppression observed at a given target concentration is a biphasic function of probe density (the number of probe DNA strands per square centimeter of electrode). At low probe densities, signal suppression decreases with increasing probe density, presumably because the increased charge density on the electrode repels the target. In contrast, at higher probe densities, the signal suppression increases dramatically with increasing probe density (Figure 13.2). The transition between these two behaviors occurs at the probe density at which the mean probe-to-probe separation approximates the length of the probe-target duplex, suggesting that molecular crowding plays a role in the effect. This result suggested that probably, in contrast with what was previously speculated, in the E-DNA signaling mechanism, even the probe-target duplex can, at certain conditions, efficiently transfer electrons to the electrode. This is particularly apparent at low probe densities, where the signal suppression observed upon hybridization is limited because even hybridized probe DNA collides with the electrode surface and transfers electrons relatively efficiently. At high probe densities, in contrast, the steric bulk of the closely packed DNA probes inhibits these collisions, leading to higher signal suppression despite presumably lower hybridization efficiency. These results provided insights into the mechanism of E-DNA signaling. As said before, the original stem-loop structure of E-DNA sensors was chosen based on the speculation that, as with the optical counterpart, the signal suppression observed upon hybridization was solely due to the increased tunneling distance between the redox moiety and the electrode [20–22]. The results obtained in this work, however, suggest an alternative mechanism for the observed signaling: hybridization induces changes in the rate with which the redox moiety collides with the electrode surface. The formation of the stem induces efficient electron transfer at both low and high probe densities. At low

315

13 Collisional Mechanism–Based E-DNA Sensors

Mean probe-to-probe separation (nm) 50

10

80 70 % Signal suppression

316

60 50 40 30 20 10 1e+ 11

1e+ 12

Surface density (molecules / cm2) Figure 13.2 We have recently demonstrated that signaling of E-DNA sensors is dependent on probe density. We have controlled probe surface density by varying the concentration of probe DNA employed during sensor fabrication. At high probe densities,

the observed signal suppression decreases monotonically with decreasing density. Shown is the dependence observed at 200 nM of a 17-base target DNA is shown. Reprinted with permission from [29]. Copyright (2007) American Chemical Society.

probe densities, however, the signal suppression observed upon hybridization is limited because even hybridized probe DNA can, on occasion, collide with the electrode surface and transfer electrons. At high probe densities, in contrast, the steric bulk of the closely packed probe DNAs precludes these collisions, leading to higher signal suppression despite presumably lower hybridization efficiency [29]. This collisional model of E-DNA signaling is further supported by studies of the sensor performance as a function of AC frequency probe and of target length and bulk [29]. For example, longer and bulkier targets produce greater signal suppression, presumably because they reduce the collision rate of the probe-target duplex (Figure 13.3). Likewise, very little target-induced signaling is observed when E-DNA sensors are probed with low frequency AC potentials (AC voltammetry), suggesting that, on a sufficiently slow time-scale, electron transfer is efficient from both the free and bound probe. As the ACV frequency increases above a threshold (presumably defined by the slower collision rate of the probe-target duplex), transfer efficiency from the bound state is reduced, producing a large, hybridization-linked reduction in the sensing current. Consistent with this argument, the frequencies at which this transition occurs are higher for lower-density sensors, presumably because reduced probe densities leads to higher collision rates for the probe-target duplex [29].


100

% Signal suppression

80 High density 60 Low density 40 Medium density 20

0 0.1

1

10 100 ACV frequency (Hz)

Figure 13.3 E-DNA signal suppression is eliminated at low ACV frequencies. This presumably occurs when the ACV frequency falls below the rate at which the redox moiety of the probe-target duplex collides with the electrode; under these conditions, electron transfer is efficient from both free and bound probes. As the ACV frequency increases, it

1000

10 000

presumably surpasses the collision rate of the bound probe, reducing transfer efficiency and producing a large, hybridization-linked reduction in the sensing current. Shown are the signal suppressions achieved at 200 nM of a 17-base target. Reprinted with permission from [29]. Copyright (2007) American Chemical Society.

We believe our results support the claim made by Anne and Demaille that the intrinsic bending elasticity of DNA controls the dynamics of electron transport in molecular layers comprised of surface-attached, redox-modified DNA [34]. We provided further evidence for this collisional signaling mechanism also demonstrating that careful optimization of probe density and measurement techniques is necessary in order to achieve maximum performance across this broad and increasingly important class of sensors. Of more importance, we have started to elucidate the possible mechanism of E-DNA sensors proposing a ‘‘collisional’’ signaling mechanism that opens the future to a new class of E-DNA sensors, which is not anymore forced to the stem-loop structure of original E-DNA platforms. All of the groups responsible for the initial development of E-DNA sensors, in fact, employed stem-loop DNA probes [20–22, 25], presumably due to the misconception that, by analogy to molecular beacons, a specific conformational (i.e., geometric) change is required in order to support robust signaling. As a confirmation of the hypothesis that a stem-loop conformation is not necessarily needed for E-DNA sensors, we have recently proposed the use of a ssDNA linear probe for E-DNA-based sensing and demonstrated that binding-induced changes in DNA dynamics are sufficient to support E-DNA signaling, thus giving further evidence to the ‘‘collisional’’ theory [35]. Linear probe

317


E-DNA sensors maintain all the positive features of their stem-loop counterparts being label-free, reusable, sequence-specific, and selective enough to employ directly in complex sample matrices such as blood serum, thus rendering them well suited for clinical applications [35]. We have fabricated E-DNA sensors using a 27-base linear probe sequence that, in order to facilitate direct comparison with earlier studies, is directly analogous to the previously characterized stem-loop E-DNA sensor used for density effect experiments [35]. In the absence of a target, the sensor gives rise to a sharp, well-defined AC voltammetry peak consistent with the ∼–0.25 V (vs Ag/AgCl) formal potential of the methylene blue redox moiety employed (Figure 13.4). Upon hybridization to a fully complementary, 17-base target this current is significantly reduced. Like the original stem-loop E-DNA architecture, the linear probe E-DNA sensor is label-free and reusable: a 30 seconds wash in room temperature distilled water or MB

MB

MB

MB

MB

MB MB

MB

eT

MB

eT

eT

Hybridization

eT MB

(a) Regeneration AC current (1e-7 A)

318

Buffer

6 Regeneration 4

Serum/buffer 50/50

Target

Target 2 −0.4

(b)

−0.3 −0.2 Potential (V)

−0.1

Figure 13.4 E-DNA signaling arises due to hybridization-induced changes in probe ‘‘collision’’ dynamics (rather than to a conformational change per se) and thus redox-modified linear probe DNAs serve as effective E-DNA sensors (a) [35]. The Faradaic current arising from such a linear probe DNA is significantly reduced in the presence of a complementary target

−0.4 (c)

−0.3 −0.2 Potential (V)

−0.1

sequence because, as demonstrated here, hybridization reduces the rate with which the terminal redox tag collides with the electrode surface and transfers electrons (b). Linear probe E-DNA sensors respond well in either (b) buffer or (c) 50% blood serum. Reproduced with permission of the Royal Society of Chemistry from [35].

