Generation and utilization of synthetic combinatorial libraries

Generation and utilization of synthetic combinatorial libraries

Generation and utilization of synthetic combinatorial libraries Jutta Eichler and Richard A. Houghten D ! Jutta Eichler Richard A. Houghten Torrey P...

3MB Sizes 1 Downloads 18 Views

Recommend Documents

Exploring immunological specificity using synthetic peptide combinatorial libraries
The definition of epitopes for human B and T cells is fundamental for the understanding of the immune response mechanism

Catalytic antibodies: Perusing combinatorial libraries
Combinatorial libraries are a promising alternative for isolating catalytic antibodies produced by the immune system in

Simplified Synthetic Antibody Libraries
Synthetic antibody libraries are constructed from scratch using designed synthetic DNA. Precise control over design enab

α-Azidoesters as divergent intermediates for combinatorial generation of glucofuranose libraries of novel N-linked glycopeptides
Epimeric azidoesters containing a glucofuranosyl moiety are readily prepared from glucoheptonolactone and may be reduced

Computer Algebra Libraries for Combinatorial Structures
This paper introduces the framework of decomposable combinatorial structures and their traversal algorithms. A combinato

Combinatorial synthesis of dihydropyridone libraries and their derivatives
Polymer-supported quench methodology has been used for parallel purification of combinatorial libraries of dihydropyrido

Soluble combinatorial libraries of organic, peptidomimetic and peptide diversities
Molecular diversity provides a source of novel lead compounds for drug discovery. In particular, soluble combinatorial l

Selection of antibodies from synthetic antibody libraries
More than 2 dozen years had passed since the field of antibody engineering was established, with the first reports of ba

High-throughput Generation of Synthetic Antibodies from Highly Functional Minimalist Phage-displayed Libraries
We have previously established a minimalist approach to antibody engineering by using a phage-displayed framework to sup

Applications of random sampling to virtual screening of combinatorial libraries
We describe statistical techniques for effective evaluation of large virtual combinatorial libraries (>1010 potential co

Generation and utilization of synthetic combinatorial libraries Jutta Eichler and Richard A. Houghten

D !

Jutta Eichler Richard A. Houghten Torrey Pines Institute for Molecular Studies, 3550 Gene.ralAtornics Court, San Diego, CA 92121, USA. Tel: +1 619 455 3803 Fax: +1 619 455 3804 e-maih [email protected]

M

,NP YP

.

VP

/:i!, "..jL '~.;!

The use of combinatorial chemistry is fundamentally changing the pace and scope of basic research and drug discovery. Since the introduction of synthetic peptide libraries several years ago, combinatorial chemistry has proven to be a powerful tool for the generation of immense molecular diversities of peptides, peptidomimetics and new organic compounds. This article briefly reviews methods for the generation and application of combinatorial libraries, with particular emphasis on soluble synthetic combinatorial libraries. The utility of these molecular diversities for basic research and drug discovery has been demonstrated through the identification of numerous highly active compounds such as antigenic peptides, receptor ligands, antimicrobial compounds and enzyme inhibitors. AS first developed for peptides 1-3,the general concept of combinatorial libraries involves the generation of all possible sequence permutations for a peptide of a given length (i.e. 64 million for a hexapeptide composed of the 20 proteinogenic amino acids), in connection with a screening and selection process that enables the identification of unique, highly active peptides in the presence of millions of less active or inactive peptides. As the individual synthesis of such large numbers of compounds is unrealistic, several approaches for the synthesis of mixtures comprising up to millions of peptides have recently been shown to be practical. Using recombinant DNA techniques, large numbers of peptides can be expressed randomly in a fusion phage vector system4. This method, however,

l

174

© 1995, Elsevier Science Ltd

rvlol.l~('til./,.i,, MI.:r)I('INI£ TC)I)AY

R

Glossary

v

i e

w

s

IB

Edman degradation - Step-wise sequence analysis of peptides,

starting from the amino terminus.

e

I Coupling

+ Y-B1

N-permethylated peptides - Peptides in which all nitrogen

atoms (i.e. the amide nitrogens, as well as the nitrogens in the side chains of histidine, lysine and arginine) are methylated. Y-B1 - ~

Peptidomimetics - Compounds that mimic the structure of

peptides, but lack the typical peptide bond (-CO-NH-), so that they are less susceptible to enzymatic degradation than peptides. Strategies for the synthesis of peptidomimetics include the use of chemically modified amino acids (for example, N-substituted glycines), as well as the post-synthetic modification of peptides (for example, chemical modification of the peptide bond).

