Trends in Drug Research III: Proceedings of the 13th Noordwijkerhout-Caminero Symposium, the Netherlands, 6-11 May 2001 (Pharmacochemistry Library)

P H A R M A C O C H E M l S T R Y L I B R A R Y - V O L U M E 32 TRENDS IN DRUG RESEARCH III Proceedings of the 13th N...

Author: H. Van Der Goot | Netherlands) Noordwijkerhout-Caminero Symposium 2001 (Noordwijkerhout

186 downloads 585 Views 14MB 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

P H A R M A C O C H E M l S T R Y L I B R A R Y - V O L U M E 32

TRENDS IN DRUG RESEARCH III Proceedings of the 13th Noordwijkerhout-Camerino Symposium

PHARMACOCHEMISTRY LIBRARY, edited by H.Timmerman Other titles in this series Volume 24 Perspectives in Receptor Research. Proceedings of the 10'hNoordwijkerhout-Camerino Symposium, Camerino (Italy), 10-14 September 1995 edited by" D. Giardinb, A. Piergentili and M. Pigini Volume 25 Approaches to Design and Synthesis of Antiparasitic Drugs edited by Nitya Anand Volume 26 Stable Isotopes in Pharmaceutical Research edited by'Thomas R. Browne Volume 27 Serotonin Receptors and their Ligands edited by B.Olivier et al. Volume 28 Proceedings XIVth International Symposium on Medicinal Chemistry edited by E Awouters Volume 29 Trends in Drug Research I1. Proceedings of the 11th Noordwijkerhout-Camerino Symposium, Noordwijkerhout (The Netherlands), 11-15 May,1997 edited by H. van der Goot Volume 30 The Histamine H3Receptor. ATarget for New Drugs edited by• R. Leurs and H.Timmerman Volume 31 Receptor Chemistry towards the Third Millennium. Proceedings of the 12'h Noordwijkerhout-Camerino Symposium, Camerino (Italy), 5-9 September 1999 edited by U. Gulini et al.

PHARMACOCHEMISTRY

LIBRARY

E d i t o r : H. T i m m e r m a n

V o l u m e 32

TRENDS IN DRUG RESEARCH III Proceedings of the ] 3'" Noordwijkerhout-Caminero Symposium, The Netherlands, 6-11 May 2001

Edited by: Henk v a n der G o o t Department of Pharmacochemistry, Free University Amsterdam, The Netherlands

2002 ELSEVIER Amsterdam - London - NewYork-

O x f o r d - Paris - S h a n n o n - T o k y o

ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 RO. Box 211, 1000 AE Amsterdam,The Netherlands ©2002 Elsevier Science B.V. All rights reserved. This work is protected under copyright by Elsevier Science, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier Science Global Rights Department, PO Box 800, Oxford OX5 1DX, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: [email protected]. You may also contact Global Rights directly through Elsevier's home page (http://www.elsevier.com), by selecting 'Obtaining Permissions'. In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90Tottenham Court Road, London WlP 0LP, UK; phone: (+44) 207 631 5555; fax: (+44) 207 631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of Elsevier Science is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier Science Global Rights Department, at the mail, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2002 Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress has been applied for. ISBN:

0-444-50760-4

The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.

PHARMACOCHEMISTRY LIBRARY ADVISORY BOARD T. Fujita E. Muts©hler N.J. de Souza EJ. Zeelen

Department of Agricultural Chemistry, Kyoto University, Kyoto, Japan Department of Pharmacology, University of Frankfurt, Frankfurt, Germany Research Centre, Wockhardt Centre, Bombay, India Heesch,TheNetherlands

This Page Intentionally Left Blank

vii

CONTENTS Preface

.....................................................................................................

ix

A. Hogner, J.S. Kastrup, J. Greenwood, S.B. Vogensen, E.H. M¢ller, T.B. Stensb¢l, J. Egebjerg and P. Krogsgaard-Larsen Towards rational design of AMPA receptor ligands: an integrated medicinal, computational, biostructural and molecular pharmacological approach ..................... 1

C. C.A. van Boeckel From heparin to synthetic antithrombotics. The pentasaccharide story and follow-up .................................................................................................... 13

A. Bruggink A new future for synthesis? ............................................................................ 21

M. T. Reetz Directed evolution of enantioselective enzymes as catalysts in the production of chiral pharmaceuticals .................................................................................. 27

R. Schoevaart and A.P.G. Kieboom At the interface of organic synthesis and biosynthesis ......................................... 39

R.E. Hubbart What can structure tell us about function in the estrogen receptors? ...................... 53

M. Kouwijzer and J. Mestres Molecular docking and dynamics simulations in the ligand binding domain of steroid hormone receptors .............................................................................. 57

S. Kliewer Peroxisome proliferator-activated receptors and reverse endocrinology .................... 67

A.J.R. Heck and C.S. Maier Biomolecular mass spectrometry related to drug research ..................................... 81

P. Angeli and G. Gaviraghi Chemical and biological diversity in drug discovery ........................................... 95

J. Gomeza, 11/1. Yamada, A. Duttaroy, W. Zhang, R. Makita, T. Miyakawa, J. Crawley, L. Zhang, H. Shannon, F.P. Bymaster, C.C. Felder, C. Deng and J. Wess Muscarine acetylcholine receptor knockout mice: phenotypical analysis and clinical implications ....................................................................................

97

P.F. Zaratin, A. Quattrini, S. Previtali, G. Hervieu and M.A. Scheideler Changes in expression of the orphan G-protein coupled receptor GPR7 in human painful peripheral neuropathies .................................................................... 115

o o °

VIII

V.J. Gillet and P. Willett Computational methods for the analysis of molecular diversity ............................ 125

D. Langley Enhancing drug discovery by acquisition of chemical diversity ............................ 135

P. Seneci Chemical diversity as a driving force to design and put in practice synthetic strategies leading to combinatorial libraries for lead discovery and lead optimization ....................................................................................... , ..... 147

H. Just, E. Stefan, C. Czupalla, B. Niirnberg, Chr. Nanoff and M. Freissmuth Beyond G proteins: the role of accessory proteins in G protein-coupled receptor signalling ......................................................................................... . ....... 161

Ph.G. Strange Mechanisms of action of antipsychotic drugs: the role of inverse agonism at the D2 dopamine receptor ................................................................................. 175

K.E.O. /lkerman, J. Niisman, T. Holmqvist and J.P. Kukkonen Agonist channeling of o~2-adrenoceptor function ............................................... 181

D. Golemi, L. Maveyraud, J. Haddad, W. Lee, A. Ishiwata, K. Miyashita, L. Mourey, S. Vakulenko, L. Kotra, J. Samama, and S. Mobashery Antibacterials as wonder drugs and how their effectiveness is being compromised .... 193

E.P. Greenberg Pseudomonas aeruginosa quorum sensing: a target for antipathogenic drug d i s c o v e r y . ................................................................................................. 207

I. Chopra, L. Hesse and A. O'Neill Discovery and development of new anti-bacterial drugs ...................................... 213

B. B. Zhan g Discovery of small molecule insulin mimetics as potential novel antidiabetic agents .................................................................. , ..... , ............................. 227

L.M. Furness Expression databases for pharmaceutical lead optimisation ................................. 237

R. G. Pertwee New developments in the pharmacology of cannabinoids ................................... 249 Author index ............................................................................................. 259 Subject index ............................................................................................ 261

IX

PREFACE Trends in Drug Research followers or setters This volume of Pharmacochemistry Library comprises the text of invited lecturers presented at the Noordwijkerhout-Camerino Symposium Trends in Drug Research, held in Noordwijkerhout, The Netherlands, from 6-11 May 2001. During the 13^^ symposium in the series the question was asked whether medicinal chemists are following trends or perhaps trendsetters. The answer was clear: trendsetters. Through the years of the series - the first one dates back to 1974 - topics of the programme have been developing into almost routine aspects of medicinal chemistry; QSAR, modelling, receptor models. The 13th symposium fitted perfectly well in this tradition. On the programme were sessions on chemical and biological diversity, on new paradigms in drug action, on new insights in receptor mechanisms. A session which got much attention - and which brought new insights - was on green chemistry, the interface between organic synthesis and biosynthesis. A special symposium was devoted to the growing problem of resistant micro-organisms and the possibilities to identify new - and better - antibiotics. In a final session on very recent developments the new finding of small molecules with insulin sensitizing properties received much attention. Would an insulin-mimetic, a small molecule, be possible? The organizers of the Noordwijkerhout-Camerino Symposia express their sincere thanks to those who supported the 2001 symposium financially: Astra Zeneca, Byk Nederland, DSM, Glaxo Wellcome, Janssen Research Foundation, Lundbeck A/S, E. Merck, Organon Research, Pfizer (Parke Davis), UCB, Yamanouchi Europe. H.Timmerman, Chairman Organizing Committee


H. van der Goot (Editor) Trends in Drug Research III © 2002 Elsevier Science B.V. All rights reserved

Towards Rational Design of AMPA Receptor Ligands: An Integrated Medicinal, Computational, Biostructural and Molecular Pharmacological Approach Anders Hogner, Jette S. Kastrup, Jeremy Greenwood, Stine B. Vogensen, Eva H. M0ller, Tine B. Stensbol, Jan Egebjerg^ and Povl Krogsgaard-Larsen*

Department of Medicinal Chemistry, The Royal Danish School of Pharmacy, 2 Universitetsparken, DK-2100 Copenhagen, Denmark and ^Department of Molecular Genetics, H. Lundbeck A/S, 9 Ottiliavej, DK-2500 Valby-Copenhagen, Denmark

1. Medicinal Chemistry, a Science Undergoing Rapid Transformations

The field of medicinal chemistry is in a state of swift development and is at present undergoing major restructuring. The molecular biological revolution and the progressing mapping of the human genome have created a new biochemical and biostructural "world order". These developments have provided new challenges and opportunities for drug research in general and for drug design in particular. The major objectives of the medicinal chemists are transformation of pathobiochemical and - physiological data into a "chemical language" with the aim of designing molecules interacting specifically with the derailed or degenerating processes in the diseased organism. Potential therapeutic targets are being disclosed with increasing frequency, and this exponential growth will continue during the next decades. In this situation, there is a need for rapid and effective target validation and for accelerated lead discovery procedures. Consequently, most industrial medicinal chemistry laboratories have built up new technologies in order to meet these demands. Key words in this regard are construction of compound libraries, high or ultrahigh throughput screening, accelerated ADME and toxicity tests, and automatized cellular assay systems. In parallel with this development, biostructure-based drug design and intelligent molecular mimicry or bioisosterism are areas of growing importance in the medicinal chemistry "playing field". Structural biology is becoming an increasingly important part of molecular biology and biochemistry, and, furthermore, organic chemists are increasingly directing their attention towards synthetic aspects of biomolecules and biologically active compounds biosynthesized by plants and animals. Thus the borderland between biology, biochemistry, and chemistry is rapidly broadening and is becoming the most fruitful working field for innovative and intuitive drug design scientists.

2. Industrial Drug Discovery - Academic Drug Design

Where are the academic medicinal chemistry departments in this area of drug research, which is now undergoing profound changes, and which is moving towards an increasing degree of integration of scientific disciplines? Furthermore, how should medicinal chemistry teaching programmes be organized and taught in this highly dynamic research area? These burning questions need to be effectively addressed, and if the responsible academics fail to meet these challenges, academic medicinal chemistry will degenerate into traditional organic synthesis, from where it originates, or into trivial service functions in relation to industrial drug design and development programmes. The equipment for automatized combinatorial chemistry and for high throughput screening procedures, now in operation in most industrial medicinal chemistry departments, is expensive, and purchase of such technical facilities is normally far beyond the financial capacity of academic departments. Furthermore, in terms of operation, these automatized procedures are predominantly technical, and although students should understand the prospects and limitations of such technologies, these aspects can only be limited parts of student courses in medicinal chemistry. The scientific challenges of the conversion of solution synthetic chemistry procedures into solid-phase synthetic methodologies are mainly of basic chemical nature, and the development of cell-based assay systems is predominantly a biochemical pharmacological task. In order to attract the attention of intelligent students, the creative and fascinating nature of drug design must be the imderlying theme of basic and advanced student courses in medicinal chemistry. In relation to industrial screening programmes and "hit-finding" procedures, students should be taught that the conversions of "hits" into lead structures and further into drug candidates require advanced synthetic chemistry supported by computational chemistry. Furthermore, these medicinal chemistry approaches should be integrated with molecular pharmacology studies using cloned target receptors, ion channels, or enzymes, expressed in appropriate model systems. It is beyond doubt that a steadily increasing number of biomolecules will be subjected to X-ray crystallographic structural analysis. The number of enzymes with established three-dimensional structure is now increasing exponentially [1], and this growth will continue during the next decades. Even oligomeric membrane-boimd receptors can now be crystallized and subjected to X-ray crystallographic analysis [2], but such analyses of mono- or oligomeric receptors are still hampered by major experimental difficulties. In recent years, however, biostructural scientists have succeeded in crystallizing recombinant versions of the binding domains of a G protein-coupled receptor [3] as well as a ligand-gated ion charmel [4]. Structural analyses of these binding domains cocrystallized with agonist and antagonist ligands have already provided insight into the structural basis of receptor-ligand interactions and of receptor activation and blockade.

These breakthroughs in biostructural chemistry have opened up new avenues in drug design. Structural information derived from X-ray analyses of enzyme-inhibitor conglomerates has been and continues to be very valuable for the design of new types of inhibitors. Similar pieces of information derived from studies of receptor binding domains co-crystallized with different types of competitive or noncompetitive ligands undoubtedly will be of key importance in receptor ligand design projects. These approaches which are in the nature of drug design on a rational basis will become important parts of student teaching programmes in medicinal chemistry. In academic research and teaching, biologically active natural products probably will play a progressively important role as lead structures. Not only do such compounds often possess novel structural characteristics, but they alsofrequentlyexhibit unique biological mechanisms of action, although naturally occurring "toxins" typically show nonselective pharmacological effects. By systematic structural modification, including molecular mimicry approaches, it may be possible to "tame" such "toxins" and convert them into leads with specific actions on biofiinctions of key importance in diseases. Biologically active natural products undoubtedly will be continue to be important starting points for academic drug design projects, and such approaches will continue to be exciting case stories in student medicinal chemistry courses.

Figure 1. Leading academic medicinal chemistry departments or centres capable of establishing innovative collaborative projects with major industrial drug discovery units will optimally have the above "four-leaf clover" integrated composition of expertises.

In conclusion, there are growing indications that industrial and academic medicinal chemistry approaches will develop differently. Industrial medicinal chemistry projects will be in the nature of drug discovery with fast and effective hit-to-lead-to-clinical candidate development as key words. Innovative academic approaches are likely to focus on long-term development of rational approaches to drug design based on biostructural analyses and molecular mimicry. In both cases computational chemistry will be a key discipline, and molecular pharmacology certainly will be an essential and fully integrated discipline in medicinal chemistry (Figure 1). Hopefully, these two major lines of medicinal chemistry research will develop in a complementary fashion, which will open up the prospects of establishing fruitful collaborative projects between industrial and academic drug design scientists. A prerequisite for the build up of productive and innovative collaborative projects along these lines is the recognition of the collaborators as equal partners. Furthermore, mutual scientific appreciation will form the basis for the extremely important participation of industrial medicinal chemists in teaching courses and training programmes in drug design. In the following we describe the development of a long-term academic medicinal chemistry project on glutamate receptor ligands from a classical drug design project based on re-design of a naturally occurring amino acid "toxin", ibotenic acid, into an integrated rational approach involving medicinal chemistry. X-ray crystallographic protein structural analysis, computational chemistry, and molecular pharmacology. 3. AMPA Receptor Ligands: Therapeutic Prospects The central excitatory neurotransmitter effects of (5)-glutamic acid [(5)-Glu] are mediated by three heterogeneous classes of ionotropic receptors named 7V-methyl-Daspartic acid (NMDA), 2-amino-3-(3-hydroxy-5-methyl-4-isoxazolyl)propionic acid (AMPA), and kainic acid (KA) receptors [5,6] (Figure 2), and a number of subtypes of metabotropic receptors [6,7]. These or perhaps distinct subtypes of these receptors have been associated with certain neurologic and psychiatric diseases and are potential therapeutic targets in such diseases [5-7]. In recent years, much interest has been directed towards the role of AMPA receptors in the mechanisms associated with cognitive functions [8], and enhancement of AMPA receptor functions has been shown to facilitate learning and memory [6,9]. Although AMPA receptor agonists may not be used therapeutically due to potential neurotoxicity, these observations have focused interest on the molecular mechanisms of receptor activation and, thus, on the structural basis of AMPA receptor-agonist interactions.

NR1 NR2A NR2B NR2C NR2D NR3A GluR1 GluR4 GluR2 GluR3

{s

NMDA

AMPA

GluR7 -1 GluR6 GluR5 KA1 KA2 51 52

20

40

60 Percent identity

80

KA

-*

] ORPHAN

100

Figure 2. Multiplicity and phylogenetic comparison of the ionotropic glutamate receptor subunits.

The AMPA Receptor Recognition Site - Hypothesis of the Presence of a Lipophilic Binding Pocket Extensive structure-activity studies (SARs) on analogues of the classical AMPA receptor agonist, AMPA [10], in which a variety of alkyl [11], aryl [12], or heteroaryl groups [ISIS] have been substituted for the methyl group, have shed some light on the structural requirements for activation of AMPA receptors by this class of ligands. A series of AMPA analogues, in which the size of the alkyl substituents in the 5-position of the 3-isoxazolol ring has been systematically increased, have been synthesized and pharmacologically characterized. Whereas the ethyl analogue of AMPA, Et-AMPA, is slightly more potent than AMPA [16], analogues containing larger alkyl groups are much weaker or inactive, and the isopentyl analogue, Pe-AMPA, does not interact detectably with AMPA receptors [11] (Figure 3). Interestingly, all of the active members of this series of compounds show full agonist effects. There are no indications of partial agonist effects of any of these analogues.

(S)-2-Me-Tet-AMPA

P

1-Me-Tet-AMPA

Tet-AMPA

P

(S)-Thio-ATPA

Figure 3. Structures of a number of analogues of AMPA. A large number of 3-isoxazolol amino acids containing a heterocyclic unit in the 5position of the ring have been synthesized and characterized pharmacologically. SAR studies disclosed that only heterocyclic substituents containing heteroatom(s) in the 2position(s) of the ring showed potent AMPA agonist effects [13,14], as exemplified by (5)-2-Me-Tet-AMPA (Figure 3), which is the most potent AMPA agonist so far tested [13,17]. On the other hand, (/?^-2-Me-Tet-AMPA and the isomeric compound, 1-MeTet-AMPA, and also Tet-AMPA possessing an imsubstituted tetrazole ring, were essentially inactive [13]. Whereas the inactivity of Tet-AMPA probably reflects the fact that the tetrazole ring carries a negative charge at physiological pH, the methyl substituent in 1-Me-Tet-AMPA appears to sterically hinder binding to the AMPA receptor. These S ARs were interpreted in terms of the presence of a lipophilic pocket of limited volume at the binding site of the AMPA receptors [16,18]. Since the demethyl analogue of AMPA is markedly weaker than AMPA itself [11], the occupancy of this proposed pocket seems to be important for binding to and activation of the receptors. Although it still is unclear, why the presence of heteroatom(s) in the 2-position(s) of the heteroaromatic substituents of the bicyclic AMPA analogues is important for AMPA agonist activity, it is tentatively concluded that the presence of a heteroatom, and thus absence of a hydrogen atom, in this position facilitate hydrogen bond formation in the proposed lipophilic pocket.

5. From AMPA Receptor Agonists to Subtype-Selective Kainic Acid Receptor Agonists Substitution of a tert-butyl group for the methyl group of (5)-AMPA to give (5)-ATPA (Figure 3) has profound pharmacological consequences. Thus, whereas (5)-AMPA is a potent and highly selective agonist at AMPA receptors, (5)-ATPA is a very potent and selective agonist at kainic acid (KA) preferring receptors of the GluR5 subtype, showing only weak agonist effects at AMPA receptor subtypes [19]. Hence, the bulky tert-hutyl group of (5)-ATPA does not seem to be easily accommodated by the proposed lipophilic cavity of the AMPA receptors. On the other hand, the unique pharmacology of ATPA [20], and in particular its (5)-enantiomer [19], suggests that the spherical structure of the tert-hutyl group almost perfectly fits into a cavity at the binding site of the GluRS receptor. The structurally related 3-isothiazolol amino acid, (5)-Thio-ATPA (Figure 3), shows very similar pharmacological effects but is markedly more potent at GluRS receptors than (5)-ATPA [21]. This observation suggests that the increased lipophilicity of the "bottom part" of the molecule of (5)-Thio-ATPA contributes to its remarkably high agonist potency at GluR5. (iS)-ATPO is a structural hybrid between (5)-ATPA and the classical competitive NMDA antagonist 2-amino-7-phosphonoheptanoic acid (AP7) [22]. Quite surprisingly, (S)ATPO turned out to be a competitive AMPA receptor antagonist showing similar potency at subtypes of AMPA receptors but substantially weaker antagonist effects at the GluR5 subtype of KA receptors (Figure 2) [23]. These observations emphasize that the structural parameters of importance for the interaction of agonists and competitive antagonists with the binding domain of the GluR2 subtype of AMPA receptors are different [24].

6. X-Ray Crystallographic Studies of a Recombinant AMPA Receptor Binding Domain Co-Crystallized with (5)-Glutamic Acid or AMPA Receptor Agonists Until recently, no structural data of (5)-Glu receptors have been available. A breakthrough did, however, occur in 1998 with the publication of the first crystal structure of the ligand-binding domain of the AMPA receptor GluR2 subtype in complex with kainic acid [4]. This achievement was based on the finding that the transmembrane region separating the two parts of the receptor protein forming the ligand binding site could be replaced by a peptide linker. These fused protein units, known as S1S2, quite remarkably retain binding affinities similar to those of the wild-type membrane-bound receptor [4,25]. The development of large-scale expression and purification methodologies for the GluR2-SlS2 constructs have paved the way for a number of highresolution structures of GluR2-SlS2 in complex with AMPA agonists and an AMPA

antagonist and in the apo state of GluR2 [24]. This series of structures provides the basis for detailed structure-function analyses. In collaboration with the group of E. Gouaux, a number of AMPA agonists and antagonists, developed at this Department, have now been co-crystallized with the GluR2-SlS2 construct. The structure of S1S2 in complex with (S)-2-Me-Tet-AMPA is illustrated in Figure 4. (S)-2-Me-Tet-AMPA is bound in the cleft between the two domains. Domain 1 is composed of segment SI and the C-terminal end of segment S2. The C-terminal end of segment SI ends in domain 2, which primarily is composed of segment 2. An analysis of the ligand-protein complex indicates that the methyltetrazole unit almost perfectly fits into a binding pocket.

N-Terminus

C-Tenninus

Linker: G-T

Figure 4. Ribbon representation of the GluR2-SlS2 structure in complex with (5)-2-MeTet-AMPA. The SI and S2 segments are in black and grey, respectively. One of the challenges, now evolving, is to interpret the structural data with the goal of designing new ligands on a rational basis. So far, the binding mode of (S)-Glu bioisosteres, such as (5)-AMPA (Figure 3), has been unclear. From the published crystal structure complexes of (S)-Glu and (S)-AMPA it is evident that the a-carboxylate and the a-ammonium group in both complexes bind to SIS2 in a consistent manner [24].

These groups form strong interactions via specific hydrogen bonds and ion-pair interactions with amino acid residues from both domain 1 and 2 (Figure 5).

Figure 5. The binding site of GluR2-SlS2 in complex with (*S)-Glu. (5)-Glu is in dark grey, and the backbones of domain 1 and of domain 2 are in black and light grey, respectively. Oxygen and nitrogen atoms are displayed as black spheres and carbon atoms as white spheres. Dashed lines indicate potential hydrogen bonds within 3.2A and water molecules are shown as dark grey spheres. Pdb id-code IFTJ (Armstrong and Gouaux [24]).

A superposition of (5)-Glu and (5)-AMPA structures shows that the fundamental interactions between S1S2 and the amino acid group are conserved (Figure 6). The difference between these two agonists lies in the different positioning of the distal carboxylate group of (iS)-Glu and the 3-isoxazolol anion of (5)-AMPA. It turns out that (5)-AMPA does not bind as a true structural bioisostere of (5)-Glu. Instead, a water molecule in the (iS)-AMPA structure occupies a similar site to one of the oxygens of the distal carboxylate group of (iS)-Glu. This piece of information emphasizes the role of tightly bound water molecules within the binding site. The binding of the 3-isoxazolol

10 anion of (iS)-AMPA is further stabilized by strong hydrogen bonds both directly to the protein but also through water-mediated hydrogen bonds. The methyl group of (5)AMPA partially fills a hydrophobic pocket in domain 1. By comparing the interactions of the 5-position substituents of (iS)-Me-Tet-AMPA and (5)-AMP A with the binding site, it becomes evident that the ethyl group of Et-AMPA (Figure 3) fits better into the hydrophobic cavity than the methyl group of (5)-AMPA. This observation may explain why Et-AMPA is slightly more potent than AMP A as an AMP A receptor agonist [16].

Figure 6. The binding site after superposition of the structures of GluR2-SlS2 in complex with (5)-Glu and (S)-AMPA. (5)-Glu and (5)-AMPA are in dark grey and white, respectively. Otherwise as in Figure 5. Pdb id-codes IFTJ and IFTM (Armstrong and Gouaux [24]).

The X-ray crystallographic data derived from structural analyses of complexes between the recombinant GluR2-SlS2 binding domain and different AMPA receptor ligands are now being exploited in terms of rational design of new types of AMPA receptor ligands.

11

7. References [I] D. Leung, G. Abbenante and D.P. Fairlie, J.Med.Chem. 43 (2000) 305. [2] K. Brejc, W J. van Dijk, R.V. Klaassen, M. Schuurmans, J. van der Oost, A.B. Smit and T.K. Sixma, Nature 411 (2001) 269. [3] N. Kunishima, Y. Shimada, Y. Tsuji, T. Sato, M. Yamamoto, T. Kumasaka, S. Nakanishi, H. Jingami and K. Morikawa, Nature 407 (2000) 971. [4] N. Armstrong, Y. Sun, G.-Q. Chen and E. Gouaux, Nature 395 (1998) 913. [5] D.T. Monaghan and R.J. Wenthold, Eds., The lonotropic Glutamate Receptors, Humana Press, Totowa, New Jersey, 1997. [6] H. Brauner-Osbome, J. Egebjerg, E.0. Nielsen, U. Madsen and P. KrogsgaardLarsen, J.Med.Chem. 43 (2000) 2609. [7] P.J. Conn and J. Patel, Eds., The Metabotropic Glutamate Receptors, Humana Press, Totowa, New Jersey, 1994. [8] H.K. Lee, M. Barbarosie, K. Kameyama, M.F. Bear and R.L. Huganir, Nature 405 (2000) 955. [9] U. Staubli, G. Rogers and G. Lynch, Proc.Natl.Acad.Sci. U.S.A. 91 (1994) 777. [10] P. Krogsgaard-Larsen, T. Honore, J.J. Hansen, D.R. Curtis and D. Lodge, Nature 284(1980)64. [II] F.A. Sl0k, B. Ebert, Y. Lang, P. Krogsgaard-Larsen, S.M. Lenz and U. Madsen, Eur.J.Med.Chem. 32 (1997) 329. [12] B. Ebert, S.M. Lenz, L. Brehm, P. Bregnedal, J.J. Hansen, K. Frederiksen, K.P. Bogeso and P. Krogsgaard-Larsen, J.Med.Chem. 37 (1994) 878. [13] B. Bang-Andersen, S.M. Lenz, N. Skjaerbaek, K.K. Soby, H.O. Hansen, B. Ebert, K.P. Bogeso and P. Krogsgaard-Larsen, J.Med.Chem. 40 (1997) 2831. [14] E. Falch, L. Brehm, L Mikkelsen, T.N. Johansen, N. Skjaerbaek, B. Nielsen, T.B. Stensbol, B. Ebert and P. Krogsgaard-Larsen, J.Med.Chem. 41 (1998) 2513. [15] B. Bang-Andersen, H. Ahmadian, S.M. Lenz, T.B. Stensbol, U. Madsen, K.P. Bogeso and P. Krogsgaard-Larsen, J.Med.Chem. 43 (2000) 4910.

12 [16] U. Madsen, B. Frolund, T.M. Lund, B. Ebert and P. Krogsgaard-Larsen, Eur.J.Med.Chem. 28 (1993) 791. [17] S.B. Vogensen, H.S. Jensen, T.B. Stensb0l, K. Frydenvang, B. Bang-Andersen, T.N. Johansen, J. Egebjerg and P. Krogsgaard-Larsen, Chirality 12 (2000) 705. [18] L. Brehm, F.S. Jorgensen, J.J. Hansen and P. Krogsgaard-Larsen, Drug News Perspect. 1 (1998) 138. [19] T.B. Stensbol, L. Borre, T.N. Johansen, J. Egebjerg, U. Madsen, B. Ebert and P. Krogsgaard-Larsen, Eur. J. Pharmacol. 380 (1999) 153. [20] V.R.J. Clarke, B.A. Ballyk, K.H. Hoo, A. Mandelzys, A. Pellizzari, C.P. Bath, J. Thomas, E.F. Sharpe, C.H. Davies, P.L. Omstein, D.D. Schoepp, R.K. Kamboj, G.L. CoUingridge, D. Lodge and D. Bleakman, Nature 389 (1997) 599. [21] T.B. Stensbol, H.S. Jensen, B. Nielsen, T.N. Johansen, J. Egebjerg, K. Frydenvang and P. Krogsgaard-Larsen, Eur.J.Pharmacol. 411 (2001) 245. [22] G.L. CoUingridge and J.C. Watkins, Eds., The NMDA Receptor, Oxford University Press, Oxford, 1994. [23] E.H. MoUer, J. Egebjerg, L. Brehm, T.B. Stensbol, T.N. Johansen, U. Madsen and P. Krogsgaard-Larsen, Chirality 11 (1999) 752. [24] N. Armstrong and E. Gouaux, Neuron 28 (2000) 165. [25] A. Kuusinen, M. Arvola and K. Keinanen, EMBO J. 14 (1995) 6327.

H. van der Goot (Editor) Trends in Drag Research III © 2002 Elsevier Science B.V. All rights reserved

13

From Heparin to Synthetic Antithrombotics The pentasaccharide story and follow-up

C.A.A. van Boeckel N. V. Organon, Lead Discovery Unit, P.O. Box 20, 5340 BH Oss, The Netherlands

Since 1936, heparin has been used in clinics for the prevention and treatment of thrombosis. Its main antithrombotic activity is explained by its ability to potentiate the activity of the serine protease inhibitor antithrombin III (AT-III), which inactivates a number of serine proteases- such as thrombin and factor Xa- in the coagulation cascaded By the end of the 1970's heparin fragments (obtained by chemical or enzymatic degradation) had been isolated by affinity chromatography on immobilised AT-III and the high affinity fractions had been analysed. From these studies it was deduced^ in 1981 that a unique pentasaccharide (PS) fragment, that occurs in about one-third of the heparin polysaccharide chains, constitutes the minimal binding domain for AT-III. The pentasaccharide fragment (also known as the DEFGH part of heparin) was synthesised^' "* a couple of years later to confirm the earlier proposal.

