QSAR and Drug Design: New Developments and Applications

PHARMACOCHEMISTRY LIBRARY- VOLUME 23 QSAR AND DRUG DESIGN" NEW DEVELOPMENTS AND APPLICATIONS

PHARMACOCHEMISTRY LIBRARY, edited by H. Timmerman Other titles in this series Volume 9

Innovative Approaches in Drug Research. Proceedings of the Third Noordwijkerhout Symposium on Medicinal Chemistry, Noordwijkerhout (The Netherlands), September 3-6, 1985 edited by A.F. Harms

Volume 10

QSAR in Drug Design and Toxicology, Proceedings of the Sixth European Symposium on Quantitative Structure-Activity Relationships, Portoro2-Portorose (Yugoslavia), September 22-26, 1986 edited by D. Had2i and B. Jerman-Bla2i~

Volume 11

Recent Advances in Receptor Chemistry. Proceedings of the Sixth CamerinoNoordwijkerhout Symposium, Camerino (Italy), September 6-10, 1987 edited by C. Melchiorre and M. Giannella

Volume 12

Trends in Medicinal Chemistry '88. Proceedings of the Xth International Symposium on Medicinal Chemistry, Budapest, 15-19 August, 1988 edited by H. van der Groot, G. Domany, L. Pallos and H. Timmerman

Volume 13

Trends in Drug Research. Proceedings of the Seventh Noordwijkerhout-Camerino Symposium, Noordwijkerhout (The Netherlands), 5-8 September, 1989 edited by V. Claassen

Volume 14

Design of Anti-Aids Drugs edited by E. De Clerq

Volume 15

Medicinal Chemistry of Steroids

by F.J. Zeelen

Volume 16

QSAR: Rational Approaches to the Design of Bioactive Compounds. Proceedings of the Eighth European Symposium on Quantitative Structure-Activity Relationships, Sorrento (Italy), 9-13 September, 1990 edited by C. Silipo and A. Vittoria

Volume 17

Antilipidemic Drugs - Medicinal, Chemical and Biochemical Aspects edited by D.T. Witiak, H.A.I. Newman and D.R. Feller

Volume 18

Trends in Receptor Research. Proceedings of the Eighth Camerino-Noordwijkerhout Symposium, Camerino (Italy), September 8-12, 1991 edited by P. Angeli, U. Giulini and W. Quaglia

Volume 19

Small Peptides. Chemistry, Biology and Clinical Studies edited by A.S. Dutta

Volume 20

Trends in Drug Research. Proceedings of the 9th Noordwijkerhout-Camerino Symposium, Noordwijkerhout (The Netherlands), 23-27 May, 1993 edited by V. Claassen

Volume 21

Medicinal Chemistry of the Renin-Angiotensin System edited by RB.M.W.M. Timmermans and R.R. Wexler

Volume 22

The Chemistry and Pharmacology of Taxol| and its Derivatives edited by V. Farina

PHARMACOCHEMISTRY

LIBRARY

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

Volume

23

QSAR AND DRUG DESIGN: N EW DEVE LO PM E NTS AN D APPLI CATI O N S

Based on Topics presented at the Annual Japanese (Quantitative) StructureActivity Relationship Symposium and the Biennial China-Japan Drug Design and Development Conference

EDITED BY:

TOSHIO FUJITA Department of Agricultural Chemistry, Kyoto University, Kyoto, and EMIL PROJECT, Fujitsu Kansai Systems Laboratory, Osaka, Japan

ELSEVIER Amsterdam

- Lausanne - New York-

Oxford - Shannon

- T o k y o 1995

ELSEVIER SCIENCE B.V. P.O. Box 1527 1000 B M A m s t e r d a m , The N e t h e r l a n d s

IS B N 0-444-88615-X

9 1995 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. Special regulations for readers in the U.S.A.-This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the U.S.A. All other copyright questions, including photocopying outside of the U.S.A., should be referred to the publisher. 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. This book is printed on acid-free paper. Printed in The Netherlands

Dedicated to

Professor Corwin Hansch Without his heartfelt encouragements, the editing of this volume would never have been completed.

This Page Intentionally Left Blank

PHARMACOCHEMISTRY LIBRARY ADVISORY BOARD T. Fujita E. Mutschler N.J. de Souza D.T. Witiak F.J. Zeelen

Department of Agricultural Chemistry, Kyoto University, Kyoto, Japan Department of Pharmacology, University of Frankfurt, F.R.G. Research Centre, Hoechst India Ltd., Bombay, India College of Pharmacy, The Ohio State University, Columbus, OH, U.S.A. Organon Research Centre, Oss, The Netherlands

This Page Intentionally Left Blank

PREFACE In this series of Pharmacochemistry Library the preceding volume dealing with the QSAR methodology and related topics is Vol. 16, QSAR: RationalApproaches to the Design of Bioactive Compounds, edited by Carlo Silipo and Antonio Vittoria, both of whom unfortunately passed away recently. Volume 16 was published as the Proceedings of the 8th European Symposium on Quantitative StructureActivity Relationships held in 1990 in Sorrento, Italy. Like the European Symposium, the Japanese Symposium on Structure-Activity Relationships has been organised annually since 1975. A bilateral symposium with Chinese scientists, the "China-Japan Drug Design and Development Conference", has been held biennially since 1989. This volume, instead of taking the form of Proceedings, is an edited volume based on topics selected from those presented at these symposia. Each chapter is thus more complete than the original presentations and includes consecutive series of the same topic originally presented separately. The structure-activity relationship (SAR) studies of bioactive compounds seem to have at least two objectives. One is to obtain insight into the pharmacological modes of action and the other is to deduce possible guiding principles for designing analogues with better bioactive profiles. The quantitative approach to the SAR (QSAR), initiated by Corwin Hansch and his co-workers some 35 years ago, opened up new possibilities in the SAR discipline. Because the Hansch QSAR expanded the Hammett-Taft paradigm in physical organic chemistry toward the biomedicinal (re)activity, the mode of action has been illustrated on the (sub)molecular level in many cases. It also revealed the critical importance of the hydrophobicity of the bioactive molecule. Before the advent of the QSAR, the mode of action had remained mostly on the level of discussions in terms of the "lock-and-key" hypothesis. Because the relationships are represented in the form of mathematical correlation equations with physicochemical (electronic, steric, hydrophobic and others when necessary) parameter terms in the QSAR, the bioactivity of non-measured analogues has sometimes been predicted by extrapolating significant parameters and proved after synthesis and biological tests. This can be regarded as the beginning of the quantitative drug design. Perhaps stimulated by the success of the traditional Hansch QSAR, a number of newer software-based methodologies have been publicized in the SAR and drug design disciplines, supported by the tremendous progress in computer technology in recent years. Among them are those based on theoretical physicochemical and/or molecular orbital calculations, those utilizing molecular modelling and graphics, those managing sophisticated statistical operations and data-base-oriented procedures. Some theoretical calculation softwares do not only deal with the stereo-electronic energy of ligands, but also extend their scope into protein molecules. Thus, the current situation is as if a successful drug design from receptor protein structures could be not entirely impossible.

In this volume topics are covered among almost every procedure and subdiscipline described above. They are categorized into three sections. Section I includes topics illustrating newer methodologies relating to ligand-receptor interactions, molecular graphics and receptor modelling as well as the threedimensional (Q)SAR examples with the active analogue approach and the comparative molecular field analysis. Note that the last two chapters also use the traditional QSAR to cross-validate the results obtained with the newer procedures. In Section II the hydrophobicity parameters, log P (1-octanol/water), for compound series of medicinal-chemical interest are analysed physico-organic chemically. New procedures for the lead generation using databases of aminoacid sequences and structural evolution patterns, as well as a newer statistical QSAR modification utilizable in cases when the bioactivity potency is represented by ratings, are also placed in this Section. Section III contains the examples based on the traditional Hansch QSAR approach. Two contributions are from China illustrating how to identify the lead structures from folk medicine and how to optimize them in clinical applications. Others in this Section are instructive examples of the Hansch approach for various series of bioactive compounds in rationalizing the potency variations, actual designing the clinical candidates and revealing the (sub)molecular mechanism of action. A variety of methodologies and procedures are presented in this single volume. It is recommended that the readers regard each of the methodologies as complementary to others. It must be confessed that editing this volume required a much longer period than I had originally expected. Apologies are due to some of the authors if their chapters have become out of date, because the speed of progress in this field is very fast. If there could be something to mitigate the responsibility, it is the fact that most of the chapters dealing with rapidly growing topics describe their methodological philosophy in some detail. With understanding the background way of thinking, further developments can hopefully be caught up without difficulty. Last but not least, the editor expresses his sincere thanks to Mrs. A. Elzabeth Ichihara for critical correction of the English in most of the original manuscripts. August 1, 1995 Toshio Fujita, at Fujitsu Kansai Systems Laboratory

XI

LIST OF CONTRIBUTORS Dr. G. Appendino Dipartimento di Scienza e Tecnologia del Farmaco via R Giuria 9 10125 Torino ITALY Dr. S.H. Chen Bristol Myers Squibb Pharmaceutical Research Institute RO. Box 5100 Wallingford, CT 06492-7660 U.S.A.

Dr. L. Landino Chemistry Department University of Virginia Charlottesville, VA 22901 U.S.A. Dr. T. MacDonald Chemistry Department University of Virginia Charlottesville, VA 22901 U.S.A.

Dr. T. Cresteil INSERM U75 Universite Rene Descartes 75730 Paris Cedex 15 FRANCE

Dr. B. Monsarrat Laboratoire de Pharmacologie et Toxicologie Fondamentales CNRS 205 Route de Narbonne 31400 Toulouse FRANCE

Dr. R.C. Donehower Division of Pharmacology and Experimental Therapeutics Johns Hopkins Oncology Center Baltimore, MD 21287 U.S.A.

Dr. E.K. Rowinsky Div. of Pharmacology and Experimental Therapeutics Johns Hopkins Oncology Center Baltimore, MD 21287 U.S.A.

Dr. V. Farina Department of Medicinal Chemistry Boehringer Ingelheim Pharmaceuticals 900 Ridgebury Road Ridgefield, CT 06877 U.S.A.

Dr. I. Royer Laboratoire de Pharmacologie et Toxicologie Fondamentales CNRS 205 Route de Narbonne 31400 Toulouse FRANCE

Dr. D. Guenard Institut de Chimie des Substances Naturelles CNRS 91190 Gif-sur-Yvette FRANCE Dr. J. Kant Bristol Myers Squibb Pharmaceutical Research Institute P.O. Box 5100 Wallingford, CT 06492-7660 U.S.A.

Dr. D.M. Was Bristol Myers Squibb Pharmaceutical Research Institute 5, Research Parkway Wallingford, CT 06492-7660 U.S.A. Dr. M. Wright Laboratoire de Pharmacologie et Toxicologie Fondamentales CNRS 205 Route de Narbonne 31400 Toulouse FRANCE

This Page Intentionally Left Blank

xIII

CONTENTS T. Fujita: Preface

SECTION I:

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

ix

Three-Dimensional Structure-Based Drug Design, Molecular Modelling and Three-Dimensional QSAR.

A. Itai, N. Tomioka, Y. Kato Rational Approaches to Computer Drug Design Based on Drug-Receptor Interactions . . . . . . . . . . . . . . . . . . . . . . . . K. Akahane, H. Umeyama

Drug Design Based on Receptor Modeling Using a System

"BIOCES(E)"

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

49

T. Matsuzaki, H. Umeyama, R. Kikumoto

Mechanisms of the Selective Inhibition of Thrombin, Factor Xa, Plasmin and Trypsin . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83

H. Koga, M. Ohta Three-Dimensional Structure-Activity Relationships and Receptor Mapping of Quinolone Antibacterials . . . . . . . . . . . . . . . . . . .

M. Yamakawa, K. Ezumi, K. Takeda, T. Suzuki, I. Horibe, G. Kato, T. Fujita Classical and Three-Dimensional Quantitative Structure-Activity Analyses of Steroid Hormones: Structure-Receptor Binding Patterns of Anti-hormonal Drug Candidates . . . . . . . . . . . . . . . . . . . .

97

125

SECTION I1: Quantitative Structure-Parameter Analyses and Database-Oriented and Newer Statistical (Q)SAR Procedures and Drug Design, C. Yamagami, N. Takao, T. Fujita

Analysis and Prediction of 1-Octanol/VVater Partition Coefficients of Substituted Diazines with Substituent and Structural Parameters . . . 153

M. Akamatsu, T. Fujita Hydrophobicities of Di-to Pentapeptides Having Unionizable Side Chains and Correlation with Substituent and Structural Parameters . . 185 T. Nishioka, J. Oda

Analysis of Amino Acid Sequence-Function Relationships in Proteins . 215

xIv

T. Fujita, M. Adachi, M. Akamatsu, M. Asao, H. Fukami, Y. Inoue, I. Iwataki, M. Kido, H. Koga, T. Kobayashi, I. Kumita, K. Makino, K. Oda, A. Ogino, M. Ohta, F. Sakamoto, T. Sekiya, R. Shimizu, C. Takayama, Y. Tada, I. Ueda, Y. Umeda, M. Yamakawa, Y. Yamaura, H. Yoshioka, M. Yoshida, M. Yoshimoto, K. Wakabayashi

Background and Features of EMIL, A System for Database-Aided Bioanalogous Structural Transformation of Bioactive Compounds . . . 235

10

I. Moriguchi, S. Hirono

Fuzzy Adaptive Least Squares and its Use in Quantitative StructureActivity Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . .

275

SECTION II1: Traditional QSAR and Drug Design. 11

12

13

14

15

16

Z-r. Guo

Structure-Activity Relationships in Medicinal Chemistry: Development of Drug Candidates from Lead Compounds . . . . . . . . . . . . . . .

299

R.-I. Li, S.-y. Wang

Chemical Modification and Structure-Activity Relationship Studies of Piperine and its Analogs: An Example of Drug Development from Folk Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

H. Terada, S. Goto, H. Hori, Z. Taira

Structural Requirements of Leukotriene Antagonists

..........

321

341

K. Mitani

Quantitative Structure-Activity Relationships of a New Class of Ca2+-Antagonistic and 0~-Blocking Phenoxyalkylamine Derivatives . . . 369

H. Ohtaka

Applications of Quantitative Structure-Activity Relationships to Drug Design of Piperazine Derivatives . . . . . . . . . . . . . . . . . . . . .

413

K. Hashimoto, H. Tanii, A. Harada, T. Fujita

Quantitative Structure-Activity Studies of Neurotoxic Acrylamide Analogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Subject index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

451

481

SECTION I: Three-Dimensional Structure-Based Drug Design, Molecular Modelling and Three-Dimensional QSAR.

This Page Intentionally Left Blank

QSAR and Drug Design - New Developments and Applications T. Fujita, editor 9 1995 Elsevier Science B.V. All rights reserved

RATIONAL A P P R O A C H E S TO C O M P U T E R D R U G D E S I G N B A S E D ON D R U G - R E C E P T O R I N T E R A C T I O N S

Akiko Itai*, Nobuo Tomioka* and Yuichi Kato Faculty of Pharmaceutical Sciences, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, Japan ABSTRACT

We have developed two novel methods and computer programs for rational drug design on the basis of drug-receptor interaction. The program GREEN is to perform docking studies efficiently and rationally, when the receptor structure is known. The main features of the program are the real-time estimation of intermolecular interaction energy and the informative visualization of the drug binding site. In addition, many functions help to find a p p r o x i m a t e l y the stable positions and conformations of a drug molecule inside the receptor cavity. The other program, RECEPS, is for rational superposition of molecules and for receptor mapping, when the receptor structure is not known. The superposition is performed through the use of spatial grid points and monitored by several goodness-of-fit indices indicating the similarities in physical and chemical properties. Based on the superposed structures, a three-dimensional receptor image can be constructed, which reveals cavity shapes, expected locations and characters of hydrogen-bonding groups, electrostatic potentials of the surface, and other features. 1. I N T R O D U C T I O N

For the development of new drugs, a tremendous number of compounds must be synthesized and assayed for biological activities. As the difficulties in synthesizing compounds have decreased with the technical advances of organic synthesis, the efficient design of bio-active molecules has become more and more important. Usually, drug development starts with the selection of a lead compound, and then the structure is modified to obtain better biological response profiles. But, starting from an appropriate lead compound is the key to success. How to find an appropriate lead compound and how to optimize the lead structure efficiently are the central problems of drug development. As yet, however, no general *Present address: Institute of Medicinal Molecular Design, 4-1-11 Hongo, Bunkyo-ku, Tokyo, Japan

methods for solving these problems are available. Indeed, finding new lead compounds is so difficult as compared with optimizing existing lead compounds that they have never been generated artificially. It has long been desired to design active structures on the basis of logic and calculations, not relying on chance or trial-and-error. Computers have been introduced into drug design for that purpose, and with the remarkable progress of computer technology in the past thirty years, computers have become widely used in drug research for maintaining databases, statistical processing, molecular modeling, theoretical chemical calculation, and so on. Since analyses of the relationships between structures and activities by using computers began more than twenty years ago (1), various approaches have been reported by many researchers. Some of them, however, have fallen by the wayside as our understanding of drug-receptor interactions has deepened.

Drug-Receptor Interactions It is well known now that a drug molecule exerts its biological activities by binding specifically to a target macromolecule, or receptor, in the body. Dozens of receptor molecules for various hormones and neural transmitters have been isolated and characterized, and their amino acid sequences have been determined. None of the three-dimensional structures of such receptors has been elucidated, whereas those of hundreds of proteins have already been elucidated to atomic resolution by X-ray crystallographic analyses. Some solutions have been obtained for complexes of protein and ligand molecules. These results have provided us with details of molecular recognition by the macromolecule as well as the three-dimensional structure of the macromolecule. Such concrete molecular images have validated the key-and-lock model for drugreceptor interaction, which had been vaguely understood for a long time. In most of the complexes, ligand molecules are non-covalently bound to proteins. The complexes are stabilized by intermolecular forces such as hydrogen bonds, electrostatic interactions, van der Waals forces, and hydrophobic interactions. The strength of binding, which is represented experimentally by equilibrium constants of binding or dissociation, can be estimated by empirical energy calculations. The sum of the intramolecular and intermolecular energy values is taken as an index for showing

the binding affinity, although the molecular recognition results from the free energy decrease upon complexation between the molecules. Accordingly, the more energetically favorable the interaction of the ligand molecule with the receptor is, the more efficiently the ligand can bind to the target receptor specifically. There are many examples where agonist and antagonist molecules with quite different chemical structures can bind strongly to the same site of the same receptor as the natural bio-active compounds. This fact is well evidenced by a number of crystallographic studies on protein-ligand or enzyme-inhibitor complexes. It can be seen that it is not the skeletal structure itself but the threedimensional array of submolecular physical and chemical properties of the ligand molecule that is recognized by proteins. As receptors consist mainly of proteins and the main functions of receptors seem to depend on the protein constituents, the molecular recognition between a receptor and drug is supposed to be very similar to that between an enzyme and substrate. The only difference is that reactions proceed in the case of enzymes, whereas signals are transduced between cells in the case of receptors. Many enzyme inhibitors are used as clinical drugs, in order to maintain biological homeostasis by controlling biochemical reactions or to prevent pathogenic microorganisms from proliferating. In this article, we use the term "receptor" in a broad sense, including not only the pharmacological receptors for hormones and neural t r a n s m i t t e r s but also enzymes or other globular proteins or nucleic acids.

Methods for Analysis of Structure-Activity Relationships Various approaches have been proposed for analyzing structure-activity relationships using computers. Among them, there are approaches in which the chemical structural formula is split up into component units. The individual substructural components are regarded as being significant to various extents for the biological activity, and the structureactivity relationships are analyzed a s s u m i n g t h a t the activity is controlled by combinations of the activity-indices assigned to the individual structural units contained in each structural formula. The activities of a series of compounds are expressed as functions of these indices by linear or non-linear combination methods. These approaches seem to be

just for the analyses, but not effective for understanding molecular recognition by biological macromolecules. Some of the substructures may indeed play important roles in interaction with the receptor. But, they can often be replaced by other groups with similar physical and chemical properties. As stated before, it is not just the existence of the particular structural units but the spatial alignments of physical and chemical properties of the units that are important. It seems to be quite difficult to reconstitute the separated pieces of a structural formula to obtain new molecules in the hope that they will have the same biological activity as the original molecule. Among approaches based on the physicochemical properties of molecules, Hansch and Fujita's method (2) is excellent. They have developed a method whereby the relationships between structures and activities can be analyzed quantitatively. In this method, biological activities are correlated with various physicochemical properties of substituent groups at specified positions of molecules in a series of derivatives with the same skeletal structure. By regression analyses, the activities of dozens of compounds can be represented by an equation consisting of a linear combination of several physicochemical variables. Usually, the physicochemical properties of substituent groups, such as inductive, resonance, hydrophobic, and other effects, and those of whole molecules, such as the partition coefficient and molar refractivity, are chosen as variables (3), since they make significant contributions to the activity. From the coefficient for each variable term in the equation, we can determine quantitatively the extent of the contribution of each property to the activity. This method is a powerful tool to indicate quantitatively the direction of subsequent structural modifications in order to improve the biological activity. Although the interpretation of the physical meanings of the variables is not always clear, the equation covers a number of interactions between drugs and biological systems. The method has been shown to be useful for performing lead optimization rationally and used worldwide. But, it is necessary to establish different methods for interpreting the structure-activity relationships for molecules with different skeletal structures, and for designing new molecules with different skeletons. For these purposes, efficient methods using three-dimensional structures, based on new concepts, seem to be essential.

Three-Dimensional Structures of Molecules The three-dimensional structure is the most realistic description of an existing molecule. The chemical structure itself cannot be directly related to biological activities and functions of a molecule, though it is an excellent graphic means to describe chemical bondings. However, all the features of a molecule, such as physical properties, chemical reactivities, dynamical behaviors and molecular interactions, should be interpretable in t e r m s of its three-dimensional structure. With the remarkable advances in techniques of solving crystal structures, it has become more and more easy to obtain three-dimensional structures of molecules. In the last three decades, techniques and equipment for measuring diffraction from crystals, and algorithms for solving the phase problem and for refining structures have made remarkable progress. In the field of small molecules, structure analyses can be routinely performed now. Even in the field of macromolecules, methods for structure analyses have been established (4) and structure elucidations have become progressively easier, although crystallization still remains a difficult problem. The analyses can now be applied to larger, more unstable, and more complicated molecules, and can be done with smaller amounts of samples, with less labor, and in a shorter period than before. The results of these crystallographic analyses have been put into generally available databases. The atomic coordinates of molecules and accompanying crystallographic data of small molecules are available in the Cambridge Crystallographic Database (5). Those of macromolecules are available in the Protein Data Bank (6) (National Laboratory Institute, Brookhaven). These databases have deepened our understanding of the three-dimensional structures of molecules and of molecular interactions. Especially, the crystal structures of protein-ligand complexes or DNA-ligand complexes have clarified the details of molecular recognition by macromolecules in general, as well as in individual cases.

Three-Dimensional Computer Graphics Three-dimensional structures and interactions of protein-ligand and DNA-ligand complexes can be better understood by using threedimensional computer graphics devices (hereafter abbreviated as "3DCG"), which can store images of three-dimensional objects in the

memory and apply three-dimensional transformations to the image, such as rotation, translation and scaling in real time (7). In the past decade, 3D-CG has become an essential tool for computer molecular modeling. Three-dimensional structures in the crystallographic databases or private data files can be displayed directly on 3D-CG and the molecules can be manipulated interactively (rotation, translation, and bond rotation) with input devices such as dials, a joystick, keys, and a mouse connected to the display. After manipulating or modeling the molecule, new atomic coordinates can immediately be stored in files and can be readily used for computation, and the picture can be reproduced at any time. In addition to various representations of molecular structures such as wire-frame, ball-and-stick and space-filling models, physical and chemical properties and virtual characters of molecules, such as electrostatic potentials, molecular orbitals, and expected sites of hydrogen bonding partners, can be displayed on 3D-CG, and compared visually with those of other molecules. Recently, high-performance 3D-CG workstations have become available in place of the combination of 3D-CG terminals with a host computer. Dozens of well-developed softwares for computer-assisted molecular design based on 3D-CG are commercially available and are now widely used (8). The main functions of the softwares are molecular modelling and theoretical calculations. In order to construct threedimensional structures, various procedures are provided with the softwares, and are usually performed interactively on graphic displays. Crystallographic databases or private structure files are referenced, if necessary, and the structures are subjected to further modification, such as addition or deletion of substituent groups, replacement of atomic elements, and conformational changes. Some theoretical calculations are applied for refining the geometries and for obtaining the stable conformation. But, a serious problem is that there are a number of possible three-dimensional structures in non-rigid molecules.

Theoretical Calculations The progress of theoretical calculations in the field of chemistry, such as molecular mechanics (9), molecular orbital (10,11), and molecular dynamics (12) calculations, has been remarkable. The methods are used

for estimating energetic stabilities, electronic properties, and molecular interactions. It is a characteristic of computational methods that they are applicable not only to actually existing molecules but also to imaginary structures. They are useful not only for interpreting various chemical p h e n o m e n a but also for predicting t h e m without experiments. Molecular mechanics and molecular orbital calculations can give us the minimum energy structure with its energy value, although it might not be the global minimum structure but only the local minimum near the starting structure because of the limitations of the energy minimization algorithm. These methods are very useful for refining structures in molecular modeling. Molecular dynamics calculations simulate the motions based on the potential energy calculation by using the force field and Newton's equation of motion, assuming each atom to be a particle. By solving the equation for each short time step in a certain period of time, a trajectory is obtained as a series of positions and velocities of atoms in the system. The dynamic behaviors of molecules can be simulated along the time course by using energy values and other structural features. Unlike the molecular mechanics calculation, the molecular dynamics calculation can override the energy barriers between local minima. But, it still has a limitation in getting over high energy barriers and the global minimum search is not easy even by this technique. Nevertheless, the calculation has come to be used for the purpose of finding the stable structures of super-flexible molecules, including those of solvated states, and estimating free energy difference between two similar states.

Active Conformation of Drugs The calculations described above have become indispensable tools not only in structural organic chemistry but also in analyses of structure-activity relationships in computer-aided drug design. They are of course useful for interpreting the chemical reactivity. For the purpose of drug design or analyses of structure-activity relationships, however, attention has to be paid to the fact that, in general, chemical reactions start from the most stable three-dimensional structures of the molecules involved in the reaction, whereas biological activities arise from the stable interaction of drug molecules with receptor macromolecules. For drug activities, we

10 must consider the stability of the drug-receptor complex, in place of the stability of the drug itself. Therefore, when the three-dimensional structures of receptor macromolecules are not known, we cannot estimate the stability and the stable structure of the drug-receptor complex computationally. Even if the receptor structure is known, it is not easy to find the stable mode of binding of the two molecules, because of the vast number of possibilities arising from the six degrees of freedom of rotation and translation. A "carpet bombing" search for the global energy minimum by changing all degrees of freedom is not realistic in a multidimensional system. A blind calculation of molecular mechanics or molecular dynamics does not yield any stably docked structures owing to the energy barriers. Therefore, we must prepare appropriate starting structures in order to avoid being trapped in unexpected local minima, before starting the calculation. The global energy minimum structure is often assumed to be the most stable structure among them, although this assumption is not necessarily correct. In the case of flexible molecules which have a number of rotatable single bonds, it is especially difficult to find the most stable structure in the complex because of the additional degree of freedom for bond rotation. The conformation which a drug molecule or a natural substrate molecule adopts on its receptor is called the "active conformation". The active conformation for each bio-active molecule is not necessarily the most stable conformation of the molecule itself. The active conformation can be determined most straightforwardly by X-ray crystallography on a crystal of the drug-receptor complex. Those of other drug molecules, which are known to interact with the same receptor, can be estimated based on the structure of the drug binding site. The main problems in docking procedure calculations are as mentioned above. Knowledge of active conformations is quite useful for evaluating structure-activity relationships and designing new structures, especially when the receptor structure is not known. But, it is very difficult to determine the active conformation of a highly flexible molecule without knowledge of the receptor structure. Theoretical calculations are less useful for these purposes.

ll 2. STRATEG1E~S OF OUR APPROACHES Background Because the background is extremely complicated and full of unelucidated factors in spite of recent advances in molecular biology, it seems to be most challenging to establish novel strategies for drug design. First of all, it is important to explore a rational way of drug design in general, r a t h e r t h a n in individual cases. To develop new concepts and new methodologies, effective and efficient utilization of computers seems to be an essential prerequisite, rather than classic procedures utilizing simple mimicry of the process or way of thinking of synthetic chemists, who previously carried out drug development. As it is receptors that hold the keys to biological activities, the most logical approach in drug design is to make use of receptor structures. Even if the receptor structure is unknown, provided that two or more active molecules are known, approaches based on an assumed common receptor are more rational than those based on simple similarities of their structures. We have been developing several program systems based on the receptor, as we will describe later. F u n d a m e n t a l Concepts The key assumptions underlying our concepts are as follows. 1) It is not the chemical structures or atomic positions that are recognized by macromolecules in biological systems. Recognition of a ligand molecule involves the overall intermolecular forces. It is the spatial arrangement of submolecular physical and chemical properties t h a t is important for the proper interaction between two molecules. These properties along with the contact surfaces should be complementary between two molecules. Among various intermolecular forces, the hydrogen bond is very important for discrimination between molecules. Hydrogen bonding works within a limited distance and direction,

whereas the electrostatic interaction works in all directions and over a long distance. In many crystal structures of protein-ligand complexes, ligand molecules have been found to be fixed firmly to the proteins through a number of hydrogen bonds as indicated in Fig. 1 as an example.

12

Fig. 1 Hydrogen bonds ( d o ~ lines) between/~ casei dihydrofolate r e d u c ~ and a potent inhibitor methotrexate (filled bonds) in the crystal structure. (Drawn with the atomic coordinates from the Protein Data Bank entry 3DFR (13)).

2) Molecules with quite different chemical structures can b i n d to the

Many examples are known of competitive inhibition between molecules belonging to different categories of structural types, as found by receptor assay with a radioisotopic ligand. These pairs of molecules, such as those shown in Fig. 2, might have a common three-dimensional shape and common physical and chemical properties such as hydrogen bonding, electrostatic, and hydrophobic interactions. The shape and the properties of these molecules must be complementary with those of the receptor. Furthermore, it is not the existence of the individual properties but their spatial arrangements on the molecule that are important for binding specifically to the receptor site. Flexible molecules must be able to adopt stable conformations that satisfy these requirements.

same site o f a receptor.

13 Natural and Synthetic Estrogens

Natural and Synthetic Retinoids

Substrate and Inhibitor of Cyclooxygenase

OH

~ Estradiol

Retinoic Acid

OH

Hi. ~ ~ N

HO Diethylstilbestrol (14)

0

AM80 (15)

H

Arachidonic Acid

COOH CH30~

N~' CH2COOHcH3 C=O CI

Indomethacin (16)

Fig. 2 Structure-pairs of natural and synthetic ligands (14,15,16) that bind to the same receptor sites. The binding to the same receptor site has been proved by receptor binding assay.

3) The whole structure of the drug molecule is not necessarily required for receptor binding. Inspection of the crystal s t r u c t u r e s of enzymei n h i b i t o r complexes elucidated by X-ray c r y s t a l l o g r a p h y indicates t h a t not all the a t o m s of an inhibitor molecule are necessarily involved in its interaction with a protein, as can be seen, for example, in Fig. 3.

Fig. 3 Three-dimensional structure of/,. case/ dihydrofolate reductase (thin line) and b o u n d inhibitor m e t h o t r e x a t e (thick line) in the crystal. Some atoms in methotrexate at the opening of the binding site may have contacts with molecules outside the protein. (Drawn with the atomic coordinates from the Protein Data Bank entry 3DFR (13))

14 As usual ligand molecules which fill the cavity of the ligand binding site are not totally buried in the protein, an opening cleft exists as an entrance into or an exit from the cavity. Even in the case where most of the atoms in a ligand directly contact protein atoms, the back surface of the ligand might be exposed to the outside. The structure of the exposed portion may be nonspecific, although the functional groups on t h a t portion would contribute to dissolution, partition, transport and permeability through the membrane, together with those in the buried portion. On the other hand, the buried portion of the ligand strongly bound to the receptor should have a specific structure corresponding to the target receptor. Therefore, structural modification for lead optimization should be applied to the exposed portion, if we can distinguish between the two portions. The a p p a r e n t molecular shapes of drugs t h a t are known to bind to the same receptor site often seem to be dissimilar because of the existence of the nonspecific portion. So, conventional shape analysis methods that use the whole three-dimensional structure of drug molecules would have no significance. Comparison of the surface electrostatic potentials between molecules with the same biological activities also seems to have no significance, unless the comparison is limited to the buried surface that is directly involved in receptor binding.

Structure-Activity Relationships and Designing New Structures To establish a correct model of structure-activity relationships is the s t a r t i n g point of designing new structures. For the optimization in a definite skeletal structure, quantitative structure-activity relationships based on two-dimensional structures of molecules (2) are useful to indicate an appropriate course of structural modification in substituents. For molecules with different skeletal structures, however, methods based on the three-dimensional structures of molecules are essential. Several methods have been proposed so far, although they are not sufficiently powerful to guarantee their success in rational drug design at present. When the receptor structure is known, examinations of relationships between three-dimensional structures and activity seem to be r a t h e r easy (8), and the design of new molecules by s t r u c t u r a l modification could be done without difficulty. But, even in these cases, the design of new molecules with different skeletal s t r u c t u r e s cannot be realized

15

easily. When the receptor structure is not known, the examination of structure-activity relationships as well as the design of new molecules becomes much more difficult. The constructed model of structureactivity relationships is necessarily less certain and less reliable because of an insufficiency of information. Each drug molecule may not be wholly complementary to the receptor cavity, only parts of the chemical and physical properties of the drug binding site being reflected. Use of information from multiple molecules with different skeletal structures can give a better image of the receptor cavity. The deduced receptor cavity or the structural requirement for binding to the receptor would give a useful hypothetical basis for structure-activity relationships, and contribute to the design of new structures, although each must be refined or modified repeatedly through synthetic trials. In any case, the design of new structures with different skeletons, so-called "lead generation", is so difficult that it can rarely be attained either by human work or by computer at present. In order to make lead generation possible, it is necessary to develop special methodologies where the h u m a n brain and computer give full play to their particular abilities.

Common Features of the GREEN and RECEPS Programs Based on the principles of drug-receptor interaction described above, we have developed new methods and computer programs for drug design. Among several systems developed for various purposes, we describe here two program systems for evaluating structure-activity relationships using the three-dimensional structures of molecules. One is the program system GREEN for efficient docking studies when the receptor structures are known (17,18), and the other is the program system RECEPS for rational superposition of molecules and receptor mapping when the receptor structures are not known (19). The GREEN program is based on the three-dimensional structures of receptor proteins. It enables the real-time estimation of intermolecular interaction energy between protein and ligand molecules throughout the docking process, describing the physical and chemical environment of the ligand binding site of the protein. It should be helpful in finding the stable relative geometry of protein and ligand molecules in explanations

15

of the m e c h a n i s m s of biochemical reactions and structure-activity relationships of drugs. Without information on receptor structures, the RECEPS program is based on the three-dimensional structures of multiple molecules which are supposed to bind specifically to the same receptor. In the RECEPS program, molecules are superposed in terms of submolecular physical and chemical properties, not in terms of the atomic positions or partial chemical structures as has so far been done conventionally. A threedimensional receptor model can be constructed according to the superposed structures. The model provides the size and shape of the bindingsite cavity, hydrogen bonding sites, the electrostatic character on the surface, and other structural indices. The common features of these two programs are that they (1) are based on the specific interactions between drugs and a target (2) (3) (4) (5)

receptor; make use of a three-dimensional grid to describe the physical and chemical properties spatially; utilize 3D computer graphics interactively, as an interface between the h u m a n brain and computer; yield numerical indices for indicating the validity of docking or superposition in real time; and are useful not only for interpreting structure-activity relationships, but also for designing new structures.

3. APPROACHES BASED ON RECEPTOR STRUCTURE

Docking Studies Techniques for isolation and identification of proteins have made remarkable progress in recent years, and a number of protein structures have been elucidated or are being elucidated at the atomic level. Some of these proteins are bound with small molecules such as inhibitors and cofactors in the crystal. Based on the three-dimensional structure of the protein in such protein-ligand complexes, we can simulate stable interaction modes of ligand molecules with the protein with the aid of computers (20). We can estimate the stability of the ligand molecule with arbitrary conformation at arbitrary relative position, search for the mode

17 of the minimum energy binding and determine its stability. Such approaches have often been called "docking studies" (21). Docking studies are used not only for investigating natural biochemical processes but also for examining the mode and stability of binding of drugs to the target receptor in drug design. Interaction and/or reaction of natural substrates may be difficult to study by crystallographic or other experimental methods, because of the rapid progress of enzymatic reactions. Substrate specificity, site-specific or stereo-specific reactivity, and stability of the possible intermediates can be evaluated by docking simulation. Furthermore, as the binding affinity and the binding mode can be predicted for molecules that have not yet been synthesized, such simulation is useful for designing molecules with enhanced affinity to a target receptor and for selecting candidate molecules for synthesis. A ligand molecule that can bind strongly to the target receptor should have energetically favorable interactions with the receptor with an appropriate relative geometry. In docking simulation, the problem of finding such geometry between ligand and target molecules is too difficult to be accomplished only by computational methods. Besides conformational freedom, six degrees of freedom for rotation and translation of the ligand may give rise to innumerable local minima, from which a global minimum cannot be easily discriminated. Therefore, for the time being, likely stable geometries usually have to be selected by visual judgment using the 3D-CG display before starting computation. To find a likely stable geometry and conformation, the ligand molecule is subjected to a series of interactive three-dimensional manipulations (rotation, translation, and bond rotation) inside the ligand binding site of the protein on the 3D-CG display. During the last ten years, many docking simulation studies for various purposes have been published, based on the known structures of proteins or nucleic acids.

Approaches by Other Research Groups In 1981, Connolly developed an algorithm for rapid calculation of the positions of a group of dots for representing a molecular surface (22) based on the definitions made by Richards (23). Electrostatic properties can be represented by color-coded dots according to electrostatic potentials calculated at the molecular surface from all the atomic charges in

18 the molecule. By using these techniques, Weiner et al. have shown that there is a good complementarity in shape as well as in electrostatic properties between partners in several protein-ligand complexes whose structures had been elucidated by X-ray crystal analyses (24). The representation is not only beautiful but also useful for understanding molecular recognition. Without numerical indices evaluating the goodness of fit, however, this method is not so significant for practical use in finding stable ligand geometry. The protein-ligand interaction energy is a good indicator in selecting or modeling ligand molecules with strong affinity to the target protein. Empirical energy function and force field parameters are usually used for estimating the intermolecular and intramolecular energetic stability of macromolecules. In order to find a stable geometry and conformation of the ligand molecule rapidly and effectively, the estimation should be made on every manipulation of the molecule to provide a guide to the direction and amplitude for the subsequent manipulation. But, because of the large number of atoms in proteins, it takes rather a long time to calculate the energies by using the conventional atom-pair type algorithm even on an efficient workstation at present. In addition to the six degrees of freedom of rotation and translation, the conformational freedom of non-rigid molecules makes the problem very difficult and time-consuming. Therefore, most of the docking processes on 3D-CG are performed without energy estimation, by monitoring only interatomic distances so that the atoms do not come too close to each other. In 1985, Goodford presented a new method to show favored sites for such functional groups as amino, hydroxy, and carboxyl groups, and water inside the ligand binding cavity of a protein (25). The favorable sites for each functional group and water, which are contoured at a certain energy level from the map of total interaction energy consisting of van der Waals, electrostatic and hydrogen bonding interactions, are shown on graphic displays as bird cage models. The method seems to be very useful for designing new structures by adding or modifying functional groups which are expected to enhance the binding. But, it is not suitable for interactive docking studies to find stable relative geometries of the ligand molecule.

19

P a t t a b i r a m a n et al. have presented another approximation method for real-time estimation of interaction energy between a protein and ligand (26). They used the square root of the product of the Lennard-Jones potential parameters of the two interacting atoms to approximate interaction energy between the pair. On each grid point defined in the ligand binding site, they precalculated two sets of data corresponding to the attracting and repulsive terms of the potential function. Although their method enables the real-time estimation of intermolecular van der Waals interaction energy, it is not so useful for practical purposes because other energies such as those of electrostatic and hydrogen-bonding interactions are ignored.

Details of the Program GREEN Intermolecular interaction energy between a protein and a ligand molecule is usually thought to consist mainly of van der Waals, electrostatic and hydrogen-bonding interactions. It can be calculated by the conventional empirical method by Eq. 1, where A and B are the LennardJones parameters, C and D are the hydrogen-bond parameters, rij is the distance between interacting atoms i and j, q is the atomic charge, s is the dielectric constant of the medium, and Nnb and Nhb are the number of atom-pairs included in the calculation of each energy term. E i . r t . . . . . tecutar = Eva,~ ar

W a a l s -3t- E e l e c t r o s t a t i c + E H - b o n d

Nnb Nnb Nhb ___ ~ ( A i j r i j--2 l _ B i j r i j--6 )_jr_ ~ qiqj "~- ~ (CijFij- 2I - - D i j r i j - o1 ) . . erij i,j i,j z,.l

[1]

The calculation takes a rather long computational time because of the large number of atoms in a protein and consequently the l a n e number of atom-pairs between the protein and ligand. We have developed an approximation which greatly speeds up the calculation of the intermolecular interaction energy for real-time use in docking studies. The energy calculations in our approximation method are performed in two phases, the calculation of grid point data by using the protein structure, and the energy calculation by using the grid point data and ligand structures. Once the grid point data have been calculated and stored in a memory or files, the second phase can be performed consecutively for various ligand structures with use of the tabulated data.

20 On each grid point in the ligand binding site, we calculate and store the van der Waals energy term for various probe atoms, electrostatic potential term, expected sites and characters of hydrogen bond partners in the ligand, surface code and other items. Calculation of the Grid Point Data Calculation of the grid point data is as follows. A three-dimensional grid with a regular interval (typically 0.4-1.0 A) is generated inside the binding pocket of the protein molecule (Fig. 4). On each grid point, the van der Waals interaction energy between a probe atom and the whole protein molecule is calculated by using the empirical potential function. Several types of atoms are used as the probe and the energy is calculated and stored separately for each probe atom type. Every atom species that exists in the ligand molecules to be studied is adopted as the probe atom (e.g. carbon, hydrogen, nitrogen, and oxygen). For the van der Waals energy term Gvdw, the Lennard-Jones type potential function as shown in Eq. 2 is used. In Eq. 2, rij is the distance between the probe position on the i-th grid point and thej-th protein atom. As the empirical potential parameters Aij and Bij, those given by Weiner et al. (27,28) are taken currently. Gvdw,i --

protein atoms E ( Z i j r ~ 12 - Bijr[j 6) J

[2]

The electrostatic potential term Gelc is calculated by using the Coulomb potential as in Eq. 3. In Eq. 3, the definition of rij is the same as in Eq. 2. qj is the atomic charge on the j-th protein atom. The value of this term is equivalent to the electrostatic interaction energy in the case that the probe atom bears a positive unit charge. K is a constant to convert the energy unit to kcal/mol. protein atoms

G~l~.i =

~

j

If qj

eriJ

[3]

Determination of the dielectric constant inside the protein molecule is a difficult but an important problem. A constant value, which is often used for simplicity, is not very realistic. We usually use a distance-dependent approximation for the dielectric constant (i.e. ~ = frij where f varies from

21 I to 4). The approximation may still be oversimplified, but it is better than a constant dielectric model when solvent molecules are not explicitly treated in the calculation. The model somehow incorporates shielding of electrostatic interaction by mediating atoms and ions.

Calculation of the Intermolecular Energy When a ligand molecule is placed and manipulated in the gridded region, the interaction energy between the protein and the ligand molecule can be estimated by using the three-dimensionally tabulated energy terms as described above. The tabulated data on the grid point nearest to each ligand atom are used for the calculation. The interaction energy between protein and ligand (Einter) is calculated by using Eq. 4. ligand a t o m s

k

Van der Waals interaction energy is calculated simply by summing up the van der Waals energy term Gvdw(k) on the nearest grid point from the k-th ligand atom. Among the van der Waals energy terms for several probe atom types, the proper term is chosen according to the atom type of each ligand atom. Electrostatic interaction energy is calculated by summing up the product of the electrostatic potential term Gelc(k) on the

ii

LL"k,

J

r

/~

9 9

~\

9.

I/

/

f

X

\

"

.

probe atom (C,H,N,O...) 9

~ f

I

L, ~ . . . j

~'1~ ) ( / \

/

----~

\

,

/•/•

/

~

/,

/

II/

~/f

~

~ %

atom acce~ Lable I /" -"~'~\ region ( ned p ~~ \ \, 9 by Gvdw) "- ~'~\"'~'--( / Il ligand l o l e c u l e ~

9

\

~

\

/

t

protein atoms ~ , . ~ .

Fig. 4 Calculation of the grid point data.

Fig. 5 Calculation of the interaction energy by using the grid point data

22 nearest grid point from the k-th ligand atom and the atomic charge qk on the k-th ligand atom. It would be better to use interpolated values derived from those on the eight neighboring grid points rather than those of the nearest grid point Hydrogen B o n d s

Hydrogen bonds play an important role in the specific recognition of molecules in biological systems. The hydrogen bonding force originates essentially from a combination of van der Waals and electrostatic interactions. But, some empirical force-field calculation methods include the hydrogen-bonding energy term in addition to the van der Waals and the electrostatic energy terms for practical reasons. Several types of potential functions have been proposed to express hydrogen bonding force, where the hydrogen atom as well as the hydrogen donor and acceptor heteroatoms are treated taking into account the atomic distances and angles among them (29,30,31). Hydrogen bonding energy in such functions could easily be calculated, if the coordinates of all atoms involved are known. The positions of hydrogen atoms in protein molecules, however, usually cannot be determined by X-ray crystallography. There are some functional groups such as hydroxy and amino groups whose hydrogen cannot take definite positions because of some degrees of free rotation. Moreover, it seems to be unnecessary to elaborate in calculations of the uncertain energy term in a docking study where the protein structure is assumed to be rigid as a first approximation. Imprecise estimation of hydrogen bonding energy is thought not to be significant, if we consider an allowed flexibility of actual protein atoms. In the GREEN system, we decided not to calculate hydrogen bonding energy using potential functions, but to count the number of hydrogen bonds possibly formed at the current position of the ligand molecule during the docking process. The GREEN system provides a function to calculate the expected region of the hydrogen bonding partner according to each hydrogenbonding functional group, such as hydroxy, primary sp 3 and secondary sp 2 amines, aromatic ring nitrogen, and carbonyl groups, taking into account the directions of lone pairs and hydrogens attached to the heteroatoms as well as the distances. For all the functional groups in a protein molecule, the expected regions are calculated and each grid point is examined to see whether it is inside the region or not. A hydrogen

23 bonding flag, which also expresses the hydrogen bond character, donor or acceptor, is assigned to the grid point inside the region, and stored as one of the grid point data. During the docking study on 3D-CG displays, the hydrogen bonding flag in the grid point data is used to detect possible hydrogen bond formation between the protein and ligand. For each functional group in the ligand molecule, the hydrogen bond flag of the nearest grid point is referenced. In order to refine the ligand geometry to the precise minimum, energy minimization by means of the Simplex algorithm (32) can be performed, where rotation, translation and bond rotation of the ligand molecule are allowed. Optionally, van der Waals and electrostatic energy terms can be calculated by the conventional atom-pair type method in the minimization. More precise energy refinement which takes into account all degrees of freedom of the protein-ligand system should be done by using an external molecular mechanics program such as AMBER (33) or CHARMm (34).

