Combinatorial Libraries: Overview of Applications in Chemical Biology
H. Peter Nestler, Combinatorial Technologies Center, Tucson, Arizona
The sequencing of genomes gave access to the complete set of building blocks for organisms of various species. A plethora of ''-omics'' technologies has been developed to investigate the dynamic interactions of the building blocks to understand the functioning of living organisms. Synthetic molecules have proven to be powerful tools to modulate that states of biological systems, but the challenges to find suitable probes are tremendous. Combinatorial libraries allow preparing and testing diverse sets of molecules with high efficiency. In this article, we will discuss how combinatorial chemistry enables to investigate and modulate biochemical function in the quest to chart chemical and biological spaces.
The scholar disciplines of chemistry and biology have for a long time lived in separate academic universes, although they heavily depend on each other. Combinatorial chemistry started to bridge this intellectual gap by introducing the concept of diversity and selection into synthesis, which mimics biological evolution. It was triggered by the frustration of chemists trying to study biological systems with insufficient chemical matter. The recognition that chemical structures cannot be tailor-made to interact with biological space stimulated the development of methods for fast synthesis and screening of small molecules. Albeit combinatorial chemistry was recognized in its infancy as the solution to all problems of drug discovery and to studying biological space, it was soon recognized that the brute-force trial-and-error approaches would be inefficient in discovering biorelevant structures. Chemical biology strives to understand and design small molecules and their interaction with proteins. One basic pillar is the grouping of the proteome into target families, assuming that members of the same protein family share structural features. This assumption holds true especially for kinases and GPCRs but less though for proteases and ion channels, which are the four most common target families in drug discovery. Other protein families such as nuclear hormone receptors and transcription factors display even higher structural variations. However, as chemical biology is also founded on the understanding of the biological function of the protein families, information about assaying as well as privileged ligand motifs that can be transferred from one target family member to another. Although combinatorial chemistry and chemical biology may seem to be opposite concepts, in reality they are highly complementary and interacting. Using the insights gained from chemical biology to bias the design of combinatorial libraries will help to assess a chemical space suitable to probe biological systems. In this article, we will discuss the roots of combinatorial chemistry, the technologies for generating and assaying compound collections, and studies performed on biochemical systems.
Chemical Biology—One Short Definition
Chemical biology covers many aspects of the influence of chemical molecules on biological function. In a broad definition, the understanding and investigation of toxicological effects, modulation of gene and protein expression, transformation and transport of the molecules in cell and whole organisms, and the change of the metabolic pathways and patterns are topics of chemical biology. With the definition below, we put a focus on one subtopic of chemical biology pertinent to the discussions in this article.
The sequencing of the human genome (1) marked the apex in the transition of biology from an observational and descriptive activity to a hypothesis-driven science. With the information about complete sets of building blocks for cells available, methods could be devised to group proteins on a structural level and to investigate the phenomenology of organisms on a molecular and structural level. Similarities in protein structures have been investigated for a long time, which cover all levels from primary (sequence) via secondary (domain folds) to tertiary (overall three-dimensional) structures. Investigations of tertiary structures help to predict functional sites and roles for uninvestigated proteins from their primary sequence. Especially through bioinformatic analysis, it has been possible to identify homologous reaction mechanisms even within proteins with lower sequence similarity and different biochemical functionality (2, 3), as highlighted by the cases of Leukotriene A4 hydrolase and angiotensin-converting enzyme (a zinc metalloprotease), which are both inhibited by bestatin but have distinct biological roles (4). Primary structure investigations have been preferentially of interest for evolutionary analyses. These phylogenetic analyses have been crucial in defining target families, which are groups of proteins that have a similar gene and therefore protein sequence, which lays the foundation for “chemical biology.” Kinases are the prototypic target family as their active sites are structurally highly homogenous and bind the same cosubstrate, adenine triphosphate (ATP). Other gene families include G-protein coupled receptors (GPCR), ion channels and transporters, and proteases, although the structural diversity in these families is higher and they therefore group in structurally and mechanistically diverse sub-families, such as cysteine or metalloproteases.
Soon after the sequences became available, researchers discussed how many of the approximately 27,000 genes that had been assigned (1) would be “druggable,” (i.e., their associated protein products could be modulated with small molecules in a directed fashion to achieve a desired physiological effect) (5). For a small molecule to exert its biological action, it is crucial that the molecular shape of the molecule complements the cast offered by the target protein. This fact has been recognized first by Emil Fischer who phrased it as a “Key-Lock” principle (6), being unaware of the dynamic and flexible nature of protein structures. Today, we examine the interactions of two molecules more in a “Hand-Glove” fashion with strong elements of induced fits (7). We term the ensemble of available interaction shapes in the proteome the “biological space,” whereas the “chemical space” is considered the ensemble of shapes offered by small molecules. With our structural understanding constantly evolving through molecular biology and crystallography, the efforts to design matching chemical structures rationally have increased and led to successes in drug development; the HIV-protease inhibitors are one prime showcase. Rational design depends on structure-activity relationship to bias the contribution of various molecular motifs to the intermolecular interactions properly; therefore, it is powerful for the optimization and understanding of activity and selectivity on the protein target level. Rational design suffers shortcomings when we attempt to address the challenges of finding novel starting structures from scratch. High-throughput screening efforts try to tackle this challenge through engaging in a high-numbers trial-and-error game. As the screening collections reflect the target history of the respective company, they often cover narrow aspects of chemical space. This challenge is addressed by combinatorial chemistry.
Considering that most novel therapies would rely on oral administration of drugs, such molecules have to fulfill requirements to achieve suitable pharmacokinetic behavior. The most quoted and commonly used guidelines are Lipinski’s “Rule-of-5” (8) and Veber’s “rotational bonds” (9) that have been based on a statistical analysis of marketed oral drugs. Considering such boundaries, it has been estimated that about 10-15% of the human genome would be “druggable” (10). Although this number may seem low, it should be considered that only one third of these mechanisms are currently targeted, and a significant fraction of drugs, even those in development, still act through undefined molecular pathways. Furthermore, the hype around the sequencing of the genome and the assumed impact on drug discovery meanwhile has vanished, as it was recognized that biological networks are too complex and redundant to allow control through one molecular dial. Systems biology tries to address this challenge by exploring the interactions of proteins and the resulting pathways of transferring biological signals and actions. Chemical biology is the matching complement in drug discovery that tries to synergize on structural relationships of proteins to address the druggable genome efficiently (11). Furthermore, the concepts of chemical biology are transferred toward the understanding of pharmacokinetics by identifying and investigating proteins that are involved in transport through the organism and the metabolic transformation of the molecules. Models to predict unwanted effects (commonly referred to as “toxicity” or “side effects”) are developed based on our increased understanding of biological pathways and the proteins involved therein that may offer interaction sites for the molecules. As those investigations do not involve combinatorial chemistry approaches, they will not be topic of this article.