Structure - Activity Org31540/SR90107 (Arixtra®)

lessential sulphates / carboxylates contributing sulphates

Figure 1 The slightly modified synthetic pentasaccharide fragment 1 ( ORG 31540/SR90107 in Figure 1) was found to elicit a very selective antithrombotic mode of action, in that it only accelerates the AT-III mediated inhibition of coagulation factor Xa but not that of thrombin^ (see Figure 2). The results of four Phase III clinical trials show that the pentasaccharide 1 provides a superior benefit over a low molecular weight heparin in

14 preventing deep-vein thrombosis (DVT) in major orthopedic surgery patients, with an overall relative risk reduction of 50% and a similar safety profile. In August 2001 Organon and Sanofi-Synthelabo have received an "approvable letter" from the U.S. Food and Drug Administration (FDA) for the registration of 1 as a new antithrombotic drug called Arixtra®

Pointa^ffiettaricte shows selective Anti Xa Activity

+ Arg +

+ +

+ Lys

Arg

factor Xa

ATin

Figure 2 The specificity of the interaction of the sulphated pentasaccharide with the protein was confirmed when heparin pentasaccharide analogues were synthesised and tested^ for inhibition of blood coagulation factor Xa. First it was established which of the charged groups play an important role in the activation of AT-III. It was found that some groups are strictly required for the activation of AT-III while other groups contribute significantly during the AT-III activation (see Figure 1). Taking into account these structure-activity relationships and by contemplating molecular modelling data we postulated a simplified AT-III/PS interaction model. On the basis of this model we introduced^ an extra sulphate group at position 3 of unit H of the naturally occurring fragment to give analogue 2 (see Figure 3). This extra-sulphated analogue displays higher affinity towards AT-III and an enhanced AT-III mediated anti-Xa activity (1250 U/mg for 2 vs. 700 U/mg for 1).

15

Tirsf ATIII/PSInl^r^ttonMQd^

OSO,

OH

V V ^ ^

AT n i BINDING SITE 1

Figure 3 Subsequently, attention was turned to a new simplified series^ in which all hydroxyl groups are methylated and in which all the N-sulphate groups are replaced by Osulphate groups . Quite to our surprise these modifications did not affect the biological activity of the PS. It should be stressed that the synthesis of these methylated analogues is much easier than that of heparin-like fragments. In this series we prepared several analogues methylated at the 2-0 and 3-0 positions of both uronic acid moieties. At first sight it was expected that such analogues would loose at least half of their biological activity as was observed for the "natural" counterparts lacking the 2-0-sulphate group of iduronic acid. However, quite unexpectedly, one of these methylated analogues, (i.e. compound 3; SanOrg 34006; see Figure 4) turned out to be highly potent^, displaying 1600 anti-Xa U/mg.

San Org 34006

aXa 700 U/mg 1600 U/mg

Ko(ATin)

tVj rat

600 nM 20 nM

0.7 hr H.Ohr

Figure 4

16 The potent compound 3 not only binds much stronger to AT-III (Kd=20nM), relative to the PS (compound 1, Kd = 700 nM), but also its elimination half-life is about fifteen fold longer. SanOrg 34006 is clinically investigated for the prevention of thrombosis using a once a week dose regimen. For many PS analogues it was found^ that the elimination half-life is proportional to the affinity of AT-III. The next challenge was to extend the concept of AT-III mediated inhibition of factor Xa by pentasaccharides towards synthetically feasible derivatives displaying both antifactor Xa and anti-thrombin activity. It is known that for AT-III mediated inhibition of thrombin a heparin fragment comprising at least 16 saccharide units is required to facilitate the binding of AT-III and thrombin to the same polysaccharide chain (the so called "bridge" or "template" mechanism).

Doftl^n of Synttette CoiftiyQattisft heparin v

\

ABD

*»»+

+

y thrombin -f

+

AT III synthetic conjugates

TBD

1 v_

ABD

t

neutra^spaceT

TBD

D-unit

Figure 5 Our model^^ of the ternary complex (see Figure 5 for a schematic representation) revealed that heparin analogues may be obtained when a thrombin binding oligosaccharide is tethered to the non-reducing terminus of the AT-III binding pentasaccharide with a neutral spacer of about 50 atoms in length. To this end glycoconjugates (e.g. compound 4 in Figure 6) were synthesised which comprise a PS as AT-III binding domain (ABD), a linear spacer and a persulphated oligosaccharide as thrombin binding domain (TBD).

17

ABD

spacer

OSOs'

TBD -OSCb'

ABD 0CH3

°f^'');^">r°'t^v^'-

TBD

anti Xa = 740 U/mg antilla= 140 U/mg

Figure 6 Compound 4 was one of the first conjugates that has been synthesised and which indeed displayed good to strong AT-III mediated anti-thrombin activity (4 = 140 U/mg; heparin =160 U/mg) besides the expected anti-factor Xa activity (740 U/mg). The potency, the anti-factor Xa/anti-thrombin ratio and half-life in circulation of this new type of heparin like molecules can be adjusted^ ^'^^ in a rational way by varying the AT-III affinity of the PS (ABD), the TBD (charge density) and the spacer (length and rigidity).

18 Design of Dual Inhibitor

ATIII-mediated anti-Xa: 885 U mg-^ Direct anti-IIa activity:

ECggsCas ^.M

Figure 7 Furthermore, a different class of conjugates was designed and prepared ^^ in which a pentasaccharide is covalently linked to a direct thrombin inhibitor (e.g. NAPAP) displaying a dual mode of action (AT-III activation and direct thrombin inhibition). Such dual inhibitors (e.g compound 5 in Figure 8) have also the advantage that the PS component (bringing about AT-III mediated anti-Xa activity) and the direct thrombin inhibitor have the same half-life in circulation. In addition such dual inhibitors are expected to neutralise clot-bound thrombin more efficiently than the heparin /AT-III complex, which because of its size is hampered to penetrate the blood clot. Several of the conjugates/ dual-inhibitors are now in (pre)-clinical development.

Nmtn Dmal Antiilhiiroinfitotiic OSO 3"

0

OSO 3-

^ OSO 3"

OSO 3-

OSO 3-

^ OSO 3-

OSO 3-

H2N

NAPAP-PS Coniugate

Figure 8

OSO 3-

19 In conclusion, the discovery of the AT-III binding pentasaccharide domain in heparin opened an avenue to various new synthetic antithrombotics showing tailor-made profiles both with respect to anti-factor Xa and anti-thrombin activity (either via AT-III or a dual mode of action) and duration of action.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Heparin (Eds.rD.A. Lane, U. Lindahl), Edward Arnold, London (1989). J. Choay et al., N.Y. Acad, Sci. 370, 644 (1981). P. Sinay et al., Carbohydr.Res. 132, C5 (1984). C.A.A. van Boeckel et al., J.Carbohydr.Chem. 4, 293 (1985). J.M. Herbert et al.. Cardiovascular Drug Reviews 15, 1 (1997). C.A.A. van Boeckel, M. Petitou, Angew.Chem.Int.Ed.Engl. 32, 1671 (1993). C.A.A. van Boeckel, J.E.M. Basten, H. Lucas, S.F. van Aelst, Angew. Chem. Int. Ed. Engl. 27, 1177(1988). G.Jaurand, J. Basten, I. Lederman, C.A.A. van Boeckel, M. Petitou, Bioorg & Med. Chem. Lett. 2, 897 (1992). P. Westerduin et al., Bioorg. & Med. Chem. 2, 1267 (1994). P.D.J. Grootenhuis et al.. Nature Struct. Biol., 2 736 (1995). J.E.M. Basten, CM. Dreef-Tromp, B. de Wijs and C.A.A. van Boeckel, Bioorg. & Med. Chem. Lett. 8, 1201 (1998). CM. Dreef-Tromp et al., Bioorg. & Med. Chem. Lett. 8, 2081 (1998). R.C Buijsman et al. Bioorg. & Med. Chem. Lett. 9, 2013(1999).

Part of this work was done in collaboration with Sanofi Recherche.



21

A new Future for Synthesis? Alle Bruggink a) Department of Organic Chemistry, University of Nijmegen, Toernooiveld 1, 6525 ED Nijmegen b) DSM Research, P.O. Box 18, 6160 MD Geleen, The Netherlands

1. Summary New scientific challenges with great implications for synthetic chemistry are developing rapidly, in particular in Life Sciences, Performance Materials and Nanotechnology. Without a drastic and radical change in approach synthetic chemistry can not be expected to deliver the required contributions, whereas the timely availability of the desired molecules, employing sustainable processes in their manufacturing, has to be at the basis of these new developments.

2. Introduction At the start of the chemical and pharmaceutical industry stoichiometric synthetic chemistry was sufficient to obtain the required molecules. With the increasing scale and volume, in particular in the petrochemicals industry, chemocatalysis was required to reach efficient and economical processes. The increasing complexity and functionality of the desired molecules, in particular in the pharmaceutical industry, could be met through the introduction of biocatalysis allowing mild reaction conditions and subtle processes. However, biocatalysis is often combined with traditional, stoichiometric chemistry to reach the desired synthesis goals. Translation of chemocatalysis from petrochemicals to a broader application in chemical and pharmaceutical industry is still rather remote from maturity and, more importantly, will not be sufficient to meet the present challenges. Moreover, all our synthetic methodology is still characterized by a "step-by-step" approach; i.e. chemical bonds are manipulated one by one, requiring several protecting groups for the remaining functionalities and various activation mechanisms for the desired transformations. Prolonged development times, increased waste streams and laborious recycle loops are important drawbacks in these processes, hampering a sustainable chemical industry for the future.

3. New Challenges With the introduction of combinatorial chemistry, high-throughput-screenings and robotics the pharmaceutical industry has greatly expanded the possibilities for generating lead compounds. The molecular complexity of these products is increasing

22 as well, stimulated by the growing impact of molecular biology and the results of genomics and proteomics research programs. The (bio-)chemical industry is highly challenged in quickly synthesizing the required molecules for further pharmaceutical development. Several small scale contract research companies and kg-shops have emerged in the past few years to meet these needs. Due to the pressure on "time to market" their synthetic methodology is very often based on the versatile diversity of traditional stoichiometric chemistry. This poses a threat to future development of sustainable, catalytic processes for industrial application as the initial syntheses are often also the start of product and process registration in drug master files. In fact, improvements in processes for existing drugs through the introduction of (bio-)catalysis might be found difficult to translate to application in processes for new drugs. Thus, there is a great challenge in applying catalytic methods right from the start in the synthesis of lead compounds. Moreover, in order to meet the demands for short development times catalyst screening and development has to be done in highthroughput-systems and process steps should be highly integrated without isolation or purification of intermediates. However, this dream of "one-time-right", i.e. direct conversion of simple starting materials in a cascade of (catalytic) reactions without activating and protecting agents into (complex) end products, is still very far away in today's organic synthesis. In fact, upcoming demands from important industrial segments such as Life Sciences, Performance Materials and Nanotechnology, are bringing us a molecular complexity that no longer can be met with the presently available tools for synthesis.

4. Learningfi-omNature In nature synthesis in microorganisms occurs in a chain of reactions each catalyzed by combined and simultaneous enzyme actions. In this cascade of reactions, concentrations of starting materials, intermediates and end products are kept very low allowing maximum selectivity and no side reactions or byproducts being formed. Starting materials are brought in in a fed-batch mode using controlled membrane transport proteins. End products are continuously removed according to the in-situ-productremoval (ISPR) principle, employing again controlled transport systems. This allows, in principle, high throughputs and efficiency. Required protecting and activating agents are kept in-situ through recyclable energy and redox carriers (ATP, NADP, etc.). The desire to utilize these features of biosynthesis in nature for a wider scope of molecules might be a historic one, but now time has come to allow a meaningful and responsible scientific program towards these ends. The first steps have been made, i.e. metabolic engineering for protein synthesis can now be done. For all other molecules the molecular biology has just started to be developed. From an industrial point of view biosynthesis is mainly limited to biocatalysis (employing the 1 enzyme/1 step approach) and fermentation processes, which are mainly empirically based (see also fig. 1). These two extremes in utilizing enzymes for syntheses should be brought together.

23 Nr. of catalyst Many

More

Full Fermentations Cascade Catalysis (in concert) Bio-transPrecursor formations Fermentations Bio-Redox

Present Catalysis

Cascade Catalysis (1 by 1)

Sto'i'chiom. Chemistry

One-pot Reactions More

Many

Nr. of bonds manipulated simultaneously

Figure 1 Synthetic Methodology in Industry Fundamental understanding of these processes is developing rapidly through metabolic flux analyses, metabolic pathway engineering and related developments in molecular biology (genomics, biomolecular informatics). In particular the intricate molecular recognition mechanisms in biosystems are slowly being revealed. Application of this knowledge in improved biocatalysis is underway. Utilization of this know-how for conversions of relative simple but often unreactive molecules is in its infancy. Studies on methane-mono-oxygenase are an outstanding example. In the end all kind of inter-atomic bonds will be manageable in an efficient and economic way employing cascades of bio- or bio-inspired-catalysts. The diversity of chemistry and the complexity of biology are brought together in a new fiiture for synthesis (see also Fig. 2).

5. Towards the Best of Both Worlds. Molecular Biology & Biosynthesis Fundamental knowledge about biological systems has increased enormously in the last 25 years. At molecular level the actions of enzymes and metabolic pathways can now be identified. To translate this know-how into new biosyntheses and biocatalysts, additional insight is needed in the operation of complete micro-organisms, individual cells, cell compartments, enzyme interactions and metabolic fluxes. Challenges are in particular the role of compartments in cells and further study of the mechanisms of the several process couplings. The control of enzyme levels and activation between genome and proteome and the associated molecular recognition systems need to be understood in detail. In order to widen the scope of fermentation processes interactions between

24 primary and secondary metabolisms should be elucidated. Transport-mechanisms into, within and out of a cell or cell compartment should be known as well for natural products as for (new or known) non-natural molecules. In particular transport-proteins and mechanisms involved in product removal should be understood in much greater detail. New or revised systems might be needed to allow high throughputs. Genetically engineered biosystems should be made available to improve biosyntheses of both enzymes for biocatalyst development and metabolic end products (natural and nonnatural). In particular fermentation of non-natural molecules will need full deployment of all available tools in molecular biology. Further exploitation of directed evolution methods will greatly enhance this (including developments in genomics, biomolecular informatics etc.) Bio-transformations Although the present approach of using one enzyme for a particular conversion can still be widened in scope the challenge should be at employing combinations of enzymes. Cascades of enzymatic reactions are already emerging but can be researched much wider. Biocatalysts consisting of a number of enzymes acting in concert and interdependent such as in redox reactions should be made available. More simple and cheaper co-factor regenerating systems should be developed and build into a biocatalyst. The present focus on single enzyme biocatalysts should further be shifted to enzyme combinations, cell compartments or even complete cells as catalyst whereby several enzymes are indeed employed. Nowadays whole cell systems are already used as biocatalyst but mainly employing only one of the enzymes present. Often these enzymes cannot be isolated or are deactivated upon isolation. Further challenges are new techniques for formulation of enzymes and enzyme systems into stable, robust and efficient biocatalysts.

Q.

E o o O) O

Joint Future

Biosyntheses Direct Fermentation Precursor Fermentation Bio-oxidation BiotrMigformations fwcatalysis Ctiemocatalysis Stoichiometric synthesis

Chemistry (after J.IVI. Lehn) Figure 2 SynthesisfromChemical and Biological Perspective

Diversity

25

Single biocatalytic conversions and bio-transformations using enzyme combinations should become a synthetic continuum with precursor fermentations and direct fermentations. Insight in interaction of enzymes with their environment (i.e. membranes) will hereby be needed. Methods to tune enzyme kinetics will be required. Artificial co-factor regenerating systems will have to be developed as well as new ISPR methods. Bio-inspired Organic Synthesis Many organic syntheses are already inspired by nature. However, the complexity of biology has forced organic chemistry to very inaccurate translations of enzyme systems into man-made catalysts (i.e. catalysts with a simple molecular structure and mol. weight below 500 vs. enzymes with highly intricate structures and mol. weights up to 500.000). The present bio-inspired trend in organic synthesis towards macromolecular systems is meeting the advancement in molecular biology at the same level. This should lead to joint design of new multifunctional (bio-)catalyst systems, which can either be used in a modified metabolic path for a fermentation process or as an efficient catalyst in a series of organic syntheses. Bio-mimetic catalyst systems will be made available designed on growing knowledge of metabolic pathways and detailed insight in bio-recognition phenomena. Biomolecular informatics will provide guidelines for the design of new and robust catalysts. When combined with mechanistic know-how of chemocatalysis 'de novo' enzyme design comes within the realm of current chemistry. Combination with directed evolution methods would be another way to new catalyst design. High selectivity for a specific target molecule can be reached. Better bio-, chemo- and hybride-catalyst formulations for a wide range of (bio-) syntheses will be the result utilizing enzymes, metals and a range of dedicated ligands. Increased knowledge of active site structures, effect of protein modifications, functional insight in enzyme systems, cell compartments and complete cells will be at the basis of these new developments together with the mechanistic insights from chemocatalysis. Specific challenges in synthetic organic chemistry are: _ direct functionalization of aromatic compounds; for instance replacing Friedel Crafts type chemistry by direct arene alkylation/acylation using olefins; _ cross coupling reactions, which play a prominent role in current synthetic repertoire, based on olefins; _ (bio-)catalytic reductions and oxidations; _ synthetic conversions without protective groups; _ catalytic methodology in heterocyclic chemistry. Numerous bioactive products in particular pharmaceuticals and agrochemicals are based on multifunctional heterocyclic compounds. Hardly any of the current catalytic methods can be employed for heterocyclic substrates due to rapid catalyst poisoning. Another major challenge is the reduction of the number of steps in common multi-step synthesis. The combination of mutually depending bio- and chemocatalysis is only one

26 of the possibilities to develop synthetic methods not depending on exhaustive protective group manipulation. Eventually multi-step, once-through processes will evolve from exceptions today to common methodology tomorrow. Combining combinatorial approaches with understand, design and build methods will be an additional challenge in reaching these new synthesis methodologies.

Acknowledgement l.This essay is the introduction of a new national Dutch research program aiming at integration of organic synthesis and biosynthesis in a joint effort of academia and the Dutch life sciences industries. The program will be organized by the Council for Chemical Sciences of NWO, the national Dutch organization for science and technology. 2.This essay has been used as a guideline for the panel discussion on "Green Chemistry" at the 13* Noordwijkerhout-Camerino Symposium on "Trends in Drug Research" on May 6-11,2001 in The Netherlands.

Further reading 1. P.S. Zurer, "Annulation Strategies (cascade reactions)", Chem. Eng. News, 79 (2001)27-30. 2. "New Voices in Chemistry", Chem. Eng. News, 79 (2001) 51-291. 3. A.I. Scott, "Towards a Total, Genetically-Engineered Synthesis of Vitamin-B12", Synlett. (1994), 871-883 4 . R. Schoevaart, F. van Rantwijk and R.A. Sheldon, "Class I fructose-1,6bisphosphate aldolases as catalysts for asymmetric aldol reactions". Tetrahedron Asymmetry, 10 (1999), 705-711. 5. R.A. Sheldon and H. van Bekkum, "Future Outlook", in "Fine Chemicals through Heterogeneous Catalysis", pg. 589-592, Wiley-VCH, 2001. 6. B. Zwanenburg (ed.), "Enzymes in Action, Green Solutions for Chemical Problems", Kluwer Academic Publishers, 2000. 7. A. Bruggink (ed.), "Synthesis of P-lactam Antibiotics, Chemistry, Biocatalysis and Process Integration", Kluwer Academic Publishers, 2001. 8. J.J. Heijnen, "Microorganisms as Micro-Chemical Factories for Sustainable Precision Production of Chemicals" (Conference Report), pg. 69-70, Gratama Workshop 2000, Osaka, Japan . 9. H.C.Kolb, M.G.Finn and K.B.Sharpless, "Click Chemistry: Diverse Chemical Functions from a Few Good Reactions", Angew. Chem. Intern. Edit., 40 (2001),2004-2021.

H, van der Goot (Editor) Trends in Drug Research III © 2002 Elsevier Science B.V. All rights reserved

27

Directed Evolution of Enantioselective Enzymes as Catalysts in the Production of Chiral Pharmaceuticals Manfred T. Reetz Max-Planck-Institut fur Kohlenforschung Kaiser-Wilhelm-Platz 1 45470 Mtilheim an der Ruhr, Germany

1

Introduction

The development of methods for the stereoselective synthesis of chiral organic compounds is of enormous academic and industrial interest [1-3]. Indeed, a major portion of research in organic chemistry during the last 30 years has been devoted to asymmetric synthesis. Such activities are certainly driven in part by the need to prepare chiral therapeutic drugs in both enantiomerically pure forms. In fact, the so-called "chiral market" of enantiomerically pure or enriched organic compounds is expanding rapidly, total sales of chiral pharmaceuticals alone exceeding 100 billion $ US in 2000 [le]. Not all, but certainly many of these compounds are prepared in the laboratories of organic chemists. Presently, classical antipode separation is used most often in industry [Id]. However, this requires stoichiometric amounts of an appropriate optically active reagent as well as large amounts of organic solvents. Due to ecological and economic reasons, asymmetric catalysis can be expected to be more efficient, provided that active and highly enantioselective catalysts can be found. Catalytic asymmetric transformations can be carried out either in the form of kinetic resolution of racemates or in reactions involving prochiral substrates. Two options are available, namely transition metal catalysis [2] or biocatalysis [3]. Success in the area of asymmetric transition metal catalysis entails a difficult goal [2], namely efficient ligand tuning (Fig. la), which in turn requires experience, intuition, knowledge of the reaction mechanism and the ability to apply molecular modeling as well as time-consuming trial and error. Numerous successful examples have been reported in the ongoing academic literature, and a few of these have been commercialized [1].

a)

I >D.

!'9^"^ ^. tuning b)

C l^W ^— site specific ^ ^ | J ^ mutagenesis near active site

VD'

Fig. 1 a) Schematic representation of ligand tuning in the design and synthesis of a chiral transition metal (M) catalyst, C2-symmetry arbitrarily being shown; the arrow symbolizes points of potential structural variation and D denote donor atoms, b) Schematic representation of "de novo design" of an enantioselective enzyme, the arrow symbolizing the exchange of amino acids on the basis of site specific mutagenesis.

28

In the general area of biocatalysis, enzymes [3] and catalytic antibodies constitute [4] the most important options. Since enzymes are inherently the more active catalysts, they have been used most often. Indeed, a number of industrial processes for the enantioselective synthesis of intermediates needed in the production of chiral pharmaceuticals are based on the application of enzymes. A prominent example is the lipase-catalyzed kinetic resolution of an epoxy-ester used in the production of the antihypertensive therapeutic Dilthiazem® [5]. There seems to be a trend in industry to use enzymes more often than in the past [6]. However, these catalysts suffer from the disadvantage that for a given synthetic transformation of interest. A—>B, enantioselectivity may well be poor. In principle, it should be possible to apply sitedirected mutagenesis [7] in order to increase enantioselectivity to an acceptable level (Fig. lb), similar to ligand tuning in transition metal catalysis (Fig. la). However, this has not turned out to be a straightforward process. We have recently proposed a radically different approach to the development of enantioselective catalysts which does not rely on any knowledge regarding the structure of the enzyme nor on any speculations concerning enzyme mechanism [8]. The combination of proper molecular biological methods for random mutagenesis and gene expression coupled with high-throughput screening systems for the rapid identification of enantioselective variants of the natural (wild-tpye) enzyme forms the basis of the concept [8-10]. The idea is to start with a wild-type enzyme showing an unacceptably low enantiomeric excess (ee) or selectivity factor (E) value for a given transformation of interest. A—>B, to create a library of mutant genes, to identify the most enantioselective enzyme-variant following expression, and to repeat the process as often as necessary using in each case an improved mutant gene for the next round of mutagenesis. Due to the fact that the inferior mutant genes and enzyme variants are discarded, an evolutionary "pressure" in the overall process builds up (Fig. 2) [8-10].

gene (DNA)

I

wild-type enzyme

random mutagenesis

library of mutated genes repeat

I expression

\

' 9 w

W"^'^-

library of enzyme-variants

I

screening (or selection) for enantioselectivity

positive mutants Fig. 2 Directed evolution of an enantioselective enzyme [9]

29 During the last 15 years molecular biologists have developed new and practical techniques for random mutagenesis. For example, Leung, Chen and Goedell described the technique of error prone polymerase chain reaction (epPCR) in which the conditions of the classical PCR are varied empirically (e.g., the MgC^ concentration) so as to attain the desired mutation rate [Ua]. Later the method was improved by Cad well and Joyce [lib]. This procedure of inducing point mutations was followed in 1994 by Stemmer's method of DNA shuffling [12] and in 1998 and 1999 by Arnold's staggered extension process [13] and random priming recombination method [14], respectively, which are all recombinative processes resulting in a high diversity of mutant genes. Since then these and other methods such as saturation mutagenesis (in which the substitution or insertion of codons is performed leading to all possible 20 amino acids at any predetermined position in the gene) or cassette mutagenesis (using DNAfragments made up of nucleotides encoding one to several hundred amino acids in a defined region of the enzyme) have been applied in the quest to obtain structurally altered enzymes with improved stability and activity [11-18]. However, enantioselectivity is a difficult parameter, and at the outset of our efforts it was not clear whether the tools of directed evolution can be applied successfully in the area. The major challenges in putting the concept described in Figure 2 into practice involve the development of efficient strategies for exploring protein sequence space with respect to enantioselectivity and the establishment of high-throughput screening or selection methods for assaying enantioselective enzymes. In this chapter we summarize the current status of this new and exciting branch of asymmetric catalysis [9]. 2

High-Throughput Screening Systems for Enantioselectivity

The determination of ee values is traditionally performed by gas chromatography or HPLC using chiral phases, but only a few dozen samples can be analyzed per day. When we initiated research in 1995 concerning the evolution of enantioselective enzymes (see below) [8], high-throughput assays for enantioselectivity were unknown. Several methods have been developed since then, including systems based on UVAîs [8, 19-21], IR-thermography [22], MS [23], and capillary array electrophoresis [24]. Since a complete review concerning the scope and limitations of these and other methods has appeared recently [10], only a few highlights as well as new developments are mentioned here. A very practical assay is a method based on electrospray ionization mass spectrometry (ESI-MS) [23]. The (/?)- and (5)- enantiomers of a given chiral product have identical mass spectra and, in the absence of chromatographic separation, cannot be distinguished. However, if one of the enantiomers is deuterium-labeled, the parent peaks appear separately in the mass spectrum of the mixture, and integration then provides the ee value. Accordingly, deuterium-labeled substrates in the form of pseudoenantiomers or pseudo-prochiral compounds are used to test a potentially enantioselective (bio)catalyst. The method is restricted to the kinetic resolution of chiral compounds and to reactions of prochiral compounds having enantiotopic groups. The system has been automated, allowing for about 1000 ee determinations per day [23]. The assay has been applied in several cases [10,23]. A new example involves the hydrolytic kinetic resolution of the epoxide 1 catalyzed by an epoxide hydrolase (Fig. 3) [25]. The chiral diol 2 is of considerable interest in the pharmaceutical industry. As in other kinetic resolutions, the reacton is allowed to reach the ideal value of 50 %.

30

Instead of employing a genuine racemate (J?)-l/(5)-l as in a normal lab-scale experiment, a 1 : 1 mixture of pseudo-enantiomers (/?)-l/(5)-(D5)-l needs to be used because at any point of the reaction the ratio of (/?)-! : (5)-(D5)-l and (/?)-2 : (S)'(Ds)-2 can be determined simply by integrating the appropriate ESI-MS peaks. This delivers the enantiomeric excess (ee) as well as the selectivity factor (£), provided the appropriate "time window" is used. If a given library of enzyme-variants contains very active as well as less active (or even inactive) enzymes, a rough pre-screening test is advisable. Several MS measurements as the reaction progresses provides time resolution if necessary.

3/^'

Ô

0

^\

H2O ^. epoxide hydrolase

(fl)-i HO

OH

^

(S)-(D5)-1

o

HO

OH

D

D^

D

\ D

(S)-(D5)-2

Fig. 3 Kinetic resolution of pseudo-enantiomers (/?)-l/(5)-(D5)-l [25] If reactions involve desynmietrization of meso-type substrates, kinetic resolution is not involved, which means that the transformation can be run to 100 % conversion [10,23], and the "time window" is of no concern. The use of the ESI-MS-based assay in such cases requires the synthesis and application of pseudo-meso substrates, i.e., mesosubstrates that contain deuterium-labeled enantiotopic groups. In sunmiary, although the MS-based assay is restricted to the two symmetry classes as delineated here, it is highly efficient. Moreover, new MS instruments utilizing an 8-channel multiplex electrospray source allow for even higher throughput, which means that about 8000 ee determinations can be performed in one day. An alternative and also practical MSbased assay requires derivatization using chiral reagents [26]. Other options include high-throughput e^-assays based on capillary array electrophoresis [24], color tests [8,19-21,27], circular dichroism [28] and DNA microarrays [29]. 3

First Example of Directed Evolution of an Enantioselective Enzyme

We initially studied the kinetic resolution of the lipase-catalyzed hydrolysis of the chiral ester 3 in which a maximum of 50 % conversion is aimed for [8]. Lipases are enzymes

31 that catalyze the hydrolysis of esters [5], the reverse reaction in organic solvents also being possible. The particular enzyme used in our case was the bacterial lipase from Pseudomonas aeruginosa, which showed an ee-value of only 5 % in favor of the (S)acid 4 at 50 % conversion.

\^

/

p. P.aeruginosa aeruginosa

rac-3 (R = n-CsH^ 7)

11

^..

(S)-4

.

r"r

OH

(^-4

Fig. 4 Lipase-catalyzed hydrolytic kinetic resolution

The first step in directed evolution is the consideration of the mutation rate, which has to do with the problem of exploring protein sequence space. The lipase from Pseudomonas aeruginosa has 285 amino acids. Complete randomization would result in 20^^^ different enzyme-variants, which is more than the mass of the universe, even if only one molecule of each enzyme were to be produced [8,9]. The other extreme entails the minimum amount of structural change, namely the substitution of a single amino acid per molecule of enzyme by one of the other 19 naturally occurring amino acids. In this case, on the basis of the algorithm A^= 19M-285!/[(285-Af)! -M!], where M = number of amino acid substitutions per enzyme molecule, the library of variants would theoretically have 5415 members [8, 9]. However, when using epPCR as the random mutagenesis method, a library of 5000 - 6000 members is not expected to contain all theoretically possible permutations. This is because the genetic code is degenerate. If two amino acids are exchanged per enzyme molecule (M = 2), then the number of enzyme-variants increases dramatically (about 14 million!). In the case of M = 3, it is more than 52 billion. We therefore initially chose a low mutation rate so as to induce an average of only one amino acid exchange per enzyme molecule [8]. Thus, in the case of the kinetic resolution of the ester 3, epPCR was adjusted to cause about 1 - 2 base substitutions per 1000 base pairs of the gene, resulting in an average of one amino acid exchange. Typically, 2000 - 3000 enzyme-variants per generation were screened. Following expression in E. colilP, aeruginosa, a screening system based on the UVAîs absorption of the liberated p-nitrophenolate at 410 nm was employed. As a consequence of the first round of mutagenesis and screening, a variant displaying an ^^-value of 31 % was identified (£ = 2.1). The corresponding mutant gene was then subjected once more to mutagenesis, and the process was repeated several times. The results after four generations of mutants led to an ee-value of 81 %, the selectivity factor being £ = 11.3 (Fig. 5).

