COMBINATORIAL
`CHEMISTRY
`
`
`
`Synthesis and Application
`
`Edited by
`
`STEPHEN R. WILSON
`
`New York University
`
`ANTHONY W. CZARNIK
`
`IRORI Quantum Microchcmistry
`
`A Wiley-Interscience Publication
`
`JOHN WILEY & SONS, INC.
`
`New York 0 Chichester
`
`- Weinheim - Brisbane
`
`' Singapore
`
`- Toronto
`
`Page 1 of 27
`
`ILMN EXHIBIT 1037
`
`
`
`Page 1 of 27
`
`ILMN EXHIBIT 1037
`
`

`
`This text is printed on acid-free paper.
`
`Copyright © 199? by John Wiley & Sons, Inc.
`
`All rights reserved. Published simultaneously in Canada.
`
`Reproduction or translation of any part of this work beyond
`that permitted by Section 107 or 108 of the 1976 United
`States Copyright Act without the permission of the copyright
`owner is unlawful. Requests for permission or further
`information should be addressed to the Permissions Department.
`John Wiley 8'. Sons, Inc., 605 Third Avenue, New York, NY
`10l58-0012.
`
`Library of Congress Cataloging in Publication Data
`
`Combinatorial chemistry : synthesis and application I edited by
`Stephen R. Wilson and Anthony W. Czarnik.
`p.
`cm.
`Includes index.
`
`ISBN 0-4?-"l-1263?-X (cloth : alk. paper)
`I, Combinatorial chemistry.
`I. Wilson, Stephen R. (Stephen
`Ross), 1946-
`.
`II. Czarnik, Anthony W.. 1957-
`RS4l9.C666
`I99‘?
`6 I 5' . l9—dc20
`
`96-44718
`
`Printed in the United States of America
`
`I0 9 8 7 6 S 4 3 2
`
`Pae2of27
`
`
`
`Page 2 of 27
`
`

`
`CONTENTS
`
`List of Contributors
`
`Preface
`
`jl Introduction to Combinatorial Libraries: Concepts
`
`and Terms
`
`Stephen R. Wilson
`
`Parallel Organic Synthesis Using Parke-Davis Diversomer
`
`Technology
`Sheila Hobbs DeWi(t and Anthony W. Czarnik
`
`Polymer-Supported Synthesis of Organic Compounds
`and Libraries
`
`Mark J’. Kurt}:
`
`Macro Beads as Microreactors: New Solid—Phase Synthesis
`
`Methodology
`
`Wolfgang E. Rapp
`
`Combinatorial Libraries in Solution: Polyfunctionalized Core
`Molecules
`
`Edward A. Winmer and Juiius Rebek, Jr.
`
`Solid-Phase Methods in Combinatorial Chemistry
`
`Irving Smhoteiki
`
`Page 3 of 27
`
`
`
`Page 3 of 27
`
`

`
`CONTENTS
`
`Radiofrequency Encoding and Additional Techniques for the
`Structure Elucidation of Synthetic Combinatorial Libraries
`
`Xian-yl Xlao and Michael P. Nova
`
`Combinatorial Synthesis Exploiting Multiple-Component
`Condensations, Microchip Encoding, and Resin Capture
`Robert W. Armstrong, S. David Brown. Thomas A. Keating, and
`Paul A. Tempest
`
`Indexed Combinatorial Libraries: Nonoligomeric Chemical
`Diversity for the Discovery of Novel Enzyme Inhibitors
`Michael C Pirrzmg, Joseph H. -L. Chan, and Jrlung Chen
`
`Strategies for Combinatorial Libraries of Oligosaccharides
`
`Carol M. Taylor
`
`Soluble Combinatorial Libraries of Peptides, Peptidomimetics,
`and Organics: Fundamental Tools for Basic Research and Drug
`Discovery
`John M. Ostresh, Barbara Déirner, Sylvie E. Blondelle, and
`Richard A. Houglzlen
`
`Combinatorial Libraries of Peptides, Proteins, and Antibodies
`
`Using Biological Systems
`Stephen Benkovic, Grove P. Miller, Wenyan Zong, and Jefi’ Smiley
`
`Index
`
`Page 4 of 27
`
`
`
`Page 4 of 27
`
`