13.3 E-DNA Sensor for DNA Binding Proteins Detection

(after deployment in blood serum) room temperature detergent solution is enough to regenerate >97% original sensor current (Figure 13.4). Of note, such linear probe sensors exhibit somewhat improved signal gain over their stem-loop counterparts presumably because target binding no longer competes with breakage of the stem structure. Despite this, however, while the signal suppression observed in serum is quite similar to that obtained in buffer the currents observed in serum are lower. This may be due to the greater viscosity and/or the reduced ionic strength of this medium, which would reduce collision rates and thus reduce electron transfer efficiency; no similar drop-off is observed for stem-loop sensors, presumably because the stem structure fixes the (MB) methylene blue near the surface, rendering its electron transfer rate relatively independent of these effects [35]. The disclosure that E-DNA signaling requires only that a target binds to an oligonucleotide probe and, in doing so, changes the efficiency with which the attached redox tag strikes the electrode could account for the approach’s generalizability, thus suggesting that the same sensing principle would also support the detection of DNA binding events that lead to the formation of bulky and/or rigid complexes. In order to test this suggestion we have recently proposed the use of E-DNA-like sensors for the detection of DNA binding proteins. These proteins are abundant and essential in cells, interacting with DNA in order to organize their packing, regulate transcription, and perform replication and repair [36, 37].

13.3 E-DNA Sensor for DNA Binding Proteins Detection

For this study we have selected two double-strand binding proteins, the eukaryotic TATA-box binding protein (TBP) (a core component of the eukaryotic transcriptional machinery), and the prokaryotic M.HhaI methyltransferase (M.HhaI) (involved in the restriction–modification system of bacteria); and two single-strand binding proteins (SSBPs) involved in the replication machinery, the prokaryotic SSBP and the eukaryotic replication protein A (RPA) [38]. We have fabricated sensors against the double-strand binding proteins using short, stem-loop probe DNAs in which the relevant recognition sequences are contained within the double-stranded stem. These probes were modified with a 3 thiol group, supporting strong chemisorption to an interrogating electrode, and a methylene blue redox tag pendant on a thymine base along the double-stranded stem (for TBP, Figure 13.5a and c) or within the single-stranded loop (for M.HhaI, Figure 13.5b and d). In the absence of a target, both probes produce a large faradaic peak at the potential expected for the methylene blue redox tag (Figure 13.5c and d). This current is due to the collisions of the relatively dynamic dsDNA probes. The relative dynamicity of the probe is ensured by the fact that this was designed to have a short (five bases) single-stranded element at the proximity of the gold electrode, thus ensuring enough freedom for the collision of the probe.

319

320


TBP

M. Hhal

MB

MB

MB MB

MB MB

MB

MB

MB

MB MB

MB

MB MB

eT

eT

M. Hhal TBP

(a)

(b) Buffer

−0.5 (c)

−0.3 V vs. Ag/AgCl, 3M NaCl

−0.1

10 nA

+ TBP

Regeneration 10 nA

Regeneration

−0.5 (d)

Figure 13.5 Schematics of E-DNA-like sensors for the detection of DNA binding proteins. (a, b) The sensor is comprised of a DNA hairpin covalently attached to a classic gold rod electrode using thiol-gold self-assembled monolayer chemistry and containing an internal methylene blue redox tag. (c, d) In the absence of target relatively

Buffer

+ M. Hhal

−0.3 V vs. Ag /AgCl, 3 M NaCl

−0.1

efficient collision between the label and the electrode produces a large faradic current. Upon target binding this faradic current is significantly reduced, presumably because the bulky structure of the protein reduces the collision rate. (Reprinted with permission from [38]. Copyright (2009) American Chemical Society.

In the presence of saturating TBP and M.HhaI, the binding of these proteins leads to a bulk and rigid DNA–protein complex, which in turns reduces the currents by 45 and 55% respectively. The two sensors support the ready detection of their target proteins at concentrations as low as 2 and 25 nM for TBP and M.HhaI respectively. Again, exactly as in the case of E-DNA sensors for hybridization detection, because all of the sensing components are strongly adsorbed to the electrode surface, this sensing architecture is readily regenerable; a short wash (30 seconds) in 8 M guanidine chloride is sufficient to regenerate 98% of the original signaling current of both sensors (Figure 13.5c and d), allowing multiple cycles of detection and regeneration (data not shown). Of note is that, it is important to stress that not only dsDNA but also ssDNA probes support this analytical approach, enabling the sensitive, convenient detection of proteins that bind to such targets. Using single-stranded, poly-thymine probes (neither of our target proteins exhibits any significant sequence specificity) we have fabricated sensors for the detection of the SSBP (Figure 13.6) and RPA. Of note, because the single-stranded DNA is thought to wrap entirely around SSBP, the gain of these sensors is dependent on the length of the single-stranded probe employed: when targeting this protein the observed signal suppression (at saturating target concentration) increases from 20 to 70% as the probe is lengthened from 20 to 70

13.4 Conclusions and Future Perspectives

321

MB

MB MB

MB MB

eT

eT

MB

SSBP

eT

(a) 40-base probe Buffer 0.2 µA

Regeneration

1 µA

+ SSBP

−0.5

−0.3

(b)

V vs. Ag/AgCl, 3 M NaCl

−0.1

70-base probe Buffer

Regeneration 0.05 µA

20-base probe Buffer

+ SSBP

−0.5


Figure 13.6 The sensing mechanism also holds for single-stranded probes, thus enabling the detection of Escherichia. coli single-strand binding protein (SSBP). (a) The faradaic current arising from such a single-stranded probe is significantly reduced upon binding with this target. (b) Of note,

−0.1

−0.5

Regeneration

+ SSBP


because the single-stranded probe wraps around the protein target, different length probes (20, 40, and 70 bases) giving rise to different responses upon protein binding. Shown are responses to 80 nM of SSBP. Reprinted with permission from [38]. Copyright (2009) American Chemical Society.

bases (Figure 13.6b). The same change in probe length is associated with a 5–70% increase in gain in the presence of RPA, an effect that could arise when multiple RPAs bind to the longer probe [39]. All three sensors are as selective as the DNA probes from which they are fabricated. For example, we do not detect any significant cross-reactivity between the two sensors directed against double-strand binding proteins (Figure 13.7a). (Neither of the two SSBPs investigated here exhibits any significant sequence specificity and thus, while the signal gain produced by the two proteins is not identical, the ability of our sensor to discriminate between these two targets is limited.) Likewise, because their signaling is linked to a binding-specific change in the probe DNA (and not simply to adsorption of target to the sensor surface), our sensors are effective in rejecting false positives arising due to the nonspecific adsorption of interferents and can be employed directly in complex samples. For example, our single-strand sensor supports the selective detection of exogenous levels of RPA directly in crude Raji cell nuclear extracts (Figure 13.7b). 13.4 Conclusions and Future Perspectives