32

81%^.»« (E=11.3)

1

2 3 mutant generations

Fig. 5 Increasing the enantioselectivity of the lipase-catalyzed hydrolysis of the model ester 3 These remarkable results constitute proof of principle. Nevertheless, a selectivity factor of £ = 11.3 cannot be viewed as industrially viable. Thus, a fifth round of mutagenesis was performed, and indeed the usual library of about 2000 mutants contained slightly improved variants. In spite of this advancement it became clear to us that we needed to develop methods which allow for even more efficient ways to explore protein sequence space with respect to enantioselectivity [9, 30]. Accordingly, DNA-analyses leading to amino acid sequence determinations of the variants were carried out as a first step. For example, the best mutants of the first four generations turned out to have the following amino acid substitutions (Fig. 6). Variant 01E4(E= 2.1):

Seri

Variant 08H3(E = 4.4):

Glyi49 Leui55

Variant 13D10{E= 9.4): Ser^g Seri55 Val47 Variant 04H3(E= 11.3):

Glyi49

Giyi49 Leui55 Gly47 Glyi49 Leui55 Gly47 LeU259

Fig. 6 Data of amino acid exchanges in the best lipase-variants of the first four generations [9, 30]

33 We then drew the following conclusions [9, 30]: 1. The process of random mutagenesis/screening identifies sensitive positions ("hot spots") in the enzyme which are responsible for improved enantioselectivity. 2. Such positions are likely to be correct, but the particular amino acid identified may not be optimal. 3. Saturation mutagenesis at the "hot spots" can be expected to generate improved mutants. Rather than continuing with epPCR in further cycles of random mutagenesis, we decided to utilize appropriate combinations of various types of mutagenesis. Saturation mutagenesis is a molecular biological method with which mutations at a given position of an enzyme can be introduced, a small library of only 300 - 400 variants being necessary to ensure that all of the remaining 19 amino acids have been introduced. Upon applying this strategy at one of the hot spots (e.g., at position 155), it was discovered that phenylalanine (F) is the best amino acid . Saturation mutagenesis using the best gene in the third generation led to the identification of a variant which showed a selectivity factor of ^ = 20, phenylalanine again "showing up" as the best amino acid at position 155. Thereafter, epPCR was applied again, which resulted in E= 25! Clearly, the combination of mutagenesis methods, namely epPCR and saturation mutagenesis, constitutes an efficient method to explore protein sequence space with respect to enantioselectivity. Thus, a small family of enzymes was created, all showing £-values of 20 - 25 and £^-values of 88 - 91 % for the model reaction [9, 30]. Following these developments, recombinant methods such as DNA-shuffling in combination with methods based on point mutations, were applied. Specifically, Combinatorial Multiple Cassette Mutagenesis using two genes obtained at high mutation rate and an oligo-cassette at two "hot spots" led to a selectivity factor of E>51 (^^>95%) [31]. Moreover, it was possible to invert the sense of enantioselectivity [9, 32]. The present results are summarized in Figure 7. (S)-selectivJty E>51 (ee> 95 %)

4o4 304

E=20-25

2o4

cassette! mutagenesis pointl mutations

1o4

l4

^ ^ = 3 4 pQjptj mutations and DNAi shuffling

wild"lype

1o4

• point| mutations and DNAi shuffling

20^ (/lO-selectlvity

E=20

©

Fig. 7 Optional (S)- or (/?)-selectivity in the lipase-catalyzed hydrolysis of ester 3 [9,30,31]

34

4

Extending the Flow of Genetic Information from DNA to Transition Metal Catalysts

Following our initial reports concerning the directed evolution of enantioselective lipase-variants, further examples of the underlying principle have been reported [33]. Thus, the general concept schematized in Figure 2 may well turn out to be general. Moreover, directed evolution of enantioselective enzymes provides a unique opportunity to study structure/selectivity relations, provided the crystal structures of the newly evolved enzyme-variants become available. Parallel to these efforts we have started to develop a concept which goes far beyond directed evolution of improved enzymes. Accordingly, we are extending the flow of genetic information from the gene (DNA) to transition metal catalysts (Fig. 8) [34]. riMA transcription^ _ . , . translation^ _ chemical DNA l i — • RNA ^^^j^^-——-^

^ i-«,„^Î f^^M Enzyme I H ^ L ^ M

Fig. 8 The flow of genetic information from DNA to transition metal catalysts (L = ligand; M = transition metal) A wild-type enzyme is chosen having a cavity large enough for potential substrate binding. It should contain a "reactive" amino acid residue (e.g., cystein) at the cavity so that chemical modification with introduction of ligands (phosphines, nitrogen-moieties, etc.) capable of binding transition metals such as palladium, rhodium, nickel, etc. Such catalysts have enzyme-like structures and therefore appropriate binding properties, yet the actual transformations are not "enzymatic" in the traditional sense because transition metal catalyzed reactions such as hydrogenation, hydroformylation, oxidation, allylic substitution, cycloadditions, etc. are the focus of interest. We have shown that such implantation of catalytically active transition metal sites is possible [34]. The next step is to perform mutagenesis on the wild-type gene as described above in the directed evolution of enzymes, to express the enzyme-variants and finally to introduce ligands/metals at defined positions for thousands of variants using robotics before testing traditional transition metal catalyzed reactions en masse employing the ê-screening systems previously developed [10]. Not only enantioselectivity, but also activity as well as chemo- and regioselectivity can then be optimized.

5

Conclusion

Directed evolution of enantioselective enzymes is emerging as a new and fascinating area of research. Two major problems have already been dealt with successfully, namely the development of strategies for efficiently exploring protein sequence space with respect to enantioselectivity, and the establishment of high-throughput eescreening systems. Nevertheless, more efforts are necessary before generality can be claimed, including the study of a wide variety of enzymes and substrates. It can be predicted that directed evolution will be used successfully to create novel enzymevariants which are highly enantioselective and active as well as stable enough to allow for a variety of applications in organic and/or pharmaceutical chemistry. This applies all the more to the idea of conceptually fusing molecular biology with transition metal chemistry in order to evolve completely new types of catalysts.

35 References [1]

[2]

[3]

[4]

[5] [6]

[7]

a) A.N. Collins, G.N. Sheldrake, J. Crosby, (Eds.), Chirality in Industry: The Commercial Manufacture and Applications of Optically Active Compounds, Wiley, Chichester, 1992; b) A.N. Collins, G.N. Sheldrake, J. Crosby, (Eds.), Chirality in Industry II: Developments in the Commercial Manufacture and Applications of Optically Active Compounds, Wiley, Chichester, 1997; c) R.A. Sheldon, Chirotechnology: Industrial Synthesis of Optically Active Compounds, Dekker, New York, 1993; d) S.C. Stinson, Chem, Eng. News 1999, 77(41), 101120; e) S.C. Stinson, Chem. Eng. News 2000, 75(43), 55-78. a) E.N. Jacobsen, A. Pfaltz, H. Yamamoto, (Eds. j , Comprehensive Asymmetric Catalysis, Vol. Mil, Springer, Berlin, 1999; b) H. Brunner, W. Zettlmeier, Handbook of Enantioselective Catalysis with Transition Metal Compounds, Vol. I-II, VCH, Weinheim, 1993; c) R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, New York, 1994; d) I. Ojima, (Ed.), Catalytic Asymmetric Synthesis, VCH, Weinheim, 1993; e) D.J. Berrisford, C. Bolm, K.B. Sharpless, Angew. Chem. 1995, 107, 1159-1171; Angew. Chem. Int. Ed. Engl. 1995, 34, 1059-1070. a) H.G. Davies, R.H. Green, D.R. Kelly, S.M. Roberts, Biotransformations in Preparative Organic Chemistry: The Use of Isolated Enzymes and Whole Cell Systems in Synthesis, Academic Press, London, 1989; b) C.H. Wong, G.M. Whitesides, Enzymes in Synthetic Organic Chemistry (Tetrahedron Organic Chemistry Series, Vol. 12), Pergamon, Oxford, 1994; c) K. Drauz, H. Waldmann (Eds.), Enzyme Catalysis in Organic Synthesis: A Comprehensive Handbook, Vol. MI, VCH, Weinheim, 1995; d) K. Faber, Biotransformations in Organic Chemistry, 3rd ed.. Springer, Berlin, 1997; e) G. Carrea, S. Riva, Angew. Chem. 2000, 112, 2312-2341; Angew. Chem. Int. Ed. 2000, 39, 22262254; f) K.M. Koeller, C.-H. Wong, Nature (London) 2001,409, 232-240. a) P.G. Schultz, R.A. Lerner, Science (Washington, D.C.) 1995, 269, 18351842: b) J.-L. Reymond, Top. Curr. Chem. 1999, 200, 59-93; D. Hilvert, Top. Stereochem. 1999,22, 85-135. R.D. Schmid, R. Verger, Angew. Chem. 1998, 110, 1694-1720; Angew. Chem. Int. Ed. 199S, 37, 1608A633. a) M. McCoy, Chem. Eng. News 1999, 77(1), 10-14; b) R.S. Rogers, Chem. Eng. News 1999, 77(29), 87-91; c) B. Schulze, M.G. Wubbolts, Curr. Opin. Biotechnol. 1999, 10, 609-615; d) A. Schmid, J.S. Dordick, B. Hauer, A. Kiener, M. Wubbolts, B. Witholt, Nature (London) 2001,409, 258-268. a) M.M. Altamirano, J.M. Blackburn, C. Aguayo, A.R. Fersht, Nature (London) 2000, 403, 617-622; b) P.N. Bryan, Biotechnol. Adv. 1987, 5, 221-234; c) J.A. Gerlt, Chem. Rev. 1987, 87, 1079-1105; d) J.R. Knowles, Science (Washington, D.C.) 1987, 236, 1252-1258; e) S.J. Benkovic, C.A. Fierke, A.M. Naylor, Science (Washington, D.C.) 1988, 239, 1105-1110; f) J.A. Wells, D.A. Estell, Trends Biochem. Sci. 1988,13, 291-297; g) Y. Hirose, K. Kariya, Y. Nakanishi, Y. Kurono, K. Achiwa, Tetrahedron Lett. 1995,36, 1063-1066; h) Z. Shao, F.H. Arnold, Curr. Opin. Struct. Biol. 1996, 6, 513-518; i) C. Heiss, M. Laivenieks, J.G. Zeikus, R.S. Phillips, / Am. Chem. Soc. 2001,123, 345-346; j) S. Patkar, J. Vind, E. Kelstrup, M.W. Christensen, A. Svendsen, K. Borch, O. Kirk, Chem. Phys. Lipids 1998, 93, 95-101; k) M. Holmquist, P. Berglund, Org. Lett. 1999,

36

[8] [9] [10]

[11] [12] [13] [ 14] [15] [16] [17] [18] [19] [20]

[21]

[22] [23] [24] [25] [26] [27] [28]

[29] [30] [31]

/, 763-765; 1) P. Kast, D. Hilvert, Cum Opin. Struct Biol 1997, 7,470-479; m) J.P. Schanstra, A. Ridder, J. Kingma, D.B. Janssen, Protein Eng. 1997, 10, 5361; n) M.H.J. Cordes, N.P. Walsh, C.J. McKnight, R.T. Sauer, Science (Washington, D.C) 1999,284, 325-327. M.T. Reetz, A. Zonta, K. Schimossek, K. Liebeton, K.-E. Jaeger, Angew, Chem. 1997,109,2961-2963; Angew. Chem, Int. Ed. Engl. 1997,36,2830-2832. M.T. Reetz, K.-E. Jaeger, Chem.^-Eur. J. 2000,6,407-412. Review of high-throughput e^-screening systems in the directed evolution of enzymes and in asymmetric transition metal catalysis: M.T. Reetz, Angew. Chem. 2001,113, 292-320; Angew. Chem. Int. Ed. 2001,40,284-310. a) D.W. Leung, E. Chen, D.V. Goeddel, Technique (Philadelphia) 1989, 1, 1115; b) R.C. Cadwell, G.F. Joyce, PCR Methods Appl. 1994,3, S136-S140. a) W.P.C. Stemmer, Nature (London) 1994, 370, 389-391; b) P.A. Patten, R.J. Howard, W.P.C. Stemmer, Curr. Opin. Biotechnol. 1997, 8, llA-m^. H. Zhao, L. Giver, Z. Shao, J.A. Affholter, F.H. Arnold, Nat. Biotechnol. 1998, 76,258-261. Z. Shao, H. Zhao, L. Giver, F. Arnold, Nucleic Acids Res. 1998,26, 681-683. a) F.H. Arnold, Ace. Chem. Res. 1998,31, 125-131; b) J. Affholter, F.H. Arnold, CHEMTECH1999,29, 34-39. M. Kanunann, J. Laufs, J. Schell, B. Gronenborn, Nucleic Acids Res. 1989, 17, 5405. A. Crameri, S.-A. Raillard, E. Bermudez, W.P.C. Stemmer, Nature (London) 1998, i97, 288-291. M.T. Reetz, K.-E. Jaeger, Top. Curr. Chem. 1999,200, 31-57. L.E. Janes, R.J. Kazlauskas, J. Org. Chem. 1997,62,4560-4561. a) L.E. Janes, A.C. Lowendahl, R. J. Kazlauskas, Chem.—Eur. J. 1998,4, 23242331; b) F. Moris-Varas, A. Shah, J. Aikens, N.P. Nadkarni, J.D. Rozzell, D.C. Demirjian, Bioorg. Med. Chem. 1999, 7,2183-2188. a) F. Zocher, M.M. Enzelberger, U.T. Bomscheuer, B. Hauer, R.D. Schmid, Anal. Chim. Acta 1999, 391, 345-351; b) E. Henke, U.T. Bomscheuer, Biol. Chem. 1999,380, 1029-1033. M.T. Reetz, M.H. Becker, K.M. Kuhling, A. Holzwarth, Angew. Chem. 1998, 110,2792-2795; Angew. Chem. Int. Ed. 1998,37, 2647-2650. M.T. Reetz, M.H. Becker, H.-W. Klein, D. Stockigt, Angew. Chem. 1999, 111, 1S72A875; Angew. Chem. Int. Ed. 1999,38,1758-1761. M.T. Reetz, K.M. KUhling, A. Deege, H. Hinrichs, D. Belder, Angew. Chem. 2000, 772,4049-4052; Angew. Chem. Int. Ed. 2000,39, 3891-3893. M.T. Reetz, A. Eipper, C. Torre, Y. Genzel, R. Furstoss, unpublished results. J. Guo, J. Wu, G. Siuzdak, M.G. Finn, Angew. Chem. 1999, 777, 1868-1871; Angew. Chem. Int. Ed. 1999,38,1755-1758. D.C. Demirjian, P.C. Shah, F. Moris-Varas, Top. Curr. Chem. 1999,200, 1-29. a) K. Ding, A. Ishii, K. Mikami, Angew. Chem. 1999, 777, 519-523; Angew. Chem. Int. Ed. 1999, JS, 497-501; b) K. Mikami, R. Angelaud, K. Ding, A. Ishii, A. Tanaka, N. Sawada, K. Kudo, M. Senda, Chem. Eur. J. 2001, 7, 730737; c) M.T. Reetz, K.M. Kuhling, H. Hinrichs, A. Deege, Chirality, 2000, 72, 479-482. G.A. Korbel, G. Lalic, M.D. Shair, /. Am. Chem. Soc. 2001, 72i, 361-362. K. Liebeton, A. Zonta, K. Schimossek, M. Nardini, D. Lang, B.W. Dijkstra, M.T. Reetz, K.-E. Jaeger, Chem. Biol. 2000, 7,709-718. M.T. Reetz, S. Wilensek, D. Zha, K.-E. Jaeger, submitted.

37 [32] [33]

[34]

K.-E. Jaeger, T. Eggert, A. Eipper, M.T. Reetz, Appl. Microbiol. Biotechnol 2001, in press. a) U.T. Borascheuer, J. Altenbuchner, H.H. Meyer, Biotechnol Bioeng. 1998, 58, 554-559; b) O. May, P.T. Nguyen, F.H. Arnold, Nat Biotechnol 2000, 18, 317-320; c) S. Pong, T.D. Machajewski, C.C. Mak, C.H. Wong, Chem, Biol 2000, 7, 873-883. M.T. Reetz, P. Maiwald, P. Wedemann, unpublished.



39

AT THE INTERFACE OF ORGANIC SYNTHESIS AND BIOSYNTHESIS ROB SCHOEVAART^ and TOM KIEBOOM^'^ Îndustrial Fermentative Chemistry, Leiden Institute of Chemistry, Leiden University, Leiden, Netherlands; ^DSM Food Specialties R&D, Delft, Netherlands

Organic synthesis^ i.e. chemistry by mankind, uses traditionally often a step-by-step approach to convert a starting material A into a final product D, in which intermediate products B and C have to be isolated and purified for each next conversion step: Conversion steps

Recovery steps

Such multi-step organic syntheses are also still quite common in to-day's fine chemical industry, i.e.: • often carried out in a non-catalytic way by using relatively large amounts of reagents, resulting in the production of multi-kg of waste per kg of final product; • requiring separation and purification steps after each conversion step in order to be able to do the next conversion, resulting in the production of heat waste due to the consumption of the appropriate amount of energy; • using extra energy to overcome eventual thermodynamic hurdles, i.e. to reach the final product D in case of higher energetic intermediates C and/or D.

40

Biosynthesisy i.e, chemistry by nature, in the cells of living organisms, goes through a multi-step cascade approach to convert a starting material A into the final product D without separation of intermediates B and C: Conversion steps

Such multi-step combined syntheses are quite common in every-day's life, i.e.: • carried out in a full-catalytic way by using enzymes with relatively limited amounts of reagents (cofactors) and so producing much less waste; • without intermediate recovery steps as possible by the mutual compatibility and high selectivity of the enzymatic conversions; • saving energy to overcome eventual thermodynamic hurdles, i.e. to reach the final product D in case of higher energetic intermediates B and/or C. For the next generation of organic synthesis^ it is the challenge to: • combine the power of the chemo-catalytic, enzymatic and microbial conversions; • search for multi-step conversions without recovery steps like nature does, i,e, to go for one-pot multi-catalytic procedures; • fine-tune reaction conditions and catalytic systems in order to allow for the right concerted cooperation without any intermediate isolation and/or purification steps. This to drastically diminish energy and waste, i.e. costs, of most of the present-day multi-step conversion processes of the fine chemical industry. For the next generation of biosynthesis., i.e. microbial conversions, it is the challenge to: • improve the productivity in terms of carbon source efficiency of micro-organisms, i.e. towards a much higher productîomass ratio; • broaden the scope of products, i.e. from natural towards modified products often required for drugs;

41

•

re-engineer metabolic pathways as most efficient way to reach these two goals; This to free fermentation processes from excessive use of renewables that end up in undesired, low-value, biomass as by-product as well as to minimize additional chemical modification steps to obtain the final product. The above mentioned challenges for both organic synthesis and biosynthesis are by no way kind of wishful thinking. This will be demonstrated by the evolution during the past 30 years of the industrial process route for the synthesis of the antibiotic Cephalexin within DSM^ Here, the integration of chemistry, enzymatic conversion and fermentation forms the basis of an efficient production method from both an economic and environmental point of view (Figure 1). Figure 1. Integration of bio- and organic synthesis in modem industrial Cephalexin making.

D-Phenylglycine

CHEMISTRY ENZYME

7-ADCA

Fe^MENTATION

The 1970's Cephalexin synthesis consisted of one fermentation (step 3) together with eight organic synthesis steps^: 1. Benzaldehyde -^ D-Phenylglycine 2. [1] "> D-Phenylglycine chloride 3. Sugar + Phenylacetic acid -> Penicillin-G 4. [3] ^ Sulfoxide of Penicillin-G 5. [4] -> Trimethylsilyl ester of Penicillin-G sulfoxide

42

6. [5] "> Ring enlargement towards full protected 7-ADCA moiety 7. [6] ^ Hydrolysis to 7-ADCA 8. [2] + [7] ^ Cephalexin This multi-step sequence used high levels of energy (e.g. low temperature conversions 5-7, many recovery and solvent recycle steps), of reagents and organic solvents (steps 2 and 4-8) and of phenylacetic acid (as no recycling was possible after the chemical hydrolysis step 7). The 2000's Cephalexin synthesis consists of one organic synthesis (step 1), together with one fermentation (step 2) and two enzymatic steps: 1. Benzaldehyde -> D-Phenylglycine amide 2. Sugar + Adipic acid ^ Adipic amide of 7-ADCA 3. [2] -> 7-ADCA + Adipic acid (the latter recycled for use in step 2) 4. [1] + [3]-^ Cephalexin This sequence avoids low temperature conversions, major part of reagents and organic solvents, and consumption of adipic acid by: • enzymatic coupling of D-phenylglycine amide with 7-ADCA in water (avoiding the synthesis of acid chloride and use of reagents and organic solvents); • metabolic pathway engineered micro-organism (GMO) that is able to ferment sugar directly into the desired 7-ADCA moiety (thus avoiding 3 chemical conversion steps) ^; • aqueous enzymatic hydrolysis of the adipyl side chain that easily allows recycling of the adipic acid to be used again in the fermentation step (Figure 2). The list of reagents (often used in more than stoichiometric amounts) and solvents (that could not be recycled completely) that are skipped by the novel cephalexin synthesis are^: Peracetic acid Trimethylsilyl chloride and Bis(trimethylsilyl)urea Pyridine.HBr and Trimethylamine.HCl Phosphorpentachloride Dichloromethane Butanol Phenylacetic acid Acid chloride HCl salt of D-phenylglycine

43

SUGAR

Figure 2. Fermentative 7-ADC A production by GMO including enzymatic recycling of adipic acid\ Apart from such a sequential cooperation of chemical, enzymatic and microbial conversions on an industrial multi-tons scale, a number of combined multi-step one-pot catalytic conversions have been described on both lab- and industrial scale, using either a combination of enzymes or a combination of an enzyme and a chemo-catalyst (Table 1). Table 1. Combined Catalytic One-pot Conversions Conversion D-Glucose -> D-Fructose -> D-Mannitol Glycerol - » - > ^ ^ D-Heptulose

Benzaldehyde -> (S)-Mandelonitrile acetate j Acetoacetate ester + D-Glucose -» (R)-3-hydroxybutanoate ester + D-Gluconate

Reaction steps & Conditions Hydrogenation Isomerisation 70°C, 50 atm H2 Phosphorylation Oxidation Aldol reaction pH 4 •> 7.5 -> 4 Cyanohydrin form. Racemisation Acylation Oxidation Reduction Cofactor-regeneration Two-phase system Ambient conditions

ChemoCatalyst Cu/Si02

BioCatalyst Glucoseisomerase Phytase Oxidase+Catalase Aldolase

HOexchanger

Ref 3

4

Lipase 5

Glucose dehydrogenase Aldehyde dehydrogenase

6

44 D/L-Hydantoin -> D-Amino acid Alkyipyruvate + Formiate -> L-amino acid + CO2 a-Keto acids -> D-Amino acids

(R/S)-epoxide -> (S)-diols

Starch -> D-Fructose

Sucrose -> D-Fructooligosacchardes Unsaturated triacylglycerols -> Hydroperoxy fatty acids ^^C-formiate + glycine ^ 3-^^C-L-Serine D-glucose 6phosphate -> D-Gluconic acid 6phosphate (R/S)-Phenylethyiamine + EtOAc -> (S)-N-Acetylphenylethylamine 1 Cephalosporin C + Methyl tetrazolylacetate + 2-Mercapto-5methylthiadiazole -> Cefazolin 4-Methylcylcohexanone + Formiate -> (S)-4-Methylcaprolacton + C02

Racemisation Hydrolysis pH8.5,50°C Reductive amination Cofactor regeneration pH8,25°C Aminotransfer Redox reaction Racemisation

HO-

Enzymatic followed by acidic hydrolysis T30"^10°C pH 7.5^1 Hydrolysis Isomerisation Double-immobilized enzyme system pH=6,70°C Fructosyl transfer Oxidation pH=6,50°C Hydrolysis Peroxidation pH=9,26"C,02, octane/water Hydroxymethylation Cofactor regeneration Oxidation Ambient conditions

H"

Acylation Racemisation Ethyl acetate

Pd/C

Oxidation Amide hydrolysis Amine acylation Thio-ether formation pH=8">6.5 T25->4^65"C Oxidation Cofactor regeneration pH8,30^C

HO-

Hydantoinase 7

Amino acid + Formiate dehydrogenases Racemase Dehydrogenases D-Aminoacid aminotransferase Epoxide hydrolase

Glucoamylase Glucose isomerase

Fructosyltransferase Glucose oxidase Lipase Lipoxygenase

8

9

10

11

12

13

Serinehydroxymethyltransferase Dihydrofolate reductase Glucose 6phosphate dehydrogenase Lipase

14

15

Aminoacid oxidase Glutaryl acylase Penicillin G acylase Cyclohexanone mono-oxygenase Formate dehydrogenase

16

17

45 (R/S)-lphenylethanol + 4Chlorophenyl acetate -> (R)-l-Phenylethyl acetate+ 4Chlorophenol (R/S)-Allylic alcohols + AcOR-> (R)-Allylic esters (R/S)-2-Phenyl-3Acetoxycyclohexene -> (R)-2-Phenyl-3hydoxycyclohexene N-Ac-D-glucosamine + a-D-Glucose 1phosphate + Phospho-enolpyruvate -> N-Acetyllactosamine

Esterification Racemisation 20-70 °C, t-BuOH, cyclohexane

Sugars -> Complex Carbohydrates and Glycoconjugates

Glycosidic bond formation Epimerisation Phosphorylation Glycosyl transfer

Racemisation Esterification Organic solvent 100% cv/ee concept Ester hydrolysis Racemisation Disaccharide coupling Phosphorylation Epimerisation Ambient conditions

Ru catalyst

Lipase

Rh2(OAc)4 18

Ru catalyst

Lipase 19

PdCb

Lipase 20

Galactose transferase Phosphokinase UDP-Glucose pyrophosphorylase UDP-Galactose 4epimerase Transferases Phosphorylases Epimerases

21

22

Two major features are apparent from the data of Table 1: • By far, multi-step one-pot conversions have been reported in the field of carbohydrates using combinations of enzymatic 4,11,12,21,22

•

conversions ' ^ ^' ^ ^^^ ^ '^^; Combinations of metal- and bio-catalytic conversions are not yet that ^^^^^^3,15,18,19,20.

•

common ' ' ' ' ; After some early examples"^'^^ in the 1980's, there has been more than a decade of 'silence', followed by a clearly increasing interest during the past few years for combined catalytic conversions.

A very first example of the combined action of an enzyme and a metal catalyst is the direct one-pot conversion of glucose into mannitol, which is trice as expensive as glucitol:

46

Glucose

I^" Fructose

V Glucitol

Mannitol

Glucose isomerase on silica Copper on silica Hydrogen pressure of 70 atm Water, pH=7-8, at 70 ^C

Here, the isomerase enzyme converts glucose into a -1:1 glucose-fructose mixture and takes care that this mixture remains in equilibrium, while at the same time the copper catalyst hydrogenates preferentially fructose from this equilibrium into mannitol. This combi approach looks, at first sight, quite simple but in practice a number of fine-tuning measures had to be taken to achieve a balanced cooperation of the two simultaneous catalytic conversion steps, e.g.: • Immobilization of the enzyme onto silica to prevent poisoning of the copper metal by protein sulphur moieties; • Protection of the enzyme by a copper ion complexing agent (EDTA) to avoid inhibition by traces of copper ions from the copper catalyst; • Right compromise of hydrogen pressure and temperature to fulfil stability and activity requirements for both catalyst systems; • Slightly basic pH to avoid that mutarotation of glucose, i.e. the interconversion of a- to p-glucopyranose forms, becomes rate limiting as the enzyme only converts the a-form. What's really happening can be seen from the quite complicated kinetic and molecular picture (Figure 3) including three 'different' types of kinetics, expressed in TON's (sec"*), i.e.: • Michaelis-Menten (enzymatic isomerisation; only two of the six sugar forms are substrates for the enzyme; KM-values for glucose, fructose, glucitol and mannitol are 0.13, 0.04, 0.4 and > 1 M, respectively); • Langmuir-Hinshelwood (heterogeneous hydrogenation on copper; adsorption constants b vary from 3-10 M"*; only --25% of the copper surface covered with fructose that, however, reacts much faster than glucose adsorbed on the copper catalyst); • Homogeneous catalysis (acid/base catalysed mutarotation, i.e. interconversion of the different glucose and fructose forms; the rate for glucose is 50 times slower than that for fructose).

47

M U T A R 0 T A T I 0 N

OH

H - C - H HO--CH HO-CH HC - O H HC - O H H , C - OH 6 5 V.

mannitol

Figure 3. Facsimile^^ of molecular and kinetic picture of one-pot glucose to mannitol conversion: enzymatic isomerisation, mutarotation and copper catalysed hydrogenation.

48

The above mentioned principle of an equilibrium of two compounds of which one is selectively converted, together with the required fine-tuning of simultaneous catalytic conversion steps in one pot, is of great importance for the so-called 100% ee-100% yield synthesis of enantiomeric pure compounds from racemic starting materials (c/Table 1, r5,7,10,15,18,19,20y

(S)-A

(R)-A

(R)-B

Some recent examples on lab-scale have been reported^^'^^"^^ for the concomitant action of transition metal catalysts with lipases for the racemisation and esterification, respectively, of racemic 1-arylalcohols into high ee esters in high yield, e,g, ^^: CH3

v^

v^

K^

Rh2(OAc)4 Lipase Racemisation Esterification 70 ^C, cyclohexane There is no doubt that these kind synthesis of enantiomeric pure compounds with 100% e.e. and in 100% yield from relatively inexpensive racemates will find its way into the fine chemical industry in the near future (as proven, already, by the industrial hydantoinase process for amino acids, in which spontaneous racemisation occurs^)

Another elegant multi-step one-pot approach recently developed is ihQ four-enzyme catalysedfour-step one-pot conversion of glycerol into a heptose sugar derivative, in which a pH switch method is applied to temporary turn off phytase enzyme during the second and third steps of the concerted synthesis'*:

49

4 steps, one-pot

,

pyrophospha* Phytase

PO4 PO4

Phytase OPO3

OH FruA butanal

OPO3

I OH

Catalase

The four consecutive enzymatic conversion steps in one and the same reactor without any separation of intermediates consist of: • Phosphorylation: Glycerol is phosphorylated with pyrophosphate by phytase at pH 4.0 at 37 ^C. Racemic glycerol-3-phosphate is obtained in 100% yield (based on pyrophosphate) in 95% glycerol after 24h. • Oxidation: By raising the pH (to 7.5) phytase activity is "switched off, hydrolysis is prevented. Oxidation of L-glycerol-3-phosphate to DHAP by GPO at 55% glycerol (v/v) is quantitative. Catalase is added to suppress the build-up of hydrogen peroxide. The D-isomer is converted back to glycerol and phosphate in the last step. • Aldol reaction: More than twenty aldehydes are known to be substrates for the aldolases from S. carnosus and S. aureus. Stereoselectivity of the aldolases must be looked at for each acceptor substrate, since isomers are formed in different proportions. The oxidation and aldol reaction can be carried out simultaneously. • Dephosphorylation: Lowering the pH back to 4 "switches on" phytase's activity, hydrolysis of the aldol adduct is initiated. Combined with the broad substrate specificity of DHAP aldolases it constitutes a simple procedure for the synthesis of a wide variety of carbohydrates from readily available glycerol and pyrophosphate'*.

50

In conclusion, the concept and various examples given show that we may foresee a renaissance in synthesis methodology by integration of biO' and organic synthesis for fine chemicals by various approaches, from one-pot multi-step (bio)catalytic procedures towards metabolic engineered microbial transformations and combinations thereof^'*. In this respect, future clean synthesis methods should be inspired by the achievements in the field of modem detergent formulations that have up to six different enzymes in them^^, i.e. an advanced multi-catalytic onepot conversion of dirty laundry -> ^ - > - > - > -> clean laundry + dirt with the washing machine as in-house catalytic reactor that simultaneously separates the product (clean laundry) from waste (dirt). Finally, investigations of such multi-step synthesis methods without isolation of intermediate products require appropriate in situ analytical methods to know what's really happening during the consecutive conversions. Quite a powerful window to this information is the use of selectively isotope {e.g. ^^C, ^^N, ^Ô^ enriched starting materials in combination with NMR^^. In this way, a sequence of conversions can be well characterized, even in complicated matrices of catalysts, reagents and mixed solvent systems, e.g. the investigation of galactose oxidase mediated cross-linking phenomena of D-galactose protein mixtures^"^.