`
`1 I
`
`NTRODUCTION TO
`COMBINATORIAL LIBRARIES:
`CONCEPTS AND TERMS
`
`STEPHEN R. WILSON
`
`Department of Chemistry, New York University, New York, New York E0003
`
`In the past few years, combinatorial chemistry has become the popular and often
`misunderstood “new wave” in drug discovery. In some cases, combinatorial chemis-
`try is presented as being in direct competition with rational, or computer-aided, drug
`design. Nothing could be further from the truth. Combinatorial chemistry encom-
`passes many strategies and processes for the rapid synthesis of large, organized
`collections of compounds called libraries. When planned intelligently, combina-
`torial methods produce collections of molecularly diverse compounds that can be
`used for rapidly screening for biological activity. Without planning,
`the GIGO
`(garbage in—garbage out) principle applies. Whether or not a library has in some
`way been designed or made more or less at random depends on the reasons for
`preparing the compounds. Combinatorial chemistry as a laboratory practice cannot
`replace computer modeling as an exercise in refining our basic understanding of
`molecular interactions. It is likely that both rational drug design and combinatorial
`chemistry will be used,
`in concert when appropriate, or directly applied to the
`problems best suited to each method.
`Combinatorial chemistry includes many research areas—new analytical meth-
`ods, new computer modeling and database—related challenges, new synthetic ap-
`proaches, new types of reagents, and new types of assays. Although this chapter
`will provide current leads into all these fields, the basic groundwork for combina-
`torial chemistry is still being laid. Many new research areas have yet to be explored.
`Combinatorial chemistry has its conceptual roots in the immune system. In the
`body, when a new antigen comes in contact with the preexisting large collection of
`antibodies, the antibody that binds best is selected and reproduced in large numbers
`
`1
`
`Page 5 of 27
`
`
`
`Page 5 of 27
`
`

`
`INTRODUCTION TO COMBINATORIAL LIBRARIES: CONCEPTS AND TERMS
`
`to effect the immune response. In a similar way, combinatorial chemistry involves
`the synthesis of a large number of compounds. (This collection can be a chemical
`mixture, a physical mixture, or individual pure components.) The collection is then
`tested for biological activity. Finally the active compound is identified and made in
`quantity as a single compound. This type of overall approach is not unlike a
`bioassay guided search for useful natural products.
`Thus, combinatorial chemistry approach has two phases: (1) making a library
`and (2) finding the active compound. Screening mixtures for biological activity has
`been compared to finding a needle in a haystack. This approach may seem unlikely
`to improve the likelihood of finding new drug leads. The fact that it appears to work
`has generated all the excitement.
`This introductory chapter will review several methods for haystack construction
`and needle searching. For example, the best way to prepare a mixture is in some
`type of array or coded form. Then, it is easy to locate and decode the structure of the
`active compound. Combinatorial chemistry has also prompted a resurgence of inter-
`est in rapid synthesis methods, particularly in the area of solid-phase synthesis.
`Many related topics such as automated synthesis, methods for identifying the struc-
`ture of the active compound on a small scale, and handling the massive amounts of
`data sometimes obtained (i.e., computer databases) are also discussed. The main
`portion of the book continues with specific topics written by experts in the field.
`
`1.] HISTORY
`
`The first report describing combinatorial chemistry appeared in 1984 in a study by
`Mario Geyson (1) titled, “Use of Peptide Synthesis to Probe Viral Antigens for
`Epitopes to a Resolution of a Single Amino Acid.” Another early pioneer was Furka
`(2) who introduced the commonly used pool-and-split methods. The 19808 was a
`period of rapid development in solid-phase peptide Syntl.'1csis—-Bruce Merrifield
`won the Nobel prize in chemistry in 1984 for his work on soIid—phase synthesis (3).
`During this time, automated peptide synthesizer technology was in its infancy, and
`the preparation of individual peptides was a challenge. ln 1985 Richard Houghten
`introduced the “tea bag” method for rapid multiple peptide synthesis (4). These and
`other advances in manual multiple-peptide synthesis (5) fed the beginnings of a
`wave of rapid bioassays based on the developing area of molecular biology. Mass
`screening of peptide ligands as a tool for drug discovery allowed the development of
`high-throughput bioassays (6,7).
`Because, historically, biological assays from natural products screening were
`often carried out on mixtures, early pioneers in combinatorial chemistry developed
`the concept of intentionally making mixtures for the purpose of testing. This ap-
`proach had the benefit of more rapid screening. For example, if a mixture showed
`no activity, then all compounds in the mixture could be assumed to be inactive.
`(This assumes that compound A in the mixture does not effect the assay of com-
`pound B or even react with compound B.) Instead of the usual method of individual
`peptide synthesis, it is possible to couple a mixture of all 20 amino acids to another
`
`Pae6of27
`
`
`
`Page 6 of 27
`
`

