Combinatorial Libraries: Life's Tinker Toys

Courtesy of UC Davis Medical Center COMBI-KING: Kit Lam brings his expertise in combinatorial chemistry, an innovative technique for creating millions of new chemical compounds in just days, to the UC Davis Cancer Center. It's the new mix and match. In the computer industry, programmers refer to modularity--the ability to shuffle different sections of computer code to create new software. In the life sciences, the modules are peptides or molecular fragments that can be combined to yield thousands of possible reactants from a given set of compounds. The result is combinatorial chemistry, aka combi-chem. "Combinatorial chemistry was only adopted broadly in industry about 10 years ago," says John Schwab, program director of the NIH's National Institute of General Medical Sciences (NIGMS), and it caused an "almost overnight change" in the way companies discover leads and develop drugs, becoming today's "predominant paradigm." The process starts with coupling "a mixture of 20 amino acids to another mixture of 20 amino acids to produce 400 dipeptides, which are then reacted with a mixture of 20 additional amino acids" to form a combinatorial library of 8,000 tripeptides, wrote Stephen R. Wilson in his introductory chapter of Combinatorial Chemistry.1 The yields of those libraries can then be increased exponentially with length, which means there's essentially no limit, save for reasonable space requirements, on the number of potential compounds, says Arno F. Spatola, professor of chemistry and director of the Institute for Molecular Diversity and Drug Design at the University of Louisville, Ky. Spatola says that some researchers are making mixtures that yield 200 billion separate compounds per vial, each of which must be characterized to be of any value. "It's been said it's like throwing 10,000 tennis balls against a wall to see if one sticks," says Spatola. Still, most combinatorial libraries are significantly smaller: well under 100,000 compounds. Libraries have been developed of antibodies, peptides, proteins, benzodiazepines, oligonucleotides, organic heterocyclics, steroids, organo-metallics and, perhaps, other compounds. Such collections can yield drug candidates. They can "be used to perturb biological systems, and therefore you can find a small molecule to turn proteins or genes on or off and figure out what it's doing," says Dennis Curran, Distinguished Service Professor and Bayer Professor of chemistry, University of Pittsburgh. In the pharmaceutical industry, libraries are used either to discover potential leads or to optimize existing leads. The former requires huge libraries, looking for as many options as possible in the hope of finding a few potential leads. Optimization, in turn, results in focused libraries based around a specific functional group. For example, Curran says, "You have a hit, but it's not a good drug candidate because it's toxic or not orally available," so libraries are built similar to that hit, like variations on a theme. Although compounds developed by combi-chem methods have yet to be commercialized, according to Schwab, because of the time from lead identification to market, virtually all pharmaceutical companies have drugs in the pipeline that were identified as lead compounds or that were optimized using combinatorial methods. THE TECHNIQUES: TEA-BAGGING IT Combinatorial chemistry has come a long way from the early days when the first combinatorial libraries were based, somewhat illogically, on peptides, which tend to be degraded in the stomach long before they can be delivered to cells. Their choice as a starting point, however, was based on early work by Bruce Merrifield, then at Rockefeller University in New York, who had written extensively about peptides in the process of developing the technique of solid-phase synthesis,2 the foundation of combinatorial chemistry. There are two basic ways to develop combinatorial libraries today: single-compound multiple parallel synthesis, and Richard Houghton's "tea bag" concept.3 Houghten's method puts many different peptides attached to substrates into "a plastic bead so all the wash, neutralization, and deprotection steps could be carried out in a single reaction vessel and resin packets could be treated separately only when an additional protected amino acid needed to be added," he wrote in 1999.4 That method, however, is not widely used today for mixtures because, says Spatola, many perceive "tea bag" results to be a sort of mixed bag of wide variations. Spatola, however, still believes in such mixtures. In drug development, he says, the basic challenge is to find tight binding sites for drug targets and thereby minimize size effects. The problem is that, even if the best fragment is found for each receptor site, the resulting compound--"a blend of the best individual fragments"--may not be active, Spatola says. Spatola is working to solve that issue by building head-to-tail macrocyclic peptide libraries for forming ring structures. The binding site usually is found on only one face of a cyclic compound, he explains, while the rest of the compound is irrelevant as far as binding properties are concerned. So, to minimize the chance of interference, the