August 27
, 2001
Volume 79, Number 35
CENEAR 79 35 pp. 49-58
ISSN 0009-2347
[Previous Story] [Next Story]

Progress in dynamic combinatorial chemistry, polymer-supported reagents, microwave heating, functional studies, and deconvolution advance the field


MANY AT A TIME Program head Andreas Marzinik (front to back) and lab specialists Raphael Gattlen and Urs Rindisbacher of Novartis Pharma AG, Basel, Switzerland, pipette coupling reagent into 96-well reaction blocks.

"The word 'combinatorial' probably means different things to different people. But it's now commonly accepted as a generic term for a variety of different technologies. Each of these now has its own momentum and appears to be separately developing at breakneck speed." So said lab head Terry Hart of the Novartis Institute for Medical Sciences, London, at the beginning of a recent conference on combinatorial chemistry--the high-throughput synthesis and screening of chemical substances to identify agents with useful properties.

The conference, "Combinatorial Approaches to Chemistry and Biology III," was organized by the Royal Society of Chemistry and held last month at Churchill College, in Cambridge, England. The purpose of the meeting--the third in a biennial series--was to bring people together to discuss new ideas and new opportunities in this rapidly changing field.

The meeting proceedings suggested that combinatorial chemistry, which has been gestating and growing for the better part of two decades now, is beginning to mature. There are those who believe the field has been overhyped and oversold. Nevertheless, a presentation on the setup and procedures of a sophisticated combinatorial research unit at a major pharmaceutical company demonstrated that combinatorial chemistry has already become a standard part of the repertoire used in the search for new drugs.

The meeting also showed that combinatorial technologies continue to evolve and develop. Dynamic combinatorial chemistry--a technique that exploits equilibrium processes in libraries--is now being used successfully to identify enzyme inhibitors and novel receptors. The use of polymer-supported reagents and scavengers is of growing interest as a possible solution to some of the drawbacks of conventional solid-phase synthetic techniques.

Microwave heating is becoming more popular as a means to reduce reaction times and reagent requirements in combinatorial syntheses. Combichem techniques are being used as basic research tools to determine the function of enzymes--a key goal in the field of proteomics--and to identify new enzyme inhibitors. And deconvolution methods--techniques used to identify active compounds in libraries--continue to be refined.

So "combinatorial" may mean different things to different people. But to many researchers, and most of those at the Cambridge meeting, it means a way to discover promising compounds by a route that's decidedly faster, better, and cheaper.

THE DEGREE TO WHICH combinatorial chemistry has become fully integrated into the drug discovery process, as opposed to being an optional add-on that has yet to prove itself, is exemplified by the sophisticated "compound factory" that has been created at GlaxoSmithKline, in Harlow, England. Researchers there have designed a flexible combinatorial chemistry environment with a wide range of automated instruments for library synthesis, analysis, and purification, and they have developed a powerful informatics system to handle the huge volume of combinatorial data generated at the center. At the Cambridge conference, David Hunter of the Discovery Research Chemistry unit at Harlow gave attendees an unusually detailed view of the combinatorial state of the art at a major pharmaceutical company.

Maintaining profitability in a highly competitive and tightly regulated environment necessitates that drug companies generate drug leads efficiently, minimize the failure rate of compounds in clinical trials by identifying strong candidates, and move drugs into the marketing pipeline quickly. Combinatorial chemistry is helping GlaxoSmithKline meet these goals, Hunter says. In the past couple of years, the company has become significantly more dependent on lead compounds generated by the combinatorial unit and less dependent on leads from the corporate compound library--the venerable mainstay of drug discovery in big pharma.

An early emphasis in GlaxoSmithKline's combinatorial group on the synthesis of very large libraries as mixtures of compounds has matured into a more focused approach on the use of solid- and solution-phase techniques to synthesize smaller libraries of well-characterized individual compounds. Automated synthesizers from Zymark, Hopkinton, Mass., and Bohdan, Vernon Hills, Ill., and the FlexChem system from Robbins Scientific, Sunnyvale, Calif., are used for most of the group's solution-phase chemistry. Synthesizers from Advanced ChemTech, Louisville, Ky., are used for training purposes at the company.