Chemical biology has reshaped all stages of the drug discovery and today is a widely used discovery paradigm in pharmaceutical industry. The focus as well as the impact of using targeted family knowledge has been on the early stages from target identification via structural understanding through lead finding efforts. The later stage of the drug-discovery process, which included the optimization of lead compounds into drug candidates, is not as amenable to technological solutions that can be provided through target family concepts as the challenges become very specific for each lead series. Still, transferring insights and understanding compound interactions with targets and other proteins help to avoid entering dead-end alleys of modification. However, the focus of this article will be on the biochemical level of chemical biology. Most chemical examples will be of peptidic nature, as most reports around the topic of this article are based on libraries derived from peptides, although this may change in the near future. The main theme of this article will be more on the general concepts and approaches than on specific molecules, but the interested reader can find a bounty of literature on specific molecules that modulate biological systems especially from a viewpoint of drug discovery.
Combinatorial Libraries—A Brief Overview
Historically, the art of chemical synthesis demanded that individual molecules are prepared efficiently with high yield and are extensively characterized. Over decades, more and more complex molecules were prepared by dissecting the target molecule at key reactive bonds, devising strategic options for the assembly, and developing novel reactions to address the synthetic needs. Although this “logic of synthesis” laid the foundation for today’s capabilities to prepare a wide variety of conceivable target structures, it did not provide the numbers of molecules necessary for and efficient investigation of the “biological space.” The resolution of this conundrum required a conceptual rethinking by chemists. Assuming that our hypotheses about evolution are correct, the natural evolution of complex organism occurred through generation of molecular diversity and selection and amplification of suitable structures. Today, we observe this process at a level of high structural complexity; genetic selection of alleles and somatic selection of antibodies are among the most prominent aspects. Thus, it was not too surprising that the concept of combinatorial chemistry was devised based on immunological challenges (12). To be successful in the game of evolution and life, four aspects have to be addressed: 1) generation of diversity, 2) compartmentalization of the individual members, 3) selection for desired properties, and 4) reproduction of selected members. The third aspect will be the topic of other articles, whereas the other aspects will be discussed here. We will not review the variety of chemistry possible for the preparation of libraries, which has tremendously developed of over the last decade and has been comprehensively reviewed on a regular basis by Dolle et al. (13).
Synthesis of combinatorial libraries
Biology handles diverse gene and protein populations by compartmentalizing the members in cells. The selected library members are amplified by growing clones of the cells that survive the selection step, which allows ultimately for the reporting of the structures for selected members through gene sequencing. The “yeast-two-hybrid” system (14) and the “phage display” technique (15) are used to study protein/protein and protein/peptide interactions in vivo, and the “SELEX” technology allows for the in vitro evolution of short oligonucleotides, which are called aptamers, that interact with small molecules and proteins (16). Such biological libraries will not be topic of this contribution.
Synthetic molecules by amplified by growing clones, and chemists who handle libraries face two major problems: the compartmentalization of the individual library members and the determination of the molecular structure of the selected species (17).
Three approaches have been devised to tackle these challenges:
1. Spatial arraying of compounds during synthesis and assaying
2. Synthesis of mixture libraries and deconvolution of active sublibraries
3. Encoding of particles carrying individual compounds to be tested
Spatial arraying has been used traditionally for the assaying of historic compound collections. As the identity of the molecule is correlated to position in the array, this approach carries the least challenge for structural assignment for active array positions. As an additional aspect to ease the implementation of arrayed libraries, arrayed libraries are commonly handled as solutions. As most biological assays are based on homogeneous test systems, providing the libraries in solution is the obvious choice. Today the preparation of arrayed libraries in solution is mostly addressed through robotic means (which basically provides high-throughput parallel syntheses), but significant approaches are available to provide arrayed libraries by synthesis at predetermined locations of solid supports. The earliest experiments used functionalized polystyrene mounted on plastic rods, which were arranged in a 96-well microtiter plate format and were used for the synthesis of the libraries and the compounds were released into the wells for their biological testing (18). Various implementations of this strategy have been developed, such as spot synthesis on cellulose membranes (19, 20) or photolithographic wafer techniques (21). However, these techniques impose limitations on the numbers of compounds that can be prepared and usually the 96-well format is used to test sublibraries and mixtures of compounds. (However, DNA and RNA libraries prepared through photolithographic technology find widespread used in genomic diagnostics.) To ameliorate the restrictions caused by the spatial addressing, Berk and Chapman (22) devised a strategy for the generation of two-dimensional spatial arrays of compounds that allows for determination of synergetic effects of residues in the same molecules, which combines the advantages of solution-based biological assays with an efficient testing of large numbers of compounds.
Several groups have reported the generation of combinatorial libraries in solution (23-26) as well as the fractionation of complex libraries by HPLC (27). Although the handling of compounds and libraries in solution makes them more amenable to the formats used for biological tests, the task of elucidating the structures of the active compounds requires complex deconvolution through sublibraries. This task can be performed by simultaneous preparation of the required sublibraries that control residues in one or two variable positions, which is called “positional scanning,” or by subsequential narrowing of the options by repeated testing and resynthesis of the most active sublibraries with additional defined positions. Both approaches are based on the hypothesis that the most active compound in a library is defined by the combination of the most active moieties in the various positions. In that respect, the deconvolution approach is conceptually related to fragment-based approaches that will be discussed later. Despite all doubts about the reliability of results that stem from the testing of complex mixtures, there have been successes to the deconvolution strategy, which have been reviewed by Houghten et al. (28). Most recently, the mixture screening and deconvolution was applied in vivo (29).