other sides of the macrocyclic structure can be filler: unnecessary moieties that make a ring but have no functional effect on the binding capacity of the structure. Courtesy of University of Pittsburgh ACADEMIC EXCELLENCE: At the University of Pittsburgh, a newly anointed Center of Excellence in Chemical Methodologies and Library Development, chemist Peter Wipf and doctoral student Jelena Janjic look at a Bohdan miniblock for chemical synthesis. GOING IT ALONE Most combinatorial chemists minimize the uncertainties of mixture-based combinatorial synthesis by using single-compound multiple parallel synthesis. In 1997, Kit S. Lam, professor of medicine and chief of the Division of Hematology/Oncology in the Department of Internal Medicine, at the University of California, Davis, developed the one bead/one compound method. The method "synthesized hundreds of thousands to millions of spatially separable, random-compound bead libraries in a few days, with each bead displaying only one compound," says Lam. "These compound beads can be concurrently screened against specific receptors, and positive beads can be isolated for structure determination." Recent improvements to that method5 include new encoding technology, which eliminates interference by the coding tag used to identify compounds from each other in a mix during screening, "and thus facilitates structural determination of the small molecule compound," he says. Lam prepares bifunctional beads with two layers from commercially available polystyrene/polyethylene glycol beads. The compounds to be tested are on the outer layer where they are exposed to target proteins, and the coding tags remain inside the bead. As a result, there is no interference from the coding tag during screening. "Decoding," Lam says, "can be accomplished with a protein sequencer or mass spectrometer. Because identical or similar chemistries are used in the synthesis of the library compound and the coding tag, many encoding steps can be eliminated, and analysis of the coding tag enables one to estimate the purity of the library compounds on each bead." THE ITCHY AND SCRATCHY SHOW When combinatorial chemists are asked which of the two methods they use to develop libraries, they generally respond in terms of choosing between solution-phase and solid-phase chemistry rather than between mixtures-based synthesis or parallel synthesis. The relative advantages are clear-cut. Solid-phase chemistry produces more compounds in smaller quantities and is geared for high-throughput chemistry. "Synthesis (solid-phase) is more time-consuming," Harbeson says. Solution-phase chemistry yields greater quantities of fewer numbers of compounds, but they're easier to purify. "The true challenge is to find effective ways to deconvolute the libraries to identify the few relevant members," says Stefan Lutz, assistant professor of chemistry at Emory University in Atlanta. DNA shuffling is one common deconvolution method. It offers rapid evolution by random fragmentation and reassembly, but "a lot of proteins have very little in common in terms of sequence," he says, thus limiting the utility of this method. An alternative approach, developed by Lutz and colleagues including Emory's Mark Ostermeier, combines DNA shuffling and incremental truncation (dubbed "itchy"6) to create hybrid enzymes in a process dubbed "scratchy."7 The set of fusions created is shuffled to augment the number of crossovers. "There's lots of room for optimization," says Lutz, particularly in protocol development. His immediate project is to develop a vector system to overcome reading frame issues by filtering out "messed-up" coding sequences, he says. The process, basically, involves cloning a gene of interest into a fusion protein. "We transfer that into E. coli and grow it. If the E. coli survive, the gene is expressed correctly," Lutz explains. Courtesy of Abbott Laboratories IMPROVING COMBI-CHEM: Ethan Hoff, a research technician in Abbott Laboratories' High-Throughput Organic Synthesis lab, conducts an automated synthesis. FINE TUNING Abbott Laboratories, in Abbott Park, Ill., has been involved with combinatorial chemistry for about a decade, according to divisional vice president of advanced technologies James B. Summers. "Our first approach was to develop large libraries of 100,000 to one million compounds." Those libraries, however, lacked structural diversity. The consequent "hit rate"--the number of potential drug leads gleaned from the library--was quite low, perhaps 5%. Rather than abandoning combinatorial chemistry, Abbott improved the technology. "We abandoned mix-and-split techniques to perform parallel syntheses." The results, he says, are "not enormous collections, but dozens or hundreds of compounds made on a theme." That theme is tightly focused upon the needs of existing drug-discovery programs to allow combi-chem to replace handcrafted compounds. "The approach is successful," Summers says. "A number of drug-development programs were advanced by finding key modifications to advance lead candidates." That process is being tweaked as Abbott adjusts its combi-chem approach to high-throughput methodologies. Chemists deliver a core molecule to the high-throughput lab, and researchers there