Synthesis and Applications of Small Molecule Libraries

Chem. Rev. 1996, 96, 555 600 555 Synthesis nd Applictions of Smll Molecule Librries Lorin A. Thompson nd Jonthn A. Ellmn* Deprtment of Chemistry, University of Cliforni, Berkeley, Cliforni 94720 Received August 17, 1995 (Revised Mnuscript Received October 18, 1995) Contents I. Introduction 555 II. Librries Synthesized on Solid Support 556 A. Librry Synthesis nd Evlution Strtegies 556 1. Discrete Compounds 556 2. Split Synthesis 557 3. Deconvolution of Soluble Librries 558 4. Structurl Determintion by Anlyticl 559 Methods 5. Encoding Strtegies 560 B. Synthesis of Orgnic Compound Librries 561 1. Introduction to Solid Supports 561 2. Post-Synthesis Peptide Modifiction 562 3. Biopolymer-Mimetic Librries 562 4. Nonoligomeric Compound Librries 566 5. Moleculr Recognition in Designed 590 Receptor Systems 6. Anlyticl Techniques 591 III. Librries Synthesized in Solution 592 A. Sptilly Seprte Synthesis 592 B. Synthesis in Pools 593 1. A Librry of Amides nd Esters 593 2. Acetylcholinesterse Inhibitors 594 3. Amides Displyed from Core Molecule 595 4. Oligoscchride Librries 595 IV. Future Directions 596 V. Acknowledgments 597 VI. Glossry 597 VII. Bibliogrphy 597 I. Introduction One of the initil steps in the development of therpeutic gents is the identifiction of led compounds tht bind to the receptor or enzyme trget of interest. Mny nlogs of these led compounds re then synthesized to define the key recognition elements for mximl ctivity. In generl, mny compounds must be evluted in both the led identifiction nd optimiztion steps. Incresing burdens hve been plced on these efforts due to the lrge number of new therpeutic trgets tht continue to be identified thorough modern moleculr biology methods. 1 To ddress this demnd, very powerful chemicl nd biologicl methods hve been developed for the genertion of lrge combintoril librries of peptides 2 nd oligonucleotides 3 tht re then screened ginst receptor or enzyme to identify high-ffinity lignds or potent inhibitors, respectively. While these studies hve clerly demonstrted the power of librry synthesis nd screening strtegies, peptides Lorin A. Thompson ws born in Lexington, KY, in 1970. He received the Bchelor of Science degree from the University of North Crolin, Chpel Hill, in 1992 where he worked under the guidnce of Joseph Desimone. He is currently pursuing his doctorte in the lbortory of Jonthn Ellmn t UC Berkeley where he is the 1994 Glxo-Wellcome fellow. His reserch interests include the development of synthetic methodology for orgnic librry construction. Jonthn Ellmn ws born in Cliforni in 1962. He received his S.B. degree from M.I.T. in 1984 where he worked in the lbortory of K. Brry Shrpless. He received his Ph.D. degree with Dvid A. Evns t Hrvrd University in 1989, where he worked on the synthesis of enntiomericlly pure nonproteinogenic mino cids, cyclopeptide lkloids, nd vncomycin. After n NSF postdoctorl fellowship with Peter G. Schultz t the University of Cliforni t Berkeley on the incorportion of unnturl mino cids into proteins, he joined the fculty t the University of Cliforni t Berkeley in 1992 s n ssistnt professor. His lbortory is engged in the development of new chemistry for the synthesis of orgnic compound librries, nd the ppliction of orgnic compound librries to different reserch problems in chemistry nd biology. nd oligonucleotides generlly hve poor orl ctivities nd rpid in vivo clernce; 4 therefore their utility s biovilble therpeutic gents is often limited. Due to the fvorble phrmcokinetic properties of mny smll orgnic molecules (<600-700 moleculr weight), 5 the design, synthesis, nd evlution of librries of these compounds 6 hs rpidly become mjor frontier in orgnic chemistry. 0009-2665/96/0796-0555$25.00/0 1996 Americn Chemicl Society

556 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn In ddition to the development of therpeutic gents, mny other pplictions of orgnic compound librries re currently being pursued. These include importnt dvnces in moleculr recognition, s well s recent efforts in mterils nd ctlysis. This review focuses on efforts towrd the synthesis nd evlution of smll orgnic molecule librries. Peptide nd oligonucleotide librries will not be described, since number of very thorough rticles hve recently reviewed this subject. 2,3 In ddition, greter emphsis will be plced upon librry synthesis methods thn upon ssy results becuse this Chemicl Reviews issue is devoted to frontiers in orgnic synthesis. II. Librries Synthesized on Solid Support The mjority of the compound librries tht hve been synthesized to dte hve been synthesized on solid support ( solid support is insoluble mteril to which compounds re covlently ttched during synthesis sequence). There re two dvntges to solid-phse synthesis strtegies. First, isoltion of support-bound rection products is ccomplished simply by wshing wy regents from the supportbound mteril, nd therefore rections cn be driven to completion by the use of excess regents. Second, innovtive methods re vilble for the mnipultion of discrete compounds nd for trcking the identity of compounds when compounds re ttched to solid support. A number of generl strtegies hve been developed for the synthesis nd evlution of compound librries synthesized on solid supports. Although most of these strtegies were initilly demonstrted with peptide librries, mny of these pproches hve now been pplied to other compound clsses. Mny of these strtegies hve been reviewed recently, 2 nd therefore only n overview of these methods will be provided with n emphsis on recent dvnces. A. Librry Synthesis nd Evlution Strtegies 1. Discrete Compounds. Sptilly Seprte Synthesis. Conceptully, the most strightforwrd method for the preprtion of compound librry is to synthesize mny compounds in prllel nd to keep ech compound in seprte rection vessel. When the finl compounds re sptilly seprte, the identity of compound t prticulr loction is known nd cn be confirmed by nlyticl methods. In ddition, biologicl evlution of the librry cn provide specific informtion bout ech compound in the librry. Of course, the compounds cn still be pooled ( pool refers to mixture of compounds) when the ssy of interest is not sufficiently high-throughput. Recently reported pooling strtegies my be prticulrly useful for this purpose. 7 A number of pproches for the prllel synthesis of orgnic compounds hve been reported. 8 The first method ws originlly developed by Geysen for peptide epitope mpping.9 Since tht time, severl improvements hve been introduced. 10 In this method, Figure 1. Diversomer TM pprtus. (Reprinted from ref 12. Copyright the Ntionl Acdemy of Sciences of the United Sttes of Americ.) 96 polyethylene pins re plced into supporting block so tht ech pin fits into seprte well of 96-well microtiter plte. Ech pin is coted with polymeric mteril tht is menble to solid-phse synthesis such s poly N,N-dimethylcrylmide. The polymeric mteril is derivtized with minolkyl groups to provide sites for substrte ttchment. During synthesis sequence ech pin is plced in seprte well of the microtiter plte so tht ech well serves s distinct rection vessel. With this pproch, on the order of 10 000 sptilly seprte compounds hve been prepred 11 in prllel by employing inexpensive lbwre, instrumenttion, nd utomtion developed predominntly for highthroughput microtiter-bsed screening efforts. Currently, pin loding levels rnge from 100 nmol to 50 µmol of mteril per pin. Even 100 nmol of mteril is sufficient for multiple biologicl ssys, s well s for nlyticl evlution of the purity nd chemicl integrity of the individul compounds. DeWitt nd co-workers hve reported the Diversomer TM pprtus, which is one of the first rection pprtus designed expressly for the prllel synthesis of smll orgnic molecules (Figure 1). 12 This pprtus is bsed on the use of porous gs dispersion tubes tht serve s continers for resin beds. Regents nd solvents re plced in up to 40 vils tht re locted in reservoir block, nd the ends of the gs dispersion tubes contining resin re plced into the vils llowing the regents to diffuse into the tubes nd contct the support. The temperture of the rection solutions cn be controlled by heting nd cooling the block. The pprtus is lso enclosed in mnifold with n injectble gsket employed for regent nd solvent dditions so tht rections cn be mintined under n inert tmosphere. Meyers nd co-workers hve recently reported conceptully relted pproch (Figure 2). 13 The rection pprtus is prepred by drilling hole into ech well of deep-well polypropylene microtiter plte. Porous frits re then plced t the bottom of ech well. The solid support nd rection solutions re plced into the wells of the pprtus. During rection step, the plte is clmped ginst Viton gsket in order to sel the hole t the bottom of the well. In between steps in the synthesis sequence, the rection solution cn be drined nd the resin cn

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 557 Figure 2. Multiple synthesis pprtus of Meyers nd coworkers. (Reprinted from ref 13. Copyright ESCOM Scientific publishers.) then be rinsed by removl of the Viton gsket. The rection pprtus employed in the three strtegies described bove re reltively inexpensive. This cn be n importnt considertion, since the prllel synthesis of hundreds to thousnds of sptilly seprte compounds will generlly require tht multiple rection pprtus be used simultneously. Clerly, when compounds re synthesized in pools, single prllel synthesis pprtus is usully sufficient for the synthesis of lrge numbers of compounds (vide infr). Commercil pprtus for prllel orgnic synthesis hve lso recently been developed. As one exmple, Advnced Chemtech hs developed the model 496 multiple orgnic synthesizer instrument for utomted chemicl synthesis. 14 The instrument is designed to produce 96 different compounds. A rnge of tempertures, mix times, nd speeds cn be employed. The instrument is lso comptible with wide rnge of rection conditions. Finlly, rections cn be performed under n inert tmosphere. b. Light-Directed, Sptilly Addressble Prllel Chemicl Synthesis. While librries of greter thn 10 000 compounds bsed upon single structurl type hve been prepred by prllel synthesis using seprte rection vessels, the synthesis of librries of hundreds of thousnds to millions of compounds per structurl type would not be prcticl by this pproch. Fodor nd co-workers hve developed strtegy bsed upon photolithogrphic methods tht cn be used to synthesize librries contining more thn 100 000 sptilly seprte compounds. 15 In this method silic wfer (borosilicte glss microscope slide) serves s the solid support (Figure 3). Aminolkyl groups or other rective functionlities tht re ttched to the surfce of the wfer re blocked with photolbile protecting groups. The photolbile protecting groups re cleved t specific regions on the silic wfer by site-specific illumintion using msks nd instrumenttion initilly developed for computer microchip construction. The silic wfer is then exposed to rection solution with rections occurring only t the regions tht were deprotected by illumintion. Only single photolbile group is necessry for the preprtion of oligomeric compounds, since monomer tht is protected with the sme photolbile group is introduced t ech step in oligomer synthesis. However, orthogonl photolbile protecting groups tht re cleved t different wvelengths of light would likely be required in order to differentite between different sites in nonoligomeric structures. The structure of finl compound t specific loction on the silic wfer is dependent upon the msking scheme nd Figure 3. Light-directed, sptilly ddressble chemicl synthesis. upon the order of ddition of the regents. Lrge numbers of sptilly ddressble compounds cn be prepred due to the exceptionl sptil resolution of photolithogrphy techniques. For exmple, Fodor nd co-workers hve ssembled peptide librries with ech peptide loclized to 50 µm squre. 15e This density llows the synthesis of 40 000 oligomers per centimeter squred, nd the current limits of photolithogrphy should llow synthesis to be performed t much higher densities. 16 In this sptilly ddressble rry, the loction of the compound on the silic wfer provides the structure of the compound. The compounds in librry therefore must be ssyed while still tethered to the wfer. Although this requirement imposes constrints on the rnge of fesible biologicl ssys, successful strtegies hve been developed for this purpose. One strtegy is to mesure the percent binding of soluble fluorescently lbeled receptor to different loctions on the silic wfer s monitored by epifluorescence microscopy. The silic support or the linker tht serves to ttch the compound to the support cn, however, hve pronounced effect on the binding ffinity. 17,15 2. Split Synthesis In generl, lrge librries of compounds re synthesized by employing pooling strtegies. The most direct method is to employ equimolr mixtures of rectnts in ech synthesis step. In fct, Geysen employed this strtegy in one of the erly peptide librry synthesis efforts. 18 One problem with this pproch is tht equimolr quntities of the finl compounds in the librry will only be obtined if ll of the rectnts in the mixture hve comprble rectivity, nd for most rection clsses, rectivity is highly dependent upon the structure of the rectnts. For the purpose of synthesizing peptide librries, severl reserchers hve demonstrted tht modest differences in rectivity cn be corrected by djusting the reltive concentrtions of the ctivted mino cids in the mixture. 19 It is lso possible to synthesize pproximtely equimolr mixtures of products by using totl of 1 equiv of mixture of regents in coupling step. 20 This strtegy, however, requires tht the concentrtion of ech regent in the mixture is firly low nd therefore is restricted to rections tht re very efficient.

558 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Figure 5. Deconvolution where X corresponds to n undefined position. Furk developed the split synthesis strtegy for the synthesis of librries of equimolr mixtures of compounds in order to overcome the complictions tht result from using mixtures of regents tht hve different rectivities. 21 This strtegy is currently the most populr method for the synthesis of compound mixtures. As illustrted in Figure 4, quntity of resin is split into equl sized portions tht re plced into seprte rection vessels. Excess of ech building block is employed to ensure tht ll of the rections re driven to completion. The resin from ll of the vessels is then recombined, mixed thoroughly, nd repportioned into the requisite number of rection vessels to perform the second synthesis step tht dds diversity. The second rection provides compounds tht incorporte ll of the possible combintions of the two sets of building blocks. By repeting the split, rect, nd mix opertions compounds incorporting ll possible combintions of the sets of different building blocks re theoreticlly generted, given tht enough beds re used. 22 The totl number of compounds tht re theoreticlly synthesized by this method is esily determined by multiplying together the number of building blocks tht re used in ech synthesis step. For oligomers the number of compounds equls X n where X is the number of compounds in the bsis set of monomers nd n is the length of the oligomer. Severl utomted systems hve lso been developed tht gretly expedite the repetitive split nd mix opertions. 6d,23 The instrument reported by Zuckermnn ws one of the first to be developed for this purpose. A number of different strtegies hve been developed for evluting librries tht re prepred by the split synthesis procedure due to the considerble chllenge of correct structurl elucidtion of the molecules hving the gretest biologicl ctivity. These strtegies re described in the subsequent sections. Figure 4. Split synthesis method. 3. Deconvolution of Soluble Librries Houghten pioneered one of the most populr strtegies for deconvoluting soluble librry fter clevge from the support. 24 Although this strtegy ws first employed for the evlution of peptide librries, it hs lso been used by mny reserchers for the evlution of smll molecule librries. In this pproch (Figure 5), pools of compounds re prepred such tht ech seprte pool hs defined building blocks t either one or two positions, nd t the remining positions ll combintions of building blocks re incorported. The optiml building block- (s) t the defined position(s) is selected by determining which pool(s) hs the gretest biologicl ctivity. A second round of synthesis is then performed with the selected building block(s) in plce t the initil defined position(s) in order to prepre pools where the next defined position is introduced. Ech pool is evluted for biologicl ctivity in order to select the optiml building block t the dditionl defined position. This deconvolution process of itertive resynthesis nd evlution is then repeted until ll of the positions re defined. For exmple, in the deconvolution sequence illustrted in Figure 5, three pools re produced tht ech contin defined building block t the lst position (G, H, or I) nd ll possible combintions of building blocks t the first nd second positions. For deconvolution, the pools re ssyed nd building block H is selected t the lst position since the pool

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 559 with H t the lst position hs the gretest ctivity. Three pools re now prepred in the second round of synthesis tht incorporte building block mixtures t the first position, defined building block t the second position (D, E, or F), nd the selected building block H t the lst position. Ech pool is evluted for biologicl ctivity nd building block D is selected for the second to lst position. Three finl pools re synthesized with defined building blocks t ech position; building blocks A, B, or C t the first position, building block D t the second position, nd building block H t the lst position. Evlution of these pools results in the identifiction of compound C-D-H. The deconvolution process described bove hs been pplied successfully to identify high-ffinity lignds, but severl issues should be considered. First, s the number of compounds in pool increses, lower concentrtions of ech compound must often be used in order to mintin the solubility of ll of the compounds in the pool. Compounds tht hve modest ctivity therefore my not be detected. Second, since the biologicl ctivity observed for given pool of compounds depends on both the ctivity nd bundnce of the ctive compounds in ech pool, the pool tht shows the gretest biologicl ctivity does not necessrily contin the most potent compound(s). In fct, for severl peptide librry studies the most ctive peptide ws not found in the most ctive mixture. 25 Third, the itertive resynthesis nd evlution of compounds cn be time-consuming nd lborious process. A number of modifictions hve been reported in order to ddress these complictions. Phrmgenics hs recently reported n ffinity selection process tht provides the ffinity of the most potent compounds in pool rther thn sum of the ffinity of ll of the compounds in the pool. 26 Jnd hs noted tht in split synthesis, resin cn be sved t ech step immeditely prior to resin mixing. 27 This resin cn lter be used s n intermedite in the itertive resynthesis nd deconvolution process resulting in considerble svings in time nd effort. Alterntive librry synthesis strtegies hve lso been reported tht provide direct determintion of high ffinity compounds without itertive resynthesis nd evlution steps, for exmple, the positionl scnning method of Houghten 25c,28 nd the orthogonl librry method of Trtr. 7 In the positionl scnning librry pproch developed by Houghten, seprte positionl librries re prepred, ech of which contins single defined building block. For exmple, if librry were prepred from three building block sets, three positionl librries would be prepred. These my be denoted s OXX, XOX, nd XXO, where O corresponds to the defined position in ech of the pools, nd X corresponds to rndomized position. Ech positionl librry is divided into seprte sublibrries with unique building block t the defined position in ech sublibrry. Ech positionl librry is then screened to directly determine the building block(s) t ech defined position tht contribute the most to biologicl ctivity. Although ctive compounds re identified without itertive resynthesis using this pproch, there is n incresed likelihood tht the most potent compound(s) will not be identified. 29 In ddition, the number of split nd mix opertions tht would be required for librry synthesis is sufficiently high tht building block mixtures re used in severl of the synthesis steps in order to reduce the number of opertions. Trtr demonstrted the orthogonl librry method by the cosynthesis of two librries tht contin orthogonl pools. 7 Using this strtegy, ny of the pools in the first librry will hve only one compound in common with ny of the pools in the second librry. During n ssy, ny highly ctive compound will cuse signl in only one pool in ech librry. By exmining which two pools re ctive, the compound cusing the ctivity cn immeditely be identified without resynthesis. This method ws demonstrted using 125 pools of 125 compounds ech, or totl of 15 625 trimers of nturl nd unnturl mino cids, nd llowed the identifiction of 63 nm inhibitor of vsopressin binding to LLCPK1 cells. However, s for the positionl scnning method, the uthors found it necessry to employ building block mixtures in ech of the synthesis steps in order to reduce the number of split nd mix opertions. 4. Structurl Determintion by Anlyticl Methods. When Assys Are Performed on Support- Bound Compounds. Methods hve lso been developed to ssy librries with compounds still ttched to resin beds. 30 This pproch is possible becuse the split nd mix opertions result in single compound structure being prepred on ech bed. Lm nd co-workers were the first to tke dvntge of the one compound per bed result of the split synthesis process. 6d,21f,31 In this pproch, the resin-bound librry is treted with lbeled soluble receptor. For mny studies fluorescent lbel hs been employed due to the sensitivity of fluorescence detection. The lbeled receptor binds to those beds tht re derivtized with molecules tht hve the highest ffinity to the receptor. The lbeled beds re then selected followed by structurl determintion of the support-bound compound. Very efficient, utomted methods hve been developed to select the lbeled beds, for exmple, by use of fluorescence ctivted cell sorting (FACS) instrument. 30b,d In order for this pproch to be successful, nlyticl methods must be vilble for the complete structurl determintion of minute quntities of the compound tht re present on single resin bed (vide infr). For this reson, this pproch hs only been employed to screen peptide nd oligonucleotide librries where very sensitive nd efficient sequencing methods hve lredy been developed, e.g. Edmn degrdtion nd DNA sequencing, respectively. A number of groups hve demonstrted tht mss spectrometry is sufficiently sensitive (electron impct, 32 electrospry, 31,33 mtrix-ssisted lser desorption, 34 imging time of flight secondry ion MS 35 )to observe moleculr ions of compounds tht re cleved from single resin beds. However, for lrge librry of molecules tht re of low moleculr weight, mny different compounds will hve the sme moleculr

560 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn weights severely complicting structurl determintion. 36 Youngquist hs developed novel strtegy tht should be pplicble to ll oligomer nd some nonoligomer librries. 34,c In his strtegy, smll percentge of cpping regent is dded t ech monomer ddition step so tht pproximtely 5-10% of the oligomer is cpped t ech step. After clevge of the finl oligomer product from the solid support, the oligomer sequence cn be red directly by mss spectrometric nlysis from the moleculr ions corresponding to the truncted nd cpped sequences tht re present in ddition to the full oligomer sequence. Youngquist hs demonstrted this strtegy in the synthesis nd evlution of peptide librries with cetic nhydride s the cpping regent. b. When Assys re Performed in Solution. As mentioned erlier, the linker or resin cn interfere with the interction of support-bound lignd with receptor. 15,17 To void these interctions, strtegies hve been developed for dividing the beds into seprte reservoirs nd then cleving portion of the compound on ech bed into solution for biologicl evlution. The collection of beds tht results in the gretest biologicl ctivity is then redistributed in smller pools. A second portion of compound is cleved from ech bed nd biologicl evlution is repeted to identify progressively smller set of beds. This process cn theoreticlly be repeted until the beds re rerryed singly for direct identifiction of the compounds responsible for ctivity. The sequence of the remining compound on the bed cn then be determined by the nlyticl methods described bove. In order to crry out this strtegy method must be vilble for the sequentil clevge of portion of the compound from the resin bed. Lebl nd co-workers hve developed multiple relese linkers for peptide synthesis. 37 One serious limittion to the generlity of this pproch is tht the rection conditions tht re employed in librry synthesis must not result in clevge of ny of the linkers. Bldwin nd co-workers hve reported tht single photolbile linker my lso be utilized for prtil relese of compounds into solution by controlled irrdition to dil in the percentge of photoclevge (eq 1). 38 The Geysen group hs lso reported photolbile linker tht llows controlled relese nd produces terminl crboxmide (eq 2). 33 5. Encoding Strtegies An lterntive strtegy to structurl determintion of compounds tht re synthesized on single beds is to employ n encoding strtegy. In this pproch, redble tgs tht record the rection sequence re Figure 6. Encoding strtegies. ttched to the resin bed concurrent with the synthesis of the compound on the bed. The tgs my be dded to provide n encoding sequence whereby the structure of the tg encodes for the building block nd the loction in the sequence encodes for the librry synthesis step (method A, Figure 6). Alterntively, ech tg my be dded individully to the resin (method B, Figure 6). In this cse the tgs must encode for both building block structure nd the step in the librry synthesis. Approches hve been developed to use encoding strtegies for the identifiction of compounds where ssys re performed on support-bound compounds or on compounds in solution fter clevge from the support.. DNA Encoding Strtegies. The encoding concept ws first proposed by Lerner nd Brenner where n oligonucleotide ws selected s the coding sequence. 39 This encoding strtegy relies on two strightforwrd but powerful tools commonly employed in moleculr biology reserch. Only minute quntity of the encoding oligonucleotide sequence is required becuse it cn be mplified employing the polymerse chin rection, nd the oligonucleotide sequence cn rpidly be red by DNA sequencing methods. A lrge number of building blocks my be encoded by this strtegy simply by employing multiple bses to encode for ech building block. Needels nd co-workers 30d nd Jnd nd coworkers 40 hve independently demonstrted this method for the synthesis of peptide librries. In both pproches orthogonl protecting group strtegies were necessry in order to differentite between the DNA tg synthesis steps nd the peptide synthesis steps. In prticulr, the bse-lbile Fmoc group ws employed for mine protection for peptide synthesis while the mildly cid-lbile dimethoxytrityl (DMT) group ws employed for hydroxyl protection for oligonucleotide synthesis. Needels employed 20:1 rtio of the binding peptide strnd to the encoding DNA strnd in order to minimize the effect of unwnted DNA-receptor interctions in binding studies. Jnd introduced linkge strtegy to cleve the peptide from support while still covlently linked to the oligonucleotide in 1:1 rtio for biologicl evlution in solution. For ppliction of these strtegies to orgnic compound librries, the comptibility of the encoding DNA strnd to the rection conditions necessry to prepre the orgnic structure should be considered nd my preclude mny synthesis pproches or limit the ccessibility of certin compound clsses. In ddition, the synthesis strtegy for librry preprtion must ccommodte protecting group scheme tht llows synthesis of the coding DNA sequence to be differentited from the synthesis of the orgnic compound. b. Peptide Encoding Strtegies. An lterntive encoding strtegy relies on the use of peptides to code for nonnturl oligomers or orgnic com-

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 561 Figure 7. Hloromtic tgs ttched through mide bond formtion. pounds. In this strtegy first demonstrted by Zuckermnn nd co-workers, coding peptide strnd ws employed to encode binding peptide-bsed oligomer strnd. 41 Orthogonl protecting group strtegies were necessry in order to differentite the synthesis of the binding strnd from the synthesis of the coding strnd. In Zuckermnn s strtegy the bselbile Fmoc group ws employed for the binding strnd nd the mildly cid-lbile N-[(4-methoxybenzyl)oxy] (Moz) or N-[[2-(3,5-dimethoxyphenyl)prop- 2-yl]oxy] (Ddz) groups were used for the coding strnd. The binding strnd, linked to the coding strnd, cn be cleved from the support by tretment with strong cid so tht the resulting complex cn be ssyed in solution. Subsequent Edmn degrdtion of the coding strnd then provides the binding oligomer sequence. One potentil compliction with this pproch is tht the coding strnd could lso interct with the receptor, lthough this ws not observed in the reported study. A relted peptide encoding strtegy hs been reported by Nikoliev nd co-workers. 42 In this study Fmoc protection ws used for the binding strnd nd Boc protection ws used for the coding strnd. At the end of the synthesis sequence the binding strnd could be relesed from the resin for biologicl evlution, while the encoding strnd is retined on the bed for lter structurl elucidtion. This procedure requires tht the solution contining the binding strnd be sptilly ddressble to the bed from which it originted to llow decoding. As for DNA encoding, the ppliction of either of these strtegies to the synthesis of orgnic compounds requires protecting group strtegy tht llows the synthesis of the coding peptide sequence to be differentited from the synthesis of the orgnic compound. c. Hloromtic Tg Encoding Strtegies. Still hs developed n encoding strtegy tht employs moleculr tgs tht re highly stble to mjority of rection conditions nd therefore should introduce few constrints upon compound synthesis. 44 Still employs hloromtic compounds s moleculr tgs tht cn be detected t levels of less thn 0.1 pmol using electron-cpture GC. Due to the sensitivity of nlyticl evlution, ech tg is ttched to the support t less thn 1% of the loding level of the synthesized compound. This minimizes the chnce of complictions in librry evlution resulting from tg-receptor interctions when support-bound ssys re used. The tgs re synthesized by lkyltion of commercilly vilble hlophenol with n ω-bromo-1-lknol, followed by ttchment of linker with relese site. In the initil reports of this strtegy, 44b photolbile o-nitrobenzyl lcohol linker served s the relese site, nd the tgs were ttched to the solid support through mide bond formtion with support-bound rective mine (Figure 7). Figure 8. Hloromtic tgs ttched through crbene insertion. In subsequent work, Nestler nd co-workers hve lso employed n oxidtively lbile linker tht is cleved with ceric mmonium nitrte, nd the tgs re directly ttched to the resin beds through rhodium-ctlyzed cylcrbene insertion (Figure 8). 44 With this pproch, in contrst to ll of the other encoding strtegies, protection scheme is not necessry to differentite the incorportion of the encoding tgs from compound synthesis. The crbene my rect with both the support nd the compound ttched to the support; however, becuse the rection is not very selective, the crbene inserts predominntly into the support due to the greter proportion of the support. Furthermore, becuse the tg loding level is only 1% of the compound loding level, compound modifiction should be miniml. The tgs must encode for both the building block nd for the rection step in the synthesis sequence. This is ccomplished by employing binry code formt. Multiple tgs re used for ech building block with the presence or bsence of tg corresponding to 1 or 0, respectively, in binry sequence. This strtegy hs been employed both for support-bound nd solution-bsed ssys. After the submission of this mnuscript, n innovtive method for encoding bsed on rdiofrequency trnsponders ws independently reported by two reserch groups. The trnsponders re encsed in glss nd re completely solvent, regent, nd temperture resistnt. They cn esily be scnned fter every rection step to record the identity of compound on btch of resin. 43 B. Synthesis of Orgnic Compound Librries 1. Introduction to Solid Supports Two types of supports were used in lrge mjority of the studies tht re described in this review. Polystyrene cross-linked with 1-2% divinylbenzene, nd polystyrene-polyethylene glycol copolymer (PEG-PS). Although other supports hve been employed in solid-phse orgnic synthesis, nd new supports re currently being developed, n overview of only these two most commonly used resins is provided. 45 In ddition, unless specificlly stted otherwise, ll synthesis sequences detiled in this review were crried out on polystyrene resin. Polystyrene cross-linked with 2% divinylbenzene ws demonstrted by Merrifield s useful support for solid-phse peptide synthesis in 1963, 46 nd polystyrene cross-linked with 1-2% divinylbenzene continues to be one of the most commonly used supports for solid-phse orgnic synthesis. Polystyrene resin is dvntgeous in tht it is the lest expensive of the resins, high loding levels cn be chieved, nd it is reltively mechniclly stble. One mjor limittion of polystyrene resin is tht the resin beds re not well solvted in protic solvents result-

