Combinatorial chemistry is revolutionizing the drug discovery process by facilitating the synthesis and analysis of an almost unlimited number of organic compounds1. Now we may wonder if the pharmaceutical field was merely a training ground for the invasion of combinatorial chemistry into all fundamental and applied areas of the biological sciences. In this issue, Schultz and colleagues2 provide a powerful example of how the use of compound libraries in chemical genetics—the small-molecule approach to gene function analysis—will influence cell biology and applied areas like tissue engineering and cell regeneration. These workers have identified a compound from a combinatorial library of simple purine analogs, called myoseverin, that induces the reversible fission of myotubes (muscle cells) into mononucleated fragments. Unexpectedly, myotube fission promoted cell proliferation after removal of myoseverin. In fact, the drug affected the expression of several genes involved in the process of tissue regeneration and wound healing. These results demonstrate that a muscle cell regeneration process commonly found in amphibians and reptiles (see Fig. 1) may also be possible with some mammals.
A more traditional approach to genetics involves the introduction of mutations in genes, analysis of the physiological effects, and identification of the encoded protein(s). In contrast, the chemical genetics3 approach makes use of small molecules that interact with proteins so as to disrupt the pathways in which they function and elucidate the protein's role in the cell. In the forward variant of chemical genetics, a large collection of small molecules is screened for an effect on a specific cellular pathway or metabolic process, or as in this case, cellular morphology. Obviously, the odds of finding hits increase with the size of the library. However, identifying a compound that provides the desired physiological effect is really just the launching point for more exciting and often surprising findings.
During the formation of skeletal muscle, proliferating myoblasts stop dividing and fuse into myotubes. To reach their goal of promoting the reverse process, Schultz and colleagues examined the ability of members of a library of 2,6,9-trisubstituted purines to disassemble multinucleated myotubes into individual cells. Mouse C2C12 myotubes derived from differentiated myoblasts were incubated with library members, and their disintegration into mononucleated fragments was followed using phase-contrast optics. Immunofluorescence studies on myoseverin, the most potent compound identified, showed that it binds to tubulin and leads to the disintegration of the microtubule cytoskeleton. Comparison of myoseverin to other tubulin binders such as taxol, vinblastine, and colchicine revealed that whereas myoseverin and colchicine have reversible effects (i.e., disassembled myotubes reorganize when the drug is removed and fresh medium is introduced), myoseverin produces non-apoptotic fragments and is less toxic to proliferating myoblasts. Surprisingly, differentiated cells freed from myoseverin and incubated in fresh growth medium synthesize DNA and proliferate at more than twice the rate of untreated cells. Transcriptional profiling would reveal, however, that the drug does not act alone in promoting cell proliferation.
Transcriptional profiling is a technique employed to monitor comprehensive changes in the expression of multiple genes simlutaneously. To which extent could micromolar amounts of a small molecule like myoseverin induce changes in the transcriptional profile of myocytes? Schultz and colleagues analyzed the expression of more than 6,000 genes by comparing mRNA levels of drug-treated cells with untreated ones. This study revealed that more than half of the 93 genes affected by myoseverin are involved in the cellular response to tissue injury. More detailed studies are needed to determine the precise molecular basis of myoseverin action, as the functions of many of the other genes affected are unknown. In fact, as genome sequencing progresses, deciphering gene function is precisely the role that small molecules from combinational libraries are expected to play.
Few rules direct the selection of a particular class of compounds in a library approach to cell-based functional screens. Purines, are key components of nucleic acids, and ubiquitous compounds in living systems. Earlier, this group had reported that several 2,6,9-trisubstituted purines inhibit cyclin-dependent kinases (CDK)4. Myoseverin, however, does not inhibit CDK1 activity and thus must affect cell differentiation by a different mechanism. Therefore, in their search for activators of myotube disassembly, it can be assumed that the authors had few reasons to use a purine library other than the practicality of having this library in hand from earlier work4. Possessing such an exclusive collection of compounds alone is a great advantage, but it comes at the price of developing an efficient synthetic route to access these novel molecules. This can be a cumbersome and time-consuming chemical venture. As shown with the current example, the bravest efforts can be rewarded with stunning results.
This article by Schultz and colleagues shows an impressively high degree of multidisciplinary research. Although barriers between chemistry and biology keep falling, the expertise and infrastructure required to undertake this type of work independently, from library synthesis to the use of genomics technology, can be limiting to several academic laboratories. Hopefully, more chemists specialized in synthesizing novel small-molecule libraries will collaborate with biologists. To the advantage of biologists less versed in library synthesis, it is not a surprise in today's world that even small-molecule libraries can be purchased. For example, the use of a commercial, structurally diverse small-molecule library was recently reported in the discovery of simple compounds that helped define the role of copper(II) ions in the activation of specific genes involved in ion homeostasis in yeast5.
These results emphasize the enormous potential of unbiased small-molecule libraries and cell-based screens to turn up compounds with novel and unpredicted biological properties. The approach is not limited to parallel libraries; a split-pool library containing mixtures of millions of new, bead-supported polycyclic compounds has been made for use in cell-based screens3. There are indeed billions of unnatural molecules waiting to be synthesized and screened for activity against a multitude of poorly understood cellular processes. This work provides a small, but exciting glimpse at the bright future of combinatorial chemistry and chemical genetics in both the study of basic cell biology and more applied disciplines such as cell regeneration and tissue engineering. Footnote 1
In the original print version of this article, the word "combinatorial" in the title was misspelt "combinational".
Plunkett, M.J. & Ellman, J.A. Sci. Am. 276, 68–73 (1997).
Rosania, G.R. et al. Nat. Biotechnol. 18, 304– 308 (2000).
Tan, D.S., Foley, M.A., Stockwell, B.R., Shair, M.D. & Schreiber, S.L. J. Am. Chem. Soc. 121, 9073–9087 (1999).
Chang, Y.T. et al. Chem. Biol. 6, 361– 375 (1999).
Stockwell, B.R., Hardwick, J.S., Tong, J.K. & Schreiber, S.L. J. Am. Chem. Soc. 121, 10662–10663 (1999).