-y

~1 Deprotection

B,tB

Radio-receptor assay - A binding assay based on the competition

for binding to a receptor between a radiolabeled ligand specific for this receptor, and the compound/library to be tested.

¥ Coupling

+ Y-B 2

Streptavidin - A 60kDa protein that binds extremely tightly (K D = 10-1SM) tO biotin (a member of the vitamin B complex). This high-affinity interaction can be utilized in binding assays by labeling one partner with biotin and using streptavidin conjugated to an enzyme as detector.

Y-B2-B~ -y

remains restricted to the 20 proteinogenic amino acids as building blocks. By contrast, chemical approaches to the generation of peptide libraries~-3-~'6 allow for the incorporation of non-proteinogenic 7 and o-amino acids s-~", as well as chemically modified amino acids and other carboxylic acidsn-~4. The generally poor oral bioavailability and rapid enzymatic breakdown of L-amino acid peptides make them generally less attractive drug candidates relative to other organic (i.e. non-peptide) compounds. Therefore, the major research focus in the field of combinatorial chemistry is currently on the development of peptidomimetic and organic chemical libraries. The majority of chemical methods used to synthesize combinatorial libraries utilize the concept of solid-phase synthesis ~7, which is based on the sequential assembly of compounds from a defined set of building blocks, in which the first building block is covalently attached to a POlymeric solid support (Fig. 1). This enables the excess of reagents to be removed by simple wash and filtration processes, and avoids the laborious isolation and purification of intermediates associated with the conventional synthesis in solution, thus greatly facilitating and accelerating the synthesis process. Mixtures of compounds can be generated by coupling each building block to separate portions of the solid support resin, followed by combining and mixing all of the resin portions before dividing them again prior to the next coupling step (Fig. 2). This method is known as 'dividecouple-recombine' (DCR) (Ref. 2), 'portioning-mixing '5, and 'split synthesis '3. An alternative means of generating compound mixtures is the incorporation of mixtures of building blocks ]'tS'l~.Libraries generated by these methods are screened either immobilized (attached to the solid support) or in solution after cleavage from the solid support.

One bead-.one compound libraries The majority of immobilized libraries reported have been synthesized using a resin as the solid support and the 'split synthesis' method, generating 'one (resin) bead--one compound' libraries3. These libraries are

I Deprotection

B2-B 1 - - ~

(n-2) Couplings and deprotections

(B)n " - ~ I Cleavage

(B)n +

B

Key

B

Solid support

B

Building block

Temporary protecting group

(B), Synthesized compound

Figure 1. The strategy of solid-phasesynthesis.After attachment of the first building block to the solid support, the compound is assembled step-wise, followed by cleavage of the synthesizedcompound from the solid support.

m

175

R e v i e w s

M()I EC'LII_ARrVll'l)l('lNl: T()I)AY

Solid support resin

Division of the resin

n

M

D U U U U m U "

.-""

/./

'~ U

Coupling of individual building blocks to separate portions of the resin

Re-combination of the resin

Mixture resin

Figure 2. The divide-couple-recombine (DCR) process for the generation of compound mixtures. Abbreviations: P, solid support resin; A, M, 1-, C, F, D, V, L, Y, N, building blocks.

typically screened in a solid-phase assay, which is based on the physical selection of positively reacting resin beads, followed by structure determination of the compounds attached to those active bead. One beadone compound libraries have been used to identify antigenic peptides 3't°, compounds that bind to streptavidin 2°'2t, enzyme inhibitors z3, as well as mapping endoprotease specificity2-~. The multiple release of equimolar amounts of compounds from the resin beads also enables one bead-one compound libraries to be screened in solution2a. For libraries composed of L-amino acid peptides, the structure of the compounds on the active beads can readily be determined by Edman degradation. One bead-one compound libraries made from building blocks other than the proteinogenic amino acids, however, require the presence of oligonucleotide24-z6,peptide27''s or other chemical tags -'9 to encode the structure of the library compound on each resin bead, which can then be decoded by sequencing or other analysis [for example, gas chromatography, high performance liquid chromatography (HPLC)] of the coding tag. This requires an additional, independent chemistry for the assembly of the coding tag, so as not to interfere with the synthesis of the library. m