References 1. DSM Magazine 147 (1998) 18. 2. J. Verweij and E. de Vroom, Reel. Trav. Chim. Pays-Bas 112 (1993) 66. 3. M. Makkee, A.P.G. Kieboom, H. van Bekkum, and J.A. Roels, J. Chem. Soe., Chem. Commun. (1980) 930; M. Makkee, A.P.G. Kieboom, and H. van Bekkum, Carbohydr. Res. 138 (1985) 237. 4. R. Schoevaart, F. van Rantwijk, and R.A. Sheldon, J. Chem. Soc. Chem. Comm. (1999) 2465 and Tetrahedron: Asymmetry 10 (1999) 705. 5. J. Oda, J. Am. Chem. Soc. 113 (1991) 9360. 6. S. Shimizu; M. Kataoka, M. Katoh, T. Miyoshi, and H. Yamada, Appl. Environ. Microbiol. 56 (1990) 2303. 7. P. Rasor and W. Tischer, in "Advances in Industrial Biocatalysis", Bio-Europe (1998) 50.

51 8.

9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

24.

25. 26.

27.

S. Rissom, U. Schwarz-Linek, M. Vogel, V.I. Tishkov, and U. Kragl, Tetrahedron: Asymmetry 8 (1997) 2523; A.S. Bommarius, K. Drautz, W. Hummel, M.R. Kula, and C. Wandrey, Biocatalysis 10 (1994) 37. A. Galkin, L. Kulakova, H. Yamamoto, K. Tanizawa, H. Tanaka, N. Esaki, and K.J. Soda, J. Ferment. Bioeng. 83 (1997) 299. R.V.A. Orru, S.F. Mayer, W. Kroutil, K. Faber, Tetrahedron 54 (1998) 859; S. Pedragosa-Moreau, C. Morisseau, J. Baratti, J. Zylber, A. Archelas, and R. Furstoss, Tetrahedron 53 (1997) 9707. Y. Ge, Y. Wang, H. Zhou, S. Wang, Y. Tong, W.J. Li, Biotechnol. 67 (1999) 33. G.S. Wang, Y.T. Liu, Taiwan Tangye Yanjiuso Yanjiu Huibao 151 (1996) 55. M. Gargouri, M.D. Legoy, Enzyme Microb. Technol. 21 (1997) 7. H. Maede, K. Takata, A. Toyoda, T. Niitsu, M. Iwakura, and K. Shibata, J. Ferment. Bioeng. 83 (1997) 113. M.T. Reetz and K. Schimossek, Chimia 50 (1996) 668. R. Femandez-Lafuente, J.M. Guisan, M. Pregnolato, and M. Terreni, Tetrahedron Lett. 38 (1997) 4693 and J. Org. Chem. 62 (1997) 9099. K. Seelbach and U. Kragl, Enzyme Microb. Technol. 20 (1997) 389. A.L.E. Larsson, B.A. Persson, and J.E. Backvall, Angew. Chem. Int. Ed. Engl. 36 (1997)1211; J.V. Allen and J.M.J. Williams, Tetrahedron Lett. 37 (1996) 1859. D. Lee, E.A. Huh, M.-J. Kim, H.M. Jung, J.H. Koh, and J. Park, Org. Lett. 2 (2000) 2377. P.M. Dinh, J.A. Howarth, A.R. Hudnott, J.M.J. Williams, and W. Harris, Tetrahedron Lett. 37 (1996)7623. C.-H. Wong, S.L. Haynie, and G.M. Whitesides, J. Org. Chem. 47 (1982) 5418. K.M. Koeller and C.-H. Wong, Chem. Rev. 100 (2000) 4465 and references cited herein. A.P.G. Kieboom, M. Makkee, and H. van Bekkum, unpublished scheme and data (1985) derived from both ref 1 and from: M. Makkee, "Combined action of enzyme and metal catalyst, applied to the preparation of D-mannitol", PhD Thesis, Delft University of Technology, Delft, Netherlands (1984). J.J. Heijnen, C.A.G. Haasnoot, A. Bruggink, R.A.L. Bovenberg, E.W. Meijer, B.L. Feringa, and A. Driessen, NWO Programme Proposal "Integration of Biosynthesis and Organic Synthesis, A New Future for Synthesis", 9 October 2000, The Hague. M. Mccoy, Chem. Eng. News, 19 Februari 2001, p. 23. J. Lugtenburg and H.J.M. de Groot, Photosynth. Res. 55 (1998) 241 and in: Stable Isotopes in Pharmaceutical Research, Pharmacochemistry Library 26(1997) and references cited therein. R. Schoevaart and T. Kieboom, Carbohydr. Res, submitted.



53

What can Structure tell us about Function in the Estrogen Receptors? Roderick E. Hubbard, Structural Biology Laboratory, Department of Chemistry, University of York, Heslington, York, YOlO 5DD. [email protected] The estrogen receptor (ER) is a ligand-inducible transcription factor that controls expression of a number of genes in a wide variety of tissues. Binding of the natural hormone, estradiol, triggers dimerisation and nuclear location of the receptor v^here it binds to a response element (ERE). It then recruits a succession of large multi-protein complexes to modify chromatin and modulate the transcriptional machinery. The classical ER is ER-a, found in female reproductive organs but also important for maintaining bone and cardiovascular function. A number of important drug therapies depend on achieving agonism or antagonism of the receptor in these different tissues. Antagonists of the receptor are effective in treating some forms of breast cancer, whereas agonists are required to maintain bone function in post-menopausal women and thus protect against osteoporosis. About five years ago, a new isoform of the receptor (ER-P) was discovered which is distributed widely in both males and females. The differential role of these receptors is currently being investigated. We have determined the structure of the ligand binding domain (LBD) of both the a and P forms of ER complexed to different ligands, including partial and full agonists and selective and full antagonists. This has provided insights at the molecular level into some key aspects of the pharmacology and function of this molecule. Our results are described in detail in the papers listed in the bibliography. The sections below provide a brief summary of the key points. The distinctive ER pharmacophore Estradiol (E2), the main female sex hormone, contains two hydroxyl groups separated by a relatively flat and hydrophobic steroid core. A large number of estrogenic compounds have been characterised which have the same general features -. the distinctive ER pharmacophore. The overall structure of the LBD is a triple sandwich of alpha helices that generate a totally enclosed, mainly hydrophobic ligand binding pocket. In all structures determined to date, the A ring phenolic hydroxyl is found in the same position, hydrogen bonding to a water molecule and an arginine and glutamic acid residue. The D ring hydroxyl (and equivalents in other ligands) forms a hydrogen bond to a histidine amino acid. There are no other polar amino acids in the ligand binding pocket, which is tightly defined around the A ring, but there is additional space around the D ring.

54 The structure of ER LBD bound to a variety of ligands has demonstrated how this additional space is exploited to accommodate ligands of different shapes. The structures also confirm that the agonist conformation of the receptor has a key helix, helix 12, lying across the putative entrance to the binding cavity. Selective antagonism by molecules such as raloxifene There are a class of ligand molecules, known as Selective Estrogen Response Modulators (SERM) - such as tamoxifen and raloxifene (RAL) - which show different agonist and antagonist properties in different tissues. The structure of ER LBD complexed to RAL shows, as expected, that the core of the ligand is bound in essentially the same orientation as E2, with the small difference in shape accommodated by a movement of the histidine residue. The large pendant side chain exits from the binding cavity, preventing helix 12 from occupying its agonist position. Instead, helix 12 is found in an alternative binding site, identified as the major coactivator binding site on the protein. The agonism exhibited by these SERM ligands in some tissues could be due to residual coactivator binding sites elsewhere on the full length receptor, which are used in different cells with different coactivators expressed. Full antagonism A more extreme antagonist is the ICI compound, which consists of the estradiol core with a very long, hydrophobic side chain. This ligand acts as a full antagonist, abolishing estrogen receptor activity and in addition, there is some evidence that it increases receptor turnover. The structure of ER LBD complexed to ICI, shows that the ligand occupies the binding pocket, but has to flip from the standard E2 conformation to allow the side chain to exit by the same route used by RAL. The side chain then occupies the coactivator binding cleft - where helix 12 is found in the RAL structure. This means that helix 12 does not have an available alternative binding site and the helix is not seen in our electron density maps. It is possible that this large conformational change also disrupts the other activation function on the protein, found in the large N terminal domain which is not a part of the LBD structure. Partial agonism in ER-^ The plant phyto-oestrogen genistein is one of a number of compounds that act as full agonists for ER-a but are only partial agonists in ER-P. In addition, there is some evidence that ER-p is more readily antagonised than ER-a by a series of compounds. The structure of ER-P complexed with genistein shows that helix 12 adopts an unusual conformation, halfway between the agonist and antagonist positions seen for E2 and RAL respectively complexed to ER-a Taken together, these results suggest that helix 12 is more easily displaced in ER-a than in ER-p.

55 Co-activator recruitment The alternate binding location for helix 12 seen in the ER-a RAL structure masks key amino acids (in particular Lys 362) which are implicated in co-activator recruitment. In addition, the buried surface of helix 12 has similar features to the LXLL motif identified as the nuclear receptor binding region on coactivators (the so-called NR-box). Structures of agonist bound LBD, complexed to representative NR-box peptides, have confirmed that the NR-box peptides adopt a helical conformation and a binding site very similar to that adopted by helix 12 in the ER RAL structure. Concluding Remarks The published structures of the ER LBD a and (3 forms, in complex with a variety of ligands, provide a structural rationale for the measured agonist and antagonist properties of the ligands, and have identified the principle co-activator binding site on the receptor. However, there are many additional features of receptor function that remain to be explained. This will require structure determinations of larger constructs of the receptor (to include the DNA binding domain and the N terminal A/B domain), possibly complexed to larger fragments of co-activators. Acknowledgements This work was funded by Karo Bio Inc and the infrastructure of the Structural Biology Laboratory at York is supported by the BBSRC. The main contributors to the work were Ashley C.W. Pike, Andrzej M. Brzozowski, Julia Walton of York, together with Tomas Bonn\ Jan-Ake Gustafsson^ and Mats CarlquistV ^Karo Bio AB, NOVUM, S14157 Huddinge, Sweden. ^Departments of Medical Nutrition and Biosciences, Karolinska Institute, NOVUM, S-14186 Huddinge, Sweden.

Bibliography Brzozowski, A.M., Pike, A.C.W., Dauter, Z., Hubbard, R.E., Bonn, T., Engstrom, O., Ohman, L., Greene, G.L., Gustafsson, J.A. and Carlquist, M. (1997), "Molecular basis of agonism and antagonism in the oestrogen receptor". Nature, 389, 753-758 Pike, A.C.W., Brzozowski, A.M., Hubbard, R.E., Bonn, T., Thorsell, A.G., Engstrom, O., Ljunggren, J., Gustafsson, J.K. and Carlquist, M. (1999) "Structure of the ligandbinding domain of oestrogen receptor beta in the presence of a partial agonist and a full antagonist" Embo J, 18 4608-4618 Hubbard, R.E., Pike, A.C.W., Brzozowski, A.M., Walton, J., Bonn, T., Gustafsson, J.A. and Carlquist, M., (2000), "Structural insight into the mechanisms of agonism and antagonism in oestrogen receptor isoforms", Eur. J. Cancer, 36, S17-S18 Pike, A. C. W., Brzozowski, A. M., Walton, J., Hubbard, R. E., Bonn, T., Gustafsson, JA. and Carlquist, A. M. (2000), "Structural aspects of agonism and antagonism in the oestrogen receptor", Bioch. Soc. Trans. 28, 396-400

56 Pike, A. C. W., Brzozowski, A. M. and Hubbard, R. E. (2000)" A structural biologist's view of the oestrogen receptor", J of Steroid Bioch and Mol. Biol 74, 261-268 Pike, A. C. W., Brzozowski, A, M., Walton, J., Hubbard, R. E., Thorsell, A.-G., Li, YL, Gustafsson, J.-A. and Carlquist, M. (2001) "Structural insights into the mode of action of a pure antiestrogen". Structure, 9, 145-153


57

Molecular Docking and Dynamics Simulations in the Ligand Binding Domain of Steroid Hormone Receptors Milou Kouwijzer-f- and Jordi Mestres§ i Molecular Design & Informatics, N.V. Organon, 5340 BH Oss, The Netherlands § Computational Medicinal Chemistry, Organon Laboratories Ltd., Newhouse, Lanarkshire MLl 5SH, Scotland

Introduction During the process of protein-ligand recognition, the structures of both the protein and the ligand undergo conformational changes as a result of a mutual induced fit aiming at optimising their interaction. Although ligand flexibility is now efficiently accounted for in docking calculations [1-8], proper treatment of protein flexibility still remains a critical issue for docking methods due to the complexity of sampling and scoring and the amount of computer power required [9-17]. The inability to account for protein flexibility can sometimes have a significant effect on the results of docking calculations. On one hand, it is known from crystal structures of proteins cocrystallized with different ligands that the backbone and/or the side chains of the residues in the binding pocket adapt to the ligand bound. This situation is illustrated in Figure la for two crystal structures of the estrogen receptor a (ERa). On the other hand, there is an intrinsic ambiguity in the orientation of heteroatoms for some residues (particularly threonine, histidine, glutamine and asparagine) due to the fact that differences in the electron density of carbon, nitrogen, and oxygen atoms are often difficult to appreciate. This situation is illustrated in Figure lb for the two independent molecules in the crystal structure of the progesterone receptor. It is thus clear that the result of a docking calculation depends on such details in the protein structure, and that scoring functions cannot correct for the inflexibility of the protein. One way of taking into account the flexibility of a protein is using molecular dynamics (MD) simulations. Because of the computer time requirements, it is not applicable for large numbers of ligands but it can be used for binding mode assessment activities. The aim of the present work is to evaluate the performance of MD simulations when applied to the ligand binding domain of steroid hormone receptors and compare the results with those obtained from flexible ligand docking methods that make use of a single rigid crystal structure.

58

Figure 1. Side chain adjustments and ambiguities in the binding pockets of steroid hormone receptors, a (left): superposition of ERa cocrystallized with estradiol [23] (light gray) or DBS [24] (dark gray), b (right): superposition of two PR structures [26] (molecule A in light gray and B in dark gray). The choice of a single rigid structure for docking purposes will clearly affect the results.

Methodology For the MD simulations, only the ligand and the residues that have at least one atom within 8 A of the ligand were allowed to move (i.e. -800 out of -4000 atoms). From a 100 ps run at 400 K (after a heating phase of 10 ps) coordinate sets were saved every picosecond. After the MD simulation, these coordinate sets were energy minimized (still taking into account the constraints) and the average interaction energy from these 100 frames was calculated. The orientation having the lowest average interaction energy was taken as the most favorable orientation. All calculations were performed with QUANTA/CHARMm [18] and charges were taken from templates. A 15 A cutoff radius was used, with a shift and a switch function between 11 and 14 A for the Coulomb and van der Waals energy, respectively. For the minimizations and MD runs, a distance-dependent dielectric constant 8 of 4r was used, whereas for the calculation of the interaction energy, 8 was set to r [19]. Docking calculations were performed with DOCK 4.0 [4]. Sphere centers were generated with SPHGEN [20]. For computational reasons, initial sphere sets were reduced to a number of around 30, ensuring appropriate coverage of the binding pocket. Energy (AMBER [21]) and chemical scoring grids of 0.3 A were generated using a united atom model and an 8 of 4r with a 10 A cut-off radius. Mol2 files of the ligands were generated with SYBYL [22] using Gasteiger charges. Virtual screening calculations were performed using uniform sampling with a maximum of 50 orientations and 5 seeds, whereas higher sampling values of 500 orientations and 25 seeds were set for binding mode

59 calculations. The minimal anchor size was set to four atoms and the maximum number of steric clashes between protein and ligand was set to three. Energy/chemical score minimization of maximally 100 iterations to a convergence of 0.1 kcal/mol was performed to each docked ligand. A database of 1000 diverse compounds selected from the Available Chemicals Directory (ACD) was used for virtual screening purposes. For validation purposes, MD simulations in the ligand-binding domains of steroid hormone receptors with their co-crystallized ligands were performed, namely, ERa__lere (estrogen receptor a, structure from pdb code lere) with estradiol [23], ERa_3erd (structure from pdb code 3erd) with diethylstilbestrol (DES) [24], ERp with genistein [25], and the progesterone receptor (PR) with progesterone [26]. In addition, MD simulations of ERa structures with the ligand that was co-crystallized in another structure were also performed, namely, ERa^lere with DES, and ERa_3erd with estradiol. The former calculation was repeated for 1000 ps at 300 K to validate the high temperature used in the simulations. The aim of these calculations was mainly to assess the effect and the extent of the accommodation of the binding pocket to a ligand other than the native co-crystallized ligand. Finally, as an application example, a prediction of the orientation of genistein in the ERa binding pocket based on MD simulations is presented (the crystal structure of the ERa-genistein complex has not yet been determined experimentally). Due to the high computer time requirements for the MD simulations, sampling of ligand orientations in the binding pocket was limited to the most probable orientations. The four starting orientations for the ligands containing a steroidal core are represented below, orientation 1 being the one observed in the crystal structures containing estradiol [23] and progesterone [26]. Analogous orientations were used for the non-steroidal ligands, DES and genistein.

1

2

3

4

Final structural details: in the ERo/p receptors, the proton of the (uncharged) histidine residue close to the steroidal D ring was placed on Ne2 and, in the PR structure (molecule A), the labels of atoms O E I and NE2 in Gln-725 were exchanged, in accordance with the hydrogen bonding scheme published [26] and the second molecule in the asymmetric unit.

60 Molecular docking In order to assess to which extent the use of a single rigid crystal structure affects the final results of a docking calculation, the structures of ERa cocrystallized with estradiol [23] and DES [24] were used in a virtual screening exercise. As shown above in Figure 1, the binding pocket of the two structures differs mainly in the conformation of four residues, namely, Met343, Thr347, Met421, and His524. While DES pushes away Met343 and His524 with respect to the position they adopt with estradiol, estradiol induces a significant conformational change to Met421 with respect to the position it adopts with DES. More interestingly, while in lere the hydroxy 1 of Thr347 points towards the ligand cavity, in 3erd it points away from the ligand cavity. All these conformational changes will define essentially a different accessible space and interaction pattem in the binding pocket and thus result potentially in a different ranking for the molecules in a database. The correlation between the energy and chemical score ranks obtained by docking 1000 diverse molecules from the ACD in the two rigid ERa structures (lere and 3erd) is presented in Figure 2. As can be visually observed, a better correlation between molecule ranks is found when using the chemical score (r^=0.82) than when using the energy score (r^=0.67). In this respect, the chemical score seems to be less dependent than the energy score on the conformational changes of the residues defining the binding pocket. This is a particularly interesting statement that would require though further investigation to other protein binding pockets to assess its generality. In conclusion, it has been shown that the use of different single rigid structures does have a significant effect in the final outcome of docking calculations and that efficient treatment of protein flexibility during the course of a docking calculation would be highly desired. BOO

'."^.s-S&l^'

••••••.• • • •••

700 600

••"••.Js^Â-

111 500 Ul •- 400

k'^-^"'0

100

200

300

'••• 300

400 3ER0

SOO

200

600

700

800

»30

100 0
0

1600 1400

300

CO 0)

^^ >

•

500

V.

B

3000

0.3 1 3 10 (mg/kg s.c.)

1000^ SKF 38393 (D1 agonist)

800-]

Quinpirole (D2 agonist)

2500 2000 1500

• WT

600 H

A M4'-/-

400

• WT M4-/-

* A

1000

200

500 0

Veh

3 10 30 (mg/kg s.c.)

100

1—I—I

Veh

0.003

1

1—r

0.03 0.1 1 (mg/kg s.c.)

10

FIG. 1. Locomotor activity of M4R'^' mice and their WT littermates. (A) Basal locomotor activity. Horizontal locomotor activity was assessed in an 'open field' test by determining the number of photobeam breaks during a 1 hr observation period (n=115). (B-D) Locomotor effects induced by dopamine receptor agonists. (B) Apomorphine (non-selective), (C) SKF 38393 (selective for Dl-type receptors), and (D) quinpirole (selective for D2-type receptors). For additional experimental details, see ref. [17]. Data are given as means ± S.E.M. *, p^

Y^

Veh

0.01

0.03

0.1

0.3

Veh

0.01

0.03 0.1

0.3

Oxotremorine (mg/kg s.c.) FIG. 2. Oxotremorine-induced analgesic responses in M2 and M4 receptor mutant mice and their WT littermates. (A, C) Tail-flick test. (B, D) Hot plate assay. Mice of the indicated genotypes were injected s. c. with increasing doses of oxotremorine or saline (Veh). Analgesia measurements were carried out as described in ref. [16]. Data are given as means ± S.E.M. (n=18-20 per dose and genotype) and are expressed as % maximum possible effect (MPE). Data were taken from refs. [16] and [17].

103 The tail-flick data suggested that spinal cord mAChRs play a key role in the suppression of pain responses. To examine the expression of M2 and M4 receptors in mouse spinal cord, we labeled spinal cord mAChRs with a saturating concentration (2 nM) of the non-selective muscarinic antagonist, [^H]N-methylscopolamine ([^H]NMS). Strikingly, the number of [^H]NMS binding sites was reduced by > 90% in preparations from M2/M4 receptor double KO mice and M2R"^' mice (A. Duttaroy and J. Wess, unpublished results). On the other hand, no significant reduction in [^H]NMS binding activity was seen in preparations from M4R"^" mice. These data indicate that the M2 receptor subtype is the predominant mAChR species expressed in mouse spinal cord, whereas M4 receptor expression appears to be too low to be detectable in radioligand binding assays. From a therapeutic point of view, the potential use of M2 receptor agonists as analgesic drugs appears problematic, primarily because of the expected cardiac side effects. On the other hand, the availability of agonists with a high degree of M4 receptor selectivity would allow the administration of relatively high agonist doses required to achieve maximum analgesia without causing unwanted side effects associated with the activation of other mAChR subtypes. Molecular characterization of central inhibitory muscarinic autoreceptors It is well known that acetylcholine (ACh), like most other neurotransmitters, can inhibit its own release via stimulation of so-called inhibitory autoreceptors present on cholinergic nerve endings [43, 44]. Physiologically, this mechanism may serve to prevent excessive neurotransmitter release and subsequent overstimulation of target cells. Classical pharmacologic studies suggest that multiple mAChRs can function as inhibitory autoreceptors in both peripheral and central tissues. However, in many cases, the identity of the specific receptor subtypes involved in this activity remains controversial, primarily due to the limited subtype selectivity of the ligands used in these studies. To address this issue in a more direct fashion, we initiated a series of in vitro ACh release studies, using brain tissues from different mAChR mutant mouse strains. Specifically, we studied hippocampal, cortical, and striatal preparations from mice which lacked either M2 or M4 receptors or which were deficient in both M2 and M4 receptors. For these studies, superfused hippocampal, cortical, and striatal slices were incubated with [^H]choline to label cellular ACh pools. Potassium-stimulated ["^HJACh release was then determined in the absence of drugs (SI phase) and in the presence of the muscarinic agonist, oxotremorine, or other drugs (S2 phase). The S2/S1 release ratio was then used as a parameter to quantitate drug effects on transmitter release. As already stated above, the M2 and M4 mAChRs (but not the Mi, M3, and M5 receptors) are efficiently coupled to G proteins of the GJGQ family [1-3]. Since mAChR-activated Gi/Go proteins mediate the inhibition of voltage-sensitive Ca^^ channels [2, 45] which are known to be intimately involved in the regulation of neurotransmitter release, we speculated that the M2 and/or M4 receptor subtypes represent the major inhibitory muscarinic autoreceptors. To test this concept in a direct and unambiguous fashion, we initially analyzed ACh release using brain tissues from M2/M4 receptor double KO mice. Incubation of hippocampal, cortical, and striatal slices prepared from WT mice with

104 increasing concentrations of oxotremorine led to concentration-dependent reductions (up to 80%) of potassium-stimulated [^H]ACh release (W. Zhang and J. Wess, unpublished results). Strikingly, in hippocampal, cortical, and striatal slices prepared from M2/M4 receptor double KO mice, oxotremorine completely lost its ability to mediate inhibition of stimulated [^H] ACh release (W. Zhang and J. Wess, unpublished results). This observation demonstrates in a very direct fashion that either M2 or M4 receptors (or a mixture of the two receptors) mediate autoinhibition of ACh release in these brain tissues. We are currently carrying out analogous studies with brain slices from M2 and M4 receptor single KO mice in order to assess the contribution of each of these receptors to autoinhibition of ACh release in different areas of the brain. Proper regulation of ACh release in the striatum is known to be critical for coordinated locomotor control [30, 46, 47]. Likewise, the maintenance of proper synaptic ACh levels in hippocampus and cerebral cortex is thought to be important for facilitating learning and memory [48-50]. Identification of the presynaptic mAChRs regulating ACh release in these brain regions is therefore of considerable therapeutic interest.

ANALYSIS OF M3 mAChR MUTANT MICE The M3 mAChR is known to be widely expressed in the CNS and in peripheral tissues [2, 3, 13, 51]. Peripheral M3 receptors are predicted to play a role in ACh-dependent smooth muscle contraction and glandular secretion [2, 4, 20, 21]. At present, little is known about the roles of the central M3 receptors. To learn more about the physiological importance of the M3 receptor subtype, we generated mice with a targeted disruption of the M3 mAChR gene [19]. Western blotting and immunoprecipitation studies confirmed the absence of M3 receptor protein in the homozygous M3 receptor mutant mice (M3R"^" mice). Moreover, immunoprecipitation studies using subtype-selective antisera demonstrated that the lack of M3 receptors had no significant effect on the expression levels of the remaining four mAChR subtypes [19]. Body weight measurements and pharmacologic and endocrinologic studies MsR'^" mice were obtained at the expected Mendelian frequency, showed no obvious behavioral abnormalities, and did not differ from their WT littermates in fertility and longevity. Immediately after birth, WT and M3R"''" mice had similar body weights (Fig. 3). However, starting at about 2-3 weeks after birth, M3R*^' mice showed a significant reduction in body weight (Fig. 3). This difference continued to increase during the following weeks and persisted throughout the life of the animals. Generally, adult male and female M3R'^' mice weighed about 25% less than their WT littermates (Fig. 3). Despite the observed differences in body weight, the body length of adult MsR'^" mice did not differ significantly form that of their WT littermates (Fig. 4A), indicating that the lack of M3 receptors does not interfere with proper linear growth. We observed, however, that the mass of peripheral fat deposits was considerably reduced in the M3R"^" mice. As shown in Fig. 4B, male M3R"" mice showed an about 50% reduction in gonadal fat pad mass, a parameter which generally correlates well with total body fat content [52].

105 Strikingly, the lack of M3 receptors also led to pronounced reductions (5-10-fold) in serum leptin and insulin levels (Fig. 4C, D). The low serum leptin and insulin levels found in the M3R"^" mice are probably primarily due to the reduction in total body fat.

B

25-1

Females

20 H 15H lO-j 51 - I — I — I — I

3 5 7 9 11 13 1517 19 21 Time (week)

3

5 7 9 Time (week)

11

FIG. 3. Lack of M3 mAChRs results in reduced body weight. Growth curves for male (A) and female (B) MsR'^" mice and their WT littermates fed with standard food pellets (n = 17 or 18 per group). Data are given as means ± S.E.M.

A

Bcxly length

B

Fat pads

i£3

10

^ ^ 2

^ 20

f 12 0)

1

8

E

i '

Leptin

3l

12

I

Insulin

E

n

M3WT(+/+)

•

M3K0(-/-)

250 ^

t

200

Glucose I

r^

JO 1501

i 100 E E 50 0} CO

0

FIG. 4. Body length, mass of peripheral fat pads, and serum leptin, insulin, and glucose levels in WT and MsR'^' mice. (A) Body length (nose-anus) measurements were carried out with 20-week old male mice (n=8). (B) Percentage of epididymal fat pad weight relative to total body weight (10-week old male mice; n=8). (C-D) Serum levels of leptin (C), insulin (D), and glucose (E). Blood was collected from freely fed male mice that were about 20 weeks old (n=6). Data are given as means ± S.E.M. *, p

li Jilliri: i

Q

Dry pellet food

§5

-^

1

r2.0

(D 0>

0.3 i ? - n

11^

20 j

£ o CO "

Q

'

D

M3WT(+/+) M3KO(-/-)

5 o

Wet paste food

FIG. 5. Lack of M3 mAChRs leads to reduced food intake. (A) Daily consumption of standard dry food pellets by 12-week old male mice. (B) Daily consumption of wet mash food by 20-week old male mice. Food intake was measured daily over a 5-day observation period. Data are expressed as food intake per mouse or as food intake per (body weight in g) [59]. Data are given as means ± S.E.M. (n=10 per group). p=0 '

+

H,N

COOH "^^ 20

^90% conversion and purity 7 monomers: >70% conversion and purity

30 selected M- monomers

54 monomers: >90% conversion and purity 8 monomers: >70% conversion and purity

87 monomer candidates

62 selected Mj monomers

c: Die. DIPEA. DIVIAP. CH2CI2, rt, 12-16 hours.

+

HOOC

98 monomer candidates 44 monomers: >90% conversion and purity 18 monomers: >70% conversion and purity

62 selected M3 monomers

Scheme 6. Monomer rehearsal for the synthesis of a large, solid phase pool library.

157 The authors prepared a small 456-member model library L l l before embarking into the whole library synthesis (Scheme 7). The synthesis was successful, as judged by the complete analytical characterization of L l l , and its biological testing on several targets highlighted some active compounds [23]. This meant drug-like properties for the components of L l l , as cellular assays were used to screen the library and penetration through the cell membrane was needed to show activity. Unfortunately, two major drawbacks were also apparent: • Deconvolution of active pools, even when containing only 8 compounds, gave rise both to non-reproducible synergistic effects and to false positives, and • The quantity of compound supported onto a single 90|im-bead was not enough to test the library on single bead-based assays.

.NH,

8 representatives

8 representatives

L11 8 representatives

456-member library 8 pools, 64 compounds/pool 90|xm Tentagel beads

Scheme 7. Synthesis of the solid phase model library L l l . The authors eventually prepared the final library L12 (Scheme 8) using larger, in house developed beads [28] and a known chemical encoding technique [29] to enable beadbased screening and structure characterization. The synthesis was again successful and a large, valuable collection of complex, natural products- and drug-like compounds was obtained.

158

^\ y

r\j^

558 compounds

^2

M2 62 representatives + skip codon

30 representatives + skip codon

Ra^ - O

34,596 compounds

px-^ÔH

L12

M3

2,179,548-member encoded library 63 pools, 34,596 compounds/pool proprietary 500|im beads

63 representatives + skip codon

Scheme 8. Synthesis of the soUd phase encoded pool Hbrary L12. The hits identified from any screening campaign are by definition showing an in vitro activity on the target, but many other features need to be optimized in order to make them useful leads. These include their physico-chemical properties, their stability, their bioavailability, their selectivity and their toxicity. This phase, here named hit optimization (Figure 1), goes usually together with hit identification under the name of lead discovery; in terms of diversity and libraries of compounds, though, the focusing on the hit structure as a model in hit optimization (similarity-driven libraries) is evident and different from diversity-driven or targeted libraries used in hit identification. It should be kept in mind that several drug discovery projects actually start from hit optimization: in fact, when the structure of a known inhibitor/modulator for a target of interest is known from the literature or from any other available source this structure becomes the model onto which to build focused libraries to progress the "external hit" status to a lead. Conversely on several occasions the hit optimization phase is not successful, that is the efforts prove how the hit structural core can not be progressed to a lead; often this happens because the compound potency and its toxicity, or aspecificity are similarly modulated and leads can not be obtained. In such a case a new screening campaign using a different screening collection, is sometimes considered to find alternative hit structures.