562 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Tble 1. Side-Chin Modifictions of Amino Acids in Permethylted Peptides prent chin side-chin modifiction nonmethylted purity, % methylted purity,,b % CGGFL-NH 2 methyl thioether 86 50 DGGFL-NH 2 methyl ester 97 60 EGGFL-NH 2 methyl ester 90 75 HGGFL-NH 2 methyl imidzole 98 40 KGGFL-NH 2 quternry slt 99 30 LGGFL-NH 2 unmodified 99 81 M[O]GGFL-NH 2 unmodified 98 70 NGGFL-NH 2 dimethyl mide 85 86 QGGFL-NH 2 dimethyl mide 85 80 RGGFL-NH 2 trimethyl gunidine 88 75 WGGFL-NH 2 methyl indole 98 70 YGGFL-NH 2 methyl ether 99 81 Estimted purity s determined by HPLC. b Sum of the mono-, di-, nd trimethylted R-mine products. ing in poor rection site ccessibility nd diminished rection rtes. There re two forms of PEG-PS resin. Rpp nd Byer hve prepred PEG-PS resin by mens of nionic polymeriztion of ethylene oxide to ttch PEG chins of controlled lengths on cross-linked polystyrene beds contining hydroxyl groups (the resin is mrketed s Tentgel resin). 47 In contrst, Millipore hs prepred PEG-PS by ttching preformed polyethylene glycol onto the polystyrene bed. Both of these resins re well solvted in protic solvents, in contrst to the polystyrene resin. In ddition, becuse the polyethylene glycol chins re not cross-linked, it hs been rgued tht the rection sites re more highly ccessible resulting in greter rection rtes. 47 Due to the lck of cross-linking in the polyethylene glycol chins, in combintion with good queous solvtion chrcteristics, PEG-PS hs been the fvored resin for support-bound ssys. Limittions of PEG-PS reltive to polystyrene include much higher cost, reduced loding levels, nd significnt mechnicl instbility (stirring or vigorous shking results in significnt loss of mteril from the resin). Scheme 1 2. Post-Synthesis Peptide Modifiction Houghten hs expnded upon his seminl work in the re of peptide librries 2b to include the postsynthesis modifiction of peptides in order to extend the diversity of the compounds tht cn be ccessed nd to provide oligomers tht hve improved phrmcokinetic properties. In his first study in this re, Houghten permethylted protected peptides while they were still ttched to the solid support. 48 Permethyltion ws ccomplished by treting the peptide librries with excess sodium hydride in DMSO followed by ddition of methyl iodide (Scheme 1). Under optimized conditions, permethyltion of the peptide AGGFL-NH 2, which does not contin rective side chin functionlity, provided the permethylted product in >90% yield nd purity. Permethyltion of peptides tht incorporte mny of the different side chin functionlized mino cids ws lso investigted. As shown in Tble 1, methyltion of the functionlized side chins generlly ws observed, s were lower levels of purity of the permethylted products. For ech of the peptides, s much s 15% of the terminl mino group hd not been fully quternized in the permethyltion product. Longer rection times could be employed to drive the quterniztion to completion, but lso resulted in pprecible degrdtion of the product. Houghten demonstrted tht miniml epimeriztion occurs in the deprotontion step by tretment of ll four possible stereoisomers of GGFL-NH 2 with NH in DMSO followed by wter quench. Subsequent clevge from the support nd HPLC nlysis showed tht less thn 0.75% epimeriztion hd occurred. Employing the bove post-synthesis methyltion strtegy, Houghten performed the synthesis nd evlution of librry tht theoreticlly contined 37 791 360 unique permethylted hexpeptides using the positionl scnning pproch (see section II.A.3). In prticulr, the librry ws screened for the bility to inhibit the growth of five different strins of bcteri or yest. A number of permethylted peptides tht incorported multiple phenyllnines were identified tht hd significnt ctivity ginst Stphylococcus ureus, methicillin-resistnt S. ureus, nd Stphylococcus snguis, with IC 50 vlues in the 1-10 µm rnge. A number of reserchers hve lso reported on the post-synthesis modifiction of individul residues of peptides within librry, such s cyltion or reductive mintion of the N-terminus, or functionliztion of rective side chin functionlity. However, these studies re outside the scope of this review nd will not be discussed here. 3. Biopolymer-Mimetic Librries Although peptides generlly hve poor phrmcokinetic properties tht limit their utility s drugs, the high level of success tht hs been chieved in identifying high-ffinity lignds to diverse receptors nd enzymes through the synthesis nd evlution of peptide librries is well documented. 2 Mny reserchers hve therefore focused on the synthesis nd evlution of biopolymer mimetics tht lthough bsed upon the peptide structure, incorporte bckbones tht my hve improved phrmcokinetic

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 563 Tble 2. Pentmer Peptoids Prepred by the Submonomer Method (Scheme 2, Method B) pentmer chrcteriztion of unpurified product side chin (R) purity, % mss recovery, b % (CH 2)CH 3 >85 90 (CH 2)Ph >85 74 (CH 2)Chx >85 79 (CH 2)CH(Ph) 2 >85 70 Ph >85 83 c-c 3H 5 >85 83 CH 2) 2indole >60 52 (CH 2) 3NH 2 >85 63 c Purity of unpurified compounds s determined by RP- HPLC. b Determined from dry weight. c Prepred with Boc- NH-(CH 2) 3NH 2. Figure 9. Unnturl biopolymers. properties s result of proteolytic stbility, more fvorble solubility chrcteristics, or unique structurl nd/or hydrogen bonding motifs. Biopolymer mimetics re lso of considerble interest to other reserch res, including the design of two- nd three-dimensionl unnturl biopolymer frmeworks with novel properties, including the design of moleculr receptors (vide infr). Peptoids. Simon nd co-workers considered number of criteri for design of new scffold including simple synthesis of monomers, incresed resistnce to hydrolytic enzymes, the bility to disply wide rnge of functionlity, high-yielding coupling steps menble to utomtion, nd the use of chirl monomers. 49 Oligo(N-substituted)glycines, or peptoids (Figure 9) were proposed to meet these requirements. The side chins of peptoids re displyed from the mide nitrogen of n oligoglycine bckbone insted of the R-crbon tom, providing protese-resistnt 50 nd chirl tertiry mine linkge. In the originl synthesis of peptoids, stndrd solid-phse peptide synthesis methods were employed with Fmoc-protected N-lkylglycines serving s the monomer components (method A, Scheme 2). 49 Scheme 2 (Benzotrizol-1-yloxy)tris(pyrrolidino)phosphonium hexfluorophosphte (PyBOP) or bromotris(pyrrolidino)phosphonium hexfluorophosphte (PyBroP) were found to be the optiml coupling gents, llowing the synthesis of up to 25-mers in high yield nd excellent purity s judged by MS nd RP-HPLC nlysis. Subsequently, simplified pproch ws reported by Zuckermnn nd co-workers. In the submonomer pproch (method B, Scheme 2), 51 ech cycle of synthesis involves two-step procedure: (1) mide bond formtion with R-bromo cetic cid employing 1,3-diisopropylcrbodiimide (DICI) s the ctivting gent nd (2) bromide displcement with suitble primry mine to provide the secondry N-lkylglycine (2) redy for the next coupling step. This pproch llows for the direct incorportion of commercilly vilble mines s building blocks; thereby eliminting costly, time-consuming monomer synthesis nd obviting the need for R-mine protection. As shown in Tble 2, the mss blnce nd purity of unpurified mteril is high for the synthesis of pentmers tht incorporte rnge of mine nucleophiles, including R-brnched mines nd unrective niline derivtives. In ddition, nonomer with five consecutive propylmine side chins followed by four consecutive butyl side chins (not listed in tble) ws obtined in greter thn 65% purity s determined by RP-HPLC nd in 86% crude mss blnce, demonstrting the efficiency of the synthesis sequence. Using the submonomer pproch, Zuckermnn nd co-workers ssembled librry on Merrifield beds using the split-synthesis pproch tht ws bised towrd the 7-trnsmembrne G-protein coupled clss of receptors. 52 The Chiron group synthesized the librry of pproximtely 5000 dimer nd trimer peptoids employing 23 different monomers s well s three terminl mine cpping regents (Figure 10). Seven monomers were selected bsed upon the structures of known lignds to this clss of receptors, while 17 monomers were selected to be s diverse s possible. The Chiron group hs recently described their experimentl design to mximize moleculr diversity for given librry or to trget librry with key fetures for given receptor. 53 Initil studies were performed to demonstrte tht ll of the mines were fully incorported into test oligomer. The librry ws then synthesized using Zymrk robot to crry out resin mnipultions nd micropipetting 54 nd ssyed s 18 pools of 216, 255, or 272 compounds for the bility to inhibit [ 3 H]- proszin binding to n R 1 -drenergic receptor preprtion. By using the itertive resynthesis nd evlution strtegy, 5nMinhibitor ws identified (3, Figure 11). Similrly, 6 nm inhibitor of [ 3 H]- DAMGO binding to µ-specific opioid receptor preprtion ws identified (4, Figure 11). The peptoid products re completely stble to proteolysis s

564 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn determined by incubting number of peptoid derivtives with crboxypeptidse A, chymotrypsin, elstse, ppin, pepsin nd thermolysin. 50b Additionl diversity hs lso been chieved through modifiction of peptoid side chins. Pei nd Moos hve described procedure for modifiction of peptoids tht hve lkenyl or lkynyl side chins by [3 + 2] cycloddition rection with nitrile oxides to produce support-bound isoxzoles nd isoxzolines. 55 The products re formed with high regioselectivity in >80% purity by HPLC nlysis, nd the uthors conclude by the lck of byproducts tht the chemicl yields re similr, lthough only one mss blnce (72%) of n unpurified product is provided. The Chiron group hs lso reported multistep modifiction of peptoid structures to provide six- nd sevenmembered ring heterocycles. These studies re described in lter sections. Oligocrbmtes. Another unnturl biopolymer-bsed librry ws reported by Cho nd coworkers. 17,56 Their method involves the use of mino cids s precursors to chirl building blocks for the synthesis of librry of oligocrbmtes. N-Protected p-nitrophenyl crbonte monomers 6 were prepred by reduction of the corresponding protected mino cids followed by tretment of the resulting lcohols 5 with p-nitrophenyl chloroformte (Scheme 3). Oligomers were initilly ssembled on polysty- Scheme 3 Figure 10. Monomer side chins nd cpping regents used in peptoid librry synthesis. Figure 11. Peptoid lignds. rene resin using the Fmoc group s the mine protecting group nd stndrd Rink mide linker. Two rections were performed in ech coupling cycle: (1) removl of the Fmoc group by tretment with 20% piperidine in N-methylpyrrolidinone (NMP) nd (2) coupling with n N-Fmoc crbonte monomer employing hydroxybenzotrizole (HOBt) s n dditive. Coupling yields were determined to be >99% by RP-HPLC nd quntittive ninhydrin tests, 57 nd oligocrbmtes were chrcterized by FAB-MS nd 1 H NMR spectroscopy. Two representtive oligocrbmtes were lso incubted with trypsin nd porcine pepsin nd were found to be stble to proteolysis. The photolithogrphy technique developed by Fodor nd co-workers (section II.A.1.b) ws then used to prepre librry of oligocrbmtes. 17 In order to introduce protecting group scheme tht ws comptible with photolithogrphy, the crbmte monomers were prepred with the photolbile [(nitrovertryl)oxyl]crbonyl (Nvoc) group in plce of the Fmoc group. An eight-step binry msking scheme ws then employed to synthesize librry of 256 oligocrbmtes round the prent sequence AcY c F c A c - S c K c I c F c L c (where X c refers to the crbmte monomer formed from the mino cid X) such tht the librry contined ll of the possible deletion sequences for the prent oligomer. The librry ws ssyed for the bility to bind the monoclonl ntibody (mab) 20D6.3, which ws rised to the keyhole limpet hemocynin conjugte of oligocrbmte AcYc K c F c L c G-OH (G-OH is terminl glycine residue). Binders were identified by scnning epifluorescent microscopy using got R-mouse fluorescein-conjugted secondry ntibody. Five out of the 10 highest ffinity oligocrbmtes, AcK c F c L c G-OH, AcF c K c F- c L c G-OH, AcY c K c F c L c G-OH, AcA c K c F c L c G-OH, nd AcI c F c L c G-OH, were resynthesized nd purified on lrge scle using Fmoc chemistry. The IC 50 vlues of ll five lignds determined in solution were in the 60-180 nm rnge. These studies lso reveled tht the support or linker cn interfere with receptor interctions, since the lignd AcY c F c L c G-OH ws lso

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 565 Scheme 5 Scheme 4 prepred on lrge scle nd ssyed in solution nd found to hve n IC 50 vlue of pproximtely 160 nm, even though this lignd rnked in the bottom 30% of lignds in the support-bound ssy. Oligoures. Burgess hs reported on initil studies towrd the solid-phse synthesis of oligoures with the gol of synthesizing oligoure librries (Scheme 4). 58 The monomers 10 for oligoure synthesis were prepred on verge in 50-60% overll yield by the three step process of reduction of the corresponding N-Boc-protected mino cid followed by converting the resulting primry lcohol 9 to the phthlimide under Mitsunobu rection conditions nd finl removl of the Boc protecting group. In the solid-phse synthesis of the oligoure, two rections re performed in ech coupling cycle: (1) removl of the phthloyl group by tretment with 60% hydrzine hydrte in DMF nd (2) coupling with monomer tht is ctivted in situ s the isocynte by tretment with triphosgene. Two oligoure/peptide hybrids, 13 nd 14 (Figure 12), were prepred by this generl pproch in unoptimized 46% nd 17% isolted yield, respectively. Vinylogous Sulfonyl Peptides. Gennri nd Still hve reported on the solid-phse synthesis of vinylogous sulfonyl peptides with the gol of mking sulfonyl peptide librries s well s in employing these biopolymer mimetics s synthetic receptors. 59 The uthors focused upon the vinylogous sulfonmide structure becuse the sulfonmide my mimic the tetrhedrl geometry of mide hydrolysis in the protese clevge of peptide bonds thereby serving s n interesting phrmcophore for protese inhibition. In ddition, the stronger polriztion of the sulfonmide bond compred to regulr peptide bond fvors the formtion of hydrogen bonds to potentilly provide more distinct preorgniztion of the sulfonyl peptides. The monomers 16 for oligomer synthesis were prepred on verge in 65-75% overll yield by three step process from the corresponding N-Bocprotected mino ldehyde (Scheme 5). A Wittig- Horner-Emmons rection provides the R,β-unsturted sulfonte 15, which is then converted into the sulfonyl chloride by deprotection with tetrbutylmmonium iodide followed by ctivtion with SO 2 Cl 2 nd PPh 3. 60 In the solid-phse synthesis of the sulfonyl peptides, two rections re performed in ech coupling sequence: (1) removl of the Boc group by tretment with trifluorocetic cid (TFA) nd (2) four cycles of coupling with 1 equiv of vinylsulfonyl chloride 16 followed by slow ddition of 1,8-dizbicyclo[5.4.0]undecne (DBU). Four cycles of coupling nd slow ddition of DBU is necessry, since excess DBU results in decomposition of the vinylsulfonyl chloride monomers, but DBU is the only bse tht ws studied tht provides clen conversion to product. By using Tentgel resin, ll five monomers, 16 to 16e, were coupled to the support-bound Gly ester followed by clevge from the support by tretment with 10% triethylmine in methnol. The methyl ester products were isolted in good overll yield, 60-70%. In ddition, two sulfonyl dipeptides were prepred to demonstrte oligomer synthesis by the bove method. Sulfonyl dipeptides 19 nd 20 were obtined in 57% nd 52% isolted yields, respectively (Figure 13). In ddition, only one distereomer of 19 ws detected indicting tht no rcemiztion occurred either in monomer synthesis or in the coupling steps. Figure 12. support. Oligure/peptide hybrids prepred on solid Figure 13. Vinylogous sulfonyl peptides. Vinylogous Peptides. In one of the erliest studies on unnturl biopolymers, Schreiber nd coworkers reported on the synthesis of vinylogous peptides nd demonstrted tht these structures cn dopt specific secondry s well s tertiry structures (Figure 14). 61 The vinylogous peptide 21 ws observed to dopt stcked rry of prllel sheets s determined by X-ry crystllogrphy, while the pep-

566 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Figure 14. Vinylogous peptide nd vinylogous peptide/ mide hybrids. tide/vinylogous mide hybrids 22 nd 23 were determined to dopt structures corresponding to ntiprllel sheets nd helicl conformtions, respectively, with turns bout the Pro-Gly bonds s determined by NMR. In this study, the compounds were prepred in solution for the purpose of structurl evlution; however, the monomer synthesis is expedient nd solid-phse synthesis strtegies for the purpose of librry construction should be strightforwrd using stndrd peptide synthesis methods. Smith nd Hirschmnn hve developed n unnturl biopolymer bsed upon linked pyrrolin-4-ones tht lso incorportes vinylogous mide into its structure (Figure 9). 62 Although these oligomers hve not been employed for librry synthesis, they do dopt secondry structure tht mimics β-strnd 62c nd hve been utilized in the synthesis of potent protese inhibitors. 62b 4. Nonoligomeric Compound Librries Nonoligomeric molecules, which re nonpeptidic in nture nd re below 600-700 in moleculr weight, hve become the mjor focus of librry synthesis efforts for the development of medicinl gents. Librries of smll nonoligomeric molecules hve been prepred both for the identifiction of led compounds nd for the optimiztion of led compounds tht hve been identified through either librry screening efforts or lterntive methods. While the compound clss for librry synthesis towrd led optimiztion is predetermined, number of fctors must be considered in the selection of compound clss for the purpose of led identifiction. To vrying degrees, three generl strtegies for compound selection hve been used for the librry syntheses outlined in this review. 63 The first strtegy is to select privileged structures, 64 where the disply of different functionlity upon the structure hs previously provided number of potent nd specific therpeutic gents or cndidtes towrd different therpeutic trgets. The second strtegy is to design compound scffolds bsed on importnt recognition elements of biologicl receptors. The finl strtegy is to select stble compounds upon which few therpeutic gents or cndidtes hve been bsed, but which re strightforwrd to prepre with multiple sites vilble for the disply of functionlity. In designing synthesis scheme to ccess librry bsed upon specific compound clss three criteri should be considered: (1) The synthesis scheme should provide the mjority, if not ll, of products in the librry in high yield nd purity. (2) The chemistry should be comptible with the disply of s much diverse functionlity s possible including heterotom functionlity tht is commonly found in drugs such s lcohols, phenols, mines, indoles, gunidines, crboxylic cids, mides, nitriles, imidzoles, nitro groups, nd hlides. (3) The building blocks for synthesis of the librry should be commercilly vilble or t lest redily ccessible, since librry cnnot be mde rpidly nd efficiently if mny of the building blocks must be prepred. 65 Finlly, it should be emphsized tht for given compound clss, we nd others hve found tht developing high-yielding nd generl synthesis sequence followed by rigorously estblishing tht diverse rry of functionlity cn be displyed cn require significnt effort. In contrst, using the optimized synthesis sequence librries cn be constructed rpidly nd efficiently s long s the building blocks re commercilly vilble or redily ccessible.. Heterocycle Librries. 1. Seven-membered Rings. All of the efforts towrd the synthesis of librries of seven-membered ring heterocycles hve focused upon the 1,4-benzodizepine structure. The 1,4-benzodizepine clss of compounds hve widespred biologicl ctivities nd re one of the most importnt clsses of biovilble therpeutic gents. In ddition to 1,4-benzodizepines such s Vlium tht hve nxiolytic ctivity, 66 there re lso derivtives tht re highly selective cholecystokinin (CCK) receptor subtype A ntgonists, highly selective CCK receptor subtype B ntgonists, 67 κ-selective opioids, 68 pltelet-ctivtion fctor ntgonists, 69 HIV Tt ntgonists, 70 reverse trnscriptse inhibitors, 71 gpiibiii inhibitors, 71b nd rs frnesyltrnsferse inhibitors. 72 1,4-Benzodizepin-2-ones. In one of the first rticles to ddress the synthesis nd evlution of smll molecule combintoril librries, Bunin nd Ellmn reported the solid-phse synthesis of 1,4- benzodizepine derivtives. 73 In this initil report, benzodizepine derivtives were constructed from three components: 2-minobenzophenones, mino cids, nd lkylting gents. By employing solution chemistry, hydroxyl-substituted 2-N-Fmoc-minobenzophenone is coupled to the [4-(hydroxymethyl)phenoxy]cetic cid (HMP) linker. 74 The linkerderivtized minobenzophenone 24 (Scheme 6) is Scheme 6 then coupled to the solid support by employing stndrd mide bond-forming methods (the linker my be ttched to either ring of the 2-minobenzophenone). Synthesis of the benzodizepine derivtive on solid support then proceeds by removl of the Fmoc

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 567 Scheme 7 Tble 3. 1,4-Benzodizepine Derivtives 30 (Scheme 7) derivtive yield entry R 1 R 2 R 3 R 4 (%) 1 4 -OH 5-Cl CH 3 H 95 2 4 -OH 5-Cl CH 3 CH 3 100 3 4 -OH 5-Cl CH 3 CH 2CH 3 97 4 4 -OH 5-Cl CH 3 CH 2CHdCH 2 90 5 4 -OH 5-Cl CH(CH 3) 2 CH 2CH 3 85 6 4 -OH 5-Cl CH 2CO 2H CH 2CH 3 95 7 4 -OH 5-Cl (CH 2) 4NH 2 CH 2CH 3 95 8 4 -OH 5-Cl CH 2Ph(4-OH) CH 2CH 3 98 9 4-CO 2H,5-Cl CH 2Ph CH 3 100 10 4-CO 2H,5-Cl CH 3 CH 2Ph 93 Yields of purified mteril bsed on support-bound strting mteril 26 (Scheme 7). protecting group from 26 by tretment with piperidine in DMF followed by coupling n R-N-Fmoc mino cid fluoride to the resulting unprotected 2-minobenzophenone (Scheme 7). The ctivted R-N-Fmoc mino cid fluoride 75 is employed in order to chieve complete conversion to the mide product 27 even for electron-deficient 2-minobenzophenone derivtives. The Fmoc protecting group is then removed, nd the resulting free mine is treted with 5% cetic cid in NMP to provide the initil benzodizepine derivtive 28. Alkyltion of the nilide of 28 then provides the fully derivtized 1,4-benzodizepine 29. To mximize synthesis generlity, lithited 5-(phenylmethyl)-2- oxzolidinone 76 or lithited cetnilide is employed s the bse since it is bsic enough to completely deprotonte the nilide of 28, but not bsic enough to deprotonte mide, crbmte, or ester functionlity. By employing these conditions 1,4-benzodizepine derivtives contining esters nd crbmtes were lkylted in high yield on solid support with no overlkyltion observed (entries 6 nd 7, Tble 3). Tretment with stndrd trifluorocetic cid clevge regents ffords the benzodizepine products 30 in high yield (85-100%, 95% v fter purifiction bsed on support-bound strting mteril 26). Finlly, no rcemiztion of the mino cid component is detected (<1%) s determined by chirl HPLC. A smll librry of 192 benzodizepines ws then prepred in order to demonstrte sptilly seprte Figure 15. Building blocks used in the synthesis of 192- member benzodizepine librry. librry synthesis of these compounds by employing the Chiron mimotopes pin support (see section II.A.1.). 77 The librry of 192 compounds ws ssembled using ll combintions of two 2-minobenzophenones, 12 mino cids, nd eight lkylting gents with vriety of functionlity being displyed (Figure 15). The chemicl integrity nd yield of mny of the compounds in the librry were determined by two nlyticl methods. For 28 of the structurlly diverse benzodizepine derivtives, FAB mss spectrometry confirmed the structure of the compound corresponding to the mjor pek (in lmost ll cses the only pek) observed by HPLC. Ech of the 2-minobenzophenones, mino cids, nd lkylting gents ws incorported into t lest one of the 28 derivtives, indicting tht ll of the building blocks were comptible with the generl solid-phse synthesis sequence. Yields were lso determined for 20 derivtives, where gin ech of the 2-minobenzophenones, mino cids, nd lkylting gents ws incorported into t lest one of the derivtives. This ws ccomplished by ddition of stock solution contining fluorenone s n internl stndrd followed by HPLC nlysis. An 86% verge yield ws then clculted on the bsis of the extinction coefficients of the 1,4-benzodizepine derivtives nd the reltive pek re of the benzodizepines to the pek re of the internl stndrd. The sptilly seprte librry of benzodizepines ws then screened to identify lignds to the cholecystokinin A receptor using competitive rdiolignd binding ssy. Detiled structure versus ctivity informtion ws obtined towrd this receptor trget.