To avoid this caveat of screening mixtures and the identification of active compounds through deconvolution, approaches to compartmentalize the library on moieties of solid supports have been devised. Houghten (30) suggested the use of labeled “teabags” that contain peptide synthesis resin for the compartmentalization and tracking of the library members (30). This approach was modified by using radio chips for the labeling of the “tea bags” (31, 32), which allow the robotic sorting of the containers for synthesis and spatial arraying. Yet, the size of the bags still put limits on the diversity of such combinatorial libraries. “Split synthesis” (33, 34) lifts the restrictions for the complexity of libraries by creating “one-bead-one-structure libraries” that display the individual library members separated on beads of a solid support. Successive cycles of distributing and mixing portions of the support lead to a combinatorial increase of the diversity of structures contained in the library. As each bead of the resin reacts with only one set of reagents per synthesis cycle, each bead carries one individual structure, which is determined by the sequence of reactions that occurred during the synthesis. Library members that interact with the targets can be isolated by selecting beads detected in assays performed on the beads. However, the application was limited to peptides and oligonucleotides because of the analytical means of determining the structure of the molecules on the beads. Molecular encoding was the breakthrough that opened the way for the application of the whole repertoire of synthetic chemistry for generating combinatorial libraries by solving the problem of structure determination from a single bead: Easily detectable molecular tags are attached to the resin beads as they proceed through the split synthesis during the library construction, which thereby records and reports the reaction history of each individual bead (35) (corresponding to the genetic coding of proteins in cells). After screening the library, these molecular tags are cleaved from each selected bead and then analyzed to report the structures of the library members on these beads. Several encoding methods are currently in use, which include microsequenceable oligonucleotide (36) or oligopeptide strands (37, 38), as well as nonoligomeric schemes that employ small molecules that can be analyzed by gas chromatography (35, 39), HPLC (40), or mass spectrometry (41).
Fragment screenings and dynamic combinatorial libraries
In recent years, two variations of combinatorial chemistry approaches have been explored that do not present preassembled molecules to the biological target (42). Both variations make use of the direct detection of interactions instead of using biochemical inhibition as a readout. One approach, “fragment screening,” uses fragments/building blocks of the molecules and relies on the detection of low-affinity interactions. Several detection approaches have been described, based on NMR (43), cocrystallization (44), or surface plasmon resonance (45, 46). The second approach, which is called “dynamic combinatorial chemistry,” relies on forming the complete molecules from fragment in presence of the target molecule and enriching molecules with high affinity in the equilibrium reaction (47). This approach has been applied to generate artificial receptors for peptides (48, 49) as well as inhibitors for enzymes with detection by cocrystallization (50).
Mapping Enzymatic Activities
One key aspect of applications of combinatorial libraries in chemical biology is the mapping of substrate specificities for enzymes. Although the early studies that use peptide libraries were focused on studying antibody epitopes (12), the substrate mapping for enzymes that transform peptides and the development of molecular probes found widespread interest.
The earliest applications focused on proteases. The major challenge that had to be tackled was the generation of a readout signal. Although most protease assays rely on generating a chromophoric of fluorophoric molecule on cleavage from the carbxy-terminal amino acid (e.g., nitrophenyl, coumarins), this method was not applicable to the screening of peptide libraries as the cleavage does not necessarily occur at the carboxy-terminus of the sequence. The problem was resolved through the introduction of fluorophor-quencher pairs at the ends of the peptide sequence: As long as the peptide is intact, the quencher at the distal end of the peptide suppresses the fluorescence. After cleavage, the quencher diffuses away from the support, and the support becomes fluorescent because of the fluorophor at the proximal end of the peptide (51, 52). Another challenge that had to be addressed for the screening of proteases was the penetration of the supports and its handling in aqueous environment during the assays. The classic supports used in peptide synthesis show a limited amount of swelling and penetration by proteins in aqueous environment, which makes only peptides on the surface accessible for study [this behavior was used to develop a molecular encoding approach by discriminating surface and interior positions of the bead (53)]. Grafting of hydrophilic chains of polyethylene-glycol and preparation of supports based on cross-linked acrylamide yielded supports that could be handled in organic solvents for synthesis and water for assaying (54). With all these different approaches in hand, substrates for all protease subfamilies have been identified from combinatorial libraries (55-59), but the application of the data has been extended the most for cathepsins (60). Based on the sequence information obtained, Greenbaum et al. (61) designed inhibitors by incorporating the epoxide motif from E-64, which is a well-known unspecific and irreversible cysteine protease inhibitor. These inhibitors allowed pulling down various cathepsins from cell extracts or labeling them with some selectively using fluorescently labeled inhibitors, which thus fingerprinted cathepsin activity in cells. The surprising observation in these studies was that specificity could be instilled into the inhibitors starting from the unspecific E-64 by attaching it to the various substrate sequences. The latest step forward was the demonstration that these substrate based inhibitors can be used to monitor cathepsin activity in living mice. The inhibitors were generated using an acyloxymethyl ketone-inhibitor motif, which acts as a suicide substrate. The two sides of the substrate were labeled with infrared fluorescent dyes and a quencher and were injected into the tail veins of mice xenografted with tumors that express high levels of cathepsins B and L. The inhibitor became fluorescent on cleavage, and the labeled cathepsins allowed a distinct imaging of the tumors (62). The use of covalent inhibitors offers the advantage of localizing the fluorescent signal to the locus of generation as the diffusion of the signal is limited by the slow diffusion of the cathepsins. Recently, a new approach based on substrates was described, which will lead to the ability for fast real-time monitoring of proteases in vivo (63).
The screening of combinatorial libraries for kinases faced another challenge. Whereas kinases allow the straightforward generation of a readout signal on peptides through the transfer of radioactive (32P)-ATP, the association of the radioactive signal to individual members of highly diverse libraries posed a problem. Autoradiography using photographic film was used to visualize groups of beads immobilized in agarose gels, and individual beads were isolated through repetitive steps of isolation of beads from the gels and dilution, which made the process work intensive (64, 65). The process was dramatically simplified by embedding the beads directly into the photographic emulsion, which allowed for the microscopic identification of individual radioactive beads (66, 67). The other approaches for library preparation and screening like positional scanning (68) or arrayed synthesis (69) were successfully applied as well. For the main purpose of the kinase chemical biology of routine profiling and substrate identification in a high-throughput mode, the array synthesis seems to be established because of the ease of readout. Recently, a novel approach lifted the need of using radioactive ATP for the screening for tyrosine-kinase substrates: The library of fluorescein-labeled peptides is encoded with PNA-tags and incubated with a kinase. The peptides are then incubated with a DNA-microarray that complements the PNA strands of the individual peptides in solution. After such attachment of the peptides to the array, the array is treated with a phosphotyrosine-recognizing antibody sandwich, which carries the fluorescent dye Cy3. The microarray is then read out at two wavelengths to allow the detection of fluorescein and Cy3. The ratiometric determination allows for a sensitive calibration against PNA/DNA complexation differences and the identification of kinase substrates (70). Unlike in the protease field, the straightforward translation of kinase substrate information into inhibitors by using non-cleavable phosphonate mimics (71) and their use for biological investigation has been limited due to the substrate size and the challenges for biological application. Most kinase inhibitors target the ATP-binding pocket and the associated hinge region, and their use has been found mostly in drug discovery with detailed pharmacophore insights (72). After the sequencing of the human genome (1), the use of substrate sequence information has been successful for the bioinformatic identification of potential protein substrates and establishment of pathway information. The information on these studies is scattered in many individual papers and databases. This use of searching protein sequence databases for protein substrates distinguishes the kinase research from the protease field, as the specificity information obtained for protease is usually limited to tetrapeptides, which will yield many irrelevant substrate candidates in proteomic searches. As the substrate information obtained for kinases usually covers sequences larger than octapeptides, the number of substrate candidates identified through bioinformatics is usually limited and can be investigated in a more detailed fashion.