159 Structural focusing becomes even more stringent in lead optimization, when the number of compounds prepared decreases but their characterization becomes extremely accurate to eventually identify one or more clinical candidates to be processed in the downstream events of the drug discovery process. It should be finally noted, though, that a successful project completion at this stage strongly depends on the correct selection of chemical diversity in the first screening campaign, on the selection of the most promising hits and on the rational design of focused libraries for the optimization of the selected hit structures. References [I] [2] [3] [4] [5] [6] [7] [8]

[9] [10] [II] [12]

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

L. Weber, Curr. Opin. Chem. Biol 4, 295-302 (2000). M. Wagener and V.J. van Geerestein, J. Chem. Inf. Comput. Sci. 40, 280-292 (2000). E.A. Wintner and C.C. Moallemi, J. Med. Chem. 43, 1993-2006 (2000). P. Willett, Curr. Opin. Biotechnol 11, 85-88 (2000). L. Xue, J.W. Godden, and J. Bajorath, J. Chem. Inf. Comput. Sci. 40, 1227-1234 (2000). N. Rhodes, P. Willett, J.B. Dunbar and C. Humblet, J. Chem. Inf Comput. Sci. 40,210-214(2000). J. Xu and J. Stevenson, J. Chem. Inf Comput. Sci. 40, 1177-1187 (2000). P. Willett, Computational methods for the analysis of molecular diversity. Trends in Drug Research III, vol. 32: Pharmacochemistry library, Ed. H. van der Goot, Elsevier, Amsterdam, 2001. D.S. Thorpe, Combi. Chem. High Throughput Screening 3, 421-436 (2000). K.-Y. To, Combi. Chem. High Throughput Screening 3, 235-241 (2000). M. Lebl and V. Krchnak, Chem. Rev. 97, 411-448 (1997). V.V. Antonenko, Automation in combinatorial chemistry. Combinatorial chemistry and technology: Principles, methods and applications, Eds. S. Miertus and G. Fassina, Marcel Dekker, New York, 1999, pp. 205-232. W.J. Coates, D.J. Hunter and W.S. MacLachlan, Drug Discovery Today 5, 521527 (2000). J.-H. Zhang,, T.D.Y. Chung and K.R. Oldenburg J. Comb. Chem. 2, 258-265 (2000). C. Barnes and S. Balasubramanian, Curr. Opin. Chem. Biol. 4, 346-350 (2000). A. Furka, Combi. Chem. High Throughput Screening 3, 197-209 (2000). Harvey, A., Drug Disc. Today 5, 294-300 (2000). A.R. Leach and M.M. Hann, Drug Disc. Today 5, 326-336 (2000). L. Xue and J. Bajorath, Combi. Chem. High Throughput Screening 3, 363-372 (2000). S.D. Pickett, I.M. McLay and D.E. Clark, J. Chem. Inf Comput. Sci. 40, 263272 (2000). D.G. Powers, D.S. Casebier, D. Fokas, W.J. Ryan, J.R. Troth and D.L. Coffen, Tetrahedron 54, 4085-4096 (1998). J.M. Ostresh, G.M. Husar, S.E. Blondelle, B. Doemer, P.A. Weber and R.A. Houghten, Proc. Natl. Acad Sci. USA 91, 11138-11142 (1994).

160 [23] [24] [25] [26] [27] [28]

[29]

D.S. Tan, M.A. Foley, B.R. Stockwell, M.D. Shair and S.L. Schreiber, J. Am. Chem. Soc. 121, 9073-9087 (1999). D.A. McGowan and G.A. Berchtold, J. Org. Chem. 46, 2381-2383 (1981). D. Keirs and K. Overton, Heterocycles 28, 841-848 (1989). O. Tamura, T. Okabe, T. Yamaguchi, K. Gotanda, K. Noe and M. Sakamoto, Tetrahedron 51, 107-118 (1995). O. Tamura, T. Okabe, T. Yamaguchi, J. Kotani, K. Gotanda and M. Sakamoto, Tetrahedron 51, 119-128 (1995). M.A. Foley, R.W. King and S.L. Schreiber, Book of Abstracts, 221'' ACS National Meeting, BTEC-40, San Diego, April l'*-5*^, 2001, ACS, Washington, DC. H.P. Nestler, P.A. Bartlett and W. Clark Still, J. Am. Chem. Soc. 118, 1813-1814 (1994).


161

Beyond G Proteins: The Role of Accessory Proteins in G Protein-Coupled Receptor Signalling Herwig Just\ Eduard Stefan^ Cornelia Czupalla^, Bemd Niimberg^, Christian Nanoff^ and Michael Freissmuth^ Institute of Pharmacology, University of Vienna, Wahringerstr. 13a, A-1090 Vienna, Austria Department of Pharmacology and Toxicology, University of Ulm, Albert-Einstein-Allee 11, D89069 Ulm, Germany Introduction The basic mechanism by which G protein-coupled receptors (GPCRs) control their cellular effectors has been clarified in the 1970s and 1980s [1,2]; the cycle of activation and deactivation comprises the receptor-mediated exchange of GDP prebound to the G protein a-subunit for GTP, the ensuing dissociation of active, GTP-liganded Ga from the Py-dimer, effector regulation by both free Ga and GPy and signal termination by the intrinsic GTPase of -Ga. Formal proof has been provided for essentially all reaction steps including the recent demonstration that dissociation and reassociation of Ga and GPY do occur in intact cells [3]. Superimposed on this reaction sequence, there is a second cycle of receptor desensitisation and resensitisation which allows for negative feed-back regulation at the level of the receptor; the individual steps include (i) phosphorylation of the agonist-liganded receptor by GRKs (G protein-coupled receptor kinases), (ii) uncoupling by binding of arrestins to the phosphorylated receptor, (iii) sequestration and intemalisation by endosomes and (iv) down-regulation by lysosomal degradation or resensitisation by dephosphorylation and recycling to the membrane [4]. These two reaction schemes, i.e. the cycle of G protein activation and deactivation and the cycle of receptor desensitisation and resensitisation, have also served as working model to analyse the regulatory input that is generated in the network of cellular signalling pathways. The basic components of the transmembrane signalling machinery are the receptor, G protein heterotrimer, and the effector. It has increasingly been appreciated, however, that these do not suffice to understand the complexity of the signal transduction process. Several additional proteins (other than G proteins) impinge on the regulatory cycle by interacting directly with receptors and G protein subunits; these include the afore-mentioned GRKs and arrestins [4]; the latter play roles in signalling other than merely turning off the signal [5]. In addition, the G protein cycle is regulated by RGS proteins, regulators of G protein signalling, which bind to GTP-liganded Ga and function as GTPase activating proteins [6,7]; signalling by Py can be quenched by phosducin and phosducin-like proteins [8]. Finally, the rate of GDP-release from Ga can be modified directly by proteins that are distinct from receptors (termed AGS, activators of G protein signalling, based on their effect in the Py-dependent pheromone pathway of yeast that was originally employed to identify them; [9-11]). Conceptually these proteins can all be placed within the framework of the central cycle, i.e. they affect the kinetics of G protein activation and/or deactivation. There are however, alternative (i.e. non-G protein regulated) effectors that bind directly to receptors (see below).

162 A list of representative accessory proteins that is not exhaustive has been compiled in Table 1. As will become evident in the subsequent sections, in many cases the distinction between components required for sorting, scaffolding proteins, regulatory proteins and effectors is arbitrary; the association of a given receptor with its interaction partner may, for instance, be required for export from the endoplasmic reticulum (ER) and for clustering this receptor at a specific site; proteins that affect receptor trafficking may also directly impinge on the mechanism of signal transduction or feed-back regulation. Receptor synthesis and sorting It is obvious that Table 1 ought to include proteins that act as adapters in the process of sorting, delivery and targeting/retention. After synthesis and insertion into the membrane of the ER, a given receptor (=cargo) is subjected to repeated sorting events; these involve a defined set of analogous steps [12]: the receptor must be concentrated in budding vesicles; this presumably involves the recruitment of coat adapter proteins and coat proteins, protein(s) that support(s) the budding of the vesicle (by pinching off the lipid bilayer), a molecular motor (to allow for movement of the vesicle along microfilaments or microtubules), a vesicular SNARE (soluble N-ethylmaleimide-sensitive factor attachment receptor; to allow for fusion with the target membrane). These steps are repeated as the receptor moves along the biosynthetic pathway from the ER through the Golgi. Surprisingly little is known about the protein-protein interactions that are required for the biosynthesis and sorting of GPCRs. It is however clear that arrestins actually not only serve to block G protein signalling but also act as scaffolds to assemble clathrincoated vesicles. Thus, arrestins are adapter proteins in the budding of endocytotic vesicles [5]. In addition, the third intracellular loop of the Pi-receptor has been found to bind endophilinsl-3 [13]; endophilin-1 is, however, also required for vesicle budding [13] because the modification of lipids by its lysophosphatidic acid (LPA) acyltransferase activity is thought to induce negative curvature of the membrane which facilitates the invagination of lipids [15]. Arrestins - and presumably endophilins - are not involved in biosynthetic trafficking (coat recruitment and vesicle budding in ERGolgi trafficking); nevertheless, these two examples illustrate that GPCRs are not only a passive cargo but can, in principle, recruit molecules required to generate a transport vesicle. It is therefore likely that GPCRs also associate with analogous molecules in the ER. Accordingly receptors that are deficient in binding appropriate components will be retained; this may be the reason why some GPCRs are constitutively retained in the ER/Golgi, if they are heterologously expressed. A prominent example is the GABAer receptor, which requires the presence of the GABAei-receptor for export from the ER [16,17]. Similarly, several mutations in the vasopressin V2-receptor (the genetic basis for X-linked diabetes insipidus) result in the retention of the protein in the ER [18]. Some of these may be due to a loss in the ability of the receptor to recruit the export machinery. Finally, in cerebellar granular cells, mGluR5 (=metabotropic glutamate receptor-5) is not efficiently exported from the ER in the absence of the Homer-1 [19], Homer (several isoforms of which exist, see [20]) is also required for clustering group I mGluRs at specific sites of the cell membrane (see below). Thus, at the very least, these observations provide an example that complexes of receptors and accessory proteins can be preassembled in the ER.

163 Table 1 Accessory proteins (=proteins other than G proteins, arrestins and GRK) that bind directly to G protein-coupled receptors (receptor abbreviated as R) Target Protein Functional effect ref. rhodopsin Tctex-l=dynein chain microtubular transport 32 mGluRl & 5 Homer targeting/retention at presynaptic 19,39,40 or dendritic membrane; clustering inGluR? PICKl retention/protein kinase C-binding 44-46 tubulin mGIuRla targeting/transport ? 52,53 sst-R-2, Shank/SSRIP Retention/clustering? 48 CLl-R 49 spinophilin retention? 50,51 D2-R, a2-R endophilins Sorting/vesicle budding (?) 13 p2-R IL8-R-B PP2A core enzyme Dephosphorylation/resensitisation/ 76 =CXCR2 signalling (?) CXCR2 RGS12 regulation of signalling/clustering 47 src signalling (MAP kinase activation) 66 p3-R signalling (ras /MAPkinase)? Grb2, Nek 65 P4-R 74 EBP50/NHERF regulation of Na^/H^-exchange PrR 36 endosomal recycling 63 signalling ATII-Rl phospholipase Cyl 64 inhibition of signalling ATII-Rl, eNOS B2-, ETB-R 62 signalling ATII-Rl JAK2 89 inhibition of signalling Ca^Vcalmodulin |Li-opioid R 79 Ca^'^-channel regulation mOluR? 81 inhibition of signalling P2-R 77,78 inhibition /G protein selectivity coupling cofactor Ai-R 72,73 Impact (?) on transcription ATF/CREB GABABI-R 99 adenosine deaminase enhanced intemalisation (?) ApR 82,87 promoted by agonist occupancy homooligomerisation PrR 88 homooligomerisation constitutive oligomers 5-opioid-R 16,17 ER-export/membrane insertion GABABI/2-R heterooligomerisation heterooligomerisation altered pharmacological specificity 83 K-/5-opioid heterooligomerisation cross-talk with ionotropic receptor 90 DI/GABAA 86 heterooligomerisation and D2/sst-R5 84,85 heterooligomerisation G protein-coupled Ai/Di 84 heterooligomerisation receptors A2A/D2 91,92 heterooligomerisation Coreceptor for axon guidance (?) A2B/DCC altered pharmacological specificity 89 adrenomedul RAMPs lin/CGRP-R Receptor delivery/targeting and retention Vesicles that are loaded with a given receptor leave the Golgi and move along tracts of microtubules (and/or along the actin cytoskeleton) and are delivered to the plasma

164 membrane. Like many other signalling components G protein-coupled receptors and G proteins are, however, in many instances not uniformly distributed over the plasma membrane. Caveolins have been proposed to play an important role in organising signalling components (for review [21]); this view, however, has been questioned [22]. Of interest, G protein subunits and receptors are co- and posttranslationally modified by lipids (myristoylation, prenylation and palmitoylation). Palmitoylation of both, receptors and Ga-subunits, is dynamic [23]. Although the precise role of these lipid modifications is not fully understood, they are likely to play a role in targeting the proteins to lipid rafts; these membrane microdomains are thought to rely, at least in part, on the propensity of certain lipids (e.g. glycosphingolipids and cholesterol) to support selforganisation and to trap lipid-modified signalling molecules [24]. In polarised cells (epithelia and neurons), however, the segregation of membrane proteins is more pronounced; in epithelia, some proteins reside exclusively in the basolateral membrane (e.g. Na^/K^-ATPase) while others are only found on the apical side (e.g. epithelial Na^-channel). Originally, it had been argued that there was a functional equivalence between membrane compartments in polarised epithelial cells and neurons with basolateral and apical membranes corresponding to somatodendritic and axonal compartments, respectively [25]. This model provided a framework to formulate hypotheses and explore common principles; e.g. in epithelial cells the tight junctional ring prevents exchange of proteins between the apical and the basolateral membrane. A similar barrier to diffusion also exists in the vicinity of the axon hillock [26]. Lipid rafts may also contribute to the polarised distribution in neurons [27]. However, these factors can only maintain the segregation of proteins; clearly, other mechanisms must be important to generate the different distribution of proteins. Because of the length of axons (up to >1 m in man) and due to extensive ramifications of dendritic trees, neurons face challenges in targeting that epithelial cells do not have to meet. For instance, some proteins must be deposited along the axon (e.g. Na"^- and K'^-channels), while others are delivered exclusively to the presynaptic specialisations (e.g. presynaptic GPCRs, Ca^^-channels, neurotransmitter transporters). Given this complex task, it is not surprising that the original proposal (apical=axonal; somatodendritic=basolateral) has repeatedly been proven to be incorrect [28]. This is also true for GPCRs [29] and presumably reflects the fact that epithelia lack components of the targeting/retention machinery. An important mechanism in the control of delivery is the selection of the appropriate motor molecule that will direct the movement of receptor-containing vesicles. In most cells, the minus ends of microtubuli originate in the perinuclear region while the plus ends point to the periphery. In axons, microtubuli are uniformly oriented with their plus ends pointing to the synapse and their minus ends to the soma. Accordingly, anterograde delivery of axonal proteins is accomplished by plus end-driven molecular motors. These are N- and M-type KIFs (=members of the kinesin family with a motor domain at the N-terminus or in the middle; [30]). In contrast, C-type KIFs and cytosolic dynein function as minus-end driven motors and are thus candidates for mediating retrograde transport and endosomal recycling. Accordingly, transport vesicles are segregated into different pools thai select different kinesins. The mechanisms, by which motor molecules are recruited, include binding to vesicular coat proteins and adapter proteins, PDZ-

165 domain containing proteins (for explanation of the acronym see below) and other scaffolds as well rab GTPases [31]. There is, however, at least one example, where the cargo, i.e. a GPCR, serves as the docking site for the motor molecule: The retinal photoreceptor rhodopsin binds directly - via its carboxy terminal extension - to a molecular motor, namely to the dynein light chain Tctex-1 [32]. This observation indicates that GPCRs can per se function as binding partners for motor molecules and thereby specify the direction of transport. Rhodopsin is apparently delivered to minus end of the microtubular rail at the base of the cilia [32]; at this point, unconventional myosin VII (=an actin-based motor) is thought to take over [33]. The example of rhodopsin also illustrates the clinical relevance of receptor transport and delivery: mutations in the C-terminus of rhodopsin have a low affinity for dynein and are mistargeted; this is associated with retinal degeneration (=retinitis pigmentosa) due to loss of photoreceptors [32]. In polarised cells, disruption of the microtubuli (e.g. by colchicine, vinca alkaloids or nocodazol) impedes the delivery of (some, but not all) GPCRs to the plasma membrane [34,35]. Similarly, disrupting cortical (=submembranous) actin filaments with latrunculin (A or B) also affects receptor recycling [36]. (It should be noted that in many cases experiments with cytochalasin D, which also destroys actin filaments, have been less revealing, presumably because cytochalasin D has only a modest impact on the cortical actin meshwork [37]). This indicates that (i) trafficking itineraries differ and (ii) that there is an interplay between kinesin (=tubulin-based) and myosin (=actin-based) motors, where a given receptor is sequentially handled by several motors. The physiological relevance is evident; during prolonged desensitisation, in many cases, receptorcontaining vesicles can be visualised in the perinuclear region (reflecting presumably endocytotic retrieval via a dynein motor). Latrunculin B-sensitive transport may allow for recycling to the plasma membrane and, hence, for resensitisation. Alternatively, a different motor molecule, delivers the receptor to the lysosomal compartment which leads to loss of receptors, i.e. desensitisation by down-regulation. In the case of the Pareceptor, this decision depends on binding of EBP50/NHERF (the ezrin binding protein of 50 kDa = NaVH^-exchange regulatory factor) to the (phosphorylated) very C-terminal end of the receptor; the association of EBP50/NHERF reroutes the receptor to endosomal recycling rather than to lysosomal degradation [36]. Receptor retention and clustering After the receptor-containing vesicle has been delivered and fused (via SNARE interaction) to its target membrane, the receptor needs to be retained at this specific site in order to maintain a polarised distribution. If this does not occur, the receptor may be redistributed [38]. In many cases that have been studied, receptors are concentrated in specialised compartments of the plasma membrane (e.g. clustered at postsynaptic densities) via interaction of their carboxy termini with scaffolding proteins that contain PDZdomains (PDZ standing for a domain found in >50 proteins but first identified in the postsynaptic density protein of 95 kDa, Drosophila discs large and the zonula occludens protein zo-1, hence the acronym). Clustering was originally reported for ionotropic receptors (e.g. the nicotinic acetylcholine receptor in the end plate of the skeletal muscle). The discovery of Homer (now termed Homer-la) represented a major breakthrough; Homer-la was originally identified as an immediate early gene [39] and proposed to contain a PDZ-do-

166 main; however, the N-terminal anchoring domain of Homer-1 (and its relatives, Homer2 and -3; see [20]) is an EVHl-domain (shared by Drosophila Enabled, VASP=vasodilator stimulated phosphoprotein, WASP= Wiskott-Aldrich syndrome protein; [40]). This EVHl-domain of Homer interacts with the carboxy terminus of group I mGluR (mGluRl and mGluRS) by recognising a proline-rich motif (PPXXFR) and this interaction is thought to target the receptor to the postsynaptic membrane. Importantly, the expression of Homer-la is regulated in a highly dynamic manner in response to synaptic activity [19,39]. Thus, in the absence of Homer, the receptor is not delivered to the synapse. The splice variants Homer-lb and -Ic differ from Homer-la by the presence of an additional C-terminal leucine zipper; depending on which splice variant is expressed, mOluRS is delivered to dendrites (in the presence of Homer-lb and -Ic), whereas expression of Homer-la also resulted in delivery to axons [19]. Thus, Homer-lb and -Ic are apparently endowed with an axonal exclusion signal. In addition, only Homer-lb and -Ic support the formation of receptor clusters via their leucine-zipper [41]. The case of Homer also illustrates the fact that it is difficult to separate the individual steps (ERexport, sorting, targeting, delivery and retention) because Homer is actually required for targeting and clustering; however, once the receptors are delivered to the appropriate sites, they remain anchored even in the absence of Homer [19]. Obviously, there must be many more proteins that target and anchor GPCRs; mGluR7, a group III mGluR, is of particular interest because it is almost exclusively targeted to presynaptic specialisations and highly enriched in the active zone [42]. The C-terminus of mGluR7 specifies axonal localisation but it does not contain a motif which supports interaction of mGluRV with Homer-la or its relatives [43]. Recently, PICKl was identified as a binding partner for mGluRV [44-46]. PICKl is a binding partner for protein kinase Ca (PKCa) and contains a PDZ that binds the very C-terminal end of mGluR7 (and, to a lesser extent, that of other group III mGluR [45]); PICKl is not required for axonal targeting of mGluR7; however, in the absence of PICKl the receptor is no longer clustered [44]. In addition, the interaction with PICKl renders the receptor less susceptible to regulation by PKCa, which phosphorylates the C-terminus of mGluRT [46]. Thus, PICKl is also likely to interfere with receptor regulation. RGS12 also has related multifunctional properties. RGS12 is a GAP (GTPase activating protein) that is specific for Gtto and Gai; in addition, RGS12 contains an N-terminal PDZ domain that specifically recognises the C-terminus of the interleukin-8 receptor B (=CXCR2) but not the interleukin-8 receptor A (=CXCR1) or other GPCRs [47]. Finally, alternative splicing generates a C-terminus that binds to the N-terminal PDZ domain and thereby supports self-association of RGS12 by generating head-to-tail RGS12 concatemers [47]. It is evident that this arrangement favours the formation of clusters. A multidomain, actin-binding protein was also found to specifically bind to the C-terminus of the somatostatin-receptor-2 (SSTR2) and hence named SSTR-interacting protein (SSTRIP [48]); this scaffolding protein contains ankyrin repeats, an SH3- and a PDZ-domain; it is highly enriched in posts)aiaptic densities, and it was also independently identified by several groups (under various names, e.g. shankl and synamon). Shank proteins (of which there are, at present, 3 isoforms) also link other GPCRs to the actin cytoskeleton of presynaptic specialisations, e.g. CLl-receptor for the spider venom a-latrotoxin [49].

167 PDZ-domains recognise the C-terminal 4 amino acids; Homer also recognises a sequence in the C-terminus; hence, currently the focus is on the C-termini of GPCRs; it is generally assumed that these specify targeting in neurons and other polarised cells. However, additional sequence elements of GPCRs, including transmembrane segments [38], may also be important for mediating retention in specific membrane compartments. In many Gi-coupled receptors, the C-terminus is short and the third intracellular loop i3 is extended. It is therefore not surprising that this segment may also serve as a site of attachment. Spinophilin (a protein that is highly enriched in dendritic spines) has, for instance, been identified as an interaction partner for the D2-dopamine receptor [50]. Spinophilin is of particular interest because it is a multifunctional protein: the Nterminus of spinophilin binds to filamentous actin (F-actin), the C-terminus contains a PDZ-domain; the middle portion of the molecule associates with PPl (protein phosphatase-1) and also binds the third intracellular loop of the D2-dopamine receptor. It is very tempting to speculate that the D2-receptor is targeted to dendritic spines (e.g. in the intermediate size spiny neurons of the corpus striatum) by spinophilin. However, the crucial evidence is still missing, i.e. it has not been possible to co-immunoprecipitate D2-receptors and spinophilin [50]. This has recently been achieved for the 3 subtypes of the a2-adrenergic receptor [51]. Interestingly, the interaction of these receptors with spinophilin is enhanced by agonist occupancy indicating that spinophilin may also regulate signalling (i.e. possibly play a role in desensitisation). It is well established that the mobility of membrane proteins is restricted by the submembranous actin cytoskeleton; the contribution of microtubuli is less clear. The C-terminus of mGluRla binds to (monomeric) tubulin [52]. Because of the abundance of microtubules, the co-purification of tubulin with a given protein is generally viewed with suspicion. Agonist-stimulation promotes tubulin depolymerisation and recruits tubulin to the plasma membrane [53]. It is however not clear, if this is related to the direct interaction of tubulin with the receptor. There are several mechanisms by which receptors may affect microtubular dynamics, e.g. in vitro both, Ga (Gtts, Gai; [54]) and GPy [55] affect tubulin polymerisation; direct effects (=tubulin depolymerisation) have also been postulated for Gaq [56,57]. In addition, GRK2 and GRK5 bind tightly to tubulin which they also recognise as substrate [58,59]. Microtubuli participate in organising ligandgated ion channels into clusters (e.g. of the glycine receptor via the bridging protein gephyrin, [60]). In contrast, there is no firm evidence that microtubuli are important in retaining GPCRs at specific sites. Finally, while it is clear that the actin cytoskeleton and microtubuli must play a role in anchoring signalling components, their role must not be overemphasised because additional scaffolding components can actually maintain clusters even after disruption of micotubuli and actin depolymerisation [61]. Association of GPCRs with (non-G protein-dependent) effectors and regulators GPCRs generate two signalling molecules (Ga and Py-dimers) that each can control several effectors. Hence, it seems counterintuitive that these receptors require additional effector mechanisms. Nevertheless, GPCRs bind and regulate alternative (i.e. non-G protein regulated) effectors, e.g. the janus kinase JAK2 (which is activated; [62]) phospholipase Cyl (which is activated; [63]), endothelial NO-synthase (which is inhibited;

168 [64]); in addition they can bind adapter molecules (Grb2 and Nek) that recognise proline-rich stretches via SH3-domains and lead to activation of the ras-cascade ([65]). Two examples illustrate the contentious nature of direct effector binding to GPCRs: (i) the non-receptor tyrosine kinase src has been reported to bind directly to the Ps-adrenergic receptor, and this reaction was considered the mechanistic basis for activation of MAP kinase (=mitogen-activated protein kinase =erk) [66]. Similarly, regulation of a neuronal cation channel (via mGluRl in hippocampal neurons) was reportedly G protein-independent (i.e. insensitive to uncaging the inhibitory GDP analogue GDPPS) but blocked by the src-inhibitor PPl [67]. However, in adipocytes, i.e. the cells in which the Ps-receptor is expressed endogenously, erk activation depends on cAMP; i.e. the signal generated by the cognate signalling pathway of the Ps-receptor (Gg/adenylyl cyclase) is necessary and sufficient to account for erk activation [68]. It is also difficult to reconcile a G protein-independent activation of src with the recent observation that src can be directly activated by Gtts and Gaj [69] (related observations have been made with other non-receptor tyrosine kinases). In addition, it has long been known that PKA can phosphorylate src (on Ser^^ in the N-terminus [70]) and that elevation of cAMP can increase src kinase activity [71]. (ii) The C-terminus of the GABABI-receptor recognises transcription factors of the ATF/CREB family; the relevance of these findings in a matter of debate; in fact, the two studies that reported on the association reached opposite conclusions, i.e. that agonist occupancy of the GABAe-receptor removed ATF4 (=CREB2) from the nucleus (and hence presumably inhibited transcription; [72]) or translocated the ATF4 from the cytoplasm to the nucleus and activated transcription [73]. It is evident that in many cases the distinction between regulatory proteins and effectors is blurred; e.g. EBP50/NHERF mediates the effect of the P2-receptor on the Na^/H^-exchanger [74] and targets internalised receptors to recycling rather than degradation [36]. Similarly, phosphorylation of the P2-receptor and subsequent binding of P-arrestin does not only target the receptor to clathrin-dependent endocytosis but also redirects the signalling pathway from activation of cAMP accumulation to activation of the MAP-kinase cascade G protein [5,75]. Receptor recycling requires that the receptor is dephosphorylated. Accordingly, binding of protein phosphatases to phosphorylated receptors is to be anticipated. Protein phosphatase 2A (PP2A) binds to the proximal portion of Cterminus of CXCR2, i.e. the region adjacent to the 7* transmembrane helix; this binding requires the core of PP2A (=regulatory 65 kDa A-subunit + catalytic 36 kDa domain) and receptor intemalisation, but it is surprisingly independent of phosphorylation [76]. It is therefore not clear, if the docking of PP2A to CXCR2 plays a role in signalling other than receptor dephosphorylatioh. There must be components that account for the difference in G protein-coupling that is observed in reconstituted systems and in intact cells; G protein-coupled receptor can display an exquisite capacity to discriminate between closely related G protein subunits, if this selectivity is assessed in intact cells; this specificity, however, is lost if the selectivity is tested in reconstitution experiments (for review see [2]). For the Ap adenosine receptor (a Gj/o-coupled receptor), a component that imposes an inhibitory constraint on receptor-G protein coupling [77] has been identified; this protein, termed

169 coupling cofactor, also enhances the complex formation between Apieceptor and Gi; in contrast, when complexed to Go, the Apreceptor is insensitive to the action of coupling cofactor [78]. Binding of additional components also allows for cross-talk, signal integration and coincidence detection. Calmodulin serves as an illustrative example; several GPCRs bind calmodulin, but the consequences differ substantially. Binding of cahnodulin to the C-terminus of mGluRTA [79] is mutually exclusive with binding of the GPy; in fact, calmodulin is required to release GPy from the receptors, which supports inhibition of the neuronal Ca^^ channels [79]. This coincidence detection limits presynaptic autoinhibition to those nerve terminals that were actively releasing glutamate because only these contain elevated Ca^^-levels and hence Ca^^-liganded calmodulin. Binding of calmodulin to two other Gi/o-coupled receptors, the |Li-opioid and the D2-dopamine receptor, inhibits signalling and thus allows for cross-talk by receptors that raise intracellular Ca^^. However, the mechanisms differ: calmodulin prevents access of Ga to the C-terminal part of i3 of the |i-receptor [80]. In contrast, calmodulin interacts with the N-terminal portion of i3 of the D2-receptor; this allows for the formation of a receptor-G protein complex but impedes G protein activation [81]. Formation of receptor homo- and heterooligomers GPCRs have the ability to form oligomers, a phenomenon originally observed with the P2-receptor [82]. Although the functional implications of oligomer formation is not fully understood, there are examples where dimerisation is important: as mentioned earlier, the GABAB rreceptor is retained in the ER and binds agonist with low affinity; concomitant expression of the GABAB2-receptor results in efficient plasma membrane insertion and formation of functional receptors that bind agonists with high affinity [16]. Opioid K- and 5-receptors can form dimers; the pharmacological specificity of theses dimers is distinct from that of each individual receptor [83]. It is highly likely that oligomerisation of GPCRs is a more general phenomenon that does not only involve closely related receptors; heterodimer formation has, for instance, been anticipated based on functional data [84]. In fact, the formation of heterodimers between Di- and Ai-receptors [85] was proposed as the basis for the mutual antagonism exerted by activation of the receptors. A similar form of (mutual) antagonism is thought to underlie the crosstalk between D2- and A2A- receptors [84]. However, it has to be pointed that these experiments did not provide formal proof for heterodimer formation because (i) co-localisation of proteins by immunofluorescence is only circumstantial evidence and (ii) because co-immunoprecipitation may still reflect the formation of a large complex, in which the two receptors are trapped due to their association with associated scaffolding proteins. The most rigorous proof is to visualise heterooligomers in intact cells by FRET (Foerster resonance energy transfer) or by related methods that rely on bioluminescence (BRET=bioluminescence resonance energy transfer) using appropriately tagged proteins (e.g. donor-acceptor pairs of cyan and yellow fluorescent protein). Because the efficiency of (quantal) energy transfer declines with the 6th power of distance, FRET can only occur if the two proteins exist in a complex (as opposed to a short-lived, non-productive collision). By using these stringent criteria, the formation of homo- and heterooligomers has been demonstrated for several receptors (see also Table 1). While heterodimer formation allows for direct cross-talk between receptors (e.g. co-activation in [86]), the physiological significance of homodimer formation remains a matter of

170 debate. A modest increase of BRET (and hence of oligomer formation) of appropriately tagged P2-receptors has been observed upon addition of agonist; this is unrelated to intemalisation because manipulations that prevent endocytosis do not affect the agonistinduced increase in energy transfer [87]. However, a study that allowed for time-resolved measurements of cell-surface receptors indicates that 5-receptors are constitutive oligomers [88] and that there is essentially no impact of agonist (or antagonist). GPCRs heterodimerise with other transmembrane proteins: RAMPs (receptor-activity modifying proteins) are single transmembrane proteins that associate with receptors for calcitonin-gene related peptides, adrenomeduUin and related agonists. The ligand specificity of a given receptor depends on the association with a RAMP [89]. The Ds-dopamine receptor (but not the closely related Di-receptor) and (ionotropic) GABAA-receptors directly associate which results in reciprocal (inhibitory) modulation of their activities [90]. The A2B-receptor has been identified as the dimeric partner of DCC {deleted in colorectal cancer= netrin-receptor); this dimerisation was reportedly essential for high-affinity binding of netrin-1 for the chemoattracting action of netrin-1 on commissural axons [91]. However, a re-analysis that used more reahstic concentrations of A2B-antagonist failed to document any requirement of the A2B-receptor for directing axon outgrowth [92]. Conclusion The diversity of additional proteins that bind to GPCRs is bewildering; in many cases, the significance of the reaction is difficult to assess because a given receptor, clearly, cannot interact with these proteins simultaneously. The structure of G protein heterotrimers and of rhodopsin is known to atomic resolution. Hence, the sizes of the two proteins can be compared. The intracellular surface of rhodopsin barely suffices to cover the necessary contact sites on transducin [93]. There is little, if any, space left to accommodate additional (bulky) proteins. The formation of receptor multimers may offer a way out; it suffices if one or a few receptors in a cluster are actually tethered to and immobilised by the scaffolding protein(s). Given their propensity to form complexes, the remaining receptors may aggregate by homo- and heterotypic aggregation. It is, in this context, interesting to note that G proteins can also form large aggregates in their basal state [94,95]. Eventually some of the G protein-independent effects of receptors that were listed in this review may turn out to be of modest physiological relevance. While the physiological relevance of some of the protein-protein interactions listed above needs to be estabhshed, it is clear that receptors interact with proteins other than G proteins. Hence, stimulation and blockage of a receptor by agonists and antagonists, respectively, is expected to yield effects other than those achieved by activating or inhibiting the down-stream G protein [96]. This is one of the arguments that supports the concept of direct G protein ligands as an opportunity for drug development [97,98]. References 1. Freissmuth M, Casey PJ, Gilman AG (1989) FASEB J, 3:2125-2131. 2. Gudermann T, Schoneberg T, Schultz G (1997) Annu. Rev. Neuroscl 20:399-427. 3. Janetopoulos C, Jin T, Devreotes P (2001) Science 291:2408-11. 4. Pitcher JA, Freedman NJ, Leflcowitz RJ (1998) Annu. Rev. Biochem. 67:653-692. 5. Miller WE, Lefkowitz RJ (2001) Curr.Opin. Cell Biol. 13:139-145.