We believe the studies we have summarized in this chapter represent a strong basis for the hypothesis that signaling in our sensors is dominated by binding-induced

−0.1

0 (a)

TBP

M. Hhal

Figure 13.7 DNA binding protein sensors are as selective as the DNA probes from which they are fabricated. (a) For example, no significant cross-reactivity is observed between the two sensors directed against double-strand binding proteins in the presence of 1.5 µM BSA and saturating concentrations of the nontargeted double-strand binding proteins (80 nM M.HhaI or 10 nM TBP). (b) Likewise our sensors are also

Raji 225 µg/ml

Raji 125 µg/ml

Comp. seq. 10 µM

10

Raji 75 µg/ml

20

Raji 25 µg/ml


SAH

TBP

BSA

Buffer

TBP M. Hhal

25

BSA

50

Buffer


75

Raji 175 µg/ml


M. Hhal

322

0

(b) effective in rejecting false positives arising due to interferents and perform well when challenged with realistically complex sample matrices. For example, we can detect the exogenous level of RPA in crude Raji nuclear extracts and partially restore initial signal by adding high concentration of a competitor sequence (polyT-70). Reprinted with permission from [38]. Copyright (2009) American Chemical Society.

changes in the efficiency with which the redox tag collides, and thus transfers electrons to the interrogating electrode. This is of primary importance and we have demonstrated that this opens the possibility to the use of DNA probe with no stringency of stem-loop conformation. As a further advancement, the collisional mechanism makes possible the monitoring of theoretically any DNA–protein or DNA–antibody interactions, which leads to the formation of bulky and/or rigid complexes. We have confirmed this suggestive hypothesis with a three-successes-out-of-three-attempts success rate in fabricating sensors targeting both sequence-specific, double-strand binding proteins and nonspecific, SSBPs. The E-DNA sensing platform has then demonstrated to be not only a promising and appealing approach for the sequence-specific detection of DNA and RNA, but also to be flexible enough to be adaptable for the detection of DNA–protein interactions. For future practical applications of this new sensing platform, however, some drawbacks have to be considered and overcome. For many applications, detection of specific DNA sequences is still not sensitive enough and will require the coupling of PCR amplification. Moreover, miniaturization of the sensing platform, although possible, has not yet been demonstrated. Recently, we have however shown the use of screen printed electrodes with E-DNA sensors, which represent a first step toward the automation and mass production of these sensors [40]. Long-term and operational stability of the sensors is another issue, which also deserves a deeper study before this platform could be applied in the real world.

13.4 Conclusions and Future Perspectives

Generalizability and flexibility of E-DNA are among the most appealing advantages. For example, the same sensing platform could be expanded to a wide range of other targets not only limited to those that bind to unmodified DNA or RNA. The collisional mechanism suggests, in fact, also a ready means of detecting proteins/enzymes that bind small molecules by simply appending the small molecule to the DNA probe, thus expanding widely, the possible E-DNA applications. We have recently demonstrated this interesting suggestion by appending a small molecule recognition element onto a relatively rigid, partially double-stranded DNA scaffold that is chemi-adsorbed to an interrogating gold electrode [41]. One of the two scaffold strands, the anchoring strand, is linked to the electrode using self-assembled monolayer chemistry via a 3 thiol group and is modified with a redox tag (methylene blue) at its 5 terminus. The second strand, the recognition strand, is complementary to a region of the anchoring strand and is covalently modified with the relevant small molecule recognition element at one of its termini. As a preliminary test for the development of this technology we have employed the small molecules biotin and digoxigenin as recognition elements to fabricate sensors directed against streptavidin and antidigoxigenein antibodies, respectively. In the absence of a target, the modified, double-stranded scaffold is free to collide with the electrode surface, and thus produces a large faradaic current at the potential expected for methylene blue. In preliminary experiments, we observed that upon target binding, this current is reduced, signaling the presence of the target. This reduction presumably occurs because target binding reduces the efficiency with which the redox tag collides with the electrode, either due to the increase in steric bulk or due to the hydrodynamic radius associated with the target. In the case of multivalent targets (such as antibodies), another cause may be that the two antigen binding sites on the antibody cross-link two scaffolds. This new sensing approach offers several significant advantages over other methods for monitoring protein–small molecule interactions or detecting proteins that bind to specific small molecule targets. For example, standard methods for protein detection, such as ELISAs and western blots, are time and capital intensive, and typically require the addition of reagents and multiple washing steps [42, 43]. Less cumbersome methods for monitoring protein–small molecule interactions, such as fluorescence polarization, and plasmon resonance, surface plasmon resonance (SPR) and quartz crystal microbalance (QCM) approaches, often fail in complex samples because of background fluorescence [44] or nonspecific adsorption [45, 46]. In contrast, the approach we have demonstrated here is rapid and convenient, and functions even when challenged directly in clinically and environmentally relevant samples. Moreover, it appears that this new approach is general; in principle, given a target macromolecule large enough to alter the collision dynamics of the scaffold, the only limiting factor is the ability to effectively conjugate the relevant small recognition molecule to a relatively rigid and dynamic molecular scaffold [41].

323

324


Acknowledgments

The work described here has been performed at the Prof. Plaxco’s lab of the University of California, Santa Barbara. Grateful acknowledgment for his scientific and economic support is given to Prof. Kevin W. Plaxco for hosting me in two different periods and for making possible this work. All the members of his research group, for helpful discussions and comments on the manuscript, are also thanked. I am also grateful to my supervisor, Prof. Giuseppe Palleschi of the University of Rome, Tor Vergata for giving me the possibility of this work.

References 1. Cao, Y.W.C., Jin, R.C., and Mirkin, C.A.

2.

3.

4.

5.

6.

7.

8.

(2002) Nanoparticles with Raman spectroscopic fingerprints for DNA and RNA detection. Science, 297, 1536–1540. Gaylord, B.S., Heeger, A.J., and Bazan, G.C. (2002) DNA detection using water-soluble conjugated polymers and peptide nucleic acid probes. Proc. Natl. Acad. Sci. U.S.A., 99, 10954–10957. Peterson, A.W., Wolf, L.K., and Georgiadis, R.M. (2002) Hybridization of mismatched or partially matched DNA at surfaces. J. Am. Chem. Soc., 124, 14601–14607. Ahn, S. and Walt, D.R. (2005) Detection of Salmonella spp. using microsphere-based, fiber-optic DNA array. Anal. Chem., 77 (15), 5041–5047. Kelley, S.O., Barton, J.K., Jackson, N.M., and Hill, M.G. (1997) Preparation of electrode-immobilized, redox-modified oligonucleotides for electrochemical DNA and aptamer-based sensing. Bioconjug. Chem., 8, 31–37. Cooper, M.A., Dultsev, F.N., Minson, T., Ostanin, V.P., Abell, C., and Klenerman, D. (2001) Direct and sensitive detection of a human virus by rupture event scanning. Nat. Biotechnol., 19, 833–837. Endo, T., Kerman, K., Nagatani, N., Takamura, Y., and Tamiya, E. (2005) Label-free detection of peptide nucleic acid-DNA hybridization using localized surface plasmon resonance based optical biosensor. Anal. Chem., 77 (21), 6976–6984. Patolsky, F., Lichtenstein, A., and Willner, I. (2001) Preparation of

9.