568 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn The dt provided by screening the librry ws confirmed by synthesizing number of the derivtives on lrge scle followed by purifiction nd IC 50 determintions. The most potent compound (IC 50 ) 0.01 µm) incorported 2-mino-4-hydroxybenzophenone, D-tryptophn, nd methyl iodide s the lkylting gent. In ddition, the structure ctivity dt correlted closely with Merck s dt on structurlly relted benzodizepine derivtives. 78 By employing the synthesis strtegy described bove, second librry of 1680 benzodizepine derivtives hs lso been prepred from three 2-minobenzophenones, 35 mino cids, nd 16 lkylting gents (structures not shown). The librry hs been screened ginst number of receptor nd enzyme trgets. Inhibitors of pp60 s-src tyrosine kinse 79 nd lignds tht block n utoimmune DNA-ntibody interction implicted in systemic lupus erythemtosus 80 hve been identified. The originl benzodizepine synthesis sequence ws bsed upon the combintion of three different building block sets: 2-minobenzophenone, n Fmoc mino cid fluoride, nd n lkylting gent. While lkylting gents re commercilly vilble, nd N-Fmoc mino cid fluorides cn be prepred in single step without purifiction from the corresponding N-Fmoc mino cids, few ppropritely functionlized 2-minobenzophenones re redily ccessible. To increse the diversity of 1,4-benzodizepin-2-ones vilble through solid-phse synthesis, Plunkett nd Ellmn utilized the Stille coupling rection to synthesize vriety of 2-minoryl ketones on solid support. 81 The Stille rection is prticulrly ppeling for this purpose since it proceeds under mild conditions, is tolernt of wide rnge of functionlity, nd well over 300 structurlly diverse cid chloride building blocks re commercilly vilble. A [[2-(4-biphenyl)isopropyl]oxy]crbonyl (Bpoc)- protected (minoryl)stnnne, which is prepred in four steps in solution, is coupled to the solid support through the HMP linker (Scheme 8). Stille coupling Scheme 8 cn then be crried out with rnge of different cid chlorides nd the ctlyst Pd 2 (db) 3 CHCl 3. The Bpoc group is cleved by brief tretment (5 min) with 3% TFA in CH 2 Cl 2. The support-bound 2-minoryl ketones re then incorported directly into 1,4-benzodizepine derivtives ccording to the previously described synthesis sequence. Using this strtegy, diverse cid chlorides were employed to prepre support-bound 2-minoryl ketones 32 tht were further incorported into 1,4- benzodizepines 33, including romtic cid chlorides tht re electron-rich, electron-poor, lkyl-substituted, polyromtic, heterocyclic, nd ortho-substituted nd liphtic cid chlorides tht cn be stericlly hindered (Tble 4). The desired benzodizepines were isolted fter the eight-step synthesis sequence yield Tble 4. 1,4-Benzodizepin-2-one Derivtives 33 (Scheme 8) derivtive entry R 1 R 2 R 3 (%) 1 C 6H 4-2-OCH 3 CH 3 CH 2CH 3 67 2 C 6H 4-3-OCH 3 CH 3 CH 2CH 3 79 3 C 6H 4-4-OCH 3 CH 3 CH 2CH 3 73 4 c-c 6H 11 CH 3 CH 2CH 3 70 5 C 6H 4-2-CH 3-3-CN CH 3 CH 2CH 3 64 6 C 6H 4-3-CF 3 CH 3 CH 2CH 3 74 7 2-thienyl CH 3 CH 2CH 3 52 8 2-furyl CH 3 CH 2CH 3 59 9 dmntyl CH 3 CH 2CH 3 80 10 C 6H 4-2-Cl CH 3 CH 2CH 3 52 11 c-c 5H 10-1-C 6H 5-4-Cl (CH 2) 2CO 2H CH 2CN 63 12 (CH 2) 2CO 2CH 3 CH 3 CH 2CH 3 58 13 C 6H 4-4-C(CH 3) 3 CH 3 CH 2CH 3 75 14 5-(1,3-benzodioxolyl) CH 2C 6H 5-4-OH CH 2CONH 2 82 15 3-nphthyl CH 3 CH 2CH 3 81 Yields of purified mteril re bsed on the initil minomethyl loding level of the resin. in >85% purity by 1 H NMR nlysis of crude products. Yields of purified benzodizepine products vried from 46% to 72% (v 61%, 15 compounds) bsed on the initil minomethyl loding of the polystyrene resin used. 82 Bunin nd co-workers hve reported the preprtion of librry of 11 200 discrete 1,4-benzodizepines 11 from 20 cid chlorides, 35 mino cids, nd 16 lkylting gents, ll of which were commercilly vilble (Figure 16). The cid chlorides were selected from set of over 300 commercilly vilble cid chlorides tht re comptible with the synthesis sequence using similrity grouping procedure developed by Dr. Steven Muskl t MDL Informtion Systems to select s diverse set s possible. The lkylting gents nd mino cids lso displyed rnge of functionlity. The librry is currently being screened by number of industril nd cdemic collbortors. Plunkett nd Ellmn hve lso demonstrted silyl linkge strtegy for the synthesis of benzodizepine derivtives. 83 Linkge of romtic compounds to solid supports through silyl groups is likely to become n importnt nd generl strtegy, since clevge from the resin is ccomplished by protodesiltion to leve no trce of the linkge site. For the synthesis of 1,4-benzodizepine derivtives, the silylsubstituted (minoryl)stnnne derivtive 34, which is synthesized in five steps in solution, is coupled to minomethylted polystyrene (Scheme 9). The syn- Scheme 9 thesis of support-bound 1,4-benzodizepines then proceeds s described previously. Tretment of the

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 569 Scheme 10 clevge (Scheme 10). The resins were purchsed s Boc-protected Merrifield resins nd deprotected with 1:1 TFA/CH 2 Cl 2. The 2-minobenzophenone imine derivtives 38 were then dded to form the supportbound imines, followed by heting to 60 C in TFA to provide the desired benzodizepine product. The cycltive nture of the clevge ws designed to enhnce the purity of the cyclized product, since incomplete products would be expected to sty bound Figure 16. Building blocks used to prepre 11 200- member benzodizepine librry. support-bound benzodizepine 36 with nhydrous HF then provides the benzodizepine product 37. Good purity of the crude product is observed, >85% by 1 H NMR, nd the isolted yield is 50-68% (v 61%, four compounds) fter purifiction bsed on the initil minomethyl loding of the polystyrene resin. In one of the erly reports of smll molecule librry synthesis, DeWitt nd co-workers described n lterntive strtegy for the synthesis of 1,4-benzodizepin-2-ones. 12 The synthesis is bsed on two-step procedure of trns-imidtion followed by cycltive Tble 5. 1,4-Benzodizepin-2-one Derivtives 40 (Scheme 10) derivtive mss blnce entry R 1 R 2 R 3 R 4 (%) 1 CH 3 C 6H 5 H H 40 2 CH 3 C 6H 5 Cl H 56 3 CH 3 C 6H 5-4-OCH 3 H H 34 4 CH 3 C 6H 5 NO 2 H 28 5 CH 3 see b below see b below H 63 6 CH 3 C 6H 5 Cl Me 18 7 CH 3 c-c 6H 11 H H 41 8 CH 3 2-thienyl H H 47 9 H C 6H 5 H H 44 10 H C 6H 5 Cl H 55 11 H C 6H 5-4-OCH 3 H H 23 12 H C 6H 5 NO 2 H 31 13 H see b below see b below H 16 14 H C 6H 5 Cl Me 20 15 H c-c 6H 11 H H 32 16 H 2-thienyl H H 41 17 CH 2C 6H 5 C 6H 5 H H 52 18 CH 2C 6H 5 C 6H 5 Cl H 46 19 CH 2C 6H 5 C 6H 5-4-OCH 3 H H 41 20 CH 2C 6H 5 C 6H 5 NO 2 H 26 21 CH 2C 6H 5 see b below see b below H 52 22 CH 2C 6H 5 C 6H 5 Cl Me 13 23 CH 2C 6H 5 c-c 6H 11 H H 39 24 CH 2C 6H 5 2-thienyl H H 48 25 3-CH 2indolyl C 6H 5 H H 43 26 3-CH 2indolyl C 6H 5 Cl H 33 27 3-CH 2indolyl C 6H 5-4-OCH 3 H H 31 28 3-CH 2indolyl C 6H 5 NO 2 H 23 29 3-CH 2indolyl see b below see b below H 23 30 3-CH 2indolyl C 6H 5 Cl Me 10 31 3-CH 2indolyl c-c 6H 11 H H 34 32 3-CH 2indolyl 2-thienyl H H 40 33 CH(CH 3) 2 C 6H 5 H H 31 34 CH(CH 3) 2 C 6H 5 Cl H 28 35 CH(CH 3) 2 C 6H 5-4-OCH 3 H H 29 36 CH(CH 3) 2 C 6H 5 NO 2 H 9 37 CH(CH 3) 2 see b below see b below H 29 38 CH(CH 3) 2 C 6H 5 Cl Me 11 39 CH(CH 3) 2 c-c 6H 11 H H 27 40 CH(CH 3) 2 2-thienyl H H 37 Mss blnce of crude mteril cleved from the resin bsed on mino cid loding level. b Prepred from 2-mino- 9-fluorenone rther thn 2-minobenzophenone.

570 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Figure 17. Structures of 1,4-benzodizepin-2-one nd 1,4- benzodipine-2,5-dione. to the support. The desired benzodizepine derivtives were purified by extrction nd isolted in crude mss blnces of 9-63% (v 34%, 40 compounds) in >90% purity s mesured by 1 H NMR (Tble 5). 1,4-Benzodizepine-2,5-diones. Boojmr nd co-workers hve lso reported generl nd highyielding method for the solid-phse synthesis of 1,4- benzodizepine-2,5-diones (Figure 17). 84 The synthesis strtegy complements the previously described 1,4-benzodizepin-2-one synthesis sequence described by Bunin nd co-workers, since wide rnge of functionlity cn be directly introduced onto the romtic core of the benzodizepine structure (R 1 ) from the greter thn 40 commercilly vilble nthrnilic cids or relted heterocyclic structures. The other two sites of diversity re introduced with R-mino esters nd lkylting gents, of which there re lso number of derivtives tht re commercilly vilble. The synthesis of 1,4-benzodizepine-2,5-diones is initited by loding n R-mino ester onto the ldehyde-derivtized support by reductive mintion employing NBH(OAc) 3 in DMF with 1% AcOH (Scheme 11). 85 Rcemiztion is not observed if the Scheme 11 Tble 6. Synthesis of 1,4-Benzodizepine-2,5-diones 46 (Scheme 11) entry R 1 R 2 R 3 (%) yield 1 8-Cl CH 2CH 3 CH 2CH(CH 3) 2 75 2 7-Cl CH 2CHdCH 2 CH 2C 6H 5 89 3 7-Br CH 2CH 3 CH 2CH(CH 3) 2 71 4 8-NO 2 CH 2CH 3 CH 2CH(CH 3) 2 92 5 6-F CH 2CONH 2 CH 2C 6H 5 62 6 8-OCH 3 CH 2c-C 3H 5 CH 2CH(CH 3) 2 79 7 b 7-(C 6H 4)-p-OCH 3 CH 2CH 3 CH 2CH(CH 3) 2 62 8 b 8-(CH 2) 5CH 3 CH 2C 6H 4-p-Ph CH 2C 6H 4OH 77 9 7-Cl CH 2CHdCH 2 (CH 2) 4NH 2 63 10 8-Cl H CH 2CH(CH 3) 2 89 11 7-Cl H CH 2C 6H 5 89 Yields of purified mterils re bsed on the loding levels of leucine nd phenyllnine ester derived resins. b Suzuki cross-coupling products. crbodiimide (EDC) is the most convenient ctivting gent since the tertiry mine hydrochloride is present in the crbodiimide structure. Cycliztion nd then subsequent lkyltion of the support-bound nilide nion 44 generted in situ is next ccomplished in single step by tretment of mide 43 with the lithium slt of cetnilide in DMF/THF (1:1) for 30 h, followed by ddition of n pproprite lkylting gent. Additionl diversity my lso be introduced onto the benzodizepine through the Suzuki cross-coupling rection s is exemplified in entry 7 (Tble 6), where cross-coupling rection ws crried out with p-methoxybenzeneboronic cid, nd in entry 8 (Tble 6) where Suzuki cross-coupling rection ws performed using B-hexyl-9-BBN. The benzodizepine products re cleved from the support by tretment with TFA/Me 2 S/H 2 O (90:5:5). Good yields were obtined for rnge of different derivtives including benzodizepines tht incorporte mino cids with side-chin functionlity such s tyrosine nd lysine, entries 8 nd 9 in Tble 6, respectively. In ddition, no detectble rcemiztion, <3%, is observed throughout the synthesis sequence s determined by chirl HPLC nlysis nd derivtiztion studies. A synthetic route to the 1,4-benzodizepine-2,5- dione clss ws lso reported by Zuckermnn nd co-workers t Chiron. 86 In this study the 1,4-benzodizepine-2,5-dione ws synthesized from the N- terminus of support-bound peptoid intermedite 47 (Scheme 12). Support-bound peptoid intermedite 47 is cylted with bromocetic cid followed by displcement of the bromide with n R-mino ester. Scheme 12 imine resulting from condenstion of the R-mino ester nd ldehyde 41 is reduced immeditely upon its formtion. Acyltion of the resulting secondry mine 42 with commercilly vilble unprotected nthrnilic cid then provides the support-bound tertiry mide 43. The optiml conditions for effecting this trnsformtion re to employ crbodiimide in conjunction with the hydrochloride slt of tertiry mine. Ethyl-3-[3-(dimethylmino)propyl]-

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 571 Tble 7. 1,4-Benzodizepine-2,5-dione Derivtives 50 (Scheme 12) derivtive entry R 1 R 2 R 3 yield (%) purity (%) b 1 CH 2CH(CH 3) 2 H H 55 >65 2 CH 2CH(CH 3) 2 CH 3 H 55 80 3 CH 2CH(CH 3) 2 CH 3 9-Cl >90 79 4 CH 2CH(CH 3) 2 CH 2C 6H 5 H 41 92 5 CH 2CH(CH 3) 2 C 6H 5 H 53 >95 6 CH 2CH(CH 3) 2 CH 2OH H 34 80 7 CH 2CH(CH 3) 2 CH 2C 6H 5-4-OH H 68 61 8 CH 2CH(CH 3) 2 CH(CH 3) 2 H 41 72 9 CH 2CH(CH 3) 2 CH 2CO 2H H 52 69 10 CH 2CH(CH 3) 2 (CH 2) 2CO 2H H 60 70 11 CH 2CH(CH 3) 2 (CH 2) 4NH 2 H 50 97 12 CH 2CH(CH 3) 2 (CH 2) 3NH 2 H 90 63 13 CH 2CH(CH 3) 2 CH 3 10-CH 3 37 61 14 CH 2CH(CH 3) 2 CH 3 8-OTf nd 88 15 CH 2CH(CH 3) 2 CH 3 8-NO 2 nd 93 16 CH 2CH(CH 3) 2 CH(OH)CH 3 H 50 59 17 myrtnyl c CH 3 H 75 93 18 myrtnyl c CH 2C 6H 5 H 55 84 19 9-fluorenyl CH 3 H 41 85 20 5-indnyl d CH 3 H 58 83 21 CH(CO 2CH 2CH 3)CH 2C 6H 5 CH 2C 6H 5 H 49 64 Yields of unpurified product bsed on the loding level of the strting resin. b Purity s estimted by HPLC. c Derived from myrtnylmine. d Derived from 5-minoindn. nd ) not determined. The resulting mine 48 is then cylted with substituted o-zidobenzoyl chloride. Rection with tributylphosphine ffords the iminophosphorne, which cyclizes upon heting to 125 C to fford the support-bound benzodizepine. Clevge with TFA provides the 1,4-benzodizepine-2,5-dione 50, which is lyophilized twice from cetic cid. A number of mino cids were evluted for comptibility with the synthesis sequence (entries 1-12, Tble 7). The study includes the synthesis of 21 compounds tht were isolted in crude mss blnces of 34-90% (v 55%) nd HPLC purities estimted t 59-97% (v 78%). Rcemiztion ws exmined for one model benzodizepine (entry 18). The distereomeric excess of this product ws 87%, nd further studies re reported to be in progress. A model study ws lso performed to demonstrte tht the synthesis sequence could be performed in split synthesis formt for librry synthesis. An equimolr mixture of eight monopeptoid resins were pooled nd trnsformed into 1,4-benzodizepine-2,5- diones using L-Phe methyl ester nd 2-zidobenzoyl chloride. Clevge of the resulting products from the resin provided smple pool of eight compounds tht ws evluted by Electrospry MS nd HPLC. All eight desired benzodizepines were present in ddition to uncyclized mteril. The uthors report tht further functionliztion of the benzodizepin-2,5- diones cn be ccomplished by Suzuki rection of triflte substituted derivtives or nitro reduction nd cyltion sequence (dt not provided). 2. Six-membered Rings. Diketopiperzines. Gordon nd Steele hve reported generl strtegy for the solid-phse synthesis of diketopiperzines employing redily vilble mino cids nd ldehydes to introduce diversity. 87 Reductive mintion of support-bound mino cid ws first ccomplished by tretment with n ldehyde nd sodium tricetoxyborohydride in CH 2 Cl 2 (Scheme 13). The reductive mintion step ws performed twice to improve conversions to 85-95%, except for stericlly hindered Scheme 13 mino cids nd/or electroniclly dectivted ldehydes which gve lower conversions. In fct, for the most difficult cse where Vl nd n electroniclly dectivted ldehyde ws employed, only 20% conversion ws observed. Prtil rcemiztion ws lso noted for some mino cids (Phe gve pproximtely 10% rcemiztion by chirl HPLC). Both romtic nd liphtic ldehydes were employed in the reductive mintion step, lthough liphtic ldehydes provided noticeble dilkyltion (1-10%). An N-Boc mino cid ws then coupled to the secondry mine employing PyBroP s the ctivting gent (double coupling ws used to drive the rection to completion). The Boc protecting group ws then removed with concomitnt clevge from the support with TFA. The resulting dipeptide cids were then dissolved in toluene nd heted t reflux for 5hto provide the desired diketopiperzine products 54. The yields of two purified diketopiperzines were reported, 42% yield when R 1 ) CH 2 C 6 H 5,R 2 )CH 2 C 6 H 4-4-OCH 3, nd R 3 ) CH 3, nd 24% yield when R 1 ) CH(CH 3 ) 2,R 2 )CH 2 C 6 H 2-2,4,6-(OCH 3 ) 3 nd R 3 ) CH 2 CH(CH 3 ) 2. A librry of 1000 DKP s ws then prepred using the split synthesis pproch from 10 N-Fmoc mino cid-derivtized resins, 10 ldehydes, nd 10 N-Boc mino cids (Figure 18). The 10 N-Fmoc mino cidderivtized resins were deprotected, pooled, nd

572 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Figure 18. Building blocks used to prepre 1000- member diketopiperzine librry. reductively lkylted with 10 different ldehydes to generte 10 mixtures of 10 support-bound N-lkyl mino cids 52. These mixtures were chrcterized using HPLC-MS nd MS-MS fter clevge from the support to confirm the presence of 96 out of the 100 expected N-lkyl mino cids. The resulting resins were then repooled, mixed, nd divided into 10 seprte pools. Ech pool ws coupled with unique N-Boc mino cid, followed by clevge from the resin nd cycliztion to provide librry of 1000 compounds s 10 mixtures of 100 compounds ech. In lter publiction, Terrett reported tht biologicl evlution of this librry resulted in the identifiction of number of ctive diketopiperzines including lignd to the neurokinin-2 receptor (IC 50 ) 313 nm). 6b Isoquinolinones. Zuckermnn nd Goff hve reported method for the solid-phse synthesis of isoquinoline derivtives, which grew from desire to increse the structurl rigidity, complexity, nd diversity of their existing peptoid librries. 88 Coupling of trns-4-bromo-2-butenoic cid to the support through the Rink linker is ccomplished with DICI (Scheme 14). Bromide displcement with primry mine is followed by cyltion of the resulting secondry mine with 2-iodobenzoyl chloride derivtive to provide the support-bound mide 57. Plldium-medited cycliztion by n intrmoleculr Heck rection using Pd(PPh 3 ) 4 followed by clevge with TFA then provides the (2H)-isoquinoline 58. However, when 2-iodobenzoyl chloride derivtive ws employed tht lso contined substituent in the 3-position (entries 4 nd 8, Tble 8), mixture of product isomers 58 nd 59 ws seen, which could fvor isomer 59 (entry 4, Tble 8). Further experiments led the uthors to conclude tht isomer 59 is produced first, with equilibrtion to isoquinoline 58 under the rection conditions, probbly through mechnism of reddition nd elimintion of Pd-H. Presumbly, n ortho substituent hinders the reddition of Pd-H. Mss blnces fter clevge nd lyophiliztion re reported for eight compounds (65-92%, v 77%), which re obtined in good to purity s estimted by HPLC (70-95%, v 87%). A pooling strtegy ws lso investigted, where seven of the different support-bound intermedites 56 (Scheme 14) were mixed nd then converted to the isoquinoline products. Clevge nd HPLC showed the presence of ll seven desired products nd the structures were verified by ES-MS. Tble 8. 2-Substituted 1-(2H)-Isoquinolinones 58 (Scheme 14) derivtive yield purity entry R 1 R 2 (%) (%) b 1 CH 2CH(CH 3) 2 H 69 83 2 CH 2CH 2C 6H 5 H 65 80 3 C 6H 5 H 85 >70 4 CH 2CH(CH 3) 2 5-CH 3 92 (1/3.2) c 94 5 CH 2CH(CH 3) 2 8-F 80 90 6 CH 2CH(CH 3) 2 6,7-(OCH 3) 2 77 95 7 CH 2CH(CH 3) 2 7-Cl 79 90 8 CH 2CH(CH 3) 2 5-OCH 3 69 (1.7/1) c 93 Yields of unpurified product fter lyophiliztion from cetic cid. b Purity s estimted by HPLC. c Vlues in prentheses re rtios of 58 to 59, otherwise only 58 ws observed. Scheme 14 1,4-Dihydropyridines. Gordeev nd co-workers t Affymx hve utilized Hntzsch-type rection to provide 1,4-dihydropyridines, which hve served s the nucleus of numerous bioctive compounds. 89 Condenstion of β-keto ester with the free mine from either the Rink (shown) or the PAL linker gives the support-bound enmino ester 60 (Scheme 15). Tretment with n ldehyde nd second β-keto ester, which forms 2-rylidene β-keto ester in situ (this regent cn lso be preformed), provides the resin-bound intermedite 61. Clevge with TFA results in cycliztion to give 1,4-dihydropyridines 62 (Tble 9). The intermedicy of support-bound cyclized 1,4-dihydropyridine is lso possible; however, in model study using 13 C-enriched ethyl cetocette (R 1 ) 13 C methyl), the 13 C NMR of the supportbound intermedite prior to cycliztion showed two signls for the R 1 methyl, tenttively ssigned to the Scheme 15

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 573 Tble 9. 1,4-Dihydropyridine Derivtives 62 (Scheme 15) derivtive entry Ar R 1 R 2 R 3 R 4 yield (%) 1 C 6H 4-2-NO 2 CH 3 CH 3 CH 3 OCH 3 65 b (70) c 2 C 6H 4-3-NO 2 CH 3 CH 3 CH 3 OCH 2CH 3 75 c 3 C 6H 4-3-NO 2 CH 3 CH(CH 3) 2 CH 3 O(CH 2) 2OCH 3 78 c 4 C 6H 4-4-NO 2 CH 3 CH 3 CH 3 CH 3 75 c 5 C 6H 4-4-NO 2 CH 3 CH 2CH 3 CH 3 OCH 3 70 c 6 C 6H 4-4-NO 2 CH 2CH 3 CH 3 C 6H 5 OCH 2CH 3 70 c 7 C 6H 4-4-CN CH 3 CH 3 CH 3 OCH 3 74 c 8 C 6H 5 CH 3 CH 3 CH 3 OCH 2CHdCH 2 72 b 9 4-pyridyl CH 3 CH 3 CH 3 OCH 3 75 c Yields re bsed on the loding level of the strting resin. b Prepred by two-component condenstion with ArCHdC(COR 3)CO 2R 4 using PAL resin. c Prepred by three-component condenstion using Rink resin, ArCHO, nd R 3COCH 2COR 4. E nd Z isomers of intermedite 61. The IR of the support-bound intermedite lso showed n bsorbnce t 1735 cm -1 while the cyclized product showed the expected bsorbnce t 1705 cm -1. Interestingly, pyridine ws lso necessry in the condenstion step to fcilitte isomeriztion of the imine to the enmine, s undesired cycloddition byproducts were observed without it. The synthesis sequence is notble becuse the rection conditions re mild, the synthesis proceeds in good overll yields (v 73%, nine compounds), nd vriety of functionlity could potentilly be displyed bout the structure from commercilly vilble strting mterils. Dihydro- nd Tetrhydroisoquinolines. Meutermns nd Alewood hve demonstrted the synthesis of dihydro- nd tetrhydroisoquinolines on support. 90 Acyltion of support-bound dimethoxyphenyllnine with either cetic cid or phenylcetic cid produced the intermedite mide 64 or 65 (Scheme 16). The support-bound dihydroisoquinolines 66 nd 67 were then obtined by tretment of mides 64 nd 65 with POCl 3 t 80 C to effect Bischler-Npierlski cycliztion rection. The dihydroisoquinolines 68 nd 69 were relesed from the resin by tretment with HF/p-cresol. The supportbound dihydroisoquinolines 66 nd 67 were lso converted to the tetrhydroquinolines 70 nd 71 by tretment with NBH 3 CN before clevge from the support. All four products were obtined in high purity s estimted by HPLC nlysis. The isolted yield fter HPLC purifiction for ech of the dihydroisoquinolines 68 nd 69 ws 40%, nd the yields of the tetrhydroisoquinolines 72 nd 73 were 25% nd 30%, respectively, s 6:1 mixture of distereomers for both compounds. In seprte experiment n equimolr mixture of eight cetic cid derivtives (R 1 ) H, phenyl, t-bu, nphthyl, p-methoxyphenyl, 3,4-dimethoxyphenyl, p-nitrophenyl, nd p-hydroxyphenyl) ws coupled to support-bound mino ester 63 nd then submitted to the cycliztion nd clevge steps. The prent ions of ll eight dihydro- nd tetrhydroisoquinolines were identified by ionspry MS, but no other chrcteriztion of the mixtures ws provided. 3. Five-Membered Rings. Hydntoins. In one of the erliest studies of orgnic compound librry synthesis, DeWitt nd co-workers reported method for the synthesis of hydntoin derivtives employing the Diversomer TM pprtus. 12 Eight resin-bound mino cids (Phe, Gly, Ile, Leu, Al, Vl, Trp, nd diphenylglycine) were condensed with five different isocyntes (structures not provided) in DMF to provide resin-bound ures 74 (Scheme 17). Heting Scheme 17 Scheme 16 the resin-bound ures 74 t 85-100 C in 6 M HCl for 2 h resulted in cycltive clevge to provide the hydntoin products 75 in crude mss blnces of 4-81%, (v 30%, 40 compounds) with chrcteriztion by TLC, MS, nd 1 H NMR. Pyrrolidines. Gllop nd co-workers hve developed solid-phse synthesis of mercptocyl proline derivtives with metllozomethine ylide cycloddition rection serving s the key step for introducing diversity. 91 Condenstion of n mino cid-derivtized resin with n romtic ldehyde in net trimethylorthoformte provided the Schiff bse 76, nd ny unrected mines were cpped with cetic nhydride (Scheme 18). The dehydrting conditions tht were employed to effect Schiff bse formtion

574 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Scheme 18 Figure 20. Cptopril nd the identified mercptocyl proline ACE inhibitor, 79. hd previously been reported by the uthors. 92 The 1,3-dipolr cycloddition ws then performed with the zomethine ylide derived from the support-bound Schiff bse 76 nd n crylte or crylonitrile s the dipolrophile under Lewis cid-medited conditions tht re nlogous to those initilly reported by Grigg nd Tsuge for the corresponding solution-phse rection. 93 The resulting rcemic substituted proline derivtives were then cylted with vrious ω-mercptocyl chlorides followed by cidolytic clevge from the support to provide the desired (mercptocyl)proline product 78 (Scheme 18). The isolted yields for six representtive products rnged from 50 to 80% nd the distereoselectivities rnged from 2.5:1 to 10:1. The uthors reported tht the products typiclly rose from n endo-selective cycloddition with the syn configurtion of the support-bound zomethine ylide. Dipolrophiles lcking crbonyl substituent showed decresed stereoselectivity. In ddition, the distereoselectivity ws dependent upon the nture of the resin support nd the concentrtion of the Lewis cid ctlyst. The uthors prepred smll librry of (mercptocyl)proline derivtives 78, employing the optimized synthesis sequence nd the split synthesis procedure. The librry ws prepred from four mino cids, four romtic ldehydes, five olefins, nd three mercptocyl chlorides (Figure 19). Becuse the librry ws trgeted towrd the identifiction of ngiotensin-converting enzyme (ACE) inhibitors, the building blocks were selected in prt upon the structure ctivity reltionships of known (mercptocyl)proline-bsed inhibitors such s cptopril. Although the totl number of combintions of the building blocks in the librry ws 240; becuse the rections were not completely regio- nd stereospecific, pproximtely 500 compounds were prepred. Figure 19. Building blocks used to prepre n 500- member mercptocyl proline librry. The librry ws decetylted by tretment with ethylenedimine nd then screened for inhibition of ACE. Deconvolution using the stndrd itertive resynthesis nd evlution protocol resulted in the identifiction of potent new inhibitor, 79 (Figure 20), with K i of 160 pm. Inhibitor 79 is 3-fold more potent thn cptopril nd is mong the most potent thiol-contining ACE inhibitors yet described. Thizolidine-4-crboxylic Acids. Utilizing Fmoc- Cys(Trt)-OH s strting point, the Selectide group hs synthesized number of N-cylthizolidines on PEG-PS resin. 94 Fmoc-Cys(Trt)-OH ws coupled directly onto PEG-PS-OH resin (Scheme 19). The Scheme 19 Fmoc nd side chin trityl protecting groups were then cleved under stndrd conditions. Addition of n ldehyde in cetic cid resulted in imine formtion followed by cycliztion to provide the support-bound thizolidine 80. Acyltion of the thizolidine 80 in pyridine with either cetic nhydride, benzoyl chloride, or n isocynte followed by clevge with NOH provided the rcemic N-cylthizolidine 82 s single distereomer (epimeriztion occurs upon hydrolysis). The products were isolted in <10-94% yield fter extrctive workup. Purity ws estimted t >97% by HPLC nlysis (Tble 10). The uthors lso report tht N-cylthizolidines prepred from electron-rich romtic ldehydes re not stble in 60-100% TFA/CH 2 Cl 2, conditions tht re necessry to cleve mny of the side-chin protecting groups Tble 10. Thizolidine-4-crboxylic Acid Derivtives (Scheme 19) derivtive entry R 1 R 2 yield (%) 1 C 6H 4-4-OCH 3 CH 3 90 2 C 6H 4-4-NO 2 CH 3 0 3 C 6H 4-4-OCH 3 NHCH 3 69 4 H CH 3 61 5 H C 6H 5 81 6 CH(CH 3) 2 CH 3 40 7 CH(CH 3) 2 C 6H 5 79 8 CH(CH 3) 2 NHC 6H 5 58 9 CH 2CH(CH 3) 2 CH 3 75 10 2-thienyl CH 3 63 Crude mss blnce fter extrctive workup.