171 6. 7. 8. 9.

Berman DM, Oilman AG (1998) J. Biol Chem. 273: 1269-1272. Wieland T, Chen CK (1999) Naunyn Schmiedebergs Arch. Pharmacol 360:14-26. Lohse MJ, Bliiml K, Danner S, Krasel C (1996) Biochem. Soc. Trans. 24:975-980. Takesono A, Cismowski MJ, Ribas C, Bernard M, Chung P, Hazard S, Duzic E, Lanier SM (1999) J. Biol Chem. 274: 33202-33205. 10. Cismowski MJ, Ma C, Ribas C, Xie X, Spruyt M, Lizano JS, Lanier SM, Duzic E (2000) J. Biol Chem. 275:23421-23424. 11. Bernard ML, Peterson YK, Chung P, Jourdan J, Lanier SM (2001) J. Biol Chem. 276:1585-1593. 12. Rothman JE, Wieland FT (1996) Science 272: 227-234. 13. Tang Y, Hu LA, Miller WE, Ringstad N, Hall RA, Pitcher JA, DeCamilli P, Lefkowitz RJ (1999) Proc. Natl Acad. Set USA 96:12559-12564. 14. Ringstad N, Gad H, Low P, DiPaolo G, Brodin L, Shupliakov O, DeCamilH P (1999) iVewrow 24:143-154. 15. Schmidt A, Wolde M, Thiele C, Fest W, Kratzin H, Podtelejnikov AV, Witke W, Huttner WB, Soling HD (1999) Nature 401:133-141. 16. Marshall FH, Jones KA, Kaupmann K, Bettler B (l999)Trends Pharmacol Sci. 20: 396-399. 17. Pagano A et al. (2001) /. Neurosci. 21: 1189-1202. 18. Oksche A, Rosenthal W. (1998) J. Mol Med. 76:326-37 19. Ango,F, Pin JP, Tu, JC, Xiao B, Worley PF, Bockaert J, Fagni L (2000) J. Neurosci. 20:8710-8716. 20. Kato A, Ozawa F, Saitho Y, Fukazawa Y, Sugijyama H, Inokuchi, K. (1998) J. Biol Chem. li:^: 23969-23975. 21. Okamoto T, Schlegel A, Scherer PE, Lisanti MP (1998) J. Biol Chem. 273:5419-22. 22. Huang C, Hepler JR, Chen LT, Gilman AG, Anderson RGW, Mumby SM (1997) Mol Biol Cell 8: 2365-2378. 23. Mumby, S.M. (1997) Reversible palmitoylation of signaling proteins. Curr. Opin. Cell Biol 9:14S-154. 24. Rietveld A, Simons K (1998) Biochim. Biophys. Acta 1376: 467-479. 25. Dotti CG, Simons,K. (1990) Cell 62: 63-72. 26. Winckler, B., Forscher, P., Melhnann, L (1999) Nature 397:698-701. 27.Ledesma MD, Brugger B, Bunning C, Wieland FT, Dotti CG (1999) EMBO J. 18: 1761-1771. 28. Craig AM, Banker G (1994) Annu. Rev. Neuroscience 17:267-310. 29. Wozniak M, Limbird LE (1998) Brain Res. 780: 311-322 30. Hirokawa,N (1998) Science 279: 519-526. 31. Klopfenstein, D.R., Vale, R.D., Rogers, S.L. (2000) Cell 103:537-540. 32. Tai AW, Chuang JZ, Bode C, Wolfrum U, Sung CH (1999) Cell 97: 877-887. 33. Wolfrum U, Schmitt A (2000) CellMotil Cytoskeleton 46: 95-107. 34. Saunders C, Limbird LE (1997) J. Biol Chem. Ill: 19035-19045. 35. Saunders C, Limbird LE (2000) Mol Pharmacol 57:44-52. 36. Cao TT, Deacon HW, Reczek D, von Zastrow M (1999) Nature 401:286-290. 37. Lamaze C, Fujimoto LM, Yin HL, Schmid SL (1997) J. Biol Chem. 272:20332-35. 38. Saunders C, Keefer JR, Bonner CA, Limbird LE (1998) J. Biol Chem. 273:2419624206. 39. Brakeman PR, Lanahan, AA, O'Brien R, Roche K, Barnes CA, Huganir RL, Worley PF (1997) Nature 386: 284-288.

172 40. Tu JC, Xio B, Yuan JP, Lanahan AA, Leoffert K, Li M, Linden DJ, Worley PF (\99S) Neuron 21: 717-726. 4L Tadokoro S, Tachibana T, Imanaka T, Nishida W, Sobue K (1999) Proc. Natl Acad. Sci. ^5y4 96:13801-13806. 42. Shigemoto R, Kulik A, Roberts JD, Ohishi H, Nusser Z, Kaneko T, Somogyi P (1996) Aârwre 378: 96-98. 43. Stowell JN, Craig AM (1999) Neuron 22: 525-536. 44.Dev KK, Nakajima Y, Kitano J, Braithwaite SP, Henley JM, Nakanishi S (2000) /. Neurosci. 20:7252-7257 45.ElFar O, Airas J, Wischmeyer E, Nehring RB, Karschin A, Betz H (2000) Eur. J. Neurosci. 12:4215-4221. 46. Boudin H, Doan A, Xia J, Shigemoto R, Huganir RL, Worley PF, Craig AM (2001) Neuron 28:485-497 47. Snow BE et al.(1998) J. Biol. Chem. 273: 17749-17755. 48.Zitzer H, Honck HH, Bachner D, Richter D, Kreienkamp HJ (1999) J. Biol Chem. 274:32997-33001 49. Tobaben S, Sudhof TC, Stahl B (2000) J.Biol.Chem. 275:36204-36210 50. Smith FD, Oxford GS, Milgram SL. (1999) J. Biol. Chem. 274:19894-19900. 51.Richman JG, Brady AE, Wang Q, Hensel JL, Colbran RJ, Limbird LE (2001) J. Biol. Chem. 276, in press 52. Ciruela F, Robbins MJ, Willis AC, Mcllhinney RA(1999) J. Neurochem. 72:346-54. 53. Ciruela F, Mcllhinney RA (2001) J. Neurochem. 76:750-757. 54. Roychowdhury S, Rasenick MM (1997) J. Biol Chem. 272: 31576-31581. 55.Roychowdhury S, Panda D, Wilson L, Rasenick MM. (1999) J. Biol Chem. 274: 13485-13490. 56.Ravindra, R, KunapuU SP, Forman LJ, Nagele RG, Foster KA, Patel SA. (1996) / Cell Biochem. 61:392-401 57. Popova JS, Rasenick MM (2000) J. Neurosci. 20:2774-2782. 58. Pitcher JA, Hall RA, Daaka Y, Zhang J, Ferguson SS, Hester S, Miller S, Caron MG, Lefkowitz RJ, Barak LS (1998) J. Biol Chem. 273:12316-12324 59. Carman CV, Som T, Kim CM, Benovic JL (1998). J. Biol Chem. 273:20308-20316. 60. Kirsch J, Wolters I, Triller A, Betz H (1993) Nature 366:745-748. 61. Allison DW, Chervin AS, Gelfand VI, Craig AM (2000) J. Neurosci. 20:4545-4554. 62. AH MS, Sayeski PP, Dirksen LB, Hayzer DJ, Marrero MB, Bernstein KE (1997) J. Biol Chem. 272:23382-23388. 63. Venema, R.C., Ju H, Venema, V.J., Schieffer, B., Harp, J.B., Ling, B.N., Eaton, DC, Marrero MB (1998)7. Biol Chem. 273: 7703-7708. 64. Marrero MB, Venema VJ, Ju H, He H, Liang H, Caldwell RB, Venema RC (1999) Biochem. J. 343:335-340. 65. Oldenhof J, Vickery R, Anafi M, Oak J, Ray A, Schoots O, Pawson T, von Zastrow M, Van Tol HH (1998) Biochemistry 37:15726-15736. 66. Cao W, Luttrell LM, Medvedev AV, Pierce KL, Daniel KW, Dixon TM, Lefkowitz RJ, Collins S (2000) J. Biol Chem. 275:38131-38134. 67. Heuss C, Scanziani M, Gahwiler BH., Gerber U (1999) Nat. Neurosci. 2:1070-1077. 68.Lindquist JM, Fredriksson JM, Rehnmark S, Cannon B, Nedergaard J. (2000) J. Biol Chem. 275: 22670-22677. 69. Ma, Y.C., Huang, J., Ali, S., Lowry, W., Huang, X.Y. (2000) Cell 102:635-46.

173 70. Yaciuk P, Choi JK, Shalloway D.(1989) Mol Cell Biol 9:2453-63. 71. Roth CW, Richert ND, Pastan I, Gottesman MM (1983) J. Biol Chem. 258:1076810773 72. Vernon E, Meyer G, Pickard L, Dev K, Molnar E, Collingridge GL, Henley JM (2001) Mol Cell Neurosci. 17:637-645 73. White JH, Mclllhinney RA, Wise A, Ciruela F, Chan WY, Emson PC, Billinton A, Marshall FH (2000) Proc. Natl Acad. Set USA 97:13967-13972. 74. Hall RA et al. (1998) Nature 392: 626-630. 75.Daaka Y, Luttrell LM, Ahn S, Delia Rocca G, Ferguson SS, Caron MG, Lefkowitz RJ(1998) J. Biol Chem. 273:685-688. 76. Fan GH, Yang W, Sai J, Richmond A. (2001) J. Biol Chem. 276, in press 77.Nanoff C, Mitterauer T, Roka F, Hohenegger M, Freissmuth M (1995) Mol Pharmacol 4S:S06-S\1 78. Nanoff C, Waldhoer M, Roka F, Freissmuth M (1997) Neuropharmacol 36:1211-19 79. O'Connor V, ElFar O, Bofill-Cardona E. Freissmuth M, Nanoff C, Karschin A, Airas JM, Betz H, Boehm S (1999) Science 286: 1180-1184. 80. Wang, D., Sadee, W., Quillan, J.M. (1999) J. Biol Chem. 274:22081-22088. Sl.Bofill-Cardona E, Kudlacek O, Yang Q, Ahom H, Freissmuth M, Nanoff C (2000) 1 Biol Chem. 275: 32672-32680. 82.Hebert TE, Moffett S, Morello JP, Loisel TP, Bichet DG, Barret C, Bouvier M. (1996) J. Biol Chem. 211: 16384-16392. 83. Jordan BA, Devi LA (1999) Nature 399: 697-700. 84. Fuxe K, Ferre S, ZoH M, Agnati F (1998) Brain Res. Brain Res. Rev. 26:258-273. 85. Gines S et al. (2000) Proc. Natl Acad. Set USA 97: 8606-8611 86. Rocheville M, Lange DC, Kumar U, Patel SC, Patel RC, Patel YC (2000) Science 288:154-157. 87. Angers S, Salahpour A, Joly E, Hilairet S, Chelsky D, Dennis M, Bouvier M (2000) Proc. Natl Acad Scl USA 97:3684-3689. 88.McVey, M, Ramsay D, Kellett E, Rees S, Wilson S, Pope AJ, Milligan G (2001) J. Biol Chem. 276:14092-14099. 89. Foord SM, Marshall FH (1999) Trends Pharmacol Scl 20: 184-187. 90. Liu F, Wan Q, Pristupa Z, Yu XM, Wang Y, Niznik H (2000) Nature 403: 274-280. 91. Corset V, Nguyen-Ba-Charvet KT, Forcet C, Moyse E, Chedotal A, Mehlen P (2000) Nature 407:747-750. 92. Stein E, Zou Y, Poo M, Tessier-Lavigne M. (2001) Science 291:1976-1982. 93. Palczewski K et al. (2000) Science 289:739-745. 94.Mixon MB, Lee E, Coleman DE, Berghuis AM, Gilman AG, Sprang SR. (1995) Science 270: 954-960. 95. Jahangeer S, Rodbell, M. (1993) Proc. Natl Acad Scl USA 90:8782-8786. 96. Hoeller C, Freissmuth M, Nanoff C (1999) Cell Mol Life Set 55: 257-270 97. Freissmuth M, Waldhoer M, Bofill-Cardona E, Nanoff C (1999) Trends Pharmacol Scl 20:237-245. 98.Niimberg B, Togel W, Krause G, Storm R, Breitweg-Lehmann E, Schunack W (1999). Eur. J. Med Chem. 34, 5-30. 99.Saura, CA, Mallol J, Canela EI, Lluis C, Franco R. (1998) J. Biol Chem. 273: 17610-17617.



175

Mechanisms of action of antipsychotic drugs: the role of inverse agonism at the D2 dopamine receptor Philip G Strange School of Animal and Microbial Sciences, University of Reading, Whiteknights, Reading, RG6 6AJ, UK

The antipsychotic drugs provide the main treatment for the serious brain disorder schizophrenia. They have enabled many people to lead more normal lives but their use is associated with serious side effects. Their precise mechanism of action is still unclear and it would help in the design of new antipsychotics if this could be defined more clearly. In this article I wish to consider some of the possible mechanisms that have been proposed for antipsychotic action.

Schizophrenia Schizophrenia is a serious brain disorder affecting about 1% of the population [1]. The symptoms experienced by schizophrenics are varied but may be categorised in to two groups. The positive symptoms are those that are additions to normal behaviour and comprise symptoms such as thought disorder, perceptual disturbances e.g. hallucinations, abnormal beliefs or delusions. The negative symptoms are those that subtract from normal behaviour and comprise symptoms such as social withdrawal, poverty of thought and speech, blunted affect, reduced motor function.

Antipsychotic drugs The antipsychotic drugs were first introduced in the 1950's with the use of chlorpromazine [2]. Subsequently a very wide range of drugs with differing chemical structures has been used for the treatment of schizophrenia. For the most part these drugs treat only the positive symptoms of the disorder, having little effect on the negative symptoms. The reduction in positive symptoms is very important for patients but it would be of great use to be able to treat the negative symptoms as well. Some drugs, notably clozapine, have been reported to reduce negative symptoms [3] and there is a considerable drive within the pharmaceutical industry to generate new drugs that can reduce negative symptoms. The antipsychotic drugs also elicit side effects (extrapyramidal side effects) that are mostly motor in character. Early in treatment patients can experience parkinsonian-like side effects with symptoms such as tremor and slowness of movement. After months or

176 years of treatment patients can experience tardive dyskinesia which is associated with abnormal involuntary movements of the tongue and lips. The therapeutic effects of the antipsychotic drugs are not immediate and in fact take up to four weeks to reach a maximum level. It seems that there may be some adaptive process occurring in the brain that needs time to occur.

The role of dopamine receptors in antipsychotic action Following intensive investigation of the properties of the antipsychotic drugs it became apparent that these drugs were affecting dopamine systems [2]. For example it was shown that these drugs would inhibit the abnormal behaviour seen in experimental animals following administration of amphetamine or apomorphine, both of which mimic the actions of dopamine. It was proposed that the antipsychotics were blocking the postsynaptic receptors for dopamine. In the late 1970's it became possible to assay the different dopamine receptor subtypes (Di, D2) defined using biochemical and pharmacological tests. In 1976, two labs reported a strong correlation between the affinities of a range of antipsychotic drugs for the D2 dopamine receptor and their daily doses for treating schizophrenia [4,5]. No such correlation was seen for other receptors including the Di dopamine receptor and it was concluded that antagonist action at D2 dopamine receptors was an essential part of the mechanism of action of the antipsychotics. In the late 1980's the dopamine receptors were subjected to attention by molecular biologists and five dopamine receptor sequences were cloned (Di, D2, D3, D4, D5) [6,7]. Subsequent analysis of the pharmacological properties of the gene products has shown that these may be divided in to two sub-families that are related to the original D1/D2 subdivision based on biochemistry/pharmacology. Thus Di and D5 are related to one another and to the original Di receptor and so are termed the Dplike receptors. D2, D3, D4 are related to one another and to the original D2 receptor and are termed the D2-like receptors. This meant that the effects of antipsychotic drugs on dopamine receptors could have been at D2 or D3 or D4 receptors. D3 and D4 receptors were of particular interest in this respect as they are found largely in limbic/cortical areas of brain that are likely to be regions where antipsychotic effects would be mediated. In principle, therefore, a drug with selective actions at D3 or D4 receptors might be an antipsychotic but without the extrapyramidal side effects as these are thought to be mediated via D2 receptors in the striatum, a brain region lacking D3 and D4 receptors. The role of these D2-like receptors in antipsychotic action is unclear but some clues may be obtained by considering the affinities of different antipsychotics for these receptor subtypes. If the affinities of antipsychotic drugs for the different receptor subtypes are considered then it seems that some drugs have much lower affinities for the D4 receptor whereas all of the drugs have high affinities for D2 and D3 receptors. It seems, therefore, that occupancy of D4 may not be mandatory for antipsychotic action.

177 In the years since these subtypes were defined the pharmaceutical industry has expended much effort to make compounds selective for the different subtypes, especially the D3 and D4 receptors. A D4 selective compound (L745870) was synthesised by Merck Sharp and Dohme and in clinical trials this proved to be without effect in schizophrenia [8]. This supports the idea that the D4 dopamine receptor does not seem to play a major role in the therapy of schizophrenia. We may conclude that the principal sites of action of the antipsychotic drugs are the D2 and D3 dopamine receptors [9] but the relative role of these two subtypes is unclear. Clinical testing of selective D2 and D3 receptor agents against psychosis in schizophrenia will help elucidate this. Several other receptors have been proposed as being important for the actions of antipsychotic drugs including the serotonin receptors (5HT2 and 5HTIA) [10,11]. It has been proposed that actions of the antipsychotics at these receptor subtypes may be important for suppressing side effects or promoting additional beneficial clinical actions of the drugs.

The antipsychotic drugs as inverse agonists It had been the dogma for many years that the antipsychotic drugs were acting as antagonists at the D2-like dopamine receptors. In the 1990's, however, for several drugs that had been considered to be antagonists at other receptors, it became apparent that these were in fact inverse agonists. This was seen in assays where the drugs were able to inhibit the agonist-independent activation of receptor-associated signalling systems. For the dopamine receptors it has been shown that the antipsychotic drugs possess inverse agonism at Di, D2, D3 and D5 receptors [12-16]. The inverse agonism is best characterised for the D2 dopamine receptor so I shall focus on this receptor subtype. The first report of inverse agonism at this receptor came from studies on the D2 receptor expressed in pituitary cells [15]. The antipsychotic drug, haloperidol, was found to stimulate prolactin release from these cells. Dopamine normally inhibits release so that haloperidol was acting as an inverse agonist. Subsequently, inverse agonism has been demonstrated from the ability of the drugs to potentiate forskolin stimulated cAMP production [13] or to inhibit stimulation of [^^S]GTPYS binding [16]. A detailed examination of the inverse agonism of the antipsychotic drugs was performed using a CHO cell line expressing the native D2 dopamine receptor at high levels (4-6 pmol/mg) [13]. Using these cells it was possible to show that all of the antipsychotic drugs tested exhibited inverse agonism at the D2 receptor based on their ability to potentiate forskolin stimulated cAMP production. This was seen for different chemical classes of drug and for drugs with different clinical profiles e.g. high or low side effects. The extent of inverse agonism was similar for each drug so that they all appear to be full inverse agonists and there is a good correlation between inverse agonist potency and binding affinity. Only one substance has been found so far that is a neutral antagonist. This is the aminotetralin (+)-UH-232. This compound was found to be a

178 neutral antagonist in studies on the regulation of forskolin-stimulated cAMP production in CHO cells expressing D2 receptors, and in fact this compound would inhibit competitively both the agonist effects of dopamine and the inverse agonist effects of (+)-butaclamol. The ability to detect inverse agonism, in this system with the native receptor, was very dependent on the expression level of the receptor in the CHO cells. In order, therefore, to provide a better detection system for inverse agonism at the D2 receptor we constructed a mutant receptor with changes in the amino acid sequence in the third intracellular loop at the base of the sixth transmembrane region. This is a region that has been found for other receptors to be "hot spot" for generating constitutively active receptors i.e. receptors active in the absence of agonist and therefore better for detecting inverse agonism. For the D2 dopamine receptor the mutation was T343R [17]. The T343R mutant D2 receptor exhibited the expected properties i.e. increased agonist affinity and potency and increased agonist-independent activation. The mutant receptor also gave responses to inverse agonists at lower expression levels, where the native receptor would not respond to inverse agonists. The range of antipsychotics that are inverse agonists has been extended using the T343R mutant receptor and again all antipsychotics tested exhibit inverse agonism including a set of atypical antipsychotics such as clozapine, risperidone, olanzapine, quetiapine. Despite the increased sensitivity of the T343R mutant receptor to inverse agonists (+)-UH-232 still behaves as a neutral antagonist in this test system. As a further test of the activity of (+)-UH-232 we examined its efficacy in CHO cells that had been treated with sodium butyrate in order to increase agonist responsiveness [18]. In these cells weak partial agonists exhibit increased efficacy but (+)-UH-232 remains a neutral antagonist in this system. In some systems with high amplification (FLIPR [19], microphysiometer [20]), however, (+)-UH-232 exhibits weak agonism. We may conclude, therefore, that (+)-UH-232 is a weak agonist or neutral antagonist unlike the antipsychotic drugs that are full inverse agonists.

The relevance of the inverse agonism of the antipsychotic drugs to their clinical effects It seems that all of the antipsychotic drugs possess inverse agonism at the D2 dopamine receptor. It is important to ask whether this is relevant to their clinical effects or whether it is a curiosity of the systems used to detect the activity. One way to approach this question would be to identify a neutral antagonist for the D2 receptor and assess its clinical effects. (+)-UH-232 seems to a suitable compound to use in this assessment as it has activity close to that of a neutral antagonist at the D2 receptor as outlined above. This compound has been used in a trial in schizophrenia and it was shown to be without antipsychotic activity [21]. Although these data relate only to one compound they at least support the proposal that the inverse agonism of the antipsychotic drugs is important for their clinical effects. A link, however, with the extrapyramidal side effects of the drugs cannot be excluded.

179 If this is correct then we need to ask what the mechanisms of the antipsychotic effect might be in relation to the inverse agonism exhibited by the drugs. Here we need to bear in mind that the effects of the drugs are not immediate and in fact take several weeks to reach a maximal level. There are many possible effects of the chronic treatment of drugs on neuronal systems but the delayed time course of the effects of the antipsychotic drugs suggest that there is some adaptive process occurring in the relevant neuronal system. One suggestion has been that the balance between the different dopamine neuronal systems is altered upon chronic antipsychotic treatment in favour of certain mesocortical neurones [1]. This would then change the balance of activity between different brain regions. One, well described, effect of the chronic administration of antipsychotics to experimental animals is an increase in the number of D2 dopamine receptors in the brain [22,23]. It has been shown that the chronic administration of dopamine agonists to animals will lead to a down regulation of D2 dopamine receptors. It has been assumed, therefore, that the up-regulation of D2 receptors following antipsychotic treatment is due to blockade of the access of dopamine to its receptors. It may, however, be that the effect is a reflection of the inverse agonist property of the drugs. There is some evidence from other G protein coupled receptor systems to show that such up-regulation might be a response to the inverse agonism of the drugs rather than receptor blockade alone [24]. Neutral antagonists, where these have been tested in parallel with inverse agonists, fail to elicit up-regulation. The neutral antagonists would still prevent access of agonist to the receptors and hence the lack of up-regulation with these compounds supports the idea that up-regulation is a reflection of inverse agonism. It has been shown that the antipsychotic drugs will cause up-regulation of D2 dopamine receptors expressed in recombinant cells [25]. There can be no dopamine present in these systems so the drugs cannot be preventing access of agonist to the receptors. In this case the up-regulation may provide an index of the effects of the drugs directly on the receptor. The effects of the drugs may, therefore, be a reflection of the conformational change that these compounds elicit in the receptor. We have examined a range of drugs for their effects on D2 receptors expressed in CHO cells and preliminary data suggest that the antipsychotic drugs tested do cause D2 receptor upregulation whereas (+)-UH-232 does not [26]. These observations provide a further suggestive link between the inverse agonism of the antipsychotic drugs and their clinical effects. Therefore a plausible theory of antipsychotic action may be proposed as follows. In this, the antipsychotic drugs are inverse agonists and this property is important for their therapeutic effects. The long-term use of the drugs changes the sensitivity of certain synapses in the brain and thus achieves an antipsychotic effect. It is still unclear as to which synapses in the brain are affected, although some suggestions have been made above. It is also unclear as to whether the up-regulation of D2 dopamine receptors in the brain is part of this change in synaptic efficacy.

180 References 1. Strange PG (1992) Brain Biochemistry and Brain Disorders, OUP 2. Leysen JE and Niemegeers CJE Handbook of Neurochemistry 9 (1985) 331 3. Kane JM. CNS Drugs 7 (1997) 347-348 4. Seeman P, Lee T, Chau-Wong M amd Wong K Nature 261 (1976) 717 5. Creese I, Burt DR and Snyder SH Science 194 (1976) 546 6. Civelli, O, Bunzow JR and Grandy DK. Ann. Rev. Pharmacol.Toxicol. 32 (1993) 281 7. Missale C, Nash SR, RibinsonSW, Jaber M and Caron MG Physiol Rev 78 (1998) 189 8. Bristow LJ, Kramer MS, Kulagowski J, Patel S, Regan CI and Seabrook GR. Trends Pharmacol Sci 18(1997)186 9. Strange PG Pharmacol Rev 53 (2001) 119 10. Meltzer HY, Park S and Kessler R Proc Natl Acad Sci USA 96 (1999) 13591 11. Millan MJ J Pharmacol Exp Therap 295 (2000) 853 12. Charpentier S, Jarvie KR, Severynse DM, Caron MG and Tiberi M J Biol Chem 271(1996)28071 13. Hall DA and Strange PG Brit J Pharmacol 121 (1997) 731. 14. Griffon N, Pilon C, Sautel F, Schwartz JC and Sokoloff P J Neural Trans 103 (1996) 1163 15. Nilsson CL and Eriksson F. J Neural Trans 92 (1993) 213 16. Kozell LB and Neve KA Mol Pharmacol 52 (1997) 1137 n.Wilson J, Fu D, Lin H, Javitch JA and Strange PG J Neurochem 77 (2001) 493 18. Marston DM and Strange PG, Brit J Pharmacol 131 (2001) 5P 19. Pauwels PJ, Finana F, Tardif S, Wurch T and Colpaert FC J Pharmacol Exp Therap 297 (2001)133 20. Coldwell MC, Boyfield I, Brown AM, Stemp G and Middlemiss DN Brit J Pharmacol 127(1999) 1135 21.Lahti AC, Weiler M, Carlsson A and Tamminga CA. J Neural Trans 105 (1998) 719. 22. Creese I and Sibley DR Ann Rev Pharmacol Toxicol 21 (1981) 357 23. Lidow MS, Williams GV and Goldman-Rakic PS Trends Pharmacol Sci 19 (1998) 136 24. Leurs R, Smit M, Alewijnse AE and Timmerman H Trends Biochem Sci 23 (1998) 418 25. Sibley DR and Neve KA (1997) in The dopamine receptors. Neve KA and Neve RL eds Humana Press, Totowa, New Jersey pp383-424 26. Marston DM, Kennedy M and Strange PG unpublished observations


181

Agonist channeling of a2-adrenoceptor function Karl E.O. Akerman, Johnny NSsman, Tomas Holmqvist and Jyrki P. Kukkonen Department of Physiology, Division of Cell Physiology, Uppsala University, BMC, Box 572, SE-75123 Uppsala, Sweden

Introduction The a2-adrenergic receptors (a2-ARs) control the function of different organs via central and peripheral effects (Table 1). Ligands for these receptors have several therapeutic applications includmg anaesthesia and treatment of hypertension and glaucoma. Potential new indications may include obesity and psychiatric disorders. The central effects are thought to mainly be a result of presynaptic mhibition of noradrenaline release. However, a2-ARs are localised at postsynaptic sites as well. The mechanisms involved in the peripheral actions of a2-AR ligands are less well understood. Their effects can be both inhibitory and stimulatory depending on the organ in question [1].

Table 1. Physiological effects of ai adrenoceptor activation Central effects • Reduction in plasma adrenaline and noradrenaline • Inhibition of neurotransmitter release • Sedation • Food intake • Hypotermia • Antinociception

Peripheral effects Inhibitory Stimulatory • Insulin secretion • Vascular tonus (e.g. • Vasopressin secretion coronary & pulmonary • Tyroxine secretion venal tonus) • Saliva secretion • TSH secretion • Gut movement (ileus) • Growth hormone • Lipolysis secretion p Glycogenolysis • Platelet aggregation • Water absorption (gut)

The variety of responses elicited by these receptors indicates that they may possess the ability to activate multiple signal transduction pathways [2]. The response elicited by the receptors depends on the tissue where these receptors are expressed as well as the subtype of receptor [3-10]. The classical cellular response observed upon activation of a2-ARs is an inhibition of sthnulated cAMP production [11]. However, changes in cAMP alone do not explain the physiological actions of a2-ARs. Three subtypes of a2ARs, a2A, 0C2B and a2c, have been identified by cloning and pharmacological tools.