10.

11.

12.

13.

14.

15.

electrode-immobilized, redox-modified oligonucleotides for electrochemical DNA and aptamer-based sensing. J. Am. Chem. Soc., 123, 5194–5205. Park, S.J., Taton, T.A., and Mirkin, C.A. (2002) Array-based electrical detection of DNA with nanoparticle probes. Science, 295, 1503–1506. Boon, E.M., Ceres, D.M., Drummond, T.G., Hill, M.G., and Barton, J.K. (2000) Mutation detection by electrocatalysis at DNA-modified electrode. Nat. Biotechnol., 18, 1096–1100. Drummond, T.G., Hill, M.G., and Barton, J.K. (2003) Electrochemical DNA sensors. Nat. Biotechnol., 21, 1192–1199. Thorp, H.H. (1998) Cutting out the middleman: DNA biosensors based on electrochemical oxidation. Trends Biotechnol., 16, 117–121. Katz, E. and Willner, I. (2003) Probing biomolecular interactions at conductive and semiconductive surfaces by impedance spectroscopy: routes to impedimetric immunosensors, DNA-sensors, and enzyme biosensors. Electroanalysis, 15, 913–947. Gore, M.R., Szalai, V.A., Ropp, P.A., Yang, I.V., Silverman, J.S., and Thorp, H.H. (2003) Detection of attomole quantities of DNA targets on gold microelectrodes by electrocatalytic nucleobase oxidation. Anal. Chem., 75, 6586–6592. Floch, L.F., Ho, H.A., Harding-Lepage, P., Bedard, M., Neagu-Plesu, R., and Leclerc, M. (2005)

References

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Ferrocene-functionalized cationic polythiophene for the label-free electrochemical detection of DNA. Adv. Mater., 17, 1251–1254. Gibbs, J.M., Park, S.J., Anderson, D.R., Watson, K.J., Mirkin, C.A., and Nguyen, S.T. (2005) Polymer-DNA hybrids as electrochemical probes for the detection of DNA. J. Am. Chem. Soc., 127, 1170–1178. Kim, E., Kim, K., Yang, H., Kim, Y.T., and Kwak, J. (2003) Enzyme-amplified electrochemical detection of DNA using electrocatalysis of ferrocenyl-tethered dendrimer. Anal. Chem., 75, 5665–5672. Kerman, K., Matsubara, Y., Morita, Y., Takamura, Y., and Tamiya, E. (2004) Recent trends in electrochemical DNA biosensor technology. Sci. Technol. Adv. Mater., 5, 351–357. Ricci, F. and Plaxco, K.W. (2008) E-DNA sensors for convenient, label-free electrochemical detection of hybridization. Microchim. Acta, 163, 149–155. Fan, C., Plaxco, K.W., and Heeger, A.J. (2003) Electrochemical interrogation of conformational changes as a reagentless method for the sequence-specific detection of DNA. Proc. Natl. Acad. Sci. U.S.A., 100, 9134–9137. Immoos, C.E., Lee, S.J., and Grinstaff, M.W. (2004) Conformationally gated electrochemical gene detection. Chembiochem, 5, 1100 . Mao, T., Luo, C., and Ouyang, Q. (2003) Studies of temperature-dependent electronic transduction on DNA hairpin loop sensor. Nucleic Acids Res., 31, 108 . Palecek, E. (2004) Surface-attached molecular beacons light the way for nucleotide sequencing. Trends Biotechnol., 22, 55–58. Thorp, H.H. (2003) Reagentless detection of DNA sequences on chemically modified electrodes. Trends Biotechnol., 21, 522–524. Lubin, A.A., Lai, R.Y., Baker, B.R., Heeger, A.J., and Plaxco, K.W. (2006) Sequence-specific, electronic detection of oligonucleotides in blood, soil, and foodstuffs with the reagentless, reusable E-DNA sensor. Anal. Chem., 78 (16), 5671–5677.

26. Lubin, A.A., Fan, C., Schafer, M.,

27.

28.

29.

30.

31.

32.

33.

34.

Clelland, C.T., Bancroft, C., Heeger, A.J., and Plaxco, K.W. (2008) Rapid, electronic detection of DNA and non-natural DNA analogs for molecular marking applications. Forensic Sci. Commun., 10 (1), 1–6. Lai, R.Y., Seferos, D.S., Heeger, A.J., Bazan, G.C., and Plaxco, K.W. (2006) Comparison of the signaling and stability of electrochemical DNA sensors fabricated from 6-or 11-carbon self-assembled monolayers. Langmuir, 22, 10796–10800. Lai, R.Y., Lagally, E.T., Lee, S.-H., Soh, H.T., Plaxco, K.W., and Heeger, A.J. (2006) Rapid, sequence-specific detection of unpurified PCR amplicons via a reusable, electrochemical sensor. Proc. Natl. Acad. Sci. U.S.A., 103, 4017–4021. Ricci, F., Lai, R.Y., Heeger, A.J., Plaxco, K.W., and Sumner, J.J. (2007) Effect of molecular crowding on the response of an electrochemical DNA sensor. Langmuir, 23, 6827–6834. Xiao, Y., Lai, R., and Plaxco, K.W. (2007) Preparation of electrode-immobilized, redox-modified oligonucleotides for electrochemical DNA and aptamer-based sensing. Nat. Protoc., 2, 2875–2880. Lubin, A.A., Vander Stoep Hunt, B., and Plaxco, K.W. (2008) The effects of probe length, probe geometry and redox-tag placement on the performance of the electrochemical E-DNA sensor. Anal. Chem., 81 (6), 2150–2158. Lai, R.Y., Lee, S.-H., Soh, H.T., Plaxco, K.W., and Heeger, A.J. (2006) Differential labeling of closely-spaced biosensor electrodes via oxidative desorption. Langmuir, 22, 1932–1936. Pavlovic, E., Lai, R.Y., Wu, T.T., Ferguson, B.S., Sun, R., Plaxco, K.W., and Soh, H.T. (2007) Specific, electromechanical detection of multiple DNA sequences in an integrated microfluidic system. Langmuir, 24 (3), 1102–1107. Anne, A. and Demaille, C. (2006) Dynamics of electron transport by elastic bending of short DNA duplexes. Experimental study and quantitative modeling of the cyclic voltammetric behavior of

325

326


35.

36.

37.

38.

39.

40.