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 575 tht re commonly used in solid-phse synthesis. To ddress this problem, oxidtion of the thioether to the sulfoxide ws ccomplished in good yield using MCPBA, providing cid-stble products. A vriety of conditions, however, filed to provide the corresponding sulfone in good yield. 4-Thizolidinones nd Relted 4-Metthiznones. Holmes nd co-workers t Affymx hve reported the synthesis of 4-thizolidinones nd relted metthiznones from support-bound mino cid, n romtic ldehyde nd mercptocetic cid derivtive (Scheme 20). 95 After Fmoc deprotection Scheme 20 Figure 21. Preprtion of imidzoles. Imidzoles. Srshr nd co-workers t Ontogen corportion hve developed method for the solidphse synthesis of imidzole derivtives bsed on one-pot procedure. 96 The imidzole derivtives my be prepred by either three-component or fourcomponent process (Figure 21). In the three-component process NH 4 OAc, 1,2-dione, nd n ldehyde re heted in AcOH to provide the imidzole product, nd in the four-component process NH 4 OAc, primry mine, 1,2-dione, nd n ldehyde re submitted to the sme rection conditions (the reltive stoichiometry of the regents is used to control primry mine introduction). The uthors perform the synthesis on solid support by linking either the mine or ldehyde component to the support (Wng resin) s shown in Scheme 21. Scheme 21 of commercilly vilble mino cid-loded resins (polystyrene, PEG-PS, nd polydimethylcrylmide/polyhipe support were ll used with equl success), one-pot rection ws performed with n ryl ldehyde nd mercptocetic cid (mercptocetic cid nd thiolctic cid were used) with heting t 70 C nd with 3 Å moleculr sieves s dehydrting gent. Clevge from the support with 50% TFA in CH 2 Cl 2 provided the thizolidinone derivtives in good purity s estimted by HPLC (84-98%, v 94%, Tble 11). Distereomeric thizolidinones re formed from chirl mino cids, nd bulky substituents were shown to provide modest distereoselectivity (e4:1 mjor/minor). Attempts to use β-mercptopropionic cid to form the nlogous six-membered ring metthiznones were not s successful. A derivtive of glycine ws isolted in good purity, but extension to Al nd Phe led to the formtion of <10% of the desired product. Tble 11. 4-Thizolidinones 84 nd Relted 4-Metthiznones Derivtives (Scheme 20) derivtive purity entry R 1 R 2 R 3 (%) yield (%) b 1 H C 6H 5 H 60 98 2 H C 6H 5 CH 3 97 96 3 H C 6H 4-2-CH 3 H 92 98 4 H C 6H 4-3-CH 3 H 91 97 5 H C 6H 4-4-CH 3 H 100 95 6 H 3-pyridyl H 55 99 7 CH 3 C 6H 5 H 82 92 8 see c below C 6H 5 H 94 96 9 CH 2C 6H 5 C 6H 5 H 86 84 10 CH(CH 3) 2 C 6H 5 H 99 88 11 d H 3-pyridyl see d below 93 64 12 d CH 3 C 6H 5 see d below 78 <10 13 d CH 2C 6H 5 C 6H 5 see d below 67 <10 Purity s estimted by HPLC. c Prepred from β-minopropionic cid. d Synthesis of 4-metthiznones from β-mercptopropionic cid. The support-bound component is then heted with the remining regents in AcOH t 100 C for 4 h. The derivtized resins 85-87 were prepred by either simple crbodiimide-medited ester formtion, or solid-phse Mitsunobu protocol originlly reported by Richter nd Gdek. 97 Rink resin ws lso exmined in ddition to Wng resin, but it ws found to be too unstble under the cidic rection conditions. The imidzoles 88-90 were isolted in 90-95% purity nd excellent yield (71-99%) s judged by 1 H NMR fter cidolytic clevge from the resin followed by chromtogrphy (Tble 12). The uthors lso reported n extension of the protocol to the preprtion of unsymmetricl bis-imidzoles (Scheme 22). Scheme 22

576 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Tble 12. Imidzoles 88-90 (Scheme 21) derivtive entry R 1 or Linker R 2 or Linker R 3 R 4 yield (%) 1 C 6H 4-4-CO 2H H C 6H 5 C 6(CH 3) 5 98 2 C 6H 4-4-CO 2H C 6H 5 C 6H 5 C 6H 5 99 3 C 6H 4-4-CO 2H c-c 5H 9 C 6H 4-4-F C 6H 4-4-F 93 4 C 6H 4-4-OH H C 6H 5 CH 3 82 5 C 6H 4-4-OH CH 2C 6H 5 C 6H 4-4-OCH 3 C 6H 4-4-OCH 3 95 6 C 6H 4-4-OH (CH 2) 5CH 3 2-furyl 2-furyl 71 7 C 6H 3-3,5-(OCH 3) 2 (CH 2) 5CO 2H C 6H 4-2-Cl C 6H 4-2-Cl 96 8 C 9H 19 (CH 2) 5CO 2H C 6H 4-3-OCH 3 C 6H 4-3-OCH 3 95 9 c-c 6H 11 (CH 2) 5CO 2H C 6H 4-4-Br-2-OH C 6H 4-4-Br-2-OH 95 Yields of chromtogrphed product re bsed on the loding level of the strting resin. Scheme 23 Figure 22. Structures of β-turn nd β-turn mimetic. Tretment of ldehyde functionlized resin 87 with NH 4 OAc nd 1,4-bis-benzil (20 equiv) under stndrd conditions provided the support-bound nd 1,2-dionesubstituted imidzole product 91. The supportbound dione intermedite 91 ws then treted with NH 4 OAc nd 4-ethylbenzldehyde under stndrd conditions followed by clevge from the support to provide the unsymmetricl bis-imidzole 92 in 75% mss recovery nd 92% purity. b. Trgeted Librries. 1. β-turn Mimetics. β-turns, 93 (Figure 22), re key structurl motif in peptides nd proteins nd often ply key role in moleculr recognition events in biologicl systems. 98 A gret del of effort therefore hs been focused on the design of smll constrined mimetics of turn structure in order to identify high ffinity nd specific lignds to receptor nd enzyme trgets. 99 However, these efforts hve met with only limited success due to difficulties in identifying the key turn residues nd the reltive orienttions of those residues in the receptor bound conformtion. To ddress these issues Virgilio nd Ellmn hve developed method for the solid-phse synthesis of librry of β-turn mimetics, 94, for the rpid identifiction of lignds tht re bsed upon the β-turn structure. 100 The turn mimetic 94 is constructed from three redily vilble components. The i + 1 side chin is derived from n R-hlo cid nd the i + 2 side chin is derived from n R-mino cid. The mimetic is constrined in turn structure by replcing the hydrogen bond between the i nd i + 3 residues with covlent bckbone linkge. The flexibility of the turn mimetic s well s the reltive orienttions of the side chins cn be vried by introducing different bckbone linkges to provide nine- or 10-membered rings. In ddition, different side-chin orienttions re obtined by introducing different bsolute configurtions t ech of the stereocenters introduced by the i + 1 nd i + 2 side chins of the turn mimetic. The β-turn mimetics re prepred with PEG-PS s the support nd using the Rink mide linker. In the initil study, p-nitrophenyllnine ws loded onto the support before the synthesis ws initited to serve s convenient UV tg for ccurte determintion of the overll purity of the turn mimetic by HPLC (vide infr). R-Bromocetic cid is first coupled to the support-bound p-nitrophenyllnine by ctivtion with DICI (Scheme 23). 51 Subsequent tretment with the minolkyl mercptn protected s the tertbutyl mixed disulfide provides the secondry mine 96. The secondry mine is then coupled with n N-Fmoc R-mino cid employing O-(7-zbenzotrizol-1-yl)-1,1,3,3-tetrmethyluronium hexfluorophosphte (HATU) to provide mide 97. 101 Tretment with 20% piperidine in DMF followed by rection with the symmetric nhydride of the pproprite R-bromo cid provides cyclic intermedite 98 tht incorportes both the i + 1 nd the i + 2 side-chin residues. Clevge of the mixed disulfide is then ccomplished by tretment of the cycliztion precursor 98 with tributylphosphine in 5:3:2 propnol/ DMF/wter comixture. It is necessry to use PEG- PS s the solid support, which is well solvted under the queous rection conditions, in order for clen reduction of the disulfide bond to occur without side rections. Cycliztion to provide the nine- or 10- membered thioether is ccomplished by tretment with tetrmethylgunidine (TMG) in DMF/H 2 O comixture. 102 Clevge of the turn mimetic from the support by tretment with 95:5:5 TFA/DMS/H 2 O then provides mimetic 99 (Tble 13). Employing this synthesis sequence, turn mimetics 99 were obtined with n verge purity of 75% (11 compounds) over the eight-step process s determined by HPLC nlysis (Tble 13). Any side

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 577 Tble 13. β-turn mimetics 99 (Scheme 23) derivtive purity (%) b entry R i+1 R i+2 bckbone (n) PEG-PS pins 1 CH 3 CH 2C 6H 5 2 90 79 2 CH c 3 CH 2C 6H 5 2 59 90 3 CH(CH 3) 2 CH 2C 6H 5 2 81 86 4 CH 2CO 2H CH 2C 6H 5 2 65 79 5 CH 3 (CH 2) 4NH 2 2 72 87 6 CH 3 CH 2CO 2H 2 63 86 7 H CH 2C 6H 5 2 85 91 8 H CH 2OH 2 82 93 9 CH 3 CH 2C 6H 4-4-OH 2 74 88 10 CH 3 CH 2C 6H 5 1 77 75 11 CH 2C 6H 4-4-OH CH 3 1 81 90 The stereochemicl configurtion t the i + 1 site is R nd t the i + 2 site is S unless otherwise specified. b Purity by HPLC. c The stereocenter hs the S configurtion. products tht were produced during the synthesis of the turn mimetic would be detected by HPLC nlysis since the UV tg, p-nitrophenyllnine, ws introduced before the synthesis of the mimetic ws initited. For ll of the turn mimetics synthesized, cycliztion provided the desired cyclic monomer with no cyclic dimer detected (<5%). 103 This includes mimetics incorporting both (R) nd (S) R-bromo cids, R-bromoisovleric cid which corresponds to the stericlly hindered mino cid, vline, nd R-chlorocetic cid which corresponds to the lest stericlly hindered mino cid, glycine, t the i + 1 site. A vriety of side-chin functionlity could be incorported successfully into the turn mimetics including lcohol, phenol, crboxylic cid, nd mine functionlity. In ddition, less thn 3% rcemiztion ws observed in the synthesis sequence. To demonstrte the utility of the synthesis sequence for the rpid construction of librry of turn mimetics 99, the 11 mimetics were synthesized simultneously employing Chiron Mimotopes pin pprtus (section II.A.1.). All 11 derivtives were obtined in very high level of purity s determined by HPLC nlysis (Tble 13). On the bsis of these results, librry of 1292 β-turn mimetics hs been prepred using the 19 R-bromo cids, 34 R-mino cids, nd two bckbone elements (Figure 23). 104 A modified librry hs lso been prepred from the initil librry by oxidtion of the thioethers to sulfoxides with hydrogen peroxide in DMF/H 2 O comixture. 105 Dr. Andrew Bry from Chiron Mimotopes evluted 7% of the librry by mss spectrometry using electrospry ioniztion, nd for ll of the derivtives tht were tested the expected moleculr ion ws observed. The librry hs successfully been evluted for the identifiction of severl specific lignds to rt somtosttin receptor subtypes I-IV by reserchers t Genentech, nd screening by Berlex identified novel lignds to the f-mlp receptor. 2. Protese Inhibitor Librries. Severl groups hve worked towrd the identifiction of protese inhibitors by the preprtion of librries of compounds tht incorporte isosteres tht mimic the tetrhedrl intermedite for peptide hydrolysis. 106 Sttine, hydroxyethylmine, hydroxylethylure, the dimino diol core nd the dimino lcohol core, hve been used to trget proteses of the sprtic cid clss, while peptidylphosphontes hve been used to trget the metlloprotese clss (Figure 24). Figure 23. Building blocks for the synthesis of β-turn librry. Figure 24. Isosteres tht hve been employed for the synthesis of protese inhibitor librries. Two distinct strtegies hve been employed. The first pproch is to incorporte the isostere directly into peptide chin. Lrge numbers of compounds cn thereby be synthesized rpidly nd efficiently using stndrd peptide synthesis methods. Alterntively, the isostere cn be ttched to the support through the functionlity tht corresponds to the site of peptide hydrolysis, such s the secondry hydroxyl

578 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn group(s) in the hydroxyethylmine, dimino diol, or dimino lcohol, thereby llowing the disply of diverse functionlity including nonpeptide functionlity from both ends of the molecule.. Asprtic Acid Proteses. Peptide Librry Incorporting Sttine. In 1991 Owens nd coworkers incorported sttine (Figure 24), known trnsition-stte nlog for the sprtic cid proteses, into peptide librry for the purpose of identifying inhibitors of HIV-1 Protese (HIV-PR). 21d This report ws the first exmple of the synthesis nd evlution of bised peptide librry tht ws designed to trget protese by the incorportion of trnsition stte isostere. In this effort, Owens synthesized tetrpeptide librry by the split synthesis method using Boc chemistry nd 22 mino cids of both D nd L stereochemistry. At the second position from the resin, the sme mixture ws employed with the ddition of sttine. The split synthesis protocol ws then continued to give finl librry of cetylted tetrpeptide mides, which were ssyed in solution for the bility to inhibit the clevge of known substrte by HIV-PR. Active inhibitors were identified by itertive resynthesis nd evlution. Not surprisingly, sttine ws found t the second site in the most ctive tetrpeptide identified, Ac-Trp-Vl-St-D-Leu-NH 2 (IC 50 ) 200 nm). On the bsis of the dt from the first librry, subsequent work resulted in the identifiction of compound (structure not provided) with n IC 50 of 5 nm. Peptides Incorporting the Hydroxyethylmine Isostere. Although not explicitly developed for the purpose of compound librry synthesis, Alewood nd co-workers hve developed n expedient solid-phse method for the incorportion of the hydroxyethylmine isostere into peptides. 107 As shown in Scheme 24, the key steps in the synthesis were to tret the Scheme 24 free mine of resin-bound peptide 100 with excess of n R-bromo ketone 101, in the presence of i-pr 2 NEt in DMF. Reduction of the resulting ketone product with NBH 4 in THF then provided the support-bound hydroxyethylmine isostere 102. 108 Three dditionl mino cids were incorported into the peptide, nd fter clevge with HF/p-cresol, the hexpeptide product ws isolted by HPLC in 52% overll yield s 6:4 rtio of (R) to(s) lcohol epimers. Nonpeptide Inhibitors. Kick nd Ellmn hve reported solid-phse method for displying functionlity from the (hydroxyethyl)mine nd (hydroxyethyl)ure isosteres for the synthesis of librry of potentil sprtic protese inhibitors. 109 Using dihydropyrn-functionlized resin, 110 the scffold 104 is coupled to support using pyridinium p-toluenesulfonte (PPTS) (Scheme 25). The primry tosyl lcohol 105 is then displced with either func- Scheme 25 tionlized or unfunctionlized primry or secondry mines. After coupling of the primry mines, the resulting secondry mine products 106 cn be converted to ures by rection with isocyntes or by stepwise tretment with triphosgene followed by mine ddition. 111 The stepwise procedure provides ccess to wide rnge of ures from the lrge pool of commercilly vilble primry mines. The secondry mine my lso be clenly cylted with cid chlorides nd sulfonyl chlorides employing stndrd methods. The synthesis bout the P 1 site of the inhibitor is initited by efficient reduction of zides 107 or 108 using thiophenol/et 3 N/SnCl 2 (4:5:1) ccording to the procedure described by Brtr for the corresponding solution-phse rection. 112 The resulting primry mine is then cylted to provide crbmte or mide products. If the mine is cylted with n Fmoc mino cid, then protecting group clevge nd further functionliztion is possible. The concomitnt removl of the side-chin protecting groups nd clevge of the mteril from the solid support is then ccomplished by tretment with 95:5 TFA/ wter. Complete clevge is observed in less thn 1 h without decomposition. By employing the previously described synthesis method number of different compounds were prepred (Figure 25). In order to demonstrte the verstility of the synthesis sequence, prticulr emphsis ws plced on incorporting functionlity tht is present in known nonpeptide-bsed HIV-1 protese inhibitors, 113 including N-tert-butylpipecolmide (120), piperzine derivtives (116), N-tertbutyl ures (112-114, 117, 118), quinldic mides (112-114, 121), nd 3(S)-hydroxytetrhydrofurnyl crbmtes (115-120). The derivtives were isolted fter four to six steps in 47%-86% yield (v 74%, nine compounds) fter chromtogrphy to provide nlyticlly pure mteril. The only uncceptble yield, 47%, ws for compound 120 tht resulted from slow tosyl lcohol displcement by the hindered

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 579 Scheme 26 Figure 25. Potentil protese inhibitors prepred ccording to Scheme 25. N-tert-butylpipecolmide. In current librry synthesis efforts towrd the identifiction of cthepsin D inhibitors, we now employ the p-nitrobenzenesulfonyl lcohol, which is n order of mgnitude more rective tht the corresponding tosyl lcohol nd provides the desired rection products under milder conditions nd for less rective mine nucleophiles, in higher yields (e.g. compound 120 ws obtined in 65% overll yield). Wng nd co-workers t Abbott hve developed solid-phse synthesis method to disply functionlity bout two clsses of C 2 symmetric trnsition stte isosteres tht trget HIV-1 protese. 114 The linkerderivtized isosteres 122 nd 123 were prepred in solution nd then coupled to 4-methylbenzhydrylmine (MBHA) resin (Scheme 26). After Fmoc deprotection, the sme N-Fmoc mino cid is simultneously coupled to both mines of the supportbound isostere 124 nd 125. The Fmoc protecting groups re then removed, nd the mines re cpped either s n mide or sulfonmide. The monolcohol linker could be cleved with 30% TFA/H 2 Ofor3h; however, the diol linker required more stringent conditions: TFA/H 2 O (95:5 overnight). The synthesis sequence ws evluted by the preprtion of 12 model compounds employing the two support-bound isosteres 124 nd 125, the mino cid Vl, nd six different cylting gents (Figure 26). The crude mss blnce ws close to theoreticl; however, the purity ws ssessed t 30-70% for the mono-ol series 126 nd 20-50% for the diol series 127 s determined by either TLC or HPLC. Mss spectrometry indicted tht the lower purity of the diol products ws in prt due to the presence of side product whereby the linker hd remined ttched to the diol product. The crude compounds were ssyed for inhibition of HIV-1 protese providing IC 50 vlues (<1 nm to 6.4 mm) tht were consistent with previous work in the re. The uthors report tht librry of 300 derivtives (structures not provided) hs been prepred in prllel using the bove chemistry nd employing n Abimed synthesizer, nd tht the librry is currently being evluted for HIV-1 protese inhibition. Figure 26. Crboxylic cid components. b. Zinc Protese Librry. Cmpbell nd co-workers t Affymx hve reported the synthesis of peptide librry tht incorportes the phosphonic cid trnsition stte isostere for the purpose of trgeting metlloproteses. 30 The librry ws synthesized on Tentgel-S-NH 2. A clevble linker ws not used, since support-bound ssy hd been developed for biologicl evlution. However, the fidelity of the synthesis sequence (Scheme 27) hd previously been demonstrted by the solid-phse synthesis of severl

580 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Scheme 27 peptidylphosphontes. 115 For both the N-Fmoc mino cid nd O-Fmoc R-hydroxy cid coupling steps, double coupling is performed first with HOBt/HBTU nd then with PyBroP in order to ensure complete rection conversion. The secondry lcohols 129 re then coupled with [(nitrophenyl)ethoxy]crbonyl (NPEOC)-protected (R-minolkyl)phosphonic cids under modified Mitsunobu rection conditions with tris(4-chlorophenyl)phosphine nd DIAD. 115 The NPEOC group is removed with 5% DBU in NMP nd the resulting free mines re cpped with Cbz-Cl. All of the rections cn be monitored by UV quntittion of the dibenzofulvene-piperidine dduct or 4-nitrostyrene relesed upon protecting group clevge. The coupling rections routinely proceed with >90% conversion. Side-chin deprotection with TFA then provides the support-bound peptidylphosphontes. By employing this synthesis sequence, librry ws prepred by the split synthesis method tht contined ll combintions of 18 of the nturlly occurring mino cids t the P 2 -position (Cys nd Asn were excluded), five R-hydroxy cids t the P 1 - position (glycolic cid, (R)-lctic cid, (R)-mndelic cid, 3(R)-phenyllctic cid, nd 2(R)-hydroxyisocproic cid), nd six (R-minolkyl)phosphonic cids t the P 1 -position (the (minolkyl)phosphonic cids corresponded to Gly, D,L-Al, D,L-Vl, D,L-Leu, D,L- Ile, D,L-Phe). The librry ws evluted employing depletion ssy whereby resin-bound mixtures were incubted with thermolysin nd then filtered to remove resinbound enzyme/inhibitor complex. Proteolytic ctivity of the filtrtes were then ssyed to rnk order the librry mixtures. Itertive resynthesis nd evlution provided qulittive structure vs ctivity reltionships s well s resulting in the identifiction of number of potent inhibitors of thermolysin, including the expected Cbz-Phe p - o Leu-Al, which ws resynthesized s the terminl mide (K i ) 49 nm) nd s the crboxylte (K i ) 122 nm). More significntly, novel thermolysin inhibitors (synthesized s terminl mides) Cbz-Phe p - o Leu-His (K i ) 57 nm), Cbz-Phe p - o Leu-Arg (K i ) 64 nm), nd Cbz-Phe p - o Leu-Gln (K i ) 127 nm) were identified tht were unexpected on the bsis of literture precedent for hydrophobic residues t the P 2 position. 3. Crbonic Anhydrse Inhibitors. Bldwin nd co-workers hve synthesized librry of 6727 cylpiperidines utilizing the split synthesis method nd Figure 27. Building blocks used in the synthesis of librry of crbonic nyhdrse inhibitors. the hloromtic tg encoding strtegy, with the tgs ttched to the polymer support using the rhodium-ctlyzed crbene insertion method (section II.A.5.c). 38,116b The synthesis ws initited by coupling four mino cids nd three mino lcohols to the PEG-PS solid support using 4-crboxy-2-nitrobenzyl lcohol linker through ester or crbonte

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 581 Figure 28. Crbonic nhydrse inhibitors. formtion, respectively. Thirty-one nturl nd unnturl N-Fmoc mino cids were then coupled to the support-bound mines by employing stndrd solid-phse mide bond forming methods. After removl of the Fmoc group, the resulting free mine ws cylted with set of 31 sulfonyl chlorides, isocyntes, crboxylic cids, nd chloroformtes (Figure 27). Thirteen tgs were sufficient to encode the 6727 compounds tht were synthesized. No compound chrcteriztion ws reported, lthough the synthesis sequence involved mide bond formtion nd ws therefore well precedented. The librry ws bised by the inclusion of some rylsulfonmide derivtives which re known phrmcophores to the enzyme crbonic nhydrse (CA). 117 Screening ws ccomplished by photoeluting compounds from either single beds or 10 bed mixtures into the wells of microtiter plte nd by using fluorescence-bsed lignd-displcement ssy (controlled relese llows repeted use of the sme bed, see section II.A.4.c). The structures of the ctive compounds were identified by reding the tg sequences nd verified by resynthesis nd independent evlution. Although detiled description of the ssy protocol nd results re beyond the scope of this review, some generl findings wrrnt comment. As might be expected, high percentge of those compounds incorporting the sulfonmide structure showed significnt ctivity, while no ctive compounds lcking sulfonmide functionlity were identified. Acylpiperidine 131 (Figure 28) ws one of the most potent compounds to be identified with K D of 4 nm for bovine CA(II). Further evlution of 217 member focused cylpiperidine sublibrry vs humn CA(I) nd CA(II) subforms lso showed interesting SAR dt. Almost ll of the compounds tht were ssyed were slightly to modertely selective for CA- (II) over CA(I), suggesting tht CA(II) selectivity my be property of the entire clss of compounds. The Phrmcopei group hs lso synthesized librry of dihydrobenzopyrns, which hs lso been screened ginst CA, with the 15 nm inhibitor 132 being identified. 116 Although few detils of librry synthesis or nlyticl evlution were provided, n overview of the synthesis sequence is provided in Scheme 28. Three dihydroxycetophenones were first coupled to the (4-crboxy-2-nitrobenzyl)oxy photoclevble linker through n ether bond, nd the resulting linker-derived cetophenones were then coupled to PEG-PS support employing stndrd mide bond-forming methods. Condenstion of the resulting support-bound dihydroxycetone derivtives with seven ketones gve the support-bound dihydrobenzopyrns. Four of the ketones (134, Scheme 28) contined Boc-protected mine for further functionliztion (dihydrobenzopyrns 135) while three ketones (cetone, cyclohexnone, nd tetrhydro-4h-pyrn-4-one) did not (dihydropyrns 137). After N-Boc deprotection of compounds 135 with TFA, the mines were functionlized to provide mides, ures, thioures, mines, crbmtes, nd sulfonmides 136 (Scheme 28). The resin btches without mines, 137, were pooled with the functionlized mine resins, 136, nd split into three btches. One btch ws treted with ethnedithiol nd BF 3 OEt 2 to form the dithiocetls 138 nd 138b; the Scheme 28

582 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Figure 29. Lvendustin A. second ws directly cleved, producing compounds 139 nd 139b; nd the third ws reduced with NBH 4, producing compounds 140 nd 140b. The chemistry ws encoded with the hloromtic tg strtegy nd resulted in the synthesis of 1143 compounds. A representtive set of these compounds ws reportedly prepred on beds for chrcteriztion, nd the uthors report tht in generl the dihydrobenzopyrns were isolted with purities of >80%; however, the exct number of compounds nd the structures of the compounds tht were exmined were not provided. The inhibitor 132 (Figure 28) ws resynthesized nd isolted in 95% purity s estimted by HPLC nlysis. 4. Tyrosine Kinse Inhibitors. Method I. Green t Arid Phrmceuticls hs crried out prllel synthesis of 60-member designed librry tht is bsed upon the known tyrosine kinse inhibitor lvendustin A (Figure 29). The librry ws synthesized by reductive mintion of resin-bound niline with substituted benzldehyde followed by lkyltion with benzylic bromide. 118 The initil studies were performed on Rink mide resin, but fter working out the rection conditions, librry synthesis ws crried out on Wng resin. The substituted benzoyl chloride is esterified onto the Wng resin with DMAP s ctlyst. The Fmoc group of the support-bound niline 141 is then removed followed by reductive mintion employing substituted benzldehyde derivtives to provide the resin-bound secondry mine 142 (Scheme 29). Alkyltion with benzyl bromide derivtive followed by clevge from the resin with TFA or BBr 3 when Rink or Wng resin is employed, respectively, provides the desired tertiry mine products 144. Clevge with BBr 3 hs the dded dvntge tht methyl phenyl ethers re Scheme 29 Figure 30. Building blocks used in the synthesis of librry of potentil tyrosine kinse inhibitors. concomitntly cleved to provide the corresponding free phenols. Severl lvendustin A derivtives were synthesized in high isolted yields. The primry mide of tetrmethyllvendustin A ws synthesized on Rink resin in 98% yield fter chromtogrphic purifiction, while tetrmethyllvendustin A ws synthesized on Wng resin with TFA clevge in 78% yield fter purifiction (the crude mteril ws of 95% purity s estimted by HPLC). Finlly, lvendustin A ws synthesized on hydroxymethyl polystyrene with concomitnt methyl ether deprotection nd clevge from the resin in 90% yield (the crude mteril ws of 87% purity s estimted by HPLC). A series of 60 sptilly seprte derivtives were synthesized by employing the building blocks in Figure 30 on Wng resin by employing the BBr 3 - medited clevge nd deprotection sequence. The purity of ech of the compounds ws ssessed by HPLC (30-97%, v 69%), nd the yield for ech compound ws determined by 1 H NMR using mleic cid s n internl stndrd (rnge from 10% to 83%, v 55%). Method II. Reserchers t Sphinx Phrmceuticls hve designed nd synthesized librry of phenol derivtives bsed upon the structures of lvendustin A nd blnol, serine/threonine kinse inhibitor, which shre the phenol s recognition element. 13 4-Amino-3-nitrophenol is first coupled to crboxylte-functionlized polystyrene resin 119 employing DICI (Scheme 30). The support-bound niline 145 is then coupled with n cid chloride or n isocynte (hydroxyl nd crboxyl groups re protected s tert-butyldimethylsilyl ethers). Reduction of the nitro group is then ccomplished with SnCl 2 in DMF, nd the resulting niline is coupled with n cid chloride or isocynte derivtive to introduce functionlity t the second position of diversity. Removl of the silyl groups with Bu 4 NF followed by clevge from the support by tretment with sodium methoxide provides the substituted phenolic derivtives. A smll librry of these derivtives ws synthesized using the previously described microtiter-bsed synthesis pprtus (Figure 2). Benzmide-derivtized resins 146 (17 unique resins were prepred)