Phosphatases and other enzymes
Conceptually, the screening for the substrates of proteases and kinases is rather straightforward, as both enzyme classes allow for detection by either fluorescence or radioactivity of their altered substrates. Other enzyme classes are not that amenable and require more elaborate assay designs. Cheung et al. (73) described an elegant scheme to identify substrates for leukocyte antigen receptor (LAR) phosphatase. Although it may be conceptually envisioned to detect the substrates for phosphatases by reduction of radioactivity after hydrolysis of a radiolabeled phosphate, the screening of combinatorial libraries for signal differences proves to be unfeasible, because of the varied concentration of the individual members. Thus, all readouts for the screening of combinatorial libraries should be laid out to yield the appearance of an effect. The screen for phosphatase substrates capitalized on the stability of phosphor-tyrosine peptides against proteolysis by chymotrypsin, whereas their dephosphorylated counterparts are very susceptible to cleavage. An on-bead combinatorial N-acetyl peptide library that features phosphotyrosine in the center of each peptide sequence was incubated with LAR. Subsequently, the library was treated with chymotrypsin, which led to the cleavage of peptides that had been dephosphorylated. The resulting free aminotermini were derivatized with amine-reactive chromophores or fluorophores to visualize the individual beads. After sequencing a coding peptide strand on the beads, six substrates were identified. Each of these substrates shows faster dephosphorylation than one native substrate, which is the D988-G998 epitope of the EGF-receptor (73). Several enzyme classes have been targeted using the positional scanning approach, such as farnesyl transferase (74) and α-glucosidase (75); the latter establishes that the screening of peptide libraries is not limited to enzymes that transform peptides.
Studying Protein-Protein Interactions/Whole-Cell Systems
The use of combinatorial libraries for the study of protein-protein interactions is currently focused on the diagnostics and the fingerprinting of interaction surfaces. The starting point for combinatorial libraries was the investigation by Lam et al. (12) of antibody epitopes with a one-bead-one-compound pentapeptide library. An anti-β-endorphin antibody recognized only 6 peptide sequences out of 2.5 million possible sequences. The intellectually intriguing aspect of the study was the use of an antibody as a biological probe; antibodies are the prime example for somatic combinatorial chemistry in biological systems. The same study also demonstrated that combinatorial chemistry can identify ligands to proteins that are unrelated to the natural ligands. The tripeptide HPQ was the privileged binder to streptavidin, and competition experiments clearly showed that it binds at the same site as biotin.
From the screening for antibody epitopes it was a short step to address another larger class of cell-surface molecules, which is called GPCR. Using a 5000-membered dimer and trimer-peptoid (N-alkylated glycine polymers) library, new ligands to adrenergic and opoid receptors were discovered. The peptoids discovered as ligands to the adrenergic receptor carry three aromatic residues and thus mimic an ephedrine polymer; the peptoids discovered for the opoid receptor resemble Met-enkephalin. Sequences for both receptors are very lipophilic, which points to a challenge that affects many library screenings. However, both types of sequences are nanomolar antagonists for their respective receptors (76). In an interesting twist, Spatola et al. (77) designed a cyclic pentapeptide library in a way that only one round of screening was necessary to deconvolute the most active pentapeptide sequence. The library contained 82,944 peptides in 48 sublibraries with various defined positions. One round of screening revealed the cyclo(Pro-D-Val-Leu-D-Trp-D-Asp) sequence as the most active antagonist of endothelin I. This sequence already had been established earlier from a microbial peptide (77). Although the concept of self-deconvolution is very intriguing, the effort in design and limitations on diversity prevented further widespread application of this approach. Combinatorial libraries have also allowed to look into the intracellular side of GPCRs. Martin et al. (78) studied the interaction of the G-protein a unit with Rhodopsin using a 11-mer peptide library displayed on phage. After several rounds of binding to rhodopsin-load membranes and amplification of the phage, several sequences similar to a sequence in the carboxy-terminus of the G-protein a unit could be identified. After resynthesis, these sequences bound to rhodopsin and helped to stabilize the meta-II rhodopsin state. Yet, phage-display libraries are not crucial to the study of intracellular protein-protein interactions. One of the most common motifs in protein-interaction domains is the Src homology 2 (SH2). This domain is an integral part in kinase-signaling binding to the phosphorylated peptides. Thus, peptides that bind to the SH2-domain could be valuable tools for dissecting the cellular-signaling cascades. Sweeney et al. (79) used a hexapeptide library while carrying a phosphoty- rosine in the center. The library beads were incubated with various biotinylated SH2-domains and the SH2-binding beads were visualized using streptavidin conjugated alkaline phosphatase. Sequences binding selectively to the SHP-1, SHP-2, and SHIP-2 SH2-domains could be identified. Furthermore, the resynthesized peptides could be used to pull down the full-length proteins from cellular extracts. The peptide sequence information was used to search for binding partners to the respective SH2-domains in genomic databases, which identified known proteins and a variety of hitherto unknown binding candidates.
For routine diagnostics in which the number of probes to be tested is smaller than in a de novo interaction screen, arrayed libraries show the higher potential because of the simplified handling. Today, these libraries are produced by spotting the individual probes on microscope slides that can be read out in fluorescence microscope readers after assaying. Reddy et al. (80) reported a study that fingerprinted three different proteins against an array of 7200 octamer peptoids. Each protein showed distinct interaction patterns, although the peptoid library was naive against the proteins. Using scatter plots of the interaction intensities, clear dependencies of detection approaches (direct labeling of protein vs. antibody detection) in the read outs could be found. However, the quality of the assays and the ease of handling makes the approach using peptoids promising as a complementation of DNA and protein arrays used in diagnostics today.