30-ferrocenyl DNA endgrafted on gold. J. Am. Chem. Soc., 128, 542–557. Ricci, F., Lai, R.Y., and Plaxco, K.W. (2007) Linear, redox modified DNA probes as electrochemical DNA sensors. Chem. Commun., 3768–3770. Kadonaga, J.T. (2004) Regulation of RNA polymerase II transcription by sequence-specific DNA-binding factors. Cell, 116 (2), 247–257. Bulyk, M.L. (2006) DNA microarray technologies for measuring protein-DNA interactions. Curr. Opin. Biotechnol., 17 (4), 422–430. Ricci, F., Bonham, A.J., Mason, A.C., Reich, N.O., and Plaxco, K.W. (2009) A reagentless, electrochemical approach for the specific detection of double- and single-stranded DNA binding proteins. Anal. Chem., 81 (4), 1608–1614. Kim, C., Paulus, B.F., and Wold, M.S. (1994) Interactions of human replication protein A with oligonucleotides. Biochemistry, 33, 14197–14206. Ricci, F., Zari, N., Caprio, F., Recine, S., Amine, A., Moscone, D., Palleschi, G.,

41.

42.

43. 44.

45.

46.

and Plaxco, K.W. (2009) Surface chemistry effects on the performance of an electrochemical DNA sensor. Bioelectrochemistry, 76 (1-2), 208–213. Cash, K., Ricci, F., and Plaxco, K.W. (2009) An electrochemical sensor for the detection of protein-small molecule interactions directly in serum and other complex matrices. J. Am. Chem. Soc., 131 (20), 6955–6957. Crowther, J.R. (2001)in The ELISA Guidebook, vol. 149 (ed. J.M.Walker), Humana, Totawa . Wild, D.(ed.) (2001) The Immunoassay Handbook, 2nd edn, Nature, London . Owicki, J.C. (2000) Fluorescence polarization and anisotropy in high throughput screening: perspectives and primer. J. Biomol. Screen., 5, 297–306. Homola, J. (2003) Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem., 377, 528–539. Cooper, M.A. (2003) Label-free screening of bio-molecular interactions. Anal. Bioanal. Chem., 377, 834–842.

327

Index

a acarbose 240 activated esters 62 adenosylcobalamin 97 – dependent reactions 97–99 5 -adenyl-β, γ -amidotriphosphate (AMP-PNP) 172 aggregation mechanism 9 alanine 69 aldehydes – oxime ligation 78–79 – site-specific modification 79 alkylation 65–66 alkyne – Cu-catalyzed [3 + 2] cycloaddition of 74 – modification 79–80 allethrin 197 Alzheimer’s disease, ubiquitin-positive amyloid plaques 14 aminoethylation 66 aminoglycosides 240 aminoimidazole-4-carboxamide ribonucleotide (AICAR) 222–224 – cyclic peptide inhibitors of 222–224 5-amino-4-imidazolecarboxamide ribonucleotide transformylase/inosine 5 -monophosphate cyclohydrolase (ATIC) 222–223 amphotericin B 197 amyloid 13 amyloidogenic proteins, toxicity of 13 anisotropy 204 anti-infectives, structure-based design to 167–187 – 4-diphosphocytidyl-2C-methyl-d-erythritol kinase (IspE) 170–182 – isoprenoids 169–174

– nonmevalonate pathway 169–174 – X-ray cocrystal structure analysis 182–184 aquocobalamin 103 aromatic hydroxylation 50, 278–280 aromatic rings, hydroxylation of 278–285 – intramolecular aromatic hydroxylation 278–280 – phenolic compounds, intermolecular ortho-hydroxylation of 280–285 ascorbate oxidase 40 asthma 228–229 atomic force microscopy (AFM) 18 atox1 7 autophagy-lysosome pathway 12 avian proteins 305 β-axial position 102–108 – cobalamin alkylation 102–103 – heterodinuclear concept 103–108 azide – Cu-catalyzed [3 + 2] cycloaddition of 74 – modification 79–80 azithromycin 197

b basic helix-loop-helix leucine zipper (bHLHZip) 216 bioconjugation 133–135 – streptavidin–biotin 134–135 biodegradable microspheres 119–120 – preparation 119–120 biological membranes 191–195 – composition of 191–193 – dynamic molecular organization 193–195 – life maintenance, role in 191 – structure of 191–193 biomimetic Cu complexes 272–285


328

Index biostable microspheres 120–121 – applications 120 – preparation 120–121 – – dispersion polymerization 120 – – emulsion polymerization 120–121 biosynthesized libraries, high-throughput screening of 220–226 biotin 134–135 blue copper proteins 33, 268 brain copper 8–9 – homeostasis 5–8 brain, metal ions in 4 Brewster angle microscopy (BAM) 201, 205

c calculated partitioning coefficient (clogP) 175, 177–178, 181 carbohydrate microarrays 257 carbohydrate synthesis 240–245 – chemical synthesis 241–243 – – automated synthesis 241 – – one-pot synthesis 241–242 – – solid-phase synthesis 242–243 – enzymatic synthesis 243–244 – – glycosidases 243 – – glycosyltransferases 243–244 – – glycosynthases 243 – glycoprotein synthesis 244–245 – – cell biological technique 245 – – enzyme remodeling 245 – – molecular biological technique 245 – – peptide ligation 244–245 carbohydrate-based drugs 240 carbon–carbon bond formation, palladium-catalyzed 76 catalytic systems 45–47 catechol 270 cecropin 197 cell-penetrating peptides 118 cell selection 130 cellular copper, redox state of 21–22 ceruloplasmin 40 chemical libraries 216–220 chemical protein modification 59–82 – challenges in 61 – posttranslational modification 59–60 – route to 60–61 – site-selective 70–81 – – dehydroalanine 71 – – dual modification 81 – – metal-mediated protein modification 71–78 – – unnatural amino acids, metal-free methods for modifying 78–81

– traditional methods for 61–70 – – aspartic acid modification 64 – – cysteine 64–70 – – glutamic acid modification 64 – – lysine modification 62–64 chemoenzymatic synthesis 251–256 chirality 293–310 – intrinsic 297 – living systems, in 293–295 – perspectives 308–309 – protein dynamics, computational techniques for 297–298 – protein motions, analysis of 298–308 – – circular dichroism 305–308 – – protein dynamics, monitoring of 302–305 – – protein structure, understanding of 300–302 – protein secondary structures 295–296 – – assignment 296–297 – – coil 296 – – α helix 296 – – poly-l-proline II 296 – – β sheets 296 – – β turns 296 chirality index 298–300 – protein dynamics, tool for monitoring 302–305 circular dichroism 305–308 cisplatin 103, 197 cleavable microspheres 132 clioquinol 8 coarse-grained model 298 cobalamin – alkylation 102–103 – enzymatic reactions, dependent 98 – gastrointestinal pathway 96 – metabolism of 98 coiled coils 308 combinatorial chemistry 167, 216 Concanavalin A (ConA) 258 conformational diseases 13 conformational movement 195 copper(III) – bioinorganic implications 33–41 – – copper enzymes 33–36 – – mononuclear monooxygenating copper-based enzymes 38–40 – – particulate methano monooxygenase (pMMO) 36–38 – – trinuclear copper models for laccase 40–41 – cuprate superconducting materials 52 – organometallic CuIII species 41–51