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 583 Scheme 30 were loded into the microtiter plte nd then cylted s previously described. Four microtiter pltes (384 compounds) of bis-mides were synthesized, one plte ech (96 compounds) of midesulfonmide nd mide-ures were synthesized, nd one plte of bis-ures 149 were synthesized. Relevnt compounds were deprotected with Bu 4 NF nd the resins cleved to fford 3-9 mg (v weight per microtiter plte, 15-45% bsed on n v moleculr weight of 400) of the desired compounds. Chrcteriztion by positive ion EI MS resulted in the observtion of moleculr ion peks for row of 12 bis-mides. HPLC nlysis of two selected polyhydroxylted phenols gve purities of 87% nd 63%. All 672 compounds were nlyzed by TLC ginst purified stndrds to get qulittive ssessment of product purity. When mixtures were observed, the compounds in the wells were chromtogrphed nd the mjor components were isolted nd chrcterized by 1 H NMR, llowing side products to be identified. The estimted purity for the mide-ure sublibrry ws 60% nd the estimted purity for the bis-ure sublibrry ws 80%. 5. Estrogen Receptor Lignds. Willird nd coworkers hve reported the synthesis of 23 hydroxystilbene derivtives trgeting the estrogen receptor. 120 Four hydroxybenzldehydes tht hd been prederivtized with the HMP linker 74 were coupled to (minomethyl)polystyrene followed by Horner- Emmons olefintion with six different benzylphosphonte nions (Scheme 31). 121 Clevge of the resulting stilbene derivtives from support with TFA provided 7-85% yield of the desired products fter chromtogrphy on the bsis of the loding level of Tble 14. Hydroxystilbene Derivtives 151 (Scheme 31) derivtive entry R 1 R 2 yield (%) 1 3-OH H 44 2 3-OH 4-NO 2 85 3 3-OH 4-Br 45 4 3-OH 3,5-OCH 3 41 5 3-OH 4-F 40 6 3-OH 3-F 35 7 3-OH,4-OCH 3 H 8 8 3-OH,4-OCH 3 4-NO 2 28 9 3-OH,4-OCH 3 4-Br 20 10 3-OH,4-OCH 3 3,5-OCH 3 25 11 3-OH,4-OCH 3 3-F 16 12 3-NO 2,4-OH H 12 13 3-NO 2,4-OH 4-NO 2 8 14 3-NO 2,4-OH 4-Br 7 15 3-NO 2,4-OH 3,5-OCH 3 7 16 3-NO 2,4-OH 4-F 10 17 3-NO 2,4-OH 3-F 14 18 2-Cl,4-OH H 55 19 2-Cl,4-OH 4-NO 2 67 20 2-Cl,4-OH 4-Br 57 21 2-Cl,4-OH 3,5-OCH 3 46 22 2-Cl,4-OH 4-F 40 23 2-Cl,4-OH 3-F 62 Yields of purified mteril re bsed on resin-bound hydroxybenzldehyde strting mteril s the limiting regent. Scheme 32 Scheme 31 the ldehyde on support (Tble 14). Although mny of the yields were modest, enough mteril ws obtined for biologicl screening in cell-culture ssy for ctivtion of the estrogen receptor. Three derivtives bsed on 2-chloro-4-hydroxybenzldehyde showed the highest ctivity, nd dose-response curves were generted tht provided EC 50 vlues of pproximtely 5-15 µm for the three most ctive derivtives. The biologicl response ws likely medited by the estrogen receptor, since the estrogenic ctivity of ll ctive compounds ws effectively inhibited by the ntiestrogenic steroid ICI 164384. 6. Antioxidnts. Kurth hs reported the synthesis nd evlution of nine pools of three ntioxidnts by the split synthesis pproch. 122 Alkyltion of Merrifield resin with the sodium slt of crboxylic cid provided support-bound ester 152 (Scheme 32). Tretment of the ester with LDA t -78 C provided the lithium enolte, which ws trnsmetlted to the zinc enolte using nhydrous zinc chloride t 0 C. Addition of n ldehyde or ketone resulted in ldol condenstion to provide the β-hydroxy ester. The zinc enolte ws employed insted of the lithium enolte in order to minimize retroldoliztion. Clevge from the solid support ws effected by DIBAL

584 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn reduction to the 1,3-diol 154. The yield ws reported for one compound, 26% fter preprtive TLC (R 1 ) CH 2 Ph, R 2 ) OCH 3 ). A librry ws then ssembled using three crboxylic cids nd nine ldehydes to provide 27 compounds which were purified s nine pools of three compounds using preprtive TLC. Chrcteriztion by GC-MS identified ll 27 desired compounds in the librry. The pools were then ssyed for their bility to inhibit oxidtion using the ferric thiocynte method. 123 The most ctive pools were deconvoluted. The three compounds incorporting electronrich 3,4,5-trimethoxybenzldehyde were determined to hve the gretest ctivity. c. Solid-phse Synthesis Methods. A number of significnt studies hve been reported on the soldphse synthesis of orgnic molecules; however, only those studies tht were expressly crried out for the purpose of estblishing rection generlity for librry synthesis will be described in this review. A number of thorough reviews hve been published tht describe the erly work on solid-phse orgnic synthesis. 124 In ddition, severl recent reports from Kurth nd Schore re worthy of note. 125 1. β-mercpto Ketones nd Esters. In one of the erly reports in the re of orgnic compound librry synthesis, Kurth reported the synthesis of nine β-mercpto ketones in three pools of three employing the split synthesis pproch nd demonstrted tht ech compound could be identified by GC-MS methods. 32 Condenstion of commercilly vilble trityl chloride resin with 1,4-butnediol followed by oxidtion of the pendnt primry lcohol with SO 3 - pyridine in DMSO provides the resin-bound ldehyde 155 (Scheme 33). 126 Condenstion with substituted Scheme 33 Horner-Emmons regent introduces the first element of diversity. Tretment of the resulting enone 156 with substituted thiophenol provides the trget β-mercpto ketone, which is cleved from the solid support using formic cid to provide the formte ester. By employing this procedure, n rry of nine compounds ws synthesized s three pools of three compounds. Pools were chrcterized by GC nd GC-MS nd ll nine compounds were determined to be of high purity by GC-MS. Structurl elucidtion of compound tht hd been cleved from one synthesis bed ws lso demonstrted using GC- MS. The isoltion of ech compound from the three pools ws crried out by preprtive TLC. The derivtives were isolted in 7-27% overll yield bsed on the initil trityl chloride loding level of the solid support. 2. Arylcetic Acids. Bckes nd Ellmn hve reported method for the solid-phse synthesis of rylcetic cid derivtives 163 (Scheme 34). 127 Al- Scheme 34 though these compounds represent n importnt clss of cyclooxygense inhibitors, the method ws predominntly developed for the purpose of evluting two importnt crbon-crbon bond-forming rections on solid support; enolte lkyltion nd plldium-medited Suzuki cross-coupling. A crboxylic cid linker ws chosen tht would be stble to the bsic rection conditions of the enolte lkyltion step nd Suzuki cross-coupling process, but tht could lso be ctivted for nucleophilic clevge to complete the synthesis sequence. A vrint of Kenner s sfety-ctch linker ws developed for this purpose. 128 Commercilly vilble 4-crboxybenzensulfonmide ws coupled to the minomethylted polystyrene by tretment with DICI nd HOBt. The pentfluorophenyl ester or the symmetric nhydride of 4-(bromophenyl)cetic cid ws then loded onto the resin employing ctlytic DMAP nd i-pr 2 NEt. Tretment of the cylsulfonmide 159 with excess LDA in THF t 0 C results in rpid deprotontion to give the trinion. Subsequent ddition of ctivted or unctivted lkyl hlides results in rpid lkyltion of the enolte trinion to provide 161. In contrst to ester 129 or crboximide 130 enolte lkyltions, ketene formtion is not observed even when employing the unrective lkylting gent isopropyl iodide (entry 6, Tble 15), since ketene formtion would require tht the sulfonmide dinion be the leving group. In ddition, miniml overlkyltion is observed (<4%). The Suzuki rection of cylsulfonmide 161 is then performed ccording to stndrd conditions using Pd- (PPh 3 ) 4 s the ctlyst, 2 M queous N 2 CO 3 s the bse, nd THF s the solvent t reflux. Deprotontion of the cylsulfonmide under the bsic rection conditions gin prevents ny hydrolysis from occurring. Good conversion is observed both for B-lkyl- 9-BBN derivtives tht re prepred by in situ hydrobortion of primry lkenes nd for rylboronic cids tht re electron poor or electron rich s well s ortho-substituted.

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 585 Tble 15. Substituted Arylcetic Acid Derivtives 163 (Scheme 34) derivtive entry R 1 R 2 nucleophile yield (%) 1 H CH 2CH(CH 3) 2 H 2O 100 2 CH 3 CH 2CH(CH 3) 2 H 2O 96 3 CH 3 CH 2CH(CH 3) 2 BnNH 2 96 4 CH 2C 6H 5 CH 2CH(CH 3) 2 BnNH 2 98 5 CH 2CH 3 CH 2CH(CH 3) 2 BnNH 2 92 6 CH(CH 3) 2 CH 2CH(CH 3) 2 BnNH 2 91 7 CH 3 CH 2CH(CH 3) 2 piperidine 96 8 CH 3 CH 2CH(CH 3) 2 niline 0 b 9 H C 6H 5 H 2O 93 10 CH 3 C 6H 5 BnNH 2 95 11 CH 3 C 6H 4-4-CF 3 BnNH 2 87 12 CH 3 C 6H 4-4-OCH 3 BnNH 2 88 13 CH 3 C 6H 3-2,4-C1 2 BnNH 2 88 Yields of nlyticlly pure mteril re bsed on the loding level of support-bound strting mteril 159. b No clevge of mteril from the resin ws observed with niline s the nucleophile. The finl step in the synthesis is nucleophilemedited clevge of the mteril from the support. Acylsulfonmide ctivtion is ccomplished by tretment with CH 2 N 2 in Et 2 O. Addition of hydroxide or mine nucleophiles provides the corresponding crboxylic cid or mide products 163 in high yield (v 93%, 12 compounds) on the bsis of nlyticlly pure mteril fter filtrtion through silic (Tble 15). Although both primry nd secondry mines resulted in efficient clevge, ttempted clevge of the mteril from the resin with niline did not provide ny nilide product nd defines the level of rectivity of the ctivted 1-methylcylsulfonmide linkge (entry 8). Bckes nd co-workers hve recently reported new method tht provides significntly more lbile ctivted cylsulfonmide linker. 131 Activtion is ccomplished by tretment of the support-bound cylsulfonmide with iodocetonitrile nd i-pr 2 NEt in DMSO or NMP t mbient temperture (Scheme 35). The resulting cynomethylted derivtive is Scheme 35 >150-fold more lbile to nucleophilic displcement thn the corresponding N-methyl derivtive (the t 1/2 for displcement with 0.007 M benzylmine in DMSO is <5 min). Both nonbsic mines nd stericlly hindered mines efficiently rect with the supportbound N-(cynomethyl)cylsulfonmide. For exmple, ctivtion nd nucleophilic displcement of cylsulfonmide 164, with tert-butylmine nd niline provides the corresponding nlyticlly pure mide products in 92% nd 96% yield, respectively, on the Figure 31. HPLC trces of () ddition of limiting mounts of five mines to cylsulfonmide resin 164 resulted in equimolr mounts (( 2%) of the five mide products prepred from (1) 4-(3-minopropyl)morpholine, (2) morpholine, (3) benzylmine, (4) piperidine, nd (5) cyclohexylmine nd (b) stndrd contining n equimolr mixture of the five mide products. bsis of the initil mine loding level of the resin. In ddition, tretment of the ctivted cylsulfonmide with limiting mounts of n mine nucleophile results in complete consumption of the mine to provide the mide product in pure form, uncontminted with excess mine. This llows novel pooling strtegies to be employed whereby limiting mount of n equimolr mixture of mines re dded to the N-(cynomethyl)cylsulfonmide resin to provide n equimolr mixture of mide products in pure form. As illustrted in Figure 31, n equimolr mixture of mide products ws obtined by ctivtion of cylsulfonmide 164 followed by ddition of limiting mounts of n equimolr mixture of piperidine, cyclohexylmine, morpholine, benzylmine, nd 4-(3- minopropyl)morpholine. 3. Plldium-Medited Processes. In ddition to the plldium-medited processes described previously (Stille rections nd Suzuki coupling rections reported by Ellmn, nd Heck rections reported by Zuckermnn), the utility of Pd(0)-medited crboncrbon bond-forming processes on solid-support hs lso been mply demonstrted by number of reserchers. Suzuki Rection. Frenette nd Friesen hve exmined the Suzuki rection on support-bound hlosubstituted benzoic cids (Scheme 36). 132 A number Scheme 36

586 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Tble 16. Biryl Derivtives 167 (Scheme 36) entry ryl hlide R 1 boronic cid R 2 yield (%) 1 4-I 4-OCH 3 >95 2 4-Br 4-OCH 3 >95 3 4-Br H 95 4 4-Br 4-CH 3 95 5 4-Br 4-NO 2 95 6 4-Br 4-F 91 7 3-Br H 90 8 2-Br H >95 9 3-Br,4-OCH 3 H >95 10 3-Br,4-CH 3 H >95 Yields of unpurified mteril re bsed on resin-bound ryl hlide s the limiting regent. All derivtives were of >90% purity s determined by 1 H NMR nd HPLC nlysis. of iodo- or bromo substituted benzoic cids were loded onto Merrifield resin (chloromethylted polystyrene) by esterifiction under stndrd lkyltion conditions (Cs 2 CO 3, KI, DMF). Coupling rections were performed with vriety of plldium ctlysts nd boronic cids to produce the biphenyl derivtives 167, which were cleved from the support using ctlytic NOMe in MeOH/THF. Frenette found tht the optiml rection conditions were 5 mol % Pd- (PPh 3 ) 4 indmewith2mn 2 CO 3 s bse. Crude mss blnces for the rection sequence employing vriety of substituted ryl boronic cids nd hlobenzoic cids were generlly greter thn 95%, nd the compounds were obtined in greter thn 90% purity s estimted by HPLC nlysis (Tble 16). Stille Rection. Deshpnde hs exmined the Stille coupling rection on support in the synthesis of 4-substituted benzmides. 133 4-Iodobenzoic cid ws coupled onto Rink resin or Al-derivtized Wng resin to provide the support-bound ryl iodide (Scheme 37). Stille rection using number of stnnnes ws Scheme 37 then performed with 5 mol % Pd 2 (db) 3 nd dded Ph 3 As to fford the support-bound styrene nd biryl products. Clevge from support using 5% TFA in CH 2 Cl 2 (Rink) or 90% TFA in CH 2 Cl 2 (Wng) provided the products 170 or 171 in crude mss blnces of 85-92% nd in greter thn 90% purity s estimted by HPLC (Tble 17). Tble 17. Stille Rection Products 170 nd 171 (Scheme 37) entry product vinyl stnnne yield (%) b 1 170 CHdCH 2 89 2 170 (Z)-CHdCH(CH 3) 91 3 170 CHdC(CH 3) 2 85 4 170 (E)-CHdCH(Ph) 89 5 170 (E)-CHdCH(Ph-3,4-di-OCH 3) 90 6 171 CHdCH 2 92 7 171 (Z)-CHdCH(CH 3) 88 Compounds 170 were synthesized on Rink resin nd were isolted s the primry mide. Compounds 171 were synthesized on Wng-Al resin. b Yields re bsed on resin-bound 4-iodobenzoic cid s the limiting regent. Scheme 38 Tble 18. Compounds Synthesized by Heck Rection (Scheme 38) entry rectnt product mss blnce (%) 1 (4-crbomethoxy)styrene 175 90 2 phenylcetylene 175 90 3 ethyl crylte 175 91 4 ethyl propenote see b below see b below 5 Ph-I 173 81 6 3-bromonphthyl 173 64 7 2-bromothienyl 173 76 8 3-bromopyridyl 173 87 Mss blnces re bsed on resin-bound 4-vinyl- or 4-iodobenzoic cid s the limiting regent. The products re >90% pure s determined by 1 H NMR nd HPLC nlysis. b A mixture of products ws obtined. Heck Rection. Yu nd co-workers hve exmined the generlity of the Heck rection on both support-bound iodides nd support-bound lkenes under vriety of rection conditions (Scheme 38). 134 Yields of the finl products 173 were excellent when the support-bound ryl iodide ws coupled with different lkene components in solution using Pd- (OAc) 2 in DMF t 80-90 C. The polymer-bound lkene ws lso coupled with four different ryl hlides using Pd 2 (db) 3 nd P(2-tolyl) 3 in DMF t 100 C to give good yields of the finl product 175, lthough Heck rections with ryl trifltes were not successful (Tble 18). Zhou nd co-workers hve extended the scope of the Heck rection on solid support to include rections tht proceed under mild conditions (Scheme 39). 135 Using phse-trnsfer system developed by Scheme 39 Jeffery 136 nd PEG-PS resin from Millipore, the uthors hve shown tht Heck rections cn be performed on support-bound 4-iodobenzoic cid in queous solvent comixtures (DMF/H 2 O/Et 3 N, 1:1:1) using Pd(OAc) 2, PPh 3, nd Bu 4 NCl t 37 C. Yields of Heck products rnge from fir to excellent (54-95%, v 79%, 6 compounds) for vriety of vinylic regents. 4. Amide Bond Formtion on Cyclic Templte. Lebl nd co-workers hve synthesized n ll-ciscyclopentne templte for the disply of mide-bsed functionlity. 137 Compound 179 is synthesized in solution by multistep route strting from com-

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 587 Scheme 40 Tble 19. Biryl Derivtives 184 or 186 (Scheme 41) entry lcohol or phenol R 1 product yield (%) purity (%) b 1 H 184 75 92 2 4-CH 3 184 79 90 3 4-OCH 3 184 75 92 4 4-OPh 184 81 89 5 4-Br 184 90 93 6 4-CO 2CH 3 184 92 96 7 4-CN 184 99 95 8 2-CH 3 184 72 90 9 2-Ph 184 99 97 10 2-CH 3, 4-CHO 184 80 94 11 C 6H 4-4-Br 186 87 92 12 C 6H 4-4-CO 2CH 3 186 77 97 13 see c below 186 94 98 14 Bu 186 66 88 15 (3-C 6H 5)-C 3H 6 186 68 81 Yields of unpurified mteril re bsed on resin-bound ryl hlide s the limiting regent. b Purities were estimted by HPLC nlysis. c 4-[(2-minomethyl)phenyl]phenyl. mercilly vilble cis-5-norbornene-endo-2,3-dicrboxylic nhydride (Scheme 40). The nhydride is opened with resin-bound secondry mine of SCAL 138 -derivtized PEG-PS resin 178 to give 9:1 mixture of the regioisomeric crboxylic cids (only the mjor distereomer is shown). After coupling the resulting free crboxylic cid to n mine using (benzotrizol-1-yloxy)tris(dimethylmino)phosphonium hexfluorophosphte (BOP), the methyl ester is sponified with 0.5% NOH. The resulting crboxylic cid is then coupled with either primry or secondry mine nd BOP. The N-Boc group is then removed by tretment with TFA, nd the resulting mine is cetylted. Finl clevge from the SCAL linker is ccomplished by tretment with 1 M TMSBr nd 1 M thionisole in TFA. The uthors report tht severl derivtives with different side chins were obtined in high yield nd purity employing the described sequence, but no structures or nlyticl dt were provided. The Selectide group hs recently extended this chemistry to include relted scffold, 1,3,5-trimethyl-1,3,5-cyclohexnetricrboxylic cid (Kemp s tricid). 139 5. Ether Formtion. Rno nd Chpmn hve optimized the Mitsunobu coupling rection on solid support for the synthesis of ryl ethers with either the phenol or lcohol serving s the support-bound component (Scheme 41). 140 Either 4-(hydroxymethyl)benzoic cid or 3-(4-hydroxyphenyl)propionic cid ws coupled to Rink-derivtized PEG-PS resin with EDC to give the support-bound lcohol or phenol. Scheme 41 Mitsunobu rection with 5 equiv of both N,N,N,N - tetrmethylzodicrboxmide (TMAD) nd Bu 3 P nd excess of the phenol or lcohol followed by clevge with queous TFA provided Mitsunobu products 184 or 186. Crude mss blnces rnged from 66-99% (v 82%, 15 compounds) nd purities rnged from 81-99% (v 92%) s estimted by HPLC (Tble 19). Mitsunobu rections using diethyl or diisopropyl zodicrboxylte were lso successful, but resulted in slightly lower purities nd yields of the finl products. Krchnk nd co-workers hve lso reported method for the modifiction of resin-bound phenol using the Mitsunobu rection (Scheme 42). 141 O- Scheme 42 Alkyltion of Ac-Tyr-OH nd 4-(hydroxybenzoyl)- glycine bound to PEG-PS resin ws ccomplished using PPh 3 nd DEAD with vriety of primry nd secondry lcohols. The Mitsunobu rection is exothermic; however, rection mixture wrming promotes decomposition of DEAD, liberting ethyl lcohol nd producing the corresponding ethyl ether s byproduct. Accordingly, the rection is performed by premixing the resin, lcohol, nd PPh 3, nd then dding solution of DEAD in THF portionwise over the period of 20 min. DIAD ws lso found to give better results in problem cses, however, the rection time is longer, typiclly 3 h insted of 1 h. The product purities by HPLC rnge from fir to excellent, lthough yields were not reported (Tble 20). This chemistry hs been used to synthesize librry of 4200 compounds using split synthesis. Twenty mino cids were coupled to PEG-PS resin, vriety of 10 romtic hydroxy cids were coupled under stndrd conditions, nd Mitsunobu rection ws

588 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Tble 20. Alkyl Aryl Ethers 188 or 190 (Scheme 42) lcohol product purity (%),b ethyl ether (%),b entry R 1 188 190 188 189 1 methnol 95 99 <1 <1 2 ethnol 96 99 min product 3 2-propnol 93 99 <1 <1 4 1-butnol 75 80 23 20 5 llyl lcohol 95 99 <1 <1 6 1,3-propnediol 94 99 <1 <1 7 benzyl lcohol 99 93 <1 <1 8 4-methoxybenzyl lcohol 98 99 <1 <1 9 4-(methylthio)benzyl lcohol 78 90 <1 <1 10 2-(hydroxymethyl)furn 66 74 9 8 11 3-(hydroxymethyl)furn 98 99 <1 <1 12 2-(hydroxymethyl)thiophene 94 93 2 <1 13 4-methyl-5-(2-hydroxyethyl)thizole 59 85 (96) 2 13 (<1) 14 2-(hydroxymethyl)pyridine 84 99 6 <1 15 3-(hydroxymethyl)pyridine 86 99 6 <1 16 4-(hydroxymethyl)pyridine 57 92 6 <1 17 2,6-bis(hydroxymethyl)pyridine 82 98 9 <1 18 1-(2-hydroxyethyl)pyrrolidine 43 (60) 53 (82) 5 (<1) 8 (<1) 19 1-(2-hydroxyethyl)-2-pyrrolidinone 42 47 (85) 10 47 (9) 20 1-(Boc-mino)ethyl lcohol 39 (52) 51 (82) 46 (29) 45 (15) 21 3-(Fmoc-mino)propyl lcohol 90 69 5 30 Purities were estimted by HPLC nlysis. b Vlues in prentheses refer to the nlogous rection performed with DIAD. performed with 21 lcohols. Structurl determintion nd screening re reported to be in progress. 6. Amine Alkyltion. Bry nd co-workers t Chiron Mimotopes hve demonstrted useful ppliction of the Chiron Mimotopes pin method (see section II.A.1.) for the rpid optimiztion of rections performed on solid support. 142 In specific, multiple rections re performed in prllel in sptilly seprte formt in order to rpidly identify the optiml rection conditions for the trnsformtion of interest. Bry demonstrted this pproch for the two-step sequence of converting support-bound 4-hydroxyproline derivtive to support-bound 4-minoproline derivtive (Scheme 43). A peptidic system Scheme 43 ws chosen to evlute the rection sequence becuse both rection conversion nd product purity hve previously been relibly monitored for peptides by ion spry MS methods. 10 To show the power of the multipin pproch, five rection prmeters were simultneously vried, including the mine structure, the mine concentrtion, the NBH 3 CN concentrtion, the ph, nd the rection solvent. Fiftysix rections were run in prllel followed by clevge nd ion spry MS nlysis. On the bsis of the trends provided by the dt, the optiml rection conditions were identified: 2.0 M [mine], 0.05 M [NBH 3 CN], methnol s the solvent, nd ph ) 5 for nonromtic mines nd ph ) 7 for romtic mines. A set of eight mines with widely different structures were surveyed using the optimized conditions nd representtive exmples of percent conversion were determined by both HPLC nd MS (Tble 21). Good correltion ws generlly observed between the two nlyticl methods, lthough the MS dt tended to give low vlue for mines with two Tble 21. Reductive Amintion Products 192 (Scheme 43) conversion (%) entry mine HPLC MS 1 β-lnine 95 80 2 5-mino-2-methoxypyridine 92 74 3 2-mino-1-propnol 90 88 4 4-bromoniline 92 72 5 3,4-dihydroxybenzylmine 93 86 6 2,2-diphenylethylmine 94 86 7 3-methoxyniline 95 91 8 piperidine 88 88 bsic sites. HPLC conversions rnged from 88-95% nd MS conversions rnged from 72-91%. 7. Ure Formtion. Hutchins nd Chpmn hve reported strightforwrd method for the synthesis of ures on PEG-PS resin (Scheme 44). 143 N-Fmoc- L-glutmic cid R-llyl ester ws ttched to support with the γ-crboxylic cid employing the 4-[4-(hydroxymethyl)-3-methoxyphenoxy]butyric cid (HMPB) hndle. After clevge of the Fmoc group under stndrd conditions, the resulting free mine ws cylted with p-nitrophenylchloroformte nd with i-pr 2 NEt s bse. The ctivted crbmte 194 ws then treted with excess of n mine in DMF t mbient temperture to provide the mide 195. Scheme 44

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 589 Tble 22. Ure Products 196 (Scheme 44) entry mine purity (%) 1 benzylmine 96 2 dmntylmine 92 3 N-(2-chlorophenyl)piperzine 95 4 N-benzylpiperzine 96 5 1,2,3,4-tetrhydroisoquinoline 98 6 niline 96 7 1,2,3,4-tetrhydroquinoline 90 8 4-nitroniline no rection Estimted purity s determined by RP-HPLC. The expected moleculr ions were observed by electrospry MS for ll of the products. Finl clevge from the support ws ccomplished by tretment with 2% TFA in CH 2 Cl 2 to provide products 196 (Tble 22). The uthors hve extended this work to the preprtion of ure-linked dimines p-nitrophenyl crbmte s described previously (Scheme 44) to provide 197. Rection with primry dimine followed by repeting the crbmte synthesis sequence produces compound 198. Ure formtion with vriety of primry or secondry mines followed by clevge produces ure-linked dimines 199 in excellent purity by RP-HPLC nlysis (Tble 23), lthough yields were not reported. 8. Aryl Substitution. Dnkwrdt nd co-workers hve synthesized vriety of ryl- nd benzylpiperzines on solid support. 145 Hloromtic compounds 200-q were coupled to support nd evluted for the bility to rect with phenylpiperzine (Scheme 46). As shown in Tble 24, fluorobenzoic Scheme 46 Scheme 45 for librry synthesis. 144 As shown in Scheme 45, 4-(minomethyl)benzoic cid is linked to PEG-PS resin, followed by ctivtion of the mine s cids ctivted by n ortho or p-nitro group rected efficiently, while less highly ctivted ryl derivtives did not undergo rection. Severl benzylic piperzines were lso formed from the corresponding support-bound benzylic hlides. On the bsis of these results, the five core structures 200-e were chosen for librry synthesis. These cores were coupled to support nd the resins mixed. The resulting resin mixture ws rected with one of 38 different substituted piperzines (structures not provided) to provide 38 pools of five compounds ech. The piperzines hd Tble 23. Ure-Linked Dimines 199 (Scheme 45) entry dimine mine (R 1R 2NH 2) purity (%) 1 1,4-phenylenedimine 1,2,3,4-tetrhydroisoquinoline 99 2 1,4-phenylenedimine niline 89 3 1,4-phenylenedimine benzylmine 99 4 p-xylenedimine 1,2,3,4-tetrhydroisoquinoline 90 5 1,3-cyclohexyldimine 2,2-diphenylethylmine 95 Estimted purity s determined by RP-HPLC. Tble 24. Benzmides 202 (Scheme 46) entry ryl strting mteril (200-q) product purity (%) 1 4-fluoro-3-nitrobenzoic cid 202 82 2 2-fluoro-5-nirtobenzoic cid 202b 81 3 [4-(bromophenyl)phenyl]cetic cid 202c 86 4 3-(chloromethyl)benzoyl chloride 202d 84 5 4-(chloromethyl)benzoic cid 202e 85 6 2-fluoro-3-(trifluoromethyl)benzoic cid 202f NR 7 2-fluoro-5-(trifluoromethyl)benzoic cid 202g 63 + 23% SM 8 4-fluoro-2-(trifluoromethyl)benzoic cid 202h NR 9 4-fluoro-3-(trifluoromethyl)benzoic cid 202i NR 10 4-fluoro-1-nphthoic cid 202j NR 11 2-chloro-4-flourobenzenesulfonyl chloride 202k MP 12 3-chloro-4-flourobenzenesulfonyl chloride 202l NR 13 2-chloro-6-methylpyridine-4-crboxylic cid 202m MP 14 6-chloronicotinic cid 202n NR 15 2-chloro-3-nitrobenzoic cid 202o NR 16 3-chloro-2-nitrobenzoic cid 202p NR 17 5-chloro-2-nitrobenzoic cid 202q NR Estimted purity s determined by RP-HPLC. NR ) No Rection. MP ) multiple products observed.