One of the most recent examples of combinatorial library applications comes back to the theme of in vivo imaging discussed earlier in the protease section. A one-bead-one-compound library of urea-capped octapeptides built around a central LDV-motif was screened against JURKAT cells. Beads were selected based on observed cell adhesion under a microscope, and the compounds on these identified beads was characterized by EDMAN sequencing of encoding peptide tags in the interior of the beads, which yields only to different sequences. Because of the biasing with the central LDV motif, these compounds showed high affinity and specificity to a4 Pi-integrin. After complexation with the NIRF dye Alexa680-streptavidin conjugate, the peptides were used for the imaging of various tumors in nude mice. Tumors that expressed the α4β1-integrin became fluorescent, whereas tumors not expressing the integrin did not provide imaging signals (81).
Most recently, the utility of combinatorial libraries in proteomics was demonstrated by enriching platelet proteins. Using columns loaded with a combinatorial peptide library, low abundant proteins from platelet extracts could be enriched through capturing and eluting them. The diversity of the library allowed capturing proteins without knowledge of their identity and binding partners. Using two different libraries, 175 new proteins could be identified and sequenced (82).
Modulating Biological Systems—Artificial Receptors
Traditionally, the approach to interrupting biological processes aims to find small molecules that fit into a binding site offered by one binding partner. This approach has been and still is very successful for drug discovery. Combinatorial chemistry opened the way to another approach for the modulation of biological processes. For the first time, it became possible to develop receptor molecules for peptides without detailed knowledge of the peptide conformation and without intricate design for the rigidification of the receptor and fitting of the intermolecular interactions. Using cheno(12-deoxy)cholic acid as a scaffold to orient combinatorial depeptide chains, Boyce et al. (83) identified receptors that would bind and distinguish Leu- and Met-enkaphalin, which are two pentapeptides. Thus, it became conceivable that one could prevent an enzymatic transformation by trapping the substrate instead of blocking the active site of the enzyme. The viability of the approach was shown on the example of the farnesylation of the H-Ras protein. H-Ras is a central signal transduction protein and in its mutated forms one of the prime players in multiple cancers. For its functioning, it has to be localized to the cell membrane, which is achieved through posttranslational farnesylation of the carboxy-terminal CaaX-box sequence. Farnesyltransferase inhibitors have been under preclinical and clinical investigation for a long time, but they suffer severe unwanted side effects, as a plethora of other proteins require the famesylation of their respective CaaX-box sequences. Although these sequences are similar enough to be transformed by the same enzyme, each protein carries its characteristic CaaX-sequence. Using a library of receptors based on four identical combinatorial tetrapeptides that grew from a trilysine scaffold, receptors that recognize the CVLS-sequence of H-Ras could be identified in an on-bead screening assay. After resynthesis, these receptors suppressed the farnesylation of H-Ras, while still permitting the farnesylation of K-Ras (CVIM) and Laminin (CAIM) (84). One receptor prevented the localization fluorescent GFP-H-Ras to the cell membrane after microinjection into xenopus oocytes, which demonstrated the use of such receptors and molecular biology tools. In another demonstration of this approach of “epitope protection,” Zhang and Kodadek (85) showed that a linear 15-mer peptide identified from phage-display library can bind to the cleavage site of interleukin-1β. This cleavage site is proteolyzed by Caspase-1, which is a key initiator of intracellular processes. As one of the key switches for inflammation and apoptosis, Caspase-1 has a broad substrate specificity, which has made the therapeutic development of Caspase-1 inhibitors very challenging. The identified peptide could suppress the cleavage of interleukin-1β, while at the same time having no impact on the autocatalytic activation of Caspase-1, one of the key factors in Caspase-1 mediated signaling (85). Yet, not only peptide epitopes can be recognized with peptide derived receptors. Using the same combinatorial receptor library as in the H-Ras study (see above, (84)), Fukase et al. discovered receptors for liposaccharides (86). Escherichia coli Lipid A, which is a strong endotoxin, is responsible for inflammatory and lethal septic events during bacterial infections. Thus, molecules that absorb the endotoxin could be of high therapeutic value. In an on-bead affinity screening, tritium-labeled Lipid A analogs were equilibrated with the library, and beads that bind the radioactive liposacchararides were visualized by microradiography. Decoding of the binding beads and resynthesis of the respective receptors yielded “molecular forceps” that can bind and distinguish the various Lipid A analogs, whereas another E. coli liposaccharide, O111:B4 LPS, was not recognized by the forceps (86).
The field of using artificial receptors to camouflage enzyme substrates or to absorb toxic biologicals is still in its infancy, and no biological activity of artificial receptors beyond the cellular level has been reported to date. Yet the ability to generate and identify receptor molecules that show binding properties similar to antibodies opens intriguing alleys to chemical tools for the study of molecular biology. Once we learn how to handle the tools provided by combinatorial libraries, we will have access to a tool chest for the investigation of cells and organisms at the molecular level, the key aspect of chemical biology.
The use of combinatorial libraries in chemical biology is plentiful. This article focused on the plethora of possibilities and challenges in a fast developing field, which highlighted some achievements. With the evolution of synthetic repertoire and analytical technologies, high-throughput screening starts to move from structurally related libraries as discussed here to libraries assembled from historic collections. With the increased diversity of structures, we will observe many more molecular tools for chemical biology. As many efforts are focused on drug discovery, one should also expect novel therapeutic approaches based on our increased understanding of chemical and biological spaces and networks.
1. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, et al. The sequence of the human genome. Science 2001; 291:1304-1351.
2. Wangikar PP, Tendulkar AV, Ramya S, Mali DN, Sarawagi S. Functional sites in protein families uncovered via an objective and automated graph theoretic approach. J. Mol. Biol. 2003; 326:955-978.
3. Koch MA, Breinbauer R, Waldmann H. Protein structure similarity as guiding principle for combinatorial library design. Biol. Chem. 2003; 384:1265-1272.
4. Orning L, Krivi G, Fitzpatrick FA. Leukotriene A4 hydrolase. Inhibition by bestatin and intrinsic aminopeptidase activity establish its functional resemblance to metallohydrolase enzymes. J. Biol. Chem. 1991; 266:1375-1378.