Index – – aliphatic C–H bond functionalizations 45–51 – – aromatic C–H bond functionalizations 45–51 – – aryl–heteroatom bond formation 44–45 – – C–C bond formation 42–44 – redox potentials 32 – stabilization of 31 copper chaperone for superoxide dismutase (CCS) 7 copper dioxygen adducts 272–275 copper ion – CuII binding, spectroscopic characterization of 15–17 – CuII -induced self-oligomerization 17–18 – oxidation states 5 copper transport (Ctr) protein 6 covalent protein modification 61 cross-ring cleavage 150–154 cuprate superconducting materials 52 cyanocobalamin (vitamin B12) 95 cyclic peptide inhibitors 224–226 cyclooxygenases (COXs) 201 cyclosporine 197 cysteine 59 – modification 64–70 – – alkylation 65–66 – – desulfurization 67–70 – – disulfides 66–67 – oxidative elimination of 72 cytidine 5 -diphosphate (CDP) binding pocket 170 – targeting of 174–182

d DEA, See dissociative electron attachment defensin 197 dehydroalanine 71 dehydrons 20 dendrimers 118 density functional theory (DFT) 147, 279 dichlorodiphenyl trichloroethane (DDT) 197 dictionary of protein secondary structures (DSSP) 296 dicyclohexylcarbodiimide (DCC) 124 dicyclohexylurea 124 diisopropylcarbodiimide (DIC) 124 dimethylbenzimidazole 102 dinuclear type-3 copper enzymes 33–36 dinucleating ligands 276 4-diphosphocytidyl-2C-methyl-d-erythritol Kinase (IspE) 170 – active site 170–174

– CDP-binding pocket, targeting 174–182 – – design 174–176 – – hydrophobic pocket, optimization of filling of 179–181 – – possible ribose analogues 175 – – ribose analogue, optimization of 176–178 – – sulfone substituents, evaluation of inhibitors featuring 180–181 – – vector design 176 – – vector, importance of 178–179 – structure 170 dispersion polymerization 120 dissociative electron attachment (DEA) 143 – electron attachment to d-ribose 148–150 – energetic position of 146 – gas-phase dna building blocks 147–148 – LIAD, use of 159 – sugar–phosphate cleavage 159–161 – tetraacetyl-d-ribose 155–159 – transient negative anions, nature of the 154–155 disuccinimidyl glutarate (DSG) 62 disuccinimidyl suberate (DSS) 62 disulfide bonds 132–133 disulfide cleavage 133 disulfides 66–67 – contraction to thioether 67 – native chemical ligation 67–69 dithiothreitol (DTT) 71, 133 DNA, radiation damage 143–162 DNA backbone – model compounds – – DEA studies on 148–161 DNA binding proteins detection 319–321 dopamine β-monooxygenase (DβM) 38 double emulsion technique 120 double strand breaks (DSBs) 144 d-ribose 148–150 – cross-ring cleavage 150–154 drug–membrane interactions 191–206 – analysis and quantification 199–201 – biological membranes 191–195 – clinical relevance of studies 195–199 – – drug development, contribution for 195–197 – – enzymatic inhibition, controlling 198–199 – – multidrug resistance (MDR) 198 – – therapeutic and toxic effect of drugs 197 – experimental techniques 200–201 – future research directions 206 – membrane model systems 199–200

329

330

Index drug–membrane interactions (contd.) – nonsteroidal anti-inflammatory drugs (NSAIDs), study of 201–206 – – drug fundamental physical–chemical studies 202–203 – – membrane dynamic studies 203–205 – – membrane structural studies 203–205 – – results 205–206 – possible effects 195–199 dual functionality 124 dual functionalized microspheres 123 dual modification 81

g

f

h

first-generation inhibitors 182 – water solubility of 182 fluid mosaic model 191 fluorescence quenching 203 fmoc chemistry 121–122 – basic steps 123 fosmidomycin 169 Franck–condon transition 158 freeze-fracture electron microscopy (FFEM) 201 fumonisin B1 198

H5N1 avian influenza 230 haloarenes, cross-coupling of 45 halothane 197 haptocorrin 96 h-bonds 20 α helix 296 hemagglutinin 230 hemocyanin, 33–34 269–270 – site of 269 hemoglobin 304 – helix of 304

gain-of-function mechanism 9 gene silencing 128 – microsphere-mediated 129 giant unilamellar vesicles (GUVs) 199 global aliphatic hydroxylation 39 glucansucrases 251–256 glutathione 132 glycans 240 glyceraldehyde 3-phosphate 168 glycolipids 192 glycomics 239–259 – bioinformatics 247 – cell, tissue, and metabolic labeling e 246–247 E-DNA sensors, collisional mechanism–based – future perspectives 259 313–324 – DNA binding proteins detection 319–321 – mass spectrometry 245–246 – microarrays 246 – features 313–314 – new insights 247–259 – future perspectives 321–323 – – acceptor specificities of glycosyltransferase – signaling mechanism 315–319 R, identification of 257–259 efflux pump effect 108 – – microwave-assisted glycosylation electron crystallography 6 247–251 electron paramagnetic resonance (EPR) 15, – – novel branched thiooligosaccharides, 269 chemoenzymatic synthesis of electrospray ionization mass spectrometry 251–256 (ESI-MS) 15 – state of the art 239–247 emulsion polymerization 120–121 – – carbohydrate-based drugs 239–240 enantiomers 294 – – carbohydrate synthesis 240–245 endoglycosidases 243 glycopeptides enflurane 197 – synthesis enhanced green fluorescent protein – – microwave-assisted glycosylation (EGFP) 128 247–251 enzymatic synthesis 243–244 glycoprotein synthesis 244–245 – glycosidases 243 glycosyl phosphatidylinositol 68 – glycosyltransferases 243–244 glycosylamino acids, synthesis 248 – glycosynthases 243 glycosylation 59 epimerization 67 glycosyltransferases 243 erythropoietin 240 grazing-incidence X-ray diffraction ester bonds 130–132 (GIXD) 204 exoglycosidases 243 green fluorescent protein (GFP) 128 expressed protein ligation (EPL) 244

