Exploring Chemical Diversity: Towards Rational Synthesis and Analysis of Large Ordered Libraries

Neal Woodbury
Department of Chemistry and Biochemistry
Arizona State University

A variety of methods have been developed for the exploration of chemical space with reference to a particular function. In the pharmaceutical industry, combinatorial chemistry has been used to generate moderately large (105 to 106) libraries of compounds that can then be tested either as mixtures or individually. Biochemical methods such as phage display, SELEX and mRNA display have made it possible to screen much larger libraries (up to about 1013) of biopolymers for members with optimal binding and in some cases catalysis. Both approaches have advantages. The combinatorial chemical libraries are small enough to allow testing of individual chemicals, providing a complete correlation between structure and function for a particular library. However, it is time consuming to create and analyze these libraries making it difficult to perform iterative improvement of molecular function in this way. The biochemical libraries allow the exploration of much more structural space, and they can be readily iterated to improve function in multiple rounds of mutagenesis and selection, but generally they provide only functional information about the molecules that emerge from the particular selection method used, rather than correlated structure/function relationships in the whole library. We have been developing technology pioneered by the DNA chip industry to explore ordered libraries of heteropolymers on surfaces. This allows the use of moderate sized libraries (105 or more) and provides structure/function information for all elements of the library. It also provides a rapid means of creating new libraries based on the results of previous libraries for iterative improvement of function. Our initial work has been done using DNA chips themselves and looking at the structure function properties of libraries of DNA aptamers. In addition, we have been developing similar high throughput synthetic methods for creating moderate sized libraries of peptides and other amide-linked heteropolymers. One application of this work has been the selection of ligands for assembly into synthetic antibodies and this work will be described. Computationally, challenges exist in developing algorithms both for creating the initial libraries, analyzing the structure function correlations that come out of these screens, and then using this information to create new libraries that facilitate the discovery of molecules with desired function.

Back to the program