176

Synthetic combinatorial libraries Synthetic combinatorial libraries (SCLs) (Ref. 2) are composed of separate compound mixtures with one or more defined positions in the sequence. As these mixtures are in solution (as opposed to immobilized libraries), SCLs can be screened in a variety of bioassays, including those involving membrane-bound receptors or whole cells. Furthermore, the systematic array [i.e. the use of defined and mixture positions (see below)] of SCLs eliminates the need for any kind of coding, as the identification of the structure of the active compounds of interest is inherent to the SCL approach. The initial SCL was composed of 400 separate soluble hexapeptide mixtures, representedas O~O,XXXX, in which the first two positions (01 and 02) were individually defined with one of the 20 proteinogenic amino acids, and the remaining four positions (X) were occupied by close to equimolar mixtures of 19 of the 20 proteinogenic amino acids (cysteine excluded). Thus, each peptide mixture is composed of 194 = 130321 peptides, and the entire library represents 400 × 130 321 = 52128 400 individual peptides~-. This SCL was synthesized using the DCR method mentioned above. Briefly, the solid support resin is contained in n (n = number of building

MOI_I-£'LILAI,~ 1",41-1)It'INI2 I O I ) A ' ~

blocks for the mixture positions; 19 in the SCL described above) polypropylene mesh packets, and each building block is coupled to a separate resin packet. After removing the excess activated building blocks by washing and removal of temporary protecting groups, the resins are taken out of the packets, combined, thoroughly mixed, and then re-divided into n packets for the next coupling step. After all mixture (X) positions have been incorporated, the resin is divided into m (m = number of peptide mixtures; 400 in the SCL described above) portions (packets), and the defined positions (O) are incorporated using standard methods of multiple solid-phase synthesis. In the initial screening of this type of SCL in a given bioassay, the most active peptide mixtures are identified, followed by an iterative process of synthesis and screening during which all positions of the active sequences are successively defined. This process involves ranking, selecting and reducing the number of mixture positions (and thereby the number of peptides in each mixture), while defining one more position at each step. It is sometimes necessary to move forward with several peptide mixtures with similar activities in the initial screening. Those mixtures not pursued from the initial screening or during the iterative process, however, can be moved forward at a later date. This SCL format has also been utilized for various other peptide and peptidomimetic libraries synthesized from o- and other nonproteinogenic amino acids, as well as carboxylic acids. Chemically modified SCLs have been generated by modifying the peptide bonds (for example, N-alkylation and/or reduction) of existing peptide libraries, thus dramatically changing the physicochemical nature of the peptides, and greatly extending the range and repertoire of molecular diversity. The components of such transformed libraries, which have been termed 'libraries from libraries 't4, are stable towards enzymatic degradation as they lack the characteristic peptide bond -CO-NH-. Two parent peptide libraries (tripeptide and tetrapeptide libraries, respectively) were synthesized using 52 L-, D- and other unnatural amino acids as building blocks. The tripeptide library was N-permethylated, and the tetrapeptide library perallylated or perbenzylated, thus generating three different N-alkylated peptidomimetic libraries 3°. An alternative approach, termed a positional scanning SCL (PSSCL) (Refs 9,18), enables the identification of active individual compounds in a single assay, thus avoiding the iterative process of synthesis and screening associated with the original SCL format. A typical PS-SCL is composed of n positional SCLs (n = number of diversity positions). Accordingly, a hexapeptide PS-SCL is composed of six independent positional SCI_.s, and is represented as O~XXXXX, XO2XXXX, XXO3XXX , XXXO4XX , XXXXOsX and XXXXXO 6 (Ref. 18). As for the SCLs described above, O represents individually defined positions, whereas X represents positions occupied by mixtures of building blocks. It should be noted that each positional SCL making up a PS-SCL, while addressing a specific position, represents the same collection of individual compounds. When used in concert, the screening of the n positional libraries making up a PS-SCL provides information about the most effective building block or functionality in each position of interest for a given ligand-acceptor interaction, as well as about the relative specificity of each position (i.e. the fewer functionalities found to be effective for a given position, the higher the specificity of that position). The synthesis of all possible combinations of the most effective building blocks at each position yields a range of individual compounds, which are then tested to establish their individual activities. Alternatively, each positional library can serve as a starting