`
`HISTORY
`
`mixture of 20 amino acids to produce 400 dipeptides. If these newly formed 400
`dipeptides are reacted with a mixture of 20 additional amino acids, 8,000 tripeptides
`are obtained. This process describes the original formulation of combinatorial
`chemistry. The mixture of 8,000 tripeptides would be called a combinatorial library.
`The combinatorial possibilities for the number of compounds is seen to increase
`exponentially as shown in Figure 1.1, that is, N reactions can produce N X N
`compounds.
`Although it is easy to see how the rapid production of many compounds in a
`mixture works, it is not so easy to see how to make this process work in drug
`discovery. There are several problems. For example, in the process shown in Fig-
`ure 1.l, what would happen i@ne or more of the amino acids does not react or
`reacts sluggishiy to give a low yield? The mixture would then not contain 8,000
`tripeptides but a lower number. The peptides
`uld not be present in the idealized
`statistical distribution (i.e., equimolar ratios
`dditionally, how does one find a
`single active peptide in a mixture with 7,999 inactive ones?
`The first problem is always a sticky one and will be discussed in more detail
`later. The second problem has the following solution, called. Imag-
`ine constructing the library not with all 20 amino acids but starting with only 19
`amino acids, leaving out one specific (known) residue. After subsequent coupling
`with 20 times 20 amino acids,
`this leads to a library of_ only 7,600 tripeptide
`compounds. If this library is now inactive, we have learned that position I of the
`
`Number at different peptides increases
`exponentially with length of molecule‘
`
`Number of
`amino acid
`residues
`
`Peptide
`
`.
`
`Number of
`distinct peptides
`
`NH2—x,x2-cooH
`
`I
`
`NH2-X1X2X3-CDOH
`
`NH2-X1X2X3X4-COOH
`
`NH2-X1X2X3X4X5-C00H
`
`400
`
`3,000
`
`160,000
`
`3,200,000
`
`NH2-X1X2X3X4X5)<fi-COOH
`
`64,000,000
`
`NH2-X1X2X3)(4X5X5X7-COOH
`
`1,230,000,000
`
`NH2-X1X2X3X4X5X5X7Xg-CODH
`
`25,600,000,000
`
`Xn represents individual amino acids residues.
`Numbers of distinct peptides are based on
`20 residues at each position
`
`Figure 1.1 Combinatorial peptide chemistry. The number of different peptides increases
`exponentially with length.
`
`Page 7 of 27
`
`
`
`Page 7 of 27
`
`