5. Drews J. Drug discovery: a historical perspective. Science 2000; 287:1960-1963.
6. Fischer E. Effekt der Zuckerkonfiguration auf die Enzymwirkung. Berichte 1894; 27:2984-2993.
7. Koshland DE Jr. The lock-and-key principle and the induced-fit theory. Angew. Chem. Int. Ed. 1994; 33:2475-2478.
8. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 1997; 23:3-25.
9. Veber DF, Johnson SR, Cheng HY, Smith BR, Ward KW, Kopple KD. Molecular properties that influence the oral bioavailability of drug candidates. J. Med. Chem. 2002; 45:2615-2623.
10. Hopkins AL, Groom CR. The druggable genome. Nat. Rev. Drug Disc. 2002; 1:727-730.
11. Wess G, Urmann M, Sickenberger B. Medicinal chemistry: challenges and opportunities. Angew. Chem. Int. Ed. 2001; 40:3341-3350.
12. Lam KS, Salmon SE, Hersh EM, Hruby VJ, Kazmierski WM, Knapp RJ. A new type of synthetic peptide library for identifying ligand-binding activity. Nature 1991; 354:82-86.
13. Dolle RE, Le Bourdonnec B, Goodman AJ, Morales GA, Salvino JM, Zhang W. Comprehensive survey of chemical libraries for drug discovery and chemical biology: 2006. J. Comb. Chem. 2007.
14. Parrish JR, Gulyas KD, Finley RL Jr. Yeast two-hybrid contributions to interactome mapping. Curr. Opin. Biotechnol. 2006; 17:387-393.
15. Gazarian K. Drug discovery and design via high throughput screening of combinatorial phage-display protein-peptide libraries. Front. Drug Des. Disc. 2005; 1:29-67.
16. Bunka DHJ, Stockley PG. Aptamers come of age - at last. Nat. Rev. Microbiol. 2006; 4:588-596.
17. Wentworth P Jr., Janda KD. Generating and analyzing combinatorial chemistry libraries. Curr. Opin. Biotechnol. 1998; 9:109-115.
18. Geysen HM, Meloen RH, Barteling SJ. Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid. Proc. Nat. Acad. Sci. 1984; 81:3998-4002.
19. Frank R. Strategies and techniques in simultaneous solid phase synthesis based on the segmentation of membrane type supports. Bioorg. Med. Chem. Lett. 1993; 3:425-430.
20. Kramer A, Schuster A, Reineke U, Malin R, Volkmer-Engert R, Landgraf C, Schneider-Mergener J. Combinatorial cellulose-bound peptide libraries: screening tools for the identification of peptides that bind ligands with predefined specificity. Methods 1994; 6:388-395.
21. Fodor SPA, Read JL, Pirrung MC, Stryer L, Lu AT, Solas D. Light-directed, spatially addressable parallel chemical synthesis. Science 1991; 251:767-773.
22. Berk SC, Chapman KT. Spatially arrayed mixture (SpAM) technology: synthesis of two-dimensionally indexed orthogonal combinatorial libraries. Bioorg. Med. Chem. Lett. 1997; 7:837-842.
23. Berg T, Vandersteen AM, Janda KD. High-throughput synthesis and direct screening for the discovery of novel hydrolytic metal complexes. Bioorg. Med. Chem. Lett. 1998; 8:1221-1224.
24. Carell T, Wintner EA, Bashir-Hashemi A, Rebek J Jr. A novel procedure for the synthesis of libraries containing small organic molecules. Angew. Chem. Int. Ed. 1994; 33:2059-2061.
25. Carell T, Wintner EA, Rebek J Jr. A solution-phase screening procedure for the isolation of active compounds from a library of molecules. Angew. Chem. Int. Ed. 1994; 33:2061-2064.
26. Han H, Wolfe MM, Brenner S, Janda KD; Liquid-phase combinatorial synthesis. Proc. Nat. Acad. Sci. 1995; 92:6419-6423.
27. Griffey RH, An H, Cummins LL, Gaus HJ, Haly B, Herrmann R, Cook PD. Rapid deconvolution of combinatorial libraries using HPLC fractionation. Tetrahedron 1998; 54:4067-4076.
28. Houghten RA, Pinilla C, Appel JR, Blondelle SE, Dooley CT, Eichler J, Nefzi A, Ostresh JM. Mixture-based synthetic combinatorial libraries. J. Med. Chem. 1999; 42:3743-3778.
29. Houghten RA, Pinilla C, Giulianotti MA, Appel JR, Dooley CT, Nefzi A, Ostresh JM, Yu Y, Maggiora GM, Medina-Franco JL, Brunner D, Schneider J. Strategies for the use of mixture-based synthetic combinatorial libraries: scaffold ranking, direct testing in vivo, and enhanced deconvolution by computational methods. J. Comb. Chem. 2007; 10:3-19.
30. Houghten RA, Dooley CT. The use of synthetic peptide combinatorial libraries for the determination of peptide ligands in radio-receptor assays: opoid peptides. Bioorg. Med. Chem. Lett. 1993; 3:405-412.
31. Moran EJ, Sarshar S, Cargill JF, Shahbaz MM, Lio A, Mjalli AMM, Armstrong RW. Radio frequency tag encoded combinatorial library method for the discovery of tripeptide-substituted cinnamic acid inhibitors of the protein tyrosine phosphatase PTP1B. J. Am. Chem. Soc. 1995; 117:10787-10788.
32. Nicolaou KC, Xiao XY, Parandoosh Z, Senyei A, Nova MP. Radiofrequency encoded combinatorial chemistry. Angew. Chem. Int. Ed. 1995; 34:2289-2291.
33. Lam KS, Hruby VJ, Lebl M, Knapp RJ, Kazmierski WM, Hersh EM, Salmon SE. The chemical synthesis of large random peptide libraries and their use for the discovery of ligands for macromolecular acceptors. Bioorg. Med. Chem. Lett. 1993; 3:419-424.
34. Furka A, Sebestyen F, Asgedom M Dibo G. General Method for Rapid Synthesis of Multicomponent Peptide Mixtures. Int. J. Pept. Prot. Res. 1991; 37:487-493.
35. Ohlmeyer MHJ, Swanson RN, Dillard LW, Reader JC, Asouline G, Kobayashi R, Wigler MH, Still WC. Complex synthetic chemical libraries indexed with molecular tags. Proc. Nat. Acad. Sci. 1993; 90:10922-10926.