590 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn been independently shown to rect with supportbound benzmide 201 nd to give products of >70% purity by HPLC. These pools re being screened for biologicl ctivity. 5. Moleculr Recognition in Designed Receptor Systems The moleculr recognition of lignds by biologicl receptors plys criticl role in virtully ll biologicl processes. To better understnd the underlying noncovlent interctions tht re responsible for the selectivity nd specificity of lrge nturl receptors, mny reserchers hve focused on chrcterizing the specificity nd ffinity of smll, well-defined host molecules. Combintoril chemistry hs rpidly become very powerful tool for ccomplishing this gol. Two generl pproches my be envisioned. In one pproch, librry of host molecules is screened ginst defined lignd or smll set of lignds in order to identify the key elements of the host structure tht re necessry to provide ffinity or selectivity. In the second pproch, defined host molecule or set of host molecules is screened ginst librry of lignds. Both of these strtegies hve been implemented nd re described below. Although outside the scope of this review, peptide nd oligonucleotide librry methods re certinly vluble tools for studying nturl receptor-lignd recognition. As prticulrly significnt exmple, Schreiber hs described the ppliction of bised peptide librries to define the recognition motifs of SH3 domins, which medite mny essentil proteinprotein interctions in biologicl systems. 146 Librries of Host Molecules. Still hs explored combintoril strtegies to develop synthetic receptors by the disply of diverse peptides from steroid scffold. 147 In this work, the steroid cheno(12-deoxy)- cholic cid ws coupled to (minomethyl)polystyrene beds by mide bond formtion. The support-bound steroid ws then selectively cylted with Fmocglycyl fluoride t the less hindered C3 position (Figure 32). All combintions of dipeptide were then synthesized from bsis set of 10 Fmoc mino cids using the split synthesis method nd hloromtic tg strtegy to record the synthesis steps. The free mino terminus ws then cpped with cetic nhydride nd the more hindered C7 hydroxyl ws cylted gin using Fmoc-glycyl fluoride, but with DMAP s ctlyst. The dipeptide synthesis nd cpping sequence ws then repeted to produce 10 4 member librry of potentil receptors. The librry of support-bound steroidl receptors ws screened for binding to four different enkephlinlike peptides tht were ech lbeled t the mine terminus with the dye Disperse Red 1 (Figure 32). The support-bound receptor librry ws incubted with the dye-linked peptide in CHCl 3 with the concentrtion of the peptide djusted such tht pproximtely 1% of the beds were stined bright red (100-150 µm depending upon the structure of the dye-linked peptide). For ech binding study, pproximtely 50 of the drkly stined beds were selected nd the tgs red to identify the structures of the tight-binding host molecules. Significnt mino cid preferences were observed for ech of the peptide lignds. Receptors tht bound lignds 206 Figure 32. Steroid receptor. nd 207 hve strong preference for Pro t the AA 2 nd AA 3 sites, while fewer thn 10% of receptors tht were selective for lignds 204 nd 205 contined Pro t those sites. In order to select which host molecules bind selectively s well s tightly to given lignd, competition experiments were performed with peptide lignds 205 nd 206, but while 205 ws still lbeled with Disperse Red 1, 206 ws now lbeled with Disperse Blue 3. Selectivity ws redily observed since bed contining nonselective host would turn purple, while bed tht contins selective host would retin the red or blue color of the respective dye. Severl receptor beds were found to turn bright blue. Upon decoding the blue beds, two predominnt consensus sequences were observed, AA 1 -AA 4 ) Phe,Pro,Pro,Leu nd Asp,Pro,- Pro,Vl. Resynthesis of these receptors followed by HPLC-bsed binding studies confirmed tht these receptors were indeed very selective, with ( G) ) -1.0 nd -1.6 kcl/mol, respectively. None of the receptors screened, however, were ble to selectively bind Leu enkephlin, (L)Tyr-Gly-Gly-(L)Phe-(L)Leu, over Met enkephlin, (L)Tyr-Gly-Gly-(L)Phe-(L)Met, with the most selective receptor identified showing ( G) )-0.2 kcl/mol. Librries of Lignds. Still hs utilized the hloromtic tg encoding strtegy for the synthesis nd evlution of librries of tripeptides in order to define the key structurl elements responsible for recognition by number of host molecules. Two generlized clsses of host molecules hve been investigted, clss of C 3 -symmetric hosts (Figure 33) 148 nd clss of hosts of either D 2 - or tetrhedrl symmetry tht re prepred by the cyclooligomeriztion of 1,3,5-benzenetricrboxylic cid (trimesic cid) nd the 1,2-dimines 216-219 (Figure 34). 149 Both host molecule clsses show high level of selectivity for specific cylted tripeptide sequences. In ddition, despite their lrge structures, compounds in both host molecule clsses cn be prepred very efficiently llowing the rpid synthesis of modified structures for further study. For ll of the host molecules except for the wtersoluble host 213, librry of 50 000 cyl tripeptides ws screened for binding. The librry ws prepred by split synthesis using the hloromtic tg strtegy. The cyl tripeptides were coupled to the minomethylted polystyrene support through n ω-minohexnoic cid linker. Fifteen mino cids were employed for ech of the three positions in the tripeptide, nd 15 cylting gents were employed to cp the tripeptide to provide theoreticl number

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 591 Figure 34. Hosts prepred by the cyclooligomeriztion of trimesic cid nd 1,2-dimines. Figure 33. C 3 -Symmetric hosts. of 50 625 unique compounds. The cyl tripeptide librry ws screened either in side chin-protected form or fter side chin deprotection by trifluorocetic cid tretment. For wter soluble host 213, tripeptide librry ws screened tht hd the generl structure AA3-AA2-AA1-NH(CH 2 ) 2 -support. This librry ws prepred on PEG-PS due to its fvorble solvtion chrcteristics in queous solvents. All possible combintions of 29 mino cids were introduced t ech position using the split synthesis process to provide theoreticl librry size of 24 389 tripeptides. The librry ws screened both in completely deprotected form nd with the side chins protected nd with the N-terminus cetylted. The generl screening procedure ws to mix the support-bound peptide librry with the dye-lbeled host molecule t set concentrtion (usully between 1 nd 100 µm). Chloroform ws employed s the solvent for the polystyrene-bound librries nd the orgnic soluble hosts, nd wter ws employed s the solvent for the PEG-PS-bound librries nd the wter-soluble host 213. After letting the mixture stnd for minimum of 24 h, percentge of the beds becme drkly stined, with the percentge of beds tht were stined depending upon the selectivity nd ffinity of the host molecule s well s the host molecule concentrtion. Multiple drkly stined beds were picked (usully between 30 nd 100) nd the tgs red to provide the structures of the tightbinding tripeptide lignds. A detiled description of the results from ech binding experiment is outside the scope of this review; however, gret del of informtion ws obtined bout the structurl fetures of the tripeptides tht were importnt for binding. Distinct selectivities were observed for different host molecules depending upon the size, shpe, hydrophobicity or hydrophilicity, nd hydrogen-bonding chrcter of the host. The screening protocol provided the positions in the cyl tripeptide tht re importnt for binding. For those positions tht re importnt determinnts for binding, both the side-chin structure nd the stereochemistry of the mino cids were lso provided. In ddition, for severl of the hosts, different fmilies of tight-binding sequences were identified. This lrge body of informtion would hve been difficult if not impossible to obtin by lterntive nonlibrry methods. 6. Anlyticl Techniques Becuse such lrge vriety of nlyticl techniques hve been used to chrcterize solid-phse chemistry or compounds bound to solid supports, only summry of the vilble methods will be presented here. A number of methods re routinely used for monitoring rections. This includes the use of trditionl methods for functionl group titrtion, mny of which cn be quntittive. Some exmples re ninhydrin, 57 picric cid, 150 trinitrobenzenesulfonic cid, 151 nd bromophenol blue 152 tests for free mines, nd Ellmn s test for free thiols. 153 In ddition, the clevge of mny protecting groups gives stble UV-

592 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn ctive byproducts which cn be quntified. This includes the piperidine-dibenzofulvene dduct tht results from Fmoc group clevge with piperidine, 154,155 the highly colored ction tht results from cid-medited dimethoxytrityl ether clevge, 156,157 nd 4-nitrostyrene tht results from deprotection of the (nitrophenyl)ethyloxy group. 115 Additionl methods continue to be developed such s the recent report of Reich nd co-workers on new regent, nitrophenyl O-tritylisothiocynte (NPIT), tht is effective for quntittion of stericlly hindered or nonbsic mines. 158 Other methods tht hve typiclly been used to clculte resin loding include Volhrd titrtion for chloride, 159 nd elementl nlysis for vriety of toms. Mss spectroscopy hs lredy been mentioned s one of the few vilble methods sensitive enough to chrcterize rection products from single synthesis bed; 32-35 however, MS hs mny other useful pplictions in librry nlysis. MS hs short smpling time, llowing rpid MS chrcteriztion of lrge number of smples. 142 Deletion products nd side products often give MS signls tht cn be used to optimize synthesis sequences. 31,34-c MS chrcteriztion of mixtures is possible, 87,160,161 nd judicious choice of compounds tht re pooled cn provide pools where ll compounds hve moleculr ions tht re seprble. 161b Finlly, the use of MS in conjunction with other nlyticl techniques (HPLC-MS, GC-MS, MS-MS) is well documented nd hs been successfully employed in librry chrcteriztion. 31,32,87 Mny NMR nd IR spectroscopic techniques hve been used for the chrcteriztion of compounds ttched to solid supports. Using PEG-PS resins, 13 C NMR spectr of compounds ttched to support cn often be collected. 162 13 C NMR of compounds on stndrd polystyrene resins cn be useful, but the resolution cn be poor reltive to the spectr obtined on PEG-PS resin nd usully require thousnds of trnsients. The use of 13 C-enriched compounds hs been reported to gretly expedite cquisition of dt nd to provide gretly improved signl to noise. 163 Other techniques hve lso been reported including mgic-ngle-spinning (MAS) solid-stte 1 H NMR, 164 MAS 13 C- 1 H correltion experiments, 165 nd recently MAS HMQC nd TOCSY experiments. 166 IR is one of the most convenient methods for evluting support-bound compounds nd hs lso been employed by mny reserchers, including recent report of FT- IR microspectroscopy llowing dt collection from single resin bed. 167 III. Librries Synthesized in Solution While the mjority of published work on smll molecule librry synthesis hs been performed using solid supports, solution-phse strtegies hve been pplied successfully by number of reserchers. A successful solution-phse pproch cn hve significnt dvntges over comprble solid-phse pproch. In prticulr, method does not need to be developed to ttch the initil strting mteril onto support or to cleve the finl product from the support. One step or short rection sequences tht proceed in high yield with stoichiometric regents re Scheme 47 menble to solution-phse strtegies, since purifiction or isoltion is not required. Quite complex structures cn be ccessed in single step by employing multicomponent rections, such s the Ugi rection (vide infr). In ddition, numerous supportbound regents nd ctlysts re vilble tht gretly fcilitte the isoltion of the products in solution nd thereby my provide ccess to multistep rection sequences for the purposes of librry synthesis. 168 Prlow hs provided recent impressive exmple of the potentil ppliction of polymer-bound regents to orgnic synthesis. 169 In his study secphenethyl lcohol ws treted simultneously with three different polymer-supported regents tht would not be comptible if they were in solution; the oxidnt poly(4-vinylpyridinium dichromte), the brominting gent perbromide on Amberlyst A-26, nd the lkyltion gent Amberlite IRA 900 (4-chloro-1-methyl- 5-(trifluoromethyl)-1H-pyrzol-3-ol) (Scheme 47). After filtrtion of the polymer regents, the desired R-lkoxyketone ws obtined in 48% isolted yield. A. Sptilly Seprte Synthesis Few reports hve ppered in the literture on the simultneous solution-phse synthesis of librries in sptilly seprte mnner, lthough number of reserch groups in cdemics nd industry hve presented on this strtegy. 170 Severl reserchers hve focused on solution-phse synthesis using the Ugi rection to provide librries of smll molecules. The Ugi rection uses four different components, n isocynide, n ldehyde, n mine, nd crboxylic cid, to provide one mjor rection product (220, Scheme 48). 171 Ugi hs recently published n over- Scheme 48 view of potentil strtegies for best pplying multicomponent rections to librry synthesis nd evlution. 172 Weber nd co-workers hve employed the Ugi rection to synthesize molecules in sptilly seprte nd prllel formt. 173 A structurlly diverse nd commercilly vilble set of 10 isocynides, 40 ldehydes, 10 mines, nd 40 crboxylic cids were chosen to provide virtul librry of 160 000 compounds. Rection conditions were optimized so tht mjor side product ws observed only when wekly nucleophilic mines, such s 4-minobenzmidine, were used in the coupling process resulting in pprecible mounts of the mine side product 221 tht does not incorporte the crboxylic cid component. An itertive serch nd synthesis procedure ws then employed to identify modertely potent throm-

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 593 Scheme 49 Figure 35. Ugi products selected by screening with genetic lgorithm. bin inhibitor by synthesizing nd evluting only smll frction of the virtul librry of 160 000 possible compounds. By ssigning binry code to the individul components, genetic lgorithm ws employed to select for incresed inhibitory ctivity of the compounds in thrombin ssy. The potentil ppliction of genetic lgorithms to combintoril synthesis hs previously been modeled. 174 Weber developed the genetic lgorithm tht ws employed for the serch nd synthesis procedure by screening the virtul librry ginst different test functions corresponding to compounds with rbitrry biologicl ctivities. The popultion size nd building block interchnges tht correspond to muttion rtes were djusted so tht the most ctive cndidte would be discovered in 10-50 cycles of synthesis nd evlution. Weber performed 20 synthesis nd evlution cycles of 20 Ugi products tht were synthesized in prllel. The first set of Ugi products were selected rndomly from the vilble set of regents. The remining Ugi products for ech synthesis cycle were selected using the developed genetic lgorithm, with building block interchnges corresponding to crossovers nd rndom muttions. From the first cycle to the lst cycle, the verge inhibitory ctivity of the set of 20 Ugi products in the thrombin ssy incresed from pproximtely 1 mm to less thn 1 µm. The most ctive compounds, 222 nd 223 (Figure 35), hd IC 50 vlues of 1.4 µm nd 0.22 µm, respectively. Interestingly, compound 223 did not incorporte the crboxylic cid functionlity nd demonstrtes the importnce of fully chrcterizing rection efficiency. It is lso significnt tht the ppliction of genetic lgorithms is in no wy limited to solution-phse librry pproches. Keting nd Armstrong hve lso investigted the Ugi rection for the purpose of librry synthesis. 175 Hving noted the limited commercil vilbility of isocynides, they developed 1-isocynocyclohexene s universl isocynide, which upon incorportion into n Ugi product (224, Scheme 49) cn be ctivted to introduce n dditionl level of diversity. (Ugi hd previously demonstrted tht Ugi products could in fct be obtined employing isocynocyclohexene. 176 ) Armstrong hs shown tht upon exposure to HCl, Ugi products tht hve incorported isocynocyclohexene become ctivted for nucleophilic displcement presumbly through cycliztion to form the ctivted münchone 226, which then rects with nucleophile to provide product 227. Tretment with H 2 O, lcohol, or mercptn provides the crboxylic cid, ester, or thioester, respectively, in 45-100% isolted yield of purified mteril (Tble 25, verge 70%, 13 compounds). This method provides ccess to vriety of esters, including the stericlly hindered tertbutyl ester (64% yield), significntly incresing the number of Ugi products tht re ccessible. Although mides re not directly vilble using this procedure, they could be synthesized in one step from the corresponding cid. B. Synthesis in Pools 1. A Librry of Amides nd Esters Two reports hve ppered in the literture on pooling strtegies for the rpid synthesis nd evlution of orgnic compound librries in solution. 177 Smith hs synthesized librry of potentilly 1600 esters or mides from the rection of 40 lcohols nd mines with 40 cid chlorides (Scheme 50). 178 Two seprte librries were prepred tht ech contin mides nd esters synthesized from ll 1600 combi- Tble 25. Compounds 227 Synthesized vi the Universl Isocynide Route (Scheme 49) Ugi Product entry R 1 R 2 R 3 XR 4 yield (%) 1 CH 3 CH 2C 6H 4-4-OCH 3 CH(CH 3) 2 OH 56 b 2 CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OH 83 b 3 CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OCH 3 100 b 4 CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OCH 2CH 3 57 b 5 CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OCH 2C 6H 5 75 b 6 CH 3 CH 2C 6H 4-4-OCH 3 CH(CH 3) 2 SCH 2CH 3 68 b 7 CH 3 CH 2C 6H 4-4-OCH 3 1,5-c-C 5H 11 OC(CH 3) 3 64 b 8 (CH 2) 10CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OCH 3 65 c 9 CH 3 CH 2C 6H 4-4-OCH 3 C 6H 5 OCH 3 79 c 10 CH 3 (CH 2) 9CH 3 C 6H 5 OCH 3 99 c 11 H CH 2C 6H 4-4-OCH 3 C 6H 5 NH 2 45 c 12 CH 2C 6H 5 c-c 6H 11 CH(CH 3) 2 OCH 3 67 c 13 CH 3 CH 2C 6H 4-4-OCH 3 1,5-c-C 5H 11 OCH 3 55 c Isolted yield of purified compound. b Yield bsed upon conversion of purified Ugi product 227. c Yield bsed upon intil strting mterils.

594 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Scheme 50 ntions of the 40 lcohol or mine regents nd the 40 cid chloride regents. The first librry is composed of 40 seprte pools of 40 compounds with ech pool contining compounds prepred from single lcohol or mine nd ech of the different cid chlorides. The second librry is composed of 40 pools of 40 compounds with ech pool contining compounds prepred from single cid chloride nd ech of the 40 different lcohols or mines. Screening the first librry set for binding to receptor provides the lcohol or mine component tht puttively contributes the most to the binding ffinity, while screening the second librry provides the cid chloride component tht contributes most to the binding ffinity. Stoichiometric quntities of the regents were used, nd ll of the nucleophiles contined n internl tertiry mine moiety in order to neutrlize HCl liberted in the rection so tht no byproducts were produced. Extended rection times (48 h) were lso employed in order to chieve rection completion. Anlyticl determintion of ll of the compounds in the librries ws not ttempted, but GC nd GC- MS ws performed on two of the pools of 40 compounds. In one pool moleculr ions tht corresponded to 25 out of 40 compounds were observed by electron impct ioniztion while in the second pool 30 out of 40 were observed. The librries were screened in number of ssys, nd two ctive compounds were identified. After synthesis on lrge scle nd purifiction, mide 229 (Figure 36) ws determined to hve n IC 50 vlue of 60 µm for the NK 3 receptor, nd mide 230 ws determined to hve n IC 50 vlue of 55 µm for inhibition of the mtrix metlloproteinse-1 (MMP-1). 2. Acetylcholinesterse Inhibitors Pirrung hs reported generl pooling strtegy for performing solution-bsed librry synthesis nd evlution tht he hs defined s indexed combintoril librries. 179 Pirrung demonstrtes this method with the solution-phse synthesis of 54 crbmtes from the combintion of nine lcohols with six isocyntes (Figure 37). Two librries re gin prepred s pools where ech lcohol is combined with ll isocyntes (librry 1) or ll lcohols re combined with ech isocynte (librry 2). Screening the two librries directly provides the structure of the lcohol nd Figure 36. Novel lignd to the NK 3 receptor nd novel MMP-1 inhibitor. Figure 37. Alcohols nd isocyntes used by Pirrung in librry synthesis.

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 595 Figure 38. Novel cetylcholinesterse inhibitor. isocynte tht contribute most to inhibitory ctivity. Equimolr quntities of ech compound in the individul pools re prepred by employing stoichiometric quntities of the respective lcohols nd isocyntes nd by ensuring complete rection conversion for ll strting mterils by heting the rection mixture in seled pressure tube. Pirrung comments tht the pproch could be extended to compounds tht re prepred from more thn two components, with n different components represented s n n-dimensionl mtrix. He lso points out tht ssy precision will put constrints on the pool size for ccurte librry evlution. The compounds re screened for inhibition of cetylcholinesterse to identify novel inhibitor, 228 (Figure 38), tht hs modest inhibitory ctivity (IC 50 ) 700 µm for cetylcholinesterse inhibition). 3. Amides Displyed from Core Molecule An lterntive strtegy for prepring librries in pools hs been reported by Rebek. 20,161,180 In this pproch, librries re synthesized by treting core molecule tht incorportes cid chloride moieties with n equimolr mixture of vriety of protected mines. The mines re chosen from group of mino cids nd smll heterocyclic mines, with the min criteri for selection being comprble rectivity of the mines towrd the cid chlorides on the core molecule. One tri- nd two tetrcid chloride core molecules hve been employed (Figure 39). A key feture in the selection of the different core molecules is tht the tri- nd tetrmide products disply functionlity in different sptil rrys in order to chieve mximl diversity. For set number of mine components, the theoreticl number of molecules produced per core is dependent upon the number of rective cid chloride sites nd on the symmetry of the core element. Therefore, for the sme set of mines, the higher symmetry of the cubne core 232 will result in smller theoreticl number of molecules thn for the xnthene core 231. Rebek prepred three seprte librries from combintion of 19 diverse mines nd the three different core elements. The theoreticl number of compounds per core were 65 341 molecules for xnthene core 231, 11 191 molecules for cubne core 232, nd 1330 molecules for benzene tricid core 233. To ensure tht ll of the mine building blocks rected in high yield, four equivlents of ech mine were treted with xnthene core 231 in sptilly seprte mnner. All of the expected tetrmide products were obtined in high yield with complete conversion in less thn 30 min. In ddition, lthough complete nlyticl evlution of the three lrge librries ws not possible, the chemistry tht ws employed to prepre the librries ws tested using both HPLC nd ESI-MS nlysis on smller sublibrries designed to cover different mine combintions. For exmple, in one study, xnthene deriv- Figure 39. Core structures utilized by Rebek for librry synthesis. Figure 40. Modified cores utilized by Rebek in model studies nd in deconvolution efforts. Figure 41. Novel trypsin inhibitor. tive 234 (Figure 40) tht contined two cid chlorides ws employed to prepre six model librries employing three sets of eight mines, two sets of nine mines, nd one set of 10 mines to provide librries of 36, 45, nd 55 compounds, respectively. The mine building blocks were grouped in sets such tht ech product in the model librry would contin different moleculr weight. Anlyticl evlution by electrospry ioniztion mss spectrometry demonstrted tht >80% of the compounds for ech sublibrry were observed. 161b The three lrge librries were screened in solution for inhibition of trypsin ctlyzed hydrolysis of N-Rbenzoyl-D,L-rginine-p-nitronilide (BAPA). 180 Although inhibition ws not observed for the librries bsed upon cores 232 nd 233, the librry bsed on core 231 did inhibit the enzyme. Deconvolution ws performed by mking pools missing three mines nd ressying to determine which set of mines ws necessry for inhibitory ctivity. Nine mines were identified, nd then nine sublibrries ech missing one mine were synthesized, nd the five most importnt mines were identified. The modified core 235, in which the two bottom crboxylic cids re protected s benzyl esters, ws then utilized to differentite the top nd bottom positions. Amines were first coupled to the top crboxylic cids, nd the benzyl esters were then cleved by hydrogention followed by coupling mines to the bottom crboxylic cids. The deconvolution procedure resulted in the identifiction of n inhibitor of moderte ctivity (K i ) 9 ( 2 µm), 236 (Figure 41). 4. Oligoscchride Librries Hindsgul hs mde considerble progress towrd utilizing solution-bsed pooling strtegies to synthesize nd evlute di- nd triscchride combintoril librries for the purpose of identifying novel biologi-