R

Ac-DVOXXX-NH

e

v

i e

w

s

2

0.08 0.06.~ Io (2_

0.040.02. 0

4XACDE

FGH

I K LMN

PQRSTVWY

Amino acid

Ac-DVPOXX-NH

2

1-

0.8

~)

0.6 0.4 0.2 m

3XA C D E F G H I K L M N

m

PQRSTVWY

Amino acid

Ac-DVPDOX-NH

0.03-

2

I

0.02O 0.01

0

2XA C D E F G H I K L M N P Q R S T V W Y Amino acid

Ac-DVPDYO-NH

0.6-

2

0.4O

I

0.2-

o

me

me

m

lXA C D E F G H I K LMNPQRSTVWY Amino acid

Rgure 3. Iterative process for DVXXXX-NH r Each graph represents one step in the process (i.e. one more defined positio0). The bars within each graph represent the reciprocal ICsovalues of the peptide mixtures defined at the 0 position with the amino acid specified on the X axis. The first bar in each graph represents the most active peptide mixture from the previous step.

m

177

R

e v

i e w

s

M { ) I . I : ( ' t I I , A R MI'I)I('INI~ T O I ) A Y

OXXXXX-NH

2

XOXXXX-NH 2

0.024 -]

XXOXXX-NH 2

0.024 -]

oo

0.035 -I

leo

o o o l 8 -I

i

o oo18 t

°.°121 0.008 t

0 0.012t ~ 0.008t

°°°41 0.=..

II,.,-,.-.-I

ADEFGH

.

-,

0.004 -1

o,,--I

I KLMNPQRSTVY

oo ii

0.025 -I

~ 0.024 < 0.015 t 0.01

,-

ADEFGH



,I.1.1..I•

o

.n .ll...m._.l

011__

I KLMNPQRSTVY

ADEFGH

I KLMNPQRSTVY

Amino acid

Amino acid

Amino acid

XXXOXX-NH 2

XXXXOX-NH 2

XXXXXO-NH 2

0.025 -1

0.012

oo, 1

0.015

ioooo

OOll

I

0.004

o.oo5 -I

o,,-- I , . . , , ADEFGH

....,,I

mm

I KLMNPQRSTVY Amino acid

0,002 n.. ,~I I I O ADEFGH

.,,

,.i

I KLMNPQRSTVY Amino acid

ooo5 0 ~

I ,__ l l l l l l . i , ADEFGH

nl,

I KLMNPQRSTVY Amino acid

Figure 4. 0pioid-receptor binding by the positional scanning hexapeptide SCL Each graph represents one of the six positional libraries. The bars within each graph represent the reciprocal ICs0values of the peptide mixtures defined at the 0 position with the amino acid specified on the X axis.

point for the iterative synthesis and screening process described above. The positional scanning format has also been used for the synthesis of a D-amino acid hexapeptide PS-SCL (Ref. 9), as well as a decapeptide PS-SCL made up of 10 positional SCLs, which each represent approximately 4 × 1012individual peptides 31. Recently, the positional scanning format has been applied to the preparation of peptidomimetic libraries. A hexapeptide PS-SCL was N-permethylated, thus generating a chemically transformed library ('libraries from libraries') of N-permethylated peptidomimeticslL A PS-SCL based on a cyclic peptide template was synthesized using 10 different carboxylic acids in addition to the 20 proteinogenic amino acids as building blocks, thus increasing the chemical diversity of this library 12. This PS-SCL comprises three positional SCLs, represented as cyclo[Lys(O1)-Lys(X)-Lys(X)-Glu]-Gly-OH,cyclo[Lys(X)-Lys(00Lys(X)-GIu]-Gly-OH and cyclo[Lys(X)-Lys(X)-Lys(O3)-Glu]-Gly-OH. The defined and mixture positions were incorporated by acylating the e-amino groups of the three lysine residues. Utilization of SCLs

Antigenic peptides Peptide SCLs of varying lengths and formats have been used to identify the antigenic determinants recognized by various monoclonal antibodies (mAbs) (reviewed in Ref. 32). The iterative synthesis and screening process described above is illustrated by the identification of the antigenic determinant recognized by monoclonal antibody 17/9 (Fig. 3), which was raised against a 36-residue fragment of the hemagglutinin of influenza virus. The antigenic determinant recognized by this antibody (-DVPDYA-) had previously been identified and characterized extensively using omission and substitution analogs of a longer antigenic peptide (Ac-YPYDVPDYASLRS-NH 2) m