Visualization Tabulated data are used not only for energy calculation but also for visualization of the physical and chemical environment of the drug binding site of the protein on the 3D computer graphic display. This facilitates the initial introduction of a new ligand molecule into the ligand binding site. By using the van der Waals energy term in the tabulated data, an "atom acceptable region" can be displayed. The region is defined as a group of grid points whose van der Waals energy term Gvdw is below a certain level (usually taken as 0.0 kcal/mol). On the 3D-CG display, the region is shown as a "bird cage" r e p r e s e n t a t i o n by threedimensionally contouring the van der Waals energy. As van der Waals energy terms are prepared for several probe atom types, the region can be defined for each atom type. The cage is usually color-coded according to the levels of the electrostatic term of grid point data. Plate 1 shows the structure of horse liver alcohol dehydrogenase, whose structure is solved as a complex with coenzyme NADH, catalytic Zn 2+ ion and inhibitor dimethylsulfoxide. Atomic coordinates were taken from the Protein Data Bank entry 6ADH (35). In Plate 1, the dimethylsulfoxide molecule at the active site was taken away from the crystal

24 structure, and grid point data were calculated on each grid point generated in and around the region which the ligand molecule occupied. The atom acceptable region is represented by a bird cage which is contoured at the energy level of 0.0 kcal/mol for van der Waals term Gvdw of the carbon probe. The color of the cage indicates the electrostatic potential term Gelc from the charges of protein atoms. It is clear that the electrostatically most positive region (red to yellow) extends near the catalytic zinc ion. In Plate 1, substrate ethanol is fitted to the "atom acceptable region" (ball and stick model). With such a cage representation, one can dock molecules much more efficiently and rationally than with the conventional docking procedure as shown in Plate 2. Furthermore, such a representation helps one to model new drug molecules which are highly complementary to the binding site cavity in shape as well as electrostatic character. The "atom acceptable region" may appear similar to the conventional molecular surface representation. But, the molecular surface representation of the ligand binding site is based only on the van der Waals radii of protein atoms, whereas the radii of the ligand atoms are also taken into account to some extent in the "atom acceptable region". The region shows spatial positions which the center of each ligand atom can occupy without severe contacts with protein atoms. The "atom acceptable region" is more useful than the molecular surface, because it clearly shows the energetically favorable region for the binding of drug molecules. The hydrogen bonding flag in the grid point data is used to display the "hydrogen bonding region" representation. The region is either shown as a "bird cage" picture by surrounding the grid points where hydrogen bonding flags are set, or as groups of small symbols at grid points. The cages or symbols are color-coded according to the type of protein functional group affecting the region. The representation shows that the displayed region is affected by the hydrogen-bonding functional group on the protein molecule. If a hydrogen bonding partner exists in this region, then a strong interaction would be expected between the partner and the protein.

25 Plate 3 shows the "hydrogen bonding region" in a part of the substrate binding site of E. coli dihydrofolate reductase (13). The colors of the cages indicate the hydrogen-bonding characters expected from the protein functional groups affecting the region. The characters are divided into three types: hydrogen donor, hydrogen acceptor and ambivalent. Red: hydrogen donor region which is affected by hydrogen-donating functional groups of protein, such as arginine and lysine side chains and main-chain amide N-H. Blue: hydrogen acceptor region which is affected by hydrogen-accepting functional groups, such as main-chain carbonyl oxygen and aspartate and glutamate side chains. Yellow: ambivalent region from functional groups which work either as hydrogen donor or as hydrogen acceptor (free-rotating hydroxy and water molecule). The protein structure is shown by a pale-colored skeleton, and the inhibitor methotrexate, which is bound in the crystal, is shown by a yellow skeleton. It can easily be seen that the functional groups of methotrexate are located at complementary positions to the hydrogen bonding regions of the protein. Representation of the "hydrogen bonding region" is useful for locating the positions of hydrogen bonding functional groups of drug molecules during the docking operation. Furthermore, the representation helps one to design positions of complementary hydrogen-bonding functional groups, when one wants to create drug molecules with more specific hydrogen-bonding capability. Plate 4 simulates the position of an inhibitor, trimethoprim, in the atom acceptable region of dihydrofolate reductase. The position of inhibitor methotrexate in the crystal structure is also shown for comparison.

Designing New Structures Using the Program GREEN. The program GREEN is useful not only for docking studies, but also for designing new structures directly based on the receptor structures. The program provides functions for model building, such as connecting fragment structures, addition or deletion of atoms or groups and replacing atomic elements. With the stable structures of the complex obtained by docking studies or the crystal structures of the drug-receptor complexes, it is possible to modify the drug structures by adding or replacing substructural fragments so as to obtain more favorable structures for interaction with the receptor. The various energy calculations and

25 visualizations provided in this program serve this purpose. In addition to lead optimization, the program is also useful for lead generation. One can construct new molecular structures interactively on 3D-CG, so as to fit well the cavity shape and properties. Structures should be constructed so that functional groups can interact with those of the receptor as much as possible, and so that the atoms can fit well inside the cavity. At the same time, the structures should be stable, or at least not unstable, intramolecularly, and not be too close to receptor atoms. The validity of the constructed structure is monitored by real-time energy estimation at eve,--] step of the procedure. In addition to this interactive approach, we are developing methods for automatic generation of new drug structures t h a t satisfy the shape and various properties of the receptor cavity. By these methods, it should be possible to obtain structures with new skeletons and new functional groups, among which a new lead compound might be found.

Summary of the Program GREEN The program GREEN has been developed for rational docking simulation and also for the construction of new structures based on the receptor structures. As regards docking simulation, the program covers almost all the necessary functions. In addition to the functions that are commonly implemented in the conventional programs for computer-aided drug design, the program GREEN provides the following features: (1) Real-time estimation of the intermolecular interaction energy by the approximation method, together with precise calculation of the energy in the conventional atom-pair-type calculation. (2) Representation of the "atom acceptable region" and physical and chemical properties, such as electrostatic potentials and expected hydrogen bonding sites in ligands. These features facilitate the initial introduction of new ligands to appropriate positions inside the receptor cavity on 3D-CG. (3) Real-time calculation of the intramolecular energy of the drug molecule, for every operation of bond rotation, by using the AMBER force field.

27

(4) Memorization of trajectories of 3D manipulation. Stable geometries can easily be retrieved after a series of interactive docking studies by use of the memorized geometries and energies. (5) Partial energy estimation, which enables a head-to-tail fitting for flexible drug molecules. (6) Interactive optimization of geometry and conformation of the drug molecule by the Simplex method. (7) Display of the contribution of each atom in the drug molecule to the total intermolecular interaction energy. (8) Display of the electron density map from crystallographic analyses of protein-ligand complexes. For determination of the position and structure of the ligand, energetically stable ones can be referenced by superposing them on the ligand electron density. (9) Interactive molecular-modeling functions which enable us to design molecules fitting well to the shape and various properties of the cavity. These are expected to be useful not only for lead optimization but also for lead generation as indicated before. In order to select the most probable structure of the protein-ligand complex, it would be desirable to compare several possible structures of the complex. If necessary, they should be fully optimized by energy minimization, taking into account the flexibility of the protein molecule. In our method, structures are refined by calculations which are done outside the GREEN program by using the AMBER or other molecular mechanics/dynamics packages developed for macromolecules. The GREEN program should provide an efficient tool not only for interpretation of the structure-activity relationships of various drug molecules, but also for the design of new structures based on the known receptor structure. 4. A P P R O A C H E S BASED ON MOI~ECULAR S U P E R P O S I T I O N

When the receptor structure is known, rational approaches seem to be feasible to some extent. However, it seems to be very difficult to find rational approaches, when the receptor structure is unknown. Nevertheless, most drug development studies have to be made without any knowledge of receptor structure, at least initially. So, drug design is done on the basis of comparison of the structures of a number of known active

28

and inactive compounds. In this situation, the elucidation of the structure-activity relationships is very important and is the starting point for designing new structures. The QSAR method has been developed mainly for this purpose. However, the method has a limitation that the design of new molecules as well as the interpretation of the structureactivity relationships must usually remain within the framework of derivatives with the same skeletal structure. It is necessary to establish approaches with three-dimensional structures of molecules, in order to compare the structures and properties of known drugs with different skeletons. The comparison of three-dimensional structures has been done for a long time by inspecting molecular models made from bamboo, metal or plastic from appropriate directions. Superposition of molecules is one of the most efficient ways to compare the structures and properties of multiple molecules. But, this is impossible with the above types of material molecular models. On the other hand, it is possible to superpose molecules on 3D-CG displays interactively or to superpose them computationally followed by visualization of the results. Such computer-aided methods enable us to store structures of the superposed molecules and to compare not only molecular structures but also physical properties with quantitative measures.

Methods for Superposing Molecules Comparison of the structures and properties of drug molecules would be meaningless, unless their biological activities are based on binding to the same receptor site in spite of their superficial similarity. This is because drugs i n t e r a c t i n g with different receptors should have different requirements for structures and properties. Molecules with apparently different chemical structures often exhibit the same kind of biological activities and pharmacological behaviors. Among them, there are many examples where bindings to the same receptor have been confirmed by receptor binding assay with radioisotopic ligands. There are many crystal structures in which a protein molecule stably binds ligand molecules whose structures are quite different from that of the natural substrate or the natural bio-active molecule. Such ligand molecules are tightly trapped inside the cavity or surface

29 cleft through hydrogen bonding, electrostatic, and van der Waals interactions, which work through space between the two molecules. This fact strongly suggests t h a t the physical and chemical properties are much more important than the chemical structure itself in these intermolecular interactions to be recognized by receptor. Therefore, the abilities of various molecules to bind to the same receptor are determined not only by similarities in molecular shape (not necessarily overall, but in part, as described before) but also more importantly by the relative arrangements of their submolecular physical and chemical properties in the threedimensional structures of the molecules. Accordingly, for the purpose of structure-activity relationships, molecules should be superposed in terms of their physicochemical properties but not in terms of their atomic positions or chemical structures. Methods for superposition conventionally used so far are: (1) l e a s t - s q u a r e s calculation specifying the a t o m - p a i r s between molecules (2) 3D manipulation of individual molecules on 3D-CG with visual judgment of the goodness of fit. The least-squares method cannot be applied easily to molecules in which the atom-pair specifications are difficult when large discrepancies exist between their chemical structures. If it can be applied, this method gives the least-squares residual as a measure of"goodness of fit". Specification of at least three atom-pairs is required for this calculation. This superposing method is routinely performed for the common skeletal part of two structures to reveal the similarities and differences in other parts. The biological activities of a series of compounds are often discussed on the basis of the similarities and differences of the volumes occupied by the two molecules. In cases where the two structures look alike, the differences in structure and properties are so clear t h a t superposing the molecules is not necessary. Superposition by the positions of heteroatoms is also often performed to examine biological equivalence, when the two structures are different from each other. But, it is not always easy to assign the corresponding atoms in the two molecules. Moreover, most of the superposition methods are done without taking into account the properties of the heteroatoms and the direction of interaction with possible partners in the

30 receptor. Although an approximate superposition might give information for substructural correspondence in a set of structurally different molecules, a significant superposition of such molecules seems to be very difficult. Another problem with the superposing method is the conformations of flexible molecules. Usually, superposition has been performed assuming the conformation of each molecule to be the same as in the crystal s t r u c t u r e , or the energetically most stable s t r u c t u r e obtained from molecular mechanics or molecular orbital calculations. But, it is doubtful whether the active conformation is the same as t h a t found in the crystal or in solution, or that of the stable state of the isolated single molecule; the active conformation may not coincide with any of these local energym i n i m u m structures. It seems to be pointless to superpose molecules with conformations other than the active conformation. In the superposition of flexible molecules, the conformations of two molecules can be varied by 3D manipulation interactively so as to fit as well as possible with each other by visual judgement. As the specification of pairs of corresponding atoms in the two molecules is not necessary, the method can be applied to very different structures. The disadvantage of such a superposition method is, however, t h a t it does not give us any numerical index of the goodness of fit. To obtain quantitative and reproducible results of superposition, appropriate indices to show the goodness of fit are necessary.

Receptor Models Three-dimensional models of the receptor cavity can be made based on the superposed structures. More accurate or more probable models would be produced based on multiple molecules which bind to the same receptor, t h a n based on a single molecule. The structure-activity relationships cannot be interpreted at all by a single active molecule. The greater the difference in structures used for the superposition, the more useful is the information obtained. In the "Active Analog Approach", Marshall et al. proposed useful definitions for the volume occupied by the receptor, based on the superposition of active or inactive molecules (36,37). They are the receptor-excluded volume defined as union of the volume of the active molecules, and the receptor-essential volume

31

defined as union of the volume of the inactive molecules minus the receptor-excluded volume. It seems to be useful for drug designers to consider the common volume, the differences in volumes of molecules, and the volume occupied by at least one molecule. The validity of the receptor model completely depends on the validity of the superposition. Therefore, superposition of molecules should be done as rationally and logically as possible. We have developed a rational method for superposing molecules based on the prerequisite of specific binding to a common receptor, and for threedimensional receptor mapping to describe the environment of the receptor cavity.

,..Program RECEPS~

Conventional Methods.)

Drug Structures

Drug Structures

in terms of spatial arrangement of physical & chemical

in terms of atomic positions

,I,

properties

9no structural correspondence required 9numerical indices to show "goodness of fit"

,I, /

\

least-squares method manual superposition specifying the atom-pairs with visual judgement 9structural correspondence required 1

Atomic Coordinates of Superposed Molecules

j

9no numerical index

Fig. 6 Superposition of molecules.

Details of the Program System RECEPS In our method, molecules are superposed in terms of physical and chemical properties by using a three-dimensional grid, whereas in the conventional methods, they are superposed in terms of the atomic positions. The specification of atom-pairs is not necessary, although a template molecule to which other molecules are superposed is required, as in other superposition methods. First, the template molecule must be chosen whose structure should be rigid or conformationally well-defined (although this limitation has been removed to some extent by the devel-

32 opment of functions for automatic superposition). On the 3D-CG, a rectangular box is set up in order to extract the essential region for specific binding to the receptor, and to determine the range of grid point calculation (Plate 5). The lengths of three edges and the position of the box are determined interactively so as not only to cover the region required by the template molecule, but also to have a sufficient reserve space for the subsequent superposition of other molecules. Then, a threedimensional grid with a regular interval of 0.4-1.0 .~ is generated inside the box. For each grid point, the following physical and chemical properties are calculated and stored: electrostatic potential, charge distribution, expected hydrogen-bonding character, flag on occupancy by each molecule, and flag for molecular surface. New molecules (hereafter called trial molecules) are superposed on the graphic expression of these three-dimensionally tabulated data. The goodness-of-fit values are calculated on the basis of spatial similarity of the physical and chemical properties of molecules by using the tabulated data. The values are displayed on the 3D-CG and updated during interactive manipulation (rotation, translation and bond rotation) of the trial molecule during the superposing process. The molecule is manipulated until satisfactory goodness-of-fit values are obtained. Trial molecules are superposed one after another, and the resultant atomic coordinates are stored in a file successively. From the atomic coordinates of every superposed molecule, the grid point data are calculated, from which united grid point data are obtained by applying weights for biological activities. These united grid point data describe the threedimensional environment of the receptor pocket. A receptor cavity model, which provides information on cavity size and shape, surface electrostatic potentials, locations of hydrogen-bonding heteroatoms and other features, can be obtained from the united grid point data. The receptor cavity model can be presented on the 3D-CG in various ways and can be further modified (including its enlargement) by superposing additional molecules. The correct superposition enables us not only to extract the structural and physicochemical requirements for the biological activity, but also to determine their required spatial arrangement. One of the major characteristics of our method is that the goodness-of-fit values can be estimated in real time t h r o u g h o u t the interactive

33

superposing process on the 3D-CG. Such values provide a quantitative measure of the extent of superposition. Goodness of Fit The current version of the grid point data file tabulates the address of each grid point, flag of occupancy by molecules, charge distribution, electrostatic potential and hydrogen bonding character. They are used to r e p r e s e n t the spatial a r r a n g e m e n t of properties of s u b s t r u c t u r e s in molecules and to calculate the goodness of fit of each molecule in real time. Goodness-of-fit values are calculated by using the tabulated data for the template molecule and the atomic data for the trial molecule, which are varied by the interactive manipulation. The goodness-of-fit terms t h a t we currently use are summarized as follows: Fshap e - - _

Number of common occupied grid points Number of occupied grid points of template tool.

Fchar9 e = __ E i

cj -

qil 2

Ei ~jl ~ j" grid point nearest to atom i

cj" charge distribution of grid point j qi" charge of atom i E i ( Vtemp,i Vtrial,i ) Felpo -- - V~/~-~i Vtemp,i 2 / ~ / E i

]Vt,-i~,,i 2

v

Vt~mp,i" electrostatic potential at the grid point i of the template molecule Vt,~ial,i" electrostatic potential at the grid point i of the trial molecule FH_bond z --

Number of common H-bonding grid points Number of H-bonding grid points of template tool.

Equations for the calculation of"goodness~f-fit" indices

The charge distributions, which we have tentatively defined from the atomic charges so as to be distributed on the grid points around the atoms in a Gaussian distribution, are calculated inside the van der Waals volume of each molecule, whereas the electrostatic potentials are calculated outside it. To improve these indices for goodness of fit, further modification of the equations, and replacement of terms or addition of new terms

34 may be required. For this purpose, the program has been designed to allow alterations to be made easily by users. Suitable terms and equations should be selected on the basis of their effectiveness by applying them to distinguish effectively the correct superposition from incorrect ones.

Hydrogen Bonds and Electrostatic Potential Atomic charges should be calculated in advance by molecular orbital calculations. In the case of a flexible molecule, the calculations are made based on the crystal structure or the energetically most stable conformation of the molecule, as the active conformation cannot easily be identified. Hydrogen-bond category numbers are assigned in advance to all hydrogen-bonding heteroatoms in the molecule. The geometries of the attached hydrogen atoms and ambiguity of their position by free rotation, as well as the hydrogen-bonding character (donor, acceptor or both) are judged according to the category number. The category number corresponds to each hydrogen-bonding functional group, such as a hydroxy O, carbonyl O, ether O, carboxyl O, amino N, amide N, aromatic N and sulfhydryl S. For the formation of hydrogen bonds, matching between the expected locations and the character of the hydrogen bonding partners of two molecules is judged during the superposition process. Allowable locations are assumed to be 2.5 to 3.1 .~ in distance and allowable deviation from the orientation vector of X-H or Y-lone-pair electrons (X, Y = N or O) is taken as 30 ~. For all hydrogen-bonding functional groups, the program provides functions for generating the positions of lone-pair electrons automatically and for predicting the possible locations of hydrogen bonding partners, taking into account the freedom of bond rotation of the C-X bond in C-X-H, and the C-Y bond in C-Y-lone-pair electrons. The correlation of electrostatic potentials between the template and the trial molecules is always calculated at the surface grid points of superposed plural molecules as discussed afterwards. The surface grid points vary at every stage of manipulation of the trial molecule.

Application to Dihydrofolate-Methotrexate System Methotrexate (MTX) is a potent inhibitor of the enzyme dihydrofolate reductase, which reduces dihydrofolic acid (DHF) to tetrahydrofolic acid

35 with the aid of the coenzyme NADPH. The structures of MTX and DHF resemble each other well, both having a pteridine ring.

H2N

N

H

(CH2)2COOH

dihydrofolate(DHF)

NH2 N H2N

N

N

I

N

II C -- N ~ CHCOOH

CH3

H

I

(CH2)2COOH

methotrexate (MTX)

Fig. 7 Chemical structures of dihydrofolate (D/IF) and methotrexate (MTX).

The enzyme has been well studied for a long time as an attractive target of rational drug design (38,39,40,41). The crystal structures of a number of isozymes from various sources and in various complexed states have been elucidated (13,42,43,44). The structure of dihydrofolate reductase

101

Fig. 8 Schematic picture of the ternary complex of dihydrofolate reductase from L. casei, the inhibitor methotrexate (MTX), and the cofactor NADPH. (Reproduced from (13) by permission of Prof. Joseph Kraut.)

35

from L. casei elucidated as a ternary complex with the inhibitor MTX and NADPH by X-ray crystallography by Bolin et al. (13) is shown in Fig. 8. The atomic coordinates are taken from the Protein Data Bank. The active conformation of MTX is assumed to be the same as in the crystal. In order to verify the validity of the program RECEPS, we have attempted the superposition of the DHF molecule on the active conformation of the MTX molecule (45). Although we can simulate the active conformation of the natural substrate DHF by means of a docking study using the known structure of the enzyme, here we discuss it by the superposition method with the MTX molecule whose active conformation is known and without using the enzyme structure. For the conformation of the DHF molecule trapped in the enzyme active site, two representative models have been proposed so far (13,40), as shown in Fig. 9 and Plate 6.

~N

TRP 21

R

~_.~TRP

N

-b - H

H

N

H--N8

~

,, H/b-H .......o\~j.~/'N~o ~ ~ factor X a .I

prothrombin

+

fibrinogen

thrombin I

fibrin I

Polymerization (Coagulation) I

fibrin polymer plasmin

+& Degradation (Fibr ino lys i s )

Fig.

1. B l o o d c o a g u l a t i o n a n d f i b r i n o l y s i s c a s c a d e s .

the

the

84 treatment

of

thrombosis

(1,2).

have been carried out Kikumoto et al. inhibitory

bleeding,

or

In a search for thrombin inhibitors,

found that arginine derivatives exhibit selective

activities

toward

different

though the active site structures o f similar amino

to that o f

acid

extensive inhibitor studies

trypsin

sequences.

judging

The

purpose

serine

these enzymes

seems

from the homology our

of

(3).

proteases

study

to be

in

their

(4,5) was

to

elucidate the mechanisms of the selective inhibition based on the three-dimensional thereby

to

structures

find

useful

of

trypsin-inhibitor

suggestions

for

the

complexes

design

of

and

enzyme-

specific inhibitors. 2.

X-RAY

ANALYSES OF T R Y P S I N - I N H I B I TO R

COMPLEXES

T a b l e 1 summarizes crystal data on the five complexes whose

In this paper, the

structures were determined by X-ray analysis.

trypsin-(2R,4R)MQPA complex will be described in detail. Table 1 Crystal

d a t a o f bovine tryps'tn

-

i n h i b i t o r complexes. .-

Inhibitor

MQPA (2R.4R)

PNPA

p31 21 6 55.34

PSI21 6 55.35 109.38 2. 5 0. 215

Space Group

Z a

b

(A)

C

Fig.

109.51 2. 5 0. 172

(A)

Resolution R-f ac t or

2 shows the

structural

MQP P21 21 21 4 55.37 56. 73 66.91 3. 0 0. 401

formula of

MQPA (2R,4s)

MQPA (2s.4R)

P21 21 21 4 55.49 56. 79 67. 07 2. 4 0.265

P21 21 21 4 63. 51 69. 07 63. 81 2. 4 0. 244

the thrombin

inhibitor,

(2R.4R)MQPA ( ( 2 R . 4 R ) - 4 - m e t h y l - 1 - I N 2 - [ ( R , S ) - 3 - m e t h y l - l , 2 , 3 , 4 - t e t r a h y d r o - 8 - q u i n o l i n e s u l f o n y l l - L - a r g i n y l l - 2 - p i p e r i d i n e c a r b o x y l i c acid)

(3).

MQPA

quinoline

is composed o f portion

and

three

portions:

piperidine

an

portion.

arginine There

portion,

are

three

asymmetric carbons in addition to the a carbon of the L-arginine. The configuration at

the 3 position of

the quinoline ring

is a

mixture of R and S .

Depending on the configurations at the 2 and

4

piperidine.

positions

of

the

(2R.4R)MQPA

is

to

be

used

there for

are

four

clinical

stereo-isomers.

purposes

as

an

antithrombotic agent. Hereafter, M Q P A is used to denote ( 2 R , 4 R ) M Q P A and other stereo-isomers are always referred to with stereo notations.

85

I b Il -I

quinol ine port ion

Fig.

2.

I

I

I.

I

Structural formula of ( 2 R . 4 R ) M Q P A .

Fig. 3 shows the formulae of PNPA( (2R,4R) - 4 - p h e n y l - l - [ N 2 - ( 7 - m e t h o x y - 2 - n a p h t h a l e n e s u l f o n y l ) - L - a r g i n y l 1 - 2 - p i p e r i d i n e c a r b o x y l i c acid) (6)

and MQP(4-methyl-l-[Nz

-[

(R, S) -3-methyl-1, 2, 3, 4-tetrahydro-8-quino-

8:"

l i n e s u l f o n y l l - L - a r g i n y l l p i p e r i d i n e ) (3).

0 - A r g -N

3

COOH

PNPA Fig.

3.

w

MQPA

has

strong

thrombin

toward

bovine

and

procedure

Crystallization I I I),

selective

(Ki=O. 0 9pM),

for

X-ray

was

inhibitory

but

still

(Ki=5.OwM).

trypsin

crystallize the complex o

0 . 01 ml

MQP

Structural formulae o f PNPA and MQP.

bovine

the

- A I- g - N >

activity

significant

We

therefore

carr ed

ys

of

is

out

by

the

the

t ryps i n-MQPA

hanging

length g r e w

in 2 weeks.

Crystals of

T h e crystals are isomorphous w i t h

the native trypsin crystal registered in the P r o t e i n D a t a Bank with

an

identification code

collected

o n an Enraf-Nonius

reflections had

measured

within

3PTN

(8).

were

calculated

with

The

2.5

A

the

resolution, in

In all

5.400

of

5.967

reflections

the calculation.

coordinates

(7)

intensity data were

C A D 4 diffractometer.

I F o ( > l . O a ( l F o ~ ) and were used

phases

A

(Sigma T y p e

1. 0 mg/ml CaCle, and 0.26 M ammonium sulfate

was kept in a reservoir w i t h 0.97 M ammonium sulfate.

1 mm

to

complex.

d r o p method.

d r o p of solution containing 60 mg/ml trypsin

2. 5 mg/ml MQPA,

tried

T a b l e 2 shows

bovine trypsin and MQPA. ana I

toward activity

trypsin

Starting in

the

86 Table 2 X-Ray analysis of the trypsin-MQPA complex at 2.5 A resolution. (1) Crystallization by the hanging drop method w i t h a-55. 3420. 0 3 , c-109. 5 1 2 0 . 2 0 P3121. Z = 6

(NH4)2SO4

A,

isomorphous w i t h the native trypsin crystal, PDB 3PTN

(2) Electron density m a p calculated with the 3 P T N parameters R=O.31 for 5400 reflections w i t h lFol 2 I.Oa(lFol) (R-0.24 for 3 P T N 1Fols) (3) Hendrickson-Konnert refinement

trypsin. MQPA’s quinoline and arginine R=O. 328 3 R=O. 258 trypsin, whole MQPA and 70 waters R=O.172. RMS deviation of bond distances RMS deviation of angle distances

native trypsin crystal.

A

reliability factor

= =

0.014 0.033

A

i\

(R-factor) of 0.31

was obtained, while the corresponding value for the 3PTN data w a s

0.24.

The

small difference

3 P T N coordinates were complex crystal.

Fig.

a good

in the R-factors approximation

indicates that

for

the

the

trypsin-MQPA

4 shows the electron density map o f

Fig. 4. Electron density map of the trypsin-MQPA complex. The quinolinesulfonyl group i s shown.

the

87 trypsin active site of the complex calculated at this stage. map quality w a s s o

good

that

the directions of

The

the two sulfonyl

oxygens and the 3-methyl group attached to the quinoline ring were easily

identified.

piperidine quinoline refined

portion and

the

But, was

(9).

and

isotropic

of

three-dimensional

group,

and

temperature

factors

of

the Hendrickson-Konnert the piperidine

After all non-hydrogen atoms of MQPA identified and

included

Fig.

in

5 shows the

structure of the trypsin-MQPA complex.

refinement

not

the carbonyl

T h e present R-factor i s 0. 1 7 2 .

the refinement.

the the

showed clear electron density for

ring and the carboxyl group.

could

first

of

fitted

MQPA with

were refined, 70 water molecules were

We

we

so

R-factor was decreased from 0.328 to 0.258 and

The

the second map

Similar

interpretat ion

possible,

arginine portions without coordinates

trypsin and the fitted part program

definite

not

see

the

is

in progress

electron

density

for of

the other

complexes.

(2s. 4S)MQPA, probably

because of its weak activity (Ki>500pM).

Gly193

Ser-2 1 7

Trp215

5. Three-dimensional structure o f the trypsin-MQPA complex. Bold line, MQPA; thin line, trypsin. Fig.

3.

THREE-DIMENSIONAL STRUCTURE OF THE TRYPSIN-MQPA COMPLEX In Fig. 6.

of

and

trypsin-BPTI trypsin-APP

the trypsin-MQPA structure is compared w i t h those (bovine basic

pancreatic trypsin

(p-amidinophenyl

pyruvate)

inhibitor)

(10) complexes.

(10)

BPTI

and A P P are bound to trypsin in a similar way through the specific hydrogen bonds at two locations ; ( 1 ) at the bottom o f active site hole and ( 2 ) other hand. MQPA

at the so-called 'oxyanion

the trypsin

hole'.

shows unique hydrogen bonding ; (1)

On the

the hydrogen

88 bonds

at

the bottom

of

the

hydrogen

bonds

at

the

carboxyl

group

and

Gly-193

active s i t e hole

oxyanion

through a water molecule,

hole

and

and

are

Ser-195

(3)

are preserved, formed

main

(2)

between

chain

the

nitrogens,

the carbonyl o x y g e n of the M Q P A

arginine is hydrogen bonded to the G l y - 2 1 6 main c h a i n nitrogen and the

nitrogen

carbonyl those

of

the

MQPA

arginine

oxygen of Gly216.

found

which

the case, does

not

8-pleated

have

the

bonded

carboxyl

the to

T h e carboxyl But this

the trypsin-MQP c o m p l e x

group

was

shown

to that of the trypsin-MQPA

scheme seems to be the result of

to

bonds a r e similar sheet.

lead to the unique scheme.

since the structure of

analysis to be similar present

hydrogen

T h e s e hydrogen

in an anti-parallel

g r o u p at the piperidine might is not

is

by

X-ray

complex.

The

the stable conformation

of the M Q P A molecule itself because the conformation is similar to that found in M Q P A single crystals.

trypsin-APP

trypsin-BPTI

trypsin-MQPA

F i g . 6. Comparison o f three-dimensional structures. Bold lines indicate hydrogen bonds.

4.

MECHANISMS OF SELECTIVE INHIBITION T h e inhibitory activities of

(3), MNP (4-methyI-l-[N2-

MQPA

( 7 - m e t h o x y - 2 - n a p h t h a l e n e s u l f o n y l ) - L - a r g i n y l ~ p i p e r i d i n e ~(6),

PNP

(4-phenyl-l-CN2- (7-methoxy-2-naphthalenesulfonyl) - L - a r g i n y l l p i p e r idine)

(6) and BAP (4-benzyl-l-IN2 - (2-anthraquinonesul fonyl) - L - l y -

syllpiperidine) trypsin,

all

(11)

toward

from bovine

a-thrombin,

sources,

are

factor shown

Xa.

plasmin

in T a b l e

3.

and

These

data c a n be interpreted o n the bases of the X - r a y s t r u c t u r e of the trypsin-MQPA c o m p l e x s h o w n in Fig. 7 and the molecular models of thrombin.

factor X a and plasmin

built

according to the homology

of the amino acid s e q u e n c e s shown in T a b l e 4. Table

3 describes

the

selective inhibition.

three

binding

sites

T h e lower half of

responsible

for

the

T h e s e are indicated by arrows in Fig. 7 and

89 Table 3 I n h i b i t o r y a c t l v i t i e s and d e s c r i p t i o n s o f b i n d i n g s i t e s . a ~

(Ki

Activity lnhibi tor

Thrombin

CH30@@SOZ-Arg-N3 MNP

or

Factor X a

0. 0 7 2

pM) toward

158'.

Plasmin

Trypsin

NA

NA

1. 4

NA

NA

0.

NA

33

23

NA

D e s c r i p t i o n o f binding site

Binding site

A1 a

Arginine binding s i t e

Q u i n o l i n e binding s i t e

Wide

A1 a Wide

Leu

T Yr

Med i u m

Deletion of 6 residues Wide

Narrow

Insertion o f 10 residues Narrow

P i p e r i d i n e binding s i t e

S er Nar r o w

Insertion

Leu99 Medium

Insertion of 5 residues Med i u m

2

of

Serl9O Na r r o w

residues Med i urn

I le63 Wide

a NA: D a t a not available.

Table

4.

T h e arginine binding

trypsin.

The

quinoline

piperidine binding s i t e data

raise several

site

binding

is near

questions

is a region near S e r - 1 9 0 o f

site

Ile-63

concerning

is (*

near

Leu-99

His-57).

and

the

T h e activity

the selectivities, w h i c h

c a n be answered as follows 4 . 1 Why a r e a r g i n i n e d e r i v a t i v e s s u c h a s MQPA a n d MNP i n g e n e r a l

more a c t i v e o n t h r o m b i n t h a n o n t r y p s i n ?

T h e reason is that the arginine binding s i t e in thrombin h a s Ala

at

the

position

arginine binding

corresponding

site o f

thrombin

of trypsin. T h e arginine portion o f

to

SerlSO

of

trypsin.

The

is consequently w i d e r than that the inhibitors c a n thus easily

90

%

Asp I94

x

Gly 193

Fig. 7. Three binding sites of the trypsin-MQPA complex r e s p o n s i b l e f o r s e l e c t i v e i n h i b i t i o n ( i n d i c a t e d by arrows). Table 4 C o m p a r i s o n o f amino a c i d sequences o f b o v i n e enzymes. T r y p s i n - l i k e r e g i o n s o f t h r o m b i n , f a c t o r Xa and p l a s m i n a r e shown. Numbering i s t h a t f o r chymotrypsin. " / " and ' I . ' ' denote a d e l e t i o n and insertion In the chymotrypsin sequence, respectively. A s t e r i s k s w i t h arrows i n d i c a t e s i t e s r e s p o n s i b l e f o r s e l e c t i v e inhibition.

THROMBIN FACT0 XA PLASMIN TRYPSIN

TH FX PL TR

TH

FX PL TR

TH FX PL TR

16

20

I VECQ IVGCR IVCGC IVGGY

80 HSRTRYERKV RNTQ//EGDE HNEKVREQSV DNINVVECNE

140 HAGFKCRVTG /QTKTCIVSG AARTECYITG /AGTQCLISG

40 50 V H L F R K ~ P Q E LL C C A S L I S D R ALLVNE.ENEC FCCCTILNEF VSL/RR.SSRH FCCCTLISPK VSL/N/.SGYH F C C C S L I N S Q

30

DAEVCLSPWQ DCAEGECPWQ VSKPHSWPWQ TCCANTVPYQ

****

.

. . . . . 150

WGNRRETWTTSVAEV F G R T H E K . . . . .C R L W G E T Q / / ..... G T F W C N T K S S . . . . . GTS

200 CECDSGCPFV CQCDSCCPHV CQGDSCGPLV CQCDSCCPVV

90

EKISHLDKIYI EHAHEVEMTVK QEIP.VSRLFR QFIS.ASKSIV

110

RDIALLKLKR FDIAVLRLKT ADIALLKLSR NDIHLIKLKS

160

170

x

********

220 CIVSWCE/CC GIVSWCE/GC GVTSWCL/GC GIVSWCS/GC

PLVERPVCKA PYVDRSTCKL PVIENKVCNR PILSNSSCKS

.....

... ..

70

YPPWNKNFTVDDLLVRIICK Q...AKRFT.....VRV/GD N I L A L S F Y K . . . . .V I L / G A S.....CIQ.....VRL/GE

I(*************

100 HPRYNWKENLO HSRF/VKETYD EP//////.SQ HPSYNSNT.LN

QPSVLQVVNL SS/TLKHLEV GECLLKEAHL YPDVLKCLKA

210 HKSPYNNRWYQN TR..FKDTYFVT CF..EKDKYILQ CS..CK////LQ

60

WVLTAAHCLL YVLTAAHCLH WVLTAAHCLD WVVSAAHCYK

120 P I E L S D YI H P PIRFR/NVAP PAIlTKEVlP AASLNSRVAS

.. 180 S..TRIRITNDM S..SSFTITPNR NEYLDGRVKPTE A..YPCQITSNM

. 230 DRNCKYGFYTH ARKCKFGVYTK ARPNKPCVYVR AQKNKPCVYTK

.. 130 VCLPDKQTAAKLL ACLPEKDWAAETL ACLPP.P..NYHV ISLPT./..SCAS

. .. . . 1 9 0 FCACYKPCEGKRGDA FCACY..DTQ.PEDA LCAGH..LIG.GTDS FCAGY..LEG.CKDS

240 VFRLKKWIQKVIDRLGS VSNFLKWIDK IHKARACAACSR VSPYVPWIEE THRRN VCNYVSWIKQ TIASN

91 enter oy

the active s i t e of thrombin, w h i l e the shorter contact w i t h

of Ser-190

The

result

in

explanation. 0 . 94

trypsin

As

results

the

present

shown

in Fig.

of

in

lower

X-ray 8.

inhibitory

analysis

activity.

supports

this

the O y atom of S e r - 1 9 0 moved

A in the d i r e c t i o n away from the active s i t e hole, leaving Serl90

I

'

F i g . 8 . Movements o f S e r - 1 9 0 Oy a n d L e u - 9 9 C 6 1 upon b i n d i n g w i t h MQPA. Thin l i n e , position i n f r e e trypsin; bold l i n e . position j n the complex; dotted line, steric repulsions leading t o the movements.

more

space

in

of

rotation

the active site.

52'

around

T h i s movement

the C a - C f l

bond.

The

resultant

A,

between S e r - 1 9 0 O y and the arginine NZ is 3.74 been

A without

2.91

explanation

that

the movement.

Ser-190

acts

Oy

This to

w a s achieved by a distance

w h i c h would h a v e

is consistent

repel

arginine

with

rather than to attract them as a hydrogen bond acceptor. case o f

lysine

critical

inhibitors

factor.

determines

the

such a s BAP,

Instead,

the

selectivity.

similar

hydrogen

Accordingly.

bond

in

to

lysine

form

where

hydrogen

inhibitor

in trypsin w h i l e

thrombin

In the

steric repulsion is not a

ability

The

hydrogen bond w i t h S e r - 1 9 0 O y

the

inhibitors

Ala

can

bonds form

cannot

it

replaces

a

form a

Ser-190.

lysine inhibitors h a v e stronger activities on trypsin

than on thrombin.

F r o m this v i e w point,

the f o u r enzymes c a n be

classified into two g r o u p s according to the a m i n o acid at position 190: thrombin and arginine Ser.

factor X a have Ala.

inhibitors

their

selectivity.

pockets The

are

enzyme

narrow

their pockets are w i d e and ;

plasmin

and

lysine

s h o w selectivity

selectivities

c a n be

and

trypsin h a v e

inhibitors explained

in

show this

way, but trypsin w a s found to

h a v e smaller K m v a l u e s for arginine

substrates

substrates

specificity

than

for

could

substrates h a v e

be

lysine

explained

two binding

sites

by (i.e.

(12).

supposing

This that

substrate arginine

two a m i n o groups)

which

92 can

form

trypsin

hydrogen

Asp-189,

bonds

while

directly

lysine

with

the

substrates

carboxyl

have

one

group

amino

of

group

which can form a hydrogen bond with the carboxyl group through a water molecule.

T h e strong hydrogen bonds formed by arginine seem

to predominate over

the steric repulsion between

for

lysine

and

similarly,

will

inhibitors

the amino group

Thus, the K m f o r arginine is smaller than that

and Oy of Ser-190. in

the

general

trypsin

be

inhibition

stronger

than

by

that

arginine by

lysine

inhibitors.

are the activities of arginine derivatives reversed, becoming stronger on trypsin. when a benzene ring is introduced at position 4 of the piperidine as i n PNP? 4.2

Why

The

reason

Thrombin has an trypsin. dine

this

lies

insertion of

in

10

the

piperidine

amino acids near

binding the

and

become

a

barrier

against

the

there is a pocket for the benzene

Ile-63 o f

benzene

the results o f

the X-ray analysis of

shows h o w the benzene ring o f

In

ring.

ring and activity is

increased by the introduction of the benzene ring. on

site.

These inserted amino acids probably surround the piperi-

ring

trypsin.

for

Fig. 9, based

a trypsin-PNPA complex,

P N P A interacts with trypsin in the

pocket near His-57.

Fig. 9. Three-dimensional structure of the trypsin-PNPA complex. 4.3 Why is MQPA not active on factor Xa? The

reason

replaces Leu prevents

in

tight

lies

in

factor X a binding

the

quinoline

position 99 o f

at

with

binding

the

present X-ray analysis supports

quinoline

this

site,

where T y r

trypsin.

This Tyr

ring

explanation,

of

MQPA. as

shown

The in

93 Fig.

The C 6 1

8.

atom of L e u - 9 9 moved

MQPA by a rotation of 2 4 '

around

A

85

0.

the C a - C y

upon binding with

bond.

T h e rotation

increased the distance between Leu-99 C 6 1 and the 3-methyl carbon o f the quinoline from 3.70 A

hindrance.

A

A

to 4 . 5 2

and eliminated the steric

modeling study in which Leu-99 w a s replaced by Tyr

that steric hindrance would occur between T y r C E ~and the

showed

quinoline C 4 with a distance o f 2.66

A. BAP. i n h i b i t p l a s m i n ?

4 . 4 Why d o e s t h e l y s i n e d e r i v a t i v e ,

BAP has

stronger

group of

plasmin

thrombin

and

and

factor

activity on

inhibitory trypsin,

than on

Xa,

explained

as

the narrow pocket

the wide in

pocket

group of

4.1.

section

The

difference in the activities for plasmin and trypsin arises from a

In plasmin,

difference in the quinoline binding site. of 6

amino acids near

able

room a t

position 99 of trypsin provides consider-

the quinoline binding

to

the

sulfonyl

site and changes the surface

T h e bulky anthraquinone can f i l l

structure from that of trypsin. this space and f i t

a deletion

itself to this surface because group

at

the

position

it

and

can

is connected change

orientation of the ring plane by a rotation around the C - S BAP will

not

inhibit

thrombin because

inserted amino acids of

thrombin near

the benzyl

the

bond.

group hits

the

I t will also not

Ile-63.

inhibit factor Xa, because the anthraquinone hits Tyr at position 99.

Thus,

the

mechanisms

of

selective

inhibition were

explained based o n

the X-ray structures of

complexes

difference

and

the

in

amino

clearly

the trypsin-inhibitor

acid

sequences

of

the

enzyme.

SUGGESTIONS FOR DRUG DESIGN A N D PROTEIN ENGINEERING

5.

The related

explanation

For example, by

adapting

hole,

(2)

carbonyl

for

the

selective

inhibition

to the design strategies for enzyme specific

by end

is

directly

inhibitors.

factor X a specific inhibitors will be obtained; (1) the

arginine

backbone

introducing a phenyl to

increase

van der

to or

fit

the w i d e

active

a benzyl piperidine at

Waal's

interaction and

site the

(3) by

introducing a benzene ring o r an alkyl chain at the amino end to avoid the steric repulsion by Tyr. Direct protein

proofs

engineering

for

the

explanation

experiments.

Table

could

5

be

obtained

by

summarizes proposed

site-directed mutageneses and the expected results.

94 Table 5 Proposed site-directed mutageneses and expected results. (I)

(‘goSer+Ala) trypsin mutant Ki for MQPA Ki for BPTI

(2)

J , t ,

(Ala+lgoSer) thrombin mutant Ki for MQPA t ,

(3) (Ala+’goSer) factor X a mutant

Ki for MQPA

(4)

(Tyr+”Leu)

.

t

K m for Arg substrates K m for Lys substrates

J

K m for Arg substrates K m for Lys substrates

t

K m for Arg substrates K m f o r Lys substrates

t

t

J.

J

factor X a mutant Ki for MQPA J. .

T h e first mutation in T a b l e 5 will

increase the activity of MQPA

by eliminating the steric repulsion between MQPA arginine and Ser

BPTI by reducing the A similar effect will be observed for

i t will decrease the activity o f

But,

Oy.

number of hydrogen bonds.

substrates w i t h arginine o r lysine at the S1 site. arginine

will

substrates

become

substrates will become weaker.

stronger

substrates, water

the

activity

a n e w hydrogen

molecule

between

will and

Oy

the

MQPA.

of bond

Ser

lysine

of

T h e r e will be steric repulsion

between MQPA arginine and O y atom by decrease

that

T h e second and third mutations are

suggested from the same purpose. will

and

T h e binding of

introduced

In be

the

the

formed. lysine

Ser

case

and

of

via a

probably

side

it

lysine

chain.

The

fourth mutation will reduce the degree of steric repulsion between MQPA quinoline and T y r by

replacing

the T y r with

less bulky Leu

and increase the activity of MQPA for factor X a mutant. 6.

CONCLUSION

X-Ray analyses of several trypsin-inhibitor complexes provided

three

novel

lines

type of

of

valuable

information.

mechanisms o f selective inhibition. for design reliable

of

First,

trypsin inhibition and lead to enzyme

atomic

specific

coordinates

a

the

Thus, i t indicated strategies

inhibitors.

to

revealed

it

elucidation o f

examine

the

Second,

it

various

provided

simulation

methods which are used to evaluate the free energy change of drugprotein

complex

activities syntheses.

of

formation.

compounds

Third,

by

could

With be

accumulating

a

valid

predicted the

simulation before

method.

their

actual

three-dimensional

struc-

95 tures o f drug-protein complexes, w e may be able to find essential factors

in

drug-protein

interaction

and

utilize

them

in

drug

design. We thank our coworkers, C. Sasaki, C. Okumura, M. and Dr. H. Kubodera.

Miyagawa

REFERENCES

1 2

3 4 5 6

7 8 9 10 11

12

R. Kikumoto. Y. Tamao. K. Ohkubo, T. Tezuka. S. Tonomura, S .

Okamoto and A. Hijikata, J. Med. Chem.. 23 (1980) 1293-1299. J. Sturzebecher. F. Markwardt. 9. Voigt. G . Wagner and P. Walsmann. Thromb. Res., 29 (1983) 635-642. R. Kikumoto, Y. Tamao. T. Tezuka. S. Tonomura. H. Hara. K. Ninomiya. A. Hijikata and S. Okamoto. Biochemistry, 23 (1984) 85-90. T. Matsuzaki. C. Sasaki and H. Umeyama, J. Biochem. 103 (1988) 537-543. T. Matsuzaki. C. Sasaki, C. Okumura and H. Umeyama. J. Biochem. 105 (1989) 949-952. R. Kikumoto and Y. Tamao. Mitsubishi Chem. R&D Rev., 1 (1987) 26-34. Protein D a t a Bank, Brookhaven National Laboratory, Upton. New York. J . Walter. W. Steigemann, T. P. Singh. H. Bartunik. W. Bode and R. Huber. (1981) 3PTN. Protein Data Bank, Brookhaven National Laboratory, Upton, New York. W. A. Hendrickson and J . H . Konnert, Biomolecular Structure, Conformat i o n . Function and Evolution, Pergamon. Oxford, Vol. 1, 1981. pp. 43-57. M. Marquart. J. Walter, J . Deisenhofer. W. Bode and R. Huber. Acta Crystal logr.. 839 (1983) 480-490. T. Naito, personal communication. C. S. Craik, C. Largman. T. Fletcher, S. Roczniak. P. J. B a r r , R. Fletterick and W. J. Rutter. Science 228 (1985) 291 -297.