COMBI LAB Combinatorial researchers at GlaxoSmithKline's Discovery Research Chemistry unit, in Harlow, England, work at a Tecan Genesis workup station (front) and a Myriad Core System robotic synthesizer.
MULTIPLEXED Micromass MUX inlet permits effluents from as many as eight liquid chromatographs to be directed sequentially into an electrospray mass spectrometer for combinatorial library analysis.
THE RESEARCHERS also make extensive use of both the Myriad Core System from Mettler-Toledo Myriad, Melbourn, England--a parallel solution- and solid-phase system for automated synthesis in 192 reaction vessels--and Myriad personal synthesizers, which are similar but have only 24 reaction vessels. The unique cap design of the Myriad reactors makes these instruments particularly useful for syntheses involving corrosive, air-sensitive, and moisture-sensitive reagents and reactive intermediates, Hunter says.

Hunter says that encoding technology from the Irori unit of Discovery Partners International, San Diego, Calif., is ideal for creating large libraries by split-and-mix synthesis--an approach in which pools of compounds are iteratively mixed, reacted, and divided up to generate large numbers of diverse products very rapidly. In the Irori system, small radio-frequency tags or two-dimensional optical bar codes are placed in or on microreactors, which are then used to synthesize compounds. The tags or bar codes can then be used to track compounds subjected to multiple split-and-mix synthesis steps. The technology combines the efficiency advantages of split-and-mix synthesis with the ability to recover and identify individual products. The methodology is used at GlaxoSmithKline to prepare large arrays of diverse compounds as well as focused libraries designed to interact with specific molecular targets.

A recent change at the company is an increased use of microwave heating for library synthesis. GlaxoSmithKline researchers are currently using the Smith Synthesizer from Personal Chemistry, Uppsala, Sweden, which permits precise control of reaction temperature and pressure and has a typical throughput of more than 10 reactions per hour.

All library compounds produced at the Harlow center are analyzed for purity by liquid chromatography-mass spectrometry (LC/MS), using the LCT with MUX technology from Micromass, Manchester, England. This instrument connects a time-of-flight mass analyzer with as many as eight LC columns--permitting an eightfold increase in throughput compared with conventional LC/MS. Company researchers also analyze the purity of compounds in focused libraries and representative compounds in large libraries by nuclear magnetic resonance (NMR) spectroscopy using the BEST-NMR system from Bruker, Rheinstetten, Germany--an automated-flow NMR instrument that makes it possible to analyze samples directly from microtiter plates. Synthesis products are weighed utilizing an automated weighing station, the Bohdan Balance Automator.

After purity analysis, the results are imported into a custom-designed computer system called RADICAL (the Registration, Analysis & Design Interface for Combinatorial & Array Libraries). RADICAL categorizes products into three groups: those that have passed quality control, those that require further purification, and those that are beyond hope and must be discarded. RADICAL also handles data regarding library design, reagent management, library synthesis, purification, quality control, and product registration in the corporate compound database, and it's used to communicate essential information about compounds to the high-throughput screening group.

THERE'S A GROWING interest in avoiding the need for product purification by generating pure compounds in the first place, Hunter notes. This can be done, for example, by carrying out combinatorial syntheses with solid support-bound reagents, which make it possible to recover pure individual compounds by simple filtration and removal of solvent.

But for compounds generated in other ways, purification is often necessary. Initial cleanup can involve liquid-liquid extraction (with the Allex system from Mettler-Toledo Myriad) or scavenging of impurities (by ion-exchange-based extraction or covalent bond formation with solid-phase-supported reagents). Subsequent purification is carried out by automated flash chromatography--using the FlashMaster II from Jones Chromatography, Hengoed, Wales, or the Quad3 from Biotage, Charlottesville, Va.--and by high-throughput, preparative high-performance LC (HPLC) using Biotage Parallex purification systems. The FlashMaster II system can purify 10 compounds sequentially, the Quad3 is a parallel 12-column instrument, and the Parallex is a fully automated system with four HPLC columns that run in parallel and can handle 200 to 300 samples per day, Hunter says.