Index heparin 240 heteronuclear single quantum coherence (HSQC) spectra 16 hexamethylphosphorus triamide (HMPTL) 67 HIF-1 226 HIF-1α 226–227 high-throughput screening (HTS) 167 HIV – lifecycle of 219 – structure of 220 HIV budding, cyclic peptide inhibitors of 224–226 homochirality 293 hot spot 215 Huisgen cycloaddition 75 human cervical cancer (HeLa) 128 human protein double minute 2 (HDM2) 217 human serum albumin (HSA) 103 Huntington’s disease 13 hydrophobic effect 193 hydroxylation 59

i iminodiacetic acid (IDA) 17 2-imino-2-methoxyethyl reagents (IME) 63 immunoglobulin antigen 304 influenza virus 230–232 inosine monophosphate (IMP) 222 interleukin-3 (IL-3) 228 – signaling pathways in asthma 229 interleukin-4 (IL-4) 228 – signaling pathways in asthma 229 intracellular sensing 126–127 intrinsic factor 95–97 iodoacetamides 65 isocyanates 62 isoprene rule 169 isoprenoids 168–169 isothiocyanates 62

j jun 217 jun activation domain-binding protein-1 (JAB1) 10 juxtanuclear compartment 12

k ketones – oxime ligation 78–79 – site-specific modification

79

l label-free electrochemical biosensors 313 laccase, trinuclear copper models for 40–41 lamellar phase 193 large unilamellar vesicles (LUVs) 199 laser-induced acoustic desorption (LIAD) 159 latex beads 119 Lawesson’s reagent 66 lectins 246 Leloir glycosyltransferase 251 – substrate 257 LIAD, See laser-induced acoustic desorption (LIAD) lindane 197 lipoplexes 192 liposomes 117–118, 199 liquid-crystalline phase 194 loss-of-function mechanism 9 low-molecular-weight heparins (LMWHs) 240 lysine 59 – iridium-catalyzed reductive alkylation 74 – modification 62–64 – – activated esters 62 – – IME reagents 63–64 – – isocyanates 62 – – isothiocyanates 62 – – reductive alkylation 62–63 – – strategies 63 lysosomes 10

m M.HhaI methyltransferase (M.HhaI) 319 magainin 197 malathion 197 maleimides 66 mammal proteins 305 maraviroc (UK-427 857) 219 mass spectrometry 245–246 matrix-assisted laser desorption and ionization (MALDI) 150 melanins 270 – synthesis of 270 membrane fluidity 195 membrane model systems 199–200 metallothionein (MT) 7 metal-mediated protein modification 71–80 – iridium-catalyzed reductive alkylation of lysine 74 – at natural residues 72–74 – – palladium-catalyzed allylation of tyrosine 73

331

332

Index metal-mediated protein modification (contd.) – – reductive desulfurization of cysteine 72 – – rhodium carbenoid alkylation of tryptophan 73 – olefin metathesis at s-allyl cysteine 77–78 – of unnatural residues 74–77 – – alkynes, cu-catalyzed [3 + 2] cycloaddition of 74 – – azides, cu-catalyzed [3 + 2] cycloaddition of 74 – – palladium-catalyzed carbon–carbon bond formation 76–77 methane, hydroxylation 38 methionine 71 methionine(Met)-rich motif 6 methylcobalamin 97 – dependent reactions 99 methylene blue 319 methylerythritol (ME)-binding pocket 170 methyl Gilman reagents 43 methylococcus capsulatus 36–37 – pMMO crystal structure 37 microarrays 246 microbeads 200 microsphere-mediated cellular delivery 117–135 – bioconjugation 133–135 – – streptavidin–biotin 134–135 – biodegradable microspheres 119–120 – biostable microspheres 120–121 – cleavable linkers 130–133 – – disulfide bonds 132–133 – – ester bonds 130–132 – delivery devices 117–119 – – cell-penetrating peptides 118 – – dendrimers 118 – – liposomes 117–118 – – nanomaterials 118–119 – noncleavable link 126–130 – – intracellular sensing 126–127 – – siRNA delivery 127–130 – solid-phase chemistry and 121–125 – – coupling agents 124–125 – – dual functionality 122–124 – – fmoc chemistry 121–122 – – preparation 121 microsphere spacers 123 model copper systems, O2 reactivity at 267–286 – biomimetic Cu complexes 272–285 – – aromatic rings, hydroxylation of 278–285 – – copper dioxygen adducts 272–275

– – ligand architecture 275–278 – – O2 binding and activation 272–285 – copper proteins involved 268–271 – – hemocyanin 269–270 – – tyrosinase 270–271 – O2 activation – – model systems and 267–268 – properties of 269 – use of 267 modified proteins, See chemical protein modification molecular chirality 293–295 molecular dynamics 297 MOLOC 170, 176, 179 monoubiquitination 10 Monte–Carlo simulations 144 multidrug resistance (MDR) 198 multifunctionalization 123–125 m-xylyl 279 mycobacterium tuberculosis 168 myristic acid 192

n nanomaterials 118–119 nanomaterials-supported lipid bilayers (nanoSLBs) 200 native chemical ligation (NCL) 67–69, 244 – limitation of 69 neuraminic acid 240 neurodegenerative disorders – biological implications 21–23 – – phospholipids 22–23 – – redox state of cellular copper 21–22 – copper ion, role of 3–24 – – brain, metal ions in 4 – – brain copper homeostasis 5–8 – – brain copper and 8–9 – ubiquitin system, role of 3–24 – – failure of 13–15 – – metal ions, interaction with 15–20 – – protein degradation 9–12 neuronal copper homeostasis 6 N-hydroxy succinimide (NHS) esters 62 nitrosylcobalamin 102 N-methylation 178 N-methyl d-aspartate (NMDA) 7 non-Leloir glycosyltransferase 251 – substrate 257 nonmevalonate pathway 168–169 nonsteroidal anti-inflammatory drugs (NSAIDs) 201–206 – drug fundamental physical–chemical studies 202–203 – membrane dynamic studies 203–205

Index – membrane structural studies 203–205 normal mode analysis (NMA) 297 novel branched thiooligosaccharides 251–256 nucleobases 147–148

o olefin metathesis 77–78 oleic acid 192 one-pot synthesis 241–242 optimer 242 organocopper 42 organocuprate(I) catalysis 42–44 – C–C bond formation in – – acetylene carbocupration 43 – – conjugate addition to α-enones 42–43 – – SN2 alkylations 43–44 – – SN2 alkylations 43–44 ortho-hydroxylation of 280–285 orthorhombic ZnII -ubiquitin crystals 20 oseltamivir 240 oxidation 59 oxime ligation 78–79

p palladium-catalyzed allylation 73 palmitic acid 192 parathion 197 Parkinson’s disease, causes 14 particulate methano monooxygenase (pMMO) 36–38 peptide ligation 244 peptidylglycine α-hydroxylating monooxygenase (PHM) 38 perivacuolar compartment 12 pernicious anemia 95 P-glycoprotein (PgP) 198 phospholipids 192 – monolayers 199–200 – polar groups of 192 photolysis 102 p-iodophenylalanine 76 piperidinyl ring 178 plasma membrane 191 plasmodium falciparum 168 poly-l-proline 296 poly-l-proline II 296 polymer colloid 119 polypeptidic chain 270 polyprenylation 59 polystyrene microspheres 121 polyubiquitination process 17 posttranslational modification 59–60 Prion diseases 305