This Page Intentionally Left Blank

QSAR and Drug Design - New Developments and Applications T. Fujita, editor 9 1995 Elsevier Science B.V. All rights reserved

97

THREE-DIMENSIONAL STRUCTURE-ACTIVITY RELATIONSHIPS AND R E C E P T O R MAPPING OF QUINOLONE ANTIBACTERIALS

HIROSHI KOGA and MASATERU OHTA Fuji-Gotemba Research Laboratories Chugai Pharmaceutical Co., Ltd. 135, 1-Chome Komakado Gotemba-shi, Shizuoka 412 J a p a n . ABSTRACT:

The q u a n t i t a t i v e structure-activity relationship (QSAR) correlation equation previously formulated for a n t i m i c r o b i a l a c t i v i t y of q u i n o l o n e - 3 - c a r b o x y l i c a c i d s i n d i c a t e s that the steric f e a t u r e s of s u b s t i t u e n t s at the I-N-, 6-, and 8p o s i t i o n s are i m p o r t a n t in g o v e r n i n g a n t i m i c r o b i a l a c t i v i t y . We r e e x a m i n e d the steric features of these s u b s t i t u e n t s by a n a l y z i n g t h e i r c o n f o r m a t i o n s by a m o l e c u l a r m o d e l i n g method. The "active" c o n f o r m a t i o n of e a c h s u b s t i t u e n t at e a c h of the p o s i t i o n s was e s t i m a t e d w i t h the c o n f o r m a t i o n a l e n e r g y c a l c u l a t e d by m o l e c u l a r o r b i t a l m e t h o d s and the a n t i m i c r o b i a l a c t i v i t y of the q u i n o l o n e c a r b o x y l i c a c i d m o l e c u l e p o s s e s s i n g that s u b s t i t u e n t . A m o d e l of the receptor supposed to accommodate substituents at the respective positions was constructed by superposing active c o n f o r m a t i o n s of s u b s t i t u e n t s in h i g h l y a c t i v e c o m p o u n d s . With this a c t i v e v o l u m e or r e c e p t o r model, the a c t i v i t i e s of c o m p o u n d s that were not predicted well by the previous QSAR were qualitatively and/or semi-quantitatively rationalized. We b e l i e v e that the p r e s e n t model is useful for p r e d i c t i o n of the a c t i v i t y of compounds to be synthesized in designing new quinolone antibacterials.

1. I N T R O D U C T I O N Nalidixic

acid

antibacterial therapy is

of

urinary

effective

effective Thus,

(NA) is the

drug

family tract

against

against

efforts

and

have

been

made

as

overcome

its m e t a b o l i c

as to

Norfloxacin 1970's

by

Koga

(NFLX, I), (2),

one

to

to

and

expand

1963

which was of

the

it

to

it

synthesized

the

is

it not

bacteria. enhance

and

authors,

for

Although

antibacterial

instability first

use

Pseudomonas

modify

present

(i) .

bacteria,

and its

of a q u i n o l o n e

clinical

gram-negative

activity

well

member

in

since

gram-positive

antibacterial

(1).

known

been

infections

most

most

first has

its

spectrum,

side

effects

in the

opened

up

late new

98

cooH R k?cooH

O

O

CH3 ~ ' c O O H

O

I

C2H5 nalidixic

acid

possibilities clinical

for use

used

However, For

is

I),

further

(Table

are orally

synthesized

developed

I) .

These

quinolone

relatively

is

enoxacin(2)

low

Subsequently,

systemic

toxicity

poorly

and

much b e t t e r

than

(4)

that

a number

are

commonly

called

share

common

a

new

(3) .

absorbed.

AM-833

(see

of NFLX,

of analogs

a few of them have been m a r k e t e d

they

P . and

quinolones

(4) or

6-fluoro-7-amino

structure.

Koga activity

(2b, c)

previously

relationships

structure

3.

From

correlations log

because

including bacteria,

against

NFLX

to

against

acid-resistant

effective has

the

superior

activity

resistant

and

(2,4).

and quite

is

nalidixic

is

absorbed

compounds

fluoroquinolones

potent

stable,

drawback,

and e x p a n d e d

NFLX

bacteria

of

against

orally,

administrated

which

have been

given

this

it has

incidences

metabolically

orally

greatly.

gram-negative

low

when

overcoming

Table

and

in that

cross-resistance

and,

infection,

were

quinolones

causes

incomplete bacteria

antibacterials

of q u i n o l o n e s

gram-positive

aeruginosa,

(norfloxacin,NFLX) (enoxacin)

of quinolone

applications

previously both

I "X=CH 2 X=N 9

(I/MIC)

the

examined

(QSAR's)

of

statistical

the

71

point

was r e p r e s e n t e d by equation =

- 0.362

quantitative

compounds

with

of view,

one

structurethe

generic

of the best

[I] 9

(LI) 2 + 3.036 L1

-

2.499

(Es6) 2 - 3.345 Es 6

-

0.205

(ZK6,7,8) 2 - 0.485 ZK6,7,8

+ 0. 986 I7 - 0. 734 I7N-CO

n=71 In mole/l

-

1.023

-

0.681 ~F6,7, 8 - 4.571

s=0.274

this

equation,

(MIC)

against

(B48)2 + 3. 724 B48 r=0. 964 the

[1]

r2=0. 929

minimum

Escherichia

F=70.22

inhibitory

coli

NIHJ

concentration is

used

as

in the

99 TABLE

i. F l u o r o q u i n o l o n e s . O

F

Ra Compound

R7

AM-833

5

ofloxacin

6

amifloxacin

7

ciprofloxacin

8

CI-934

9

AM-1091

11

difloxacin

activity.

parameters (Es)

R6.

constants terms is

an

carbonyl

the

zero

maximum

and

H

is

width

(R7

inductive

8-substituents.

the

(L) the

to

be

to

V e r l o o p 's

of

the

unity the

B48

of

side) .

when

7-amino

is

it

F6,7, 8

is

the

hydrophobic site(s).

I7 is

not.

is

as

sum (F's)

I7N-CO

unity

when

a

moiety,

and

parameter

for

direction the

quadratic

7-substituent

STERIMOL

parameters almost

The

substituent

another

perpendicular

with

is

expressed

at

parameter

hydrophobic

action the

STERIMOL

substituent steric the

to

their

when

is

electronic

Equations

sum

related to

zero

which

of R8 in the

R 1 substituent

F

and R 8 substituents.

as

within not.

of

compounds

variable

does

for

equal

expressed

exists

one

length

seem of

variable

it

Lupton-Hansch

N-X___/

R 6, R7,

is

group

>

of the

indicator when

F

stands

transport

and

--

the T a f t - K u t t e r - H a n s c h

parameter

the

H

C2H 5 -

7, 8

indicator

another

the

this

on

hydrogen

as

(~'s)

of

effect

~6,

CH3NH--

F

the

Es 6 r e p r e s e n t s

H

_

L1

representing

i.

of

CH3-N

F CH2 CH2 --

--OCH2CH (CH3) - -

/--A N-k__/

~'~N

H

R1

F

/--A N-k__/

~ NH~ N

PDI17558

position

HN

~N

10

biological

CH3-N

R1 R8

/--I CH3-N N -k__Y /---A CH3-N N-k__/

4

/COOH

II

of

opposite the

of the

equivalent,

or

to

Swain6-,

7-,

slightly

100 poorer the

statistical

significance

hydrophobicity

whole

of

Equation

[i] indicates

compounds

having 4.2

fluoroethyl, substituent

piperazinyl,

a

of

to

be

NFLX.

al.

(5)

almost much

this

These since

few

some

[i]

N-l-aryl

activity

analogs,

the

been

good

elaboration effects

quinolone

of

8-positions

of

analyzing

their

the

effects

rationalize quinolones

in

has not been

steric

steric

three-dimensional

of

found

detail

by

are

For

predicted

be

has

predicts

irrelevant However,

and

the

the

cases

of

a

due

general of

applicability

compound

quantitative

substituents

of

(II)

[i]

et

to

attempted.

map

Chu

ii

and

structure-

In this

chapter,

at

i-,

the

systematically and

substituents a possible

6-, in

more

attempted

in

terms

receptor

we and

of

region

to the of

3.

Equation and

not

substituents.

conformations

to

and/or

example,

equation

to

the

antibacterials

structure

been

compounds

ciprofloxacin(7),

2. ANALYSIS OF THE STERIC EFFECT OF 1-SUBSTITUENTS alkyl

have

of

(ii) .

assumed

except

R8

or more

synthesized

activities

compound been

an

methyl,

to

predictions

However

parameters

have

is

relationship

reexamined

have

steric

deviants

equation

and

to p o s i t i o n

bromo,

activities

developed.

NFLX.

for this

deviations of

as

(e.g.,

-1.4

,

of

an R7

(Table i) .

whose

been

of

E s value

comparable

(i0)

that

(L)

l-(p-fluorophenyl)-fluoroquinolone

activity

lower a c t i v i t y

assignments of

that

high

an

opposite

amifloxacin(6), (4)

compounds have

the

the

methylamino,

nitrogen),

chloro,

these

of

PDI17558

industries

some

found as

fluoro,

evaluations

equation

with

oxygen,

activities

ofloxacin(5),

d e v e l o p e d by various Recently,

(e.g.,

and

length

aminopyrrolidinyl)

exhibit

A M - 1 0 9 1 (9) ,

a

approximately

Subsequently, by

such as AM-833(4),

by

i

could

valid

C I - 9 3 4 (8) ,

well

of

for

of the RI, R6,

methoxy,

R 6 substituent

value

P

K7 for

It p r e d i c t s

with

vinyl, chloro,

either log

factors

(B4) in the d i r e c t i o n

1.8

oxygen)

that

ethyl,

aminopiperidinyl,

approximately

shown

substituent an

or

for activity.

fluoro,

z

with a width

methylene, than

R1

(e.g.,

with

substituent

an

(e.g.,

with

[I].

that the steric

important

cyclopropyl),

approximately-0.65

in e q u a t i o n

are A

formulated

R 7 substituent,

instead of ~ 6 , 7 , 8

approximately

of

the

molecule

and R 8 s u b s t i t u e n t s

1

only

were

[i]

was

substituted

derived alkyl

from

groups

compounds as

the

(6)

having

R1

only

simple

substituent,

not

lO1 including

any

prediction of

R1

aryl

of the

should

factor

for

considered

group.

This

activity

be

more

alkyl

could

be

the

of N l - a r y l q u i n o l o n e s .

complex

groups,

than

when

that

reason

expressed

these

for

The

the

steric by

mis-

effect

the

length

Rl-arylquinolones

are

together.

2.1 Compounds and Biological Activity The

listed of

compounds

in T a b l e

biological

negative

2.

E.

relative

to

were

not

five

groups 2.

that

of

overall

activities

to

lowest

ranging

activity.

activities

of t h e s e

a

quinolone

activity

compounds

from

4 to

There

is

an

index

each was

were

calculated was as

tested

The

into

shown

activity,

almost

other

compound

classified

activities

0.5.

gram-

against

of

activity

highest an

as

are

antibacterials

NFLX(1) ,

relative the

chosen

representative

biological

have

activities

activities

drug,

The

their

1 compounds

of

The the

comparable.

is

their

standard which

was

coli

coli

(2,4) .

the

biological

E.

E.

parallel

under

activities

their

against

the

according

Class

the

MIC

bacteria

always

relative shows

and

conditions

and

because

roughly

coli

gram-negative

Table

The

activity

bacterium

against

because

analyzed

and

class

in

their

5 compound

8000-fold

range

in

compounds.

2.2 Conformational Analysis and Molecular Modeling 2.2.1 General Procedure:

quinolone

ring

oxolinic structures the of

was

acid

and

compound force

the

was

The were

in

using as

as

bonds

rotatable of

analyses

were

and

(15). with

5~ .

was

built

initial

of

orbital

standards

for

Nl-substituents

ethyl,

by

from use

Gaussian

the

with

of 82

cyclopropyl,

such

MO

program

the

each

Tripos

minimum-energy for

further

R6=F,

R7=Rs=H)

molecular

rotated

energy methods

(16)

modeling.

conditions

were

and p h e n y l

bond

of

method.

under

minimum

from model

standard

these

(MO)

taken

energy

6-fluoroquinolones(3"

within

The

with

of The

primary

coordinates

preliminarily

continued

were

mechanics

the

(12) .

The

conformational

of

structure

R8=H)

(13).

examined Starting

X-ray

compound

coordinates

primary

all

each

molecular

The

the m o l e c u l a r

were

compounds

by

the

the

structure

-OCH20-,

system

compound

Then,

used

of

SYBYL

7,8-unsubstituted

used

AM1

each

(13) .

Conformations increment

the

minimized were

optimization

of

from

R6-R7 =

substituents

angles.

field

structures

constructed

library

substituents

lengths

three-dimensional

(B : R I = C 2 H 5 ,

of

fragment

The

was

groups

where

with

an

conformations, as

CNDO/2

also

used

as RI.

(14) for

STO-3G

102 TABLE 2 . R e l a t i v e A c t i v i t y of F l u o r o q u i n o l o n e s .

R e lat i v e

1 14

-

1/21

R

R8

1

H

H

5(S)

Me

6 7

11 12 13 14 15

16

17

18 2

[1/4-1/16]

Me H Me H H Me Me H H

Me

H H H H H H H

H H H

30

[1/32-1/128]

31

H

2

H

7

MeNHcyc-Pr4-F-PhFCHzCH2CH2=CH2,4-F2-Ph-

H H H H H

H ~-oH-P~n -CH=CH -S -

-C

*

(CH2)=CH-S-CHJCH7S-

8t9 9 5 2 2 5 5 10

10 11

Me Me Me Me Me Me H Me

Me

-OCHzCH (CH,) (R) P h2-F-Ph2 -Me-Ph4-C1-Ph4 -Me-Ph3 , 4 - ( O C H 2 0 ) -Ph-

32

MC

H H H

33 [1/250-1/1000]

Me

H

4

Et

n-PrCH2=C13CH2HOCH2CH2PhCHz-

Me

3

ref

(S)

20 21 22 23

27 28 29

R2

(CH3)-

-OCH2CH

19

24 25 26

*

R1

H

(Me)>NCH2CH73-F-Ph4-Br-Ph-

H

H

2 5 5

4-MeO-Ph-

H

5

n

basis set w a s used for t h e G a u s s i a n 8 2 c a l c u l a t i o n s .

a n a l y s e s , all r o t a t a b l e bonds i n N 1 - s u b s t i t u e n t s

I n t h e MO w e r e rotated w i t h

103 15 ~ i n c r e m e n t minimum. each

After

of the

group

and

was

determination

introduced

program. of

or

The

using

was

the

the

by

from

Conformational compound

(35) . 35,

quinolones shown

in

energy. three

(Table

bond

Fig.

i.

There

was

methods. where

quinolone

plane.

With

decide

which

the

ring, this of

results of

the

the

information

these

two

the

the

two

energy

moiety

other

to

alone,

is the

more

each

conformation

Activity:

values

of

above

the

it

by

the

to

the

plane

is

is

are

identical

results

corresponds

it

of

conformation, and

where

however,

was

derivative

minima

is

that

active

class,

Nl-substituted

in the

minima

The

classes,

and

calculated

difference

der each

Nl-cyclopropyl

the

of

the

van for

compound

among

showed

with the

7-piperazinyl

energy

cyclopropyl

and

of

from

was

site

system.

Conformation

the

the AM1

optimized

different

in a c t i v i t y .

substantial

One

SYBYL

active

activity

and the the

compounds

volume

The

at

the

the of

rotated,

The no

having

the

(7),

of

it.

of

less

2).

was

using

superposing

started

highest

by

by

Active

was

compounds

Nl-substituents

total

Ciprofloxacin the

that

in

the

between

examined

conformation the

of

local

conformation

compound

volumes

the v a r i a t i o n s

has

NI-RI

MO

the

that

model

to

by

routine

total

analysis

compound

as

Nl-substituents

2.2.2 Relationship

the

energy

the

or N - m e t h y l - p i p e r a z i n y l

each

close

calculated

the

around

optimized

of

occupied

MVOLUME

reflect

the

were

defined

subtracting

class

where

minimum

7 of

value

was

between

to

position

volume

of

difference compound

of the

or an e n e r g y

"total"

estimated

5~ i n c r e m e n t

conformation

conformation

classes

with

structures

receptor

volume

assumed

at

whole

"active"

energy

"active" Waals

of

The

action

scanned

Nl-substituents , a piperazinyl

conformations

minimum

then

below

of the

impossible

responsible

for

to the

activity. Conformation structure more

and

insight

compounds. of

the

(7) .

S

into

isomer

The

S

(17).

[3: R6=F,

two

stable

R7=H,

ofloxacin activity,

the

Ofloxacin

CH2CH2CH(CH3)-] isomer

of

potent

(5)

5(S) isomer is

is of

also

(5) , was

active has

optical than

S-25930 reported

Conformational

analysis

without

in

of

more

of

the

showed

to

the R

5(R)

R8-RI =than

the

R

of

5

compound

the

energy

S

1

activity

isomer

R7=CH3, model

rigid obtain

class

the

active

that

significant

fairly

and

the

R6=F,

be

a

order of

isomers

that

[3 : to

R 8 - R I = - O C H 2 C H ( C H 3)-]

conformations

has

conformation

two

higher

which

analyzed

isomer

has

difference.

104

10

8

6

v

4

35 The d i h e d r a l

2

a n g l e was d e f i n e d as 2-1-1'-3'

0 -60

-120

0

60

120

Rotation 8 (degree)

A : Energy c u r v e c a l c u l a t e d by G a u s s i a n 82 B : Energy c u r v e c a l c u l a t e d by AM1 c : Energy c u r v e c a l c u l a t e d b y CND0/2

(STO-3G)

F i g . 1 . R o t a t i o n a l Energy Map of t h e N1-R1 Bond o f t h e 1C y c l o p r o p y l Compound35 (Reproduced from r e f . 6b by p e r m i s s i o n o f t h e American Chemical S o c i e t y ) .

is

One

that

where

the

branched

methyl

moiety

is

p e r p e n d i c u l a r t o t h e quinolone r i n g p l a n e and t h e o t h e r i s t h a t where it

i s oblique t o t h e plane. From t h e s i m i l a r i t y o f t h e t h e l a t t e r was s e l e c t e d t o match

o v e r a l l shape of N1-substituents,

one o f t h e c o n f o r m e r o f t h e N 1 - c y c l o p r o p y l

compound 35,

i n which

t h e c y c l o p r o p y l g r o u p i s l o c a t e d above t h e p l a n e o f t h e q u i n o l o n e ring.

Consequently,

the

matched

c o n f o r m e r s were

regarded

having t h e "active conformation" f o r t h e N1-substituents

as

of 5 (S)

and 7 ( F i g . 2 ) . F o r t h e N1-ethyl

g r o u p i n compound 3, i n which R1=C2H5, Rg=F,

R7=Rg=H a s t h e model o f n o r f l o x a c i n

(l), t h e r e a r e t h r e e e n e r g y

minima where t h e e t h y l g r o u p i s a b o v e , b e l o w a n d p a r a l l e l t o t h e quinolone r i n g . R1)

The b o n d - r o t a t i o n a l b a r r i e r s a r o u n d t h e bond

b e t w e e n t h e s e t h r e e c o n f o r m e r s were n o t h i g h .

which

the

e t h y l g r o u p was

above t h e

r i n g was

(N1-

The model i n

selected as

the

105 active of

conformation

5(S)

and

For

the

Nl-phenyl

conformational which and

the

showed

those

of the N l - s u b s t i t u e n t s

of

compounds

compounds

I,

5(S),

compounds

of

class

conformers

as

substituents

and

R6=F,

two

the

quinolone

and

benzene

The

latter and

(Fig.

best

highly

RI=C6H5, are

1 were

those of

that

24 7

(3" there

between

respectively.

conformer

it m a t c h e d

derivative

search

angles

i00 ~

because

7 best.

Ii

2) .

selected matching

active

was by

active

as

the

with

from

active

compounds

are

the

the

and

80 ~

active of

of o t h e r

low

energy

conformation

(i,5(S),

in

those

conformers

similarly

a

minima,

rings

selected

comparison

The

R7=R8=H),

energy

7)

of

NI-

(Figs.

2-

4). The

active

superposed

by

total

volume

model

an

the

be

is

activity energy

the

of

2

they

occupies

selected

in

compounds. as

For

the

region

are

of

two

The

form.

conformation the

of

substituents (Figs. We

class

2

former this

L

with

a

well, be

biological

because was

equation the

the

of

a

activity

the

20,

21,

extended as by

22

was

least

and

their

30) ,

a

the

active

the

suggesting in

as

active

and

significant

represented

receptor

this

dimethylaminoethyl

most

[i],

of

low

of the NI-

compound

selected the

of

fairly

occupies

(19, an

end

low is

and

compounds

a number

position

have

6.

arbitrarily

in

of

should

high

are

(22)

conformations"

was

interact

to

hydroxyethyl, 3

to

model

model

model

which

substituent

analysis

length

in

conformation

of N l - s u b s t i t u e n t s

the

there

compound

in Fig.

and

This

prediction

with

The

accommodating

receptor

This

to the m e t a

seemed

active

allyl,

ring.

calculated

5).

and

this

(22),

close 24

shown

(Fig.

were

found.

conformer

Nl-benzyl as

was

receptor

compounds

Nl-benzyl

plausible

in

effect

function

The

fits

compounds

quinolone

activity.

been

a region

1

verification

derivative

The

Nl-propyl ,

substituents there

possible

for

high

have

The

their

antibacterials

compound

later.

unfavorable

the

Nl-substituted

Nl-benzyl

group

described

in

class

Nl-substituents

If a c o m p o u n d show

conformations.

phenyl

the

atoms

standard

to

novel

substituent

for

a

whenever

For

the

quinolone

as

activity.

for

of

superposed

volume"

expected

amended

class

the

active

used

biological it

matching

of

"active

highly

could

conformers

bent

factor

quadratic that

these

extended

forms

6-9). calculated

the

difference

between

the v o l u m e s

occupied

by

106

Fig. 2. S t e r e o v i e w of the s u p e r p o s i t i o n of the p r o p o s e d a c t i v e conformers of I (green), 5(S) (yellow) , 7 (blue), a n d II ( o r a n g e ) ( R e p r o d u c e d f r o m ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society) .

r

Fig. 3. S t e r e o v i e w of the s u p e r p o s i t i o n of the p r o p o s e d a c t i v e conformers of 6 (yellow) , 13 (green) , 15 (orange) , a n d 16 (cyan) ( R e p r o d u c e d f r o m ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society) .

Fig. 4. S t e r e o v i e w of the s u p e r p o s i t i o n of the p r o p o s e d a c t i v e conformers of 12 (green), 14 (red), 17 (yellow) , a n d 18 ( v i o l e t ) ( R e p r o d u c e d f r o m ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society) .

107

Fig. 5. S t e r e o v i e w of the t o t a l v o l u m e (orange) of the NIs u b s t i t u e n t s of the class 1 c o m p o u n d s ( R e p r o d u c e d f r o m ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society).

Fig. 6. S t e r e o v i e w of the s u p e r p o s i t i o n of the p r o p o s e d a c t i v e c o n f o r m e r s of 19(green), 2 2 ( y e l l o w ) , 5(R) (orange), and 2 4 ( v i o l e t ) and the d i f f e r e n c e (orange) b e t w e e n the t o t a l v o l u m e s of the set of 19, 22, 5 (R) , and 24 and t h o s e of the c l a s s 1 compounds ( R e p r o d u c e d f r o m ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society) .

108 Nl-substituents of

class

2

occupied

in class

compounds

volumes

are

increases

where

repulsions

steric

8

shows

(28),

and

the

class

(24,

resulting

ends

of

25,

in

the

fact, seen

occupy

for

meta

R8-R 1

for

the

too

compound

quinolone

31.

moiety

ring,

as

receptor.

to

the

fit

the

are

in Fig.

and

compounds in the

of

of

are

19,

the

N l-

wall, The

and

the

meta

to be

21.

In

activity, methyl

the

6, d i s t u r b i n g

(21)

the

assumed

20,

branched

below

region b e l o w the plane

and

p-methyl

receptor

one

reduce

fixed

the

7,

hydroxyethyl

to

regions

Nl-phenyl

and

1 compounds.

and

compounds

6,

those

occupied

p-hydrogen

(20) ,

methylene

(29),

not

of class

These

of

wall

the

regions be

Figs.

Nl-phenyl

corresponding

5 (R)

shown The

the

in

The to

receptor

and

those

group.

The

of

of

(23)

(19) , allyl

activity

8. seem

activity.

regions

small

than

on

and

conformers

increase

m-oxymethylene

occupy

regions

substituents

cyclic

to the

are

activity

of the N l - p h e n y l

unfavorable

(26) ,

Nl-methyl

Nl-propyl

substituents

positions

26)

7,

volume the

the

substituents

The

the

6,

between

reducing

substituent

lower

Figs.

The a c t i v e

and

occupied

occur

(27)

(22)

1 compounds.

phenyl

the

o-methyl

p-chloro

and N l - b e n z y l

in

in

Nl-substituent

that

superposed,

shown

representing end of the

1 and 2 compounds.

were

as

in

plane

the

of

the p r o p e r

the

binding

of the q u i n o l o n e

ring

s h o u l d reduce the activity. The

difference

substituents together the

is

shown

These

at

the

regions

activities

para

In

of the

region

by

occupied compounds.

I0 in

shows the

The

Nl-phenyl

(33)

relevant

binding

is

of

to

Nl-phenyl

cause

are

is too

meta

fluorine

causes

seem

work

31 in class

difference 4

compound

region

occupied

probably the

I, 2, and 3 compounds.

more receptor

between and

the

by the

reduction

the

1 and the

and the of

the in

3. the

total

class

I,

p-methoxy the

in

simultaneously

significantly than

(32) .

(31) ,

small,

by

and the

of class

derivative

factors

the

NI-

occupied

(30)

reductions

to those

position to

by

2 compounds

substituents

further

Nl-(m-fluorophenyl)

of c o m p o u n d

with

1 and

regions

30 and 32 r e l a t i v e

the

class new

occupied

class

Additional

at the para

two

the a c t i v i t y

Fig.

class

thought

the

These

9.

volumes

and the

of the N l - d i m e t h y l a m i n o e t h y l

position

hydrogen

occupied

activity. lowering

in Fig.

groups

were

the

3 compounds

of c o m p o u n d s

compounds.

volume

between

class

two N - m e t h y l

bromine

2

of

volumes 2,

group

and

3

of the

unfavorable

Nl-substituents

for of

109

least

Finally,

the

active

derivative

(34) ,

examined methyl

(Fig.

inhibitory

below

substituents that

of

of

occupied

the

The the

i,

region

total

volumes

of

the

compounds

was

N I- (2, 6 - d i m e t h y l p h e n y l )

2,

3,

and

occupied

quinolone

ring

4

by

one

seemed

provide

to

for

by

of

the

exert

ortho

a marked

the

the

Nl-phenyl

We

of the q u i n o l o n e above

the

group

to the

of the Nl-phenyl.

We

into

one

above

fluorine also

that

NI-

there

corresponds

the

and

propose

the

for

propose

activity"

Nl-cyclopropyl other

insights

relationships

antibacterials.

increasing

and the

the plane

important

structure-activity

quinolone

ring,

position

below

class

analyses

regions

quinolone para

the

the

the

effect upon the activity. present

two

between

compound,

and

three-dimensional are

5

Ii) .

groups

The

difference

class

plane

of

hydroxyl that

to

the

at the

the

regions

ring and a r o u n d the m e t a p o s i t i o n

quinolone

ring

plane

prevent

proper

r e c e p t o r binding. Fig.

12

shows

a modified

volume

occupied

of

Nl-(p-hydroxy)-phenyl.

the

(length) QSAR

is best

equation

phenyl

toward

at

4.2

For

receptor allyl,

activity

as

used

optimum

they w o u l d

and

not

para p o s i t i o n

methyl further

Fig.

12

derive

volume reach

to

has

the

other

of the Nl-phenyl

as

the

the

of

NI-

has

cyclopropyl L

of

L

in

an and

compounds

value

changes

in these

compounds

in terms

receptor is

too

of the

The

group,

the

to

volumes

onto the L to

fit

that

the

but

n-propyl,

the

forbidden

region.

The

favorable

for

Nl-substituents

region

in

dimethylaminoethyl

forbidden

could

to

model

small

extrude

[i]

group.

[i]

group

does,

optimum

optimum

the

situation

groups

above.

why

The

group

equation

of

activity

cyclopropyl

corresponding

the

the

of the

into

total

activity

equation

length

value.

two

explain

the

in

that

group

benzyl

the

hydroxy

for N l - S u s t i t u e n t s

decreases

methyl

on

the

can

predict

variations

described

to

to

shows

projection a

penetrate in

to

based

and

model

parameter

optimum

activity

terminal

model

compounds

[i]

corresponding

The

substituent

L1

the

example,

wall

receptor

the

of

hydroxyethyl

region.

in

the

unable

substituents

side

one-dimensional

axis.

is The

model

group

parameter

corresponding

Nl-phenyl

explain

This

a steric

but

i

receptor

Nl-cyclopropyl

Equation

either

of the

the

as

[i]

groups.

without fact,

the

derivatives.

optimum ethyl

by

be of

of

accommodated cyclopropyl,

corresponding

the only but

to the

110

Fig. I . Stereoview of the superposition of the proposed active conformers of 2 0 (orange), 23 (green), 25 (blue), and 2 7 (yellow) and the difference(0range) between the total volumes of the set of 20, 2 3 , 2 5 , and 2 7 and those of the class 1 compounds. Since the benzene rings of the N1-substituents of 2 5 and 2 7 overlap, this region appears white. The N1-methyl of 23 and N1-ally1 of 20 also overlap and the N1-methyl appears white or yellowish-green (Reproduced from ref. 6b by permission of the American Chemical Society).

Fig. 8 . Stereoview of the superposition of the proposed active conformers of 21 (green), 26 (yellow), 28 (violet), and 2 9 (blue) and the difference (orange) between the total volumes of the set of 21, 2 6 , 28, and 29 and class 1 compounds (Reproduced from ref.6b by permission of the American Chemical Society).

Fig. 9. Stereoview of the superposition of the proposed active conformers of 30 (green), 31 (yellow), and 32 (cyan) and the difference between the total volumes (orange) of 30, 31, and 32 and class 1 and class 2 compounds (Reproduced from ref.6b by permission of the American Chemical Society).

111

~)__...... ~

~

0

.I~N~ ~

0

Fig. I0. S t e r e o v i e w of the p r o p o s e d active c o n f o r m e r of 33 and the d i f f e r e n c e b e t w e e n the v o l u m e s (orange) of c o m p o u n d 33 and class I, 2, and 3 c o m p o u n d s (Reproduced from ref. 6b by p e r m i s s i o n of the A m e r i c a n Chemical Society).

Fig. Ii. S t e r e o v i e w of the p r o p o s e d active c o n f o r m e r of c o m p o u n d 34 and the d i f f e r e n c e b e t w e e n volume (orange) of c o m p o u n d 34 and the total of class i, 2, 3, and 4 c o m p o u n d s (Reproduced from ref. 6b by p e r m i s s i o n of the A m e r i c a n Chemical Society).

Fig. 12. S t e r e o v i e w of the m o d i f i e d r e c e p t o r model for the volume occupied by Nl-substituents of quinolone antibacterials. ( R e p r o d u c e d from ref. 6b by p e r m i s s i o n of the A m e r i c a n C h e m i c a l Society) .

112

3. ANALYSIS OF THE STERIC EFFECT OF 6-SUBSTITUENTS

(18)

According to equation [l], formulated for the entire series of quinolones 3, the effect of substituents at the 6-position on the activity is represented by the Taft-Kutter-Hansch Es Equation [l] reflects equation [2] for the subset of

parameter.

6-monosubstituted compounds 36, the activity of which varies parabolically with the Es of the R g substituent (Fig. 13A) (2). log(l/MIC)

= -3.318(+0.59) ( E S ~ -4.371(?0.85) ) ~ Es6 +3.924 n=8 S = 0 . 1 0 8 r=0.989 F=112.29

[21

In equation [21, the Es value adopted for the nitro group is the one

(-1.01) evaluated from its half-thickness representing the

steric effect in the perpendicular direction and that of methoxy is approximated by the value of the ethyl group

-2

0

-1

1

-2

-1

E s6 Fig. 13

(2).

6 ES

For the

0

Parabolic relationships for the effect of

6-substituents with the Es6 Parameter.

1

113 corresponding

use

of

the

reasonable for

7-piperazinyl

same

E s value

(Fig.

its

13B) .

coplaner

significant significant

correlation

set

of

compounds

for

the

the

greatest

deviation,

observed

calculated

value

changes

of

the

vicinal

is

piperazinyl

relationship

between

0.61) .

log(I/MIC)

the

This

from the

not

For

the

(Es6)2

+1.426(+0.29) s:0.250

6-nitro-7-

for the

Although

being

be

due

value

to

and

conformational

confirmation

of

with

was

the

this,

analyzed

and a c t i v i t y

higher

observed

interaction

36 and 37 was

showed

much

the

a

combined

compound

between

(Es6) 2 -2.682(+0.93) r=0. 984

No

(half-thickness)

steric

conformation

- -2.587 (+0.89)

the

obtained

activity

the

not

half-width

[3].

Es

the

effective.

unless

could

by

R 6 in c o m p o u n d s

s=0.079

n:15

[4] was

however,

apparently

equation

difference

group

:-2.026(+0.68) n=6

give

predicted

(the

the

also

is

6-nitro-7-piperazinyl

group.

of the

to

using

its

6-nitro

conformation

Iog(I/MIC)

37)

value

estimated is

37,

group

formulated

equation the

compounds

6-nitro

E s value

omitted

group,

of

the

was

is

(36 and

6-nitro

subset

effect

correlation compound

the

The

steric

piperazinyl

than

for

the

and

the

examined. [3]

Es 6 +5.561

F-45.50 -3.351 (+1.25)

Es 6

[4]

17 +4. 088 r=0.971

F=60.84

3.1 C o n f o r m a t i o n and Steric Parameters 3.1.1 C o n f o r m a t i o n a l receptor similar

mappings

As R 6 : N O 2)

ring

Thus,

analysis The

should

the

36 for

and

37

were

carried

l-substituents

6-nitro

almost

hand,

group

plane

6-substituents,

14, is

other

6-substituents

substituent. of

Fig.

nitro

used

analyses

and out

in 3.

by

The AM1

for the MO method.

energy

the

the

Conformational

compounds

to those

used in

low

On of

quinolone some

was

shown at

plane. angle

of

procedures

Hamiltonian

Analysis:

in

of

is about could

be

the

steric

low

compound 55 ~ .

of

with (37:

(36: ring

conformation,

the the

the

on

with

conformations by

of the

based

38

quinolone

R6=N02),

influenced

analysis

parameter

compound

the

energy 39

Likewise,

markedly

for q u a n t i t a t i v e a

group

coplaner

the

steric

effect

conformational

be used.

6-methoxy

compound

40

(36:

R6=OCH 3)

has

of

adjacent

two

conformers

114 with

energy

group

is

moiety shown of

minima.

almost

of

the

methoxy

in Fig.

the

14.

methoxy

m e t h o x y group

Fig. 14. right) , 39

The

One

corresponds

coplaner The

group

locates

with

group

locates is that

is

opposite

conformer

was

the

methylthio

conformation

at

group

is

as

the

and

the

the

and

methoxy

the

5-position only

methyl

the

methyl side

as

direction

moiety

of

the

side.

c o n f o r m a t i o n of c o m p o u n d s 38 ( u p p e r (lower right), and 41 (lower left) .

6-methylthio-7-piperazinyl upward,

in w h i c h plane

in w h i c h

at the 7 - p o s i t i o n

the

turning

to that

quinolone

other

Proposed active (upper left), 40

first

the

taken

almost

modeled

as

the

active

compound

41

coplaner

with

in

Fig.

(37"

structure

R6=SMe)

the

14,

since

in

which

quinolone

has

lower

plane energy

than the other.

3.1.2 Quantitative Structure-Activity Relationship using Conformational StericParameters: conformations calculated. sphere the of

Each

with

the

quinolone the

along

New

steric p a r a m e t e r s

for the atom

van

ring

Waals

plane.

projection

substituents

in the

der

6-substituent

the

6-substituents

from of

The the

the

based

on the p r o p o s e d

of q u i n o l o n e s

6-substituent radius. length carbon

bond

The

P.

represented

plane

L is the atom

between

and the C6 onto the plane

was

at

P

is

farthest the

the

active

36 and 37 were a as

extension

6-position

(~ a t o m

as

defined

of

the

A box w h i c h t o u c h e s

(C6) 6-

the

115 van

der

Waals

through 15.

The

sides H2

the

values

of

are

the

the

tangentially

was

widths

defined

of

the

as

from

and

shown

passes in

substituent

respectively.

substituent

compound

H26

The

The the

that

reliable

36. [7]. of

n=8

situations

the

activity

the

6-NO2

plane

for

With

[6]

in

the

are

the

H 1 and

P

and

are

WI,

W2,

HI,

mainly

due

works

well

[5]

the

and

the

the

to

fact

[6] of

new

by

steric

variable,

quality

[4],

than

the the

the

steric

of

Figs.

16A of

the

correlation parameter

about

for

the

is gives

nature

at

give lower more of the

of R 6.

: -5.806(+2.67) s=0.255

r=0.937

(H26) 2 +17.67 (+8.42)

H26

-8.235

[S]

F:18.08

P

P

COOH

H1

H2 H2 --

15.

(half-

[2] to

R6

from

substitution

combined

R6 and

the

Es

equation

for

the

thickness

that

in

I7,

H2

effect

the

were

Es p a r a m e t e r

statistically

selected.

illustrated

with

L,

the

parameter

is

group

and

were

the

that

indicator

Although equation

37

show

accord

equations

information

effect

log(i/MIC)

Fig.

Fig.

to

values

plane

parameters,

36 and

represents

and

on

steric

respectively,

[5]

7-position,

equation

these

compounds

[6],

quinolone

thickness)

with for

[5] and

substituents

steric

are

the

equations,

Equations

than

of

examined

substituents.

the

W2

5-positions,

correlations

equations In

the

and

6-substituent 6-position

as H 2 _> H I.

The

16B.

the

the

W 1 and

7-

H2 w e r e

best

of at

thicknesses

defined and

radii

carbon

Definition

of the

new

steric

parameters.

H1

116 n

R6D3 OoH

7.0-

36

31

-

B

A

6.5-

4

I

GH5

6.0-

6.0

5.5-

5.5

\

a, 4

5.0

m

4.5

0

4

1 .o

1.5

2.0

1 .o

2.5

1.5

2.0

2.5

H2 Fig. 16. Parabolic relationships for t h e effect of 6 - s u b s t i t u e n t s w i t h t h e newly d e f i n e d H

l/MIC)

=

l/MIC)

s=O.211 =

parameters

( H z ~ )+ 6~ . 9 5 9 ( + 6 . 5 2 ) H26 + 0 . 8 8 6

-2.222(+1.99)

n=7

2

r=0.863

-3.427(+1.86)

[61

F=5.81

( H 2 6 ) 2 t 1 0 . 5 7 2 ( + 6 . 0 0 ) H26

+ l .7 0 5 (+O . 3 9 ) I 7 - 3 . 2 8 8

s=O . 3 3 1

n=15

r=O . 9 4 9

[71 F=33.07

V a r i o u s t y p e s of s t e r i c p a r a m e t e r s e t s h a v e b e e n e m p l o y e d f o r

QSAR

analyses.

Although

various

parameter

s u c c e s s f u l l y u s e d d e p e n d i n g upon t h e t y p e o f

sets

have

been

steric i n t e r a c t i o n s

i n v o l v e d , t h e y sometimes d o n o t r e f l e c t t h e s i t u a t i o n based o n t h e biologically

a c t i v e form.

T h e new

s t e r i c parameters

proposed

a b o v e i n a way s i m i l a r t o t h e STERIMOL v a l u e s seem t o be v e r s a t i l e i n o t h e r examples, conformation

from

s i n c e t h e y are b a s e d on t h e p r o p o s e d " a c t i v e " conformational

m a n i p u l a t e d on t h e c o m p u te r g r a p h i c s .

analysis

and

appropriately

1 I7 TABLE 3 . S t r u c t u r e a n d A c t i v i t y o f q u i n o l o n e s a n d fluoroquinolones having 8-substituent.

n

d b - " " O H I RR

log 1/MIC

(mole/l) a g a i n s t E . c o l i

-1

obsd.

'ZH5

calcda)

dif.

43')

H

3.939

4.489

-0.55

44')

F

4.575

4.586

-0.01

4SC)

c1

4.606

4.449

0.16

Me

4.868

4.818

0.05

4 7 ')

OMe

3.694

3.881

-0.19

48')

Et

3.088

3.149

-0.06

2.514

2.386

46

C)

4 gC) OEt

l o g 1/MIC RNJ

R8

R8

1

R

obsd.

ref

0.13

(mole/l) a g a i n s t E . c o l i

R,

R1

b)

calcd?)

dip) calcd?)

difb) ref

C)

H

Et

H

6.629

6.375

0.25

2

50c)

F

Et

H

6.873

6.564

0.31

2

51c)

c1

Et

H

6.892

6.801

0.09

2

H

7.184

7.007d) 0 . 1 8

5.581e)1.60

2

Me

6.859

6.69gf) 0.16

5.798')1.06 h) 6.880 -0.04

2,7

5 2 ')-CH2 5

CH2CH (CH3)

-0CH2 CH (CH3) -

-

53

OMe

Et

H

6.844

5.759

1.08

54

Br

Et

H

6.600

6.746

-0.15

55

CN

Et

H

6.236

6.506

-0.27

56

NO2

Et

H

5.970

6.154

-0.18

20 21 21

i) 6.532 -0.56

C a l c u l a t e d by e q u a t i o n [l] . D i f f e r e n c e between observed and c a l c u l a t e d v a l u e s . I n c l u d e d t o d e r i v e e q u a t i o n [l]. C a l c u l a t e d w i t h B 1 of t h e e t h y l g r o u p i n p l a c e o f B 4 ( 2 b ) C a l c u l a t e d u s i n g B 4 of t h e e t h y l g r o u p f o r B 4 8 . C a l c u l a t e d u s i n g B 1 of t h e 8 - m e t h o x y g r o u p i n p l a c e o f B 4 C a l c u l a t e d u s i n g B 4 o f t h e methoxy g r o u p f o r B 4 8 . C a l c u l a t e d u s i n g B 2 o f t h e methoxy g r o u p f o r B 4 8 . C a l c u l a t e d u s i n g B1 of t h e n i t r o g r o u p i n p l a c e o f B 4 .

21

118

3.2 Proposed Receptor Model The

active

conformers

of

norfloxacin

droxacin,

tioxacin,

and DJ-6783

of

quinolone

rings.

their

substituents 17.

The

positions receptor these

and

total

should

compounds

HN~ . J

helpful

>--S

to the

are very including

volume

for

active

at

of the

against

E.

oxygen,

as

the

estimating

vicinity

fluorine,

occupied

calculated

compounds

oxolinic

acid,

by m a t c h i n g by

shown 5-,

the

6-,

shape

6-position,

coli

with

atoms the and of

a variety

C2H5

C2H5

C2H5

O

i

C2H 5

tioxacin

Fig. 17. Active volume (cyan) of quinolone antibacterials.

acid

the of

and nitrogen. O

oxolinic

7-

because

O

(I)

6-

in Fig.

O

norfloxacin

O

be

was

these

(1),

superposed

total

groups

of

corresponding

6-substituents

The

adjacent volumes

were

droxacin

O

I

C2H5

DJ-6783

of the

6-substituents

and vicinity

119

4. ANALYSIS OF THE STERIC EFFECT OF 8-SUBSTITUENTS

( 19)

The activity (MIC) of 8-substituted quinolones 4 2 ( 4 3 - 4 9 in Table 3 ) has previously been reported as being parabolically related with B48, one of the STERIMOL parameters for the maximum width of the Re as indicated by equation [81 and Fig. 18 (2). The B4 value as the steric parameter of Re substituents also applies to 1, 6, 7, 8-tetra-substituted quinolones 3 ( 5 0 - 5 2 in Table 3) since

the activity

of these

equation

(2).

111

compounds has been

The 8-substituent

well

predicted by

is thought to interact

sterically with the 1-ethyl-substituent in compound 4 2 . Therefore, the maximum width of the Re expressed by B4 has been believed to be that in the direction opposite to the 1-substituent ( R 7 side) and to recognize the receptor wall as such. Depending upon the structure, however, the 8-substituent may be directed above or below the quinolone ring plane with steric repulsions of substituents at positions 1 and 7. log(l/MIC)

=

-1.016(*0.46) (B48)2 +3.726(+2.04) B48 +1.301

n=7

s=O.221

r=0.978

F=44.05

Me

1 .o

2.0

3.0

B48 Fig. 18 Parabolic relationship for the effect of 8-substituents with the STERIMOL B4 parameter.

181

120

Fig. 19. (pink), 48

S t e r e o v i e w of the p r o p o s e d (green), and 49 (blue) .

active

conformers

of

47

proposed

active

conformers

of

53

Stereoview of the a c t i v e of q u i n o l o n e a n t i b a c t e r i a l s .

volume

model

the

8-

Fig. 20. S t e r e o v i e w of (yellow) and 56 (green).

Fig. 21. substituent

the

of

121 Since structure and

equation 3,

50-52

in

quinolones the

[I]

including Table

3,

(5 and 5 3 - 5 6

activities

of

was

formulated

8-substituted some

symmetrical

top

8-substituents

a ring

the

l-substituent

[i],

with

that

group

was

of c o m p o u n d

53

not.

may

conformations

This of

the

were

be

due

by the

We

compounds

with 44-49

some

reported.

Although

spherical

5 with

the

well

by the

or

R 8 forming equation

unsymmetrical

methoxy

differences

between

the

1,8-disubstituted

(47-49)

and

compounds

8-substituents.

quinolones

as

having

predicted

to in

been 55

compound

substituted

8-substituents

8-substituents

and

and

71

such

i, 6, 7, 8 - t e t r a - s u b s t i t u t e d

3) have 54

i, 6, 7, 8 - t e t r a - s u b s t i t u t e d unsymmetrical

new

in Table

compounds

for

ones

such

analyzed

as

the

53

having

conformations

of

to examine this p o s s i b i l i t y .

4.1 Active Conformation and Activity The

compounds

analysis

of

described

above

As ethoxy

the

analyzed

8-substituted

(2.2.1)

shown

in

direction

the

1-ethyl

e),t.',.'.,

19,

opposite

k. " " " . '




( E q . 12)

Obsd.

0. 00 0. 55 0. 96 1. 19 0. 47 0.95 0. 99 1. 54 2 10 1. 43 0 . 25 0. 03 0. 54 0. 46 1. 19 -0.24 0. 21 0. 23

Calcd.

Dev.

0. 01 0. 52 1. 06 1. 19 0. 51 1. 08 0. 92 1. 43 2 12 1. 40 0 . 23 - 0 . 04 0.57 0.53 1. 13 -0.35"' - 0 . 26"' - 0 . 03"'

-0.01 0. 03 -0. 10 0. 00 -0.04 -0. 13 0. 07 0. 1 1 -0.02 0. 03 0. 02 0. 07 -0.03 -0.07 0. 06 0.11 0. 47 0. 26

Calcd.