178

(Ref. 33). Through the screening of an N-acetylated hexapeptide library (Ac-O,O_,XXXX-NH2, Ac-DVXXXX-NH, was found to be the most active peptide mixture (IC~, = 58~M). Twenty new peptide mixtures (Ac-DVOXXX-NH2) were synthesized and tested. The most effective inhibiting peptide mixture was Ac-DVPXXX-NH 2 (IC50 = 151~M).This process was repeated twice more in order to identify the most effective amino acids at the fourth (Ac-DVPDXX-NH 2, IC5o = 1.1 tzM) and fifth (Ac-DVPDYX-NH 2, IC5, = 40nM) positions, respectively. Upon defining the sixth and final position of the sequence, a set of 20 individual peptides (Ac-DVPDYO-NH2) was synthesized and assayed. Ac-DVPDYA-NH 2 was identified as the most active individual peptide (IC50 = 2 riM). This sequence exactly matches the antigenic determinant found in earlier studies to be recognized by mAb 17/9.

Receptor ligands The use of combinatorial libraries to identify opioid-receptor ligands has been reviewed extensively elsewhere 34. In order to verify the utility of PS-SCLs for identifying receptor ligands, a non-acetylated hexapeptide PS-SCL was screened in an N-specific opioid radio-receptor assay 18.The screening results for the six positional libraries making up this PS-SCL are shown in Fig. 4. A set of individual peptides representing all possible sequence permutations of the most effective amino acid residues at each position (i.e. Tyr in position one, Gly in position two, Phe and Gly in position three, Phe in position four, Phe, Leu, Met and Tyr in position five, and Phe, Arg and Tyr in position six), composed of 1 × 1 x 2 X 1 × 4 X 3 = 24 individual peptides, was then synthesized and tested. The first five amino acid residues of the two most active of these peptides (i.e. YGGFMY-NH 2, IC5o= 17 nM and YGGFMR-NH2, IC50 = 24 nM) exactly match the sequence of the natural opioid-receptor ligand methionine-enkephalin (YGGFM).

M()I.I:C'LJI AI~. M I ~ r H ( ' I N I - r ( ) l ) A Y

Ac-RWXXXX-NH: and Ac-RFXXXX-NH, were the most active peptide mixtures found by screening an N-acetylated hexapeptide SCL (Ac-O~OzXXXX-NH2) in the I~-opioid-receptor assay. It should be noted that prior to this study, no acetylated opioid peptides were known. The iterative process for Ac-RFXXXX-NH, resulted in the identification of a new group of potent opioid receptor antagonists, termed acetalins (Ac-RFMWMO-NH2, O = K,T,R, ICs, = 5--6riM) (Ref. 35). An N-acetylated hexapeptide SCL prepared entirely from D-amino acids (indicated in lower case)(Ac-o~o~xxxx-NH:) was also screened in the opioid receptor assay, followed by the standard iterative process, which resulted in the identification of a potent all D-amino acid opioid peptide (Ac-rfwink-NH2, ICs0 = 18 riM). This peptide was found to be a ~-receptor-specific agonist and to produce analgesia in mice when administered intracerebroventricularly (i.c.v.) or intraperitoneally s. These analgesic effects were antagonized by i.c.v, injection of the opioidreceptor antagonist naloxone, indicating that these effects were centrally mediated and consequently, that Ac-rfwink-NH., crosses the blood-brain barrier.