Prion protein (PrP) 7, 305 proof-of-principle approach 232 propionibacterium shermanii 95 protein degradation, ubiquitin, role of 9–12 protein–protein interactions 215–232 – biosynthesized libraries – – AICAR transformylase activity 222–224 – – high-throughput screening of 220–226 – – HIV budding, cyclic peptide inhibitors of 224–226 – chemical libraries – – high-throughput screening 216–220 – future direction 226–232 – – asthma 228–229 – – networks of influenza virus 230–232 – – tumor hypoxia response network 226–228 protofibrils 13 proton-coupled electron transfer mechanism (PCET) 282 pseudomonas denitrificans 95 pseudo-π sandwich 173 purine synthesis 222–223 pyrazolyl 178 pyruvate 168

r radiation damage – to DNA 143–162 – – chemical bonds, breaking of 145–147 – – DEA studies on model compounds 148–161 – – low-energy electrons role 143–148 – – nucleobases 147–148 Raman spectrum 269 reductive desulfurization 72 reductive slkylation 63 remodeling 228 resonance 145 resorufin 131 reverse two-hybrid system (RTHS) 221 – heterodimeric 222 rhodium carbenoid alkylation 73 ribose moiety 101–102 ripple phase 194 RNA interference (RNAi) pathway 127 RNA-induced silencing complex (RISC) 127 root mean square deviations (RMSDs) 305 55% rule 179–180

s 26S proteasome SAXS 204

10

333

334

Index scattering methods 204 shape resonance 155 β sheets 296 Shilov system 45 sialic acid 230 SICLOPPS mechanism 221 signal transducer and activator of transcription factor 6 (STAT-6) 228 single strand breaks (SSBs) 144 small interfering RNA (siRNA) 127 small unilamellar vesicles (SUVs) 199 sodium cyanoborohydride 62 solid-phase chemistry 121–125 solid-phase synthesis 242–243 soluble methane monooxygenase (sMMO) 36 solvent polarity 18–19 Sonogashira coupling 76 sonolysis 102 spin-forbidden process 267 spray-drying 120 Staudinger ligations 79 – site-selective modification 80 steap proteins 6 stearic acid 192 sterols 192 stoichiometric systems 47–51 streptavidin 134–135 STRIDE algorithm 297 structural analysis 184 structure-based design 167 substantia nigra 5 substrate microarrays 257–259 superoxide dismutase (SOD) 7 supported phospholipid bilayers (SPBs) 200 Suzuki–Miyaura cross-coupling 77 α-synuclein (α-syn) 9

t tartaric acid 293 TATA-box binding protein (TBP) 319 tetraacetyl-d-ribose 155–159 tetrahydrothiophenyl ring 178 tetrazine ligation 81 tetrazole-containing proteins, selective modification of 80–81 thiooligosaccharides 254–256 toluene 248 transcobalamin I 96 trans-Golgi network (TGN) 7 translational movement 195 triazamacrocyclic ligand 50

trifluoroperazine 197 trinuclear copper models 40–41 tryptophan modification, rhodium carbenoid alkylation 73 tumor hypoxia response network 226–228 β turns 296 tyr 33–34 tyrosinase 268, 270–271 – function of 270 – structure of 270 tyrosine, palladium-catalyzed allylation 73

u ubiquitin 3 – metal ions, interaction with 15–20 – – CuII-binding and solvent polarity, cooperativity between 18–19 – – CuII-induced self-oligomerization 17–18 – – polyubiquitination process, possible implications for 17 – – thermal stability 15 – neurodegenerative disorders, failure in 13–15 – phospholipids and 22–23 – protein degradation 9–12 ubiquitination 10 ubiquitin-proteasome system (UPS) 10, 11 Ullmann chemistry 41

v vanadium 102 vertical attachment energy (VAE) 146 vinyl sulfones 66 virtual screening 167 vitamin B12 95–109 – derivatives 99–108 – – β-axial position 102–108 – – b-, d-, e-cobalamin derivatives 99–101 – – molecular structure 100 – – ribose moiety, modifications on 101–102 – – structure 99 – discovery of 95 – metabolism of 97–99 – – adenosylcobalamin-dependent reactions 97–99 – – methylcobalamin-dependent reactions 99 – transport mechanism 96–97

w water-soluble inhibitors – design of 182–183

Index – – enzyme assays of inhibitors WAXS 204 Wilson disease 5

183

– water-soluble inhibitors, design of 182–183 – – enzyme assays of inhibitors 183

x

z

X-ray cocrystal structure analysis 182–184

zanamivir

240

335

Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life

Ideas in Chemistry and Molecular Sciences: Advances in Synthetic Chemistry

Ideas in Chemistry and Molecular Sciences: Advances in Nanotechnology, Materials and Devices

Molecular Biology in Medicinal Chemistry

Molecular Biology in Medicinal Chemistry

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Gold Chemistry: Applications and Future Directions in the Life Sciences

Chemistry: The Molecular Science

Stilbenes: Applications in Chemistry, Life Sciences and Materials Science

When topology meets chemistry: A topological look at molecular chirality

Molecular physical chemistry

Partial Order in Environmental Sciences and Chemistry

Partial Order in Environmental Sciences and Chemistry

Microsystem Technology in Chemistry and Life Sciences (Topics in Current Chemistry)

Where Sea Meets Sky

Where sea meets sky

Where Sea Meets Sky

Where Earth Meets Sky

Where Sea Meets Sky

Where Earth Meets Sky

Where sea meets sky

The Chemistry of Life

Where Sea Meets Sky

Where Sea Meets Sky

The Chemistry of Life

Biology - Molecular Biology In Medicinal Chemistry

Molecular Pathology (Modules in Life Sciences)

Art in Chemistry; Chemistry in Art:

Ideas in Chemistry and Molecular Sciences: Where Chemistry Meets Life

Ideas in Chemistry and Molecular Sciences: Advances in Synthetic Chemistry

Ideas in Chemistry and Molecular Sciences: Advances in Nanotechnology, Materials and Devices

Molecular Biology in Medicinal Chemistry

Molecular Biology in Medicinal Chemistry

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Physical Chemistry for the Life Sciences

Gold Chemistry: Applications and Future Directions in the Life Sciences

Chemistry: The Molecular Science

Stilbenes: Applications in Chemistry, Life Sciences and Materials Science

When topology meets chemistry: A topological look at molecular chirality

Molecular physical chemistry

Partial Order in Environmental Sciences and Chemistry

Partial Order in Environmental Sciences and Chemistry

Microsystem Technology in Chemistry and Life Sciences (Topics in Current Chemistry)

Where Sea Meets Sky

Where sea meets sky

Where Sea Meets Sky

Where Earth Meets Sky

Where Sea Meets Sky

Where Earth Meets Sky

Where sea meets sky

The Chemistry of Life

Where Sea Meets Sky

Where Sea Meets Sky

The Chemistry of Life

Biology - Molecular Biology In Medicinal Chemistry

Molecular Pathology (Modules in Life Sciences)

Art in Chemistry; Chemistry in Art:

Recommend Documents