0.02 0. 49 1.00 1. 12 0.49 1. 03 0. 97 1. 46 2. 1 1 1. 40 0. 20 0. 00 0. 58 0.55 1. 25 -0.17 0. 04 0. 32

Dev.

-0.02 0. 06 -0.04 0. 07 -0.02 -0. 08 0 . 02 0. 08 -0.01 0. 03 0. 05 0. 03 -0.04 -0.09 -0. 06 -0.07 0. 17 -0. 09

a) Not included in the c o r r e l a t i o n b u t c a l c u l a t e d by Eq. 6.

TABLE 5

Hydrophobicity Parameters of 2-Substituted Pyrimidines(I1) ~~

nZPM

0. 0 0

0. 46 0. 80 0. 94 0 . 39 0. 67 1. 18 1. 45 0. 13 -0.27 0. 13 1. 51 -0. 76 0 . 24 -0.37

0 . 03

0. 32 0 . 82 0. 94 0. 44 0 . 69 1. 17 1. 18"' 0 . 17 -0.03"' 0 . 51"' 0.90"' -0.32"' - 0 . 31"' -0.10"'

Dev.

-0.03 0. 14 -0. 02 0. 00 -0.05 -0.02 0. 01 0. 27 -0.04 -0. 24 -0.38 0. 61 -0.44 0. 55 -0.27

a) Not i n c l u d e d i n the c o r r e l a t l o n b u t c a l c u l a t e d by Eq. I .

162

TABLE 6 Hydrophobicity P a r a m e t e r s of 4 - S u b s t i t u t e d PyrimidinesCIII) z q p M Substltuent

Calcd.

H C1 Me OMe

OEt CN COzMe NMez CONHz NH2 NHAc

(Eq. 1 3 )

(Eq. 8 ) Obsd.

0. 0 0 0. 9 1 0. 39 0. 98 1. 4 1 0. 36 0. 17 1. 0 2 -0.24 0. 1 9 0. 47

Dev.

-0. -0. -0. 0.

0. 06 1. 0 4 0. 5 1 0. 81 1. 3 8 0. 36 0. 08 1. 0 2 -0.25"' - 0 . 42"' - 0 . 11"'

0.

0. 0. 0. 0. 0. 0.

06 13 12 17 03 00 09 00 01 61 58

Calcd.

Dev.

-0. 03 0. 01 -0.05 0 . 11 0. 03 0. 0 8 0. 0 6 -0. 16 -0.25 0. 1 3 0. 04

0. 03 0. 90 0.44 0. 8 7 1 . 38 0. 28 0 . 11 1. 18 0. 0 1 0. 06 0. 43

a) Not i n c l u d e d in t h e c o r r e l a t i o n b u t c a l c u l a t e d b y Eq. 8.

TABLE 7 Hydrophobicity P a r a m e t e r s of 5-Substituted Pyrimidines(IV)

n g p M < l o g P ( p y r i m i d i n e ) = -0. 44> (Eq. 9)

Substituent

Calcd. H

F C1

Br Me OMe

OEt C02Me COzEt NMez CONHz

NHAc

0. 0 0 0. 4 1 0. 9 1 1. 10 0. 4 5 0 . 51 1. 0 0 0.47 0. 96 0. 90 -0.48 0.22

0. 08

Dev.

-0. 08 0. 05 0 . 9 4 -0. 03 1.19 -0.09 0 . 6 1 -0. 16 0. 47 0. 04 0. 8 9 0 . 11 0 . 41 0. 06 0. 9 6 0. 0 0 0. 8 0 0. 10 - 0 . 78"' 0 . 3 0 -0.09"' 0.31 0. 3 6

(Eq. 14)

Calcd.

0. 03 0. 37 0. 9 1 1. 1 4 0.50 0.53 0. 9 1 0. 4 7 0.98 0. 8 6 -0.52 0. 28

(Eq. 17)

Dev.

-0. 0. 0. -0. -0. -0. 0. 0.

-0.

0. 0.

-0.

03 04 00 04 05 02 09 00 02 04 04 06

Calcd.

Dev.

0.10 0.42 0. 97 1.11 0. 5 9 0. 47 0 . 81 0. 45 0. 9 3 0. 78 -0.58 0 . 40

-0.10 -0. 01 -0. 06 -0.01 -0. 14 0. 04 0. 1 9 0. 02 0. 0 3 0. 12 0. 10 - 0 . 18

a) Not included in t h e c o r r e l a t i o n but c a l c u l a t e d b y Eq. 9.

163 TABLE 8 Hydrophobicity Parameters of 3-Substituted Pyridazines(V1 z g p D

0.09 1. 0 1 0.47 0.83 1. 2 3 0. 12 0. 9 5 -0.14"'

Dev. -0.09 - 0 . 18 -0. 09 -0. 02 0 . 13 0 . 18 0. 07 0. 14

Calcd. 0.12 1. 0 1 0.49 0.84 1. 2 1 0. 17 0. 94 -0. 08

Dev. -0.12 - 0 . 18 -0.11 -0. 03 0 . 15 0. 1 3 0. 08 0. 08

~

a) Not included in the correlation but calculated by Eq. 10.

TABLE 9 Hydrophobicity Parameters of 4-Substituted Pyridazines(V1). R ~ P D

(Eq. 16)

Calcd. 0. 05 0.42 0.46 0.19 0. 42 0. 81 0. 70 0. 28 0. 08 0.43

Dev. -0. 05

-0. 0 1 -0. 04 -0.09 0. 04 0 . 15 -0. 06 0. 05 0. 12 -0.11

a) Not included in the correlation but calculated by Eq. 11

164 t w o t y p e s of s u b s t i t u t e d p y r i d i n e s s h a r i n g a common s u b s t i t u e n t e x c e p t f o r

F o r such 2 - s u b s t i t u t e d d i a z i n e s as 2PR(I). 4PM

t h e symmetric 2PM and 5PM

(111). and 3PD ( V ) , t h e c o r r e l a t i o n with

the alternative ( n3

or

p

~

A ~ P Y )

KZPY

w a s always b e t t e r t h a n t h a t with

a s shown i n Table 10

In p a r t i c u l a r . f o r

t h e 2PR (I) system, t h e s i n g l e c o r r e l a t i o n with n p P y is much b e t t e r t h a n t h a t with

i n c l u d i n g s u c h o u t l i e r s as OR, SMe, A c . COzR, CONHz. NH2, and N M e z

K 3py,

in Eqs

2 and 3 f o r t h e 2PY system

This i n d i c a t e s t h a t t h e

v a l u e of 2-

A

s u b s t i t u e n t s i n d i a z i n e s u l t i m a t e l y s h a r e s components from complex proximity i n t e r a c t i o n s with

z

K

p

~

The a b o v e o b s e r v a t i o n s l e d u s t o u s e t h e

Then, Eq. 1 c a n b e r e w r i t t e n as

series as t h e r e f e r e n c e independent v a r i a b l e Eq

value f o r t h e pyridine

A

4 c o n s i d e r i n g t h e backward e f f e c t of t h e ring-N by t h e u P ( p ,

Z(ortho)-X-substituents

A s a matter of f a c t , t h e u s e of

0 )

p x

~

o n hydrogen-bondable

term

~

as t h e r /e f e r e n c e means ~ t h a t ~t h e

~

i n t h e r e f e r e n c e p y r i d i n e system is r e g a r d e d as a

s u b s t i t u e n t X p l u s -N=

" s i n g l e " u n i t (N+X), and t h a t t h i s (N+X) u n i t and t h e second -N= function(Y) are bidirectionally interacting partners.

X,

u

x",

but

not

that

of

(N+X)

In Eq. 4, however, t h e forward e f f e c t of

to

is assumed

work

on

the

second

-N=

function(Y1 and t h e backward e f f e c t of Y o n l y o n X is c o n s i d e r e d depending upon t h e r e l a t i v e p o s i t i o n s of X and Y.

TABLE 10 S i n g l e Correlation C o e f f i c i e n t (r) between DiazineA Z P R ( I )

nppy ~ 3 p y

0.954 0. 814

0 . 940

-

-

0. 678 16

0. 680 15

4PY

~ p h x

ne'

nZPM(I1) -

n4PM(III)

x

a n d Reference x

r5PM(IV)

n3PD(V)

-

0. 923 0. 785

0. 881 -

0 . 956 -

0. 783 0 . 589 10

-

0. 840 12

0 . 556 8

Values.

n4pD(vI) -

0. 8 6 5 0. 8 8 5 0 . 710 10

a) Number of s u b s t i t u e n t s whose n v a l u e s are a v a i l a b l e i n common among m o n o s u b s t i t u t e d b e n z e n e and p y r i d i n e systems.

The a n a l y s i s was made n

~

for

each series

For

t h e reference

v a l u e , t h~e c h o i c e ~ was made a~c c o r d i n g t~o which is~ b e t t e r asd t h e

parameter set i n t h e f i n a l c o r r e l a t i o n s

4PM a n d 3PD). 4PD series.

(I-VI)

KZPY

For

a -substituted

w a s , of c o u r s e . t h e parameter of c h o i c e

Z ~ P Yand

A ~ P Y ,

r e s p e c t i v e l y , were s e l e c t e d

d i a z i n e s (2PR. 2PM, F o r t h e 5PM a n d

In f a c t , t h e select-

~

e d s e t of pyridine-Ir

v a l u e s was always b e t t e r c o r r e l a t e d with t h e set of

corresponding diazine

(

v a l u e s even i n t h e s i n g l e c o r r e l a t i o n

K

For non-hydrogen bonding s u b s t i t u e n t s where

p

=

0.

T a b l e 10

)

and some hydro-

g e n a c c e p t i n g s u b s t i t ~ i e n t swhere p x is n o t v e r y l a r g e , t h e "backward" e f f e c t (a

pp

of Y o n X is n o t v e r y s i g n i f i c a n t

x)

e a c h d i a z i n e series. Eq

F o r s u c h a s u b s t i t u e n t set In

4 c a n b e simplified t o E q

5

Thus, w e f i r s t used Eq. 5 f o r e a c h d i a z i n e series e x c l u d i n g t h e amphiprotic

NHz

substituents, NHAc.

examined t h e u s e of a s u b s t i t u e n t series the

u

and CONHz. possessing

large

values

p

4, i n c l u d i n g amphiprotic s u b s t i t u e n t s

p r o c e e d e d t o u s e Eq

I

and a

parameters i n p l a c e of

a

term was

we

in E q

5 f o r each

A s i n t h e case f o r s u b s t i t u e n t s i n t h e p y r i d i n e series.

p a r a m e t e r g a v e almost e q u i v a l e n t statistics with

I

Then,

W e preliminarily

insignificant

The

coefficient

of

the

x

a

but

Dyrldlne

term.

the

a

R

however.

d e v i a t e d from u n i t y s i g n i f i c a n t l y f o r some series In T a b l e 11. t h e c o r r e l a t i o n e q u a t i o n s (Eqs 6-11) formulated with u s e of

E q 5 are shown

values according t o t h e s e

In T a b l e s 4-9, t h e c a l c u l a t e d x

e q u a t i o n s are l i s t e d f o r comparison with t h e o b s e r v e d values. series. N M e z ,

SMe and

From t h e 2PM

COzR had t o b e d e l e t e d , i n a d d i t i o n t o t h e amphiprotic

s u b s t i t u e n t s , t o o b t a i n t h e a c c e p t a b l e q u a l i t y of t h e c o r r e l a t i o n t h e number of

Although

d a t a r e l a t i v e t o t h a t of t h e independent v a r i a b l e terms w a s

n o t s u f f i c i e n t i n some series, E q

5 seems t o h o l d as f a r as

~r

v a l u e s within

e a c h system are c o n c e r n e d A s e x p e c t e d , t h e c o e f f i c i e n t of t h e p y r i d i n e - x

t e r m was close t o u n i t y

b e i n g c o v e r e d almost completely i n t h e r a n g e of t h e 95% confidence i n t e r v a l i n Eqs

6-11

f u n c t i o n (-N=),

The p y v a l u e , t h e s u s c e p t i b i l i t y c o n s t a n t of t h e s e c o n d aza v a r i e d , however, depending on t h e system

j u s t i f i e d o n l y a t t h e 92% l e v e l i n E q

Although it w a s

10 f o r t h e 3PD system, t h e

p y

value

c o u l d b e c a t e g o r i z e d i n t o two g r o u p s : o n e g r o u p l o c a t e d a r o u n d 0 8 7 (Systems I, 111, a n d V), and t h e o t h e r a t a b o u t 0 5 3 (Systems 11, IV. and VI)

I. 111, a n d V. t h e r e were common s t r u c t u r a l f e a t u r e s s u c h t h a t as t h e r e f e r e n c e

In systems

I I ~ P Y

was used

For systems IV and VI where x B P y and n I P y were used,

t h e p y v a l u e was lower t h e reference, t h e p

Although system I1 is one of t h o s e i n which 2PY is

v a l u e w a s lower i n E q 7

t h i s is a v a i l a b l e a t p r e s e n t

much h i g h e r t h a n t h e c o r r e s p o n d i n g s t i t u t e d p y r i d i n e system

No a d e q u a t e e x p l a n a t i o n f o r

I t w a s a l s o noted t h a t t h e s e P Y

values in E q s

p y

v a l u e s are

2 and 3 f o r t h e sub-

The v a r i a t i o n s i n t h e p y v a l u e s with d i f f e r e n t

systems i n d i c a t e t h a t t h e assumptions made t o f o r m u l a t e Eqs

4 and 5 a p p l y

TABLE 11 Correlations of z ( d i a z i n e ) w i t h E q . 5 ” ’

I : 2PR

1I:ZPM

System I I1

III:4PM

IV:5PM

r

Correlation If

ZPR

=

1. 1 9 3 z ~

P

Y+

(0.0 9 3 ) 7C 2 P M

=

1. 0 4 8 n

ZPY

(0. 195)

I11

Z ~ P M =

IV

*SPM

V

~

VI

Z ~ P D=

0 . 8 2 6 u x” ( m )

(0. 241)

*

0 . 555U ; ( p ) (0. 303)

1. 2 4 5 x 2 ~+ ~0 . 8 7 2 u E ( p ) (0.544)

(0. 345) =

1. 0 0 5 n s p y

(0.2 4 1 ) S

P

D=

+

0. 4 7 5 u E ( m ) (0. 398)

0 . 9 5 3 z z ~+ ~0 . 8 8 3 u f ( m ) (0. 385) (1. 055)

0 . 9 9 4 ~ 4 p y+ 0 . 5 7 3 u f ( m ) (0. 331)

(0. 245)

VI : 4PD

V : 3PD

S

F

n”’

E q . No

+

0.014

0.994

0. 073

456. 0

15

6

+

0. 033

0. 988

0. 072

105. 8

8

7

0. 0 5 5 (0. 214)

0.979

0.119

57. 2

8

8

0.081 (0. 158)

0. 9 6 9

0.100

53. 0

10

9

0. 087 (0. 312)

0. 961

0. 1 6 0

23. 8

7

10

0.020 (0. 144)

0.985

0 . 069

63. 6

7

11

(0. 097) (0. 131)

+

+

+

-

a) Reproduced from r e f . 13 with permission o f t h e copyright owner, VCH Publishers, Inc. b) S u b s t i t u e n t s n o t included: I: CONHZ. NHz. NHAc. 11: SMe. COzR. NMez. CONHZ. NHz. NHAc. 111: CONH2, NHz. NHAc. IV: CONHZ. N H A c . NHz(not measured). V: CONHz.NHz(not measured),NHAc(not measured). VI: CONHZ, NHz. NHAc.

167 Eq. I a l s o differed from t h e o t h e r s in t h a t it w a s

only within each series.

unable t o accommodate w e l l non-amphiprotic

s u b s t i t u e n t s such a s NMez, SMe

and COZR. Table

Eq. 4 for

12 lists, t h e c o r r e l a t i o n substituents

including

Tables 4 and 6-9 a l s o show t h e x

equations Eqs.

12-16 formulated

amphiprotic ones having higher

P X

with

values.

values calculated according t o E q s . 12-16.

For t h e 2PM system, no reasonable c o r r e l a t i o n equation was derived.

In

addition t o o u t l i e r s u b s t i t u e n t s from Eq. I , no amphiprotic s u b s t i t u e n t s w e r e accommodated and so t h e c o r r e l a t i o n with t h e bidirectlonal procedure w a s n o t formulated f o r t h i s series.

For t h e 2PR. 4PM. and 5PM systems, however, NHz. w e l l accommodated in Eqs. 12, 13, and 14.

CONHz.

and N H A c were

Again, t h e number of d a t a r e l a t i v e

t o t h a t of t h e independent variable terms w a s n o t enough in some systems. Since t h e c o r r e l a t i o n f o r 3PD was derived from t h e d a t a set including only one amphiprotic s u b s t i t u e n t , t h e significance of t h e coefficient of t h e

p x

t e r m in Eq. 15, which was Justified only a t t h e 90% level, i s somewhat uncerThe coefficient of t h e

tain.

term in Eq. 16 f o r t h e 4PD system was

xqpy

considerably lower t h a n t h a t in t h e corresponding Eq. 11 in Table 11. substituted

pyridazine series, t h e two vicinally

s t r o n g l y i n t e r a c t wlth each other.

N in t h e r e f e r e n c e X-substituted

situated

ring

In t h e

N-atoms may

The e l e c t r o n i c i n t e r a c t i o n s between X and pyridines could be severely d i s t o r t e d be-

tween t h e corresponding X and N

in t h e X-substituted

pyridazines.

Under

such conditions, t h e regression coefficients of t h e x d p y term i n Eq. 16 would b e highly p e r t u r b e d by t h e N-N e n t is amphiprotic.

interactions, especially when t h e X-substitu-

Since t h e numbers of compounds included in t h e s e pyrida-

zine systems are not l a r g e enough, examination of t h i s type of p e r t u r b a t i o n a f f e c t i n g t h e bidirectional

model requires f u r t h e r study, with additionally

synthesized compounds Except f o r t h e s e minor uncertainties, however, t h e general p a t t e r n of v a r i a t i o n s in t h e P Y value is s i m i l a r with t h e sets of equations in Tables 11 and 12. ic terms

The sign of t h e regression coefficient of t h e bidirectional e l e c t r o n in Eqs.

12-16 can be rationalized a s being analogous t o t h a t

in

c o r r e l a t i o n s f o r d i s u b s t i t u t e d benzenes (8). The forward effect of e l e c t r o n withdrawing

X

substituents

tends

to

decrease

the

basicity

of

-N=.

being

unfavorable t o s o l v a t i o n with t h e more acidic water, enhancing t h e p a r t i t i o n ing toward t h e less acidic 1-octanol phase. P

Y

value (coefficient of

u

This process explains why t h e

Z ) i s positive in a l l t h e correlations.

The positive sign of t h e

u?

value (coefficient of

p

f o r t h e back-

ward effect reflects t h e electron-withdrawing n a t u r e of t h e aza (-N=) The c o e f f i c i e n t of t h e P

x

group.

t e r m in Table 12 seems t o be categorized i n t o two

TABLE 12 Correlations of n (dlazlne)

IC

( d i a z i n e ) w i t h Eq. 4

a z (pyridine)

=

-

pya,"

+

O F p x

+

C ~~~~

Systcm

r

Corrclation

F

E q No.

n

0 . 3 8 1 ~+ ~0.016 (0. 175) (0.101)

0. 994

0. 078

355. 8

18

12

0 . 5 9 9 P x - 0.0'27 - (0. 2 1 9 )

0. 973

0. 138

42. 1

11

13

0.9 3 1 ~ +3 0. ~ 6~5 9 a t ( m ) - 0. 1 1 6 +~ 0. ~ 031 (0.0 7 8 ) (0.219) (0. 1 5 8 1 ( 0 . 083)

0.995

0 . 054

278. 0

12

14

0. 026 (0. 239)

0 . 984

0. 1 1 7

40. 5

8

15

0 . 4 6 6 ~- ~0 . 0 4 6 (0. 328) (0. 176)

0.966

0 . 107

27. 8

10

16

I

XZPR

111

X ~ P W=

IV

X 5PM

Y

X ~ P D=

+

vI

z4PD

- 0.447aPIm)

=

1 . 1 3 0 n z ~+ ~0 . 7 5 6 u p ( r n ) (0. 242)

(0.080)

I.115nzpy

(0.2 6 7 )

-

O.7448P(p)

+

+

(0. 4 3 0 )

(0.5 0 1 1

=

0. 8 9 8 x 2 ~ ~0 . 9 4 1 0 Z ( m ) (0.2 4 0 ) (0. 774) 0.706X,,~ (0.191)

=

S

( 0 . 4221

*

+

0. 4 7 4 p x (0.6 3 5 )

+

169 g r o u p s : o n e is 0 6 i n Eq a r o u n d 0 43 i n E q s second

13 f o r t h e 4PM series and t h e o t h e r is c e n t e r i n g

12, and 14-16

13. t h e backward effect of

In E q

atom is d i r e c t e d t o t h e " p a r a " X s u b s t i t u e n t s .

ring-N

systems, it acts i n t h e "meta" d i r e c t i o n group

These

the

in o t h e r

The magnitudes of t h e s e t w o t y p e s ( p ) and

o f c o e f f i c i e n t c o u l d c o r r e s p o n d t o t h o s e of t h e u t h e second -N=

but

( m ) v a l u e s of

u

values, 0 6 and 0 4, are, however, c o n s i d e r a -

u

b l y lower t h a n t h o s e of 0 99 and 0 9 3 , r e s p e c t i v e l y . o b s e r v e d f o r s u b s t i t u t e d pyridines The a b o v e b i d i r e c t i o n a l

p r o c e d u r e with t h e use of

x z P y v a l u e as t h e r e f e r e n c e worked v e r y well f o r t h e

t h e corresponding

~

Z

ries i n c l u d i n g " i r r e g u l a r " s u b s t i t u e n t s i n t h e c o r r e l a t i o n of

x q p M se-

and

P

R

x

~

3,

with E q

P

Y

s u p p o r t i n g t h e a b o v e mentioned a n t i c i p a t i o n t h a t t h e components a t t r i b u t a b l e

t o t h e proximity

effects

between

similar among t h e s e 2 - s u b s t i t u t e d

the

a-substituent

diazine-x

p r o c e d u r e did n o t work. however, f o r t h e c o r r e l a t i o n between x from t h e pyridine- x

2 P M

and x

zpy

are

N-atom

and pyridine-

K ~ P Mseries

s u b s t i t u e n t s were considered

a n d amphiprotic

and

IC

very

values

The

if hydrogen-acceptable

together

was n o t v e r y low ( r

Though t h e simple =

0 94 1, t h e s h i f t s

v a l u e s f o r t h e s e hydrogen-bonding s u b s t i t u e n t s were t o o

i r r e g u l a r t o explain

3.1.2 Physicochemical Meaning of t h e C o r r e l a t i o n s : A s mentioned above, t h e p r e s e n t model r e p r e s e n t e d by E q

4, in which t h e p y r i d i n e - x ,

are u s e d as i n d e p e n d e n t v a r i a b l e s , assumed t h a t t h e (N+X) e n c e p y r i d i n e system a n d t h e second - N = with

each o t h e r

p and

p

unit in the refer-

(Y) i n t e r a c t b i d i r e c t i o n a l l y

The " b i d i r e c t i o n a l " i n t e r a c t i o n s

c o n s i d e r e d f o r t h e second -N=

(N+X) g r o u p

function

u

were,

however,

actually

o n l y with t h e s u b s t i t u e n t X b u t n o t with t h e

The i n t e r a c t i o n between t h e two N-atoms w a s t a k e n as being un-

changed. a t least within e a c h series of s u b s t i t u t e d d i a z i n e s

These assump-

t i o n s a p p a r e n t l y i n c l u d e o v e r s i m p l i f i c a t i o n s as s u g g e s t e d above, i n p a r t i c u l a r , f o r t h e 2PM. 3PD, and 4PD systems We p r e l i m i n a r i l y a t t e m p t e d t o a n a l y z e diazine- x

e a c h system

v a l u e s using x

PhX

for

The 5PM series w a s t h e o n l y one i n which a n a c c e p t a b l e c o r -

r e l a t i o n was found as shown i n E q

17 by r e g a r d i n g t h e d i a z a g r o u p (-N=(C)-N=)

a s t h e i n v a r i a b l e " s u b s t i t u e n t Y"

The c a l c u l a t e d

II

v a l u e s are also shown i n

T a b l e 7.

x n

5 P M

=

=

12

0.9 3 9 x ~ h x+ 0.5 6 2 ;~( m ) ( 0 . 188) (0. 510) r

=

0. 974

s

=

The a c c e p t a b l e q u a l i t y of E q

+

0. 126

1. 2 5 3 +~ 0. ~ 095

(0. 477)

F

=

( 0 . 188)

1171

48. 4

17 a s a c o u n t e r p a r t of Eq. 14 c o u l d b e under-

s t o o d by t h e f a c t t h a t t h e 5PM series

is

unique among o t h e r isomers i n being

170 symmetric, where t h e s u b s t i t u e n t and two N-atoms have "meta" relationships, perhaps without significant d i r e c t resonance e f f e c t s among t h e t h r e e functions. The "forward" e f f e c t of would,

in f a c t , r e p r e s e n t

i n t e r a c t i o n s between t h e two N-atoms a r e neglected. uPp

e f f e c t on s u b s t i t u e n t X in terms of proximately twice t h e e f f e c t of

u

t (m)

values in Eq.

in Eq.

each of

u

x"

in Eq.

17

t h e two N-atoms

if

X in terms of

substituent

twice t h e e f f e c t on each of

p

Likewise, t h e "backward"

17 could account f o r ap-

t h e N-atoms.

When t h e

and

p

17 a r e divided by two, t h e forward and backward ef-

fects between each p a i r of X and -N=

s u b s t i t u e n t s a r e 0.28 and 0.63. respec-

These values are c l o s e r t o t h e corresponding values, 0.26 and 0.94 in

tively.

Eq. 3 f o r

x x

values in t h e monosubstituted pyridine series on t h e basis of

z P h X . This seems t o s u p p o r t t h e relevancy of Eq. 1 in analyzing t h e substituent

II

values of N-heterocycles.

on t h e basis of

Eq. 14.

II S P Y .

p

Y

and

0

When t h e analysis of

S(m)

n,,,,

w a s made

were 0.66 and 0.42, respectively

The "enhanced" forward e f f e c t of X in Eq. 14 in terms of

p yu

in

x" ( m )

compared with 0 . 2 6 ~P in Eq. 3 could be due t o t h e e f f e c t of t h e N-atom

in

t h e (N+X) u n i t on t h e second N-atom in augmenting t h e effect of X so t h a t t h e use of t h e r e g u l a r u P would under-estimate t h e t o t a l effect of (N+X). ing t o a higher

value.

p y

lead-

The reduction of t h e backward e f f e c t could be

due t o a competition f o r e l e c t r o n s of hydrogen-bondable

s u b s t i t u e n t X be-

tween two N atoms a p p a r e n t l y non-additive in n a t u r e so t h a t t h e p

x

value in

Eq. 14 could be over-estimated, leading t o a lower u t ( m ) value. The above t y p e of i n t e r a c t i o n s among X-substituent and two N-atoms a r e c e r t a i n l y r e f l e c t e d in c o r r e l a t i o n equations o t h e r than Eq. 14 f o r t h e o t h e r diazine systems. The forward effect of X on t h e second N atom in terms of p y is higher t h a n t h a t f o r 5PM f o r systems in which t h e reference

(N+X) group

w a s t a k e n as t h e ZPY, such a s 2PR (Eq. 12). 4PM (Eq. 13), and 3PD (Eq. 15). r e v e r s e is t h e case f o r t h e 4PD system (Eq.

The

16) where t h e (N+X) group was

This t r e n d is understandable since t h e e f f e c t of X on

r e g a r d e d a s t h e 4PY.

N within t h e (N+X) group i t s e l f is highest in systems with t h e 2PY r e f e r e n c e and lowest in t h a t with t h e 4PY reference.

The higher t h e e l e c t r o n i c e f f e c t

of t h i s type, t h e g r e a t e r is t h e e x t e n t of t h e under-estimation ward e f f e c t by t h e use of t h e r e g u l a r

d

of t h e for-

P , leading t o t h e higher

The backward e f f e c t of t h e second N atom on t h e (N+X) in terms of coefficient of t h e t h e o t h e r systems. higher t h a n t h e u t p

x

P x

p

value.

u p , the

term. i s higher f o r t h e 4PM system (Eq. 13) t h a n f o r

This is understandable because t h e

u t h e t a ) f o r t h e -N=

uF(para) value i s

" s u b s t i t u e n t " , although t h e " i n t r i n s i c "

term f o r t h e e f f e c t of two N-atoms in diazine systems i s n o t additive

as mentioned above.

171 In E q

7 for t h e ZPM series. OR s u b s t i t u e n t s were w e l l i n c o r p o r a t e d . b u t

s u c h g r o u p s as SMe. N M e z a n d NHz w e r e p o s i t i v e o u t l i e r s and CONHz.

COZR were n e g a t i v e o u t l i e r s between

N H A c and

P o s s i b l e proximity steric and i n d u c t i v e e f f e c t s

X and t h e s e c o n d N atom were examined by i n t r o d u c i n g t h e Taft-

Kutter-Hansch

E,

and t h e C h a r t o n

u

I

terms f o r X. s i n g l y or t o g e t h e r , b u t The OR s u b s t i t u e n t s i n t h e 2PM

t h e y did n o t r a t i o n a l i z e t h e deviations

system would s h a r e proximity i n t e r a c t i o n s with j u s t o n e of common with t h o s e i n t h e 2PY system 2PM series. t h e

the N

atoms i n

For o t h e r outlying su b stitu e n ts in t h e

z Z P M v a l u e seems t o d e v i a t e from t h e c a l c u l a t e d v a l u e i n a

manner similar t o b u t n o t n e c e s s a r i l y t h e same as t h a t o c c u r r i n g with P a r a m e t r i z a t i o n of

t h e proximity e f f e c t s involved

h e t e r o c y c l i c compounds as

2-substituted

pyridines

K z p y

i n such 2 - s u b s t i t u t e d and

pyrimidines

N-

require

f u r t h e r elaboration

3.2 Di- and P o l y - s u b s t i t u t e d P y r a z i n e s Analysis and p r e d i c t i o n of t h e l o g P v a l u e s would b e more d i f f i c u l t f o r m u l t i - s u b s t i t u t e d d i a z i n e s , i n which a d d i t i o n a l e l e c t r o n i c and steric

interac-

2.0-

Q

1.0-

0 3

0 -

0 H

-0.5

I

I

I

I

I

I

Total Carbon Number Fig. 1. P l o t of log P a g a i n s t the c a r b o n number in a l k y l p y r a z i n e s Closed circles. 2-X-substituted p y r a z i n e s : t h e symbols r e p r e s e n t t h e X-substituents Open circles. p o l y s u b s t i t u t e d p y r a z i n e s : t h e numerals indicate t h e substituent positions S l o p e A: , S l o p e B: -------, S l o p e C: - - (Reproduced from r e f 14 with permission of t h e copywrite owner. t h e American Pharmaceutical Association)

172 t i o n s between individual p a r t n e r s of s u b s t i t u e n t s and ring-N atoms a r e generally

involved

A s a f i r s t t r i a l , we investigated possible procedures

t o t h e above mentioned bidirectional Hammett-Taft-type

similar

analyses f o r t h e log P

values of d i s u b s t i t u t e d pyrazines and extended t h e analysis t o p o l y s u b s t i t u t pyrazines (14)

ed

pyrazines

The experimentally measured log P values of

a r e l i s t e d in Tables 13-15, while t h o s e f o r

disubstituted

polysubstituted

pyra-

z i n e s are given in Table 16 In di- and poly-substituted systems. s t e r i c e f f e c t s may o p e r a t e in a

manner t h a t t h e s u b s t i t u e n t k ) n e x t t o t h e ring-N atom would

such

hinder

the

pair

elec-

solvent

molecule

trons

To test t h i s possibility. we first examined various a l k y l p y r a z i n e s

from forming a hydrogen bond with t h e N-lone

in

which t h e e l e c t r o n i c c o n t r i b u t i o n was thought t o be very low Log P values including mono- and poly-alkylsubstituted plotted shown

against in

Fig

pyrazines were

t h e t o t a l carbon number contained in t h e a l k y l 1

F o r t h e s e r i e s of 8-n-alkyl

substituted

chains

pyrazines,

p y r a z i n e t o 2-Bu-pyrazine, t h e increment per C1-unit was very r e g u l a r , mated

a s 048+0 01 (slope A)

When a methyl group was introduced

as from

esti-

in

their

o r t h o position, t h e increment for each alkylpyrazine was 0 3810 03 h = 4 , s l o p e

B).

which

was

lower t h a n t h a t f o r monoalkyl

pyrazines

With

successive

methyl s u b s t i t u t i o n s a t t h e f o u r available pyrazine positions, t h e log P v a l u e increased almost r e g u l a r l y from mono-Me t o Me4 derivatives (omitting with

a s l o p e (C.036i004) c l o s e t o s l o p e B

2,5-Mez)

These r e s u l t s suggest t h a t

di-

s u b s t i t u t i o n s of f a i r l y bulky groups a t t h e 2,3- and Z6-positions. b u t n o t t h e 2,5-position have s t e r i c e f f e c t s

3.2.1 C o r r e l a t i o n Analyses for Disubstituted Pyrazines: The a n a l y s e s were made f o r t h e

A

( d i s u b s t ) ~value ~ defined by E q

For comparison. t h e sum of t h e corresponding s u b s t i t u e n t from

( 2K 15

monosubstituted

It

pyrazines

(2A

zPR ) ,

and 2-substituted pyridines ( 2A

P h X ) ,

monosubstituted

benzenes

a r e presented in Tables 13-

Although t h e s i t u a t i o n is considerably improved with

d e v i a t i o n s from t h e observed values a r e still g r e a t decreased when t h e

XZPR

P

~

A ~ P Y .the

The deviations a r e much

values a r e used, although t h e a d d i t i v i t y t e n d s t o

r e s u l t i n over-estimation of values JC

A

of t h e log P v a l u e f o r most com-

v a l u e s r e s u l t s in g r e a t under-estimation

pyrazine

values derived

c l e a r l y seen t h a t t h e prediction based on t h e a d d i t i v i t y of

IS

pounds

Z P Y )

A

18

Thus, we s e l e c t e d t h e monosubstituted

values a s t h e r e f e r e n c e parameter

For applying E q

1 t o analyze t h e

A

( d i s u b s t ) p R values, some modifica-

X

173 t i o n s were r e q u i r e d

In u s e of

( d l s u b s t ) p R values r e l a t i v e t o t h e log P

II

v a l u e of t h e u n s u b s t i t u t e d p y r a z i n e . no d i s t i n c t i o n was made between X and Y The a n a l y s e s were

s u b s t i t u e n t s as t o which is fixed and which is v a r i a b l e made

for

individual

t y p e s of

disubstituted

pyrazines

t r e a t t h e sum of t h e b i d i r e c t i o n a l e l e c t r o n i c terms, a n i n d e p e n d e n t v a r i a b l e . as shown i n E q

where 2

r e p r e s e n t s t h e sum of

II 2~~

II

we

First, p yu

P and

tried

u pp

to as

19,

f o r X and Y s u b s t i t u e n t s .

ZPR

In t h i s

t r e a t m e n t , b i d i r e c t i o n a l i n t e r a c t i o n s are assumed t o be p r o p o r t i o n a l t o t h o s e i n m-

and p - d i s u b s t i t u t e d b e n z e n e s

The r e g r e s s i o n c o e f f i c i e n t "b" c o u l d b e

a n i n d i c a t i o n of t h e " t r a n s m i t t i n g e f f i c i e n c y ~ 'of t h e b i d i r e c t i o n a l i n t e r a c t i o n s of s u b s t i t u e n t s compared with t h e c a s e of t h e c o r r e s p o n d i n g l y d i s u b s t i t u t e d benzenes

Preliminary

analyses

of

the

values

II

d i s u b s t i t u t e d series l i s t e d in Tables 13 and 14 with

for

v a l u e s i n T a b l e 1. however. demonstrated t h a t Eq

and p

the

u X and

2.6- and 2.5u

constants

19 w a s n o t s u f f i c i e n t

t o s i m u l a t e t h e s i t u a t i o n in "meta" and " p a r a " d i s u b s t i t u t e d p y r a z i n e s W e n e x t t r i e d t o r e p r e s e n t t h e e l e c t r o n i c e f f e c t s by i n d u c t i v e and resou

n a n c e components s e p a r a t e l y using the

P X U I C Y ) ,

P Y U I C X ) ,

I

P Y ~ R ( X ) ,

and

u

and

P X ~ R C Y ,

electronic constants

With

terms, a n a l y s e s were

made by c o n s i d e r i n g whether t h e s u b s t i t u e n t s are amphiprotic (H-donor) or Hacceptor H-donor

W e found t h a t t h e s e t e r m s w e r e s t a t i s t i c a l l y signiflcant only f o r

X and Y s u b s t i t u en t s

In o t h e r words,

v a l u e s f o r H-acceptors

p

c o u l d b e r e g a r d e d as z e r o l i k e t h o s e for non-hydrogen-bonders

w a s t r a n s f o r m e d t o Eq

where p

20,

is t h e o r i g i n a l l y r e p o r t e d P

(AM,

Then, Eq. 19

v a l u e f o r amphiprotic g r o u p s b u t

is t a k e n t o b e z e r o f o r H-acceptors and non-hydrogen b o n d e r s (Table 1). a n d

2

u

p :An,

represents

for amphiprotic X and Y

+

P

:An,

Analysis of a l l t h e 2.6-disubstituted

R

( 2 , 6-PR) v a l u e s i n Table 13 yield-

p

1

Or

R

0 ;

Or

R

0 :

Or

R

substituent pairs

ed Eq. 21

II

( 2 , 6-PR)

=

1. 0 0 8 2

R z p R

+

0. 6 1 5 2 p

(AM) u I

(0.4 8 4 )

(0.0 6 3 ) +

0. 5832 P

(AM,

u

(0.2 5 0 ) n

=

39

r

=

0.987

s

=

0. 125

R

-

0 . 103

(0.1 0 8 ) F B . 35

=

446.2

1211

TABLE 13 Hydrophobicity Parameters of 2.6-Disubstituted Pyrazines (2.6-PR) K

( 2 , 6-PR)

Substituents

x, Y

C1,

F

c1, C1. C1. C1. C1.

c1 Me OMe OEt NH2 NHAc CN C02Me CONHz NMe2

C1. C1. C1. C1, C1, C1. OPr Me, Me Me, OMe Me, N H 2 Me. N H A c Me, C N Me, C02Me Me, COzEt Me. C O N H 2 Me. NMez Me, O P r

log P

Obsd.

Calcd.”’

1 . 15 1. 5 3 1. 03 1. 6 5 2. 2 2 0. 9 5 1. 10 0. 79 0. 4 7 0. 28 1. 9 5 2. 71 0. 5 4 1. 2 9 0 . 35 0 . 38 0. 4 4 0. 1 0 0. 51 -0.13 1. 57 2 . 38

1.41 1.79 1.29 1. 9 1 2. 4 8 1. 2 1 1. 3 6 1. 05 0. 7 3 0. 5 4 2. 21 2. 97 0.80 1. 5 5 0. 6 1 0. 6 4 0. 70 0. 36 0. 7 7 0. 1 3 1. 8 3 2. 64

1.47 1.82 1. 3 4 1 . 88 2. 4 0 1. 2 1 1. 2 5 1. 18 0. 8 2 0. 61 2. 07 2. 9 3 0. 85 1. 4 1 0.5 7 0. 57 0. 70 0 . 30 0.78 -0. 01 1. 5 9 2. 46

dev. -0. -0. -0. 0.

06 03 05 03 0. 0 8 0. 0 0 0 . 11 - 0 . 13 -0. 0 9 -0. 07 0. 1 4 0. 04 -0. 05 0. 1 4 0. 0 4 0. 07 0. 00 0. 06 -0. 01 0. 14 0. 24 0. 18

Z x 2 p R b ’ dev. 1 . 51 1. 9 2 1.43 1. 95 2. 50 1. 1 7 1. 1 9 1.21 0. 9 9 0. 72 2. 15 3. 06 0.94 1 . 46 0.68 0. 7 0 0. 7 2 0. 50 1. 0 1 0. 2 3 1. 6 6 2. 5 7

-0. 10 13 14 04 02 0. 04 0. 1 7 -0. 16 -0. 26 - 0 . 18 0. 06 -0.09 -0.14 0. 0 9 -0.07 -0. 06 -0. 02 -0. 1 4 -0.24 -0. 1 0 0. 17 0 . 07

-0. -0. -0. -0.

Zxphxc’ 0 . 85 1. 4 2 1. 27 0. 6 9 1. 0 9 -0. 5 2 -0. 2 6 0. 1 4 0. 70 - 0 . 78 0. 8 9 1. 76 1. 1 2 0. 5 4 -0. 67 -0.41 -0. 0 1 0.55 1. 07 -0. 9 3 0. 74 1. 6 1

dev.

0. 5 6 0. 3 7 0. 0 2 1. 2 2 1. 3 9 1. 7 3 1. 6 2 0. 9 1 0. 0 3 1. 32 1. 3 2 1. 2 1 -0. 32 1. 0 1 1. 28 1. 0 5 0. 7 1 -0.19 -0. 30 1. 0 6 1. 0 9 1. 0 3

Zzapyd’

0 . 81 1 . 35 1. 0 8 1 . 31 1. 7 8 0. 4 5 0. 4 9 0. 37 0 . 33 0. 1 2 1. 6 2 2. 3 5 0. 9 2 1 . 15 0. 29

0 . 33 0. 21 0. 1 7

0. 6 8 -0. 04 1. 4 6 2. 19

dev. 0. 0. 0. 0. 0.

60 44 21 60 70 0. 7 6 0. 8 7 0. 6 8 0. 4 0 0. 42 0. 59 0. 6 2 -0. 1 2 0. 40 0. 3 2 0. 31 0. 4 9 0. 19 0. 09 0. 1 7 0. 3 7 0. 45

OMe. OMe OMe. O E t OMe. N H 2 OMe. NHAc OMe, C N OMe. C02Me OMe. C O z E t OMe, C O N H 2 OMe. N M e . 2 C N . NMez NMe2. CO2Me NMe2. C 0 2 E t NMe2. C O N H 2 CONHz. O E t F. F OEt, OEt NH2, NH2

1. 5 8 1. 9 8 0.73 0. 82 0.95 0 . 69 1 . 20 0. 13 1. 99 1 . 12 0 . 90 1. 24 0. 38 0. 61 0 . 74 2. 55 -0.45

1.84 2. 24 0.99 1. 08 1.21 0. 95 1. 4 6 0. 39 2. 2 5 1. 38 1. 1 6 1. 5 0 0. 64 0.87 1.00 2.81 -0.19

1. 93 2.45 1.02 1. 0 0 1.23 0. 93 1. 42 0. 58 2. 1 2 1. 41 1. 11 1. 59 0. 6 3 1. 10 1.10 2. 97 -0.22

-0. 09 -0. 21 -0.03 0. 08 -0.02 0 . 02 0 . 04 -0. 19 0 . 13 -0. 03 0. 05 - 0 . 09 0. 01 -0. 23 -0.10 -0. 16 0. 03

1. 9 8 2. 5 3 1.20 1.22 1. 24 1.02 1.53 0. 75 2. 1 8 1. 4 4 1. 2 2 1.73 0.95 1. 3 0 1.10 3. 08 0.42

- 0 . 14 -0. 29 -0.21 -0.14 -0. 0 3 -0.07 -0.07 -0. 36 0. 07 - 0 . 06 - 0 . 06 -0.23 -0.31 -0. 43 -0.10 - 0 . 27 -0.61

- 0 . 04 0 . 36 -1.25 -0.99 -0.59 -0. 03 0. 49 -1. 51 0. 16 -0. 39 0. 17 0. 69 -1.31 -1. 11 0. 28 0. 76 -2. 46

1. 8 8 1. 8 8 2. 24 2.07 1. 8 0 0. 98 0. 97 1. 9 0 2. 09 1. 7 7 0. 99 0 . 81 1. 9 5 1. 98 0. 72 2. 05 2. 27

1 . 38 1. 8 5 0. 52 0. 56 0. 44 0 . 40 0. 91 0. 19 1. 6 9 0. 75 0. 7 1 1. 22 0 . 50 0. 66 0. 38 2. 32 -0.34

0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.

46 39 47 52 77 55 55 20 56 63 45 28 14 21 62 49 15

a) C a l c u l a t e d b y E q . 23. F o r example, t h e v a l u e for 2-C1-6-NH2-pyrazine 1s c a l c u l a t e d as f o l l o w s : H (2-Cl-6-NHz) = 0 . 9 4 7 { H z p n ( C l ) + a z p n ( N H z ) } + 0. 5 6 8 ( p ( A M ) ( N H ~ ) * u I ( C ~ + ) P ( A M ) ( C ~ ) * ~ I ( N H ~ ) } + 0 . 6 8 0 ( p ( A M ) ( N H 2 ) * o , ( C l ) + P ( A M ) ( C l ) * o , ( N H z ) } - 0 . 080Ee(C1)*E,(NH2) + 0 . 078 = 0. 9 4 7 ( 0 . 9 6 + 0 . 2 1 ) + 0 . 5 6 8 ( 0 . 7 4 * 0 . 4 7 + 0) + 0 . 6 8 0 { 0 . 7 4 ( - 0 . 2 5 ) + 0) - 0 . 0 8 0 ( - 0 . 9 7 ) (-0.6 1 ) + 0 . 0 7 8 = 1 . 2 1 b) For t h e H 2 P R v a l u e of t h e component s u b s t i t u e n t s . see T a b l e 4. c) F o r t h e H~~~ v a l u e of t h e component s u b s t i t u e n t s . see T a b l e 1. d) For t h e H 2 p y v a l u e of t h e component s u b s t i t u e n t s . see T a b l e 1.

176 Although E q

21 seems t o b e s t a t i s t i c a l l y a c c e p t a b l e , examinations of

r e s i d u a l s showed t h a t t h e

v a l u e s of

K

compounds with bulky s u b s t i t u e n t s

s u c h as COzR a n d CONHz g a v e s i g n i f i c a n t n e g a t i v e d e v i a t i o n s

Moreover, t h e

l e v e l of s i g n i f i c a n c e of t h e i n t e r c e p t w a s c l o s e t o 95% in E q

21 These facts

were

effect

thought

Lo

reflect

a

contribution

d i s u b s t i t u t i o n i n lowering t h e derivatives

E

(11)

of

the

steric

of

2.6-

v a l u e as described above f o r 2.6-dialkyl

K

This c o n t r i b u t i o n was examined with u s e of t h e steric 7arameter Eq

ZE, ( 2 , 6 ) v a l u e as an a d d i t i o n a l term

22 o b t a i n e d with t h e

For E, v a l u e s of p l a n a r r-bonded s u b s t i t u -

w a s s l i g h t l y b e t t e r t h a n E q 21

e n t s . t h e v a l u e s f o r t h e steric effect on t h e c o p l a n a r , b u t n o t t h e perpend i c u l a r d i r e c t i o n were used x ( 2 , 6-PR)

=

0. 9502 +

0. 5692

+

K ZPR

(0.0 7 0 )

0. 6 7 1 2 ~ CAM)

+

O R

(0.2 3 5 ) r

n = 39 In E q

0.990

=

u

(AM)

u

I

(0.4 4 0 )

s

0. 0722E, (0.0 5 0 )

0. 1 1 3

=

+

0 . 140 ( 0 . 193)

1221

411. 1

=

22, however, t h e i n t e r c e p t t u r n e d o u t t o be p o s i t i v e and t h e d e v i a t i o n

from z e r o w a s i n c r e a s e d from t h a t i n Eq. 21. effect

of

2.6-substituents.

being

This suggested t h a t t h e steric

overestimated f o r s u b s t i t u e n t s

of

smaller

size, d i d n o t o p e r a t e e x a c t l y additively. b u t p r o p o r t i o n a l l y with t h e bulk of t h e two s u b s t i t u e n t s To e x p r e s s t h i s s i t u a t i o n , t h e cross-product

of

E,(X)

and E,(Y),

Es',

would b e r e l e v a n t , s i n c e it t a k e s v a l u e s c l o s e t o z e r o when one or b o t h of t h e s u b s t i t u e n t s X a n d Y idare) small enough, b u t it i n c r e a s e s Lo a signific a n t l y l a r g e size when b o t h X and produced E q 23.