`
`4
`
`INTRODUCTION TO COMBINATORIAI. LIBRARIES: CONCEPTS AND TERMS
`
`..._..--.__
`
`‘T’
`
`g‘-,1/ULi|¢I..
`
`tripeptide has the residue that was omitted from the synthesis. If you continue this
`process, methodically omitting one amino acid after another, a single combination
`still has the active compound. Thus a maximum of 20 experiments are needed to
`define position 1. Now, with the active first residue fixed, 20 more experiments are
`carried out to define the best residue in position 2. There are 20 + 20 + 20 = 60
`omission experiments required to define the complete tripeptide sequence of the
`active component in 8,000 compounds. This mathematical principle is the parallel-
`ism advantage: 20 + 20 + 20 = 60 (linear) versus 20 X 20 X 20 = 8000
`(exponential) that underlies the principle of combinatorial chemistry (8,9). This
`iterative synthesis and screening strategy has been discussed in detail by Dooley and
`Houghten (8). A careful analysis of methods for deconvolution of libraries and their
`intrinsic problems is described by Freier and co-workers (10).
`As you can see from this example, this strategy for combinatorial chemistry
`relies on two important points. First, the synthesis of the mixture must be fast and
`efficient because you need to prepare many variations of the libraries. Second, the
`testing must also be fast and easy, because you need to rapidly test many compound
`mixtures to find the active sequences. Because it is not always the case that both
`synthesis and testing are fast and efficient, these two key points return again and
`again. In addition, complications involving multiple active components (more than
`one needle in each haystack) often confuse the issue. On the other hand, the classic
`combinatorial situation in Figure 1.1 is a basic method with which one can Contrast
`other approaches.
`Fast and easy synthesis and testing suggests automation and robotics. Over the
`past few years, combinatorial chemistry has grown so quickly, in part because good
`automation exists in both areas. Section 1.7 on robotic instrumentation will review
`
`some advances in robotics that have enabled fast and easy preparation of com-
`pounds. A discussion of the automation of bioassays,
`that is, high-throughput
`screening,
`is beyond the scope of this book, and readers are referred to other
`reviews of advances in assay processes (6).
`
`1.2 REVIEW OF THE LITERATURE
`
`Combinatorial chemistry is an approach to synthetically produce molecular diver-
`sity. Traditionally, molecular diversity was primarily obtained from natural prod-
`ucts. Two early reviews contrast natural products diversity (11) with molecular
`diversity obtained from peptide combinatorial chemistry (12). Two other early re-
`views in 1993 on applications of combinatorial chemistry barely mention organic
`chemistry (13,14). It was not until 1994 that new comprehensive literature reviews
`of combinatorial chemistry were published covering both peptide combinatorial
`methods (15) and organic chemical applications (16). Since that time many other
`reviews have appeared directly addressing organic combinatorial chemistry (17-
`26). A comprehensive bibliography of combinatorial chemistry references is avail-
`able on the World Wide Web (http:Nvesta.pd.com) (27).
`
`Page 8 of 27
`
`
`
`Page 8 of 27
`
`

`
`SOLID-PHASE ORGANIC SYNTHESIS
`
`1.3 SOLID-PHASE ORGANIC SYNTHESIS
`
`Until the revolution in high-speed bioassays, the testing of new compounds was the
`rate-limiting step in the drug discovery process. While it might take a chemist one
`or two weeks to synthesize a single compound, it required a much longer time to
`carry out the biological assays. Now modern high-speed assays using robotic sarn-
`plers can screen more than 10,000 compounds per week. The biological assay has
`evolved from the rate-limiting step to the driving force in the need for large numbers
`of compounds, and thus the driving force of combinatorial chemistry. In some
`cases, where compounds are prepared attached to polymer supports, bioassays are
`still not very well adapted.
`In the area of high-speed automated synthesis, peptide chemistry has already
`arrived. In 1963, Merrifield reported the first examples of solid-phase synthesis of
`peptides using chloromethylated-polystyrene containing immobilized N-protected
`amino acid building blocks (3). This chemistry developed over the ensuing decade
`and became the basis for much of the progress in peptide chemistry.
`Figure 1.2 shows the Merritield approach. An insoluble polymer bead contain-
`ing—CH2Cl groups is prepared by chloromethylation of the copolymer of styrene
`and p-divinylbenzene. These chloromethyl groups can be esterified with N-pro-
`tected amino acid building blocks. The amino group is iteratively deprotected and
`coupled with new N-protected amino acids to build the growing chain. Merrifield
`used the rerr-butyloxycarbonyl protecting group (BOC group) to protect the free
`amino terminus during the coupling step. This allows depmtection with acid
`(CF3COOH) before coupling with the next N-protected residue. The cycle can be
`continued until the desired sequence is obtained. By-products and excess reagents
`are not bound by the resin and can be removed from the resin beads during washing
`cycles.
`An entire industry has developed to serve the peptide synthesis field. There are
`many instrument companies that provide automated peptide synthesizers, as well as
`other companies that provide peptide building blocks, reagents, and supplies (17).
`Solid-phase synthesis techniques do not, per se, mean fast synthesis. On the other
`hand, there are many years of experience in automating bead—based synthetic ap-
`proaches because it is easy to extend the principles to organic applications.
`Although peptide chemistry advanced using solid-phase techniques, applications
`of polystyrene bead-supported approaches to traditional organic molecules did not
`go far in the earlier days. Some work appeared in the 1970s, but an early review by
`Rapoport had a negative tone (28). Although later successes by Leznolf (29),
`Neckers (30), Frechet (31), and others clearly showed that polystyrene bead-
`supported organic chemistry worked quite well, interest in the field largely disap-
`peared (32). This might have been due to several factors. First, solid-phase chemis-
`try often produced only small quantities of compound. Pre-1980 nuclear magnetic
`resonance (NMR) instrumentation required much larger amounts of compound for
`characterization than is needed with today's high—field Fourier transform (FT) ma-
`chines. Second, the high-speed testing of small amounts of compound that launched
`
`Page 9 of 27
`
`
`
`Page 9 of 27
`
`