Antimicrobial compounds The increasing emergence of bacterial strains that are resistant to commonly used antibiotics has led to a critical need for new antimicrobial agents. SCLs represent a powerful means with which to develop new antimicrobial leads for pharmaceutical research. Several microorganisms have been used in studies involving peptide SCLs, including Gram-positive bacteria such as Staphylococo~s aureus and Staphylococcus sanguis, Gram-negative bacteria such as Escherichia coli and Pseudomonas aeruginosa, and yeast such as Candida albicans. A variety of potent antimicrobial peptides have been identified by screening L-amino acid peptide SCLs of different formats -''36. As L-amino acid peptides are prone to proteolytic degradation, which limits their pharmaceutical utility to topical or intravenous application, compounds that are resistant to proteolytic degradation are expected to be of higher therapeutical value than L-amino acid peptides. Consequently, an N-permethylated peptidomimetic PS-SCL, the components of which have been shown to be resistant to enzymatic breakdown, was screened against S. aureus ~4. On the basis of the screening data, 144 individual N-permethylated compounds were synthesized and tested, resulting in the identification of pm[LFIFFF-NH2] and pm[FFIFFF-NH2] as the compounds with the strongest antibacterial activity (IC5o= 6 ~g mlfol" both) against S. aureus. Similar activities were found against S. sanguis, whereas none of the N-permethylated compounds or mixtures were active against the Gram-negative bacteria E. coli, or yeast C. albicans. Enzyme inhibitors Enzyme inhibitors are of broad interest for biomedical research, as well as for their therapeutic potential. Inhibitors of HIV protease have been identified by screening a tetrapeptide library 6. A thrombin inhibitor was found in a non-peptide one bead--one structure library 13. Trypsin and chymotrypsin were used as model enzymes in our studies. A hexapeptide SCL was specifically designed to represent all possible reactive sites for trypsin inhibitors with the P~ position (lysine or arginine) in every position of the sequence, except the last. This library, which is composed of 10 sublibraries (represented as Ac-KOXXXX, Ac-ROXXXX, Ac-XKOXXX, Ac-XROXXX... through Ac-XXXXKO and Ac-XXXXRO), was screened in a trypsin inhibition assay, followed by an iterative synthesis and screening process. The most active trypsin inhibitor found was Ac-AKIYRP-NHz (IC50 = 46 ~M) ~7.

r

e

v i

e

w

s

Another hexapeptide SCL, represented as XXOOXX-NH 2, was screened in a chymotrypsin inhibition assay, resulting in the identification of WFLYYC-NH: as the most active chymotrypsin inhibitor. This sequence contains four aromatic amino acid residues (tryptophan,

The outstanding questions • If you have a synthetic mixture made up of, for example, more than 100000 compounds, how can you be certain that every compound expected to be present is in fact present? While the synthetic methods used (such as the divide, couple aM recombine resin method) give one a very high level of confidence to ~wect that all the compounds will be present, there are currently 11oanalytical procedures (HPLC, mass spectral analysis, capillary electrophoresis, etc.) that are capable of characterizing such comple_~ mixtures on the level of individual compounds. However, in the cases studied in which an existing compound was known to be present in a library, those compounds were found. • As the number of compounds in the mixtures making up combinatorial fibraries is often very large (i.e. 100 000), will you miss active compounds if an agonist and antagonist are present in the same mixture? This could be a problem if combinatorial libraries were screened primarily in vivo. However, combinatorial libraries are commonly screened hi vitro;for example, using radioreceptor-bbuting assays. When using this assay format, one cannot distinguish agonists from antagonists, as both bind to the receptors, and agonists and antagonists will be found. • As the number of compounds present in a library can exceed ten to hundreds of millions (far exceeding the number of compounds tested worldwide over the past 100 years), will a single such library be sufficient for all applications? One library would definitely not be appropriatefor all indications, as various types of libraries may have very different chemical characters. A simple analogy can be made with locks and keys - one can make 100 000 different keys of a particular size, but if the lock is smaller than the keys, none of the keys will open the lock. • Which research or discovery efforts will benefit from the use of combinatorial chemistry? The breadth of combinatorial chemistry means that virtually all scientific disciplines will benefit immensely from its use. Thus, basic biomedical research will be accelerated greatly by the thoroughness of the studies now possible with combinatorial technology. Medicinal chemists can rapidly increase biological activities more efficiently than was even imagined five years ago. • If combinatorial chemistry is so powerful, why haven't more drugs been identified? Combinatorial chemistry is capable of rapidly identifying hits or lead compounds. It is important to understand that the time frame required to carry out toxicology, and the FDAapproved Phase I, II and III clinical trials typically takes seven to ten years. Practicalcombinatorial chemistry is less thanfive years old. Compounds identified using the combinatorial methods are just now starting to enter toxicology or Phase I trials in many companies. A few have been moved through to Phase II studies. These numbers will dramatically increase over the next five years.