Y are large

The a d d i t i o n of t h e E,'

term

Although t h e q u a l i t y of t h e c o r r e l a t i o n w a s almost t h e same

as t h a t of E q 22, t h e i n t e r c e p t of E q 23 was much c l o s e r t o z e r o (2. 6-PR)

=

0. 9 4 7 2 II (0.0 7 2 ) -

2PR

0. 080E,'

+

(0.0 5 7 ) n

=

r

39

=

0 . 990

s

=

O. 5682 P (0.443)

+

(AM)

U I

+

0. 6 8 0 2 p

(AM)

0 . 078 ( 0 . 161)

0. 1 1 3

u

R

(0.2 3 8 ) [231

F4. 3 4

=

405. 7

I t is e x p e c t e d t h a t t h e a n a l y s i s f o r 2.5-disubstituted p y r a z i n e s c o u l d b e d o n e in a manner similar t o Eq. 21, s i n c e no s i g n i f i c a n t steric effect would b e induced i n t h e r e l a t i v e s o l v a t i o n of t h e ring-N atoms by i n t r o d u c i n g t h e s e c o n d s u b s t i t u e n t in t h e " p a r a " position. a n d 25 of methoxy

approximately

derivative

was

equivalent omitted

For t h e d a t a in Table 14, Eqs. 24

quality

from t h e

were formulated.

correlation

analyses,

The 2.5-dibecause

it

177 showed l a r g e d e v i a t i o n s i n a l l preliminary t r i a l s

A

( 2 , 5-PR)

0. 9622

=

0. 4752 P

A PPR

(0.0 7 0 ) n

A

n

r

15

=

(2, 5 P R )

15

=

=

0. 994

s

0.9 5 4 2

=

=

K Z P R

+

=

12

0. 5 6 2 2 ~ (AM) O

0. 994

s

0. 0 8 2

R

=

12

493.6

From t h e s t a t i s t i c a l p o i n t of view. it is d i f f i c u l t two e q u a t i o n s 2 p

=

-0 78) among t h e s e f i f t e e n compounds

cal p o i n t of view. however, w e c o n s i d e r t h a t E q

of

the

t o choose between t h e s e

There was a f a i r l y high c o l i n e a r i t y between

(r

u

[251

(0.1 0 9 )

Fz.

0. 077

=

1241

473. 7

=

(0.3 6 2 )

(0.0 6 9 )

r

0. 096 (0.1 0 9 )

(AM) u I

(0.3 2 0 ) 0. 0 7 9 Fz,

2p u

term

is p o s i t i v e ,

2 p

(AM)

u

I

and

From t h e physicochemi-

25, i n which t h e c o e f f i c i e n t

is p r e f e r a b l e

The

electron-withdrawing

effect of s u b s t i t u e n t s t e n d s t o i n c r e a s e t h e a c i d i t y of t h e second s u b s t i t u e n t s . enhancing t h e s o l v a t i o n with 1-octanol, which is and l e a d i n g t o a h i g h e r The b e h a v i o r of t h o s e of from

A

A

( 2 , 5-PR)

m o r e b a s i c t h a n water,

value

A

( 2 , 3-PR)

and

w a s e x p e c t e d t o b e more complicated t h a n

(2. 6-PR),

A

steric i n t e r f e r e n c e as w e l l

because it c o u l d i n v o l v e components possible

as

proximity

electronic

effects

between s u b s t i t u e n t s

Since t h e compounds i n o u r 2,3-PR series d o n o t in-

c l u d e hydrogen-donor

s u b s t i t u e n t s , t h e c o r r e c t i o n for t h e e l e c t r o n i c e f f e c t s

2 p c

i n t e r m s of

2 p

and

AM) u I

(AM) 0 R

was of

no significance

Further

measurements a r e needed t o o b t a i n a g e n e r a l i z e d c o r r e l a t i o n e q u a t i o n t h a t c o u l d a p p l y t o a wider r a n g e of t h e analysis for

A

n

( 2 . 3-PR) =

=

A

1. 0 6 0 2 A z p ~

(0.1 2 6 ) r = 0. 9 8 4

13

2,3-disubstituted pyrazines

( 2 . 3 - P R ) v a l u e s i n Table 15 y i e l d e d Eq 0. 125

C261

(0.2 1 4 ) s

=

0. 131

FI.

1 1

=

343. 3

The c o r r e l a t i o n was improved by t h e a d d i t i o n of t h e E,' a n a l y s i s of

A

( 2 , 3-PR)

=

A

(2. 6-PR),

0. 9 9 2 2

=

The

r

13,

p u

I

=

0. 992.

XZPR

-

0 . 050E,'

+

(0.0 3 4 ) s

=

0 . 096, F P .

term for hydrogen-accepting

test for a p a r t i c i p a t i o n

t e r m as done i n t h e

and gave Eq. 27,

(0.1 0 4 ) n

Nevertheless,

26

=

0 . 082 (0.2 1 3 )

1271

327. 1

s u b s t i t u e n t s was added t o Eq

27 t o

of t h e proximity e l e c t r o n i c (inductive) e f f e c t s be-

tween 2- and 3 - s u b s t i t u e n t s . b u t no improvement was observed.

TABLE 14 Hydrophobicity Parameters of 2,5-Disubstituted Pyrazines (2,5-PR)

K

(2.5-PR)

Substituents

x. Y

c1. C1. C1, C1. C1. C1. C1. C1, Me, Me, Me, Me, Me, OMe, OMe. OMe,

c1 Me OMe

OEt NHz NHAc CN NMez Me CN COzMe COZEt CONHZ OMe NH2 NMe2

log P

Obsd.

1. 58 1. 08 1. 52 1. 99 0. 67 0. 56 0.9 2 1. 70 0. 63”’ 0. 26 0. 17 0. 61 -0.25 1. 14 0. 63 1. 65

1. 84 1 . 34 1. 78 2. 25 0. 93 0.82 1. 18 1. 96 0.89 0.52 0. 43 0. 87 0.01 1. 40f’ 0. 89 1.91

Calcda’

dev.

1. 75 1. 28 1. 78 2.30 0. 93 0. 92 1. 07 1. 97 0.82 0.61 0 . 40 0.88 0.10 1.81 0. 82 2.00

0. 09 0. 06 0. 00 -0. 05 0. 00 -0. 10 0. 1 1 -0.01 0. 07 -0.09 0. 03 -0.01 -0.09 -0.41 0.07 -0.09

Z n

Z p ~ b ’

1. 92 1.43 1.95 2. 50 1. 17 1.19 1.21 2. 15 0. 94 0. 72 0. 50 1.01 0. 23 1. 98 1.20 2. 18

dev

-0. 08 -0.09 -0.17 -0. 25 -0. 24 -0.37 -0.03 -0.19 -0.05 -0.20 -0. 07 -0.14 -0. 22 -0. 58 -0.31 -0. 27

2

K PhXC’

1. 42 1. 27 0. 69 1. 09 -0.52 -0. 26 0. 14 0. 89 1.12 -0.01 0.55 1. 07 -0.93 -0.04 -1.25 0. 16

dev.

0. 42 0. 07 1. 09 1. 16 1. 45 1. 08 1. 04 1. 07 -0.23 0. 53 -0.12 -0.20 0 . 94 1. 44 2. 14 1. 75

2

zpyd’

1. 24 1. 08 1. 31 1. 78 0. 45 0. 49 0. 37 1. 62 0. 92 0. 21 0. 17 0. 68 -0.04 1. 38 0. 52 1. 69

dev.

0. 60 0. 26 0. 47 0. 47 0. 48 0. 33 0. 81 0. 34 -0.03 0. 31 0. 26 0. 19 0. 05 0. 02 0. 37 0. 22

-~ ~~

a) Calculated by Eq. 25. For example, t h e value f o r 2-0Me-5-NHZ-pyrazine 1s c a l c u l a t e d a s follows: n ( 2 - 0 M e - 5 - N H Z ) = 0. 954{n z p R (OMe) + K z p ( N~ H 2 ) } + 0 . 562{p ( A M ) ( N H z ) * U R (OMe) + p ( A M ) (OMe) * - 0 . 082 = 0. 954(0. 99 + 0. 21) + 0. 562{0.74(-0. 57) + 0) 0. 082 = 0.82 b) For t h e n Z P R value of t h e component s u b s t i t u e n t s , see Table 4. c) For t h e K P h X value of t h e component s u b s t i t u e n t s , see Table 1. d) For t h e n z p y value of t h e component s u b s t i t u e n t s . see Table 1. e ) From ref. 6. f ) Omitted from t h e a n a l y s i s b u t calculated by Eq. 25. -

u

R

(NHz) }

TABLE 15 H y d r o p h o b i c i t y P a r a m e t e r s of 2 . 3 - D i s u b s t i t u t e d P y r a z i n e s (2.3-PR)

R

Substituents Y

x.

Me, Me, Me, Me,

Me

Et

Pr n-Bu Et. Et M e , OMe Me, O E t Me, O C H M e z E t . OMe Me, SMe CN. CN COZMe. COzMe COzEt. C O z E t

log P 0. 54 1. 0 7 1. 5 7 2. 1 0 1 . 51 1. 2 4 I. 82 2. 24 1. 8 0 1. 81 0. 38 -0. 02 0. 65

Obsd. 0.80 1. 3 3 1.83 2. 36 1. 77 1. 5 0 2. 08 2.50 2. 06 2. 07 0. 64 0. 24 0. 91

( 2 . 3-PR) Calcda’

dev.

0.94 1.41 1.87 2. 38 1.88 1. 50 2. 04 2.54 1. 9 7 1. 9 0 0. 56 0.26 0. 8 4

-0. 14 -0. 08 -0. 04 -0. 02 - 0 . 11 0. 00 0. 04 -0.04 0. 09 0. 17 0. 08 -0. 02 0. 07

Z

K 2pRb’

0.94 1. 4 2 1.90 2. 4 2 1.90 1. 4 6 2. 0 1 2.51 1. 9 4 1. 9 0 0. 5 0 0. 50 1. 0 8

dev. -0.14 -0. 09 -0.07 -0. 06 -0.13 0. 04 0. 07 -0. 01 0. 12 0. 17 0. 14 -0. 26 -0. 17

Z

A PhXC’

1. 1 2 1. 5 8 2 . 11 2. 6 9 2. 04 0. 54 0. 94 1. 4 1 1. 0 0 1. 1 7 -1. 1 4 -0. 02 1.02

dev.

-0. 32 -0. 25 -0. 28 - 0 . 33 -0. 27 0. 96 1. 14 1. 0 9 1. 06 0. 90 1. 7 8 0. 26 -0.11

Z

dev.

A 2pyd’

0. 92 1.41

-0. 12 -0. 08

-e

-

-e

-e

1.90 1. 1 5 1. 6 2

e

-0.13 0. 3 5 0. 4 6

-e

-e

0. 4 2 0. 55 1. 1 4 0.82 0. 4 7

1. 6 4 1. 5 2 -0.50 -0.58 0. 44

~~

a) C a l c u l a t e d b y Eq. 27. F o r e x a m p l e . t h e value f o r 2-Me-3-OMe-pyrazme IS calculated as f o l l o w s : I[ ( 2 - M e - 3 - O M e ) = 0 . 9 9 2 I Z~P R ( M ~ ) + A ZPR (OMe) } - 0 . 0 5 0 E , (Me) *E, ( O M e ) + 0. 0 8 2 = 0. 9 9 2 ( 0 . 47 - 0 . 0 5 0 ( - 1 . 2 4 ) ( - 0 . 5 5 ) + 0 . 0 8 2 = 1. 50 b) For t h e X ~ P Rv a l u e of t h e c o m p o n e n t s u b s t l t u e n t s . see T a b l e 4. c) F o r t h e z p h X v a l u e of t h e c o m p o n e n t s u b s t i t u e n t s , see T a b l e 1. d) For t h e R Z P Y v a l u e of t h e c o m p o n e n t s u b s t l t u e n t s , see T a b l e 1. e) T h e c o m p o n e n t n 2 p Y v a l u e s are unknown.

+

0. 9 9 )

180 The c a l c u l a t e d v a l u e s according t o E q s

23. 25 and 27 a r e l i s t e d in

Tables 13-15

3.2.2 Physicochernical Meaning of t h e Correlations: The r e s u l t s described above demonstrated t h a t t h e l o g P value of d i - s u b s t i t u t e d p y r a z i n e s having non-amphiprotic s u b s t i t u e n t s could be approximately p r e d i c t e d by t h e additivi t y model of

k 2 P R

v a l u e s provided t h e i r steric effects a r e n o t s i g n i f i c a n t

The b i d i r e c t i o n a l e l e c t r o n i c c o r r e c t i o n terms expressed by p u

p u

and/or

I

a r e needed o n l y f o r amphiprotic substituents(H-donors), possibly because t h e electron-withdrawing

effect of

r i n g N-atoms in t h e pyrazine to

ring tends t o

enhance t h e hydrogen-donating

ability but

reduce t h e proton-accepting

c a p a b i l i t y of t h e s u b s t i t u e n t s

The n e t hydrogen-donating a b i l i t y of c e r t a i n

amphiprotic g r o u p s could be higher t h a n t h a t in s u b s t i t u t e d benzene systems Thus, t h e p

For t h e hydrogen-

v a l u e is s i g n i f i c a n t f o r amphiprotic groups

a c c e p t o r s with lower o r i g i n a l

values. t h e s i t u a t i o n could be d i f f e r e n t

p

Under s u c h conditions, hydrogen a c c e p t o r s would behave as if t h e y were nonis z e r o o r v e r y c l o s e t o zero.

hydrogen b o n d e r s in which t h e "effective" p

T h e r e f o r e , o n l y t h e terms f o r t h e amphiprotic s u b s t i t u e n t s p a r t i c i p a t e i n t h e

20

c o r r e l a t i o n of t h e t y p e of Eq

I t should be noted t h a t t h e bR v a l u e is comparable with br in E q f o r 2.6-PR

Considering t h a t

t h e resonance

importance f o r meta-derivatives.

almost e q u i v a l e n t with t h a t from t h e p u intervention

through

the

ring-N

Eq

P

s u b s t i t u e n t s a r e " p a r a " t o each o t h e r 23 is c l o s e t o t h a t in E q

O R

by

the

two

those

having

two

because

The reason why t h e bR v a l u e i n

25 is n o t c l e a r bulky

minor

substituents

term is significant,

In f a c t . however, examination

of U V s p e c t r a in water showed t h a t t h e absorption of most 2.6-PR

including

of

t e r m t h a t is

p U R

effect might i n d i c a t e a resonance

I

atom sandwiched

For 2.5-PR, it is n o t unexpected t h a t t h e the

is g e n e r a l l y

effect

a c o n t r i b u t i o n from t h e

23

groups.

was

almost

derivatives,

identical

with,

s l i g h t l y more bathochromic t h a n t h a t of t h e corresponding 2.5-PR

but

compounds,

s u g g e s t i n g n o s i g n i f i c a n t d i f f e r e n c e in t h e resonance s t r u c t u r e between t h e two series In any case, considering t h a t

u

=

u

I

+

R

and

u

u

=

+

I

( 1 0 1 , t h e o b s e r v a t i o n t h a t t h e br and bn c o e f f i c i e n t s of t h e ., p u

"

0 4u

t e r m s in

a l l t h e e q u a t i o n s obtained were considerably smaller t h a n u n i t y lead u s t o conclude t h a t t h e p a r t i t i o n i n g behavior in t h e pyrazine system is p e r t u r b e d by t h e X-Y e l e c t r o n i c i n t e r a c t i o n s t o a "lesser" e x t e n t t h a n in t h e benzenoid system

This means t h a t

s u b s t ) PR

ring-N

some of

v a l u e s including t h a t

t h e e l e c t r o n i c component attributable

atoms and s u b s t i t u e n t s has been

reference

t o the

simulated

by

in t h e

x (di-

i n t e r a c t i o n between using

K

z P R as

the

The i n t r o d u c t i o n of b u l k y s u b s t i t u e n t s d e c r e a s e d t h e l o g P v a l u e i n t h e and 2,6-PR series

2,3-

water molecule

Besides t h e fact mentioned above t h a t t h e smaller

is more e a s i l y

a c c e s s i b l e t o t h e crowded ring-N

atom,

the

steric h i n d r a n c e t o t h e a t t a i n m e n t of a p l a n a r conformation might b e considThis p o s s i b i l i t y c o u l d p r o b a b l y b e eliminated. however, a t least i n 2.6-

ered

PR. j u d g i n g from t h e a b o v e mentioned U V spectral f e a t u r e s

2.5-Dimethoxypyrazine was t h e o n l y o u t l i e r among 68 compounds examined here

I n t e r e s t i n g l y , t h e summation of

g a v e a b e t t e r p r e d i c t i o n t h a n t h a t of a n a l y s i s of

(Eq

3)

K

2py,

~ Z P Yv a l u e s

K PPR,

for t h e two OMe g r o u p s

as shown i n Table 14

g r o u p was shown t o b e o n e of

t h e 2-OMe

Our e x p l a n a t i o n is t h a t 2-methoxy-pyridine

In t h e

the outliers

u n d e r g o e s a 1-to-1

che-

s o l v a t i o n i n water b u t n o t i n o c t a n o l t h a t enhances t h e log P

lating-type

value irregularly

Since t h e e f f e c t of s u c h s o l v a t i o n o c c u r r i n g a t b o t h of

two methoxy s u b s t i t u e n t s c o u l d also be simulated on t h e b a s i s of

KZPR.

no

r e a s o n a b l e e x p l a n a t i o n was made o n t h e behavior of t h i s compound

3.2.3 P o l y s u b s t i t u t e d P y r a z i n e s : The empirical c o r r e l a t i o n e q u a t i o n s f o r di-substituted

pyrazine

n

v a l u e s were a p p l i c a b l e t o p r e d i c t

of i n t e r a c t i o n s (2,6-,

t h o s e of t h e

Taking i n t o a c c o u n t a l l t y p e s

p o l y s u b s t i t u t e d p y r a z i n e s given i n Table 16

2.5- and 2 , 3 - i n t e r a c t i o n s ) , t h e g e n e r a l form of t h e p r e -

d i c t i o n e q u a t i o n is w r i t t e n a s E q

28

TABLE 16 Hydrophobicity P a r a m e t e r s of P o l y s u b s t i t u t e d P y r a z i n e s (Poly-PR)

Substituents log P

Obsd.

Calcd?’

H

0 . 95

Me

1. 2 8 1. 9 5 1 . 50

1. 2 1 1 . 54 2. 2 1 1. 7 6 1.21 1.55 2 . 05 2. 56 3. 1 5

1. 1 9 1. 44 2. 1 0 1. 6 9 1.35 1.70 2. 12 2. 51 3. 04

x2

x3

x5

XI3

Me Me

Me Me Et C1 CN CN CN CN CN

Me Me Me

Et Me

CN CN CN CN CN

Me C1 Me Et

Me Bu

H H H C1 C1 Bu C1

0 . 95 1. 2 9 1. 7 9 2 . 30 2. a 9

dev.

I: K

0. 0 2

2pRb)

1.41 1.88 2. 37 1.90 1.46 1.93 2.41 2. 92 3.41

0. 1 0

0. 11 0. 0 7 -0.14 -0.15 - 0 . 07 0 . 05 0 . 11

dev.

-0. 20 - 0 . 34 -0.42 -0.14 -0.25 -0.38 - 0 . 36 - 0 . 36 - 0 . 25

a) C a l c u l a t e d by Eq. 28. For example, t h e v a l u e f o r Z3-di-CN-5-Me-6-CIp y r a z i n e is c a l c u l a t e d as follows: n ( C N , C N , Me. C1) = n ( 2 . 3 - P R ) (CN. C N ) + n (3. 5 P R ) ( C N , Me) JT ( 5 . 6 - P R ) (Me. CI) n ( 2 . G P R ) ( C l , CN) K ( 2 . 5 - p ~ () C N , Me) + (3. 6 P R ) (CN, C l ) - 2(2* ~ z P R ( C N ) n2PR(Me) a Z P R ( C l ) } = Eq. 27(CN. C N ) + Eq. 23(CN, Me) + Eq.27(Me8CI) Eq.23(CI,CN) Eq.25(CNSMe) Eq.25(CN,C1) 2(2nZPR(CN) + nzzpR(Mf?) z z p n ( c 1 ) ) = 0. 56 0.70 1 . 4 5 + 1. 18 + 0 . 6 1 + 1 . 0 7 - 2 ( 2 * 0 . 25 + 0.47 0 . 9 6 ) = 1. 70, i n whlch Eq.N(X,Y) r e p r e s e n t s t h e II v a l u e c a l c u l a t e d by Eq. N f o r X.Y-disubstituted p y r a z i n e s . b) F o r t h e n 2 P R v a l u e o f t h e component s u b s t i t u e n t s , see Table 4. +

+

+

+

+

+

+

+

+

+

+

+

~

182

R

(poly-PR)

r e p r e s e n t s t h e difference in l o g P values between t h e given

compound and t h e u n s u b s t i t u t e d pyrazine, a v a i l a b l e positions, and

K (1,

one of t h e c o r r e l a t i o n E q s t i o n s of XI and X,

J-PR)

X is any s u b s t i t u e n t on t h e f o u r

is t h e value c a l c u l a t e d with t h e use of

23, 25 and 27. depending on t h e r e l a t i v e o r i e n t a is t h e sum of

R 2 P R ( X )

R ZPR

values o v e r a l l s u b s t i t

uents. and n is 1 and 2 f o r tri- and t e t r a - s u b s t i t u t e d derivatives, r e s p e c t i v e ly

By t h e addition of

x (I,

J-PR)

in E q

28, we counted each

R

(X-PR)

one

o r two times e x t r a depending on whether t h e compound is t r i s u b s t i t u t e d or

t e t r a - s u b s t i t u t e d , so t h e negative c o r r e c t i o n was needed

The c a l c u l a t e d

l o g P v a l u e s simulated t h e observed values f o r 9 compounds very w e l l (r

0 99) and a r e l i s t e d in Table 16 of

matter

The good predicting power of E q

course, since t h e s u b s t i t u e n t s included

bonders or "weak" hydrogen a c c e p t o r s

=

28 was a

a r e e i t h e r non-hydrogen

Although t h e r e l i a b i l i t y of

Eq

28

should b e examined f u r t h e r on various pyrazines s u b s t i t u t e d by amphiprotic s u b s t i t u e n t s . it indicates t h a t E q s 23, 25 and 21 a r e p r e d i c t i v e enough as f a r as t h i s d a t a set is concerned 4.

CONCLUSIONS The above analyses are believed t o indicate t h a t bidirectional Hammett-

Taft-type t r e a t m e n t s can be applied t o r a t i o n a l i z e t h e v a r i a t i o n s in

K

values

of s u b s t i t u e n t s o r increment in log P with polysubstitutions in each of t h e

various types

diazine systems, unless

of

solvations

and/or

the

substituents are

intramolecular

involved

interactions

For

monosubstituted diazines, t h e p a i r of one of t h e two ring-N substituent

was

regarded

as

a

"single" s u b s t i t u e n t

between t h e " s u b s t i t u e n t set" and

t h e o t h e r ring-N

and

in

specific

analyses

the

interactions

atom t o r e g u l a t e

r e l a t i v e s o l v a t i o n s with p a r t i t i o n i n g s o l v e n t s were analyzed

of

atoms and t h e

The

the

v a l u e of

K

corresponding s u b s t i t u e n t s in s u b s t i t u t e d pyridines was used a s t h e r e f e r ence

For t h e di- t o p o l y - s u b s t i t u t e d pyrazines. t h e i n t e r a c t i o n s between t h e

s u b s t i t u e n t p a i r s were primarily considered. t h e sum of s u b s t i t u e n t

R

values

in t h e monosubstituted pyrazine system being t h e r e f e r e n c e hydrophobicity Although t h e f a c t o r s governing t h e hydrophobicity of s u b s t i t u t e d diaz i n e s w e r e s e p a r a t e d t o components a s f a r a s well-behaved

compounds a r e

concerned, t h e y were r a t h e r complex and t h e analyses should be done v e r y carefully

These f a c t o r s should be considered in c o n s t r u c t i n g computer-aided

automatic systems t o p r e d i c t t h e log P values of heteroaromatic compounds

183 Because

of

restrictions

relating

to

the

stability

of

compounds, t h e

number of compounds in some monosubstituted diazine systems were low regular

u

v a l u e was t h e b e t t e r

d e s c r i p t o r of

monosubstituted d i a z i n e systems, whereas t h e f o r t h e disubstituted pyrazines

CJ

I

the electronic and

u

The

effect

for

v a l u e s were b e t t e r

This discrepancy may be a t t r i b u t a b l e t o t h e

r e s t r i c t i o n s in accumulating r e l i a b l e d a t a

Since t h e u s e of r e g u l a r

p a r a m e t e r s i n h e t e r o a r o m a t i c systems is sometimes s u b j e c t

a-type

t o skepticism.

a

uniform t r e a t m e n t of t h e e l e c t r o n i c e f f e c t on t h e r e l a t i v e s o l v a t i o n s h o u l d b e examined with

compound sets including g r e a t e r

numbers

of

having v a r i o u s d e g r e e s of e l e c t r o n i c effect and hydrogen-bonding

substituents behaviors

Outlying b e h a v i o r s of s u b s t i t u e n t s in some of t h e 2 - s u b s t i t u t e d series should also be c l a r i f i e d in terms of experimental physical-organic chemistry

REFERENCES 1 2 3 4 5 6 7 8

9 10 11 12 13 14 15

T. F u j i t a , J. Iwasa, and C. Hansch, J. A m e r . Chem. SOC. 86 (1964) 5175. A. Leo, C. Hansch, and D. Elkins, Chem. Rev. 71 (1971) 525. C. Hansch and A. J. Leo, S u b s t i t u e n t Constants f o r C o r r e l a t i o n Analysis i n Chemistry and Biology, John Wiley and Sons, N e w York 1979. J. Iwasa, T. F u j i t a , and C. Hansch. J. Med. Chem., 81 (1965) 150. S. J. L e w i s , M. S. Mirrlees, and P. J. Taylor, Quant. S t r u c . Act. Relat. 2 (1983) 1. S. J. L e w i s . M. S. Mirrlees, and P. J. Taylor. Quant. S t r u c . Act. Relat. 2 (1983) 100. J. Bradshaw and P. J. Taylor, Quant. S t r u c . Act. Relat., 8 (1989) 279. a)T. F u j i t a , Progr. Phys. Org. Chem.. 14 (1983) 75. b)Y. Nakagawa, K. Izumi. N. Oikawa. T. Sotomatsu, M. Shigemura, and T. F u j i t a , Environ. Toxicol. Chem., 11 (1992) 901. T. F u j i t a and T. Nishioka, Progr. Phys. O r g . Chem.. 12 (1976) 49. M. Charton. Prog. Phys. Org. Chem.. 13 (1981) 119. E. K u t t e r and C. Hansch, J. Med. Chem. 12 (1969) 647. M. Charton. in: N. B. Chapman and J. S h o r t e r (Eds.). C o r r e l a t i o n Analysis i n Chemistry. Plenum Press, N e w York, 1978, pp. 175-268. C. Yamagami. N. Takao and T. F u j i t a , Quant. S t r u c . Act. Relat., 9 (1990) 313 C. Yamagami. N. Takao and T. F u j i t a , J. Pharm. Sci., 80 (1991) 772. 0. Exner, in: N. B. Chapman and J. S h o r t e r (Eds.). C o r r e l a t i o n Analysis in Chemistry, Plenum P r e s s , N e w York, 1978, pp.439-540.

This Page Intentionally Left Blank

QSAR and Drug Design New Developments and Applications T. Fujita, editor 9 1995 Elsevier Science B. 9'. All rights reserved -

185

H Y D R O P H O B I C I T I E S OF DI- TO PENTAPEPTIDES H A V I N G UNIONIZABLE SIDE CHAINS AND C O R R E L A T I O N WITH SUBSTITUENT AND STRUCTURAL P A R A M E T E R S MIKI AKAMATSU and TOSHIO FUJITA Department of Agricultural Chemistry, Kyoto University, Kyoto 606-01, Japan A B S T R A C T : Under standardized experimental conditions, we measured the partition ratio P' in a 1-octanol/pH 7.0 aqueous phosphate buffer system of a large number of zwitterionized di- to pentapeptides composed of amino acids having unionizable side chains as an approximate "molecular" partition coefficient P. The variations in log P' value of peptides were analyzed with free-energy-related physicochemical parameters for the side chain substituents and substructures. The side chain parameters representing the intrinsic hydrophobicity, the steric effect on the relative solvation of functional groups on the backbone, and the conformational potential index derived from the ChouFasman p-turn propensity parameters were shown to be significant. For polar side chains, specific indicator variables attributable to intramolecular hydrogenbond formations and the "polar proximity effect" for augmentations of hydrophobicity observed when polar groups are crowded together were required in addition. The proline residue was shown to participate in the log P' value depending not only upon its location on the backbone but also upon the total number of residues included in peptides. 1.

INTRODUCTION

The hydrophobicity of component amino acids and peptide segments is believed to govern not only the three-dimensional structure of proteins determining their biological functions (1), but also the affiliation properties of their partial domains into hydrophobic biomembraneous phases (2). In addition, peptides and their analogs have been attracting interest as potential drugs (3). The hydrophobicity of drugs is regarded as a highly important parameter to control transport behaviors from their site of administration to their site of action through a number of biomembranes as well as their binding with hydrophobic receptor sites (4). The log P value, P being the partition coefficient of neutral molecules measured with the 1-octanol/water system, has been widely used to represent molecular hydrophobicity (3, 4). In this article, we show that the log P' values, P' being the partition ratio in the system of 1-octanol/pH 7.0 aqueous buffer at 25~ for a number of di- to

186 pentapeptides having unionizable side chains are analyzable with free-energyrelated physicochemical parameters of the side chain substituents of the component amino acids by the regression technique (5, 6). The log P' value under such conditions is believed to be close to the log P of zwitterionized "neutral" peptides. In tetra- and pentapeptides, such conformational factors as the [3-turn formation (7) are shown to contribute to the net molecular hydrophobicity in addition to factors considered for di- and tripeptides. The correlation equation should be able to predict the log P' values of peptides, at least up to pentapeptides consisting of amino acids with unionizable side chains. 2.

H Y D R O P H O B I C I T Y OF PEPTIDES

2.1 Measurement of Partition Ratio Each peptide dealt with here showed a pH-log P' profile taking a "flat" parabolic form (5). The maximum log P' value, which is expected to be observed at the isoelectric point, should be the "true" hydrophobic parameter, log P, for the zwitterionized molecule in which the electric charges are cancelled. Unfortunately, the isoelectric points of most peptides were difficult to measure because of their limited solubility. Because of the flat form, the pHprofile of the log P' value is almost horizontal within the pH range between 5 and 7. We acertained that the log P' measured at pH 7 indeed parallels that measured at pH 6 for a subset of compounds (5). A theoretically reasonable pH-log P' profile with a "flat" parabolic form was not obtained when NaC1 was used to adjust the ionic strength of the acidic to neutral buffer solutions (5). Therefore, the buffer solution should be prepared from sodium hydrogen phosphate and dihydrogen phosphate only. This was considered to be due to the partitioning of the ion-pairs with counter anions. The phosphate anion, probably existing as a mixture of mono- and divalent species under acidic to neutral pH conditions, is perhaps much less hydrophobic than chloride. 2 . 2 Physicochemical Side-chain and Structural Parameters In the course of preliminary analyses, we found that the variations in the log P'(pH 7) value are governed at least by the hydrophobic and steric effects of side chain substituents of component amino acids. As the hydrophobic parameter of side chain substituents, we used the n value of general utility for aliphatic substituents evaluated under conditions free from components such as intramolecular stereoelectronic and hydrogen-bonding interactions (5, 6). For side chains with a polar group or heteroatom, the "intrinsic" aliphatic n value was evaluated from the log P value of related (but not peptidic) compounds in

187 which the polar group is separated by at least two methylene units from a chromophore (8). Our rc value for alkyl side chains is equivalent to that proposed by Hansch and Leo (9) under consideration of the branching factor, being close to that of Fauchbre and Pligka (10). Our ~ value for polar side chains in serine, threonine, methionine, tryptophan, glutamine, and asparagine is more negative than the corresponding value of Fauchbre and Pliska as will be shown later. For the steric effects of side chain substituents except "that" of proline, we used either the E's or E's c parameter depending upon the situation. The E's is the Dubois steric parameter (11). The E's parameter was defined to improve the Taft Es parameter (12). The E's c is the "corrected" Dubois steric parameter related to the original E's by Eq. 1, where n is the number of o~-hydrogen atoms in aliphatic substituents. E's c = E's - 0.306(3 - n)

[1]

The "correction" term in Eq. 1 takes the same form as that for the Taft Es made by Hancock and coworkers (13) to eliminate possible hyperconjugation effects of alkyl substituents on the reference reaction rate from which Es is defined. As indicated previously (14), however, the E's c (improved Es c) value is the parameter not corrected for the hyperconjugation effect attributable to the (xhydrogen atoms of substituents, but that representing not only the steric bulk but also the effect of a-branching. The coefficient of the correction term is fixed as -0.306 in Eq. 1, but values between -0.25 and -0.35 were found to be equally good. The relevance of the use of E's c for the steric effect of aliphatic substituents is discussed in detail in our previous analyses of the log P value of aliphatic amines and the ion-pair formation-partition constant of aliphatic ammonium ions (14). By definition, the bulkier, as well as the more o~-branched the substituents, the more negative the E's c value becomes. For most side chain substituents dealt with in this article, the E's value has been defined (11). The E's values of the indole-3-methyl group in tryptophan, the aminocarbonyl-methyl group in asparagine, and its higher homolog in glutamine were estimated using a highly linear relationship (5) between the E's value and Charton's ~ steric parameter (15, 16). That of the 4-hydroxybenzyl group in tyrosine was taken to be equivalent to that of the benzyl group in phenylalanine. The reference points of E's and E'sC were shifted so that E's(H) and E'sC(H) of the "side chain" in glycine were zero (5, 6, 17). To deal with the conformational effect arising from the possible 13-turn structure in tetra- and pentapeptides, we adopted the 13-turn "potential" index for

188 component amino acids proposed by Chou and Fasman (7). Their 13-turn index is defined statistically for each amino acid in each of the four consecutive positions from the data for 457 [3-turned backbone substructures found in 29 proteins of known sequence and crystallographic structure. As will be discussed later, the logarithm of the 13-turn index, f, for the i-th amino acid in the four consecutive positions, log fi, was regarded as a free-energy-related 13-turn potential parameter of each amino acid. For the inductive electronic parameter of side chain substituents, the Charton civalue was used (16). The relevant parameter sets are listed in Table 1. Factors governing the value of log P'(pH 7) shown in Table 2 were analyzed by the multiple regression technique in terms of the above-mentioned physicochemical free-energy-related parameters for the side chain substituents and indicator variables for particular substructures. T a b l e 1. Hydrophobicity Scale, Steric and Electronic Parameters, and [3-Turn Potential Indices of Amino Acid Side Chains Amino Acid Gly Ala Val

Leu Ile Phe Tyr Trp Met Ser Thr Asn Gln

Pro

na 0.00 0.32 1.27 1.81

E's b E's c c (YI d 0.00 0.00 0.00 -1.12 -0.20 -0.01 -1.60 -1.29 0.01 -2.05 -1.44 -0.01

log fie log fi+l e log fi+2 e log fi+3 e log ft e 0.09 -0.19 -0.19 -0.18

0.00 -0.02 -0.26 -0.54

0.31 -0.37 -0.46 -0.40

0.25 -0.17 -0.13 -0.09

0.19 -0.19 -0.28 -0.24

1.81 1.95 1.20 1.92

-2.12 -1.51 -1.51 -1.47f

-1.81 -0.90 -0.90 -0.86

-0.01 0.03 0.03 0.00

-0.17 -0.07 0.03 -0.10

-0.39 -0.37 -0.10 -0.89

-0.57 -0.17 0.09 -0.10

-0.14 -0.07 0.18 0.30

-0.27 -0.19 0.05 0.00

0.61 -1.49 - 1.18 -1.95 -1.41 0.86

-1.64 -1.09 - 1.04 -1.60 f -1.43 f -

-1.03 -0.48 -0.73 -0.98 -0.82 -

0.04 0.11 0.04 0.06 0.05 -

-0.07 0.14 0.02 0.25 -0.08 0.13

-0.07 0.17 0.09 -0.02 0.01 0.53

-0.85 g 0.12 -0.09 0.33 -0.35 -0.23

-0.14 0.03 -0.03 0.004 0.09 -0.11

-0.19 0.13 -0.00 0.18 -0.01 0.19

a b c d e

From ref. 5 and 17. From ref. 11, unless noted. The reference point is shifted so that E's(H) = 0. From ref. 5 and 14. The reference point is shifted so that E'sC(H) = 0. From ref. 16. Calculated from ref. 7. f See Text. g Not reliable. The corrected value -0.33 was used in Eq. 17.

2.3 Di- and Tripeptides First, we examined the log P' values of di- and tripeptides composed of the nonpolar amino acids; glycine, alanine, valine, leucine, isoleucine, and phenylalanine, in terms of the summation of the side chain ~ value of component amino acids and derived Eq. 2 with an indicator variable Itri.

189

Table 2. Log P and Physicochemical Parameters of Di- to Pentapeptides log P

~~

No. Compounds 1 2 3 4 5 6 7 8 9 10

FL LF FF LL LV VL A1 I1 LI

vv

11

ww

12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

WF WA WL WY LY YL VY

FY YY LM ML MV FM SL PF PL PI FP LP IP FFF GFF FVG FVF FVA LVV LII LVL LAL LLL WGG WFA WWL LLY VFY GFY YLV YVF

Zrr

3.76 3.76 3.90 3.62 3.08 3.08 2.13 3.62 3.62 2.54 3.84 3.87 2.24 3.73 3.12 3.01 3.01 2.47 3.15 2.40 3.12 3.12 2.58 3.26 0.32 2.81 2.67 2.67 2.81 2.67 2.67 5.85 3.90 3.22 5.17 3.54 4.35 5.43 4.89 3.94 5.43 1.92 4.19 5.65 4.82 4.42 3.15 4.28 4.42

E,C(RN) ZE,C(RM)

-0.90 -1.44 -0.90 -1.44 -1.44 -1.29 -0.20 -1.81 -1.44 -1.29 -0.86 -0.86 -0.86 -0.86 -0.86 -1.44 -0.90 -1.29 -0.90 -0.90 -1.44 -1.02 -1.02 -0.90 -0.48 -

-

-0.90 -1.44 -1.81 -0.90 0.00 -0.90 -0.90 -0.90 -1.44 -1.44 -1.44 -1.44 -1.44 -0.86 -0.86 -0.86 -1.44 -1.29 0.00 -0.90 -0.90

0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 -0.90 -0.90 -1.29 -1.29 -1.29 -1.29 -1.81 -1.29 -0.20 -1.44

0.00 -0.90 -0.86 -1.44 -0.90 -0.90 -1.44 -1.29

E',C(Rc) log Sum log f,+q

-1.44 -0.90 -0.90 -1.44 -1.29 -1.44 -1.81 -1.81 -1.81 -1.29 -0.86 -0.90 -0.20 -1.44 -0.90 -0.90 -1.44 -0.90 -0.90 -0.90 -1.02 -1.44 -1.29 -1.02 -1.44 -0.90 -1.44 -1.81

-

-0.90 -0.90 0.00 -0.90 -0.20 -1.29 -1.81 -1.44 -1.44 -1.44 0.00 -0.20 -1.44 -0.90 -0.90 -0.90 -1.29 -0.90

0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Obsd.

-1.17 -1.15 -0.85 -1.46 -2.05 -2.07 -2.60 -1.82 - 1.64 -2.82 -0.27 -0.47 -1.98 -0.73 -1.13 -1.94 -1.75 -2.52 -1.68 - 1.87 - 1.87 - 1.84 -2.53 -1.59 -2.49 -2.07 -2.41 -2.56 - 1.36 -1.76 -1.79 -0.02 - 1.33 -2.33 -0.76 -2.19 -2.10 -1.11 -1.57 -2.03 -0.94 -2.72 - 1.oo 0.36 - 1.34 -1.50 - 1.96 - 1.45 -1.37

Calcd. (Q. 18)

(!3+19)

-1.23 -1.36 -0.93 -1.66 -2.12 -2.09 -2.50 -1.98 -1.77 -2.55 -0.21 -0.56 -1.89 -0.86 -1.14 -1.93 -1.80 -2.36 -1.51 -2.08 -2.01 -1.91 -2.37 -1.58 -2.66

-1.24 -1.38 -0.95 -1.67 -2.13 -2.09 -2.50 -1.98 -1.78 -2.56 -0.22 -0.58 -1.91 -0.87 -1.14 -1.94 -1.80 -2.37 -1.51 -2.08 -2.02 -1.91 -2.37 -1.59 -2.67 -1.95 -2.24 -2.35 -1.37 -1.79 - 1.99 0.04 -1.31 -2.29 -0.72 -2.05 -1.90 -1.19 -1.43 -2.01 -0.97 -2.73 -0.92 0.48 -1.24 -1.38 -1.87 -1.57 -1.28

0.05 -1.29 -2.27 -0.71 -2.03 -1.90 -1.20 -1.44 -2.00 -0.97 -2.71 -0.90 0.48 -1.25 -1.38 -1.87 -1.58 -1.28

190 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

102 103

YGF YYL AYI IYV MLF LSL ISL IS1 SLI SLL FIT LIT IIT LTI TLI TVL PLL LPL LLP IPI FGGF VAAF LLVF LLLV VGFF AVLL IAGF FFFF LLGF LLAF LLLF IlVV IIGF IAAI FFGF VLVL WLLV WGLL YILG FVYF IYIV VFLT MIL1 VMFI PLLL LPLL LLPL LLLP IPGI VPVL VPGV YPGW YPGI GGFVF

3.15 4.21 3.33 4.28 5.07 2.13 2.13 2.13 2.13 2.13 2.58 2.44 2.44 2.44 2.44 1.90 4.48 4.48 4.48 4.48 3.90 3.86 6.84 6.70 5.17 5.21 4.08 7.80 5.57 5.89 7.38 6.16 5.57 4.26 5.85 6.16 6.81 5.54 4.82 6.37 6.09 3.85 6.04 5.64 6.29 6.29 6.29 6.29 4.48 5.21 3.40 3.98 3.87 5.17

-0.90 -0.90 -0.20 -1.81 -1.02 -1.44 -1.81 -1.81 -0.48 -0.48 -0.90 -1.44 -1.81 -1.44 -0.73 -0.73 -1.44 -1.44 -1.81 -0.90 -1.29 -1.44 -1.44 -1.29 -0.20 -1.81 -0.90 -1.44 -1.44 -1.44 -1.81 -1.81 -1.81 -0.90 -1.29 -0.86 -0.86 -0.90 -0.90 -1.81 -1.29 -1.02 -1.29 -1.44 -1.44 -1.44 -1.81 -1.29 -1.29 -0.90 -0.90 0.00

0.00 -0.90 -0.90 -0.90 -1.44 -0.48 -0.48 -0.48 -1.44 -1.44 -1.81 -1.81 -1.81 -0.73 -1.44 -1.29 -1.44

-1.44

0.00 -0.40 -2.73 -2.88 -0.90 -2.73 -0.20 -1.80 -1.44 -1.64 -2.88 -3.10 -1.81 -0.40 -0.90 -2.73 -2.88 -1.44 -3.25 -2.19 -2.71 -2.34 -3.25 -1.92 -2.88

-2.88

-

-2.19

-0.90 -1.44 -1.81 -1.29 -0.90 -1.44 -1.44 -1.81 -1.81 -1.44 -0.73 -0.73 -0.73 -1.81 -1.81 -1.44 -1.44 - 1.44

-

-1.81 -0.90 -0.90 -0.90 -1.29 -0.90 -1.44 -0.90 -0.90 -0.90 -0.90 -0.90 -1.29 -0.90 -1.81

-0.90 -1.44 -1.29 -1.44 0.00 -0.90 -1.29 -0.73 -1.81 -1.81 -1.44 - 1.44 - 1.44

-

-1.81 -1.44 -1.29 -0.86 -1.81 -0.90

0.00 0.00 0.00 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.18 -0.65 - 1.25 -1.24 -0.43 -0.94 0.06 -0.68 -0.48 -1.16 -1.19 -1.14 -0.31 -0.69 -0.20 - 1.28 -1.16 -0.58 -0.51 -0.31 -0.96 -0.99 -0.99 -0.58 -0.89 -0.14 - 1.04 O.0Od 0.54 -0.21 0.53 1.17 0.73 -0.21

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.31 -0.37 -0.46 -0.40 -0.17 -0.40 0.31 -0.17 0.31 -0.37 -0.40 -0.46 0.31 -0.37 0.31 -0.46 -0.40 -0.40 -0.40 0.09 -0.57 -0.40 -0.40 -0.17 -0.40 -0.40 -0.23 0.00a 0.31 -0.46 0.31 0.31 0.31 -0.17

-1.86 -1.38 -2.04 -1.77 -1.03 -2.35 -2.28 -2.64 -1.99 -2.03 -1.95 -2.14 -2.23 -2.30 -1.66 -1.97 -1.64 -1.56 -1.58 -1.65 -1.51 -1.91 -0.25 -0.51 -0.51 -1.74 -1.78 1.63 -0.42 -1.00 0.24 -1.41 -0.99 -2.82 0.17 -1.23 0.23 0.06 -1.49 -0.32 -1.09 -1.32 -0.49 -0.63 -1.06 -0.92 -1.00 -1.18 -1.69 -1.91 -2.83 -1.25 -1.65 -1.40

-2.08 -1.39 -2.09 -1.92 -0.92 -2.21 -2.41 -2.52 -2.09 -1.97 -1.68 -2.10 -2.31 -2.10 -1.93 -2.28

-1.34 -2.22 -0.27 -0.53 -0.99 -1.25 -1.73 1.43 -0.50 -0.77 0.23 -1.35 -0.82 -2.41 0.23 -1.00 0.27 -0.53 -1.59 0.29 -1.25 -1.21 -0.54 -0.49

-1.26

-2.09 -1.38 -2.08 -1.91 -0.92 -2.21 -2.41 -2.52 -2.08 -1.97 -1.68 -2.10 -2.31 -2.10 -1.93 -2.28 -1.89 -1.44 - 1.44 -1.75 -1.36 -2.25 -0.28 -0.52 -1.01 -1.25 -1.75 1.41 -0.51 -0.78 0.23 -1.34 -0.82 -2.42 0.21 -1.00 0.27 -0.54 -1.58 0.30 -1.24 -1.22 -0.52 -0.48 -1.32 -0.88 -0.75 - 1.09 -1.92 -1.81 -2.50 -1.10 - 1.86 -1.27

191

104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 a

VFVGL VGFVF GAALL AFGVF AGFVF LIIGA GLLGF ALLGF IIIIG IVVVI FGAGI FAAAL WGGFV WLFAA IAYWG GLSVL SLAIV YTGFL LVGTF L-enk b M-enk c

6.30 6.44 4.26 5.49 5.49 5.75 5.57 5.89 7.24 7.43 4.08 4.72 5.14 6.32 5.25 3.40 3.72 3.78 3.85 4.96 3.76

-1.29 - 1.29 0.00 -0.20 -0.20 -1.44 0.00 -0.20 -1.81 -1.81 -0.90 -0.90 -0.86 -0.86 -1.81 0.00 -0.48 -0.90 -1.44 -0.90 -0.90

-2.19 -2.19 -1.84 -2.19 -2.19 -3.62 -2.88 -2.88 -5.43 -3.87 -0.20 -0.60 -0.90 -2.54 -1.96 -3.21 -3.45 -1.63 -2.02 -0.90 -0.90

-1.44 -0.90 -1.44 -0.90 -0.90 -0.20 -0.90 -0.90 0.00 -1.81 -1.81 -1.44 -1.29 -0.20 0.00 -1.44 -1.29 -1.44 -0.90 -1.44 - 1.02

-0.11 -0.49 -0.39 -0.37 -0.48 -0.42 -0.48 -0.48 -0.87 -1.01 0.24 -0.63 0.15 -0.98 0.21 -0.45 -0.89 0.36 -0.16 0.28 O.28

0.31 -0.17 -0.37 0.31 -0.17 0.31 0.31 0.31 -0.57 -0.46 0.31 -0.37 0.31 -0.17 0.09 0.12 -0.37 0.31 0.31 0.31 0.31

-0.97 -0.50 -2.55 -0.59 -1.10 -1.65 -0.18 -0.63 -0.97 -0.89 -1.87 -2.23 -0.44 -0.32 -1.47 -1.64 -1.94 -1.18 -1.18 -0.80 -1.39

-0.70 -0.78 -2.33 -0.71 -1.08 -1.36 -0.73 -0.53 -1.30 -1.11 -2.10 -2.02 -0.75 -0.18 -1.12 -1.59 -1.95 -0.97 -1.30 -1.22 -1.57

-0.70 -0.77 -2.32 -0.70 -1.07 -1.36 -0.73 -0.54 -1.31 -1.13 -2.09 -2.0O -0.75 -0.16 -1.12 -1.60 -1.96 -0.98 -1.30 -1.22 -1.57

These parameter terms were not counted, see text.

b [Leu]enkephalin(YGGFL) c [Met]enkephalin(YGGFM)

log P'= 0.804 En - 0.689 Itri - 4 . 4 2 5 (0.217)

(0.417)

[2]

(0.752)

n = 20

s = 0.333

r = 0.892

F2,17 = 32.9

In Eq. 2 and the following correlation equations, n is the number of compounds, s is the standard deviation, r is the correlation coefficient, F is the ratio of regression to residual variance, and the figures in parentheses are the 95% confidence intervals. Itri is zero for dipeptides and unity for tripeptides. In trying to improve the correlation, we noticed that the log P' values of peptides containing 13-branched amino acids with a-branchings in the side chain, such as valine and isoleucine, were more negative than the value calculated by Eq. 2. Since the steric effect of the "crowded" structure of the branched side chain on the relative solvation of the NHCO moiety and terminal NH3 + and COO- groups with partitioning solvents was anticipated to contribute to these deviations, we introduced steric terms into Eq. 2. Among various steric parameters (18) examined, the E's c parameter worked best, yielding Eq. 3.