`
`-»
`
`INTRODUCTION TO COMBINATORIAL LIBRARIES: CONCEPTS AND TERMS
`
`I
`
`R1
`
`R‘
`
`O
`
`3oc\
`
`N
`H
`T--pr
`
`GOOH
`
`°
`
`NHBOC
`CF-_-,COOH
`----——I-
`
`O
`
`*3
`
`mi:
`
`rmeoe
`
`R:
`
`CF3COOH
`T».
`
`D
`
`O
`
`R
`
`1
`
`N
`H
`
`0
`
`R2
`
`""2
`
`P
`
`R,
`
`O
`
`°\")\N
`
`H
`
`‘3
`
`P
`
`Figure 1.2 Merrifield solid-phase peptide synthesis (3).
`
`the field of combinatorial chemistry had not emerged. The late 1970s and 19803 was
`a period of tremendous growth in the development of selective reagents and meth-
`ods for solution-phase organic synthesis. There were plenty of new targets and new
`synthesis methodology to occupy the attention of organic chemists. Thus, while
`solid-phase peptide methods rapidly matured and led to instrumentation for automa-
`tion of chemistry, solid-phase methods for organic synthesis all but faded away.
`Until 1992, the combinatorial library field was exclusively the domain of peptide
`and oligonucleotide-based chemistry.
`The situation changed in 1992 with the report by Bunin and Ellman of the
`preparation of combinatorial libraries of organic molecules (33). The report of solid-
`phase synthesis of Diversomers by the Parke-Davis group (34) shortly thereafter
`stimulated considerable interest in reexamination of solid-phase organic synthesis.
`Within the past few years, a number of reports have appeared that point to great
`potential for solid-phase organic synthesis. An excellent recent review of the litera-
`ture of solid-phase organic synthesis has recently been published (35).
`Bunin and Ellman’s (33) solid-phase synthesis of 1,4-benzodiazopines is an
`excellent illustration of the process of combinatorial library production using organ-
`ic templates (Fig. 1.3). The chemistry is very similar to the Parke—Davis Diversomer
`
`Page 10 of 27
`
`
`
`Page 10 of 27
`
`

`
`SOLID-PHASE ORGANIC SYNTHESIS
`
`Figure 1.3 Ellman solid-phase l,4-benzodiazcpine synthesis.
`
`approach (34) detailed in Chapter 2, except that Bunin and Ellman’s compounds
`were not cleaved from the solid support. 1,4-Benzodiazepines are constructed from
`three components: 2-aminobenzophenones, N-protected amino acids, and alkylat-
`ing agents. As one can see in Figure 1.3, the preparation of a combinatorial library
`on solid—phase synthesis beads requires several things. First, a selection of building
`blocks are needed with R1, R2, R3, and R4 that can be permuted. Second, a reaction
`scheme that is suitable for interative conversion, in a similar manner to peptide
`synthesis, is required. Third, a method for linking one of the building blocks to a
`solid support and conditions for cleaving the compound(s) off the support at the end
`of the cycle must be found.
`The variation of groups Ry-R4 give a matrix of different structures (Fig. 1.4).
`These compounds are the combinatorial library and the chemistry is sometimes
`called matrix chemistry. Besides chemistry issues, keeping track of a large collec-
`tion of structures is often difficult. Section 1.8 discusses some of these points as
`well as software solutions that attempt to deal with multidimensional collections of
`pharmacophores.
`From only a few studies in 1992, the number of publications reporting applica-
`tions of the use of solid-phase organic chemistry has increased dramatically. So
`many new applications of organic reactions on a solid support have been reported
`that only a partial list of the types of reactions carried out on a solid support can be
`provided here (Table 1.1).
`Besides the reaction itself, so1id—phase synthesis methods also require consider-
`ation of several new technical issues. Table 1.2 reports some characteristics of the
`
`Page 11 of 27
`
`
`
`Page 11 of 27
`
`