596 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn Scheme 51 Figure 42. Lctosylmine linked to Tentgel resin. clly ctive oligoscchride lignds. 181 Becuse ech monoscchride crries t lest three hydroxyl groups, the development of n orthogonl protecting group strtegy tht would provide ccess to ll possible monoscchride combintions s well s ll possible connectivities nd both R- nd β-isomers would be extremely chllenging, prticulrly for triscchrides. Hindsgul therefore chose to investigte rndom glycosyltion pproch whereby librries of smll pools of di- or triscchrides would be generted tht include ll possible connectivities between the monoscchride strting mterils (e112 compounds). Literture precedent suggests tht rndom glycosyltion would be difficult to chieve since glycosyltion rte is highly dependent on the steric nd electronic environment of the lcohol cceptor. However, two thoroughly chrcterized exmples demonstrte the potentil utility of this pproch. As shown in Scheme 51, rndom glycosyltion of N-cetylglucosmine (GlcNAc) derivtive 238 with the percetylted glctosyl (Gl) donor 237 provided significnt mount of ech of the six possible discchride products (three different connectivities with both R- nd β-isomers). The rection ws terminted t 40% conversion to minimize the formtion of tri- nd higher order oligoscchride derivtives. The hydrophobic p-methoxyphenoxyoctyl glycon of cceptor 238 fcilitted seprtion of the discchride products from rection byproducts using reverse-phse HPLC nd lso provided chromophore for ccurte isomer quntittion. Similrly, rndom fucosyltion of βgl(1f3)βglcnac-or 240 provided ll six possible R-fucosylted products (8-23%), lthough only minor quntities of the β-fucosylted isomers were observed (Scheme 52). In these two studies, ll of the isomeric products were rigorously chrcterized by NMR, MS, nd methyltion nlysis. Whether or not this strtegy cn be pplied to the mjority of the other mmmlin monoscchrides remins to be demonstrted. Scheme 52 Severl reserchers hve reported very elegnt pproches for the solid-phse synthesis of oligoscchrides nd peptide-oligoscchride hydrids. 182 Although these methods were developed for importnt pplictions independent of librry synthesis, they provide the groundwork for ny future efforts in the solid-phse synthesis of oligoscchride librries. Vetter nd Gllop hve lso developed solutionphse synthesis of glycosylmines. By using procedure described by Kochetkov, 183 commercilly vilble reducing sugrs re treted with sturted mmonium crbonte solution to convert the nomeric hydroxyl to n mine. 184 This simple, onestep rection provides the β-nomer in >95% isomeric purity in lmost ll cses, nd ws demonstrted on 54 chrged, neutrl, nd di- nd oligoscchrides. Monocyltion of the glycosylmine with excess disuccimidyl suberte provides n ctive crbohydrte conjugte which cn be purified by precipittion nd linked to minomethyl Tentgel resin for screening purposes 30b (Figure 42). The recognition of crbohydrte-conjugted beds by lectin ws demonstrted. Preprtion of seven crbohydrtebed conjugtes from the glycosylmines of mltose, chitobiose, lctose, silyllctose, LcNAc, GlcNAc6SO 3, nd GlcNAcPO 2-3 ws followed by mixing the beds. Incubtion with fluorescently lbeled lectin (whet germ gglutinin or ricinus communis gglutinin) followed by sorting using FACS instrument showed selective binding of the lectin to different frctions of the pool. Although the structures of the sugrs bound from the pool were not directly determined, chitobisylmide-derivtized beds bound to whet germ gglutinin nd lctosylmide beds bound ricinus communis gglutinin, s determined by ssying ech sugr-bed conjugte individully. These discchrides re known lignds for the corresponding lectins. This work hs been extended to the synthesis of N-linked glycopeptides. 185 By incorporting llyl ester protection of Glu or Asp into peptide librry, these side chins cn be selectively deprotected with plldium nd ctivted with pentfluorophenyl trifluorocette. Condenstion with vriety of glycosylmines gives the support-bound glycopeptide, which cn be deprotected nd cleved from the support. Glycopeptides were synthesized from 18 different mono- nd oligoscchridrylmines nd five support-bound peptides, which were derivtives of the Leu-enkephlin sequence (Tyr-Gly-Gly-Phe-Leu) with n Asp or Glu dded in vrious plces to bind the sugr. Mono- nd most discchride-derived glycopeptides were isolted in quntittive yields, nd unchrged oligoscchrides gve yields in the 50-80% rnge. Chrged sugrs gve considerbly lower conversions (30-50%). IV. Future Directions A number of importnt new dvnces continue to be mde in the synthesis nd pplictions of com-

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 597 pound librries. Although much of the chemistryperformed on solid support will continue to be bsed upon nlogous chemistry in solution, new chemistry tht does not hve current solution-phse counterprt will lso be developed. There is lso considerble enthusism for combintoril biosynthesis strtegies for chieving librries of diverse nd complex structures, prticulrly through genetic mnipultion of polyketide biosynthetic pthwys. 186 The synthesis nd evlution of librries will incresingly be used not only to study ligndreceptor interctions, but in ny re of chemistry where the identifiction of the optiml chemicl structure for prticulr ppliction typiclly requires the synthesis nd evlution of mny different compounds. Two exmples where reports hve lredy ppered re the evlution of symmetric ctlysts for the synthesis of opticlly ctive compounds 187 nd the development of new mterils. 188 V. Acknowledgments Support to J.A.E. ws provided by the Ntionl Institutes of Helth, the Ntionl Science Foundtion, the A. P. Slon Foundtion, nd the Burroughs Wellcome Fund. L.A.T. is grteful to Glxo-Wellcome for predoctorl fellowship. The uthors thnk the mny reserchers who provided copies of mnuscripts tht were in press t the time of submission. VI. Glossry AC BOP Bpoc CA DCC DEAD DIAD DICI DBU EDC Fmoc HATU HMPB HMP HOBt MBHA NMP NPEOC Nvoc PAL PEG-PS PPTS PyBOP PyBrOP Rink SCAL Tentgel TFA cid clevble linker, [4-(hydroxymethyl)-3- methoxyphenoxy]cetic cid (benzotrizol-1-yloxy) tris(dimethylmino)phosphonium hexfluorophosphte [[2-(4-biphenyl)isopropyl]oxy]crbonyl crbonic nhydrse 1,3-Dicyclohexylcrbodiimide diethyl zodicrboxylte diisopropyl zodicrboxylte 1,3-diisopropylcrbodiimide 1,8-dizbicyclo[5.4.0]undecne ethyl-3-[3-(dimethylmino)propyl]crbodiimide [(fluorenylmethyl)oxy]crbonyl O-(7-zbenzotrizol-1-yl)-1,1,3,3-tetrmethyluronium hexfluorophosphte 4-[4-(hydroxymethyl)-3-methoxyphenoxy]butyric cid linker derived from [4-(hydroxymethyl)phenoxy]- cetic cid hydroxybenzotrizole 4-methylbenzhydrylmine N-methylpyrrolidinone [(nitrophenyl)ethoxy]crbonyl [(nitroveritryl)oxy]crbonyl peptide mide linker derived from 5-[4-(minomethyl)-3,5-dimethoxyphenoxy]vleric cid polystyrene-polyethylene glycol grft copolymer pyridinium p-toluenesulfonte (benzotrizol-1-yloxy)tris(pyrrolidino)phosphonium hexfluorophosphte bromotris(pyrrolidino)phosphonium hexfluorophosphte linker derived from 4-[[(2,4-dimethoxyphenyl)- mino]methyl]phenol sfety ctch mide linker, 4-[4,4 -bis(methylsulfinyl)-2-oxy-(9-fluorenylmethyloxycrbonyl] benzhydrylmino]butnoic cid PEG-PS polymer mrketed by Rpp Polymere trifluorocetic cid THP TMAD TMG Trt Wng linker derived from 6-(hydroxymethyl)-3,4-dihydro-2H-pyrn N,N,N,N -tetrmethylzodicrboxmide tetrmethylgunidine resin derived with trityl chloride linker derived from (hydroxymethyl)phenol VII. Bibliogrphy (1) Grcheck, S. J.; Miller, P. F.; Mrks, J. S. Annu. Rep. Med. Chem. 1993, 28, 161-167. (2) () Gllop, M. A.; Brrett, R. W.; Dower, W. J.; Fodor, S. P. A.; Gordon, E. M. J. Med. Chem. 1994, 37, 1233-1251. (b) Pinill, C.; Appel, J.; Blondelle, S.; Dooley, C.; Dorner, B.; Eichler, J.; Ostresh, J.; Houghten, R. A. Biopolymers (Pept. Sci.) 1995, 37, 221-240. (c) Pvi, M. R.; Swyer, T. K.; Moos, W. H. BioMed. Chem. Lett. 1993, 3, 387-396. (d) Jung, G.; Becksickinger, A. G. Angew. Chem., Int. Ed. Engl. 1992, 31, 367-383. (e) Dower, W. J.; Fodor, S. P. A. Annu. Rep. Med. Chem. 1991, 26, 271-280. In ddition, n excellent bibliogrphy of rticles in the field of librry synthesis (both peptide librries nd orgnic compound librries) is mintined on the world wide web by the journl Moleculr Diversity. The site is currently ccessed t the URL http://vest.pd.com/index.html nd is edited by Dr. Michl Lebl. (3) () Ecker, D. J.; Vickers, T. A.; Hneck, R.; Driver, V.; Anderson, K. Nucleic Acids Res. 1993, 21, 1853-1856. (b) Gold, L.; Polisky, B.; Uhlenbeck, O.; Yrus, M. Annu. Rev. Biochem. 1995, 64, 763-797. (4) () Smith, A. B.; Hirschmnn, R.; Psternk, A.; Akishi, R.; Guzmn, M. C.; Jones, D. R.; Keenn, T. P.; Sprengeler, P. A.; Drke, P. L.; Emini, E. A.; Hollowy, M. K.; Schleif, W. A. J. Med. Chem. 1994, 37, 215-218. (b) Conrdi, R. A.; Hilgers, A. R.; Ho, N. F. H.; Burton, P. S. Phrm. Res. 1992, 9, 435-439. (c) Hirschmnn, R. Angew. Chem., Int. Ed. Engl. 1991, 30, 1278-1301. (d) Humphrey, M. J.; Ringrose, P. S. Drug Metb. Rev. 1986, 17, 283-310. (e) Bell, J.; Peters, G. E.; McMrtin, C.; Thoms, N. W.; Wilson, C. G. J. Phrm. Phrmcol. 1984, 36, 88P. (f) Veber, D. F.; Sperstein, R.; Nutt, R. F.; Freidinger, R. M.; Brdy, S. F.; Curley, P.; Perlow, D. S.; Pleved, W. J.; Colton, C. D.; Zcchei, A. G.; Tocco, D. J.; Hoff, D. R.; Vndlen, R. L.; Gerich, J. E.; Hll, L.; Mndrino, L.; Cordes, E. H.; Anderson, P. S.; Hirschmnn, R. Life Sci. 1984, 34, 1371-1378. (5) Kim, E. E.; Bker, C. T.; Dwyer, M. D.; Murcko, M. A.; Ro, B. G.; Tung, R. D.; Nvi, M. A. J. Am. Chem. Soc. 1995, 117, 1181-1182. (6) () Gordon, E. M.; Brrett, R. W.; Dower, W. J.; Fodor, S. P. A.; Gllop, M. A. J. Med. Chem. 1994, 37, 1385-1401. (b) Terrett, N. K.; Grdner, M.; Gordon, D. W.; Kobylecki, R. J.; Steele, J. Tetrhedron 1995, 51, 8135-8173. (c) Blondelle, S. E.; Perezpy, E.; Dooley, C. T.; Pinill, C.; Houghten, R. A. TrcsTrends Anl. Chem. 1995, 14, 83-92. (d) Lebl, M.; Krchnk, V.; Sepetov, N. F.; Seligmnn, B.; Strop, P.; Felder, S.; Lm, K. S. Biopolymers (Pept. Sci.) 1995, 37, 177-198. (e) Desi, M. C.; Zuckermnn, R. N.; Moos, W. H. Drug Dev. Res. 1994, 33, 174-188. (f) Jnd, K. D. Proc. Ntl. Acd. Sci. U.S.A. 1994, 91, 10779-10785. (g) Moos, W. H.; Green, G. D.; Pvi, M. R. Annu. Rep. Med. Chem. 1993, 28, 315-324. (7) () Deprez, B.; Willird, X.; Bourel, L.; Coste, H.; Hyfil, F.; Trtr, A. J. Am. Chem. Soc. 1995, 117, 5405-5406. (b) Bry hs recently demonstrted the use of similr pproch for the identifiction of lignds from hexpeptide librry contining pproximtely 24 million peptides: Bry, A. M. 1996, submitted for publiction. (8) Strtegies hve been developed specificlly for the prllel solidphse synthesis of peptides nd oligonucleotides. These hve been recently reviewed: See refs 2 nd 2d. (9) Geysen, H. M.; Meloen, R. H.; Brteling, S. J. Proc. Ntl. Acd. Sci. U.S.A. 1984, 81, 3998-4002. (10) () Bry, A. M.; Vlerio, R. M.; Dispsqule, A. J.; Greig, J.; Meji, N. J. J. Peptide Sci. 1995, 1, 80-87. (b) Vlerio, R. M.; Bry, A. M.; Meji, N. J. Int. J. Pept. Protein Res. 1994, 44, 158-165. (c) Geysen, H. M.; Rodd, S. J.; Mson, T. J.; Tribbick, G.; Schoofs, P. G. J. Immun. Methods 1987, 102, 259-274. (e) Meji, N. J.; Bry, A. M.; Vlerio, R. M.; Wng, W. Pept. Res. 1995, 8, 33-38. (11) Bunin, B. A.; Plunkett, M. J.; Ellmn, J. A. Methods Enzymol. 1996, in press. (12) () DeWitt, S. H.; Kiely, J. S.; Stnkovic, C. J.; Schroeder, M. C.; Cody, D. M. R.; Pvi, M. R. Proc. Ntl. Acd. Sci. U.S.A. 1993, 90, 6909-6913. (b) Cody, D. R.; DeWitt, S. H. H.; Hodges, J. C.; Kiely, J. S.; Moos, W. H.; Pvi, M. R.; Roth, B. D.; Schroeder, M. C.; Stnkovic, C. J. United Sttes Pt. 5,324,483, June 28, 1994. (13) Meyers, H. V.; Dilley, G. J.; Durgin, T. L.; Powers, T. S.; Winssinger, N. A.; Zhu, H.; Pvi, M. R. Moleculr Diversity 1995, 1, 13-20. (14) Advnced Chemtech Inc., 5609 Fern Vlley Rd., Louisville, KY. Phone (502)-969-0000. (15) () Holmes, C. P.; Adms, C. L.; Kochersperger, L. M.; Mortensen, R. B.; Aldwin, L. A. Biopolymers (Pept. Sci.) 1995, 37, 199-211. (b) Jcobs, J. W.; Fodor, S. P. A. Trends Biotechnol. 1994, 12,19-26. (c) Pirrung, M. C.; Red, J. L.; Fodor, S. P. A.; Stryer,

598 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn L. United Sttes Pt. 5,143,854, 1992. (d) Rozsnyi, L. F.; Benson, D. R.; Fodor, S. P. A.; Schultz, P. G. Angew. Chem., Int. Ed. Engl. 1992, 31, 759-761. (e) Fodor, S. P. A.; Red, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Sols, D. Science 1991, 251, 767-773. (16) Newmn, T. H.; Willims, K. E.; Pese, R. F. W. J. Vc. Sci. Technol. B 1987, 5, 88. (17) Cho, C. Y.; Morn, E. J.; Cherry, S. R.; Stephns, J. C.; Fodor, S. P. A.; Adms, C. L.; Sundrm, A.; Jcobs, J. W.; Schultz, P. G. Science 1993, 261, 1303-1305. Although support-bound ssys hve been successful t identifying high-ffinity lignds, they re usully used for peptide librries: see ref 2. (18) Geysen, H. M.; Rodd, S. J.; Mson, T. J. Mol. Immunol. 1986, 23, 709-715. (19) () Ostresh, J. M.; Winkle, J. H.; Hmshin, V. T.; Houghten, R. A. Biopolymers 1994, 34, 1681-1689. (b) Rutter, W. J.; Snti, D. V. United Sttes Pt. 5,010,175, 1991. (20) A nice exmple is the synthesis of mixtures of mides by coupling mixtures of mines with cid chlorides: Crell, T.; Wintner, E. A.; Bshirhshemi, A.; Rebek, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 2059-2061. (21) () Sebestyen, F.; Dibo, G.; Kovcs, A.; Furk, A. BioMed. Chem. Lett. 1993, 3, 413-418. (b) Furk, A.; Sebestyen, F.; Asgedom, M.; Dibo, G. Int. J. Pept. Protein Res. 1991, 37, 487-493. (c) Lm, K. S.; Slmon, S. E.; Hersh, E. M.; Hruby, V. J.; Kzmeierski, W. M.; Knpp, R. J. Nture 1992, 360, 768-768. (d) Owens, R. A.; Gesellchen, P. D.; Houchins, B. J.; Dimrchi, R. D. Biochem. Biophys. Res. Commun. 1991, 181, 402-408. (e) Houghten, R. A.; Pinill, C.; Blondelle, S. E.; Appel, J. R.; Dooley, C. T.; Cuervo, J. H. Nture 1991, 354, 84-86. (f) Lm, K. S.; Slmon, S. E.; Hersh, E. M.; Hruby, V. J.; Kzmierski, W. M.; Knpp, R. J. Nture 1991, 354, 82-84. (22) Burgess, K.; Liw, A. I.; Wng, N. Y. J. Med. Chem. 1994, 37, 2985-2987. (23) () Zuckermnn, R. N.; Bnville, S. C. Pept. Res. 1992, 5, 169-174. (b) Zuckermnn, R. N.; Kerr, J. M.; Sini, M. A.; Bnville, S. C. Int. J. Pept. Protein Res. 1992, 40, 497-506. (24) () Dooley, C. T.; Chung, N. N.; Wilkes, B. C.; Schiller, P. W.; Bidlck, J. M.; Psternk, G. W.; Houghten, R. A. Science 1994, 266, 2019-2022. (b) Blondelle, S. E.; Tkhshi, E.; Weber, P. A.; Houghten, R. A. Antimicrob. Agents Chemother. 1994, 38, 2280-2286. (25) () Eichler, J.; Luck, A. W.; Houghten, R. A. Proceedings of the 23rd Europen Peptide Symposium, Brg, September 1994; Poster 198. (b) Wytt, J. R.; Vickers, T. A.; Roberson, J. L.; Buckheit, R. W.; Klimkit, T.; Debets, E.; Dvis, P. W.; Ryner, B.; Imbch, J. L.; Ecker, D. J. Proc. Ntl. Acd. Sci. U.S.A. 1994, 91, 1356-1360. (c) Dooley, C. T.; Houghten, R. A. Life Sci. 1993, 52, 1509-1517. (26) Beutel, B. A.; Bellomy, G. R.; Wldron, J. A.; Voorbch, M. J.; Bertelsen, A. H. 1996, submitted for publiction. (27) Erb, E.; Jnd, K. D.; Brenner, S. Proc. Ntl. Acd. Sci. U.S.A. 1994, 91, 11422-11426. (28) () Pinill, C.; Appel, J. R.; Blondelle, S. E.; Dooley, C. T.; Eichler, J.; Ostresh, J. M.; Houghten, R. A. Drug Dev. Res. 1994, 33, 133-145. (b) Houghten, R. A. Gene 1993, 137, 7-11. (c) Pinill, C.; Appel, J. R.; Blnc, P.; Houghten, R. A. Biotechniques 1992, 13, 901-905. (29) Employing oligonucleotide librries nd RNA hybridiztion s the ssy, Freier hs demonstrted tht positionl scnning is less effective thn deconvolution strtegies for identifying the most ctive compound in librry when multiple ctive compounds re present: Freier, S. M.; Konings, D. A. M.; Wytt, J. R.; Ecker, D. J. J. Med. Chem. 1995, 38, 344-352. (30) () Cmpbell, D. A.; Bermk, J. C.; Burkoth, T. S.; Ptel, D. V. J. Am. Chem. Soc. 1995, 117, 5381-5382. (b) Vetter, D.; Tte, E. M.; Gllop, M. A. Bioconjugte Chem. 1995, 6, 319-322. (c) Lm, K. S.; Lebl, M. Methods, A Compnion to Methods in Enzymology 1994, 6, 372-380. (d) Needels, M. C.; Jones, D. G.; Tte, E. H.; Heinkel, G. L.; Kochersperger, L. M.; Dower, W. J.; Brrett, R. W.; Gllop, M. A. Proc. Ntl. Acd. Sci. U.S.A. 1993, 90, 10700-10704. (e) Fodor, S. P. A.; Red, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Sols, D. Science 1991, 251, 767-773. (31) Stnkov, M.; Isskov, O.; Sepetov, N. F.; Krchnk, V.; Lm, K. S.; Lebl, M. Drug Dev. Res. 1994, 33, 146-156. (32) Chen, C. X.; Rndll, L. A. A.; Miller, R. B.; Jones, A. D.; Kurth, M. J. J. Am. Chem. Soc. 1994, 116, 2661-2662. (33) Brown, B. B.; Wgner, D. S.; Geysen, H. M. Mol. Diversity 1995, 1, 4-12. (34) () Youngquist, R. S.; Fuentes, G. R.; Lcey, M. P.; Keough, T. J. Am. Chem. Soc. 1995, 117, 3900-3906. (b) Egner, B. J.; Lngley, G. J.; Brdley, M. J. Org. Chem. 1995, 60, 2652-2653. (c) Youngquist, R. S.; Fuentes, G. R.; Lcey, M. P.; Keough, T. Rpid Commun. Mss Spectr. 1994, 8, 77-81. (d) Zmbis, R. A.; Boulton, D. A.; Griffin, P. R. Tetrhedron Lett. 1994, 35, 4283-4286. (35) Brummel, C. L.; Lee, I. N. W.; Zhou, Y.; Benkovic, S. J.; Winogrd, N. Science 1994, 264, 399-402. (36) Structurl determintion is further complicted for oligomer librries where observtion of the moleculr ion cn t most provide the structure of the individul monomers tht re incorported into the oligomer, but not the sequence of the oligomer, lthough tndem MS-MS sequencing methods could provide oligomer sequences. (37) () Lebl, M.; Ptek, M.; Kocis, P.; Krchnk, V.; Hruby, V. J.; Slmon, S. E.; Lm, K. S. Int. J. Pept. Protein Res. 1993, 41, 201-203. (b) Slmon, S. E.; Lm, K. S.; Lebl, M.; Kndol, A.; Khttri, P. S.; Wde, S.; Ptek, M.; Kocis, P.; Krchnk, V.; Thorpe, D.; Felder, S. Proc. Ntl. Acd. Sci. U.S.A. 1993, 90, 11708-11712. (c) Kocis, P.; Krchnk, V.; Lebl, M. Tetrhedron Lett. 1993, 34, 7251-7252. (38) () Bldwin, J. J.; Burbum, J. J.; Henderson, I.; Ohlmeyer, M. H. J. J. Am. Chem. Soc. 1995, 117, 5588-5589. (b) In ddition, Holmes hs described new photolbile linker for solid-phse synthesis which llows more rpid relese of supprt-bound mteril nd produces nitoso ketone insted of recteve nitroso ldehyde: Holmes, C. P.; Jones, D. G. J. Org. Chem. 1995, 60, 2318-2319. (39) Brenner, S.; Lerner, R. A. Proc. Ntl. Acd. Sci. U.S.A. 1992, 89, 5381-5383. (40) () Nielsen, J.; Jnd, K. D. Methods: A Compnion to Methods in Enzymology 1994, 6, 361-371. (b) Nielsen, J.; Brenner, S.; Jnd, K. D. J. Am. Chem. Soc. 1993, 115, 9812-9813. (41) Kerr, J. M.; Bnville, S. C.; Zuckermnn, R. N. J. Am. Chem. Soc. 1993, 115, 2529-2531. (42) Nikoliev, V.; Stierndov, A.; Krchnk, V.; Seligmnn, B.; Lm, K. S.; Slmon, S. E.; Lebl, M. Pept. Res. 1993, 6, 161-170. (43) () Nicolou, K. C.; Xio, X. Y.; Prndoosh, Z.; Senyei, A.; Nov, M. P. Angew. Chem., Int. Ed. Engl. 1995, 34, 2289-2291. (b) Morn, E. J.; Srshr, S.; Crgill, J. F.; Shhbz, M. M.; Lio, A.; Mjlli, A. M. M.; Armstrong, R. W. J. Am. Chem. Soc. 1995, 117, 10787-10788. (44) () Nestler, H. P.; Brtlett, P. A.; Still, W. C. J. Org. Chem. 1994, 59, 4723-4724. (b) Ohlmeyer, M. H. J.; Swnson, R. N.; Dillrd, L. W.; Reder, J. C.; Asouline, G.; Kobyshi, R.; Wigler, M.; Still, W. C. Proc. Ntl. Acd. Sci. U.S.A. 1993, 90, 10922-10926. (45) Jnd hs recently reported on n interesting ppliction of liquid phse synthesis for the construction nd evlution of librries: Hn, H.; Wolfe, M. M.; Brenner, S.; Jnd, K. D. Proc. Ntl. Acd. Sci. U.S.A. 1995, 92, 6419-6423. (46) Merrifield, R. B. J. Am. Chem. Soc. 1963, 85, 2149-2154. (47) Byer, E. Angew. Chem., Int. Ed. Engl. 1991, 30, 113-129. (b) Byer, E.; Rpp, W. Ger. Offen. DE Pt. 3,500,180, 1986. Tentgel resin is commercilly vilble from Rpp Polymere: Eugenstrsse 38/1 D 7400 Tübingen, Germny. Phone: x49-7071-35107. (48) Ostresh, J. M.; Husr, G. M.; Blondelle, S. E.; Dorner, B.; Weber, P. A.; Houghten, R. A. Proc. Ntl. Acd. Sci. U.S.A. 1994, 91, 11138-11142. (49) Simon, R. J.; Kni, R. S.; Zuckermnn, R. N.; Huebner, V. D.; Jewell, D. A.; Bnville, S.; Ng, S.; Wng, L.; Rosenberg, S.; Mrlowe, C. K.; Spellmeyer, D. C.; Frnkel, A. D.; Snti, D. V.; Cohen, F. E.; Brtlett, P. A. Proc. Ntl. Acd. Sci. U.S.A. 1992, 89, 9367-9371. (50) () Miller, S. M.; Simon, R. J.; Ng, S.; Zuckermnn, R. N.; Kerr, J. M.; Moos, W. H. Drug Dev. Res. 1995, 35, 20-32. (b) Miller, S. M.; Simon, R. J.; Ng, S.; Zuckermnn, R. N.; Kerr, J. M.; Moos, W. H. BioMed. Chem. Lett. 1994, 4, 2657-2662. (51) Zuckermnn, R. N.; Kerr, J. M.; Kent, S. B. H.; Moos, W. H. J. Am. Chem. Soc. 1992, 11, 10646-10647. (52) Zuckermnn, R. N.; Mrtin, E. J.; Spellmeyer, D. C.; Stuber, G. B.; Shoemker, K. R.; Kerr, J. M.; Figliozzi, G. M.; Goff, D. A.; Sini, M. A.; Simon, R. J.; Bnville, S. C.; Brown, E. G.; Wng, L.; Richter, L. S.; Moos, W. H. J. Med. Chem. 1994, 37, 2678-2685. (53) Mrtin, E. J.; Blney, J. M.; Sini, M. A.; Spellmeyer, D. C.; Wong, A. K.; Moos, W. H. J. Med. Chem. 1995, 38, 1431-1436. (54) Zuckermnn, R. N.; Kerr, J. M.; Sini, M. A.; Bnville, S. C. Int. J. Pept. Protein Res. 1992, 40, 497-506. (55) Pei, Y. H.; Moos, W. H. Tetrhedron Lett. 1994, 35, 5825-5828. (56) Morn, E. J.; Wilson, T. E.; Cho, C. Y.; Cherry, S. R.; Schultz, P. G. Biopolymers (Pept. Sci.) 1995, 37, 213-219. (57) () Srin, V. K.; Kent, S. B. H.; Tm, J. P.; Merrifield, R. B. Anl. Biochem. 1981, 117, 147-157. (b) Kiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anl. Biochem. 1970, 34, 595-598. (58) Burgess, K.; Linthicum, D. S.; Shin, H. W. Angew. Chem., Int. Ed. Engl. 1995, 34, 907-909. (59) () Gennri, C.; Nestler, H. P.; Slom, B.; Still, W. C. Angew. Chem., Int. Ed. Engl. 1995, 34, 1763-1765. (b) Gennri, C.; Nestler, H. P.; Slom, B.; Still, W. C. Angew. Chem., Int. Ed. Engl. 1995, 34, 1765-1768. (60) Gennri, C.; Slom, B.; Potenz, D.; Willims, A. Angew. Chem., Int. Ed. Engl. 1994, 33, 2067-2069. (61) Hgihr, M.; Anthony, N. J.; Stout, T. J.; Clrdy, J.; Schreiber, S. L. J. Am. Chem. Soc. 1992, 114, 6568-6570. (62) () Smith, A. B.; Guzmn, M. C.; Sprengeler, P. A.; Keenn, T. P.; Holcomb, R. C.; Wood, J. L.; Crroll, P. J.; Hirschmnn, R. J. Am. Chem. Soc. 1994, 116, 9947-9962. (b) Smith, A. B.; Hirschmnn, R.; Psternk, A.; Akishi, R.; Guzmn, M. C.; Jones, D. R.; Keenn, T. P.; Sprengeler, P. A.; Drke, P. L.; Emini, E. A.; Hollowy, M. K.; Schleif, W. A. J. Med. Chem. 1994, 37, 215-218. (c) Smith, A. B.; Holcomb, R. C.; Guzmn, M. C.; Keenn, T. P.; Sprengeler, P. A.; Hirschmnn, R. Tetrhedron Lett. 1993, 34, 63-66. (d) Smith, A. B.; Keenn, T. P.; Holcomb,