Postpurification processing is then carried out with a series of instruments from Biotage; Micromass; Tecan, Männedorf, Switzerland; Zymark; Genevac, Suffolk, England; and Bohdan. In this step, purified compounds are put in bar-coded vials and transferred to the high-throughput screening group for testing.

Combinatorial chemistry and high-throughput screening have become an integral part of daily operations in medicinal chemistry groups at GlaxoSmithKline in the past few years, Hunter says. Has the changeover from the traditional one-compound-at-a-time approach been worth it? As a result of the changeover, the productivity of biochemists and medicinal chemists at the company has been enhanced, Hunter says, while the quality of compounds going into the drug development pipeline has been maintained.

DYNAMIC COMBICHEM In a preliminary study of binding specificity, Thomas and coworkers found that iminium products of two aldehydes (R5CHO and R8CHO) in a library of eight aldehydes bound with highest affinity to the enzyme -galactosidase.

"HIGH-THROUGHPUT synthesis has had a substantial impact on the progress of research programs, with numerous examples of instant SARs, reduced cycle time, and more potent leads," Hunter says. SARs (structure-activity relationships), which measure the influence of compound structural changes on biological activity, can help researchers identify more highly active agents.

GlaxoSmithKline's combinatorial program, Hunter says, has also led to "increased serendipity"--the generation of compounds of unexpectedly high potency and the discovery in libraries of leads for unexpected targets. Most importantly, he adds, more than 50% of drug leads identified at the company in 2000 originated from high-throughput chemistry--more than twice as many as in 1999.

One attendee asked Hunter when he thought people would see case studies of compounds from high-throughput chemistry that had actually gone into human clinical trials. "It's taking longer than anticipated," Hunter replied. "But eventually the majority of drugs going into clinical trials will originate from high-throughput chemistry."

Another presentation at the Cambridge conference covered dynamic combinatorial chemistry (DCC), a technique in which a molecular target compound is introduced into a mixture of library compounds that can interconvert with each other chemically. Some of the library constituents bind to the target selectively and are thus removed from the pool of interconverting species. This causes the equilibrium of the library solution to shift, favoring the production of species that bind to the target and minimizing the concentration of poorly binding library compounds.

"Addition of an enzyme, receptor, or template should perturb the composition of the library to favor the formation or stabilization of the library member for which it has highest binding affinity," explains Neil R. Thomas, a lecturer at the school of chemistry at the University of Nottingham, in England. DCC can thus be used to determine which compounds in a library bind selectively to a target.

DCC has been practiced up to now primarily by researchers such as chemistry professor Jeremy K. M. Sanders of Cambridge University; Alexey V. Eliseev, director of chemistry at Therascope AG, Heidelberg, Germany (who is on leave from the State University of New York, Buffalo); and chemistry professor Jean-Marie Lehn of Université Louis Pasteur, Strasbourg, France, one of the founders of Therascope.

For example, Lehn and coworkers showed a few years ago that a series of aldehydes and amines that formed 12 iminium products could be mixed with carbonic anhydrase to find compounds that bound most tightly to the enzyme. And Eliseev tells C&EN that "Therascope is making really good progress in finding new enzyme inhibitors by DCC--one of the first targets being neuraminidase, a key enzyme involved in influenza virus activity."

But interest in DCC seems to be spreading. At the Cambridge conference, Thomas reported some preliminary DCC results obtained recently by his group. Thomas, postdocs Gilles Quéléver and Andrew Walsh, and coworkers prepared a dynamic system of a type developed earlier by Lehn and Sanders--a mixture of aldehydes that undergoes an enzyme-catalyzed reaction with 4-hydroxypiperidine to form iminiums. The reaction was carried out, and the researchers "fixed" the iminium products--that is, removed them chemically from the equilibrium system.

Their analysis of the results suggested that iminium products of two of the aldehydes bound with the highest affinity to the enzyme, -galactosidase. These aldehydes could therefore have potential as galactosidase inhibitors. Thomas and coworkers are currently carrying out additional experiments to further the study.