m

179

Reviews phenylalanine and two tyrosine residues), each of which may serve as the P1 position of the reactive site, as aromatic amino acids are known to be specific for the P1 position of chymotrypsin inhibitors. Trypsin and chymotrypsin inhibitors were also identified by screening a hexapeptide PS-SCL synthesized entirely from D-amino acids. The most active inhibitors found were Ac-ryrpwp-NH 2 for trypsin (IC50=62 p.M) (Ref. 9) and Ac-ygyyyr-NH2 for chymotrypsin (IC5o=36 IxM). As expected, both peptides are completely stable towards proteolyticdegradation. Conclusions The broad utility of SCLs relies upon the fact that they can be screened in virtuallyany bioassayof interest.This is illustratedby the various levels of complexity of biological systems that can be studied using SCLs, which include soluble receptors (i.e. antibodies,enzymes), membrane-bound receptors (opioid receptors), whole cells (bacteria, yeast) and directly/n vivo. The versatility of the SCL concept is also reflected by the range of libraries with different formats and compositions. Starting with 'simple' L-amino acid peptide libraries, the SCL concept has since been used to synthesize peptide libraries composed of D- and other non-proteinogenic amino acids and carboxylic acids. Conformationally defined libraries have also been prepared, which enable the custom design of complex molecules with defined secondary structures that can perform specific biological functions. The concept of 'libraries from libraries', which is based on the post-synthetic chemical modification of parent peptide or non-peptide libraries, opens the door to novel chemical diversities with great therapeutic potential. Acknowledgements, We thank Eileen Silva for her editorial assistance. The work performed in this laboratorywas hmded by HoughtenPharmaceuticalsInc., San Diego, CA, USA. References 1 Geysen, H.M., Rodda, S.J. and Mason, T.J. (1986) A pr/or/delineation ofa peptide which mimics a discontinuous antigenic determinant, Mol. Immunot. 23, 709-715 2 Houghten, R.A. eta/. (1991) Generation and use of synthetic peptide combinatorial libraries for basic research and drug discovery, Nature 354, 84-86 3 Lam, ICS. et ol. (1991) A new type of synthetic peptide library for identifying figand-binding activity, Nature 354, 82-84 4 Smith, G.P. and Scott, J.K. (1993) Libraries of peptides and proteins displayed on filamentous phage, Methods Enzymol. 217, 228-257 5 Furka, A., Sebestyen, E, Asgedom, M. and Dibo, G. (1991) General method for rapid synthesis of multicomponent peptide mixtures, InL J. Peptide Protein Res. 37, 487-493 6 Owens, R.A., Gesefichen, P.D., Houchins, B.J. and DiMarchi, R.D. (1991) The rapid identification of HIV protease inhibitors through the synthesis and screening of defined peptide mixtures, Biochent Biophys. Res. Comrawt 181,402--408 7 Blondelle, S.E., Takahashi, E., Weber, P.A. and Houghten, R.A. (1994) Identification of antimicrobial peptides using combinatorial libraries made up of unnatural amino acids, Antimicrob. Agents Chemother. 38, 2280-2286 8 Dooley, C.T. et aL (1994) An all D-amino acid opioid peptide with central analgesic activity from a combinatorial library, Science 266, 2019-2022 9 Pinilla, C. etal. (1994) Versatility of positional scanning synthetic combinatorial libraries for the identification of individual compounds, Drug D ~ Res. 33, 133-145 10 I_am, K.S. et ol. (1993) Discovery of D-amino-acid-containing ligands with Seleclide technology, Gene 137, 13-16 11 Simon, R.J. et al. (1992) Peptoids: A modular approach to drug discovery, Proc. NatlAcad. Sci. USA 89, 9367-9371 12 Eichler, J., Lucka, A.W. and Houghten, R.A. (1994) Cyclic peptide template combinatorial libraries: synthesis and identification of chymotrypsin inhibitors, Peptide Res. 7, 300-307 13 Lebl, M. et at (1995) One-bead--one-structure combinatorial libraries, Biopolymers/Peptide Sci. 37, 177-198