Synthesis nd Applictions of Smll Molecule Librries Chemicl Reviews, 1996, Vol. 96, No. 1 599 R. C.; Sprengeler, P. A.; Guzmn, M. C.; Wood, J. L.; Crroll, P. J.; Hirschmnn, R. J. Am. Chem. Soc. 1992, 114, 10672-10674. (63) In designing librries for led identifiction, one would prefer to hve mny different modertely sized librries bsed upon compounds tht disply functionlity with different topologies nd vrying degrees of rigidity thn to hve single lrge librry bsed upon one compound clss. (64) Ariens, E. J.; Beld, A. J.; Rodrigues de Mirnd, J. F.; Simonis, A. M. In The Receptors: A Comprehensive Trety; O Brien, R. D., Ed.; Plenum: New York, 1979; Vol. 1, pp 33-41. (65) It is therefore not surprising tht mny of the synthesis schemes described in this review incorporte mino cids not only becuse mny re commericlly vilble, but becuse they re commericlly vilble in side chin-preprotected form. (One would ssume tht the commericl vilbility of diverse building blocks will continue to grow under the impetus of librry synthesis efforts.) (66) Sternbch, L. H. J. Med. Chem. 1979, 22, 1-7. (67) Bock, M. G.; Diprdo, R. M.; Evns, B. E.; Rittle, K. E.; Whitter, W. L.; Veber, D. F.; Anderson, P. S.; Freidinger, R. M. J. Med. Chem. 1989, 32, 13-16. (68) Römer, D.; Buschler, H. H.; Hill, R. C.; Murer, R.; Petcher, T. J.; Zeugner, H.; Benson, W.; Finner, E.; Milkowski, W.; Thies, P. W. Nture 1982, 298, 759-760. (69) Korneki, E.; Erlich, Y. H.; Lenox, R. H. Science 1984, 226, 1454-1456. (70) Hsu, M. C.; Schutt, A. D.; Holly, M.; Slice, L. W.; Shermn, M. I.; Richmn, D. D.; Potsh, M. J.; Volsky, D. J. Science 1991, 254, 1799-1802. (71) () Bondinell, W. E.; Cllhn, J. F.; Huffmn, W. F.; Keenn, R. M.; Ku, T. W.-F.; Newlnder, K. A. Pt. WO 93/00095, 1993. (b) Puwels, R.; Andries, K.; Desmyter, J.; Schols, D.; Kukl, M. J.; Breslin, H. J.; Reymeckers, A.; Vn Gelder, J.; Woestenborghs, R.; Heyknts, J.; Schellekens, K.; Jnssen, M. A. C.; Clercq, E. D.; Jnsen, P. A. J. Nture 1990, 343, 470-474. (72) Jmes, G. L.; Goldstein, J. L.; Brown, M. S.; Rwson, T. E.; Somers, T. C.; McDowell, R. S.; Crowley, C. W.; Lucs, B. K.; Levinson, A. D.; Mrsters, J. C. Science 1993, 260, 1937-1942. (73) Bunin, B. A.; Ellmn, J. A. J. Am. Chem. Soc. 1992, 114, 10997-10998. (74) Shepprd, R. C.; Willims, B. J. Int. J. Pept. Protein Res. 1982, 20, 451-454. (75) Crpino, L. A.; Sdtlee, D.; Cho, H. G.; Deselms, R. H. J. Am. Chem. Soc. 1990, 112, 9651-9652. (76) The pk of 5-(phenylmethyl)-2-oxzolidone is 20.5 in DMSO s determined by Bordwell. Evns, D. A.; Britton, T. C.; Ellmn, J. A.; Dorow, R. L. J. Am. Chem. Soc. 1990, 112, 4011-4030. Lithited 5-(phenylmethyl)-2-oxzolidinone is employed rther thn unsubstituted 2-oxzolidinone due to its greter solubility in THF. (77) Bunin, B. A.; Plunkett, M. J.; Ellmn, J. A. Proc. Ntl. Acd. Sci. U.S.A. 1994, 6, 4708-4712. (78) Evns, B. E.; Rittle, K. E.; Bock, M. G.; DiPrdo, R. M.; Freidinger, R. M.; Whitter, W. L.; Gould, N. P.; Lundell, G. F.; Homnick, C. F.; Veber, D. F.; Anderson, P. S.; Chng, R. S. L.; Lotti, V. J.; Cerino, D. J.; Chen, T. B.; King, P. J.; Kunkel, K. A.; Springer, J. P.; Hirschfield, J. J. Med. Chem. 1987, 30, 1229-1239. (79) Buddie, R. A.; Levin, V. Anderson Cncer Center, Houston, TX, unpublished results, 1995. (80) Glick, G. University of Michign, unpublished results, 1995. (81) Plunkett, M. J.; Ellmn, J. A. J. Am. Chem. Soc. 1995, 117, 3306-3307. (82) These yields represent the lower limit of the rection efficiency, s when compound 31 is deprotected, cetylted, nd cleved, 4-hydroxycetnilide is isolted in only 92% yield on the bsis of the minomethyl loding level of the resin, indicting tht cyltion to provide 31 did not proceed to completion. (83) Plunkett, M. J.; Ellmn, J. A. J. Org. Chem. 1995, 60, 6006-6007. (84) Boojmr, C. G.; Burow, K. M.; Ellmn, J. A. J. Org. Chem. 1995, 60, 5742-5743. (85) Lndi, J. J.; Rmig, K. Syn. Commun. 1991, 21, 167-171. (86) Goff, D. A.; Zuckermnn, R. N. J. Org. Chem. 1995, 60, 5744-5745. (87) Gordon, D. W.; Steele, J. BioMed. Chem. Lett. 1995, 5, 47-50. (88) Goff, D. A.; Zuckermnn, R. N. J. Org. Chem. 1995, 60, 5748-5749. (89) Gordeev, M. F.; Ptel, D. V.; Gordon, E. M. J. Org. Chem. 1996, in press. (90) Meutermns, W. D. F.; Alewood, P. F. Tetrhedron Lett. 1995, 36, 7709-7712. (91) Murphy, M. M.; Schullek, J. R.; Gordon, E. M.; Gllop, M. A. J. Am. Chem. Soc. 1995, 117, 7029-7030. (92) Look, G. C.; Murphy, M. M.; Cmpbell, D. A.; Gllop, M. A. Tetrhedron Lett. 1995, 36, 2937-2940. (93) () Brr, D. A.; Grigg, R.; Gunrtne, H. Q. N.; Kemp, J.; McMeekin, P.; Sridhrn, V. Tetrhedron 1988, 44, 557-570. (b) Tsuge, O.; Knems, S.; Yoshiok, M. J. Org. Chem. 1988, 53, 1384-1391. (94) Ptek, M.; Drke, B.; Lebl, M. Tetrhedron Lett. 1995, 36, 2227-2230. (95) Holmes, C. P.; Chinn, J. P.; Look, G. C.; Gordon, E. M.; Gllop, M. A. J. Org. Chem. 1995, 60, 7328-7333. (96) Srshr, S.; Mjlli, A. M. M.; Siev, D. 1996, submitted for publiction. (97) Richter, L. S.; Gdek, T. R. Tetrhedron Lett. 1994, 35, 4705-4706. (98) Rose, G. D.; Giersch, L. M.; Smith, J. A. Adv. Protein Chem. 1985, 37, 1-109. (99) For reviews on β-turn mimetics see: () Bll, J. B.; Alewood, P. F. J. Mol. Recogn. 1990, 3, 55-64. (b) Khn, M. (guest editor) Tetrhedron 1993, 49, (Symposi 50) 3433-3677. (100) Virgilio, A. A.; Ellmn, J. A. J. Am. Chem. Soc. 1994, 116, 11580-11581. (101) Crpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397-4398. (102) Reserchers t Genentech hve reported the synthesis of 15- membered ring RGD mimetics by thiolkyltion: Brker, P. L.; Bullens, S.; Bunting, S.; Burdick, D. J.; Chn, K. S.; Deisher, T.; Eigenbrot, C.; Gdek, T. R.; Gntzos, R.; Lipri, M. T.; Muir, C. D.; Npier, M. A.; Pitti, R. M.; Pdu, A.; Qun, C.; Stnley, M.; Struble, M.; Tom, J. Y. K.; Burnier, J. P. J. Med. Chem. 1992, 35, 2040-2048. (103) Cycliztion to provide 99 ws initilly ttempted by mcrolctmiztion between the i + 1 nd i + 2 residues rther thn by thiolkyltion; however, cycliztion with rnge of ctivting gents nd solid supports provided significnt mounts of cyclic dimer. These results re in ccord with the well-precedented difficulties in mcrolctmiztion to provide nine- nd 10- membered ring structures. () Story, S. C.; Aldrich, J. V. Int. J. Pept. Protein Res. 1994, 43, 292-296. (b) Kemp, D. S.; Stites, W. E. Tetrhedron Lett. 1988, 29, 5057-5060. (104) Virgilio, A. A.; Ellmn, J. A. In Combintoril Chemistry nd Moleculr Diversity in Drug Discovery; Gordon, E. M., Kerwin, J. F., Jr., Eds.; Wiley: New York, in press. (105) Virgilio, A. A.; Ellmn, J. A. Unpublished results. (106) As detiled in this review, severl groups hve lso identified protese inhibitors using lterntive librry templtes. Purely peptide-bsed inhibitors hve lso been identified but re outside the scope of this review. (107) Alewood, P. F.; Brinkworth, R. I.; Dncer, R. J.; Grnhm, B.; Jones, A.; Kent, S. B. H. Tetrhedron Lett. 1992, 33, 977-980. (108) Other reserchers hve previously performed nlogous work in solution: () Rich, D. H.; Green, J.; Toth, M. V.; Mrshll, G. R.; Kent, S. B. H. J. Med. Chem. 1990, 33, 1285-1288. (b) Arrowsmith, R. J.; Dvies, D. E.; Fogden, Y. C.; Hrris, C. J.; Thompson, C. Tetrhedron Lett. 1987, 28, 5569. (c) Godfrey, J. D.; Gordon, E. M.; Von Lngen, D.; Engebrecht, J.; Pluscec, J. J. Org. Chem. 1986, 51, 3073. (109) Kick, E. K.; Ellmn, J. A. J. Med. Chem. 1995, 38, 1427-1430. (110) Thompson, L. A.; Ellmn, J. A. Tetrhedron Lett. 1994, 35, 9333-9336. (111) Mjer, P.; Rndd, R. S. J. Org. Chem. 1994, 59, 1937-1938. (112) Brtr, M.; Rome, P.; Urpi, F.; Vilrrs, J. Tetrhedron 1990, 46, 587-594. (113) West, M. L.; Firlie, D. P. Trends Phrmcol. Sci. 1995, 16,67-75. (114) Wng, G. T.; Li, S.; Wideburg, N.; Krfft, G. A.; Kempf, D. J. J. Med. Chem. 1995, 38, 2995-3002. (115) Cmpbell, D. A.; Bermk, J. C. J. Am. Chem. Soc. 1994, 116, 6039-6040. (116) () Chbl, J. C.; Bldwin, J. J.; Burbum, J. J.; Chelsky, D.; Dillrd, L. W.; Henderson, I.; Li, G.; Ohlmeyer, H. J.; Rndle, T. L.; Reder, J. C.; Rokosz, L.; Sigl, N. H. In Genomes, Moleculr Biology nd Drug Discovery; Acdemic Press: New York, in press. (b) Burbum, J. J.; Ohlmeyer, M. H. J.; Reder, J. C.; Henderson, I.; Dillrd, L. W.; Li, G.; Rndle, T. L.; Sigl, N. H.; Chelsky, D.; Bldwin, J. J. Proc. Ntl. Acd. Sci. U.S.A. 1995, 92, 6027-6031. (117) Ponticello, G. S.; Freedmn, M. B.; Hbecker, C. N.; Lyle, P. A.; Schwm, H.; Vrg, S. L.; Christy, M. E.; Rndll, W. C.; Bldwin, J. J. J. Med. Chem. 1987, 30, 591-597. (118) Green, J. J. Org. Chem. 1995, 60, 4287-4290. (119) Frrll, M. J.; Frechet, J. M. J. J. Org. Chem. 1976, 41, 3877-3882. (120) Willird, R.; Jmmlmdk, V.; Zv, D.; Benz, C. C.; Hunt, C. A.; Kushner, P. J.; Scnln, T. S. Chem. Biol. 1995, 2, 45-51. (121) Leznoff, C. C. Acc. Chem. Res. 1978, 11, 327-333. (122) Kurth, M. J.; Rndll, L. A. A.; Chen, C. X.; Melnder, C.; Miller, R. B.; Mclister, K.; Reitz, G.; Kng, R.; Nktsu, T.; Green, C. J. Org. Chem. 1994, 59, 5862-5864. (123) () Imi, S.; Morikiyo, M.; Furiht, K.; Hykw, Y.; Seto, H. Agric. Biol. Chem. 1990, 54, 2367-2371. (b) Nktni, N.; Intni, R. Agric. Biol. Chem. 1983, 47, 353-358. (c) Osw, T.; Nmiki, M. Agric. Biol. Chem. 1981, 45, 735-739. (124) () Frechet, J. M. J. Tetrhedron 1981, 37, 663-683. (b) Leznoff, C. C. Acc. Chem. Res. 1978, 11, 327-333. (c) Crowley, J. I.; Rpoport, H. Acc. Chem. Res. 1976, 9, 135-144. (125) () Beebe, X.; Schore, N. E.; Kurth, M. J. J. Org. Chem. 1995, 60, 4196-4203. (b) Beebe, X.; Chippri, C. L.; Olmsted, M. M.; Kurth, M. J.; Schore, N. E. J. Org. Chem. 1995, 60, 4204-4212. (c) Duffy, J. L.; Kurth, M. J. J. Org. Chem. 1994, 59, 3783-3785. (d) Moon, H. S.; Eisenberg, S. W. E.; Wilson, M. E.; Schore, N. E.; Kurth, M. J. J. Org. Chem. 1994, 59, 6504-6505. (e) Moon,

600 Chemicl Reviews, 1996, Vol. 96, No. 1 Thompson nd Ellmn H.; Schore, N. E.; Kurth, M. J. Tetrhedron Lett. 1994, 35, 8915-8918. (f) Beebe, X.; Chippri, C. L.; Kurth, M. J.; Schore, N. E. J. Org. Chem. 1993, 58, 7320-7321. (g) Moon, H. S.; Schore, N. E.; Kurth, M. J. J. Org. Chem. 1992, 57, 6088-6089. (h) Beebe, X.; Schore, N. E.; Kurth, M. J. J. Am. Chem. Soc. 1992, 114, 10061-10062. (i) Kim, H. R.; Kim, H. J.; Duffy, J. L.; Olmsted, M. M.; Ruhlndtsenge, K.; Kurth, M. J. Tetrhedron Lett. 1991, 32, 4259-4262. (j) Schore, N. E.; Njdi, S. D. J. Am.Chem. Soc. 1990, 112, 441-442. (k) Njdi, S.; Reichlin, D.; Kurth, M. J. J. Org.Chem. 1990, 55, 6241-6244. (126) Monoprotecion of diols using trityl chloride resin hs been reported by Leznoff: Leznoff, C. C.; Fyles, T. M. J. Chem. Soc., Chem. Commun. 1976, 251-252. (127) Bckes, B. J.; Ellmn, J. A. J. Am. Chem. Soc. 1994, 116, 11171-11172. (128) Kenner, G. W.; McDermott, J. R.; Shepprd, R. C. J. Chem. Soc., Chem. Commun. 1971, 636-637. (129) Rthke, M. W.; Woodbury, R. P.; Sullivn, D. F. J. Org. Chem. 1977, 42, 2023-2039. (130) Evns, D. A.; Mthre, D. J.; Ennis, M. D. J. Am. Chem. Soc. 1982, 104, 1737-1739. (131) Bckes, B. J.; Virgilio, A. A.; Ellmn, J. A. 1996, submitted for publiction. (132) Frenette, R.; Friesen, R. W. Tetrhedron Lett. 1994, 35, 9177-9180. (133) Deshpnde, M. S. Tetrhedron Lett. 1994, 35, 5613-5614. (134) Yu, K. L.; Deshpnde, M. S.; Vys, D. M. Tetrhedron Lett. 1994, 35, 8919-8922. (135) Hiroshige, M.; Huske, J. R.; Zhou, P. Tetrhedron Lett. 1995, 36, 4567-4570. (136) Jeffery, T. Tetrhedron Lett. 1994, 35, 3051-3054. (137) Ptek, M.; Drke, B.; Lebl, M. Tetrhedron Lett. 1994, 35, 9169-9172. (138) Ptek, M.; Lebl, M. Tetrhedron Lett. 1991, 32, 3891-3894. (139) Kocis, P.; Isskov, O.; Sepetov, N. F.; Lebl, M. Tetrhedron Lett. 1995, 36, 6623-6626. (140) Rno, T. A.; Chpmn, K. T. Tetrhedron Lett. 1995, 36, 3789-3792. (141) Krchnk, V.; Flegelov, Z.; Weichsel, A. S.; Lebl, M. Tetrhedron Lett. 1995, 36, 6193-6196. (142) Bry, A. M.; Chiefri, D. S.; Vlerio, R. M.; Meji, N. J. Tetrhedron Lett. 1995, 36, 5081-5084. (143) Hutchins, S. M.; Chpmn, K. T. Tetrhedron Lett. 1994, 35, 4055-4058. (144) Hutchins, S. M.; Chpmn, K. T. Tetrhedron Lett. 1995, 36, 2583-2586. (145) Dnkwrdt, S. M.; Newmn, S. R.; Krstennsky, J. L. Tetrhedron Lett. 1995, 36, 4923-4926. (146) For n excellent recent review, see: Chen, J. K.; Schreiber, S. L. Angew. Chem., Int. Ed. Engl. 1995, 34, 953-969. (147) Boyce, R.; Li, G.; Nestler, H. P.; Sueng, T.; Still, W. C. J. Am. Chem. Soc. 1994, 116, 7955-7956. (148) () Crrsco, M. R.; Still, W. C. Chem. Biol. 1995, 2, 205-212. (b) Yoon, S. S.; Still, W. C. Angew. Chem., Int. Ed. Engl. 1995, 33, 2458-2460. (c) Borchrdt, A.; Still, W. C. J. Am. Chem. Soc. 1994, 116, 373-374. (d) Borchrdt, A.; Still, W. C. J. Am. Chem. Soc. 1994, 116, 7467-7468. (e) Erickson, S. D.; Simon, J. A.; Still, W. C. J. Org. Chem. 1993, 58, 1305-1308. (149) () Wennemers, H.; Yoon, S. S.; Still, W. C. J. Org. Chem. 1995, 60, 1108-1109. (b) Torneiro, M.; Still, W. C. J. Am. Chem. Soc. 1995, 117, 5887-5888. (c) Yoon, S. S.; Still, W. C. J. Am. Chem. Soc. 1993, 115, 823-824. (150) Gisin, B. Anl. Chim. Act 1972, 58, 248-249. (151) Hncock, W. S.; Bttersby, J. E. Anl. Biochem. 1976, 71, 260-269. (152) Krchnák, V.; Vágner, J.; Sfár, P.; Lebl, M. Czech. Chem. Commun. 1988, 53, 2542-2548. (153) Ellmn, G. L. Arch. Biochem. Biophys. 1959, 82, 70-77. (154) () Heimer, E. P.; Chng, C.-D.; Lmbros, T. L.; Meienhofer, J. Int. J. Pept. Protein Res. 1981, 18, 237-241. (b) Chng, C.-D.; Wki, M.; Ahmd, M.; Meienhofer, J.; Lundell, E. O.; Hug, J. D. Int. J. Pept. Protein Res. 1980, 15, 59-66. (c) Meienhofer, J.; Wki, M.; Heimer, E. P.; Lmbros, T. J.; Mkofske, R. C.; Chng, C.-D. Int. J. Pept. Protein Res. 1979, 13, 35-42. (155) These nd other methods for monitoring mine coupling re summrized in n excellent review on the utility of Fmoc mino cids in solid-phse peptide synthesis: Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990, 35, 161-214. (156) Cruthers, M. H.; Brone, A. D.; Beucge, S. L.; Dodds, D. R.; Fisher, E. F.; McBride, L. J.; Mtteucci, M.; Stbinsky, Z.; Tng, J.-Y. In Methods in Enzymology; Wu, R., Grossmn, L., Eds.; Acdemic Press: Sn Diego, 1987; Vol. 154, pp 287-313. (157) For recent review, see: Beucge, S. L.; Iyer, R. P. Tetrhedron 1991, 48, 2223-2311. (158) Chu, S. S.; Reich, S. H. BioMed. Chem. Lett. 1995, 5, 1053-1058. (159) Lu, G.-S.; Mojsov, S.; Tm, J. P.; Merrifield, R. B. J. Org. Chem. 1981, 46, 3433-3436. (160) () Jung, G.; Becksickinger, A. G. Angew. Chem., Int. Ed. Engl. 1992, 31, 367-383. (b) Metzger, J. W.; Wiesmuller, K. H.; Gnu, V.; Brunjes, J.; Jung, G. Angew. Chem., Int. Ed. Engl. 1993, 32, 894-896. (161) () Dunyevskiy, Y.; Vouros, P.; Crell, T.; Wintner, E. A.; Rebek, J., Jr. Anl. Chem. 1995, 67, 2906-2915. (b) Crell, T.; Wintner, E. A.; Sutherlnd, A. J.; Rebek, J.; Dunyevskiy, Y. M.; Vouros, P. Chem. Biol. 1995, 2, 171-183. (162) Byer, E.; Albert, K.; Willisch, H.; Rpp, W.; Hemmsi, B. Mcromolecules 1990, 23, 1937. (163) Look, G. C.; Holmes, C. P.; Chinn, J. P.; Gllop, M. A. J. Org. Chem. 1994, 59, 7588-7590. (164) Fitch, W. L.; Detre, G.; Holmes, C. P.; Shoolery, J. N.; Keifer, P. A. J. Org. Chem. 1994, 59, 7955-7956. (165) Anderson, R. C.; Jrem, M. A.; Shpiro, M. J.; Stokes, J. P.; Ziliox, M. J. Org. Chem. 1995, 60, 2650-2651. (166) Anderson R. C.; Stokes, J. P.; Shpiro, M. J. Tetrhedron Lett. 1995, 36, 5311-5314. (167) Yn, B.; Kumrvel, G.; Anjri, H.; Wu, A.; Petter, R. C.; Jewell, C. F., Jr.; Wreing, J. R. J. Org. Chem. 1995, 60, 5736-5738. (168) () Desi, M. C.; Strmiello, L. M. S. Tetrhedron Lett. 1993, 34, 7685-7688. (b) Hodge, P. In Innovtion nd Perspectives in Solid Phse Synthesis, 1990; Epton, R., Ed.; Collected Ppers, First Interntionl Symposium, 1989; Oxford, Eng. SPCC (UK) Ltd.: Birminghm, 1990. (c) Hodge, P. In Syntheses nd Seprtions Using Functionl Polymers; Hodge, P., Sherrington, D. C., Eds.; John Wiley nd Sons: New York, 1988. (d) Hodge, P. In Synthesis nd Seprtions using Functionl Polymers; Sherrington, D. C., Hodge, P., Eds.; Wiley: Chichester, 1988. (e) Bergbreiter, D. E. In Polymer Regents nd Ctlysts; Ford, W. T., Ed.; ACS Symposium Series 308; Americn Chemicl Society: Wshington, DC, 1986. (f) Hodge, P. In Polymer-Supported Rections in Orgnic Synthesis; Hodge, P., Sherrington, D. C., Eds.; Wiley-Interscience: Chichester, 1980. (g) Pittmn, C. U., Jr. In Polymer-Supported Rections in Orgnic Synthesis; Hodge, P., Sherrington, D. C., Eds.; Wiley-Interscience: Chichester, 1980. (169) Prlow, J. J. Tetrhedron Lett. 1995, 36, 1395-1396. (170) () Peterson, J. R. Pnlbs. Presenttion t Explioting Moleculr Diversity: Smll Molecule Librries for Drug Discovery. Jnury 23-25, L Joll, CA, 1995. (b) Hogn, J. C., Jr. Presenttion t Explioting Moleculr Diversity: Smll Molecule Librries for Drug Discovery. Jnury 23-25, L Joll, CA, 1995. (171) Ugi, I.; Domling, A.; Horl, W. Endevour 1994, 18, 115-122. (172) Ugi, I.; Goebel, M.; Gruber, B.; Heilingbrunner, M.; Heiss, C.; Hörl, W.; Kern, O.; Strnecker, M. Res. Chem. Intermed. 1995, in press. (173) Weber, L.; Wllbum, S.; Broger, C.; Guberntor, K. Angew. Chem., Int. Ed. Engl. 1995, 34, 2280-2282. (174) Sheridn, R. P.; Kersley, S. K. J. Chem. Inf. Comput. Sci. 1995, 35, 310-320. (175) Keting, T. A.; Armstrong, R. W. J. Am. Chem. Soc. 1995, 117, 7842-7843. (176) Rosendhl, F. K.; Ugi, I. Ann. Chem. 1963, 666, 65-67. (177) A recent publiction describes new screening method tht is demonstrted with peptides: Sepetov, N. F.; Krchnk, V.; Stnkov, M.; Wde, S.; Lm, K. S.; Lebl, M. Proc. Ntl. Acd. Sci. U.S.A. 1995, 92, 5426-5430. (178) Smith, P. W.; Li, J. Y. Q.; Whittington, A. R.; Cox, B.; Houston, J. G. BioMed. Chem. Lett. 1994, 4, 2821-2824. (179) Pirrung, M. C.; Chen, J. J. Am. Chem. Soc. 1995, 117, 1240-1245. (180) Crell, T.; Wintner, E. A.; Rebek, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 2061-2064. (181) Ding, Y.; Knie, O.; Lbbe, J.; Plcic, M. M.; Ernst, B.; Hindsgul, O. In Glycoimmunology; Alvid, A., Axford, J. S., Eds.; Plenum Press: New York, in press. (182) For leding references, see: () Hlcomb, R. L.; Hung, H. M.; Wong, C. H. J. Am. Chem. Soc. 1994, 116, 11315-11322. See Also: (b) Roberge, J. Y.; Beebe, X.; Dnishefsky, S. J. Science 1995, 269, 202-204. (c) Yn, L.; Tylor, C. M.; Goodnow, R.; Khne, D. J. Am. Chem. Soc. 1994, 116, 6953-6954. (d) Dnishefsky, S. J.; Mcclure, K. F.; Rndolph, J. T.; Ruggeri, R. B. Science 1993, 260, 1307-1309. (183) Likhosherstov, L. M.; Novikov, O. S.; Derevitskj, V. A.; Kochetkov, N. K. Crbohydr. Res. 1986, 146, C1-C5. (184) Vetter, D.; Gllop, M. A. Bioconjugte Chem. 1995, 6, 316-318. (185) Vetter, D.; Tumelty, D.; Singh, S. K.; Gllop, M. A. Angew. Chem., Int. Ed. Engl. 1995, 34, 60-63. (186) () Rohr, J. Angew. Chem., Int. Ed. Engl. 1995, 24, 881-885. (b) Tsoi, J.; Khosl, C. Chem. Biol. 1995, 2, 355-362. (187) Liu, G.; Ellmn, J. A. J. Org. Chem. 1995, 60, 7712-7713. (188) () Xing, X. D.; Sun, X. D.; Briceno, G.; Lou, Y. L.; Wng, K. A.; Chng, H. Y.; Wllce-freedmn, W. G.; Chen, S. W.; Schultz, P. G. Science 1995, 268, 1738-1740. (b) Briceno, G.; Chng, H. Y.; Sun, X. D.; Schultz, P. G.; Xing, X. D. Science 1995, 270, 273-275. CR9402081