Among notable developments in DCC, Thomas points particularly to a recent paper by Lehn and coworkers in which a dynamic library was used to identify acetylcholinesterase inhibitors [ChemBioChem, 2, 438 (2001)]. In the study, Lehn and coworkers identified a potent bis-pyridinium inhibitor of acetylcholinesterase, synthesized it, and characterized the influence of some of its structural features on inhibitory potency. In addition to the potential applicability of dynamic combinatorial chemistry to drug discovery, "I like to insist also on its potential in materials science, such as in the case of supramolecular polymers," Lehn tells C&EN.

Another recent DCC achievement was the first isolation of a new receptor using this approach [Angew. Chem. Int. Ed., 40, 423 (2001) and C&EN, Jan. 22, page 14]. Sanders and coworkers identified, from a library of cyclic peptides, a tripeptide receptor to which acetylcholine binds with high affinity.

"The approach that Lehn and Therascope are taking is precisely complementary to ours," Sanders says. "They are trying to find the optimum small ligand to fit the active site of an enzyme or natural receptor, while we are trying to find large optimum synthetic receptors for interesting templates."

Sanders and coworkers have now carried their receptor discovery efforts one step further--by amplifying and isolating a new receptor for lithium ions from a DCC library of 10 different macrocycles [J. Am. Chem. Soc., published Aug. 18 ASAP,]. The cyclic peptidelike receptor they isolated has a structure that would have been unpredictable beforehand. The study therefore led "to the efficient synthesis and isolation of a complex molecule that would be difficult to prepare by conventional chemistry," Sanders says--a demonstration that DCC may sometimes have advantages over conventional combinatorial techniques.

SOLID-PHASE organic synthesis (SPOS) has long been an integral component of the synthetic repertoire employed by combinatorial chemists. In SPOS, solid support-bound substrates are elaborated synthetically by using an excess of reagents to drive reactions to completion. Desired products can then be isolated easily by simple filtration and removed from the support material. SPOS has many advantages but also many important drawbacks: It can be difficult to adapt conventional solution-phase chemistry to a solid-phase format, some solid-phase reactions are significantly slower than they are when run in solution, and the progress of solid-phase reactions can be difficult to monitor.

Professor of organic chemistry Steven V. Ley of Cambridge University suggests that the use of polymer-supported reagents and scavengers may be more practical than conventional SPOS for many combinatorial syntheses--and for syntheses of individual organic compounds and natural products as well [J. Chem. Soc., Perkin Trans. 1, 2000, 3815]. His group has been developing strategies in which reagents, catalysts, and cleanup agents attached to solid supports facilitate reactions of substrates in solution.

HIS GROUP'S PROJECT on polymer-supported reagents and scavengers "was born out of sheer frustration," Ley says. "Anybody who's tried to generate libraries on immobilized supports knows how difficult it can be to optimize these reaction processes." A major goal of the project is "to generate chemistry that's practically reasonable--chemistry that requires only the process of filtration and evaporation to generate pure product," even in multistep reactions.

For example, a solid-supported reagent can be used to drive to completion a coupling reaction between two reactive substrates, one of which is present in excess. A support-bound scavenger can then be used to remove the excess reagent, yielding pure product. Or a support-bound capture reagent can be used to remove the coupling product directly, which is then cleaved from the support--a technique called catch and release.

With immobilized reagent systems, Ley says, "we can do long linear syntheses in which each step is easily optimized, because we're running these reactions in solution. But the real beauty for us is that when you have an immobilized reagent you can sometimes combine one, two, or three of these in a single reaction vessel. These reagents, by being immobilized, won't react with one another. They'll simply react with the components and therefore allow you to do one-pot syntheses. For example, we could have an oxidizing agent and a reducing agent--which would in solution be incompatible--running at the same time in a linear route." This ability to run multiple reactions in single reaction vessels with mutually incompatible reagents "really adds power to synthesis," he says.

Convergent routes also are possible with polymer-supported reagents. Convergent synthesis "is how we make big molecules," Ley says. By using different polymer-supported reaction channels, "we can make elaborate building blocks that can then be recombined by yet another immobilized-reagent system to give us more and more complex materials." In a recent convergent synthesis, Ley and a coworker constructed sildenafil (Viagra) from n-butanal using several polymer-supported reagents [Bioorg. Med. Chem. Lett., 10, 1983 (2000)].