devise ways of decorating that molecule. In a few days, they can produce 96 fully characterized, purified compounds. Initially, "many believed it was impossible to devise a wide variety of structurally different molecules in this way," Summers says. Abbott proved them wrong by evaluating 20 of the most commonly synthesized compounds and reactions to produce a few sets of conditions used generally to produce active molecules. "The result," Summers says, "is smaller libraries based on specific chemistries." In the first year, productivity for lab groups using this approach increased 50-fold compared to those producing handcrafted compounds. "A large number of patent applications are out that include these compounds," Summers says. At Vertex Pharmaceuticals in Cambridge, Mass., combinatorial libraries begin as virtual libraries. "We then pick which compounds should be synthesized based upon product properties, solubility, structure, their fit into active sites, etc.," says Scott Harbeson, senior research fellow at Vertex. The compounds are docked to their targets on the computer to determine the best fits. As a result, Harbeson says, "We have a much smaller set and use parallel synthesis to make discreet, individual compounds." Library size, he says, typically ranges from a few hundred to about 1,000 compounds. WINNING WAYS The NIH has taken notice of combi-chem. Last October, it awarded a $9.6 million grant to the University of Pittsburgh to advance the methodology used to create combinatorial libraries, making Pitt the second Center of Excellence in Chemical Methodologies and Library Development in the United States, following Boston University. "As many as four more grants are planned for the next two years," according to the NIH's Schwab. In those labs that want to win combi-chem grants from the NIGMS, "biologists and chemists need to talk," says Schwab. Grantees also must establish core facilities to make libraries, refine methodologies, and engage in outreach to the biological community. The grants, Schwab emphasizes, are not for library development, per se, but for method development to support the continued advance of combinatorial chemistry. "Most organic chemistry is done on pure compounds in specific conditions. With massively parallel processes [such as those required by combinatorial chemistry] you can't tailor conditions as easily for each step in the synthesis," Schwab explains. "There is a vast array of reactions that work for pure compounds, but relatively few that work dependably and generally for combinatorial chemistry. Industry, therefore, is essentially making the same libraries over and over," and is missing large areas of potential diversity. "We're very anxious to provide a greater array of tools for combi-chem," says Schwab. "The more tools, the more diversity and higher-quality libraries chemists can make." Proponents are sold. "Combinatorial chemistry opens up new areas of protein engineering, such as for sequences previously not accessible because of the limits of DNA shuffling," says Lutz. And while it "hasn't yet led to the expected flood of new drugs," says Spatola, "it's an exciting technique that's adding new dimensions to chemistry." Gail Dutton (gaildutton@gi-wireless.net) is a freelance writer in Mayne Island, British Columbia. References 1. S.R. Wilson, "Introduction to combinatorial libraries: Concepts and terms," In: Combinatorial Chemistry: Synthesis and Application, S.R. Wilson, A.W. Czarnik, eds., New York: John Wiley & Sons, 1997, p. 2-3. 2. R.B. Merrifield, "Solid-phase peptide synthesis. I. The synthesis of a tetrapeptide," J Am Chem Soc, 85:2149-54, 1963. 3. R.A. Houghten, "Generation and use of synthetic peptide combinatorial libraries for basic research and drug discovery," Nature, 354:84-6, 1991. 4. M. Lebl, "Parallel personal comments on 'classical' papers in combinatorial chemistry," J Comb Chem, 1:3-24, 1999. 5. R. Lui et al., "A novel peptide-based encoding system for 'one-bead one-compound' peptidomimetic and small molecule combinatorial libraries," J Am Chem Soc, 124:7678-80, 2002. 6. G.L. Moore et al., "Predicting crossover generation in DNA shuffling," Proc Natl Acad Sci, 98:3226-31, 2001. 7. S. Lutz et al., "Creating multiple-crossover DNA libraries independent of sequence identity," Proc Natl Acad Sci, 98:11248-53, 2001. function sendData() { document.frm.pathName.value = location.pathname; result = false if (document.frm.score[0].checked) result = true; if (document.frm.score[1].checked) result = true; if (document.frm.score[2].checked) result = true; if (document.frm.score[3].checked) result = true; if (document.frm.score[4].checked) result = true; if (!result) alert("Please select") return result; } .myradio{background-color : #94bee5} Please indicate on a 1 - 5 scale how strongly you would recommend this article to your colleagues? Not recommended 1 2 3 4 5 Highly recommended Please register your vote

Combinatorial Libraries: Life's Tinker Toys

Interested in reading more?

Become a Member of

Courtesy of UC Davis Medical Center
COMBI-KING: Kit Lam brings his expertise in combinatorial chemistry, an innovative technique for creating millions of new chemical compounds in just days, to the UC Davis Cancer Center.