`
`INTRODUCTION TO COMBINATORIAL LIBRARIES: CONCEPTS AND TERMS
`
`Template
`
`R,
`
`OH
`
`OH
`
`OH
`
`OH
`
`OH
`
`OH
`
`R,
`
`R,
`
`Cl
`
`CH,
`
`CH, CH,
`
`CH, CH, CH,
`
`CH,
`
`CH,
`
`CH, CH,
`
`C]
`
`C]
`
`CH, CH,
`
`CH, CH,
`
`CH, CH,
`
`CH, CH, CH,
`
`Figure 1.4 Typical molecular spread sheet of benzodiazepine derivatives.
`
`most common form of resin bead: crosslinked polystyrene. In planning a synthesis,
`one must consider loading, the level of substitution on the beads and swelling
`characteristics of the resin. As can be seen in Table 1.2, crosslinked polystyrene
`beads swell quite a bit in volume with different solvents. The physical size of the
`beads used in a solid-phase synthesis, of course, also effects the weight of each bead
`
`TABLE 1.1 Organic Reactions on Polystyrene
`Solid Supports (35)
`
`Reaction
`
`Reference
`
`Dieckman cyclization
`Diels-Alder reaction
`Micheal reaction
`Aldol condensation
`
`Mitsunobu reaction
`
`Pausen-Khand Cycloaddilion
`Organometallic additions
`Stille reaction
`
`Suzuki reaction
`Heck reaction
`
`Suzuki coupling
`Urea synthesis
`Thiazoline synthesis
`Solid-phase Steroid synthesis
`Wittig reaction
`
`28
`36
`37
`38
`
`39
`
`40
`41
`42
`
`43
`44 and 45
`
`4-6
`47
`48
`49
`37
`
`Pae12of27
`
`.._ -
`
`
`
`Page 12 of 27
`
`

`
`SOLID—PHASE ORGANIC SYNTHESIS
`
`TABLE 1.2 Data for Polystyrene-Bead Solid Support (50, 51)
`
`Styrenewdivinylbenzene copolymer:
`
`l—2% crosslinked; 1 g of 200-400 mesh resin has ~4-
`IO X 105 beads
`
`l05—l30°C
`Thermal stability:
`Bead sizes: 100 mesh (212 um), 140 mesh (150 pm), 200 mesh (106 um), 325 mesh
`(75 pm), 400 mesh (45 um)
`Swelling: DMF (3.5 mLfg), TI-IF (5.1 mLfg), CH30H (1.5 mug), H20 (1.5 n1LIg),
`CI-IECI2 (5.3 mug), CHCI3 (5.9 mI..;’g)
`Loading: 0.1-0.4 rnmollg resin, ~I00—400 pmol per head
`
`and therefore the amount of product on each head. Although crosslinl-ted poly-
`styrene resin beads (Merrifield-type resin) are the most common for historical
`reasons, a popular new material called TentaGel is now coming into wide use (52).
`This new material contains a polystyrene core with polyethylene glycol spacer arms
`(PS—PEG). TentaGel (although physiomechanically less robust than polystyrene)
`has very desirable characteristics for synthesis because the attached reacting groups
`project out in solution rather than being anchored close to the polymer back-
`bone. This provides for better reactivity and for reactivity more closely paralleling
`
`R1"CH3,R;=H
`K1=t-Bu,R2=H
`R]=Ph, R1'='H
`R1=CH3»R2=CH3
`R1"t-Bu.R2='CH3
`R1=Fh. Rz"CH3
`R1"'CH3,R1“"Cl
`R]‘='I,'B|),R2=Cl
`R]‘=Pl'l, Rz=‘CI
`
`Figure 1.5 Solid-phase synthesis of small molecule combinatorial libraries (37).
`
`Page 13 of 27
`
`
`
`Page 13 of 27
`
`