Polymer-supported reagents can also be used in split routes, where simple starting materials are converted to advanced building blocks or monomer sets. "The real advantage is you can start to rapidly build proprietary building block sets that could be very useful in all sorts of other synthesis programs," Ley says.

Further work is needed to make polymer-supported reagent synthesis even more useful. "There's no doubt we need more reagents--but reagents designed for immobilized supports, not reagents taken from known solution-phase chemistry and just applied to a bead," Ley says. "We've got to have a lot better polymers and support materials. Beads are awful to use in synthesis. They pick up water, they pick up static electricity, they're very difficult to weigh out, and they're not always robust and stable when you try to stir them."

Other needed improvements include better scavengers; improved aids to automation, such as reagent chips; and novel synthetic approaches such as the use of stacked reactors and flow reactors. The tremendous potential of polymer-supported reagent chemistry for complex multistep combinatorial syntheses and for conventional syntheses of individual organic compounds justifies further efforts in all these areas, Ley says.

POLYMER-SUPPORTED reagents can also be used in conjunction with microwave heating, which has become an increasingly popular tool for combinatorial synthesis in the past few years. At the conference, lab head Christopher T. Brain of the Novartis Institute for Medical Sciences discussed combining polymer-supported reagents with microwave heating to design rapid and "clean" combinatorial syntheses.

"The use of polymer-supported reagents effectively reduces workup and purification to a simple filtration, and microwave promotion reduces the reaction time for these transformations, typically from hours to minutes," Brain says. "Microwave technology has now reached a stage where it is appealing and convenient for many synthetic and medicinal chemists as a frontline rather than a remedial methodology." He and his colleagues began working on microwave chemistry about three years ago, in collaboration with Personal Chemistry.

Microwave heating not only increases reaction rates but also reduces reagent requirements in some cases--for example, by permitting a 1:1 reagent ratio in a reaction where a large excess of one reagent is normally needed to drive the chemistry. "This has very profound implications for doing reactions back to back without purification between steps of a synthetic sequence," Brain says. The greater control over reaction processes afforded by the use of microwave energy can be used to minimize by-products, leading to cleaner product profiles. The shorter reaction times involved in microwave heating can also potentially prevent excessive thermal decomposition of thermally sensitive reagents and facilitate rapid optimization of new reactions, making it possible "to investigate many new chemistries very quickly and efficiently," Brain notes.

Early work on microwave synthesis "really relied on the expedient of modified domestic microwave instruments," he says. "We know these are very limited in their applications." They have poor reproducibility, crude controls, heterogeneous microwave fields, and "they're intrinsically hazardous when used with organic solvents." A new generation of scientific microwave instruments offers safe operation with accurate controls; real-time monitoring of power, temperature, and pressure parameters; and the ability to interface with other lab instrumentation, such as liquid-handling robots, he says.

The synthesis of 1,3,4-oxadiazoles via the dehydration of 1,2-diacylhydrazines has traditionally been carried out with harsh reagents under forcing conditions--limiting the range of substrates that could be used, Brain says. However, by using a new polystyrene-supported dehydrating agent (for easier purification) in conjunction with microwave heating, he and his coworkers were able to synthesize a collection of 1,3,4-oxadiazoles in excellent yields under much milder conditions, broadening the range of compatible substrates [Synlett, 2001, 382]. In one experiment, oxadiazole synthesis with a support-bound reagent was incomplete after three hours of conventional thermal heating but fully complete after less than five minutes of microwave heating.

"Controlled microwave heating clearly provides advantages in synthetic efficiency," Brain says, "and interfacing this with polymer-supported reagents is particularly attractive" because it makes it possible for one to simply irradiate a mixture of reagents and filter off the desired products.

THE SEQUENCING of the genomes of humans and other organisms is "certainly one of the most exciting events in science over this past century," says chemistry professor Jonathan A. Ellman of the University of California, Berkeley. As a result, "we're now at the stage where, in the life sciences for many years to come, the major goal will be the identification of the function of the hundreds of thousands of proteins coded by these genomes." Combinatorial chemistry can play a major role in this task, he says.