/

180

MOLECULAR

MEDICINE

TODAY

14 Ostrcsh, J.M. et al. (1994) 'Libraries from libraries': Chemical transformation of combinatorial libraries to extend the range and repertoire of chemical diversity, Proc. Natl Acod. Sci. USA 91, I1138-11142 15 Merrificld, R.B. (1963) Peptide synthesis. I. The synthesis of a tetrapeptide, J. Am. Chem. Soc. 85, 2149-2154 16 Geysen, H.M., Meloen, R.H. and Barteling, S.J. (1984) Use ofa peptide synthesis to probe viral antigens for epitopes to a resolution of single amino acid, Proc. Natl Acod. Sci. USA 81, 3998-4002 17 Houghten, R.A. (1985) A general method for the rapid solid-phase synthesis of large numbers of peptides: specificity of antigen-antibody interactions at the level of individual amino acids, Proc. NatlAcad. Sci. USA 82, 5131-5135 18 Eichler, J. and Houghten, R.A. (1993) Identification of substrate-analog trypsin inhibitors through the screening of synthetic peptide combinatorial libraries, Biochemistry 32, 11035-11041 19 Dooley, C.T. and Houghten, R.A. (1993) The use of positional scanning synthetic peptide combinatorial libraries for the rapid determination of opioid receptor ligands, LifeSci. 52, 1509-1517 20 Lam, K.S. and Lebl, M. (1992) Streptavidin and avidin recognize peptide ligands with different motifs, Immunomethods 1, 11-15 21 Stankova, M. et al. (1994) Application of one-bead one-structure approach to identification of nonpeptidic ligands, Drug Dev. Res. 33, 146-156 22 Meldal, M., Svendsen, J., Breddam, K. and Auzanneau, F-I. (1994) Portionmixing peptide libraries of quenched fluorogenic substrates for complete subsite mapping of endoprotease specificity, Proc. NatlAcad. Sci. USA 91, 3314-3318 23 Lebl, M. et aL (1993) Multiple release of eqnimolar amounts of peptides from a polymeric carrier using orthogonal linkage-cleavage chemistry, Int. J. Peptide Protein Res. 41,201-203 24 Brenner, S. and Lemer, R.A. (1992) Encoded combinatorial chemistry, Proc. Natl Acad. Sci. USA 89, 5381-5383 25 Nielsen, J., Brenner, S. and Janda, K.D. (1993) Synthetic methods for the implementation of encoded combinatorial chemistry, J. Ant. Chem. Soc. 115, 9812-9813 26 Needels, M.C. et al. (1993) Generation and screening of an oligonucleotideencoded synthetic peptide library, Proc. Natl Acad. ScL USA 90, 10700-10704 27 Kerr, J.M., Banville, S.C. and Zuckermann, R.N. (1993) Encoded combinatorial peptide libraries containing non-natural amino acids, J. Ant. Chem. Soc. 115, 2529-2531 28 Nikolaiev, V. et al. (1993) Peptide-encoding for structure determination of nonsequenceable polymers within libraries synthesized and tested on solid-phase supports, Peptide Res. 6, 161-170 29 Ohlmeyer, M.H.J. et al. (1993) Complex synthetic chemical libraries indexed with molecular tags, Proc. NatlAcad. ScL USA 90, 10922-10926 30 D6mer, B., Ostresh, J.M., Husar, G.M. and Houghten, R.A. (1995) Extending the range of molecular diversity through amide alkylation of peptide libraries, in Peptides 94 (Maia, H.L.S., ed.), pp. 463-464, Escom 31 Pinilla, C., Appel, J.R. and Houghten, R.A. (1994) Investigation of antigeHntibody interactions using a soluble nonsupport-bound synthetic decapeptide library composed of four trillion sequences, Biochem. J. 301,847-853 32 Pinilla, C. et al. (1995) A review of the utility of soluble peptide combinatorial libraries, Peptide Sci./Biopolym. 37, 221-240 33 Pinilla, C., Appel, J.IL and Houghten, R.A. (1993) Functional importance of amino acid residues making up peptide antigenic determinants, Mol. ImmunoL 30, 577-585 34 Dooley, C.T. and Houghten, R.A. A comparison of combinatorial library approaches for the study of opioid compounds, Perspectives in Drug Discovery and Design (in press) 35 Dooley, C.T., Chung, N.N., Schiller, P.W. and Houghten, R.A. (1993) Acetalins: new opioid receptor antagonists determined through the use of synthetic peptide combinatorial libraries, Proc. NatlAcad. Sci. USA 90, 10811-10815 36 Blondelle, S.E., Takahashi, E., Dinh, K.T. and Houghten, R.A. (1995) The antimicrobial activity of hexapeptides derived from synthetic combinatorial libraries, J. AppL Bacteriol. 78, 39-46