36. Needles MC, Jones DG, Tate EH, Heinkel GL, Kochersberger LM, Dower WJ, Barrett RW, Gallop MA. Generation and screening of an oligonucleotide-encoded synthetic peptide library. Proc. Nat. Acad. Sci. 1993; 90:10700-10704.
37. Nikolaiev V, Stierandova A, Krchnak V, Seligmann B, Lam KS, Salmon SE, Lebl M. Peptide-encoding for structure determination of nonsequenceable polymers within libraries synthesized and tested on solid-phase supports. Pept. Res. 1993; 6:161-170.
38. Kerr JM, Banville SC, Zuckermann RN. Encoded combinatorial peptide libraries containing non-natural amino acids. J. Am. Chem. Soc. 1993; 115:2529-2531.
39. Nestler HP, Bartlett PA, Still WC. A general method for molecular tagging of encoded combinatorial chemistry libraries. J. Org. Chem. 1994; 59:4723-4724.
40. Maclean D, Schullek JR, Murphy MM, Ni ZJ, Gordon EM, Gallop MA. Encoded combinatorial chemistry: synthesis and screening of a library of highly functionalized pyrrolidines. Proc. Nat. Acad. Sci. U.S.A. 1997; 94:2805-2810.
41. Geysen HM, Wagner CD, Bodnar WM, Markworth CJ, Parke GJ, Schoenen FJ, Wagner DS, Kinder DS. Isotope or mass encoding of combinatorial libraries. Chem. Biol. 1996; 3:679-688.
42. Nestler HP. Combinatorial chemistry and fragment screening - two unlike siblings? Curr. Drug Disc. Technol. 2005; 2:1-12.
43. Olejniczak ET, Hajduk PJ, Marcotte PA, Nettesheim DG, Meadows RP, Edalji R, Holzman TF, Fesik SW. Stromelysin inhibitors designed from weakly bound fragments: effects of linking and cooperativity. J. Am. Chem. Soc. 1997; 119:5828-5832.
44. Hartshorn MJ, Murray CW, Cleasby A, Frederickson M, Tickle IJ, Jhoti H. Fragment-based lead discovery using X-ray crystallography. J. Med. Chem. 2005; 48:403-413.
45. Dickopf S, Frank M, Junker HD, Maier S, Metz G, Ottleben H, Rau H, Schellhaas N, Schmidt K, Sekul R, et al. Custom chemical microarray production and affinity fingerprinting for the S1 pocket of factor Vila. Anal. Biochem. 2004; 335:50-57.
46. Metz G, Ottleben H, Vetter D. Small molecule screening on chemical microarrays. Meth. Princ. Med. Chem. 2003; 19:213-236.
47. Ladame S. Dynamic combinatorial chemistry: on the road to fulfilling the promise. Org. Biomol. Chem. 2008; 6:219-226.
48. Hioki H, Still WC. Chemical evolution: a model System that selects and amplifies a receptor for the tripeptide D-Pro-L-Val-D-Val. J. Org. Chem. 1998; 63:904-905.
49. Corbett AD, Cheeseman JD, Kazlauskas RJ, Gleason JL. Pseudodynamic combinatorial libraries: a receptor-assisted approach for drug discovery. Angew. Chem. Int. Ed. 2004; 43:2432-2436.
50. Congreve MS, Davis DJ, Devine L, Granata C, O’Reilly M, Wyatt PG, Jhoti H. Detection of ligands from a dynamic combinatorial library by X-ray crystallography. Angew. Chem. Int. Ed. 2003; 42:4479-4482.
51. Meldal M, Breddam K. Anthranilamide and nitrotyrosine as a donor-acceptor pair in internally quenched fluorescent substrates for endopeptidases: multicolumn peptide synthesis of enzyme substrates for subtilisin carlsberg and pepsin. Anal. Biochem. 1991; 95:141-147.
52. Meldal M, Svendsen I, Juliano L, Juliano MA, Del Nery E, Scharfstein J. Inhibition of cruzipain visualized in a fluorescence quenched solid-phase inhibitor library assay. D-Amino acid inhibitors for cruzipain, cathepsin B, and cathepsin L. J. Pept. Sci. 1998; 4:83-91.
53. Vagner J, Barany G, Lam KS, Krchnak V, Sepetov NF, Ostrem JA, Strop P, Lebl M. Enzyme-mediated spatial segregation on individual polymeric support beads: application to generation and screening of encoded combinatorial libraries. Proc. Nat. Acad. Sci. 1996; 93:8194-8199.
54. Meldal M. PEGA: a flow stable polyethylene glycol dimethyl acrylamide copolymer for solid phase synthesis. Tetrahedron Lett. 1992; 33:3077-3080.
55. Murphy MM, Schullek JR, Gordon EM, Gallop MA. Combinatorial organic synthesis of highly functionalized pyrrolidines: identification of a potent angiotensin converting enzyme inhibitor from a mercaptoacyl proline library. J. Am. Chem. Soc. 1995; 117:7029-7030.
56. Eichler J, Houghten RA. Identification of substrate-analog trypsin inhibitors through the screening of synthetic peptide combinatorial libraries. Biochem. 1993; 32:11035-11041.
57. Roussel P, Bradley M, Kane P, Bailey C, Arnold R, Cross A. Inhibition of the tissue factor/factor Vila complex-lead optimization using combinatorial chemistry. Tetrahedron 1999; 55:6129-6230.
58. Al-Obeidi F, Ostrem JA. Factor Xa inhibitors by classical and combinatorial chemistry. Drug Disc. Today 1998; 3:223-231.
59. Blackburn C, Pingali A, Kehoe T, Herman LW, Wang H, Kates SA. Libraries of angiotensin converting enzyme inhibitors: solid-phase synthesis and affinity selection. Bioorg. Med. Chem. Lett. 1997; 7:823-826.
60. Wood WJL, Huang L, Ellman JA. Synthesis of a diverse library of mechanism-based cysteine protease inhibitors. J. Comb. Chem. 2003; 5:869-880.
61. Greenbaum DC, Arnold WD, Lu F, Hayrapetian L, Baruch A, Krumrine J, Toba S, Chehade K, Bromme D, Kuntz ID, Bogyo M. Small molecule affinity fingerprinting a tool for enzyme family subclassification, target identification, and inhibitor design. Chem. Biol. 2002; 9:1085-1094.