Ellman's group has been focusing on using combinatorial chemistry in a systematic way to establish the function of proteases, which play a critical role in regulating biological processes. Combichem can be used to establish the function of proteases in two different ways, he says:

  • Establishing substrate specificity--the amino acid sequence at which a protease cleaves--expedites the identification of a protease's physiological substrates and provides tremendous insight into its biological function.
  • Identifying potent, cell-permeable inhibitors of a protease permits the use of those inhibitors as research tools for function studies.

One type of combinatorial library used to establish substrate specificity is the fluorogenic positional scanning substrate library pioneered by Kevin T. Chapman, Nancy A. Thornberry, and coworkers in the departments of molecular design and diversity and of biochemistry at Merck Research Laboratories, Rahway, N.J. Positional scanning is a combinatorial technique for identifying optimal substituents at different positions within a set of molecules in an efficient manner.

The Merck group's use of fluorogenic substrates in positional scanning libraries made it possible for them to profile enzymes, categorize them into groups based on their substrate specificities, and establish their functions. However, the method was restricted to fluorogenic substrates incorporating an aspartic acid residue at cleavage sites, limiting the proteases that could be profiled.

Ellman and coworkers--in collaboration with professor Charles S. Craik's group in the department of pharmaceutical chemistry at UC San Francisco--expanded the method so it was compatible with substrates containing any amino acid residues at cleavage sites. They have used it to profile the specificities of over 50 proteases to date. "The key point of this strategy is that we can go after many proteases rapidly," Ellman says.

KEY REACTIONS Ellman, Bergman, and coworkers used combinatorial means to determine if the synthetic potential of C–H activation is on a par with that of palladium-based carbon-carbon formation and olefin metathesis--perhaps the most important organic synthetic reactions introduced in the 1980s and '90s.

IN ADDITION, mechanism-based combinatorial libraries can be used to develop inhibitors for all members of specific enzyme families, Ellman notes. Mechanism-based libraries are collections of compounds containing known groups that act as key binding elements that interact with conserved active sites of enzyme families.

Ellman's group, in collaboration with that of chemistry professor Irwin D. Kuntz Jr. of UC San Francisco, prepared and screened a mechanism-based library of 1,000 support-bound compounds for inhibitors of the aspartyl protease cathepsin D, which is implicated in neurodegenerative diseases. The study turned up a number of potent nonpeptide inhibitors of the enzyme. In a further collaboration with the group of neurobiology professor Gary Lynch of UC Irvine, Ellman's group used those inhibitors to establish that cathepsin D is most likely responsible for a proteolytic process leading to the formation of neurofibrillary tangles in Alzheimer's disease and a related neurodegenerative condition--suggesting novel therapeutic strategies for these diseases.

Using mechanism-based libraries, Ellman and coworkers have also identified nanomolar inhibitors of cruzain, which is associated with Chagas' disease, a parasitic infection, and rhodesain, which is implicated in sleeping sickness. "What we want to do is develop potent reversible inhibitors that improve selectivity and minimize toxicity relative to previously reported irreversible inhibitors" of these enzymes, Ellman says. The researchers currently have a crystal structure of one of their cruzain inhibitors, "which is going to provide a great deal of insight into design efforts." And some of the rhodesain inhibitors have shown promising activity in culture studies against sleeping sickness.

Combinatorial techniques can be used to target not only biological function but also new chemistry. Important organic synthetic reactions developed in the past two decades include palladium-mediated carbon-carbon bond formation in the 1980s and olefin metathesis in the '90s. They're both carbon-carbon bond-forming reactions that help researchers construct desired compound skeletons, they both use readily available substrates or feedstocks, and they are both compatible with many types of functional groups, "so we can introduce amines, acids, esters, amides, and indoles," Ellman says--"the kinds of functionality we find in drugs and natural products."