`
`I0
`
`INTRODUCFION TO COMBINATORIAL LIBRARIES: CONCEPTS AND TERMS
`
`solution-phase chemistry. Applications of TentaGeI in combinatorial chemistry is
`the subject of Chapter 4.
`A good example to illustrate the use of well-known organic reactions is Chen and
`co—workers’ report of the preparation of a small collection of compounds from a trityl
`resin-linked alcohol (Fig. 1.5) (37). The diol was linked to trityl polystyrene and then
`oxidized with S03!pyridine to the aldehyde. The resulting polystyrene-linked al-
`dehyde could be reacted with three different Homer-Emmons reagents (R, = Me,
`I-Bu, and Ph). Using the split-mix method, these enones were combined in equal
`amounts, split into three separate reaction vessels, and reacted in a Michael addition
`reaction with three different R2-SH compounds. Nine different compounds were
`formed, a very small but prototypical combinatorial library. After product cleavage,
`each of the nine compounds could be isolated and completely characterized. In
`addition, the use of Fourier transform—infrared (FT-LR) and gas chromatography-
`mass spectroscopy (GC—MS) for analysis of products attached to beads was demon-
`strated. This work and its extensions are more fully described in Chapter 3.
`
`1.4 ANALYTICAL TECHNIQUES FOR SOLID-PHASE SYNTHESIS
`
`The principle of solid-phase synthesis requires that the product remain attached to
`the resin head. A disadvantage of the technique is that it is more difficult to
`characterize the product at intermediate stages of the synthesis. Methods used to
`identify reaction products from solid-phase synthesis typically involve elemental
`analysis, titration of reactive groups, or simply weight gain (35). Sometimes, a
`small amount of the intermediate resin beads are removed, and the attached product
`cleaved and analyzed in the usual manner. For product identification, organic chem-
`ists usually use spectroscopic methods such as IR, NMR, or MS. Scheme 3. i5 in
`Chapter 3 shows the FT-IR of starting material and product in a typical solid-phase
`reaction;
`the conversion to product can be clearly observed by changes in the
`spectrum.
`
`is NMR. Unfortunately,
`The mainstay of organic structural analysis, however,
`the heterogeneous environment of resin—attached compounds leads to very broad
`‘H-NMR lines. Recent work suggests that the use of magic—angle spinning ‘H-NMR
`(53,134) and '3C—NMR (34) with polystyrene resin-bound compounds may be use-
`ful. The use of ‘3C-NMR with synthetic compounds bound to TentaGel resin beads
`(PS—PEG) was shown to give excellent line widths for compounds where no signal
`could be observed for the same compounds directly attached to polystyrene (55).
`The use of 13C-enriched building blocks also provided high-quality spectra with less
`interference from the polymer and solvent.
`Finally, the use of mass spectroscopy for evaluating solid-phase synthesis has
`been explored. Methods usually require cleavage of the product from the resin and
`analysis of the resulting microsamplc in the usual manner. Several examples have
`been reported involving direct GC—MS of the product from a single head. A single
`100-200 mesh polystyrene “Merrifield” bead weighs approximately 0.1-0.2 p.g.
`At a usual loading, each head contains r-400 pmol of compound, more than enough
`for a mass spectrum. Figure 3.20 in Chapter 3 shows a typical single-bead result.
`
`a
`
`, Pae
`
`
`
`Page 14 of 27
`
`

`
`ENCODING OF BEADS
`
`11
`
`Several reports have also appeared using direct matrix-assisted laser desorption
`(MALDI) of combinatorial library beads (56-58). The most elegant approach to
`mass spectral analysis of synthesis beads was the report of the molecular weight
`imaging of individual beads using time-of-flight secondary ion mass spectrometry.
`The structure of small peptides could be seen directly from individual 30-60 um
`polystyrene beads after cleaving the resin—peptide link with trifluoroacetic acid
`(TFA) (59).
`The combination of afiinity selection for active receptor ligands with electro-
`spray mass spectroscopy shows promise for the rapid identification of active com-
`pounds in a combinatorial library (Fig. 1.6) (60). A small molecule combinatorial
`library containing 600 compounds was incubated with purified receptor and then
`separated by size-exclusion chromatography. The receptor containing bound ligand
`was collected, and the ligand was released and analyzed by electrospray mass
`spectroscopy. Figure 1.6 shows the low-molecular-weight (MW) region of the spec-
`trum reveals a tight-binding ligand from the library at MW = 708. Specific binding
`of this compound can be determined by evaluating its competition with a known
`receptor ligand. It is likely that further developments of this “affinity selection”
`technique could speed up isolation and identification of leads from combinatorial
`libraries.
`
`1.5 ENCODING OF BEADS
`
`The combinatorial method for producing mixtures of thousands or millions of
`compounds has always presented an awkward deconvolution process. In principle, a
`better strategy would be to prepare the compounds in pure form, in separate, labeled
`bottles. What if each compound was attached to a separate bead—-one compound
`per head (61)? The mixture of beads then represents a physically separable collec-
`tion of compounds. This approach has been widely used and only requires a means
`to determine what is attached to a single bead. Instead of direct structure determina-
`tion of the compound attached to a single head, the synthesis history of the head is
`recorded on the bead—-encoding. Each head can be derivatized with a “tag” that
`contains information on the structure attached to the bead. Therefore, instead of
`identifying the compound, one identifies the code. This concept was first described
`for beads encoded with a genetic DNA tag (62). This elegant approach allows
`amplification of the encoded “signal” by polymerase chain reaction (PCR) for
`analysis.
`A more practical approach for organic chemistry is the tagging strategy devel-
`oped by Ohlmeyer and co-workers (63). The use of the split-synthesis method
`provides a final library wherein each head has a single compound attached. This
`method (Fig. 1.7) involves attachment (at the level of a few percent of the real
`compound) of a binary code of chlorophenyl groups at each synthesis step. Using
`electron-capture GC analysis, 0.1 pmol of code attached to a single 50-80 pm head
`is more than enough for detection (64). The presence or absence of one of the four
`GC separable chlorophenyl derivatives (Fig. 1.7, n = l—4) gives a binary code of 0
`or 1. Thus the binary codes such as 0000, 0001, 0010, 0100, and so forth can be
`
`Page 15 of 27
`
`
`
`Page 15 of 27
`
`