62. Blum G, Degenfeld Gv, Merchant MJ, Blau HM, Bogyo M. Noninvasive optical imaging of cysteine protease activity using fluorescently quenched activity-based probes. Nat. Chem. Biol. 2007; 3:668-677.
63. Watzke A, Kosec G, Kindermann M, Jeske V, Nestler HP, Turk V, Turk B, Wendt KU. Selective activity-based probes for cysteine cathepsins. Angew. Chem. Int. Ed. 2008; 47:406-409.
64. Wu J, Ma QN, Lam KS. Identifying substrate motifs of protein kinases by a random library approach. Biochem. 1994; 33:14825-14833.
65. Lou Q, Leftwich ME, Lam KS. Identification of GIYWHHY as a novel peptide substrate for human p60c-src protein tyrosine kinase. Bioorg. Med. Chem. 1996; 4:677-682.
66. Nestler HP, Wennemers H, Sherlock R, Dong DLY. Microautoradiographic identification of receptor-ligand interactions in bead-supported combinatorial libraries. Bioorg. Med. Chem. Lett. 1996; 6:1327-1331.
67. Chan PM, Nestler HP, Miller WT. Investigating the substrate specificity of the her-2/Neu kinase using peptide libraries. Cancer Lett. 2000; 160:159-169.
68. Songyang Z, Blechner S, Hoagland N, Hoekstra MF, Piwnica- Worms H, Cantley LC. Use of an oriented peptide library to determine the optimal substrates of protein kinases. Curr. Biol. 1994; 4:973-982.
69. Tegge W, Frank R, Hofmann F, Dostmann WRG. Determination of cyclic nucleotide-dependent protein kinase substrate specificity by the use of peptide libraries on cellulose paper. Biochem. 1995; 34:10569-10577.
70. Pouchain D, Mochon JJ, Bialy L, Bradley M. A 10,000 member PNA-encoded peptide library for profiling tyrosine kinases. ACS Chem. Biol. 2007; 2:810-818.
71. Burke TR Jr., Lee K. Phosphotyrosyl mimetics in the development of signal transduction inhibitors. Acc. Chem. Res. 2003; 36:426-433.
72. McGregor MJ. A pharmacophore map of small molecule protein kinase inhibitors. J. Chem. Inf. Model. 2007; 47:2374-2382.
73. Cheung YW, Abell C, Balasubramanian S. A combinatorial approach to identifying protein tyrosine phosphatase substrates from a phosphotyrosine peptide library. J. Am. Chem. Soc. 1997; 119:9568-9569.
74. Wallace A, Koblan KS, Hamilton K, Marquis-Omer DJ, Miller PJ, Mosser SD, Omer CA, Schaber MD, Cortese R, Oliff AI, Gibbs JB, Pessi A. selection of potent inhibitors of farnesyl-protein transferase from a Synthetic tetrapeptide combinatorial library. J. Biol. Chem. 1996; 271:31306-31311.
75. Eichler J, Lucka AW, Pinilla C, Houghten RA. Novel a-glucosidase inhibitors identified using multiple cyclic peptide combinatorial libraries. Mol. Diversity 1995; 1:233-240.
76. Zuckermann RN, Martin EJ, Spellmeyer DC, Stauber GB, Shoemaker KR, Kerr JM, Figliozzi GM, Goff DA, Siani MA, Simon RJ, et al. Discovery of nanomolar ligands for 7-transmembrane G-protein-coupled receptors from a diverse (N-Substituted)glycine peptoid library. J. Med. Chem. 1994; 37:2678-2685.
77. Spatola AF, Crozet Y, de Wit D, Yanagisawa M. Rediscovering an endothelin antagonist (BQ-123): a self-deconvoluting cyclic pentapeptide library. J. Med. Chem. 1996; 39:3842-3846.
78. Martin EL, Rens-Domiano S, Schatz PJ, Hamm HE. Potent peptide analogues of a G protein receptor-binding region obtained with a combinatorial library. J. Biol. Chem. 1996; 271:361-366.
79. Sweeney MC, Wavreille AS, Park J, Butchar JP, Tridandapani S, Pei D. Decoding protein-protein interactions through combinatorial chemistry: sequence specificity of SHP-1, SHP-2, and SHIP SH2 domains. Biochemistry 2005; 44:14932-14947.
80. Reddy MM, Kodadek T. Protein “fingerprinting” in complex mixtures with peptoid microarrays. Proc. Nat. Acad. Sci. 2005; 102:12672-12677.
81. Peng L, Liu R, Marik J, Wang X, Takada Y, Lam KS. Combinatorial chemistry identifies high-affinity peptidomimetics against a4b1 integrin for in vivo tumor imaging. Nat. Chem. Biol. 2006; 2:381-389.
82. Guerrier L, Claverol S, Fortis F, Rinalducci S, Timperio AM, Antonioli P, Jandrot-Perrus M, Boschetti E, Righetti PG. Exploring the platelet proteome via combinatorial, hexapeptide ligand libraries. J. Proteome Res. 2007; 6:4290-4303.
83. Boyce R, Li G, Nestler HP, Suenaga T, Still WC. Peptidosteroidal receptors for opoid peptides. sequence-selective binding using a synthetic receptor library. J. Am. Chem. Soc. 1994; 116:7955- 7956.
84. Dong DLY, Liu R, Sherlock R, Wigler MH, Nestler HP. Molecular forceps from combinatorial libraries prevent the farnesylation of ras by binding to its carboxy-terminus. Chem. Biol. 1999; 6:133-141.
85. Zhang Z, Ly T, Kodadek T. An inhibitor of sequence-specific proteolysis that targets the substrate rather than the enzyme. Chem. Biol. 2001; 8:391-397.
86. Fukase K, Wakao M, Izumi M, Ueno A, Oikawa M, Suda Y, Kusumoto S, Nestler HP, Sherlock R, Liu R. Identification of molecular forceps that recognize immunostimulating glycoconju- gate lipid A from encoded combinatorial libraries. In Proc. International Electronic Conferences on Synthetic Organic Chemistry, 5th, 6th, 2001. pp. 374-379.
Schreiber SL, Kapoor TM, Wess G. Chemical Biology: From Small Molecules to Systems Biology and Drug Design 1st edition. 2007. Wiley-VCH: Weinheim, Germany.
Chemical Libraries: Screening for Biologically Active Small Molecules
High Throughput Screening (HTS) Techniques: Overview of
Proteins, Chemistry and Chemical Reactivity of Synthetic Peptides and Proteins to Elucidate Biological Function