192 log P' = 1.031 2;~: - 0.778 Itri + 0.521 E'sC(RN) + 0.337 E'sC(RM) (0.081) (0.225) (0.131) (0.188) + 0.335 E'sC(Rc) - 4.068 (0.128) (0.264) n=20

[31 s=0.113

r=0.990

F5,14 =141

R N and RC represent the side chains of amino acids at the N- and C-termini, respectively. RM is for the side chain of the central amino acid of tripeptides. For dipeptides, the E'sC(RM) term is not counted, i.e., E'sC(RM) = 0. This does not mean that the "phantom" central amino acid in dipeptides is regarded as being glycine. The effect attributed to the "phantom" glycine is compensated by the Itri term. Equation 3 indicates that the steric effects of the side chains on the relative solvation with partitioning solvents depend upon their location in the molecule. When including peptides with polar amino acids, no relevant correlation equation was derived unless indicator variable terms for the presence of respective polar amino acids were introduced to give Eq. 4 (Y, W, M, S, and T are one letter notations for tyrosine, tryptophan, methionine, serine, and threonine, respectively). log P ' = 0.960 z n - 0.635 Itri + 0.561 E'sC(RN) + 0.337 E'sC(RM) (0.075) (0.136) (0.096) (0.123) + 0.255 E'sC(Rc) + 0.165 Iy + 0.352 Iw + 0.637 IM (0.097) (0.079) (0.096) (0.149)

[4]

+ 1.665 (Is + IT) - 3.912 (0.219) (0.210) n=59

s=0.138

r=0.982

F9,49 = 1 4 8

Since the slope values for Is and IT terms were very close in the preliminary calculation, they were combined in Eq. 4. Peptides containing proline were not included in Eq. 4, because the E's c value for the cyclic "side chain" of proline is difficult to estimate. Peptides containing glutamine and asparagine were also not included, because their log P' values were too low to measure accurately. The fact that the slope of the zn term is very close to unity shows that the intrinsic hydrophobic factor of the side chains of constituent amino acids contributes to the total hydrophobicity of peptides almost as such after factors attributed to other effects are separated. The positive sign of the E's c terms

193 indicates that the solvation of backbone functional groups with the bulkier 1octanol is less favorable than that with the smaller water as the side chain substituents are bulkier and more c~-branched, resulting in lower log P' values. Equation 4 also shows that the steric effects of the side chains on the relative solvation depend upon their locations in the molecule. The coefficient of the E'sC(RN) term is the highest and that of the E'sC(Rc) term is the lowest, that of the E'sC(RM) term being intermediate. The NH3 + group as the strongest hydrogen donor in the molecule is solvated more effectively than other sites such as CONH and COO- groups with the more basic octanol than the less basic water. The solvation of the NH3 + group with the bulkier 1-octanol favorable for enhancing the log P' value would suffer the steric hindrance from the N-terminal side chain substituent most sensitively. The coefficient of the Itri term, the Itri value taking unity for tripeptides, indicates that the log P' value decreases by about 0.64 with introduction of one more peptide unit into the dipeptide backbone, other things being unchanged. It also corresponds to the rc value of the CH3CONH group. The value -0.64 is, however, considerably more positive than that [-2.17 (9, 19)] expected under conditions without any intramolecular stereoelectronic effects. This could be rationalized by the "polar proximity factor" (9) for the enhancement of hydrophobicity observed when polar groups are close to each other. Change from a di- to tripeptide backbone increases the proximity interaction between two CONH groups. This factor could be evaluated approximately taking the interaction between two CONH groups separated by CH2 as about 1.73 (9), which is close to the real situation, the increase being about 1.6 (-0.6 + 2.2).

I

I

HC~CH2~O~

I C---'-O I NH

I

HC~CH2~O~ I,

,

R~ ~'

C - - ' - O .... H--O~. NH

I

I

HC~CH2~O~ I

... H

R

ROH

C~O

"~

NH

I

"~ _ H o

..... H----O~"

R

I

Fig. 1. Intramolecular Bridging-Solvation of the Hydroxyl Group of a Serine Residue and Carbonyl Group on the Backbone. R is Either 1-Oct or H. Reproduced from ref. 5 by permission of VCH Verlagsgesellschaft mbH.

Indicator variable terms specific to polar amino acid side chains are always positive. The log P' value of peptides with these side chains is higher than that predicted by the hydrophobicity of the side chains and backbone, as well as

194

factors attributed to the steric effect on the relative solvation. For the side chains of serine and threonine, the size of the coefficient is remarkably high. This is probably due to the fact that the hydroxyl group in the side chain and the carbonyl group in the backbone are well positioned for intramolecular bridgingsolvation as shown in Fig. 1. This type of bridging hydration has been observed in glutamine in the crystal structure of human deoxy-haemoglobin (20). Abraham and Leo (21) discussed the possibility that serine and threonine take this type of bridging-solvation structure in rationalizing the side chain rc value of Fauchbre and Pliska (10), which is significantly higher than the value usually used for aliphatic systems. This type of solvation was estimated to make the log P' value 0.6---0.9 unit higher than that of the structure without such "intramolecular" solvation (22, 23). Subtracting the value attributed to the bridging solvation, the size of the regression coefficient for Ser and Thr residues is about 0.8---1.1. The "corrected" regression coefficient value seems to decrease with the number of bonds separating the polar heteroatom on the side chain from the backbone more regularly than the uncorrected value, as shown in Table 3. When the number of bonds increases, the net inductive electronwithdrawing effect of the side chain polar groups on the backbone functional group is gradually reduced. The electron-withdrawing effect of substituents raises the partition coefficient in series of substituted compounds regardless of whether the functional group is hydrogen-donating or hydrogen-accepting (22). Thus, the greater the number of bonds, the lower should be the increment in the log P' value assigned as the polar proximity factor (9). Table 3. Regression Coefficient of Indicator Variable Terms Amino Acid Ser Thr Met Trp Tyr Asn Gin a b c d

Side Chain RegressionCoefficient a -CH2OH 1.665 (0.8--1.1) c -CH(CH3)OH 1.665 (0.8,--1.1) c -CH2CH2SCH3 0.637 -CH2-(3-Indolyl) 0.352 -CH2-(4-OH-Phenyl) 0.165

nb 2 2 3 4 6

-CH2CONH2 -CH2CH2CONH2

3

1.971 d (1.1 ~- 1.4) c

1.337d (0.5,--0.8) c 4

Unlesss noted, the regression coefficient value of indicator variable terms in Eq. 4. The number of bonds separating the polar heteroatom in the polar group from the a-carbon of the peptides. The value in parentheses is "corrected" by subtracting the intramolecular bridging-solvation factor. Estimated from Eqs. 20 and 21.

195 The intercept of Eq. 4 should correspond with the log P' value of glycylglycine where every independent variable is zero. The very good correlation quality of Eq. 4 could be taken to mean that quite a few component factors contributing to variations of the log P' value are almost completely separated from each other. The contributions of the side chains of asparagine and glutamine residues to the log P' value were analyzed indirectly with use of protected peptides as described later.

2.4 Tetra- and Pentapeptides With the above results for di- and tripeptides in mind, we analyzed the log P' values of tetra- and pentapeptides using parameter terms corresponding to those used in Eq. 4, and formulated Eq. 5. log P ' = 1.025 %rt -0.262 Ipent + 0.575 E'sC(RN)+ 0.491 [ZE's C(RM) (0.157) (0.226) (0.205) (0.137) + E'sC(Rc)] + 0.329 Iw + 0.887 IM + 1.772 (Is + IT) - 4.544 (0.335) (0.432) (0.476) (0.670) n =46

s =0.335

r=0.926

[5]

F7,38 = 32.6

Ipent is an indicator variable taking zero for tetrapeptides and unity for pentapeptides. %E'sC(RM) means the sum of E's c parameters for side chains other than those of the two terminal amino acids. Preliminary examinations indi-cated that the steric effect of RM substituents is almost position-independent, so their E's c values were added together. The coefficients of %E'sC(RM) and E'sC(Rc) terms were also so close that they were combined. Equation 5 might be acceptable, but the quality of the correlation in terms of r and s is consider-ably poorer than that of Eq. 4. The Iy term for the tyrosine side chain is insignificant over the 95% level in Eq. 5. The Iw term for the tryptophan is also only justified over the 94.5% level. Moreover, the coefficient of Ipent corresponding to Alog P with introduction of one more peptide unit is significantly more positive than that of Itri in Eq. 4. The intercept is about 0.6 unit more negative than that in Eq. 4, reflecting the difference between the reference peptide series in Eqs. 4 and 5: dipeptides and tetrapeptides. Since physicochemical factors governing the log P' value of lower peptides could at least be involved as factors for tetra- and pentapeptides on the same standards, these discrepancies should indicate that variables other than those used in Eqs. 3 and 4 are required for log P' of tetra- and pentapeptides. We considered that a specific conformational feature such as the 13-turn formation could be a factor required for tetra- and pentapeptides but not for di- and tripeptides.

196 R3 (i+2)

R4 (i+3)

[3-Turns, classified into at least three types, have been observed as regular conformational patterns in regions of backbone chain reversals of globular proteins (24). ~-Turned substructures consist of four consecutive amino acid residues, mostly with hydrogen-bonding formation between the CO-oxygen of the residue at position i and the NH-hydrogen of the residue at position (i + 3). One of the [3-turn structural types (named type I by Venkatachalam) is shown in Fig. 2 (25).

R2

(i+

i)

Fig. 2. One of the [3-turn Structures (Type I) of Peptides. Reproduced from ref. 25 by permission of the Journal of Biological Chemistry.

We assumed that tetra- and pentapeptides exist as an equilibrium mixture of random and [3-turned structures depending upon the ~-turn potential of component amino acids in partitioning solvents, and so the partition of peptides can be depicted as shown in Fig. 3.

1-Octanol Phase: [C~ Water

Phase:

Koct ~ [Coct]~

[Cw]R ~

[Cw][~

Kw Fig. 3. Partition and Conformational Equilibria of Peptides; [C] represents the concentration, and suffixes, R and [3, express the random and ~-turn structure, respectively. Reproduced from ref. 6 by permission of the American Pharmaceutical Association. The net P' value is expressible by Eq. 6. p

t

.--

[Coct]R+[Coct]13 = [Coct]R [ l + K o c t ] [Cw] R + [Cw][3

[Cw] R

1 + Kw

[6]

In Fig. 3 and Eq. 6, Koct and KW are the conformational equilibrium constants in the 1-octanol and water phases, respectively, being reflected by the [3-turn

197 potential of four consecutive amino acids. Di- and tripeptides are unable to take the [3-turn structure and so Koct=Kw=0 in Eq. 6. Thus, Eq. 7 holds for these lower peptides.

[Coct]R

log P ' = log ~ = log [Cw]R

P'R

[7]

P'R is the P' value for molecules with random structures. It has been shown that, the more hydrophobic the environment, the easier is the intramolecular hydrogen-bond formation (22). For oligopeptides, intramolecular hydrogenbonding could lead to the formation of conformationally fixed structures such as [3-turns and o~-helices. Consistent with CD spectra measured in aqueous buffer (pH 7) and 2,2,2-trifluoroethanol (6), the tetra- and pentapeptides studied here were considered to exist almost entirely as random conformers in the aqueous phase, but to take the [3-turn structure in aliphatic alcohols to various extents according to the [3-turn potential of their component amino acids. Thus, such conditions as I>>Kw and l>l for Eq. 9, but the procedure was admissible at least as a first approximation. In Table 2, the log P' values calculated using Eq. 18 are

202 shown for 105 peptides.

2.6

Peptides Containing Proline

Peptides containing proline were not included in the above correlations, since the E's c value for the "side chain" of proline is not easily estimated. By substituting the values of available parameters for peptides including proline such as En, Ipep, log fi+2, and Iturn into Eq. 18, we calculated the summation of these parameter terms and examined the difference, Alog P', from the observed value. The Alog P' value should correspond with the component of the log P' value attributable to the steric effect together with other effects specific to the Pro residue. As shown in Table 4, the effects seem dependent not only Table 4. Alog P' and Indicator Variables on the location but also on the of Peptides Containing Proline number of residues involved. Compounds ~xlogP' Ip(N) Ip(#pep) When the Pro residue is at the NPI -0.683 1 -1 terminus, the Alog P' value is PL -0.648 1 -1

invariably negative, being -0.5 -0.9. At the C-terminus, however, it shows the reverse effect only in dipeptides. For tripeptides without N-terminal proline, the Alog P' is nearly zero. For tetrapeptides, the Alog P' is always negative. We considered that the effect of a Pro residue at a position other than the Nterminus is to lower the log P' value almost regularly with increase in number of total residues from dipeptides regardless of its location. Although the variation patterns of the Alog P' value looked rather

PF FP IP LP IPI PLL LPL LLP PLLL LPLL LLPL LLLP IPGI VPGV VPVL YPGW YPGI

-0.605 0.325 0.531 0.354 0.090 -0.562 -0.128 -0.148 -0.888 -0.407 -0.607 -0.437 -0.123 -0.688 -0.457 -0.509 -0.140

1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0

-1 -1 -1 -1 0 0 0 0 1 1 1 1 1 1 1 1 1

complex, we assumed that they are represented by two indicator variables. The one is for the effect when the Pro residue is located at the N-terminus, Ip(N), and the other is for the effect of the number of residues, Ip(#pep). The values of these indicator variables were set as zero for tripeptides without N-terminal proline, since their Alog P' value is closest to zero. The values of indicator variables are shown in Table 4. With these two additional indicator variable terms for the Pro residue, Eq. 19 was finally formulated for 124 peptides

203 without any significant decrease in the correlation quality. log P' = 0.942 Zrt - 0.582 Ipep+ 0.546 E'sC(RN) + 0.295 [ZE'sC(RM) (0.064) (0.096) (0.089) (0.071 ) + E'sC(Rc)] + 0.516 Iturn + 0.764 log fi+2 + 0.144 Iy (0.172) (0.211) (0.089) + 0.378 Iw+ 0.659 IM + 1.581 (Is + IT) - 0.807 Ip(N) (0.106) (0.165) (0.197) (0.225) -0.346 Ip(#pep)- 3.866 (0.118) (0.190) n = 124 s =0.209

[191

r=0.967

F12,111 =

134

In Table 2, the log P' values calculated by Eq. 19 are also listed. For Leu-LeuLeu-Pro, where no 13-tum formation with intramolecular hydrogen bonding is possible, the Iturn and log fi+2 terms were ommitted in calculating the log P' values. At the protonated amino group of the N-terminus working as the hydrogen donor, the solvation with the more basic 1-octanol could effectively compete with that with the less basic water. Since the number of polarized N+-H bonds in peptides including proline is lower by unity than that in others without cyclic amino acids at the N-terminus, the solvation with 1-octanol is less significant in peptides including proline than that in other regular peptides, leading to lower log P' values. The slope of the Ip(N) term, -0.81, was in the same order as that previously observed (-0.52) for the effect of the decrease in the number of N+-H bonds on the ion-pair formation-partition equilibrium for various aliphatic ammonium ions and picrate in the 1-octanol/water system (14). At positions other than the N-terminus, one of the amide NH sites working as the hydrogen-donor is reduced by replacing the regular primary amino acid residue with proline. By the same token as that for the N-terminal N+-H sites, the reduction of the NH sites would induce reduction of log P'. On the other hand, the steric inhibition effect of the "side chain" of the Pro residue on the hydrogen-bonding solvation of a neighboring CONH or COO- group could be lowered by the cyclization. This reduced steric effect would be favorable to the solvation of the bulkier 1-octanol leading to the augmentation of log P'. For tripeptides, these two oppositely operating factors may be balanced. The positive effect is predominant for dipeptides, but the negative effect gradually becomes dominant for higher peptides with increase in the number of residues. No theoretical rationalization for variations in the balance between these two

204 opposite factors is available at the moment. Measurements of the log P' values of more peptides containing proline at various positions are needed before drawing definite conclusions.

2.7

Peptides Containing Glutamine and Asparagine

Because the log P' value was very low, it was not always easy to measure the value for zwitterionic peptides including Gln (Q) and Asn (N). To understand the effects of these residues on the hydrophobicity of peptides, we measured the log P values of a number of N-acetylpeptide amides containing these residues under conditions equivalent with those for free peptides (data not shown), and formulated Eq. 20 as the counterpart of Eq. 4 (17). log P = 1.044 1;re- 0.570 Itri + 0.237 XE's c + 0.073 Iy + 0.258 Iw (0.047) (0.054) (0.046) (0.075) (0.080) + 1.476 (Is + IT) + 1.162 IQ + 1.753 IN - 2.375 (0.106) (0.121) (0.154) (0.074) n=53

s=0.072

r=0.997

[20]

F8,44 = 8 4 0

In Eq. 20, the E's c terms for side chain substituents are combined into a single XE's c term. This is due to the fact that, in N-acetylpeptide amides, it is invariably the CONH group toward which the side chain substituents exert the steric effect on the relative solvation. The intercept should correspond with the log P value of Ac-Gly-Gly-NH2. Except for these, the corresponding terms are very similar in Eqs. 4 and 20. Although the indicator variable terms for side chains in Eq. 20 are slightly smaller than the corresponding terms in Eq. 4, the correspondence is very good. In fact, for side chains of Ser, Thr, Trp, and Tyr, Eq. 21 was derived, in which RC is the regression coefficient of the indicator variable terms. RC(Eq. 4) = 1.010 RC(Eq. 20) + 0.092 (0.077) (0.082) n=4

s=0.024

[21]

r=0.9997

F1,2=3146

The slope of the side-chain indicator variable terms for Asn and Gln in Eq. 20 was adjusted to conform to the slope for residues in free peptides with use of Eq. 21 and is indicated in Table 3. Indicator variable terms for Asn (N) and Gln (Q) residues are very large. An intramolecular bridging-type solvation

205 between the side chain amide group and the backbone CONH similar to that shown in Fig. 1 is likely to occur in peptides including these residues (20). In fact, they are even larger than those expected from the simple relationshop with the number of bonds between the side chain heteroatom and the backbone after the correction for the intramolecular solvation is made. This indicates that the size of indicator variable terms is also governed by such factors as the number of hydrogen-bonding sites and electronic effect of the polar groups. In any case, by introducing these indicator variable terms in Eq. 18 or 19, the log P' value of free peptides including Asn and Gln should be estimated with considerable accuracy. 0

A N E W E F F E C T I V E H Y D R O P H O B I C I T Y S C A L E O F SIDE CHAINS

3.1

Definition From the results shown in Eqs. 19 and 20, we propose a new effective hydrophobicity scale, ha, for unionizable amino acid side chains as shown in Eq. 22. The na value is defined as the summation of such factors contributing to the "overall" hydrophobicity of each side chain unit as the "intrinsic" hydrophobicity, steric effects on the relative solvation of backbone functional groups, intra-residue hydrogen-bond formation and the proximity polar effect. In Eq. 22, 8 is 0.55 for N-terminal residues and 0.30 for others. The conformational factors are not included since they are attributable to not only the types of amino acid residues, but also their locations in the sequence. Moreover, they are not applicable to di- and tripeptides or to peptides larger than pentapeptides in which other conformational effects such as a - h e l i x formation should be considered. For proline, the nc~ value varies depending upon its situation. = 0.94 [intrinsic n] + ~5E's c + [coefficient of I for each polar side chain and proline]

[22]

The newly defined na values are listed in Table 5. ha(N) and ncdMC) mean the rta values for N-terminal residue and for others, respectively. The value calculated by Eq. 23 with the na value for the nonconformational components is supposed to be the log P' for an imaginary random form. Comparison of the log P'(random) with the experimentally observed log P' should be useful to obtain information on the component attributable to the effect of the conformation. log P'(random) = Y~na- 0.58

Ipep -

3.87

[23]

206 T a b l e 5. Hydrophobicity Scales for A m i n o Acids or Their Side Chains a,b Amino Acid Gly gla Val Leu Ile Phe Tyr Trp Met Ser Thr Asn Gln Pro d

na n (N) (MC) (FP) 0.00 0.00 0.00 0.19 0.24 0.31 0.48 0.81 1.22 0.91 1.27 1.70 0.71 1.16 1.80 1.34 1.56 1.79 0.78 1.00 0.96 1.71 1.92 2.25 0.67 0.92 1.23 -0.08 0.04 -0.04 0.07 0.25 0.26 -0.51 -0.26 -0.60 -0.51 -0.31 -0.22 e e 0.72

Af (NT) 0.0 0.5 1.5 1.8 2.5 c 2.5 2.3 3.4 1.3 -0.3 -0.4 -0.8 c -0.5 c 0.8

Ef (R) 0.00 0.53 1.46 1.99 1.99 2.24 1.70 2.31 1.08 -0.56 -0.26 _1.05 -1.09 1.01

AG (W) 0.00 -0.45 -0.40 -0.11 -0.24 -3.15 -8.50 -8.27 -3.87 -7.45 -7.27 -12.07 -11.77 -

AHS (C) 0.00 0.05 0.43 0.22 0.58 0.34 -0.68 -0.25 0.10 -0.41 -0.37 -0.84 -1.19 -0.56

AHS AHP (J) (KD) 0.0 0.0 0.0 2.2 0.3 4.6 0.2 4.2 0.4 4.9 0.2 3.2 -0.7 -0.9 0.0 -0.5 0.1 2.3 -0.4 -0.4 -0.5 -0.3 -0.8 -3.1 -1.0 -3.1 -0.6 -1.2

AHS A[-Z] (E) (H) 0.00 0.00 0.09 2.16 0.38 4.92 0.37 6.42 0.57 6.67 0.45 7.15 -0.14 3.62 0.21 6.98 0.10 4.72 -0.42 0.27 -0.34 1.31 -0.80 -0.99 -0.85 0.05 -0.23 3.45

a The reference point is shifted so that each value for Gly is zero. The values for Gly are: G(W) = 2.39, HS(C) = -0.34, HS(J) = 0.3, HP(KD) = -0.4, HS(E) = 0.16, and -Z(H) = -2.23. b For symbols, see text. c Estimated from the value in ref. 31. d Not included in regression analysis. e n~(location, number of residues) of proline; n,x(N, 2): 0.35, n~(MC, 2): 1.16, n~(N, 3): 0.00, na(MC, 3): 0.81, ha(N, 4): -0.34, na(MC, 4): 0.46.

For tetra- and pentapeptides, the conformational effect was represented by Iturn (= 1) and log fi+2 terms. Thus, examination of the difference between experimental log P' and calculated log P'(random) should allow us to predict the 13-turn potential parameter of any amino acids included in tetra- and pentapeptides. Although it does not apply to peptides including proline at the moment, this procedure may be extended to higher peptides in which secondary structural factors differ from those included in tetra- and pentapeptides. To estimate the log P'(random) value for partial domains of proteins, we recommend the use of 8 = 0.30 for the RM and RC side chains to calculate each n~ value by Eq. 22.

3.2

Comparison with Various Hydrophobicity Scales for Amino Acids and Their Side Chains Quite a few sets of parameters supposedly representing the "hydrophobicity" scale of amino acid residues have been proposed. Comprehensive lists of these parameters have been reported by Eisenberg (27), Charton (28) and Nakai and coworkers (29). These parameters are defined and/or estimated on the bases of various standards that are not always consistent.

207 They are broadly categorized into three groups. Parameters in the first group are defined from phase-transfer properties similar to that used in this study but with individual amino acids and their derivatives or related compounds. The scales in the second group are based on the probability of finding a certain amino acid residue in the interior of globular proteins relative to the probability of finding it in the surface. The third group is a composite of parameters of the above two types of scales. The values are listed in Table 5 and the relationships with the rta(MC) are drawn in Fig. 4. Fauchbre and Pligka (10) have measured the log P' value of N-acetylamino acid amides with a system of 1-octanol/aqueous buffer (pH 7), from which they defined the rt value of side chains as the difference from that of N-acetylglycineamide. Because their rt value inherently includes factors such as steric effects on the solvation of backbone CONH functions, the proximity polar effect between the side chain polar group and the backbone CONH functions and the internal hydrogen-bonding in addition to the intrinsic hydrophobicity, our rta(MC) value for 13 unionizable side chains was expected to correspond with theirs. Eq. 24 was formulated for this correspondence.

[24]

rt(FP) = 1.254 rta(MC) - 0.010 (0.175) (0.167) n=13

s=0.198

r=0.979

FI,ll = 2 4 8

The well known classic scale, Af, of Nozaki and Tanford (30) is based on free energy of transfer (kcal/mol) of amino acids from ethanol to water relative to that of glycine, for which Eq. 25 was obtained.

[25]

Af(NT) = 1.819 rta(MC) - 0.080 (0.277) (0.265) n=13

s=0.313

r = 0.975

F I , l l = 208

In the original publication (30), the Af values for lie, Gin, and Asn are not given. In Table 5 and Eq. 25, the values for these residues were estimated from the work of Segrest and Feldman (31). Rekker (32) has proposed a scale, f(R), named the hydrophobic fragment constant for each structural fragment. It is estimated from the 1-octanol/water log P values of a number of organic compounds including substructures appearing in the amino acid side chains statistically based on the additiveconstitutive nature of log P. The summation of the fragment constant values, Y_,f(R), for constituent substructures of amino acid side chains is related to rt~ as shown in Eq. 26.

208 Af (NT) [kcal/mol]

n (FP)

00

% I

I

|

i

i

na(MC)

na(MC) Ef (R)

AG (W) [kcal/mol]

N

0

-10 I

~

I

o

|

]

i

na(MC)

i

2

2

na(MC)

AHS (J) [kcal/mol]

AHP (KD)

-% II

u 9

II

I

,

i

.

1

o

-4

i

I

AHS (E)

i

o

na(MC)

2

na(MC)

zX[-Z (H)] 9

I

I

o0

0 i

I

0

i

I

1

|

71;ot(MC)

i

2

1

I

0

|

i

1

|

71:et(MC)

Fig. 4. Relationships of Various Hydrophobicity Scales with the na(MC) Parameter

209 Zf(R) = 1.538 rca(MC) - 0.673 Ip + 0.088 (0.130) (0.179) (0.161) n=13

s=0.137

r=0.995

[26] F2,10 = 5 0 6

The Ip is an indicator variable taking unity for polar side chains of Ser, Thr, Met, Trp, Asn, and Gln. Because the Rekker fragment parameter is estimated from log P values of compounds without structural characteristics of amino acids or peptides, it seems to underestimate the contribution of such factors in increasing the molecular log P value comprised in the regression coefficient of indicator variable terms for polar side chains listed in Table 3. The Tyr residue did not require the value of Ip = 1 in Eq. 26. This is in accord with the fact that the regression coefficient for the Tyr residue in Eq. 19 is very low. The slope of the rta term is considerably higher than unity. This is due to the fact that the Rekker value neglects the participation of the steric effect of side chains on the relative solvation in lowering the log P' leading to overestimation of the effective hydrophobicity. A set of phase-transfer parameters somewhat special among the category has been proposed by Wolfenden and coworkers (33). Their parameter, G(W), is the free-energy of transfer (kcal/mol) of RH, in which R is the side chain substituent in amino acids, H2NCH(R)COOH, from a gaseous to aqueous phase. Again, this parameter is based on the property of molecules without characteristic features of peptides. Moreover, the phases dealt with in estimation of the parameter are drastically different from the systems in which the above three types of parameters are defined. Therefore, their parameter is not expected to be related with rta. Preliminary examinations showed that G(W) is correlated only with the number of hydrogen-bondable hydrogens, I H D , in the polar groups on the side chain. Eq. 27 shows the situation in which the reference point of G(W) is shifted to that of Gly and so AG(W) = G(W) G(W)gly.

[271

AG(W) = - 5.661 IHD - 1.405 (1.146) (1.101) n = 13 s = 1.385

r = 0.957

FI,ll

=

118

The addition of the rta term to Eq. 27 did not improve the correlation. Eq. 27 indicates that the water-affinity or the hydration potential of the side chains is governed most significantly by the number of hydrogens capable of hydrogenbonding. The higher the number, the less hydrophobic is the residue. Recently, Radzicka and Wolfenden (34) suggested that the vapor phase resembles cyclohexane rather than octanol in its lack of polarity.

210 As parameters of the second category, those proposed by Chothia (35) and Janin (36) are well known. Janin has defined his parameter as the free-energy of transfer (kcal/mol) from the inside to the surface estimated from the ratio of mol fractions in buried and accessible states of each residue in globular proteins. The original Chothia parameter (35), the proportion of each residue 95 % buried in globular proteins, has been modified to place it on the free-energy-related background (37) similar to the Janin parameter. These two parameters were of course very well correlated with the slope of 1.128, s = 0.115, and r = 0.978 taking the Janin parameter as the independent variable. Wolfenden and coworkers showed that their parameter G(W) and the Janin parameter are well correlated with a correlation coefficient of r = 0.90 (33). This is not unexpected because Eq. 28 formulated here for the Janin parameter, HS(J), shows that it is also heavily dependent on IHD. [28]

zxHS(J) = 0.210 rca(MC) - 0.422 IHD - 0.012 (0.117) (0.110) (0.140) n=12

s=0.109

r=0.975

F2,9 =88.0

In Eq. 28, the Tyr residue is not included. Its HS(J) value was significantly lower than that expected. In spite of this, the interior/surface preference of amino acid residues tends to be governed in part by the phase-transfer energy between the two liquid phases. Because the general feature of hydrophobicity scales in terms of the freeenergy of transfer is quite different between scales with the two liquid phases and scales with the gaseous/aqueous system or the interior/surface preference, and also because it seems unrealistic to expect that all aspects of the "hydrophobicity" of residues can be summarized in a single manner, quite a few parameter sets have been proposed by combinations of different categorical parameters for each amino acid residue. One of these third-category parameter sets is the hydropathy scale, HP, proposed by Kyte and Doolittle (2). They defined their scale by somewhat arbitral amalgamation of the Wolfenden G(W) and the Chothia parameters. Because both the G(W) and the Chothia parameters are strongly dependent on the IHD, the HP value is of course related with the IHD as well as the bulkiness in terms of-E's c but not with the rca as shown in Eq. 29. AHP(KD) = -2.271 E's c - 3.039 IHD + 0.882 (0.878) (0.552) (0.965) n-13

s-0.656

r=0.976

[29] F2,10-100

211

Another set of parameters has been put forward by Eisenberg and coworkers (37) as the "consensus" hydrophobicity scale, HS(E). In this scale, not only the gaseous/aqueous, [G(W)], and the interior/surface parameters of Chothia [HS(C)] and Janin [HS(J)], but also a phase-transfer parameter between organic and aqueous liquids theoretically evaluated by von Heijine and Blomberg (38) are amalgamated with normalization and averaging. Because the consensus HS(E) scale involves the component for the phase-transfer between liquids, Eq. 30 shows the significance of our rta as a component factor. In fact, Eq. 30 is very similar to Eq. 28 for the Janin parameter. The amalgamation for the consensus parameter seems to correct the outlying behavior of the Tyr residue from Eq. 28.

[30]

zxHS(E) = 0.303 na(MC) - 0.393 IHD + 0.012 (0.106) (0.099) (0.130) n=13

s=0.103

r=0.979

F2,1o = 113

Hellberg and coworkers (39) have examined a number of descriptors for the amino acid residues characterizing chemical, spectral, phase-transfer, and chromatographic properties statistically by using the principal component analysis and extracted a principal component supposedly related to the hydrophobicity. For their scale, -Z(H), Eq. 31 was formulated.

[3~]

zx[-Z(H)] - 3.638 na(MC)- 1.166 E's c -0.103 (0.767) (1.139) (0.986) n=13

s=0.745

r=0.974

F2,10 = 92.9

Depending upon the selection of the original parameters for the amalgamation, the third-category scales are heavily governed by either the hydration potential represented by IHD and/or the phase-transfer property represented by rta. In Eqs. 29 and 31, a negative E'sC term is significant. The more negative the E ' sC , the more "hydrophobic" is the side chain. This reflects the fact that the steric inhibition effect of side chain substituents on hydration of backbone CONH groups works to make the side chains more burried inside the globular proteins for the hydropathy scale in Eq. 29. In Eq. 31 for the -Z(H) scale, the principal component analysis of the amino acid descripters probably extracted the scale as rta- 8E's c (8 = 0.30) in Eq. 22, because there is no backbone CONH function upon which the steric effect of side chains is exerted in single amino acid residues.

212 4.

CONCLUDING REMARKS

The above examinations are believed to show that the hydrophobicity of peptides, at least up to pentapeptides, that is estimated from the partitioning behavior in an alcohol/aqueous system such as 1-octanol/pH 7.0 buffer, can be analyzed and predicted by combinations of well-defined side chain and substructural parameters. The composition of the hydrophobicity scale was rather complex but each component was rationalized physicochemically very well except for the composition attributable to the Pro residue. The extensions of the present approach toward peptides including ionizable side chains as well as higher peptides should be future projects. The rt (rta) value defined here as the "effective" hydrophobicity index of side chains or residues is unique in that it was estimated from the experimentally measured net "hydrophobicity" of oligopeptides existing in solutions as such. Most of the hydrophobicity indices of amino acid side chains so far published are defined from partition or phase transfer parameters of single amino acids or their analogs or calculated from the solvent-accessible surface area of each residue in globular proteins or composites of these two types of indices, as indicated in the preceding section. We examined the relationship between our rta and each of the existing parameters somewhat in detail because we would like to propose our rta value as the standard hydrophobicity index of amino acid side chains as components of peptides. In this respect, it should be noted that a recent publication of Eisenberg and McLachlan (40) indicates that the solvation energy of globular proteins in water is well rationalized not only by the solvent accessible surface area but also by an "atomic solvation parameter" of each atom included in amino acid side chains accessible to water. The simple ratio of molecular fractions in buried and water-accessible states for amino acid side chains is obviously an oversimplification in estimating the hydrophobicity. The atomic solvation parameter assignable to each atom is very well estimated from the phase-transfer free-energy based on values with the 1-octanol/water system rather than a gaseous/aqueous system. Eisenberg and McLachlan proposed that the interior environment of globular proteins is adequately modeled by nonaqueous but amphiprotic liquids. In a more recent publication of Sharp et al. (41), the changes in the partition free energy of component amino acid residues in a 1-octanol/water system corrected for solute-solvent size differences were shown to agree well with the changes in unfolding free-energy of a variety of mutant proteins. These publications seem to support our proposal that our rta value could be used as the standard hydrophobicity scale.

213 REFERENCES

1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

Kauzman, W., Adv. Protein Chem. 14 (1959) 1-63. Kyte, J. and Doolittle, R.F., J. Mol. Biol. 157 (1982) 105-132. a: Hadzi, D. and Jerman-Blazic, B. (Eds.) QSAR in Drug Design and Toxicology, Elsevier Science Publishers, Amsterdam, 1987, pp. 221-297; b: Claassen, V. (Ed.)Trends in Drug Research, Elsevier Science Publishers, Amsterdam, 1990, pp. 73-108. Hansch, C. and Fujita, T., J. Am. Chem. Soc. 86 (1964) 1616-1626. Akamatsu, M., Yoshida, Y., Nakamura, H., Asao, M., Iwamura, H., and Fujita, T., Quant. Struct.-Act. Relat. 8 (1989) 195-203. Akamatsu, M. and Fujita, T., J. Pharm. Sci. 81 (1992) 164-174. Chou, P.Y. and Fasman, G.D., J. Mol. Biol. 115 (1977) 135-175. Iwasa, J., Fujita, T., and Hansch, C., J. Med. Chem. 8 (1965) 150-153. Hansch, C. and Leo, A.J., Substituent Constants for Correlation Analysis in Chemistry and Biology, John Wiley and Sons, Inc., New York, 1979, pp. 17-43. Fauchbre, J.-L. and Pli~ka, V., Eur. J. Med. Chem. -Chim. Ther. 18 (1983) 369-375. MacPhee, J.A., Panaye, A., and Dubois, J.-E., Tetrahedron 34 (1978) 3553-3562. Taft, R.W., Jr., in: Newman, M.S. (Ed.), Steric Effects in Organic Chemistry, John Wiley and Sons, Inc., New York, 1965, pp. 556-675. Hancock, C.K., Meyers, F.A., and Yager, B.J., J. Am. Chem. Soc. 83 (1961) 4211-4213. Takayama, C., Akamatsu, M., and Fujita, T., Quant. Struct.-Act. Relat. 4 (1985) 149-160. Charton, M., Topics Curr. Chem. 114 (1983) 57-91. Charton, M. and Charton, B.I., J. Theor. Biol. 102 (1983) 121-134. Akamatsu, M., Okutani, S., Nakao, K., Hong, N.J., and Fujita, T., Quant. Struct.-Act. Relat. 9 (1990) 189-194. Fujita, T. and Iwamura, H., Topics Curr. Chem. 114 (1983) 119-157. Calculated from the hydrophobic fragmental constants: f(CH3CONH) f(H) = -1.94 - 0.23. The f values were from ref. 9. Fermi, G., Perutz, M.F., Shaanan, B., and Fourme, R., J. Mol. Biol. 175 (1984) 159-174. Abraham, D.J. and Leo, A.J., PROTEINS: Structure, Function, and Genetics, (1987) 130-152. Fujita, T., Prog. Phys. Org. Chem. 14 (1983) 75-113. Leo, A., J. Chem. Soc. PERKIN TRANS. II (1983) 825-838. Venkatachalam, C.M., Biopolymers 6 (1968) 1425-1436. Dickerson, R.E., Takano, T., Eisenberg, D., Kallai, O.B., Samson, L., Cooper, A., and Margoliash, E., J. Biol. Chem. 246 (1971) 1511-1535. Lewis, P.N., Momany, F.A., and Scheraga, H.A., Proc. Nat. Acad. Sci. USA 68 (1971) 2293-2297. Eisenberg, D., Ann. Rev. Biochem. 53 (1984) 595-623. Charton, M., Progr. Phys. Org. Chem. 18 (1990) 163-284. Nakai, K., Kidera, A., and Kanehisa, M., Prot. Eng. 2 (1988) 93-100.

214 Nozaki, Y. and Tanford, C., J. Biol. Chem. 7 (1971) 2211-2217. Segrest, J.P. and Feldman, R.J., J. Mol. Biol. 87 (1974) 853-858. Rekker, R.F., The Hydrophobic Fragmental Constant, Elsevier Science Publishers, Amsterdam, 1977. 33. Wolfenden, R., Andersson, L., Cullis, P.M., and Southgate, C.C.B., Biochemistry 20 (1981) 849-855. 34. Radzicka, A. and Wolfenden, R., Biochemistry 27 (1988) 1664-1670. 35. Chothia, C., J. Mol. Biol. 105 (1976) 1-14. 36. Janin, J., Nature 277 (1979) 491-492. 37. Eisenberg, D., Weiss, R.M., Terwillger, T.C., and Wilcox, W., Faraday Symp. Chem. Soc. 17 (1982) 109-120. 38. von Heijine, G. and Blomberg, C., Eur. J. Biochem. 97 (1979) 175-181. 39. Hellberg, S., Sj/Sstrom, M., Skagerberg, B., and Wold, S., J. Med. Chem. 30 (1987) 1126-1135. 40. Eisenberg, D. and McLachlan, A.D., Nature 319 (1986) 199-203. 41. Sharp, K.A., Nicholls, A., Friedman, R., and Honig, B., Biochemistry 30 (1991) 9686-9697. 30. 31. 32.

QSAR and Drug Design - New Developments and Applications T. Fujita, editor @ 1995 Elsevier Science B.V. All rights reserved

ANALYSIS

OF

PROTEINS

TAKAAKI

ACID

SEQUENCE-FUNCTION

RELATIONSHIPS

IN

N I S H I O K A and JUN' ICHI ODA

Institute

Uji,

AMINO

215

Kyoto

for Chemical

Research,

611, Japan

Kyoto University,

ABSTRACT:

A n e w s t r a t e g y for d r u g d e s i g n is p r o p o s e d , in w h i c h relationships between the function and structure of proteins that are p o s s i b l e t a r g e t s of d r u g s are a n a l y z e d and u t i l i z e d . The f u n c t i o n of p r o t e i n s to r e c o g n i z e the m o l e c u l e s was e x a m i n e d in t e r m s of t h e i r a m i n o a c i d s e q u e n c e r a t h e r t h a n t h e i r t h r e e dimensional structure. Target proteins recognize ligand m o l e c u l e s by f u n c t i o n a l a m i n o a c i d s e q u e n c e s c o r r e s p o n d i n g to chemical substructures of t h e l i g a n d s . The new procedure "Homology Graphing", in c o m b i n a t i o n w i t h the E n z y m e - R e a c t i o n database, could detect sequence segments conserved among a set of sequences of functionally related proteins. Examples of analyses of a m i n o acid sequence-ligand structures showed a great p o t e n t i a l i t y in the lead identification phase in drug design.

1.

INTRODUCTION

Recently,

and-see

drug

complementary

model of

to the

of a t a r g e t

the

binding.

design

approaches

protein For

of

target

protein

and

strategy

finding

the

instance,

site

have been c a l c u l a t e d

advances,

de novo

hundred

be r e a l i z e d

design

in the near

by t r a d i t i o n a l screening.

picoseconds

This

structure-function molecular the

of

(i).

the

supported

Lead-structures

and/or

to

lack

especially

rotations

and

drug

of

by

drug

to

found

large-scale

information

amino

for

technological

is u n l i k e l y

are still

of proteins.

proteins

in

dihydrofolate

these

methods due

static

motions

of 1 f e m t o s e c o n d

structure

relationships of

of

look-

structures

involved

lead

probably

motions

translations

Even w i t h

of a new

relationships,

recognition

simulations like

is

positions

at i n t e r v a l s

future.

beyond

of the m o l e c u l a r

processes

atomic

trial-and-error

chemical

in a t h r e e - d i m e n s i o n a l

dynamic

the

progressed

with

to s i m u l a t i o n

reductase several

has

drugs

acid

on

sequence-

In other words, molecules

of CPK m o d e l s

are

of organic

216 reagents

to find out the

knowledge

of

reaction

relationships, state

"best" rules

no c h e m i s t

structures

of

complementary in

terms

In

this

functions

reactants

proteins,

several are

of o t h e r

relationships structures

chemical 2.

WHY

between

their

of

acid

transition

the

reaction

bases

sequences

must

"molecular of t h e i r

of the

acid

by t h e s e

ANALYZE

proteins:

the the

we are

sequences

that

RELATIONSHIPS,

STRUCTURE-FUNCTION

RELATIONSHIPS

(ligands)

searching

and is,

NOT

amino

acid s e q u e n c e s

their been

in

genes.

in t h e

Protein

Identification

Research

Foundation).

of i n t e r e s t

as targets

crystallographic Laboratory) were

data

rapidly: only

available

number

of

sequence-

THREE-

?

14,372

Protein

Resource

at

DNA

techniques,

M o r e than

90% of the known

from the DNA s e q u e n c e s amino

acid

Sequence the

sequences

Database

National

of

had

(NBRF;

Biomedical

S e q u e n c e data are i n c r e a s i n g

for p r o t e i n s

of drugs and a g r o c h e m i c a l s .

In contrast,

on p r o t e i n s

are

still

in the Protein Data Bank

585

proteins

1989,

NBRF

coordinate

in O c t o b e r

have been registered

biology

have been d e d u c e d

By D e c e m b e r

registered

increasing

molecular

of genes has b e c o m e easy.

for

chemical

2.1 A v a i l a b i l i t y of sequence data progress

as

signals.

to d i f f e r e n t i a t e

sequences;

SEQUENCE-FUNCTION

DIMENSIONAL

by

have

such

of cell

and h o r m o n e s

Therefore,

amino

Proteins

functions

receptors

substrates

chemicals.

between

be defined.

recognition"

and h o r m o n a l

recognized

With

and

sequences

the a p p l i c a t i o n

and transduction

structure relationships.

sequencing

for

three-dimensional

physiological

reactions

in

structures

from t h o s e

amino

We a l s o d i s c u s s

kinds

of e n z y m e s

chemical

biochemical

between

function,

of chemical

interested

ability

the

simulate

in drug design.

the term,

different

catalysis

or

Without

Then, we show how to analyze

relationships.