`
`won3
`
`____.._._._...._____.____..___.__._.____O
`
`mdmm
`
`
`
`
`
`
`
`....dam:EE.mamm....._..........mwa......nmmvN53.33"an
`
`
`
`
`
`
`
`Tan.TamNmmm.1
`
`...m.mENd?Nmmmcm"1.
`nmnm3%_mumGa2.9:1
`
`mmq
`
`rr
`an
`
`(N
`IN
`
`r-4
`-4
`
`O
`
`‘*5!Ln
`N
`
`“E0'‘!
`.—I
`
`(9%) "WI '|3H
`
`(96) ‘W1 '|9H
`
`
`
`Page 16 of 27
`
`
`

`
`ENCODING OF BEADS
`
`polymer
`
`eleavable linker
`
`Figure 1.7 Still bead encoding method from combinatorial libraries (41). The synthesis
`polymer is derivitized with a photocleavablc linker and a tag containing a choloroaryl group.
`The tags can be removed from the beads, the resulting alcohols silylatcd and then detected
`by GC.
`
`used to define the synthetic steps applied to the head. The compound synthesis
`proceeds on the beads as usual. After each step a coding reaction is carried out on
`the bead at a trace level (~1%). Usually the code is linked using a photocleavable
`linker so that the code can be removed without damaging the valuable (perhaps
`active) compound attached to that bead. Details of the use of an encoded library to
`locate ligands for a synthetic (unnatural) receptor molecule has been reported (65).
`Further development and commercialization of this coding method is now in the
`hands of Pharmacopoeia (66) and a review of the Still coding method has just
`appeared (67).
`Another coding method uses peptide sequences for a reaction code (68). Two
`orthogonally protected points of extension allow the primary peptide (or organic)
`molecule to be formed at one point and a coding sequence built at the second. The
`code can be “read" by selective cleavage of the coding peptide, followed by se-
`quencing. The coding sequence was claimed not to effect binding of the primary
`structure .
`
`(a) Electrospray MS of a 600—compound small molecule library. (b) Library
`Figure 1.6
`from (a) after affinity selection with an excess of receptor. Note the strong enrichment and
`selection of the compound at MW == 708 (MI-l+). (c) Mixture after afiinity selection in the
`presence of a strong competitive ligand. This figure shows that the compound at 708 MH+ {Z
`can be competed off the receptor by this peptide ligand.
`
`Page 17 of 27
`
`
`
`Page 17 of 27
`
`

`
`14
`
`INTRODUCTION TO COMBINATORIAL LIBRARIES: CONCEPTS AND TERMS
`
`Finally, an intriguing new coding method has just appeared. TentaGel polymer
`beads containing a memory storage device are used in synthesis. The synthesis code
`can be written into the memory chip with radiofrequency pulses (69). The memory
`chips are similar to EPROM chips (erasable, programmable, read-only memory.) A
`more comprehensive discussion of this technology is covered in Chapters 7 and 8.
`
`1.6 POSITIONALLY ADDRESSABLE SPATIAL ARRAYS
`
`Another approach to separationfidentification of combinatorial library collections is
`based on preparing each compoun