182 Many of the systemic eflfects of a2-AR ligands are due to actions on presynaptic receptors, which inhibit transmitter release from in neuronal cells via inhibition cf voltage-gated N- or P/Q-type Ca^^ channels [12-14]. Like the inhibition of cAMP production this response is sensitive to pertussis toxin and therefore likely to be mediated by Gi/o-type G proteins. Stimulation of cAMP accumulation Activation of a2-AR has also been shown to cause an increase rather than a decrease of cAMP production both through recombinantly expressed [5,9,10,15] and endogenous a2-ARs [16-18]. In certain cells this stimulatory response is seen only at high agonist concentrations or if Gi-type G proteins are inactivated by pertussis toxin-treatment [6,10]. In many cell types the primary response seems to be stimulation of cAMP production and in many cases particularly the a2B-AR subtype has been implicated in the stimulatory coupling [10,19,20]. With all the subtypes, there are some agonists (e.g. clonidine, oxymetazoline, UK 14,304), though different for each subtype, that prefer coupling to an inhibitory response and have only a very weak stimulatory action [20-22]. The phenomenon where certam agonists can selectively activate specific signal pathways has been termed agonist trafficking (of receptor signals) [23]. The basis for this is that different active receptor conformations will preferentially activate specific G proteins. Adenylyl cyclase (AC) can be stimulated directly through GSOL and directly or indirectly by other messengers [24] (Table 2). In many studies the stimulation cf cAMP production by a2-ARs and other Gj-coupled receptors has thus been interpreted as Gs coupling. However, many other possibilities exist such as effects via G protein p y subunits (Gjiy)? Ga^^ and protein kinase C (PKC). The effect of receptor activation would be dependent on the transduction mechanisms it utilises in the cells in question and on the respective AC expressed in the cells studied. In PC-12 cells, for instance, chelation of intracellular Ca^^ with BAPTA reduces the a2-AR-stimulated cAMP accumulation [5]. Table 2. Activation of adenylyl cyclase isoforms by different effectors Isoform activator Inhibitor Typel Type III, VIII Type II, IV, VII Type V, VI Type IX

Gsa, Ca'* Gâ, Ca^"^ G,a, PKC, GpY

Gsa Gsa

GpY, Gitt

Gia Gia (?) Gitt, Ca^^

Gia

To test whether py subunits would be involved in the stimulation of cAMP accumulation via the human a2B-AR in Sf9 cells the effects of Gpy and P-ARK (which functions as a Gpy scavenger) on forskolin-stimulated cAMP accumulation was tested (Fig. 1). While p-ARK enhanced forskolin-stimulated cAMP accumulation the GPy coexpression considerably reduced cAMP accumulation. This means that the adenylyl cyclase of Sf9 cells is inhibited by Py rather than stimulated. The a2B-AR-stimulated AMP accumulation in Sf9 cells can thus not be due to Py. The basal and forskolin-

183 stimulated cAMP accumulation are unaffected by stimulation of protein kinase C while a considerable stimulation of coexpressed type II adenylyl cyclase can be observed (manuscript in preparation). Ca^"*" enhances the forskolin-stimulated cAMP accumulation [25]. However, stimulation of cAMP accumulation is seen also in cells depleted of intracellular Ca^"*" [10]. Thus, the typical second-messenger- or CPy-niediated effects on adenylyl cyclase can be excluded in the case of the a2B-AR.

p-ARK

3-1

^2.5!

control

0

3

21 —

0 j = .

0.5^ n J

n basal forsk

r^ basal

forsk

n basal

forsk

Fig. 1. Effect of P-ARK and Gpy expression on basal and forskolin-stimulated (forsk) cAMP accumulation in Sf9 cells. The methods used are essentially those described in references [10,22,25]. Cells were infected with recombinant baculovirus harboring the genes for either PARK (gift from Dr. A. DeBlasi, Consorzio Negri Sud, Santa Maria Imbaro, Chieti, Italy) or pi and Y2 (gifts from Dr. T. Haga University of Tokyo, Tokyo, Japan) subunits in the same recombinant virus. The infection time was 26 hours. Activated ai-ARs have been shown to coprecipitate Gs proteins [6]. Mutagenesis studies on a2A-AR and a2B-AR suggest that the stimulatory response can be specifically modified by structural changes in the second intracellular loop [22] and in the Nterminal or C-terminal portion of the third intracellular loop [26]. Certain basic residues in the C-terminal portion are important for the stimulatory coupling [27]. The approximate positions of these domains in the intracellular loops of rhodopsin [28] are illustrated in Fig. 2. These findings, together with the fact that certain agonists can preferentially couple to inhibition of adenylyl cyclase, suggest that the inhibition and stimulation of adenylyl cyclase via a2-ARs require different structural determinants or are induced by different conformations of the receptor.

184

Fig. 2. The approximate position of domains important for stimulatory coupling super imposed on the intracellular surface of rhodopsin deduced from its crystal structure . The outline of the Cterminus second (i2) and third (i3) loops as well as the approximate positions of helices III, IV, V and VI are marked. Note that the i3 loop of a2-ARs is much longer than the corresponding loop in rhodopsin. The areas where stimulation of cAMP accumulation by a2-ARs are affected by mutagenesis are dotted black. The ability of different a2-AR agonist to enhance cAMP accumulation as compared to their ability to activate coexpressed Gi, determined by ^^S-GTPyS binding, is shown in Fig. 3. Many of the ligands, like UK14,304 and clonidine, showed a weaker ability to stimulate cAMP accumulation while they were almost as effective as noradrenaline in coupling to Gi. With Gi coexpression these ligands inhibited rather than stimulated cAMP accumulation (Fig. 4). A small inhibiton was seen even with those ligands, which showed a high degree of stimulation when the Gi expression levels were further increased with longer incubation times. These results demonstrate that Gi can override the stimulatory coupling. They also indicate that the ability of the different ligands to couple to Gi is similar but the coupling to Gs is ligand-dependent. Thus when the receptor is activated by these ligands its ability to couple to stimulation of cAMP production is weaker than the coupling to inhibition. This would suggest that the Gs "affinity" of the receptor conformation induced by these ligands is lower. The results also suggest that Gi coexpression considerably attenuates the stimulatory coupling.

185 OH

H

Y

HO

adrenaline

CH3

HjC

^

(CH^C^

noradrenaline

^ oxymetazoline

CH3

N

^^xL ^CH

HO

dexmedetomidine

a-methyl-noradrenaline

guanabenz

NH2

t> ^"CP" BHT933

UK 14,304

clonidine

r 1.2 O

3

5'

crq

o

o

N

C

I O

7^ O

•«

N

£

o

B o

0)

^

C

s

to/)

^ 0

m en

^ ^

H

m T-H

;D

G^

UJ «

S X T3

Fig. 3 Effect of different a2-AR agonists on cAMP accumulation (empty bars) and S-GTPyS binding (gray bars) to co expressed Q (Qa and Gpy). The methods used are essentially those described in references [10,22,25]. The data is normalised with respect to noradrenaline.

186

u s

0.3

o

.s o

I o

c3

N

2

*| o

s s

T3

o

X

S 1 d

S

2~ E-OH + p

O

CO2"

Nu-E

H N.â)02

CO2 0(?

•' co2-oyKl^

'CO2 strand #2 ^ I QlcNAcI

l^cc strand #1

H H "

O

»

r I MurNAcp^V^N"^ ^002 strand #2

Figure 1. The crosslinking of the peptidoglycan stands in the bacteria cell-wall synthesis. The cross-linking of the bacterial cell wall is catalyzed by transpeptidases, members of the penicillin-binding protein (PBP) family. These enzymes pursue an active site-serine strategy in their catalytic reaction. As depicted in Figure 1, the active site serine in these enzymes reacts with the amide carbonyl group of the penultimate DAla in the peptidoglycan structure, giving rise to an acyl-enzyme species. On binding of

196 the second strand of peptidoglycan to the active site, the side chain amine is promoted to approach the ester carbonyl to give the amide bond in the product Transpeptidases are the targets of p-lactam antibiotics, such as penicillins and cephalosporins. Tipper and Strominger [23] had proposed that the conformation of acyl-D-Ala-D-Ala portion of peptidoglycan mimics the conformation of the backbone of the p-lactam antibiotics. As such, the antibiotic would acylate the active site serine of these enzymes, and the acyl-enzyme species resists deacylation, resulting in the bacterial cell death. We proposed that this property of the cephalosporin nucleus could be used to design a novel molecule (1) that would acylate the active site of these enzymes, as would the first strand of the peptidoglycan. However, we incorporated in the design of the molecule features that would mimic the incoming second strand of peptidoglycan (2). On acylation of the active site of the transpeptidation, the complex would have features of the two peptidoglycan strands sequestered in the active site, as would the two strands of the peptidoglycan in the course of the cross-linking reaction. Essentially, we would have a snapshot of the two strands of the peptidoglycan just prior to the cross-linking event.

mimics the first strand of peptidoglycan after acylation of the active site serine

^

mimics the approaching second strand of peptidoglycan

PBP E-OH

CO2 -^ ^ ^^NH 1+ CO2 ^ O0.^^h ""V^ H"

"

O

I

A

N"^C02

Synthesis of compound 1 has been reported [24]. This compound indeed modifies the active site serine of the bifunctional transpeptidase/DD-peptidase from species Streptomyces R61. The structure of the complex that was solved at 1.2 A resolution provided a detailed knowledge of the specific interactions in the active site of this bacterial enzyme that leads to the cross-linking event, as reported elsewhere [24]. This structure was the building point for the generation of a model for the two full peptidoglycan strands up to the NAG-NAM repeat units. The energetics of the model were evaluated in the course of dynamics simulations and energy minimization procedures [24]. These models provide information on the binding sites for the two substrates and they reveal two grooves on the surface of the enzyme that are the likely locations for the binding of the polymeric NAG-NAM repeat backbones [24]. Paul Ehrlich prophetically stated that "drug resistance follows the drug like a faithful shadow" [11]. There are seven major classes of antibacterials that are currently in clinical use. These are p-lactams, fluoroquinolones, aminoglycosides, glycopeptides, macrolides, tetracyclines and sulfonamide [25]. Cases of resistance to all are known.

197 Indeed, resistance to any one class by itself is a relatively rare occurrence, and one actually sees often resistance to multiple classes of antibiotics in pathogens. There are five common strategies that microorganisms have evolved for resistance to antibacterials. These are resistance due to reduced permeability into the organisms (plactams, fluoroquinolones, and folate inhibitors), altered target site (p-lactams, aminoglycosides, tetracyclines glycopeptides, fluoroquinolones, and folate inhibitors), resistance enzymes (p-lactams, aminoglycosides, macrolides, chloramphenicol), efflux mechanisms (tetracyclines and fluoroquinolones), and target by-pass (sulfonamide and trimethoprim) [25]. Of these, the only two that are amenable to intervention by medicinal chemists are the strategies by efflux mechanisms and by resistance enzymes. In essence, if one were to inhibit the efflux pumps or the resistance enzymes, the viability of the original antibiotic could be restored. The Mobashery laboratory has had a long-standing interest in enzymes of antibacterial resistance. We have devised strategies to either inhibit or circumvent these enzymes over the years. One recent effort centered on developing new classes of antibiotics that would bind to the validated targets that are known for the existing classes of antibiotics. In one such work, we commenced by investigating the structure of the complex of paromomycin, an aminoglycoside antibiotic, bound to the acyltransfer site ("A site") of bacterial ribosome [26]. We retained as essential elements two rings of paromomycin for binding to the RNA structure. These two rings made important electrostatic interactions with the RNA moieties and produced the core of the structure within the A site that was retained. The remainder of the structure of paromomycin was eliminated in silico. We explored a number of three-dimensional compound data banks for molecules that would bind to the space that was within the A site, which was created by the elctronic elimination of the functionalities that were removed. This exercise was carried out using the program DOCK, which sampled 380,000 such molecules. The top 100 scoring molecules that were shown to have affinity for binding at the subsites contiguous with those of the two retained rings were scrutinized closely. Several revealed the possibility of attachment to the retained core of the two sugars. A number of the molecules were synthesized and they were shown to bind to the A site in in vitro experiments. We also showed that some of the synthetic molecules were not susceptible to the action of three of the common enzymes of resistance to aminoglycosides. The project disclosed here is one step in the incremental process necessary to move away from the structure of aminoglycosides, while retaining the ability to bind to a validated target in bacteria for antibiotics, namely the A site of the bacterial ribosome. The use of aminoglycoside antibiotics has facilitated evolution of a number of enzymes that structurally modify these antibiotics, whereby the affinity for binding to the bacterial ribosome is lost [27], [28], [29], [30]. This is indeed a very common mechanism for resistance to these antibacterials. The most common mechanism for resistance to these antibiotics is by the activity of aminoglycoside 3 ' phosphotransferases [APH(3')s]. These enzymes transfer the y-phosphoryl group of ATP to the 3'-hydroxy 1 of aminoglycosides. This modification interrupts the essential aforementioned electrostatic interactions between the ribosomal RNA and aminoglycoside, disfavoring the binding process. These enzymes are usually expressed in high copy numbers and often operate at the diffusion limit [27]. Despite the fact that

198 they are cytoplasmic [27], as is the target ribosome, they confer resistance extremely effectively. It occurred to us that if we could manipulate the structure of the antibiotic such that the 3' carbon were to bear a geminal dihydroxy species (a hydrated ketone), we may be able to reverse the process of drug phosphorylation. The concept is disclosed in Figure 2. Compound 3 was synthesized from kanamycin A in 11 steps [31]. The mixture of the hydrated version of compound 3 (compound 4) to the ketonic version was 3:1 at neutral pH and at room temperature. We demonstrated that compound 4 underwent phosphorylation by APH(3')s, followed by the spontaneous release of the inorganic phosphate. Hence, a "futile cycle" ensues, in which ATP is hydrolyzed. It is noteworthy that the minimal inhibitor concentration of kanamycin A, the parental compound to 4, went up by 1000-fold in the organisms that produced the APH(3')s over the background strain that did not express the enzyme. The similar situation increased the MIC for compound 4 by a mere four-fold [31]. If ATP could be set up for recylcing, as shown in Figure 2, the process of phosphorylation and elimination of phosphate could go on for as long as there exists NADH. Once exhausted, the process could be resumed by the addition of NADH, and as expected, the event was dose dependent [31]. The major cause of resistance to p-lactam antibiotics comes about as a consequence of the catalytic action of p-lactamases. As depicted in Figure 3, these enzymes hydrolyze the P-lactam moiety of these antibiotics, the product of the reaction lacks antibiotic property. Whereas the first p-lactamase was discovered in the early 1940s, prior to the beginning of the use of penicillins in the clinic, the discoveries of the novel forms of these enzymes did not take place until much later in the 1980s and 1990s [32], [33]. There are over 340 of these enzymes known to date and they fall into four classes, classes A, B, C, and D. Class B enzymes are zinc-dependent, whereas the remaining three classes are active-site-serine enzymes.

199

A

NH2

0^^0^g;^NH2

-a

OH

}H

'^-^^ y( NH2

> NH2

/

NH2

^8h ^

ATP

-oV„-

a.Po-

OH

ADP

-Xf

pyruvate kinase

o

NAD"^

lactate dehydrogenase

PEP (excess)

B

NADH

îÔH

80 ^xM NADH

/

72^MNADH

/ e

0.4 40 ÂM NADH

en

/

CD

16 H,M NADH

0.?.

/

o

1000

2000

3000

4000

Time (s) Figure 2. (A) The coupled spectrophotometric assay of pyruvate kinase and lactate dehydrogenase recycles ATP, which is consumed by the APH(3')-mediated phosphorylation of 4. (B) The gradual disappearance of the chromophore for NADH, as a function of phosphorylation of 4, was monitored by the decline in absorbance at 340 nm. The process is continuous until PEP (180-fold excess) is entirely depleted (details in ref 20). Each run was initiated/reinitiated by the addition of the given quantity of NADH to the recycling mixture.

200

There is clear evidence by now that p-lactamases and PBPs are related to each other [34], [35]. The issue of kinship of this family of enzymes is addressed elsewhere in greater depth [35]. But, it would appear that nature has selected four distinct mechanisms for resistance to P-lactam antibiotics [36]. The zinc-dependent enzymes are distinctly different, but so are the serine enzymes. In the case of class A enzymes, the hydrolytic water approaches the acyl-enzyme species from the a direction and the water molecule is promoted by the invariant residue Glu-166 as the general base [37], [38], [39], [40]. In the case of the class C enzymes the water is promoted from the opposite P direction. It is believed that the collective environment of the hydrolytic water promotes it in the reaction. We and others have shown that there is an electrostatic interaction with the nitrogen of the acyl-enzyme species, hence substrateassisted catalysis applied here [41], [42]. The mechanism of class D enzymes is interesting. These enzymes utilize a carbamylated lysine residue in the active site for both the acylation and deacylation chemistries [43]. Hence, the catalytic events in class D enzymes, in contrast to enzymes of classes A and C, are synunetrical [44]. Only a mere handful of proteins are known to have carbamylated lysine in their structures.

p-lactamase

CO2

>-'

N—1^^

further degradation

p-lactamase

CO2

Figure 3. Hydrolysis of P-lactams by P-lactamases

These mechanistic differences underscore the clinical difficulties in fighting the deleterious action of p-lactamases. It is desirable to have inhibitors that work against more than one class of enzyme. This expectation has not been fulfilled to date, and indeed the examples of inhibitors that would inhibit two classes of these enzymes are extremely rare [45], [46], [47]; none have made it to the clinic to date. There exist combinations of three inhibitors for class A P-lactamases, clavulanate, sulbactam or tazobactam, with penicillins that are being used in fighting infections [32]. These inhibitors all work by the same mechanism in inhibiting the enzymes of class A [32]. Over the past few years, variants of these class A enzymes have been discovered that are capable of performing their function as resistance enzymes, but they have become resistant to inhibition by these clinically used inhibitors

201 [48], [49], [50], [51]. Since the mechanism of inhibition by all three inhibitors is the same, there is the potential that resistance to inhibition by one inhibitor might impart the same property to others. Therefore, there is a need for additional classes of inhibitors that inhibit the enzyme by distinct mechanisms. We have reported on additional types of inhibitors for p-lactamases recently. The chemistry in each case takes advantage of the reactivity of the active site serine for the onset of inhibition. Monobactam inhibitors are exquisitely effective in inhibition of class A plactamases, and they are competitive non-covalent inhibitors of class C enzymes [52], [53], [54], [55]. These inhibitors acylate the active site serine rapidly, followed by the elimination of the good living group. There are four possibilities for the inhibitory species (6-9), two of which, 6 and 7 have been observed crystallographically.

/^^3

cJrrk ,0>O.Ts Ser-70 |-^^

PHa

0=( ^^0-Ts 3 H Ser-70 5 — ' ^ p-elimination

/)H3

r—J^ 0=(

O "S 0-Ts Ser-70

NH2

Ser-70|-^

tautomerization

/=\ ^

0==(^NH2

Ser-701—^

6

8 hydrolysis /

hydrolysis tautomerization

of\ Ser-70 \ - ^ 7

^"3

0==PbH Ser-70

\—^ 9

AT O 0-Ts 12 Ser-70

202 Compound 10 is of special interest. This compound inhibits the class A TEM1 P-lactamases within seconds and resists deacylation of the inhibitory species for days [53]. The x-ray structure revealed the presence of the iminium species 11 for the inhibitory species [53]. Compound 12 proved extremely effective in inhibition of the broad-spectrum class A enzyme NMCA from Enterobacter cloacae [55]. X-Ray analysis revealed that the carbonyl species of 12 exists in two conformations after acylation of the enzyme. One ensconced in the active site oxyanion hole, the other with the carbonyl flipped out of this space [55]. We explored the process of this conformational change in the active site by computational simulations. Within the time scale of picoseconds, one sees dynamics motion of the inhibitor from one species to the other, a process that repeats itself, indicative of the facility of the transition from one species to the other. It was also demonstrated that the clinically used carbapenem antibiotic imipenem also undergoes the same motion of the acyl-enzyme species in an out of the oxyanion hole [56]. Another class of inhibitors for the class A enzymes are based on the structure of penicillanate. These compounds acylate the enzyme active site and the acyl-enzyme species resists deacylation. In essence, the hydroxyalkyl moiety at the 6a position prevented the approach of the hydrolytic water from that direction in the acyl-enzyme species. Therefore, these moelcules acylate the enzyme and resist deacylation for a number of hours [40], [57], [58].

COf-

O

cor

o

-N02

r ^ N

1 further degradation

In the U.S. alone 50 million pounds of antibiotics were used in 1996, and the trends over the past decades would indicate that these numbers increase on an annual basis. A key general feature of antibiotics is that they arfe often not metabolized in the body as they are eliminated from patients. Hence, these biologically active molecules are introduced into the sewer system and ultimately to the environment with plenty of exposure to bacterial. Antibiotics have the ability to select for resistance well after their

203

elimination from patients.lt occurred to us that one could design antibiotics that could self-destruct after use in humans. Many such antibiotics can be envisioned, but as a proof of concept, we have recently reported on cephalosporin 13 [59]. This compound unmasks a hydrazine derivative from its side chain on exposure to either visible or ultraviolet light. The so-generated hydrazine destroys the p-lactam moiety of the antibiotic, rendering it inactive (15). Cephalosporin 13 has activity against both Gramnergative and Gram-positive bacteria. It also serves as substrate to P-lactamases, so it functions like a typical p-lactam antibiotic. Development of antibiotics is a labor-intensive process and expensive. It is incumbent upon us to make certain that their clinical viability is extended as long as possible. To begin to understand the processes that are necessary in development of antibiotics and occurrence of mechanisms of resistance, it is necessary that these be studied at the molecular level. This manuscript described some of our efforts in this direction. ACKNOWLEDGEMENTS The work in France was funded in part by the Programme de Recherche Fondamentale en Microbiologic (MENRT) and CNRS. The work in the USA was supported by grants from the National Institutes of Health (AI33170 and GM61629) and the National Science Foundation (SM).

REFERENCES [I] [2] [3] [4] [5] [6] [7]

[8] [9] [10] [II] [12] [13] [14]

J. Lederberg, Science 288 (2000) 287. The World Health Report 1996. Fighting Disease, Fostering Development. World Health Organization, Geneva, 1996. J. Lederberg, R. E. Shope, Eds., Emerging Infections: Microbial Threats to Health in the United States, National Academy Press, Washington, D. C , 1992. S. Mobashery, E. Azucena, Encyclopedia Life Sci. London. Nature Publishing Group, UK, 2000 (www.els.net). J. Davies, Trends Curr. Biol. 9 (1999) M2. 283. K. J. Poole, J. Pharm. Pharmacol. 53 (2001). K. K. Wong, D. L. Pompliano, Resolving the Antibiotic Paradox: Progress in Understanding Drug Resistance and Development of New Antibiotics: Peptidoglycan biosynthesis: Unexploited antibacterial targets within a familiar pathway, Eds. B.P. Rosen, and S. Mobashery, New York: Plenum Press, 1998. W. M. M. Kirby, Science, 99 (1944) 452. R. D. Gonzales, P. C. Schreckenberger, M. B. Graham, S. Kelkar, K. DenBesten, and J. P. Quinn, The Lancet, 357 (2001) 1179. http://www.ncbi.nlm.nih.gov L. P. Kotra, S. Vakulenko, S. Mobashery, Microbes Infect. 2 (2000) 651. A. E. Allsop, Curr. Opin. Microbiol. 1 (1998) 530. D. T. Moir, K. J. Shaw, R. S. Hare, G. F. Vovis, Antimicrob. Agents Chemother. 43(1999)439. D. J. C. Knowles, F. King, Resolving the Antibiotic Paradox: Progress in

204

[15] [16] [17]

[18] [19] [20] [21] [22] [23] [24] [25] [26] [27]

[28] [29] [30] [31] [32]

[33] [34] [35] [36]

Understanding Drug Resistance and Development of New Antibiotics: The Impact of Bacterial Genomics on Aantibacterial Discovery, Eds. B.P. Rosen, S. Mobashery, Plenum Press, New York, NY, 1999. J. H. Gate, M. M. Yusupov, G. Zh. Yusupova, T. N. Earnest, H. F. NoUer, Science, 285 (1999) 2095. G. C. Agalarov, G. Sridhar Prasad, P. M. Funke, C. D. Stout, J.R. Williamson, Science, 288 (2000) 107. B. T. Wimberly, D. E. Brodersen, W. M. Clemons Jr., R. J. Morgan-Warren, A. P. Carter, C. Vonrhein, T. Hartsch, and V. Ramakrishnan, Nature, 407 (2000) 327. M. M. Yusupov, G. Zh. Yusupova, A. Baucom, K. Lieberman, T. N. Earnest, J. H. D. Gate, and H. F. Noller, Science, 292 (2001) 883. J. M. Ogle, D. E. Brodersen, M. C. William, Jr., M. J. Tarry, A. P. Carter, V. Ramakrishnan, Science, 292 (2001) 897. S. C. Blanchard, and J. D. Puglisi, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 3720. C. T. Walsh, Nature, 406 (2000) 775. C. Goffin, J. M. Ghuysen, Microbiol. Mol. Biol. Rev. 62 (1998) 1079. D. J. Tipper, and J. L. Strominger, Proc. Natl. Acad. Sci. U.S.A. 54 (1965). W. L. Lee, M. A. McDonough, L. P. Kotra, Z. H. Li, N. R. Silvaggi, Y. Takeda, J. A. Kelly, S. Mobashery, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 1427. L. P. Kotra, D. Golemi, S. Vakulenko and S. Mobashery, Chemistry & Industry (2000) 339, J. Haddad, L. P. Kotra, B. Llano-Sotelo, E. F. Azucena Jr., M. Liu, S. B. Vakulenko, C. S. Chow and S. Mobashery, (unplublished results). G. D. Wright, A. M. Berghuis, S. Mobashery, Resolving the Antibiotic Paradox: Progress in Understanding Drug Resistance and Development of New Antibiotics: Aminoglycoside Antibiotics: Structures, Functions and Resistance, Eds. B.P. Rosen, S. Mobashery, Plenum Press, NY, 1998. L. P. Kotra, J. Haddad, S. Mobashery, Antimicrob. Agents Chemother. 44 (2000) 3249. E. Azucena Jr., I. Graspas and S. Mobashery, J. Amer. Chem. Soc. 119 (1997) 2317. G. D. Wright, Curr. Opin. Microb. 2 (1999) 499 J. Haddad, S. Vakulenko, S. Mobashery, J. Am. Chem. Soc. 121 (1999) 11922. K. Bush and S. Mobashery, Resolving the Antibiotic Paradox: Progress in Understanding Drug Resistance and Development of New Antibiotics: How pLactamases Have Driven Pharmaceutical Drug Discovery: from Mechanistic Knowledge to CHnical Circumvention. Eds. B.P. Rosen and S. Mobashery, Plenum Press, NY, 1998 K. Bush, Clin. Met. Dis. 32 (2001) 1085. J. A. Kelly, A. P. Kuzin, P. Charlier, and E. Fonze, Cell. Mol. Life Sci. 54 (1998) 353. I Massova and S. Mobashery, Antimicrob. Agents Chemother. 42 (1998) 1. L. P. Kotra, J. P. Samama, S. Mobashery, Bacterial Resistance to Antimicrobials, Mechanisms, Genetics, Medical Practice and Public Health: Structural Aspects of

205

[37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59]

p-Lactamase Evolution, Eds. A. Lewis, Salyers, H. Haber, R. G. Wax, Marcel Dekker, Inc, 2001 (in press). O. Herzberg. and J. Moult, Science, 236 (1987) 694. J. R. Knox, and P. C. Mowes, J. Mol. Biol. 220 (1991) 435. N. C. Strynadka, H. Adachi, S. E. Jensen, K. Johns, A. Sielecki, C. Betzel, K. Sutoh, and M. N. James, Nature, 359 (1992) 700. K. Miyashita, I. Massova, P. Taibi and S. Mobashery, J. Am. Chem. Soc. 117 (1995)11055. A. Bulychev, I. Massova, K. Miyashita, and S. Mobashery, J. Am. Chem. Soc. 119(1997)7619. A. Patera, L. C. Blaszczak, B. K. Shoichet, J., Am. Chem. Soc. 122 (2000) 10504 L. Maveyraud, D. Golemi, L. P. Kotra, S. Tranier, S. Vakulenko, S. Mobashery, J. P. Samama, Structure, 8 (2000) 1289. D. Golemi, L. Maveyraud, S. Vakulenko, S. Tranier, A. Ishiwata, L. P. Kotra, J. P. Samama, S. Mobashery, J. Am. Chem. Soc. 122 (2000) 6132. A. Bulychev, I. Massova, S.A. Lerner, S. Mobashery, J. Am. Chem. Soc. 117 (1995) 4797. T. Nagase, D. Golemi, A. Ishiwata, S. Mobashery, Bioorg. Chem. (2001) (in press). J. D. Buynak, V. R. Doppalapudi, G. Adam. Bioorg. Med. Chem. Lett. 10 (2000) 853. S. B. Vakulenko, B. Geryk, L. P. Kotra, S. Mobashery, S. A, Lerner, Antimicrob. Agents Chemother.43 (1998) 1542. E. B. Chaibi, D. Sirot, G. Paul, R. Labia, J. Antimicrob. Chemother. 43 (1999) 447. R. A. Bonomo, and L. B. Rice, Front. Bioscience. 4 (1999) e34. C. Therrien, R. C. Levesque, FEMS Microb. Rev. 24 (2000) 251. A. Bulychev, M.E. O'Brien, L Massova, M.Teng, T.A. Gibson, M.J. Miller, S. Mobashery, J. Am. Chem. Soc. 117 (1995) 5938. P. Swaren, I. Massova, J. Bellettini, A. Bulychev, L. Maveyraud, L. P. Kotra, M. J. Miller, S. Mobashery, J. P. Samama, J. Am. Chem. Soc. 121 (1999) 5353. A. Bulychev, J. R. Bellettini, M. O'Brien, P. J. Crocker, J. P. Samama, M. J. Miller, S. Mobashery, Tetrahedron, 56 (2000) 5719. L. Mourey, L. P. Kotra, J. Bellettini, A. Bulychev, M. O'Brien, M.J. Miller, S. Mobashery, J. P. Samama, J. Biol. Chem. 274 (1999) 25260. L. Maveyraud, L. Mourey, L. P. Kotra, J. D. Pedelacq, V. Guillet, S. Mobashery and J. P. Samama, J. Am. Chem. Soc. 120 (1998) 9748. L. Maveyrad, I. Massova, C. Birck, K. Miyashita, J. P. Samama, S.Mobashery, J. Am.Chem. Soc. 118 (1996) 7435. L. Mourey, K. Miyashita, P. Swaren, A. Bulychev,J. P. Samama, S. Mobashery, J. Am. Chem. Soc. 120 (1998) 9383. W. L. Lee, Z. H. Li, S. Vakulenko, S. Mobashery, J. Med. Chem. 43 (2000) 128.