First,

show

not

and functions.

relationships

We

we

relationships

of

structures

to find

chapter,

the

fitting.

structure-function

can e i t h e r p r e d i c t p o s s i b l e

the

p a t h w a y by energy calculation. examining

of

1989.

entries

known

on

Moreover,

for several proteins with

limited

and

are not

(Brookhaven National crystal

at least

structures

two entries

in the database,

three-dimensional

so the

structures

is

217 actually

less

than

for

drug

available

120. design

interest

for d r u g

localized

in pathogens,

small

design,

quantities,

or

protein

sequences,

"target

(c) h a v e

methods

from

its

the

to

need

for

deduce

sequence, sequence

optimization active

structures,

information

low as

predictions for

relationships

prediction

through

crystal

is

and main-chain

structure

Glutathione

coli enzyme

to

data

on

of

three-

model

of

At present,

the

have

energy

in

the

prediction

We

are

success of

interested

structure

reductase

to

only

-COOH,

catalyzes with

in

not

moiety,

the

from the sequence.

including

two

to

by

the

complex,

to

the

in

the

protein

hydrogen-bonding

misconception

spatial

the r e d u c t i o n

coenzymes,

only at the

where

of

of o x i d i z e d

NADPH

specificity

engineering

and

FAD.

of the E.

(6,7).

The

i/i00 of that to

2'-OH p o s i t i o n

a phosphate

that

orientation

and -NH 2 groups.

enzyme to NADH is only

and NADH d i f f e r

in

sequence

function,

bind

tried to change the c o e n z y m e

adenosine-ribose

of

and e s t i m a t e d

The poor

and p r o t e i n

leads

occurs

glutathione

of and

b e t w e e n the ligand and s i d e - c h a i n

This

such as -OH,

a

a tertiary

molecule,

observed

from NADPH to NADH by protein

NADPH

of

the a v a i l a b i l i t y

interactions

a f f i n i t y of the wild-type NADPH.

as

predictions

drug

(4,5).

assumed

interactions

by p r o t e i n s

P e r h a m et al.

such

W h e n no c r y s t a l l o -

of a p r o t e i n - l i g a n d

generally

groups.

groups

glutathione

in

of a drug to fit into the

between

sequence

a

Even

design.

the

point-to-point

and c h a r g e - c h a r g e

functional

way.

limited

usually roles

structure is recognized by local sequence

molecule

recognition

structure

of the three-dimensional

2.2 Chemical In the

drug

between

has

(2,3).

and

% at most

of

of

of some other protein with

evaluation

protein

the m a t c h

50-60

use

are available,

reliable

the

than

prediction

involve

as a model

in a s a t i s f a c t o r y

is as

structural

ligand

for

less

is increasing.

structure

of the c h e m i c a l

site

secondary values

low

are

weights

a three-dimensional

too

between

far

accurate

however,

so

been

are

for drug design

for related proteins

far

proteins

(a)

larger molecular

graphic data interactions

structures

because

proteins",

data

such as the crystal

a closely related

less

centers and virus coat proteins.

structure

practical template

far

crystallographic

dimensional protein

three-dimensional

are

(b) play important p h y s i o l o g i c a l

those of p h o t o r e a c t i o n Since

The

group

of the

is p r e s e n t

in

218 TABLE

1

Arginines

conserved

in N A D P H - b i n d i n g

NADPH-binding reductases Glutathione reductase E. coli 196 Human 216 Mercuric reductase S__~. aureus 279 S. f l e x n e r i i 298 Trypanothione reductase T__~. c o n q o l e n s e 221

reductases.

*)

F V R K H A P L R S F D M I R H D K V L R S F D m

M Q R S E R L F K T Y D L A R S T L F F R E - D C Y R N N P I L R G F D

NADH-binding reductases Dihydrolipoamide dehydrogenase E. col____!. 202 V E M F D Q V I P S S D Yeast 231 V _E F Q p Q I G A S M D Human 242 V E F L G H V G G V G I *) M o d i f i e d from ref. 6 with p e r m i s i o n of the o r i g i n a l authors. A m i n o acid residues are r e p r e s e n t e d by o n e - l e t t e r symbols. N u m b e r s i n d i c a t e p o s i t i o n s of the first r e s i d u e of each sequence. NADPH, less

but

at

not

the

the

human

ray

analysis

(8,9).

NADH

of

residues recognize residues

NADPH. which with

I) .

might

be

Perham

Arg198

al.

the m u t a n t

positions

was

showed Next,

mutant

less

only to

coenzyme-binding secondary

structures the

site

the

enzyme

of

around

13).

This

requiring

NADH

and

to

in

leucine

mutagenesis. at

the

As two

enzyme,

NADH.

of

NADH

those

site

with around of

reductase

the N A D P H - b i n d i n g type

enzyme

sequence

with

in that

affinity

coli

with

acid

with

arginine

the w i l d - t y p e

glutathione

fold";

the

charges

activity

amino

2'-

to neutral

These

E.

no p o s i t i v e than

the

residues

by m e t h i o n i n e

c a ta l y s i s

by X-

residues

the

site-directed

to N A D P H

human

be

a mutant

the

the

arginine

suppressing

catalytic

"dinucleotide

(12,

dehydrogenases

the

near

replaced

of

charged

dehydrogenases

are

and NADH.

replaced

increased

determined

located

to

compared

In

gy c a l l e d

for

with

catalytic

they

dehydrogenases.

beta-sheet

enzyme

improve

two

NADPH

by u s i n g

slightly

enzyme,

the

charges

strucutre

positively

In o t h e r

constructed were

negative

was

are

concluded

between

side-chains

two

the

residues

unnecessary

et

expected, but

Thus,

and A r g 2 0 4

neutral

NADPH.

were

the d i f f e r e n c e

that

residues

arginine

reductase

are

complex

showed

arginine

these

(Table

glutathione

there

three-dimensional

enzyme-NADPH

of the bound

as coenzyme,

is, The

Results

two

group

that

in NADH.

erythrocyte

side-chains phosphate

in NADH;

2'-OH

the the

other

(i0,ii),

form a topolo-

a beta-sheet-turn-alpha-helixof or

fold

is

NADPH.

found They

commonly showed

in that

219 T A B L E 2 A l i g n m e n t of s e q u e n c e ~ of N A D P H the d i n u c l e o t i d e - b i n d i n g fold.-) NADPH-binding reductases Adrenodoxin reductase Human 151 Octopine synthase Aqrobacterium 8 Malic enzyme Rat 300 Glutamate dehydrogenase Yeast 224 Mercuric reductase S. f l e x n e r i i 276 Glutathione reductase E. coli 174 Human 194 Thioredoxin reductase E. coli 152

and N A D H - e n z y m e s

around

G Q G N V A L D V A R I G A G N V A L T L A G D G A G E A A L G I A H L G S G N V A Q Y A A L K G S S V V A L E L A Q A G A G Y I A V E L A G V G A G Y I A V E M A G I G G G N T A V E E A L Y

NADH-binding reductases Dihydrolipoamide dehydrogenase E. coli 180 G G G I L G Alcohol dehydrogenase Rat 15 G L G G V G_ Lactate dehydrogenase Mouse 25 G V G A V G Glyceraldehyde phosphate dehydrogenase Yeast 7 G F G R I G

L E M G T V L S V V I G M A C A I S R L V M R I

*) M o d i f i e d from ref. 6 with p e r m i s i o n of the authors. Amino acid r e s i d u e s are r e p r e s e n t e d by o n e - l e t t e r symbols. The numbers i n d i c a t e the p o s i t i o n of the first residues of each sequence. dehydrogenases

requiring

GIy-X-GIy-X-X-GIy Gly

is

in the the

replaced E.

by Ala

coli

above

shown

at A l a 1 7 9 ,

mutant

enzyme

have

X is a n y

a highly amino

in d e h y d r o g e n a s e

glutathione

alignments

mutations

NADH

(where

reductase

in T a b l e s Ala183, and

2,

Val197,

finally

they

Lys199,

obtained

the

NADPH

2).

By

further and

a

sequence

while

requiring

)(Table

1 and

conserved

acid),

third

(Ala179

comparing introduced

His200

mutant

in the

enzyme,

Ala179Gly/Ala183Gly/Val197Glu/Arg198Met/Lys199Phe/His200Asp/Arg 204Pro,

enzyme

with

activity

to NADPH.

This

example

phosphate

group,

NADH,

is

phosphate by

group

side-chains

the

structural on

the

interactions

environmental

comparable

illustrates the

recognized

charge-charge

to NADH

enzyme

interactions

and m a i n - c h a i n

important difference not

between

and p o s i t i v e l y of

to that

only

the

fact by

Arg

phosphate

the

2'-

NADPH

and

point-to-point,

negatively

of the

that

between

charged the

of the w i l d - t y p e

charged

side-chains, group

dinucleotide-fold.

2'-

but also with The

the

helix

220 in the

fold

stabilizes

the positive

the n e g a t i v e

helix

by dipoles

tight

turn

that

(14,15). allows

the

The

first

fold

to m a k e

contact with NADH by van der Waals 2.3 Loops are responsible The p e p t i d e

segment

does not always the

a

determined antibody

loop

light

affinity the

of

DNA-binding that,

in

crystallographic

a strict

molecular

Gly

six h y p e r v a r i a b l e

as a s t r u c t u r a l

of

recognition

structure

proteins

(17),

cases,

analysis.

recognition

sites

loops

unit

six

(18,19).

for m o l e c u l a r

For example,

an

the

three-dimensional

structure

forming the loops.

loops

but

to

the

chemical

called the of the heavy

specificity

proteins

synthetase

are r e l a t e d

acid

of

identified

in ras

amino

and

structures

are also

recognition

so

be

of

and tryptophan

of the f l e x i b l e

is

cannot

by the

Loops

such as

but

it

site,

The

are g o v e r n e d

(20), glycogen phosphorylase(21), functions

loops.

a

steric)

(16).

The a n t i g e n - b i n d i n g

consists

form

(less

for m o l e c u l a r

some

by

side of the

residues

a close

region located in the variable domains

chains,

of the b i n d i n g

The

two

interaction

responsible

of its antigen.

hypervariable and

coenzyme

for m o l e c u l a r r e c o g n i t i o n

structure

by X - r a y

makes

structure

of the

take a fixed t h r e e - d i m e n s i o n a l

helix-turn-helix

flexible

charges

charge that is induced at the N-terminal

not

(22).

to the

sequences

2.4 Sequence segments of functional importance are conserved The

related locally

amino

to

each

similar

recognize

and

substitutions importance protein

acid

sequences

other

in

terms

in the regions bind

their

of

fatal

proteins

where

ligand

by r a n d o m m u t a t i o n s

cause

of

their

that

polypeptide

is not inherited.

are

only

fold

Amino

at the p o s i t i o n s

that are conserved among the proteins

closely

chains

molecules.

loss of the p h y s i o l o g i c a l

and this mutation

are

functions,

to

acid

of f u n c t i o n a l

f u n c t i o n of the

Therefore,

are sequences

sequences

of functional

importance. Here,

we

"dinucleotide

briefly fold"

dehydrogenases. evolution

refer

to the b i o l o g i c a l

reasons

is conserved among the sequences

According

of proteins,

to a r e c e n t

the gene

theory

why

the

of different

of the m o l e c u l a r

of a n e w p r o t e i n

evolves

not by

r a n d o m mutations

of the gene of some other protein with different

function,

"exon-shuffling"

but by

(23-26).

In the exon-shuffling

221 theory, acid

exons

coding

residues

rearrangement ferred for

from

of

similarities separate

are

acid

genes,

identical

in

to

residues

function.

same

ancestral

each

other.

sequence, by

find

random

the

not

only

Then

proteins

after begin

mutations. of

the

importance

of m o l e c u l a r

30-50

exons

a new

exons

they

boundaries

of

are

gene

show

trans-

that

codes

whose

genes

local

sequence

divergence

from

amino

of e v o l u t i o n a r y

form

exon would Just

but

of p h y s i o l o g i c a l

a unit

to

the two d u p l i c a t e d

substitutions

difficult are

with

composed

to be a unit

duplications,

and m i x e d

novel

the

segments

supposed

By g e n e

genes

with

inherited

are

genes.

other

a protein

have

sequence

in length

into

the a n c e s t r a l

to

accumulate

With

time,

two exon

nucleic

it b e c o m e s

duplicated

exons.

But,

conserved.

Thus,

exons

are

evolution,

but

also

of

protein

function. 2.5 M o t i f s

are too small

The p o s i t i o n a l different

proteins

are

GIy-X-GIy-X-X-GIy (27,28)

and

motifs.

functionally

to b u i l d only

actions

a peptide

a structural

one

a bound

motif

is u s u a l l y

tif.

Protein

beta-sheets of

five

lowing

supported

secondary

but by the

the m o t i f

composed in t h e called

of

ras the

secondary

27 r e s i d u e s and

including

molecule,

not

of

the

loop"

(35,36).

by w h i c h

is too

Although

than

does the

of

by Chou

and

composed

and

preceding

Fasman

and

fol-

GIy-X-GIy-X-X-GIy

hand,

the

in a l o o p not

motif

a

the mo-

sequence

the m o t i f

and

inter-

folding

"dinucleotide-binding is

short

alpha-helices

local

of

detected

direct

longer

supposed

kinase

Thus,

make

as

On the o t h e r

adenylate

(30),

peptide

far

the

in g r o u p s

a motif

sequences

For e x a m p l e ,

and

recognition.

such

The

zipper

well-known

successfully

the

by a short of the

are

motifs

only

by a s e q u e n c e

(15).

leucine

developed

are

among

patterns).

proteins,

in a motif

as o r i g i n a l l y

"glycine-rich structure

been

residues

find

for m o l e c u l a r

is a p a r t

protein

have

to

residues

"context"

(34).

in d e h y d r o g e n a s e s

DNA-binding

structures

are d e t e r m i n e d

conserved

(or fingers,

methods

ligand

are

serine-proteases

chain

or six r e s i d u e s

(32,33),

of

acid

acid

unit

that

dehydrogenases,

of

unit

or two amino with

(29)

proteins

three-amino

However,

motifs of

several

related

of

(31).

called

sequence

Recently,

patterns

of residues

sequence

zinc-finger

GIy-X-Ser-X-GIy

as a s t r u c t u r a l

patterns

make itself

fold"

same m o t i f structure any

fixed

does

not

222 have

any

fixed

assigned

secondary

depending

is involved.

We have to search

than that of the motifs 3. H O M O L O G Y

GRAPHING:

FUNCTIONAL

If

we

METHOD

could

find

of

molecules

containing

these

nition

common

paring

of a c o m m o n

aligned

There

of

common

segments

similarity addition, of

low

computer

-

may

acid

3.1

amino

be as

(or s u b s t r u c t u r e ) .

as c o n s e r v e d

acid

segments in

to a l i g n m o r e

been

developed

available

by w h i c h among

for

a

set

are

than

usually and

three of

sequences

as

their

In

sequences

although pairs

we d e v e l o p e d regions

of

such

sequences

in p o s i t i o n .

alignment

conserved

acid

length,

(37,38),

Recently,

by com-

in d e t e c t i n g

30 % i d e n t i t y

alignment).

(39,40).

Such

These

sequences

a set of a m i n o

residues

20-

method

are

difficulties

within

low as

has

sequence

of

many

se-

a method

within are

a given

detected

Homology graphing Homology

lative (target

local

graphing

Window

of an a m i n o

and

sequence

from the N H 2 - t e r m i n a l along

eral

The

residues.

step is d e f i n e d 3.1.2 search

acid

as segment-i

with

sequence

segments:

The

stepwise

segment (Figure

against

the cumu-

to be a n a l y z e d

sequences. target

sequence

with

is

a window.

at intervals

in the w i n d o w

of sev-

at the

i-th

i).

of h o m o l o g y value:

is p e r f o r m e d

graphically

to the C O O H - t e r m i n a l

the sequence

sequence

Calculation

is a l i g n e d

and shows

to a set of r e f e r e n c e

The w i n d o w moves

ment-i

calculates

similarity

sequence)

3.1.1 scanned

ity

the

recognize

structure

Graphing"

quantitatively

among

can

chemical

proteins.

programs

OF

present that

longer

(or s u b s t r u c -

related

(pairwise

are

for the recog-

segments

40

that

SEGMENTS

proteins

chemical

it

structure

no p r a c t i c a l

"Homology

SEQUENCE

is

should be r e s p o n s i b l e

certain

similarity

quences amino

20

segments

role

with which

segments

however,

sequence

as

a common

functional

sequences with each other.

functionally

short

its

commonly

related

c o u l d be d e t e c t e d

are,

conserved

TO FIND

segments

functionally

and

of the s e q u e n c e s

sequence

only.

IMPORTANCE

sequences ture),

structure

on the c o n t e x t

For segment-i,

a reference

one of the r e f e r e n c e

sequence sequences,

similar-

set.

Seg-

sequence-

223 Window

NH2-Terminal

COOH-Terminal

i

Target sequence

|

Segment-i

Similarity search Reference sequences m

u

Sequence 1 - - ~

score-i,1

Sequence j - - ~ score-i,j

m

Sequence n - - - ~ score-i,n ~

Homology value of segment-i in Score-ij { if score-ij>Maxd } Score-i1 Fig.

I.

H o m o l o g y graphing.

j, by u s i n g

found,

IDEAS s y s t e m

the d e g r e e

calculated

dent on the amino factors.

limit

acid

If score-ij

is not saved.

is slightly,

composition

is h i g h e r

similarity),

segment-i

than the threshold value.

(42).

length

than a given

the v a l u e

The degree of

of

segment-i.

threshold

(a lower

If not,

it

sequence-(j+l),

and saved if score-i(j+l)

This process

depen-

as to these two

is saved.

reference

is

is

is higher

is repeated until all the

sequences have been compared pairwise with segment-i.

The sum of the score-ij

number

(score-ij)

but significantly,

and the

is a l i g n e d w i t h

then similarity is calculated reference

for the a l i g n m e n t

the v alue is c o r r e c t e d and n o r m a l i z e d

of d e t e c t i n g

Next,

When the best local a l i g n m e n t

from the amino acid mutation data

similarity thus calculated

Therefore,

(41).

of s i m i l a r i t y

of r e f e r e n c e

value of segment-i

(from j=l to n, where n is the total

sequences)

[Equation i].

saved

is d e f i n e d

as the h o m o l o g y

224 H o m o l o g y value of segment-i Score-ij

J

{ if score-ij

The h o m o l o g y v a l u e similarity

and

number

homology

of

of

alignments

3.1.3

segment-i segment.

is

until

showing

calculated

the

for

[i] in the d e g r e e

higher

of

similarity

is r e p e a t e d at each step

COOH-terminal.

each

segment

in

Thus,

the

the

target

Graphing:

To show g r a p h i c a l l y ,

the h o m o l o g y value of

is p l o t t e d

against

at

the

residue

By v a r y i n g three p a r a m e t e r s

movement

}

increase

This p r o c e s s

the w i n d o w

value

sequence.

> threshold

increases with

than the t h r e s h o l d value. of m o v e m e n t

=

of the window,

and t h r e s h o l d

we can detect any sequence

the

center

(window size,

for d e t e c t i n g

segments differing

of

the

step size of

similarity),

in length and simi-

larity. 3.2 H o m o l o g y graphing of glutathione reductase Here, homology human

we

show

graphing.

glutathione

an e x a m p l e The

reductase,

u n d e r the e n t r y name of RDHUU the c r y s t a l composed NADPH-

structure 293),

to 478) domains enzyme

FAD-

central-

(43,44).

of 1.54-2 A

as a t a r g e t Three sequence

sequence

registered

(from r e s i d u e (294 to

analysis acid

364),

using

sequence

of

X-Ray analysis

of

in t h e

(478 residues).

(8-10).

NBRF

database

that this enzyme is 19 to r e s i d u e and

157),

interface-

structures

sequences

(365

of the

includes

those

coenzyme;

19

sequence

are c o m p o s e d

of the F A D - r e l a t e d

reductase

could

detect

are

prepared

from

the

the s e q u e n c e s

NBRF of the

that require NADPH or NADH as a coenzyme; enzymes.

enzymes

of 14 FAD-related

of the s e q u e n c e s

for the c o n t r o l

sets

NAD(P)H-related

27 sequences

graphing

selected

for coenzyme binding.

The first one c o m p r i z e s

enzymes

of

The enzyme was therefore

to test how h o m o l o g y

reference

database.

NAD(P)H-related

enzymes

amino

The three-dimensional

the segments of importance

thione

sequence

the

c o m p l e x e d w i t h FAD and NADPH have also been a n a l y z e d at a

resolution

30

the

is

of the enzyme r e v e a l e d

of four domains:

(158 to

of

target

requiring experiment;

not requiring NADPH,

that

enzymes.

The

second

require

set

FAD as a

These two sets

f u n c t i o n a l l y r e l a t e d to the gluta-

both N A D P H and FAD. sequences NADH,

The third

set is

of n u c l e o t i d e - n o n r e l a t e d

or FAD.

This

set is to detect

225

omain

200

(a)

tO > Cn 0

S

100

f'

100

200

300

400

Residue number

500

NADPH-domain

150

(b) g

100

qJ

ii-,

o 0

E o 50

-r

100

200

,

300

I i

400

Residue number

Fig.

2.

H o m o l o g y graphs of human g l u t a t h i o n e

500

reductase.

A n a l y t i c a l conditions: w i n d o w length = 50 residues, step size = 5 residues, and threshold = 45. R e f e r e n c e sequence sets are ( ) F A D - r e l a t e d and (---) n u c l e o t i d e - n o n r e l a t e d enzymes in graph (a) and ( ) N A D ( P ) H - r e l a t e d and (---) n u c l e o t i d e n o n r e l a t e d enzymes in graph (b). M o d i f i e d from Ref (39) with permission, C o p y r i g h t 1989, A m e r i c a n Chemical Society.

226 the

regions

similar

binding.

A homology

with

a reference

major

peak

130-150,

when

by

graph

170-250,

the

66,

129,

130,

localized

the

other

331,

domains,

peak regions With

homology 245-330. 337,

339,

370

(8-9).

regions

interacting

tively,

as r e f e r e n c e for

and FAD.

tool to detect

cal structures.

4.1

enzyme

cal

combination

unit

of

homology

( i0 ) .

as

These

19 to

The

but

are

51,

two m a j o r peaks

primary 197,

contact

198,

201,

except

all

with

the

the

extracted

chemical

protein

sequence

ligand

at

and nicotinamide

NADPH

molecule

(substructures).

structure

using

respec-

are those

to p r o v i d e

of

a

and chemi-

OF S E Q U E N C E - C H E M I C A L

glutathione

the

290,

structures

sequences

structure recognized

with

and

the bound

224,

enzymes,

is b e l i e v e d

FOR A N A L Y S I S

of h u m a n

the

of

the

370 are e x t r a c t e d

and F A D - r e l a t e d

graphing

on

in the

at 190-245

218,

not

usually

enzymes,

extracted

The regions

57,

are

in the homology graph.

NAD(P)H-related

acid

FAD in

spread

reactions

residues

at

reference

residues

157),

of c a t a l y t i c

between

DATABASE

of m o i e t i e s

chemical

467

recognition

phosphodiester, of

50,

the

with the bound

identified

successfully

RELATIONSHIPS

structure

31,

195,

relationships

interacts

phosphate,

amino

(residues

sequences.

the

complex

The

been

enzymes

Units of chemical In the

the

set.

for

appear

with the bound NADPH and FAD separately

Thus,

ENZYME-REACTION

STRUCTURE

sequence

in the graph.

graphs

sets of N A D ( P ) H - r e l a t e d

responsible

significant

graph

All these residues

regions

homology

as

2a) gave one

are

the

interactions

of

(Figure

These

that make

assigned

reductase

410-460.

2b) showed

residues

been

of g l u t a t h i o n e

peaks

sites

set

to

small

in

segments)

(Figure

The

These

4.

and

coenzyme-

Other

of d o m a i n s .

a reference

graph

have

and

have

because

(conserved

as conserved

NADPH

enzyme

in the F A D - d o m a i n

are on the b o u n d a r i e s

NADPH

and

the

enzymes

80.

peaks

which make primarily complex

related

sequence

50 to

300-340,

with

FAD-enzyme

not

set of F A D - r e l a t e d

nucleotide-nonrelated the

of the

at r e s i d u e s

compared

residues

chance,

recognizable

by proteins

reductase adenine,

moieties.

is

NADPH,

ribose,

3'-

The chemi-

recognized

This by

with

suggests

proteins

as

that is

a a

a

227

I

o

o

', O - P - - O - P - O

/ Fig. into

3. Various possible ways of dividing the structure of NADPH substructures.

substructure twenty).

composed

of

several

atoms

(probably

less

than

The size of substructures recognized by proteins would

be limited by the length of the sequence segments coded by one or two exons. The c o n s e r v e d

graph

of

sequence

glutathione

regions

reductase

detected

are

the

in the h o m o l o g y

sequence

responsible for the recognition of the substructures the NADPH molecule.

segments

contained in

To find the conserved sequence segments for

the r e c o g n i t i o n of the p h o s p h o d i e s t e r moiety, we have to compile a reference

sequence

dehydrogenases,

set

including

but also synthetases,

the

sequences

kinases,

p h o s p h o d i e s t e r m o i e t y is c o m m o n l y p r e s e n t NADPH,

NADH,

substrates

chemical

FAD,

structure

sequence-chemical 4.2

ATP,

and

of these enzymes.

GTP,

relationships

Enzyme-Reaction

only

in the s t r u c t u r e s

are

the

cofactors

the p r o t e i n

we are a n a l y s i n g

"substructure" relationships.

The

of

and

sequence-

are a c t u a l l y

database

There are many possible ways of dividing the chemical struc-

ture of NADPH into substructures tures

which

Therefore,

of not

and ligases.

(Figure 3); from small substruc-

such as -OH and -NH 2 to large ones including the adenosyl-

phosphate problems

moiety,

of

evolutionally

which

and

their

combinations.

substructures

significant,

are recognized by proteins.

are

Here

arise

physiologically

and how many d i f f e r e n t

the

and

substructures

228 /// ENTRY NAME

EC 6.3.1.2 Glut amat e-ammoni a ligase Glu tamine S y n t h e t a s e Lig a s e s bonds For m i n g c a r b o n - n i t r o g e n (or amine) ligases Aci d - a m m o n i a (am i d e s y n t h a s e s ) L-G l u t a m a t e : a m m o n i a ligase (AMP-forming) ATP + L - G l u t a m a t e + NH3 = ADP + O r t h o s p h a t e + L-Glutamine ATP L-Glutamate NH3 ADP Ort h o p h o s p h a t e L-G lutamine L-M e t h i o n i n e s u l f o x i m i n e L-2 - A m i n o - 4 - ( h y d r o x y m e t h y l p h o s p h i n y l ) b u t a n o a t e AJEBQT AJAIQ AJZJQ2 AJAAQ AJE CQ A24714 A05079 A05097 A23970 AJF BO A22 947

CLASS

SYSNAME REACTION SUBSTRATE PRODUCT INHIBITOR NBRF-ENTRY ///

Fig.

4. To

Contents study

database amino This

these

called

acid

types

problems,

we

contains

including

the

their

structure

common

as

classified

by

structures

of s u b s t r a t e s ,

NBRF

inhibitors sequence The

base of

collected

in the

entries

collected

enzymes

by July

1991.

with

each

of

known

2,477

version-up The

the

each

Union

products,

and

is

enzymes.

about

We

IUB

keep

entry

41.5 the

%

codes

the

datanumber

for

and

number

database

in the

The

5,864

a name the

effec-

in the N B R F

was

of

reaction

Databank.

Database.

gave

(46),

45).

the names

of B i o c h e m i s t r y ) ,

registered

1984

(40,

4):

activators,

Protein

database

the

in

of

and E C - n u m b e r s ,

cofactors,

our

a

analysis

1,027

EC-number of

enzymes

biochemically

updated

with

the

of the NBRF database.

total

Enzyme-Reaction with

Since

enzymes

sequences

characterized

in

construct

(Figure

Enzyme-Reaction

NBRF

for

names

the e n z y m e s

to the

relationships

and the B r o o k h a v e n

of all

for

items

(International

reaction

database

entries

are

and

IUB

started

Database

following

chemical tors,

database. have

Enzyme-Reaction

sequence-chemical

database

of e n z y m e s

of E n z y m e - R e a c t i o n

number

updating

compounds

of

Database are

of

the

stored

chemical was

compounds

1,554

database. by

molfile

in

July

The

registered 1991

chemical

format

and

in

the

increases

structures

(Molecular

of

Design

229 Ltd.,

San

MACCS

system

format

Chemical search

Leandro,

are

stored

Chem

A,

of

32

FAD,

Software

coordinates The

substructures.

ring

is

system,

form a n e w

as

compounds

substructures

database.

the

all

the of

into

found

hetero the

another This

now

substructures

to

System.

datafile by

atom

the

hetero

database. atom

are

result

only

connected,

rules

those

in a they

2,764

that

other

of

a

to a set of

project,

out of the

apply

and if

bonds,

(3)atoms

suggests

listed to

(2)

by multiple

these

in

their

substructures in

research

trying

CONCORD

Software

a

substructures

(49). are

using

three-dimensional

substructure,

can be a u t o m a t i c a l l y We

by

to

compounds

indexed

(i)

Pomona

substructure,

When we a p p l i e d in

the

possible

a

a Med-

started

of

registered

form

to the

substructure.

are

follows: it

have

a substructure

list

to

their

database

to

We

in the M e d C h e m

if two or more

were

store

using

Project,

3100. database

and

structures

is i n c l u d e d

different

reduce

have

connected

(4)

substructures 400

the

attached

atom

gives

from m o l f i l e

space

structures

construct

in

substructures

atom

Reaction

to

chemical

atoms

carbon

4,733

We

in the

define

carbon the

is

structures

Chemistry

on V A X s t a t i o n

at Austin)

step

for

substructure-

database

to save disk

Enzyme-Reaction

compounds

on a

Institute

acyl-derivatives

chemical

(Medicinal

CA)

the

next

hydrogen

the

(47,48)

into a THOR d a t a b a s e

the

included We

format

of Texas

the

including

three-dimensional

in

(University

in the

For e x a m p l e ,

against

molfile

and s u g a r - n u c l e o s i d e .

System

the

registered

which

compounds

Claremont,

generate

ester

in

related-enzymes

on F A C O M - 3 8 0

translating

SMILES

structures

of the

University.

NAD(P),

we are

into

College,

Kyoto

chemical

EC-numbers

pyrophosphate

list

Now,

The

the

installed

Research,

Coenzyme format

(MDL)

with

output

CA).

with

about

Enzyme-

rules

to

biological

significance. 5.

APPLICATION STRUCTURE

Previously, tures

of

drugs

OF S E Q U E N C E - S U B S T R U C T U R E

we

showed

supposed

sequence

similarity

segments

detected

substructure

RELATIONSHIPS

IDENTIFICATIONS

and in

to

our

strategy

interact

homology the

relationships

with of

be u s e d

identify target

graphing

analysis could

to

(39,

amino

TO

lead

proteins 50). acid

as f u n c t i o n a l

LEAD

strucusing

Sequence sequencetemplates

230 that

specifically

a sequence a

region matching

protein,

sequence

the

to

with

substructures. listed,

many

a high

be a b l e

chemical

that

combination

of

some

recognizes chemical of

constituting

structure

but

the

modifications

corresponding the

target

suggested

together three

for

on

together.

a

as

For

These

This

combinations

of

structure 5).

phosphate

to

called an

broad

is

more

binding

bind

substrate

be

of

which

"effector"

has

no

a

ligand

new

"modulator".

no

segment

by s c a n n i n g

on

accepts

and

oxidized

from

site

with

various

compound,

and

binding

site

by c y t i d i n e

tri-

to the binding

site

similarity

CTP binds

are

either

binding

A

the

the

structural

nicotineamide,

to

with

by the

structure

site

is i n h i b i t e d

from that of aspartic

or

of

separately

FAD,

of

structure

reductase

a broad

binding

structural

of the enzyme.

domain

to

so

lead

substructures

or

NADPH,

alloxan,

carbamyltransferase

(CTP),

When

All

using

strictly

of the substrates.

a new

by

of the

is d e t e c t e d

site

of

of

A protein

Part

drastic

the

the

than two c o m p o u n d s

us to construct

composed

may

part

glutathione

substrates,

for

interactions

design.

templates,

give

on the w a y

molecule.

by the protein.

substrates

the

chemical

latter

set

substructures

structures

somewhat

example,

substructures

the substrate

on a d i f f e r e n t

its

candidates

lead

structure

is

the

structures.

be r e q u i r e d

substructures the

a

substructures

is r e c o g n i z e d

recognizes

prompts

moieties,

Aspartate

acid,

with

single

binding-affinity

(Figure

lead

usually

compounds

cysteine

the

obtain

lead

of

ligand

The

accept

to c e r t a i n

sites.

glutathione.

whose

may

sequence

ligand

is not.

protein

are as f o l l o w s .

through

the

not to be recognized

An e n z y m e

different

of the

the rest

ligand

find

relationships

its ligand molecule

substructures

as

should

to

The

of

template

containing

of these

the

set

structures

strategies

sequence-substructure

protein,

to i d e n t i f y

of substructures.

Additional

could

combinations

constraint

the

of the t a r g e t p r o t e i n

we

a given

by

compounds

the s e q u e n c e

Among various

different

structure

to

When

in the sequence

by the p r o t e i n .

templates,

combinations

is found

of the leads.

characterized

affinity

By s c a n n i n g

various

we w o u l d

possible

a template

be r e c o g n i z a b l e

show

substructure.

drugs

substructures

substructure

would

expected

characterize

acid

to a s p a r t i c

(51,52).

Since proteins

of

CTP is

interest

231

O I

O

I

NH

HO

j....

-\---/----I

t/),

may

have

scanning

templates,

known The

a binding

the target

L

.

.

.

.

.

.

HN

I

for

sequences

cases,

O

research

Research

the M i n i s t r y

was

from the three substrates

not well

an u n k n o w n

with various

we may find new binding

present

sites

supported

on Priority Areas,

of Education,

oll

HS

, I I

ligands.

Scientific

0

O

in most

site

0

2

Fig. 5. C o m b i n a t i o n of substructures gives n e w lead structures. are,

o

N ~NH

H3C H3C

for drug d e s i g n

o

characterized,

effector

molecule.

conserved

sequences

for compounds

by

a

"Genome

Science and Culture

they

as

other than

Grant-in-Aid

Informatics",

of Japan.

By

for

from

REFERENCES

1 2 3 4 5 6 7 8 9

U.C. Singh, in: The Third Alliant C h e m i s t r y Colloquium in Tokyo, 1989. T.L. Blundell and M.J.E. Sternberg, Trends Biotech., 3 (1985) 228-235. T.L. Blundell, B.L. Sibanda, M.J.E. Sternberg, and J.M. Thornton, Nature, 326 (1987) 347-352. W. Kabsch and C. Sander, FEBS Lett., 155 (1983) 179-182. K. Nishikawa and T. Ooi, Biochem. Biophys. Acta, 871 (1986) 45-54. N.S. Scrutton, A. Berry, and R.N. Perham, Nature, 343 (1990) 38-43. S. Greer and R.N. Perham, Biochemistry, 25 (1986) 2736-2742. E.F. Pai, P.A. Karplus, and G.E. Schulz, Biochemistry, 27 (1988) 4465-4474. P.A. Karplus and G.E. Schulz, J. Mol. Biol., 210 (1989) 163180.

232 i0 Ii 12 13 14 15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

40 41

P.A. Karplus and G.E. Schulz, J. Mol. Biol., 195 (1987) 701729. P.A. Karplus, E.F. Pai, and G.E. Schulz, Eur. J. Biochem., 178 (1989) 693-703. M.G. Rossmann, A. Liljas, C.I. Branden, and L.J. Banaszak, Enzymes, ii (1975) 61-102. C.I. Branden, Q. Rev. Biophys., 13 (1980) 317-338. W.G.J. Hol, P.T. Van Duijinen, and H.J.C. Beendsen, Nature, 273 (1978) 443-446. R.K. Wierenga, M.C.H. de Maeyer, and W.G.J. Hol, Biochemistry, 24 (1985) 1346-1357. R.K. Wierenga, P. Terpstra, and W.G.J. Hol, J. Mol. Biol., 187 (1987) 101-107. R. Schkeif, Science, 241 (1988) 1182-1187. P.T. Jones, P.H. Dear, J. Foote, M.S. Neuberger, and G. Winter, Nature, 321 (1986) 522-525. C. Chothia, A.M. Lesk, A. Tramontano, M. Levitt, S.J. SmithGill, G. Air, S. Sheriff, E.A. Padlan, D. Davies, W.R. Tulip, P.M. Colman, S. Spinelli, P.M. Alzari, and R.J. Poljak, Nature, 342 (1989) 877-883. M.V. Milburn, L. Tong, A.M. deVos, A. Brunger, Z. Yamaizumi, S. Nishimura, and S.-H. Kim, Science, 247 (1990) 939-945. E.J. Goldsmith, S.R. Sprang, R. Hamlin, N.-H. Xuong, and R.J. Fletterick, Science, 245 (1989) 528-532. C.C. Hyde, S.A. Ahmed, E.A. Padlan, E.W. Miles, and D.R. Davies, J. Biol. Chem., 263 (1988) 17857-17871. C.C.F. Blake, Nature, 273 (1978) 267. J. Rogers, Nature, 315 (1984) 458-459. M. Cornish-Bowden, Nature, 313 (1985) 434-435. M. Marchionni and W. Gilbert, Cell, 46 (1986) 133-141. W.H. Landschulz, P.F. Johnson, and S.L. McKnight, Science, 240 (1988) 1759-1764. C.R. Vinson, P.B. Sigler, and S.L. McKnight, Science, 246 (1988) 911-916. A. Klug and D. Rhodes, Trends Biochem. Sci., 12 (1987) 464. R.F. Smith and T.F. Smith, Proc. Natl. Acad. Sci. USA, 87 (1990) 118-122. H.O. Smith, T.M. Annau, and S. Chandrasegaran, Proc. Natl. Acad. Sci. USA, 87 (1990) 826-839. P.Y. Chou and G.D. Fasman, Adv. Enzymol., 47 (1978) 45-148. J. Garnier, D.J. Osguthorpe, and B. Robson, J. Mol. Biol., 88 (1978) 873-894. W. Kabsch and C. Sander, Proc. Natl. Acad. Sci. USA, 81 (1984) 1075-1078. E.P. Pai, W. Kabsch, U. Krengel, K.C. Holmes, J. John, and A. Wittinghofer, Nature, 341 (1989) 209-214. E.F. Pai, W. Sachsenheimer, R.H. Schirmer, and G.E. Schulz, J. Mol. Biol., 114 (1977) 37. M. Murata, J.S. Richardson, and J.L. Sussman, Proc. Natl. Acad. Sci. USA, 82 (1985) 7657-7661. D.J. Lipman, S.F. Altschul, and J.D. Kececioglu, Proc. Natl. Acad. Sci. USA, 86 (1989) 4412-4415. T. Nishioka, K. Sumi, and J. Oda, in: P.S. Magee, D.R. Henry, and J.H. Block (Eds), Probing Bioactive Mechanisms, ACS Symposium Series, No. 413, American Chemical Society, 1989, pp.i05-122. K. Sumi, T. Nishioka, and J. Oda, Protein Eng. 4, (1991) 413420. W.B. Goad and M. Kanehisa, Nucleic Acids Res., 10 (1982) 247263.

233 42

43 44 45 46 47 48 49 50 51 52

M.O. Dayhoff, R.M. Schwartz, and B.C. Orcutt, in: Atlas of Protein Sequence and Structure, Vol. 5, Suppl. 3, National Biomedical Research Foundation, Washington, D.C., 1978, pp. 345-352. G.E. Schulz, J. Mol. Biol. 138 (1980) 335-347. R. Thieme, E.F. Pai, R.H. Schirmer, and G.E. Schulz, J. Mol. Biol. 152 (1981) 763-782. M. Suyama, T. Nishioka and J. Oda, unpublished. International Union of Biochemistry, Nomenclature Committee, Enzyme Nomenclature, Academic Press, Orlando, FL., 1984. D. Weininger, J. Chem. Info. Comp. Sci., 28 (1988) 31-36. D. Weininger, A. Weininger, and J.L. Weininger, J. Chem. Info. Comp. Sci., 29 (1989) 97-101. T. Nishioka and J. Oda, unpublished data. H. Kato, M. Chihara, T. Nishioka, K. Murata, A. Kimura, and J. Oda, J. Biochem., i01 (1987) 207-215. K.L. Krause, K.W. Voltz, and W.N. Lipscomb, J. Mol. Biol., 193 (1987) 527-553. K.H. Kim, Z. Pan, R.B. Honzatko, H.-M. Ke, and W.N. Lipscomb, J. Mol. Biol., 196 (1987) 853-875.

This Page Intentionally Left Blank

QSAR and Drug Design - New Developments and Applications T. Fujita, editor 9 1995 Elsevier Science B.V. All rights reserved

235

BACKGROUND AND FEATURES OF EMIL, A SYSTEM FOR DATABASEAIDED B I O A N A L O G O U S S T R U C T U R A L T R A N S F O R M A T I O N OF BIOACTIVE COMPOUNDS Toshio Fujita, Michihiro Adachi, Miki Akamatsu, Masaaki Asao, Harukazu Fukami, Yoshihisa Inoue, Isao Iwataki, Masaru Kido, Hiroshi Koga, Takamitsu Kobayashi, Izumi Kumita, Kenji Makino, Kengo Oda, Akio Ogino, Masateru Ohta, Fumio Sakamoto, Tetsuo Sekiya, Ryo Shimizu, Chiyozo Takayama, Yukio Tada, Ikuo Ueda, Yoshihisa Umeda, Masumi Yamakawa, Yasunari Yamaura, Hirosuke Yoshioka, Masanori Yoshida, Masafumi Yoshimoto, and Ko Wakabayashi EMIL Working Group, Department of Agricultural Chemistry, Kyoto University, Kyoto 606-01, Japan* ABSTRACT : Various structural transformation processes observed in a number of past developmental examples of pharmaceuticals and agrochemicals are regarded as being invaluable precedents for the prospective analog design. In certain cases, (sub)structural transformation patterns are interchangeable among various compound series in spite of differences in their pharmacological category. Thus, the patterns extracted with a computer-readable format could be accumulated and integrated as a database for potential "rules" for bioanalogous molecular transformations. EMIL is a system that incorporates the database and a data-processing engine constructed to release "higher-ordered" candidate structures from a "lower-ordered" input structure "automatically". Conceptual background for the database construction and the procedure for the database collection are presented on the basis of some lead evolution examples among pharmaceutical and agrochemical series of compounds. 1. INTRODUCTION There are numerous series of compounds exhibiting specific biological effects. Examples exist among such pharmaceuticals as those acting to nervous, circulatory, respiratory, digestive, and immunoregulatory systems and chemotherapeutics including antimicrobial and anticancer agents as well as among such agrochemicals as insecticides, herbicides, and fungicides. In each series, an ultimate prototype lead compound has been identified or disclosed first. In certain cases, bioactive principles in natural products, including secondary metabolites of animals and plants and endogenous participants such as hormones and signal-transmitters, are the origin of *The corresponding author and the business addresses of authors are listed at the end of this article.

236 the lead compound. In many instances, it is selected from organic compounds synthesized intentionally or unintentionally. The structure of the prototype lead compound is usually modified variously so as to improve the profiles of biological activity and to potentiate the target activity as well as to eliminate undesirable side effects including chronic toxicities and environmentally hazardous behaviors. There seem to exist two aspects in the structural modification processes. The one is the optimization of the lead structure with a systematic replacement of substituents keeping the skeletal structure (almost) unchanged. This is often called the "lead optimization" (1). The other is the structural transformation usually associated with more or less "drastic" variations in the skeletal structure. The structural transformation is usually performed into more elaborated or "higherordered" lead structures one after another consecutively, quite often in different institutions independently and/or competitively. These consecutive structural transformations could be called the "lead evolution" (2). Of course, the lead optimization can be made starting from the "intermediary" lead structure in each step of the consecutive lead evolution processes. How to make the lead evolution, i.e., the lead evolution strategy is also called the analog design (3). Although the disclosure or identification of the ultimate prototype structure is the prerequisite for the structural modifications, the lead evolution is perhaps most important from the synthetic chemical points of view to obtain patentable pharmaceuticals and agrochemicals having newer generation skeletal structures. In the structural transformation or lead evolution series, a majority of individual steps may originally be attempted on trial-and-error bases. However, because structural transformation patterns included in these steps have eventually been "utilized" in improving or at least in retaining the bioactivity profile, they are well regarded as being invaluable precedents for the analog design or "bioanalogous" molecular transformation (4). If these precedents are integrated and organized as a database for the bioanalogous transformation "rules" and the database is incorporated into a system so that any prototype or "lower-ordered" lead structures introduced into the system are processed with the rules to release elaborated or "higher-ordered" candidate structures as the output "automatically", the system could be a great benefit for the synthetic medicinal and agricultural chemists. We have been working on a project to construct a computerized system for the lead evolution or analog design, named EMIL : Example-Mediated-lnnovation-for-Lead-Evolution (5, 6). In this article, after showing some lead evolution examples, we demonstrate that certain (sub)structural transformation pattems are interchangeable among various series of bioactive compounds in spite of differences in the pharmacological category. Then, we illustrate how to collect the database and how to operate the EMIL system for the analog design.

237 2. LEAD EVOLUTION EXAMPLES From among a number of examples, we selected two each for pharmaceuticals and agrochemicals of current interest. In each example, the lead evolution processes were examined according to a "tree" in which structures are arranged not necessarily in the chronological order but from the most primitive (but not always simplest) structure toward the more elaborated (but not always the more complex) one somewhat concisely. If bioactive compounds before and after a certain structural transformation in lead evolution processes elicit analogous biological responses, the transformation could be bioisosteric and the two compounds or two interchangeable substructures be bioisosters in a broader sense. Here, we adopt the terms, "bioanalogous" and "bioanalog", instead of "bioisosteric" and "bioisoster", respectively, as proposed by Floersheim and coworkers (4). The term "bioanalogy" can be used more flexibly than "bioisosterism" without being restricted by the basic definition of the isosterism including isometricity in terms of various physicochemical parameters (7 - 9). 2.1 Cromakalim and Related Potassium Channel Activators. Figure 1 is a simplified lead evolution tree of cromakalim analogs, which are potassium channel activators exhibiting smooth muscle relaxation effects such as antihypertensive and anti-bronchial asthmatic activities (10 - 12). The very prototype was synthesized at Beecham (now SmithKline Beecham) in the early 1980's with an idea that the cyclization of the side chain in such I]-adrenoceptor antagonists (13blockes) as alprenolol (1) to restrict its conformational freedom may give compounds retaining the antihypertensive activity lacking side effects associated with l-blockers (10). The ring-closured compound of the structure 2 was found to indeed show an antihypertensive activity without 13-blocking effects. The geminal dimethyl at the 2position and the nitro group at the 6 position of compound 2 were necessary for the activity but introduced to enhance the cyclization reaction to form the dihydrobenzopyran skeleton originally (10). During the structural modification trials, the pyrrolidine compound 3 was shown to be highly active in vivo but only moderately in vitro. Thus, cromakalim (4) with a lactam ring was designed and synthesized as a possible metabolite of the pyrrolidine compound 3 and proved to be highly active (10). In the course of lead evolution processes starting from cromakalim (4), the lactam structure was successively transformed via the acyclic amide (in 5) and urea (in 6) structures into the cyanoamidine (in 7), cyanoguanidine (in 8), and triazolediamine (in 9) structures. These transformation patterns are shared by quite a few series of compounds of different pharmacological categories as will be shown later in section 3.2.2.

to

2

1: alprenolol

3

4 9cromakalim (lemakalim)

5

6

/ NCN

NCN

~ ~.~ 7 9KP 293

H3C~N--N

~ ~~

~~'~~

8

9

~o

~o

10" NIP 121

~o 12 "bimakalim

11 9emakalim

,

e.~.