What will be the most important synthetic reaction developed in this decade? "It's my belief that carbon-hydrogen (C–H) activation really could be the chemistry that comes to the fore," Ellman says. C–H activation and follow-up reactions can be used to form carbon skeletons. And nearly every molecule has a carbon-hydrogen bond, "so you really can't beat C–H activation in terms of substrate availability." But only a limited amount of information is available on the technique's overall scope and limitations. "Can we apply it to a range of different heterocycles, or is it limited?" Ellman asks. "Can we apply it in the presence of a wide range of functionality, in the presence of different electronic requirements, or in the presence of different steric situations? Does this chemistry have a level of generality that is required for it to be really powerful?"

To answer such questions, one must test a reaction on a large number of compounds, and Ellman's group, in collaboration with that of chemistry professor Robert G. Bergman of UC Berkeley, set out to do that in an efficient way, using combinatorial chemistry. They used a C–H activation-based reaction sequence to convert single substrates or mixtures of substrates into ketone products. They derivatized the products with a mass spectrometric label and analyzed them by electrospray MS. This enabled them not only to identify the products but also to obtain reaction yield data and structure-versus-rate data on those products.

They were able to identify a wide range of C–H activation reaction products, including primary amides, free amines, tertiary amines, nitriles, halides, alcohols, esters, and acidic compounds. "Through this combinatorial effort we became very excited about C–H activation and became convinced that it had sufficient functional group compatibility and generality that it could be applied in productive ways," Ellman says. "In collaboration with Bob Bergman, we're currently applying this C–H activation technology in parallel synthesis applications."

BAR CODES Fenniri (front to back), Laboratory for Chemical Nanotechnology Director Alexander E. Ribbe, and graduate student Yegor Zyrianov of Purdue University examine spectroscopic bar codes of newly designed resins that could effectively reduce the number of combinatorial library deconvolution steps to zero.
A MAJOR CHALLENGE in combinatorial chemistry is finding ways to reduce the effort and cost required to identify promising compounds from combinatorial libraries. With the completion of the Human Genome Project, it's been estimated that every major pharmaceutical company will be screening over a million compounds a year by 2005, says assistant professor of chemistry Hicham Fenniri of Purdue University--and new strategies will be needed to help researchers identify the active ones.

Compound identification strategies used in the screening of large solid-phase libraries include deconvolution techniques (such as positional scanning) and encoding techniques. In deconvolution, researchers figure out the identity of an active compound by tracing it back through the synthetic mixtures used to create it. Encoding is a strategy for marking compounds with chemical or physical tags so the compounds can be easily identified.

Last year, Fenniri and coworkers introduced an approach called dual recursive deconvolution (DRED) that is a hybrid of deconvolution and encoding. DRED approximately doubles the number of synthetic steps required to create a library, but it identifies all the compounds in the library as it synthesizes them. That's well worth the additional synthetic steps, Fenniri maintains.

The beads formerly used in DRED were heterogeneous mixtures of resins selected from various commercial sources, and specialized instrumentation was required to analyze them--factors that limited the practicality of the technique. Fenniri and coworkers have now enhanced the DRED technique by synthesizing resin beads with built-in spectroscopic bar codes. Vibrationally active groups that they incorporated into the beads are chemically inert and readily identifiable with standard Raman or Fourier transform infrared spectrometers. Each bead's unique vibrational signature can also be converted to a bar code for rapid visual or standard laser-bar-code reader identification [J. Am. Chem. Soc., 123, 8151 (2001)].

Fenniri and coworkers are currently using directed sorting strategies, initially introduced by Irori, to eliminate library redundancy, so each bar-coded bead is assigned to a single library member. "The deconvolution process in this case boils down to simply collecting the active beads, reading their bar code, assigning a chemical structure, and confirming the result," Fenniri says. "The number of deconvolution steps is effectively reduced to zero as each bar code encodes and thus identifies a unique compound."

Advantages of the DRED method and spectroscopic bar coding, Fenniri says, include its low cost, its adaptability to any solid-phase chemistry, and the ability to prepare beads in one-step polymerization reactions from readily available commercial starting materials. "Resins with built-in bar codes will soon be available commercially," he says. "Several chemical companies are currently reviewing this technology, and the first generation of bar-coded resins could reach the market within the next few months."


[Previous Story] [Next Story]


Chemical & Engineering News
Copyright © 2001 American Chemical Society