207

PSEUDOMONAS AERUGINOSA QUORUM SENSING: A TARGET FOR ANTIPATHOGENIC DRUG DISCOVERY Everett P. Greenberg Department of Microbiology, University of Iowa, Iowa City, Iowa 52242 USA

It was once held that most bacteria function only as individuals designed to compete with one another and to multiply rapidly under appropriate conditions. This concept has given way to the view that like animals, bacteria can communicate with each other and form communities that represent more than the sum of the individuals [1-3]. Bacteria use chemicals to signal each other, and to coordinate their activities. Many Gram-positive bacteria use small peptides in signaling one another [4, 5], Gram-negative bacteria appear to use small molecule signals of various sorts [1, 6-9]. Perhaps the best-studied signaling system is the Gram-negative acyl-homoserine lactone (acyl-HSL) system. This type of bacterial cell-to-cell communication was first discovered in the context of microbial ecology but it is now evident that acyl-HSL signaling is important in plant and animal (including human) diseases. Acyl-HSL signaling is a dedicated communication system that is used by bacteria to control specific genes in response to population density. Acyl-HSLs are small molecule signals with no other known function. These chemical signals are produced by specific enzymes, and they are detected by specific receptors. Because acyl-HSL signaling provides a mechanism by which a bacterial species can monitor its' own population density, this type of signaling and other signaling systems that achieve the same purpose have been termed quorum sensing systems [10]. Acyl-HSL signals are generated by the activity of a single enzyme. The enzyme uses as substrates, 5-adenosylmethionine, and an intermediate of fatty acid biosynthesis, acyl-acyl carrier protein [11-15]. The enzyme is generally a member of the Luxl family of acyl-HSL synthases. Different Luxl homologs generate different acyl-HSLs. Thus the Pseudomonas aeruginosa Rhll primarily catalyzes the synthesis of A^-butyryl-HSL (C4-HSL) and the P. aeruginosa LasI directs the synthesis of N-(3-oxododecanoyl)-HSL (30C12-HSL). The acyl side-chain length and the substitutions on the side chain provide signal specificity. Acyl side chains of these signals can be fully saturated, or they can have hydroxyls or carbonyls on the third carbon. Acyl-HSLs with side chain lengths of 4 to 16 carbons have been identified [7], Short-chain signals like C4HSL diffuse freely through the cell membrane [16, 17], and 30C12-HSL partitions into cells, presumably in the membrane. This signal can diffuse into the surrounding environment but export is enhanced by the mexAB-oprM, and perhaps other efflux pumps [17, 18]. Regardless, the cellular concentration of an acyl-HSL is defined by the environmental concentration, and environmental

208 concentrations can rise only when there is a sufficient population of the signal producing bacterium. The specific receptors for acyl-HSL signals are members of the LuxR family of transcriptional regulators. LuxR family members have been proposed to consist of 2 domains, a C-terminal DNA-binding domain, and an N-terminal acyl-HSL-binding domain [19]. Quite often the two regulatory genes (the R and I genes) are linked but not always. The orientation of the two genes with respect to each other is variable. Acyl-HSL quorum sensing was first discovered to control the luminescence of Vibrio fischeri, a bacterium that forms a mutualistic Ught organ symbiosis with certain marine animals [20, 21]. Here quorum sensing is critical to the symbiosis. Acyl-HSL signaling is critical for virulence of the plant pathogen Erwinia carotovora [22], and for virulence of P. aeruginosa in mouse models of lung [23] and bum infections [24], in invertebrates [25-27], and in plants [28]. In this report P. aeruginosa serves as a model for the role of bacterial communication in community behaviors important in pathogenesis. That the P. aeruginosa quorum sensing system controls virulence makes it a target for ani-pathogenic drug development. Quorum Sensing in Pseudomonas aeruginosa P. aeruginosa can be isolated from soil and water. It is also an opportunistic pathogen of humans, other animals, and plants. One of the reasons P. aeruginosa is a successful opportunistic pathogen is that it produces a battery of secreted virulence factors. These virulence factors include exoproteases, siderophores, exotoxins, and lipases. Many of these virulence factors are regulated by quorum sensing [1, 29, 30]. Of what advantage to P. aeruginosa is quorum sensing control of virulence factors? First it is economical to produce extracellular factors only after a critical population has been achieved. A mass of cells is required to produce sufficient quantities of these factors to influence the surrounding environment. Furthermore, in the host, timing of the deployment of virulence factors may be critical. The pathogen can amass without displaying its virulence factors, and then the pathogen can mount a surprise attack in which the arsenal of virulence factors is deployed in a coordinated and overwhelming fashion. Genetic studies have revealed two quorum-sensing systems in P. aeruginosa. Both of these systems have linked R and I genes. They are the LasR-I and RhlR-I quorum sensing systems [31-36]. In addition, the recently completed P. aeruginosa genome sequencing project has revealed a third LuxR homolog that is adjacent to a cluster of quorum sensing controlled (qsc) genes [37]. However, a third Luxl homolog is not evident from the sequence, and the function of the third LuxR homolog is as yet unknown. LasR is a transciptional regulator that responds primarily to the Lasl-generated signal, 3-OC12-HSL, and RhlR is a transcriptional regulator that responds best to the Rasl-generated, C4-HSL. The current model for the quorum sensing in P. aeruginosa is as follows: at low population densities LasI produces a basal level of 3-OC12HSL. As density increases, 3-OC12-HSL builds to a critical concentration at which it interacts with LasR. This LasR-3-OC12-HSL complex then activates transcription of a number of genes including rhlR [29, 32, 36, 38, 39]. The

209 activation of rhlR by LasR results in a quorum sensing regulatory cascade, in which activation of the rhl system requires an active las system. RhlR responds best to the Rhll-generated C4-HSL. RhlR then activates expression of genes required for production of a variety of secondary metabolites such as hydrogen cyanide and pyocyanin [29]. A DNA sequence with dyad symmetry called a /wx-box-like sequence can easily be identified in the promoter regions of many quorum-sensing controlled (qsc) genes [10, 37, 40, 41]. By analogy to other acyl-HSL quorum sensing systems we deduce that the lux-box like sequences function as binding sites for LasR and RhlR. It is not yet clear how RhlR and LasR discriminate between their respective binding sites. In fact many genes show partial activation with either LasR or RhlR and the appropriate acyl-HSL [for example see 30, 37]. One explanation for this is that binding site discrimination is less than perfect and either LasR or RhlR can bind with varying efficiency to any lux box-like element. However, lux box-like sequences are not apparent in the promoter regions of all qsc genes. This suggests that LasR or RhlR may also bind to yet to be identified sequences, or that some qsc genes are controlled by LasR or RhlR indirectly. A recent study used a random mutagenesis approach to identify 39 gene that were highly regulated (minimum 5-fold induction, maximum 740-fold induction) by quorum sensing [37]. The genes were divided into 4 different classes, two of which respond to 3-OC12-HSL, and two of which required both C4-HSL and 3-OC12HSL for maximum induction. The qsc genes map throughout the P, aeruginosa chromosome, confirming the view that quorum sensing in this bacterium represents a global regulatory system [29]. The 39 genes revealed by the random mutagenesis study represent only a subset of the qsc genes in P. aeruginosa. It was estimated that as many as 4% of the roughly 6,000 P, aeruginosa genes are controlled by quorum sensing [37]. One report indicates that transcription of rpoS, a gene encoding an RNA polymerase subunit involved in expression of stationary phase factors is activated by RhlR and C4-HSL [42]. This raises the possibiUty that many genes may be controlled indirectly rather than directly by quorum sensing. It is also an enticing hypothesis because it lends itself to the idea that one specific cue that enables a cell to anticipate stationary phase is crowding. Unfortunately, quorumsensing control of rpoS transcription is an example for which there is limited evidence. It is also an example for which there are low levels of induction at best (3-fold). In fact recent investigations suggest that quorum sensing may have no significant influence on rpoS transcription in P. aeruginosa [43]. Biofilms and quorum sensing Bacteria often tend to attach to surfaces and form communities enmeshed in a self-produced polymeric matrix. These communities are called a biofilm [2, 44]. P. aeruginosa is often found in naturally occurring biofilms. Under the appropriate laboratory conditions, P. aeruginosa forms characteristic biofilms that can be several hundred micrometers thick. Development of a mature biofilm proceeds through a programmed series of events [2]. After attachment, cells multiply to form a layer on a solid surface. Individuals in the layer then exhibit a surface motiUty called twitching. Twitching is dependent on

210

Type IV pili. As a result of twitching motility small groups of P. aeruginosa called microcolonies form. Microcolonies then differentiate to form a mature biofilm. Microcolonies in a mature biofilm have tower and mushroom-shaped architectures. The cells in these structures are encased in an extracellular polysaccharide matrix. Water channels that allow the flow of nutrients into and waste products out of the biofilm innervate these structures. There is a significant physiological heterogeneity within biofilms. This heterogeneity in physiological activity makes studying biofilms with traditional molecular microbiological techniques difficult. Bacteria in these mature biofilms are phenotypically resistant to microbiocidal agents including antibiotics. Thus biofilms cause many different types of chronic or persistent bacterial infections [2]. Recent studies have linked quorum sensing and biofilm maturation [45]. This is a particularly gratifying finding because quorum sensing functions to control gene expression in groups of bacteria, and biofilms are just that, organized groups of bacteria. A mutation in lasl has a dramatic affect on biofilm maturation. Lasl mutants are incapable of 30-C12-HSL synthesis, and the development of Lasl mutant biofilms is arrested after microcolony formation but prior to maturation of the microcolonies into thick structured assemblages. Thus Lasl mutant biofilms appear flat and undifferentiated. The normal biofilms architecture can be restored to the mutant by addition of the Lasl-generated quorum sensing signal 30-C12-HSL. A Rhll mutant exhibits normal biofilm development and architecture. The 30-C12-HSL-responsive qsc genes involved in biofilm maturation remain unknown. ACKNOWLEDGEMENTS Work on quorum sensing in the author's laboratory has been funded by the NIH, NSF, Office of Navel Research, and the US Cystic Fibrosis Foundation. Similar accounts of quorum sensing have been published elsewhere.

REFERENCES [1] Fuqua, C , and Greenberg, E. P. (1998) Curr, Opin. Microbiol. 1,183-189. [2] Costerton, J. W., Stewart, P., and Greenberg, E. (1999) Science 284, 13181322. [3] Salmond, G. P. C , Bycroft, B. W., Stewart, G. S. A. B., and Williams, P. (1995) MoL Microbiol 16, 615-624. [4] Dunny, G. M., and Leonard, B. A. {l991)Ann, Rev. Microbiol. 51, 527-564. [5] Novick, R. P., and Muir, T. W. (1999) Curr. Opin. Microbiol. 2,40-45. [6] Fuqua, W. C , and Winans, S. C. (1994b) / Bacteriol. 176,2796-2806. [7] Fuqua, W. C , Winans, S. C , and Greenberg, E. P. (1996) Annu. Rev. Microbiol. 50,121'15l. [8] Pesci, E. C , Milbank, J. B., Pearson, J. P., Kende, A. S., Greenberg, E. P., and Iglewski, B. H. (1999) Proc. Natl. Acad. Sci. USA 96,11229-11234. [9] Bassler, B. L. (1999) Curr. Opin. Microbiol. 2,582-587.

211

[10] Fuqua, W. C , Winans, S. C , and Greenberg, E. P. (1994a) J. Bacteriol. 176, 269-275. [11] Schaefer, A. L., Val, D. L., Hanzelka, B. L., Cronan, J. E., Jr., and Greenberg, E. P. (1996) Proc. Natl Acad. ScL USA 93, 9505-9509. [12] Parsek, M. R., Val, D. L., Hanzelka, B. L., Cronan, J. E., Jr., and Greenberg, E. P. (1999) Proc. Natl. Acad. Sci. USA96,4360-4365. [13] More, M. I., Finger, D., Stryker, J. L., Fuqua, C , Eberhard, A., and Winans, S. C. (1996) Science 272, 1655-1658. [14] Hanzelka, B., Parsek, M. R., Val, D. L., Dunlap, P. V., J. E. Cronan, J., and Greenberg, E. P. (1999) /. Bacteriol. 181, 5766-5770. [15] Hoang, T. T., Y. Ma, R.J. Stem, M.R. McNeil, and Schweizer, H. P. (1999) Gene 237, 361-371. [16] Kaplan, H. B., and Greenberg, E. R (1985) J. Bacteriol. 163,1210-1214. [17] Pearson, J. P., Van Delden, C , and Iglewski, B. H. (1999) J. Bacteriol. 181, 1203-1210. [18] Evans, K., Passador, L., Srikumar, R., Tsang, E., Nezezou, J., and Poole, K. (1998) J. Bacteriol. 180, 5443-5447. [19] Stevens, A. M., and Greenberg, E. P. (1998) in Cell-Cell Signaling in Bacteria (Dunny, G., and Winans, S. C , Eds.) pp 231-242, ASM Press, Washington, DC. [20.] Nealson, K. H., and Hastings, J. W. (1979) Microbiol. Rev. 43,469-518. [21.]Ruby, E. G. (1996) Anna. Rev. Microbiol. 50, 591-624. [22] Pirhonnen, M., Flego, D., Heikiheimo, R., and Palva, E. T. (1993) EMBO J. 12, 2467-2476. [23] Tang, H. B., DiMango, E., Bryan, R., Gambello, M., Iglewski, B. H., Goldberg, J. B., and Prince, A. (1996) Infect. Immun. 64, 37-43. [24] Rumbaugh, K. P., Griswold, J. A., and Hamood, A. N. (1999) J. Burn Care Rehabil. 20,42-49. [25] Tan, M. W., Rahme, L. G., Sternberg, J. A., Tompkins, R. G., and Ausubel, F. M. (1999) Proc Natl Acad Sci USA96, 2408-2413. [26] Tan, M. W., Mahajan-Miklos, S., and Ausubel, F. M. (1999) Proc. Natl. Acad. Sci. f / 5 ^ 96, 715-720. [27] Mahajan-Miklos, S., Tan, M. W., Rahme, L. G., and Ausubel, F. M. (1999) Ce//96,47-56. [28] Rahme, L. G., Stevens, E. J., Wolfort, S. F., Shao, J., Tompkins, R. G., and Ausubel, F. M. (1995) Science 268, 1899-1902. [29] Pesci, E. C , and Iglewski, B. H. (1997) Trends in Microbiol. 5,132-135. [30] Van Delden, C , and Iglewski, B. H. (1998) Emerg Infect. Dis. 4, 551-560. [31] Brint, J. M., and Ohman, D. E. (1995) J. Bacteriol. 177, 7155-7163. [32] Gambello, M. J., Kaye, S., and Iglewski, B. H. (1993) Infect. Immun. 61, 1180-1184. [33] Latifi, A., Winson, K. M., Foglino, M., Bycroft, B. W., Stewart, G. S. A. B., Lazdunski, A., and WiUiams, P. (1995) Mol. Microbiol. 17, 333-344. [34] Ochsner, U. A., Koch, A. K., Fiechter, A., and Reiser, J. (1994) J. Bacteriol. 176, 2044-2054. [35] Ochsner, U. A., and Reiser, J. (1995) Proc. Natl. Acad. Sci. USA92, 64246428.

212 [36] Passador, L., Cook, J. M., Gambello, M. J., Rust, L., and Iglewski, B. H. (1993) Science 260,1127-1130. [37] Whiteley, M., Lee, K. M., and Greenberg, E. P. (1999) Proc. Natl. Acad. Sci. USA 96,13904-13909. [38] Gambello, M. J., and Iglewski, B. H. (1991) J. Bacteriol. 173, 3000-3009. [39] Seed, P. C , Passador, L., and Iglewski, B. H. (1995) J. Bacteriol. Ill, 654659. [40] Pearson, J. P., Pesci, E. C , and Iglewski, B. H. (1997) J. Bacteriol. 179, 5756-5767. [41] Pesci, E. C , Pearson, J. P., Seed, P. C , and Iglewski, B. H. (1997) J. Bacteriol. 179, 3127-3132. [42] Latifi, A., Foglino, M., Tanaka, K., Williams, P., and Lazdunski, A. (1996) Mol. Microbiol. 21,1137-1146. [43] Whiteley, M., Parsek, M., and Greenberg, E. P. Submitted for Publication. [44] Reimmann, C , Beyeler, M., Latifi, A., Winteler, H., Foglino, M., Lazdunski, A., and D, D. H. (1997) Mol. Microbiol. 24, 309-319. [45] de Kievit, T., Seed, P. C , Nezezon, J., Passador, L., and Iglewski, B. H. (1999) J. Bacteriol. 181, 2175-2184. [46] Albus, A. M., Pesci, E. C , Runyen-Janecky, L. J., West, S. E., and Iglewski, B. H. (1997) J. Bacteriol. 179, 3928-3935. [47] Costerton, J. W., Lewandowski, Z., Caldwell, D. E., Korber, D. R., and Lappin-Scott, H. M. (1995) Ann. Rev. Microbiol. 49, 711-745. [48] Davies, D. G., Parsek, M. R., Pearson, J. P., Iglewski, B. H., Costerton, J. W., and Greenberg, E. P. (1998) Science 280,295-298. [49] Stickler, D. J., Morris, N. S., McLean, R. J., and Fuqua, C. (1998) Appl. Environ. Microbiol. 64, 3486-90.


213

Discovery and Development of New Anti-Bacterial Drugs Ian Chopra, Lars Hesse and Alexander O'Neill Antimicrobial Research Centre and Division of Microbiology, School of Biochemistry and Molecular Biology, University of Leeds, Leeds LS2 9JT, United Kingdom

INTRODUCTION The ability to treat bacterial infections with chemotherapeutic agents, introduced vyîth the discovery of penicillin and prontosil in the 1930s (Figure 1), represents one of the most important medical achievements of the twentieth century. Indeed, the rapid advances made in the discovery of new antibiotics and other antibacterial agents during the so-called "golden" period between 1940 and the mid-1960s (Figure 1) led to widespread optimism that bacterial infections could be completely conquered. Figure 1. Discovery of antibacterial agents Empiric screening has been based on the identification of antibacterial agents by their ability to inhibit bacterial growth. Synthetic approaches comprise chemical modification of existing drug classes to improve their properties e.g. circumvention of resistance mechanisms to earlier members of the class. Only representative antibacterial agents are indicated. Empiric screening Cycloserine Erythromycin Ethionamide Isoniazid Metronidazole Pyrazinamide Rifamycin —Trimethoprim Vancomycin Virginiamycin Chlortetracycline •

1930

^

±

1940

- ^

•^ ^

^

Imipenem

Semi-synthetic penicillins & cephalosporins Oxazolidinones Defensins

Ethambutol Fusidic acid -^ Mupirocin Nalidixic acid

1960

Glycylcyclines

- ^ ^ Newer cartiapenems

Newer aminoglycosides > - Fluoroquinolones

\ 1950

Rifabutin Rifalazil Rifapentine

Semi-synthetic glycopeptides Semi-synthetic streptogramlns

Minocycline

- ^ Chloramphenicol Neomycin Polymixin Streptomycin — Thiacetazone

Newer macrolides & ketolides

- Rifampicin

^

Cephalosporin Cy

Penicillin Prontosil

Synthetic approaches

1970

1980

\

I

1990

2000

214 This period of optimism is captured by the famous remark made in 1969 by the US Surgeon General who testified to Congress that "the time has come to close the book on infectious disease" [1]. However, even from the very earliest period of the antibiotic era the potential for the emergence of drug resistant bacteria has been recognised (Figure 2), [2]. Figure 2. Emergence of antibacterial resistance (1944 - 2001). The first significant reports of resistance in clinical isolates are indicated.

Unfortunately the selection of organisms resistant to antibacterial agents has continued to the present day (Figure 2) and the next millennium has arrived with the dramatic emergence of resistance to antibacterial agents in all significant bacterial pathogens [1,3-6]. In some cases, organisms multiply resistant to virtually all chemotherapeutically useful antibacterial agents have been identified. The latest developments in this increasingly alarming situation have been the emergence of vancomycin intermediate resistant strains of Staphylococcus aureus (so-called "VISA" strains) [6] and linezolid resistant strains of Enterococcus faecium [7]. As noted by Cookson [8], the emergence of VISA strains may herald the appearance of high-level vancomycin-resistant strains of S. aureus. Such an event would have serious consequences for the control of nosocomial staphylococcal infections. The rapid emergence of linezolid resistance in the enterococci is equally disturbing since linezolid, a member of the new oxazolidinone class of antibiotics, was only approved by the US Food and Drug Administration (FDA) in April 2000 [6]. Although sensible measures to limit antibiotic usage to valid therapeutic indications and to reduce the spread of resistant organisms are of value in limiting the emergence of resistant organisms, the resistance problem continues to require renewed effort by the

215 pharmaceutical industry to create products that will prevent or treat infections caused by antibiotic-resistant pathogens [3,4]. Currently there are four principal drug discovery approaches: • Expansion of knov^n drug classes to include organisms resistant to earlier members of the class • Protection of known classes by resistance mechanism inhibitors • Re-evaluation of earlier pharmacophores • Discovery or design of new agents through rational selection of novel targets underpinned by genomics This chapter considers these approaches for the discovery of new antibacterial agents presenting the advantages and disadvantages of each strategy. Strategies will be illustrated by specific examples. Since the emphasis for future research will be the discovery of novel agents active against new molecular targets, it will be important to pursue approaches that minimise the emergence of bacterial resistance to newly discovered agents. This chapter also addresses this issue. 1.

EXPANSION OF KNOWN DRUG CLASSES TO INCLUDE ORGANISMS RESISTANT TO EARLIER MEMBERS OF THE CLASS

L1 Overview and examples Since the mid-1970s industrial approaches to the development of new antibacterial agents have been dominated by this paradigm (Figure 1). Indeed, in recent years, the oxazolidinones and defensins represent the only new classes of antibacterial agents to be developed. Expansion of known drug classes has been the principal strategy adopted by the pharmaceutical industry to combat the emergence of bacterial resistance to antibacterial agents. Essentially, this strategy depends on the synthesis of repertoires of new analogues related to known antibacterial agents to create structural modifications that circumvent the resistance phenotype. Early examples of this approach include development of semi-synthetic penicillins such as methicillin and the isoxazolyl penicillins which were developed on the basis of their stability to staphylococcal penicillinases [4]. More recent examples include the third generation cephalosporins, exemplified by agents such as cefotaxime (introduced in 1981) and ceftazidime (introduced in 1985), which v/ere developed on the basis of their stability to the TEM-1 and SHV-1 P-lactamases which are broadly dispersed amongst clinical isolates of Gram-negative bacteria [4]. The strategy has continued to the present day [9-13]. For example the glycylcyclines, exemplified by tigilcycline (GAR936) (Figure 3) which is currently in Phase II clinical trials, represent a new class of tetracycline analogues [12]. Glycylcyclines exhibit activity against bacteria expressing resistance to earlier tetracyclines by efflux and ribosomal protection mechanisms [12,13].

216 Figure 3. Structure of 9-t-butylglycyIamido minocycline (GAR-936;tigilcycline) N(CH3)2

N(CH3)2

1.2 Limitations of developing analogues of existing drug classes Expansion of existing drug classes to meet the clinical challenges imposed by resistant organisms has undoubtedly led to the introduction of a number of clinically successful agents (Figure 1). However, this approach can now only be considered at best a temporary solution to the problem of resistance. Unfortunately, the existence of resistance mechanisms to earlier members of the drug class often provides the organisms with a "head-start" for mutational adaptation by which expression of resistance to the newest member of the class also rapidly emerges. Notable examples include the relatively limited number of amino acid changes in the TEM-1 and SHV-1 P-lactamases required to confer resistance to the third generation cephalosporins [13] and the observation, at least under laboratory conditions, that single-site mutations in genes encoding tetracycline efflux pumps confer resistance to members of the new glycyIcycline group of antibiotics [13,14]. 2. PROTECTION OF KNOWN CLASSES BY RESISTANCE MECHANISM INHIBITORS 2.1 Overview and examples The bacterial enzymes that degrade or modify antibiotics, as well as the efflux systems that remove antibiotics from bacteria are themselves potential targets for drug action. The objective of this approach is the introduction of combination products which contain an antibiotic and a specific inhibitor that protects the antibiotic from enzymatic inactivation or removalfromthe cell [4]. The concept of using resistance mechanism inhibitors in conjunction with antibiotics has been most successfully applied in the area of P-lactamase inhibitor-P-lactam combinations [4]. The p-lactamase inhibitor clavulanic acid was the prototype molecule that established the value of this strategy. It exhibits weak antibacterial activity, but binds with high affinity and essentially irreversibly to many bacterial p-lactamases [4]. Clavulanic acid thus protects p-lactam antibiotics from destruction and is available commercially in combination with amoxycillin and ticarcillin (Table 1). Other Plactamase inhibitors, e.g. tazobactam and sulbactam, have also been introduced as therapeutic agents in combination with p-lactam antibiotics (Tablel) [4,15,16].

217 Table 1. Commercially available P-lactam/p-lactamase inhibitor combinations. (i.v. = intravenous; i.m. = intramuscular) Inhibitor

P-lactam antibiotic

Clavulanic acid

Amoxycillin

Augmentin

Oral, i.v.

Clavulanic acid

Ticarcillln

Timentin

I.v.

Tazobactam

Piperacillin

Zosyn

I.v.

Sulbactam

Ampicillin

Unasyn

I.m., I.v.

Antibiotic efflux is now recognised as a major mechanism of bacterial resistance to antibiotics. Some efflux pumps selectively extrude specific antibiotics (e.g. tetracyclines), while others, classified as multidrug resistance (MDR) pumps, mediate efflux of a variety of structurally diverse compounds with differing antibacterial modes of action [17,18]. In keeping with the paradigm of p-lactamase inhibitor-P-lactam combinations the discovery and development of efflux pump inhibitors would lead to products containing an antibiotic and an inhibitor that would prevent efflux of the drug from the bacterial cell. A number of new research programmes have recently been initiated in several pharmaceutical companies to identify and develop appropriate bacterial efflux pump inhibitors. These programmes have already identified a number of promising research leads. Limitations of space prevent complete discussion here. However, readers are referred to a recent comprehensive review [18] for further information. 2.2. Limitations associated with resistance mechanism inhibitors Although the existing P-lactamase inhibitor-p-lactam combinations have gained widespread acceptance as valuable therapeutic agents [15,16], their spectrum of activity encompasses only some of the clinically relevant P-lactamase enzymes (Table 2 ). Thus the three inhibitors developed to date are at best only weak inhibitors of the molecular class B and C p-lactamases. The deficiencies in the spectrum of activity of existing plactamase inhibitors has led to intensive research efforts to discover new inhibitor classes that also encompass the molecular group B and C enzymes. To date, these approaches have only identified reaearch leads. Readers are referred to the recent excellent review by Payne et al [16] for further details on the discovery of new plactamase inhibitors. A further limitiation of existing commercially available plactamase inhibitors, has been the development of inhibitor-resistant p-lactamase variants in clinically relevant organisms [16].

218 Table 2. Clinically important ^-lactamases and activity of serine p-lactamase inhibitors. Based on reference [16]. Molecular class Class A

Class B

Class C

Class D

P-lactamase type TEM-1

Clinical significance of enzymes

SHV-1

Most commonly found plactamases

Extended spectrum TEM/SHVs

Resistance to 3rd and 4th generation cephalosporins

Narrow spectrum carbapenemases

Carbapenem resistance

Broad spectrum carbapenemases (metallo enzymes)

Resistance to majority of plactams

Stably derepressed mutants

Resistance to 3rd generation cephalosporins

Plasm id-mediated Class C

As above, but with capacity for mobility

Extended spectrum OXA enzymes

Resistance to 3rd generation cephalosporins

Inhibition (ICjo, uM) by | Clavulanic Tazobactam Sult>actam 1 acid

Trends in Drug Research III: Proceedings of the 13th Noordwijkerhout-Caminero Symposium, the Netherlands, 6-11 May 2001 (Pharmacochemistry Library)

Trends in drug research II: proceedings of the 11th Noordwijkerhout-Camerino Symposium, 11-15 May 1997, Noordwijkerhout

Perspectives in Receptor Research: Proceedings of the 10th Camerino-Noordwijkerhout Symposium, Camerino, Italy, 10-14 September 1996 (Pharmacochemistry Library)

The Histamine H3 Receptor (Pharmacochemistry Library)

Proceedings of the 37th International Symposium on Archaeometry, 13th - 16th May 2008, Siena, Italy

Stable Isotopes in Pharmaceutical Research, Volume 26 (Pharmacochemistry Library)

May 1940: the battle for the Netherlands

The Economist - 19 May 2001

The Economist - 05 May 2001

The Economist - 12 May 2001

Symposium on Turbulence (13th)

The Economist - 26 May 2001

Trends in the Transport Sector, 1970-2001

Programs as Data Objects: Second Symposium, PADO 2001, Aarhus, Denmark, May 21-23, 2001, Proceedings

Trends in Catalysis Research

Trends in Genome Research

Drug Education Library Opium Drug Education Library

Drug Education Library - LSD (Drug Education Library)

Proteomics In Drug Research

Proteomics in Drug Research

Atmospheric Ozone Research and Its Policy Implications: Proceedings of the 3rd Us-Dutch International Symposium, Nijmegen, the Netherlands, May 9-13 (1989) (Studies in Environmental Science)

Scientific American (May 2001)

Architecture in the Netherlands

Peptides for Youth: The Proceedings of the 20th American Peptide Symposium (Advances in Experimental Medicine and Biology Vol 611)

Molecular Pathomechanisms and New Trends in Drug Research

Molecular Pathomechanisms and New Trends in Drug Research

Architecture in the Netherlands

Sinope, The Results of Fifteen Years of Research. Proceedings of the International Symposium, 7-9 May 2009

Proceedings of the Third Scandinavian Logic Symposium

Proceedings of the Second Scandinavian Logic Symposium

Advances in Plant Ethylene Research: Proceedings of the 7th International Symposium on the Plant Hormone Ethylene

Trends in Drug Research III: Proceedings of the 13th Noordwijkerhout-Caminero Symposium, the Netherlands, 6-11 May 2001 (Pharmacochemistry Library)

Trends in drug research II: proceedings of the 11th Noordwijkerhout-Camerino Symposium, 11-15 May 1997, Noordwijkerhout

Perspectives in Receptor Research: Proceedings of the 10th Camerino-Noordwijkerhout Symposium, Camerino, Italy, 10-14 September 1996 (Pharmacochemistry Library)

The Histamine H3 Receptor (Pharmacochemistry Library)

Proceedings of the 37th International Symposium on Archaeometry, 13th - 16th May 2008, Siena, Italy

Stable Isotopes in Pharmaceutical Research, Volume 26 (Pharmacochemistry Library)

May 1940: the battle for the Netherlands

The Economist - 19 May 2001

The Economist - 05 May 2001

The Economist - 12 May 2001

Symposium on Turbulence (13th)

The Economist - 26 May 2001

Trends in the Transport Sector, 1970-2001

Programs as Data Objects: Second Symposium, PADO 2001, Aarhus, Denmark, May 21-23, 2001, Proceedings

Trends in Catalysis Research

Trends in Genome Research

Drug Education Library Opium Drug Education Library

Drug Education Library - LSD (Drug Education Library)

Proteomics In Drug Research

Proteomics in Drug Research

Atmospheric Ozone Research and Its Policy Implications: Proceedings of the 3rd Us-Dutch International Symposium, Nijmegen, the Netherlands, May 9-13 (1989) (Studies in Environmental Science)

Scientific American (May 2001)

Architecture in the Netherlands

Peptides for Youth: The Proceedings of the 20th American Peptide Symposium (Advances in Experimental Medicine and Biology Vol 611)

Molecular Pathomechanisms and New Trends in Drug Research

Molecular Pathomechanisms and New Trends in Drug Research

Architecture in the Netherlands

Sinope, The Results of Fifteen Years of Research. Proceedings of the International Symposium, 7-9 May 2009

Proceedings of the Third Scandinavian Logic Symposium

Proceedings of the Second Scandinavian Logic Symposium

Advances in Plant Ethylene Research: Proceedings of the 7th International Symposium on the Plant Hormone Ethylene

Recommend Documents