P~.o

9~.o N

.~

o

N

~o NC I ~ ~ O . ~ , .

S.-c-N~cN H

O 2 N ~ --- CH2F

13" Ro 31-6930

14" TCV 295

15" YM 099

16" EMD 57283

17" SR 44994

Fig. 1. Simplified Structural Evolution Tree of Cromakalim Analogs.

18" KC 399

239 One of the other pathways is an elaboration of the lactam moiety leading to compounds 10, 11, 12, and 17 and to pyridine N-oxides 13, 14, and 15. A recently reported acyclic thioamide KC 399 (18) from Chugai (12e) is one of members designed and synthesized (13) with a combination of structural features of bimakalim (12), in which the dihydropyranol structure of the preceding compounds is dehydrated into the benzopyran (11), and aprikalim (19) belonging to an independent S~c.NHCH3 series of potassium channel activators (12a), in which a thioamide 6~sk..v o structure is attached at the c~-position to the aromatic system. The compound 18 was reported to be some 1000-fold more potent than 19: aprikalim cromakalim in relaxation of precontracted rat aorta (12e).

2.2 Non-peptide Angiotensin II Receptor Antagonists. The title compound series are recently attracting enormous attention to develop antihypertensive agents which are orally active with a prolonged duration (14). In the course of structural transformations leading to increasingly potent antagonists, it has been shown that there are at least two subtypes of the receptor, AT1 and AT2 (15). Structures arranged in Fig. 2 showing a summarized evolution tree are mostly those of the AT1 antagonists (16 - 25). The ultimate lead compound in this series is CV 2198 (20) which was synthesized by scientists at Takeda in the late 1970's in a series of projects for derivatization and screening of 1-benzylimidazole-5-acetic acid analogs (16). Because this compound 20 and its close analogs were among the first as the nonpeptide angiotensin II receptor antagonists, a number of research groups over the world started projects for transformation of the structure of compound 20 as the lead (14). Among intensive efforts, a great break-through is likely to be the disclosure of DUP 753 (23: losartan) at DuPont (now DuPont Merck) publicized in the late 1980's (17), because numerous analogs developed following losartan either share the 2'tetrazolyl-biphenyl-4-yl-methyl structure in common (in 24 - 26, 30, 31, 36, and 37) or have closely related biarylylmethyl structures carrying an acidic group bioanalogous to the tetrazolyl at the position corresponding to that in the biphenylyl structure (in 28, 29, 32 - 35, and 38) as an indispensable moiety. The imidazole moiety originally included in CV 2198 (20) has been variously transformed into spiro (in 30), oxy-aryl (in 26), and condensed bicyclic (in 31 - 38) systems as well as ring-fissioned structures (in 24 and 25). Candesartan cilexetil (31) is a prodrug. The ester moiety of this compound is metabolized into the free carboxylic acid, candesartan, as the active form in vivo (21a). One of the most recently reported compounds, L 162313 (35), has been revealed to be a partial

( - ~ , N,'r C1 X~/

N

~u-~. ,'r

.N~CH2COOH

C1

N

~u~. ,y

.N~CH2COOH

C1

,

N.~CH2COOMe

N

,'r

C1

N

.o. "v"

~ _ ~.~

~

.N~'~CH20 H ,~,,,,~,,~N,,~COO H V ' ~ N ' ~ ~

N

"~

O ---t~ ~-1~

20:CV 2198 /

~

21 :EXP 6155 O: ~ /2"EXP6803 /

~_~.~,~ ~COOH

~

~

Tet~ j ~ 23ilosartan

~

TetI ~ TetI ' ~ 24"valsartan / 5 " A 8 1 9 8 8

Vet I ~ ~TM 26"ICID8731

~u-~'~~u~ ~u-~'~~~-~o~o~~~. ~ ~. ~o ~.~ ~o-~~z~,

CF3SO2NH

27 9eprosartan

I

H --.t~.~.

28 "saprisartan

3

HO(~" ",,a,"

29 9SC 52458

30 9irbesartan

~-'~o

~ PhC

31 9candesartancilexetil

32 9TAK 536

N~'~Me

BuOC BuOC

33 9telmisartan

34 9MK 996

35 9L 162313

36 9tasosartan

37 9CL 329167

38" L 162393

Fig. 2. Simplified Structural Transformation Tree of Non-peptide Angiotensin II Receptor Antagonists (Tet 9tetrazol-5-yl).

1".9

241 antagonist acting also as the agonist to the AT1 receptor (22). This compound is the first non-peptide agonist of peptide receptors outside the opiate system. Another, L 162393 (38), is one of the balanced angiotensin II antagonists capable of potent binding to both AT1 and AT2 receptor subtypes (23). The AT1 binding potency of this compound in vitro is about 100 times higher than that of losartan at a subnanomolar level. The structure of compound 26 is unique as is that of eprosartan (27). In compound 26, the acidic biarylylmethyl group is attached to the heteroaromatic ring via oxygen. Eprosartan (27) has an acrylic acid side chain and the carboxyphenyl instead of the acidic biarylyl. In leading to these and related structures, threedimensional superimposition pattems of the small-molecule antagonist candidates on a putative pharmacophore model of angiotensin II has been examined iteratively (24, 25). The angiotensin II model has been constructed with structure-activity studies of its peptide analogs containing conformationally constrained replacement of key amino acid residues and conformational analyses of active analogs. The structural modification of this series of compounds is a typical example for the lead evolution associated with the lead optimization from the intermediary lead structures. Substituents at various positions in each structure of compounds shown in Fig. 2 are mostly those optimized with the more or less systematic modifications of the substituent structure in terms of the in vitro binding as well as the oral activity and its duration. The activity potentiation of the order of 10- to 50fold in the optimization phase is not unusual, if the substituent selection has been done appropriately.

2.3 Fungicidal [~-Methoxyacrylates and Analogs. o~-Substituted-aryl-[~-mcthoxyacrylatcs and their analogs such as o~methoxyiminophenyl-acetates and -acetamides are now being developed as agricultural fungicides with a systemic as well as a broad spectrum activity. Figure 3 shows a simplified lead evolution scheme of this series of compounds (26, 27). The original lead compound, strobilurin A (39), is a fungicidal principle included in small agarics belonging to species of Strobilurus and Oudemansiella which grow on decaying woods. There arc a number of analogs differing in substitution patterns on the conjugate polyene moiety and the benzene ring (28). The toxophoric structure of compounds in Fig. 3 is likely to be the "[3-methoxyacryloyl" or "methoxyiminoacetyl" moiety, but the corresponding free acids are known to exhibit only a very low activity. The fungicidal activity is due to the inhibition of the respiratory chain of fungi (29). The target site is believed to be the cytochrome bcl complex located in the inner membrane of fungal mitochondria.

242

OMe

OMe

!

OMe 39 9strobilurin A

40

~ 42

O~oMe I OMe

[~O

OMe !

OMe

41 OMe ~

~

[ ~O

O

i

Ooe OMe~ ~

M

NHMe 43" SSF 126

OMe 44" BAS 490F

,, N,,.Y-'N.o CN

O

OMe I

45" ICIA 5504

OMe

I~NSJ

OMe |

46

OCH3

Fig. 3. Structural Transformation Tree of 13-Methoxyacrylates and Analogs. The structural transformations from strobilurin A (39) to ICIA 5504 (45) have been made to increase the photostability and to decrease the phytotoxicity as well as to increase the systemicity into the plant body suffering from fungal diseases by adjusting the molecular hydrophobicity (26). Although the design principle of SSF 126 (43) is its own being from the ring fission trials of fungicidal carbamoyl isoxazoles (30), it is reasonable to locate this compound following the ICIA compound 41 in the lead evolution tree. Currently (August, 1994), besides ICIA 5504 (45) by Zeneca and SSF 126 (43) by Shionogi, BAS 490F (44) is being under extensive trials for commercialization by BASF (26). 2.4 Arylsulfonylureas and Related Herbicides. The ultimate lead compound of this series, INU 3373 (47), was serendipitously found to show a modest plant-growth retardant activity in the mid-1970's by Levitt and his coworkers at DuPont (31). The discovery of sulfonylureas such as chlorsulfuron (48: a wheat/barley herbicide), metsulfuron methyl (49: a wheat/barleyl/rice herbicide) and thifensulfuron methyl (52: a wheat/barley herbicide) shown in Fig. 4 was the fruits of extensive efforts of DuPont scientists (32). These and a number of analogous DuPont sulfonylureas are characterized by unprecedentedly low dose rates (generally 5 to 50 g a.i./ha with the lowest of 2 g a.i./ha) to eradicate various species of weeds (32). Depending upon structural

~1

.CH3

,COOCH3

SO2NHCONH---(, N - - ~ 47

~

d ON(CH3)2

48 :chlorsulfuron

/

~-

N._ ~,C1 OCH3 ff'-~" g N_--~ ~./~ N~SO2NHCONH-'~q_~ 55 9imazosulfuron

_N~ -N~I~"

Cl

/

~

~~~

N_ OCH3

CH3

OCH3

r

53 9pyrazosulfuron ethyl

~1~

54 9NC 330

I ~

. - >

Fig. 7. Benzocycloalka(di)ene-l-carboxylic Acids as Antiinflamatory Agents (98- 102) and Plant Growth Regulators (103 - 107). >>,---, and > compare the potency between two compounds of both sides in each series in common. We used to study structure-activity relationships of the same type of cyclized arylalkanoic acids (103 - 107) as plant growth regulators (54) the structures of which are also shown in Fig. 7. 1,4-Dihydro-l-naphthoic acid (104) was most potent among them. As the antiinflammatory agent, the indane-l-carboxylic acid derivative (98) was most potent and compound 108 named clidanac was selected as a clinical drug (52a, 55). Of course, the structure-potency patterns need not completely coinside between the two series of compounds. Among partially COOH hydrogenated 1-naphthoic acid series, however, coincidence in C l ~ the potency variations is remarkable suggesting a similarity at ~ J least in the substructural features of the receptor sites between [ 1 the two pharmacologically different series of compounds. ~108: clidanac

3.2.2 Urea, Thiourea, Cyanoguanidine, Nitroethenediamine, and Related Structural Components in Various Bioactive Compound Series. The bioanalogous relationship among the title "polar hydrogen-bonding groups" has been well known since most of them and other related groups were shown as being "interchangeable" with each other in various series of histamine H2antagonists (56). Their general structural feature, as indicated in Table 3, is to consist of the aromatic ring (R), flexible chain (C), and polar hydrogen-bonding grouping (H). Along with thiourea, cyanoguanidine, and nitroethenediamine structures, some other polar hydrogen-bonding groups are arranged in Table 3 as representatives in respective H2-antagonist series in which the aromatic ring (R) and flexible chain (C) are fixed (56, 57). Many of these polar hydrogen-bonding groups are found in various R-C series simultaneously. Although not every combination between the R-C and H moieties is congenial in giving potent compounds, the H structures for the polar hydrogen-bonding group in Table 3 are regarded as being potentially interchangeable. Interestingly, a very similar bioanalogous set of structural components is found in Fig. 1 for the cromakalim series of potassium channel openers. In the consecutive steps from the ring-fissioned acetamino-compound (5) to the methyltriazolediamine

T A B L E 3. Representative H2-Receptor Histamine Antagonists. J R " Aromatic ] Ring j

t C "Flexible Chain k

Ring "R" and Chain "C" H

H 9Polar ] H-Bonding Group

Polar H-Bonding Groups "H" S II

iCH3

)

mNHCNHCH3

109

NCN II

--NHCNHCH 3

CHNO 2

II

---NHCNHCH 3

110: cimetidine

NNO 2

II

--NHCNHCH 3

111

112 o

S

NCN

II

II

mNHCNHCH3

113 NH2

H2N-'J~NANN ~~S~'r

~

O II

--NHCCH2OCCH 3

120 9roxatidine s

~

--CNH 2

N'S'N --NH

~ I! NH 2

115 9ranitidine

116

O ii

II

117 9tiotidine

i

II

--NHCNHCH 3

NSO2NH 2

--NHCNHCH3

O II

---NHCNHCH 3

114

NCN II

S

CHNO2

N"S'N --NH

118" famotidine

~ /1 NH 2

119 o

H3C~N_ N --NH-~NN~.--NH 2

121 9lamtidine

N,,S-N --NH

,, I/' NH 2

122

CHNO 2

II

123

N H

CHNO 2

II mNHCNHCH3 124 9nizatidine

t'~

250 (9), structural components which are replaced one after another are those included in Table 3 as the hydrogen-bonding polar groups. A similar bioanalogous set such as compounds 125 - 127 exhibiting various degrees of smooth muscle relaxant activity have been explored in the synthetic project of compound 18 (12e, 13, 58).

O....C"NHCH3

NCN..~.,NHCH3

O.:.C"~ -

125

126

CN

u CH2F 127

Examples are also found in other series of potassium channel openers, pinacidil (128) and its analogs (129 - 132) (59) and nicorandil (133) and its analogs (134 and 135) (60).

~

N,NcC~_~.Bu

1~ NCN lq@N,, C,,N_~-Bu

128

[~ CHNO2 N J ~ N-C',N_.~t-Bu

129

130

O

N ~ ~ N,i~_~N._~t.Bu H2N,~ NCN ~ N,.C.. N,,.@ 131

~ONO2

132

NCN J~N~ONO2

133

NCN f ~ H2N~I~N~'~'~ -N

135

C1

Further examples exist in imidacloprid and related compounds (136 - 139) which are potent insecticides acting as agonists of the nicotinic receptor of acetylcholine in the insect nervous system (61) and in artificial sweeteners such as cyanosuosan (140 - 142) and superaspartame (143 - 145) series (62).

NNO2 N~NH 136: imidacloprid

A

l -2'Y

137

CHNO2 CI....~N~ C2H5 138: nitenpyram

CHNO2 NXNH NCN

CI 139: acetamiprid

251

N

~ C ~ ~ C O O H

HOOC

140:X=O 141 : X = S 142 : X = NCN

K,~ I

143 : X = O 144 : X = S 145 : X = NCN

It should be noted that, in compounds 5, 7, and 18 in Fig. 1,118 and 120 in Table 3, 125 - 127, 133 - 135, and 139, structural units, which are interchangeable with (thio)urea, N-cyanoguanidine, nitroethenediamine and related structures, have either (thio)amide or N-substituted amidine structures which lack one of the two N atoms in (thio)urea-related structures. The bioanalogous relationship between amide and N-cyanoamidine structures is likely to be disclosed first in penicillins such as 146 and 147 showing an antibacterial activity at comparative levels (63). The possibility for the cyanoamidine compound 147 to be active after hydrolysis giving the amide was excluded. The cyanoamidine is stable enough chemically and tolerable against enzymatic hydrolyses. NCN

O/~-'N ~.,SCOOH 146 :penicillin G

o,~N 147

I,,,COOH

3.2.3 F r o m " A m i d e s " to Cyclic D i c a r b o x i m i d e s a n d R e l a t e d Structural Transformation

Patterns

in A g r o c h e m i c a l s ,

Anticancer

Agents, and

Anticonvulsants.

Compounds having the N-phenyl-amide moiety such as anilides (148),Nphenylcarbamates (149) and N-phenylureas (150) are herbicidally active exhibiting various degrees of the Hill reaction (a component of the photosynthetic system) inhibitory potency (64). The most conventional substitution pattern on the benzene ring in these compound series, 148 - 150, is X = 3,4-C12. Propanil (148: X = 3,4-C12, R = Et), swep (149: X=3,4-C12, R = Me) and diuron (150: X = 3,4-C12, R = R ' = Me) are among representatives. They are regarded as being bioanalogous to each other.

148

149

150

There is a family of agricultural fungicides the structual feature of which is that they are N-phenyl cyclic dicarboximides, such as procymidone (151:R1 - R4 =

252 Me, R2 - R3 = -CH2-), vinclozoline (152: R 1 = Me, R2 = CH=CH2) and iprodione (153:R1 = CONHCHMe2, R2 = R3 = H), sharing the 3,5-dichloro-substitution on the benzene ring in common (65). They are particularly effective on Sclerotinia and Botrytis diseases in vineyards and greenhouses.

R2 3 C

_ CI

151

N

1 O

C1

152

O

2 3

153

Structures of the cyclic imide moiety of above fungicidal compounds, the pyrrolidinedione (in 151), oxazolidinedione (in 152), and imidazolidinedione (in 153), can be regarded as being generated through the cyclization of the side chain structures of the Hill reaction inhibiting anilides (148), carbamates (149) and ureas (150), respectively, with the insertion of another carbonyl component. Structures 151 - 153 are bioanalogous. Regardless of the type of atoms next to the carbonyl function, the open chain "amides" ( 1 4 8 - 150) are the Hill reaction inhibiting herbicides and the ring-closured dicarboximides (151 - 153) are fungicides. N-Phenylcarbamates 154 and 155 having structural features common with the herbicides (149) are also fungicidal against gray mold diseases of vines, vegetables, and beans caused by Botrytis strains resistant against benzimidazole-fungicides (66). Thus, in spite of some differences in the target of the biological activity and the optimum substitution pattern on the benzene ring, the open chain "amides" and cyclic "dicarboximides" can be regarded as being bioanalogous. Examples supporting this respect will be shown below. Cl CH3CH20--~ Cl

154

CH3CH20

NHCOCH(CH3)2 155

Among anilides (148), chloranocryl (X = 3,4-C12, R = -C(Me)=CH2) and pentanochlor (X = 3-C1, 4-Me, R = CH(Me)C3H7) have been used practically to exterminate annual grass and broad-leaved weeds in various crop fields (67). They have the 3,4-disubstitution patterns as X as well as the branched chain alk(en)yl groups as R. Interestingly, a member of compound series 148 similar to the above herbicides, but having X = 3-CF3,4-NO2 and R = CH(Me)2 named flutamide from Schering, is an antiandrogen (68) and has been used as an antiprostatic cancer agent for some 15 years. Flutamide, having the 3,4-disubstitution pattern on the benzene ring and the branched alkyl as R, is reasonably considered to show some Hill reaction inhibitory activity. Although no description about the herbicidal activity has been

253 found, some higher homologs of flutamides in the acyl moiety have been observed to show a potent antibacterial activity (69). Quite interestingly moreover, compound 156 named nilutamide from RousselUCLAF is also a potent and selective antiandrogen being used as an antiprostatic cancer agent (70). The bioanalogous relationship between anilides and N-phenyl cyclic dicarboximides very similar to that described above in agrochemicals is observed in entirely different pharmacological category.

_ ~ O2N F3C

156

O )I.-~H O2N~ N ~ (~-CH3 ~ O CH3 F3C

H _ ~ N OH NC ' ~ (~-CH3 O CH3 F3C

O cH NHC-- CH2SO2- ' ~ F ~H3

157

158

The dicarboximide heterocycle of nilutamide (156) belongs to the imidazolidinediones (in 153). The structural differences of nilutamide (156) from the fungicidal compound series 153 are the substitution patterns on the benzene and imidazolidinedione tings. Flutamide works as its hydroxylated metabolite 157 in vivo (71). The hydroxy group in the metabolite 157 corresponds well with the NH group in nilutamide (156). Thus, nilutamide is regarded also a ring-closured bioanalog of the metabolite 157. By the way, bicalutamide (158) modified further from the "hydroxyflutamide" is now being extensively investigated for clinical use by Zeneca (71).

O~

H

HN _C=O ,C'--~ Et O Ph 159 :phenobarbital

O~

,H

QC--O

f-'


(3_

('I>

~o ~l-

(I)

,--o

~:~

0 ~-~

~---

~'~" ~;~ :~" r e

"0

0

(D 0

(I)

:~

II

I

0

OCD

I

~-b --b

0

~" CI)

C~) O -

~

I I I C-2 C"2 C'~

0

O)

I

~

I

I

I

~.l,O

0

~CD

~

o

:::~

Z ~

0

(I)

~-b ,-b

0

O(I)

:::~

~

0

(I)

~

C..~

0

O" Cl)

Z

~

~['o

~

"~ 0

0

(I)

,-b ,-b

0

O" (I)

C:)

(1)

~'=~

~" (I)

0

(I)

I

--b "~

0

CT" (I)

~

Z

0

(I)

C~

w-6

O" CI) 0

0

~"-

~" ~o

0

(I)

~.=,o

,-b

0

~" (1)

0

00(I)

(w~

~

C~

0

(I)

(1)

(I)

o-

,--,. ~:::~

"~

g)

(I)

Ca

E~

~...,o

~.,o

Ee

w,,o

(9

w-.,

0

w.,o

(3

(0 ~-~

0

0 ~-

CLCD O0

0

~D

"13

c~

~

w-- ,--.

(D

e-~'~

(~ ~'~

::~- ~:~0 CI)

C~

~-~:~

00~

"~D

C~

0

~-. 0

CD

-'.

CL

~. ::~- ~ ,.--.- '-~ ~:~

~0

0

0

0

0

(I)

~:~ ~"

O0

::3"W'~

(3- ~-+

(I) (I)

re~


( - ~ ('~ = 0

=I=

.

[mO ~---~

~-.-["0

~

0

::x:~ CI)

4~.

.

/ ~

.

O0

O0 ['.0

~

.

03

['x3 [~0

-IV

Z

r.O

4:~

03 ~.]

.

~.~

~

*-~

'~

0

~

O'J

~ L'~

.

"~

t"B

~:~ --

"0 :

~

OO -.~]

I

0

"~

~

(I)

('D

~

0

.

.

..--.~ O o

r.O C.O

I

C.O

O'J

O00q r.O ~:;~

rd~

IV ~

--

I C'~ C'~ =:Z:: Z[~

["0

O0

0

ICE>

=~

..

X

X

I

I'

~

,I=~ Oo

=:Z='.

I

~

(--2

(3"J 0"I

-....]

~ Oq

0-1

O0 Oq

"

Z

+-~

0

"'0

0

CI)

L~O ~

(3 0 (9 ~b

o

C~

~o

o

"o

~-o

(9 00 (3

Z o

0

'0

iJo

i'D

C3

,..,o

0

b-,o

0

II ,...,o

i,--o

i-j

OB

l=.,o

L~

294 c o n t r i b u t e to t h e toxicity r e g a r d l e s s of t h e i r n u m b e r

in a m o l e c u l e .

The

t h r e e k i n d s of v a r i a b l e s are listed in Table 11. FALS c a l c u l a t i o n d e r i v e d a d i s c r i m i n a n t f u n c t i o n w i t h p r e t t y good d i s c r i m i n a n t a n d predictive ability. i n c l u d e d in t h e f u n c t i o n .

As s h o w n in Table 12, 4 0 v a r i a b l e s are

F r o m t h e sign of t h e d i s c r i m i n a n t coefficient for

e a c h variable, it is inferred t h a t n u m b e r s of N, O, S, a n d CI a t o m s , b e n z e n e and naphthalene

rings, h y d r o p h o b i c i t y ,

etc. c o n t r i b u t e to e n h a n c i n g t h e

toxicity, w h e r e a s n u m b e r s of sp 3 c a r b o n a t o m s , carboxylic a c i d s a n d esters, etc. p r o b a b l y c o n t r i b u t e to lowering t h e toxicity.

TABLE

13

Results

of r e c o g n i t i o n

and p r e d i c t i o n

Recognition

Calcd 1

Obsd

1

152

2 3 N

=

324

Correct

MMG

recog

Leave-one-out prediction

12 0

= 0.859 = 87.3%

3

16

0

1

1

142

20 0

MMG = 0 . 8 0 2 pred = 80.2%

Nmi s

1 32

41(0)

Calcd

2 3 N = 324 Correct

2

99 12

Nmi s =

Obsd

u s i n g 40 d e s c r i p t o r s

2

3

26

0

89 13

= 64(2)

Rs

= 0.859

(p 12 j=l

(4)

In eq. 4, vector uij is the coordinate of the j-th atom of these chains on the moment of inertia, vector < u > is the coordinate of the gravity center of these chains (c]. Fig. 12), and m is the total number of O and C atoms in the chain. The distributions of alkoxyl chain lengths are shown in Fig. 13. Then, the effective lengths(RL) of the alkyl and alkoxyl chains were expressed relative to that of the w-chain of LTE4 defined by eq. 5.

354

.~ :--~-x

I

\.,.2~/

,;

7

/

,'

,' , ; ,,

,u2

~

"

v , ~, : '

.

~

~

~

-i,,.'

t ,,~

"~/,.



u5 u6

U'_/

/~ Moment

z

Gravity center

i'.~.

~"~,

, ,'

Fig. 12. Moments of inertia and gravity center < u > of alkoxyl

~.

of i n e r t i a

chains.

(5)

RL = )~LT

In eq. 5, )~LT is the covariance value of the w-chain of LTE4 in its most stable conformation. We took the length of the w-chain in its most stable conformation as the reference, because this chain can be assumed to take the most extended conformation when it binds to its receptor(38). A value of RL - 1 indicates an identical effective length to t h a t of the w-chain at certain 0

9

,

9

,

9

,

3

o~

, k

20 i

i

t

f

10

=:" / -/

CnH2n,1-0

L)

5

:,~ 6 :~: '7 "~......;, !..i i

8

9

10

hi'...

X" - ' ",,.,t-.,".'$,,";>"7%-,: A'Ja-". t ' """ " " " ""-" .t-

..

~::

::"

:: .... "';

20

40

""..:i'"

.... ...

60

Z# Fig. 13. Distribution probability (Pi) of i-th conformation of the alkoxyl chain length as a function of the covariance value ()q). Numbers besides traces represent chain lengths n.

356 1.0

i

0.5

0.0 ,,J

-0.5 -1.0

-2.0

J!

qp

-1.5

i

f 4

5

6

7

8

9

10

Fig. 15. C h a n g e of DL with length of the alkoxyl chain of B P O s . n: C n u m b e r of the chain.

lengths with the w-chain length: the highest similarity corresponds to Fs 1.0 and D L = oe, and 50% similarity to Fs = 0.5 and D L = O. Values of D L are summarized in Table 1. The D L value of the alkyl group of C, was about the same as that of the alkoxyl group of OC,_I, indicating that the feasibilities of these corresponding chains to have an effective length similar to the w-chain are very close. The highest D L values were observed with chains of C7-C8 and OC6-0C7. Fig. 15 shows the change in D L with the alkyl chain length n of BPOs. The D L value was maximal between n - 6 and 7, at which maximal activity was observed. It is noteworthy that the curve was not symmetric, change of D L with n being very steep in the region of less than n - 6 and relatively less in that of more than n - 7, indicating that the flexibility of an alkoxyl chain (and alkyl chain) becomes greater with increase in its chain length.

0

EFFECTS OF EFFECTIVE LENGTHS OF ALKYL AND ALKOXYL CHAINS OF BENZAMIDES ON ANTI-LT ACTIVITIES

It was of interest to know whether antagonists with chains of similar effective lengths to that of the w-chain exhibit potent anti-LT activities. Thus, we analyzed the antagonist activities of these benzamides in terms of DL, and obtained the significant correlation shown by eq. 8. plso-

7.088 + 0.478 D L + 0.735 Io + 1.096 IBp (+0.240) (+0.215) (+0.261) (• (n -- 18, r -- 0.957, s = 0.259)

(s)

355

BPMs

I0.0 7.5

o~

~

__~

--~i

7.5

.o_

-T"I

5.0

BPOs

I0.0

::::::::::::::::::::::::::

..

~.~

. ..

g:::,.:.....::l r ~..::~.~-~:~ ::.. 9 ::.~

~ 5 - 0 F.:i:!:::::~::"

2.5

2.5

-

i !

0

ii i

U,~I ~:~:~:~i.a . . . . . . . . .

0.4

/

i::..

.~.:.. :::~ !

0.8

!

12

!

.6

0 2

0

0

0.4

0.

O

!.2

1.5

2.0

RL

RL

Fig. 14. Conformation probability (P~) of alkyl chain (left) and alkoxyl chain (right) of B P s as a function of their effective lengths relative to that of the w-chain of LTE4. N u m b e r s besides traces are C numbers.

conformations of alkyl and alkoxyl moieties. Some of the results with BPMs and BPOs are shown in Fig. 14. Binomial distributions of conformations were observed, the distributions being the same for BPMs and BPOs with the same chain lengths. To quantify the similarities in the effective lengths of alkyl and alkoxyl groups to that of the w-chain of LTs, we estimated the feasibilities of these chains to take a certain range of lengths similar to the length of the w-chain of LTE4. The feasibility is expressible by summation of the probability of occurrence (P) of conformations in the RL range of 0.8 to 1.2, shown by the shaded area in Fig. 14. The sum of the areas under the distribution curves in this range is referred to as Fs as shown in eq. 6.

(i "0.8 _< RL

~

, ~o o

"--4-)

0

4-)

(d

~

~ ~9

"o @ O

H

@ r

~

~

(1) ~

~ O

4m

o @ .~--) 0 ~

o9

~

>

"~

~ o B.0

. ~ (D .~

O q-~

r.q q) ~ O

~ ~ 0 , "~

~ 0 "J

O

~ 0

cO ~9

(D 4-)




(D (~ ,-~

(D

.~ "q a::l r..q 09 (D

~ ~

~ ~

0

~

r::~

O

~

~

4-) O

,~

~

ad 0 ~

-~

4-) O ~

>)

cO ~

,~ O

~ @

0 .~

~1

O o

~ @

~

~

~

q-~

o .~

,~

~'~

~ "~ ~2 bO

O ..I-n

Oq-~

,~9 ~ ~

O

~_) ~

~-~ o . ~ ~'~

>. > )

~~

4~

0

0 0 (D > ,--~ @ ~1 @ .>~ (D (D . ~ q q_~ .,_~

~

O"~

~ j:z~.'~

0

O

.~ (D 094o ~ ccI 4-)

~-~ Z ~.~

~=

~ "0 ~;Z: s.-C~ s ~ r,q "~-) :Z~ ~3: (]) led -H .,-~

3:

~'~

(D.~ ~ ~

~. O

4.~0 ~ @ ~ (D ~

~4m 0 ~

~

'-'-'

~

~.~'~ . ~ .~.4~H ~ " ~ O (b ._~

F.--, " ~

"~

t~ ~d (D-~ O ~

~ . 0~ O.-, O I:D~~c::~ ~

~

A

q_~ O . H ~ @ .~ O O ~ F._., rc~0 (D a::l ( D . ~

~

(I) ~

4m 0 .~'~

O

"~

~

~

=

~ ~::~ (b

Z~9

~=

@ q-" i:::z,.~ ..o ~ ~ ~ {,q ~>..~[/~ ~ (b ~

ca~

.1_~4_.~ 0 [/..) O ~ .~ ~

~ ~.~

O'~

N

~

O..,

[n.O ..~ ~

2~o >~:~o=.~='~

--~L) ~

~

o

.ID O

'-,-Oh

f::~

r~

O'~

~

""~ O .~~_~ ~.< ~ Q O "~ (D o ~ 4 . ~ ~ O-'4~ - ~ ~ I Q~D ~ ~-~ ~-~ @ ' ~

~

~,~

o

~ =~

~ ~ ~ ~

.

.~

(D ~

~

4--) @ ;:> o c~ (]) ~ (b ,..c:; (b . aJ

~

"~'c0:0

,r_,

(b

O.

;;"

~.,

~z~

o

~-~

z

O = I

"~ I

O 4-) ~ s

~

o

f-~

--

m= o

~

~-4 ~d L~ 4m

F..q

'--1

~

~'~

~,~

,~

~

_

~

~

[-~ Z ~

~: ~>~ ~

~.~

m,,,,,

=

[.~

[-~ ~'~ Z

"0

~

(d

~-~

.H

~:;

9

o

(b S:,,

cd

~ ~

~

~ ~

H H

.,-.-i 4-) ~

s ,,~

(~ 4:= 4-)

O ~

(1) .D

g

"~ ~ O

~

~

"

"

0 4-)

H

@

(lJ ~ a~

O

2

0

~ ,~

O

0

0 .~ r

~ O

~>)

ad ~

~

~

0

~

~

~

~:;

~

s 4--) (D

,----I

(d

~

(1] 4--) ad ,-~

(b

O

"

,.~

.I~ "~

(])

(D o ad

('d

9

O9 I

~

~ _~I

aJ

(D

(D ~ ~

ral3

~ H

~-~

"

s @

H

~"

H H m

~= ~

4_.)

"~

(])

.I-)

O

~

~ I or) ad .- '~ ~

(D ,q

o9

(])

"~

>)

O ~

..~

,~

O ~

O O

~ @ ~

"~D

.~ 4--) ~

O

O ~---"4~ ~-I 9 (1.)

r.,q :

~

(D .~ O

O

~

o O

~: (D ~ ~

~ O

I

4-)

.D

O

q-~

~

~ ~

4-.~

~

,..~

.,~ >

0 (b

r.q

~

0

:>~

ccl ~

4-) O (D

Q~

(D 4-) O

o

@

@

4--)

.~

(b

(])

"u

.,~

,-~

~

O

~

~

X..sd

~

"O

@

~

.s

"~

..,o

.~

~

~

~ ~.~

~

cd ~

O 0 ~

(1)

~

(b ~ b9

~

~ 4--) 0

.~

0

~

X:::::

*

~

s

0

(b

" '--

(D

Q) > ~

~aO ~9

~

.~

R

.

~-~

;>~ ~ ~

I

~ a~ ~

~ (6 O .~

O

~

0

L~ @

"~ E

~>).~

;>~ 2h4

O_, ~

cd

0

"~ -~ ~_~

~ "~

.~

=

>'~ (b ~

.~

~

r.._)

~

~

0

c-f"

:z~

c~

~ ~ C~ ~ c-I" ~ ' ~ 9

b~

~

(1) ~ c-~ r~

~

~

c-~

0

=

~

=0

9

c~

~

Ct

c-~

(I)

~

(1)

~-~

0

~ c-~ ,~ 9 ~I" :z~-"

t-,. c_l.

ct Ca

< ~.

~ ~-

0

~

:z~ ('D

C~

~.

:~-,

~

o

c~ 0

~-~

0

I C~

~ ~ 0 ~ (1) ~

~-~

~

~

I~

0

(I)

c~

0

~ 0

,--3 ~

c-~ :z~ ~)

0 ~

< I~

(-~

.

~ ('D ~ ~

0

(1)

"

(-r :z%

(1) ~

~

~ c~

t-.,.

~

~. C3

~.

1~ ~ c'~ I~ O~ 0

C~ ~ I

(I)

~.

0

~)

~ .,.

~)

(1) ~

~"

0 ~ c-f l~

0

c-l" ~ 0 ~ I~ 0

Cr C)" ~. c-f"

~

:z~

~

~_~

~.

~.

ct

~-' 0

~

~.

=

~"

0 0

c-I" ~ (1) ~ t~. ~

I ~Z~ t--,. C~

9

~.

~

~

~-~

~ ~.

~"

~

~

~.

I~

~" 0

~. ~

~ ~ c-l" ~ C~ 0

C-~ I~ '

~D

~

~-~-

~

~

:~

~

~ ~

~ r~

I---I

~

~ ~

c-l" ~.




~J ~ c?

'~" ,--t ~ 0

{I} ~z: 4-~

~ ~ ~ O I ~ CO ~'~

~ ~ r~

4-) 1:2, l~ ~

F~ O o

.,--I 4-~ ~ ~

~ (D "O

~ (D .~

4-) ~ .~

~ O

~ 9 .~

4~ 4-} ~1

I cO I:~

~ ~z: 1:2,

"u ~ ccl

.~ ~

~

O

~ (D

~

"" ::> 4_)

~ ~

"U (1.}

.,~ 4-) ~

.~ .z~ :3: "O

,z~

,.D '-I

O ~-~ ~

60

4..) ~ ~9

O ~

(D

{1:1

I H ~-4

~

~

r

(1.}

,_c::; E~

~ O

~ ~

.~ ~

~.)

m.~ ~ ~ ~

~

~ I

~ od

~ I

"~E~ ~ ~

~ .,~ (1.}

~

~

4-~

4-) ~

(~ ,--t

O

"~

(D

~

O q-~ (D

(D 4-) O 1:2,

4~

,_~ ~

O

.H E~ ~

.~

~

O

(D

~

q-t

{1:1 Ix: O

(D (D C.)

~9 Od II (D

(~ "~

1:2,

,--I

"

(D

"O

or) II

4~ (D

,---I

(D ..~

~

~

.,--I

q-~ ~

4-~

. eo

O

~ O

-,-q

~ ~ ~1

.,-i

~ "C::l

~

~ ~J ~

II

~ ~1

O

4~ ~ ~

~ I--I

~

4~ ~ .~

"(~

~ (I.} >

~ (3.} ~= ~:

I

o ..~

4.~ .,-I

4D

r~

9

i__1

9

o

or} cq

il

Om

Om

9

~T

,5

c:)

~--

O,.T

II

c)

coo odb-9 .. o ,.s c:) o,J cO

-t-

9

['-- o,J o'~oO

+

I-- c )

o,d

coo or~ ("q

i

C ) C)

('X.I

1:2,

Cb

~

rc'~

E~

o

(D

~r

~ ~D

H ::~

W

~D ~

~ ~"

0

~

~-~ 0

ct

d)

D3 ~ ~.

ct

0

c~ Z:r

0

~D

c-f ~

0

t---' ~

~ ~.

~ ,---3

~D

X ~. E~ ~

(1)

~--~

~

0

~

~

~D

h~

0

~

~

,~

-

9

cr

c-t

~

0

~

~

~

c-I" ~"

zz

ce

~

~

~

~

dD ~

~

~

~

~

c-f

c-l"

CD

0

~ I:D

~

.

0

~

~

c~

~.

~

ct

~ ~0

~ ~

~-~

9

:_~

~

~:~

~.

~

~ 0

~

~ ~1"

r~

0

~

i_,"

c-I"

< ~.

CD

~

~

('D o

0

c-f t:r

~

~D O~

~

r..~

~

Crq

'-~

~D

__~

~

~



"

,---, PO

-.--,CO

0k33

0~0 C.O C)~

~13>

"

I

II

II

I

~='

:~'

I

I

0

PoI~

I

0

~9. . ~ .

I

I

I

I

.~r

~

I

I

I

~ I

I

O. r

~

I

I

I

X

~

I

I

I

0 I

I

I

~ I

I

I

~ I

I

I

I I

I

I

I "0 _~.

I I

I

I

0

-3

~D ---~r

I

g')

~

c ~ O0 4:~ o o 0

0 0 m ~ c'+~D

--~

0 -~.. N

- l - - t - G)

I

v

~

~D ~D ~D ~

~9- ..~.

~D ~

I I

I

I

I

I

t

I

~-

0

I

~-

~

I

-~

~

c+r

-1~ 0 7 Po -.q 4~ -.q ~

~

~

~D ~D CD CD ~D ~D fD

C~ 0-1 4::~ O0 r,o t--~ 0

~

I'D

:::~

~ 00

1

_..j

~:

"-

~

0

I

Z-'~ ~-"J

PO

"~"

~D I

0 7 4:~ r,o P~ 4::~ 0 O0 -I:::~ 4:~ 4::~ ~ 0 CY~ PO ~ 0

rD 0

~D - - C ~ , - -

"~ I'D

-IDO I

I'~0 ~-, 0 0 Cr~ 4::~ 0 4:~ O0 Po

A CY~ CY~ CY~ Cr', r.,.rl Lrl C~ CY~ Cr~ CY~ C~ C~ 0 7 . . . . . . . . . . . .

I

I

c_~.

N ~D O_

-$

0

~D

DO

X~

C~

"~ I'D

~

.-J.

o

,-~-

~ ,,

~

I

~

I

_...a

o

•

o

('D ~

i

r---

v~ c-

O0

('D

-S --.~ I

_.1.

"~"

..J. 0

fl)

~

CD

- o .-J. ::5-

o

" o9 ~ - ~

o

--J. ~ Po

376 molecule. used

and

as

For the sake

values

of simplicity,

relative

to

that

these

of

H:

steric

A MR(X)

parameters

= MR(X)

were

- MR(H)

A B5(X ) = B5(X ) - B5(H ).

Table 3 Ca-antagonistic activity and physicochemical parameters of R3-substituted compounds (II) Me0,

CN

Me

Me0~C-(CH

2 )3N (CH2) 3 0 0 ~

MeOr--- R3

Me

PA2 Compd. No.

R3

~

a)

) AMRb) AB5C

11-5 H 0.00 0.00 11-6 Me 0.54 0.46 11-7 Et 1.08 0.93 11-8 n-Pr 1.62 1.39 11-4 iso-Pr 1.49 1.40 11-9 n-Bu 2.16 1.86,, 11-10 iso-Bu 2.03 1.86~! II-11 n-Hex 3.24 2.79~! 11-12 , n-Oct 4.32[! 3.72!! II-13 g) n-dodecyl 6.48t) 5.58t) 11-14 benzyl 2.22 2.90 II-15 (CHg)~OMe-0.32~! 1 57f) 11-16 (CH~i~OEt 0.50t) 2 03f) a) b) c) d) e) f) g)

0.00 1.04 2.17 2.49 2.17 3.54 3.45 4.96 6.39 9.27 5.02 3.49 3.81

A c) Obsd.d) Eq. 1 B1 Calcd.(A )e) 0.00 0.52 0.52 0.52 0.90 0.52 0.52 0.52 0.52 0.52 0.52 0.52 0.52

Eq.3

Eq.2 Calcd.(A )e)

5 . 5 6 6.28(-0.72) 6.76 6.91(-0.15) 7.44 7.33 (0.11) 7.79 7.52 (0.27) 8.05 7.49 (0.56) 7 . 2 1 7.50(-0.29) 7.53 7.52 (0.01) 7.46 6.79 (0.67) 5 . 0 6 5.21(-0.15) 5.33 -0.80 6.48 7.48(-1.00) 6.80 6.22 (0.58) 6.68 6.56 (0.12)

Calcd. ( A )e) 5.76 (-0.20) 6.77(-0.01) 7.38 (0.06) 7.46 (0.33) 7.38 (0.67) 7.43(-0.22) 7.45 (0.08) 6.68 (0.78) 5.10(-0.04) -0.49 6.63(-0.15) 7.44(-0.64) 7.35(-0.67)

5.83(-0.27) 6.61 (0.15) 7.15 (0.29) 7.43 (0.36) 7.43 (0.62) 7.47(-0.26) 7.47 (0.06) 6.79 (0.67) 5.11(-0.05) -1.28 6.64(-0.16) 7.48(-0.68) 7.42(-0.74)

From ref. i i unless otherwise noted. Scaled by 0.i and from ref. 12 unless otherwise noted. Calculated from the values cited from a brochure given by Dr. A. Verloop. pA9 values in the KCl-depolarized guinea-pig taenia coli. A~ the difference between observed and calculated values. Estimated from those of closely related substituents, see ref. I0 for the detail. Omitted from the correlation.

In Eqs. because reason

of was

another length

]-3 compound its not

quality

in T a b l e

especially

large

the

in terms

deviation

in Eqs.

2 and

from

extra have

correlation

particularly 3,

an

might

of the

was omitted

deviation

but

site

of the R 3 chain.

satisfactory, shown

clear,

receptor

The

(II-]3)

pronounced

for

from the c a l c u l a t i o n

the

correlations.

binding

arisen

interaction

due

of Eqs.

to

the

]-3 was

of the standard alkoxyalkyl

3. We o r i g i n a l l y

The with

increased by no m e a n s

deviation.

derivatives thought

that

As was the

"

~< .

E~

I

~ ~

I:1)

0

c-t.

~" ~

d)

~ 0 ~

~-, CD

~ c~" ~ (D

1~

l::::Z,

I

of" ~

~

,-~

b~

~

C-~

0

C~

~~ .

~

~.

0"g :

~

~

~. O

c-t

~ o

:~.

~

~

(-I"

D~...

(-~

o

O l:::z, (D

~

~

I-~.

CD

~

0

c-t

d)

c-t"

~

~.

~

O

I

o O ::=t

O c-I" ~z~ (D

~-'

dD O

I:1)

~

~ .

0

"~

c-t"

~-]

CD ~ ~

O E~ ~ dD ~"

~ O (-I"

,,.

~ D~

Po

r

~

0

c--t"

""

~

0

"1

O

~c'-t" "

I

(-t~ ~ ~

~-

~.

I k,n ~ ~

co

CD

I

~0

~

.~j

H

~m

~

T

~0 ~, ,-

~

m

PO

O ~ ~

~~.~

~

~"

~ o

m

_~>/c~_ a o

o'o

zo,-

Z

[]

u1 I

2 2 PA2+0.30ALpara-0.73ALpara+0.77AMRmeta-1.12AMRmeta

(D

c-f

~-'-

O ~

I- ~

(1)

Ca

(I)

c~

~z~

~"

~-'.

~

~ ~

c-I"

(D

ct

~

~a

~"

~

~

~

ct ,.

C~

Ca

~

ct

~

(1)

"

~

~.

~

t- ~

0

~

~

-~

~D

c~

(1)

(1)

~

c-I" ~

(1)

c#

~.

Ca ~

E

~

0

(D ~

c-t

~

c~

~

~

(1)

~

~.

0

c-t"

0 ~

~

(D

0

0

c-t" ,. 0 ~

O

~0

0

~"

0

(1)

(1)

c-~

~.~

0 ~

~

c-~

0

(-t, ~'~

O

(I)

c7"

~

"

~ ~

~"

(D

c~

hb

~

~

~

~

9 b,~

Ca

0

~

ct

~.

[]

~=

----4

C)

I

"-~

0

~ :~

cf

0

(D

~

Oq

0 (D

'

~

H~ 0

Ca

'

--~ Po

(D

c#

~

(1)

~"

c-t"

~

o

~

C/)

0 ~

Ca

~z~

~

0

~.

=~'

0

~-b 0 ~

Ca

Ca

~

~

~

~

0

~,

9

c-~

0

~

(D



I=7" (I)

0

~" < r ~

< ~

-q ~

,~ ~

0 ~ ~) ~

~I

.,~ ~-1

~

~.

IS~

~

.~ '~,

,

~--, ~-~

~ ~z~ (I)

< ~. ct" H. (I)

0 c-~

I

" o I~0

II I

CO ~ .. I-~.


r"-

bd

P

P

r--

t> r---

I>

. ".4. r~)

0

XD

bd

0

P

P

. k~). -.j

0

X~)

bd

,, ~,,

.C~ t'o

P

P

I>"

X=)

bl

.--J

~,,

P

P

X=)

bd

,,

--J

P

P

XD

bl

(.~

,~,

P

P

bl

m

O -h

~F

F~z

CD

-q

=~

.mII hO

CD

9

II



9

+

9

+

+

--~

ho

c=zn ---,

CDLn

.

CD--~ . . . O~ ~ --~WI

P~

~ O O

~13 ~

.

I

[-~

~-j

~U ---

c-~ O

~ O

LIl - q

(1) c-t ~) ,~,

:z~

~

~U ~ (1) C-~

CD--~

~

O

-~

hO