In this continuation application, the investigator proposes to continue to develop and refine a genomic version of the SELEX technique, a method invented in his laboratory almost 10 years ago. The original SELEX technique used a very large number of synthetic oligonucleotides in a multi-cycle procedure to select for those nucleic acids that bind specifically and tightly to purified protein. The original protein and others that have been studied by this procedure were logical choices for this method because they were known to be nucleic acid binding proteins (""""""""professional"""""""" n.a. binding proteins according to the investigator). However, the premise that RNAs and DNAs could bind to all sorts of proteins, perhaps for as-yet known biological reasons, was put forward, and much of SELEX has been developed to search for these types of interactions. Even without a biological purpose, nucleic acids that bind specifically and tightly to particular parts of proteins could be very useful modulating agents, both in research and perhaps clinically. Genomic-SELEX is a modification of the original procedure in which the nucleic acid pool from which the oligonucleotides are derived is a naturally occurring genome or part of a genome. Thus, real, natural nucleic acids inside the same cell as the interesting proteins could be selected, amplified, purified, and sequenced by Genomic-SELEX. The investigator and colleagues have performed proof-of-principle experiments in a variety of settings with this modification of SELEX. This proposal is an ambitious one to continue to develop Genomic-SELEX, with a particular emphasis on making several key improvements: 1) by applying a modified protocol for separating protein:nucleic acid complexes from unbound n.a., the cycle time for performing the entire procedure is vastly reduced, by as much as 30-fold. 2) attempts will be made to develop a standard set of conditions for performing the binding reactions that approximate physiological conditions inside a cell. 3) the approach will be applied to single proteins as well as complex mixtures from both yeast and humans. The use of many proteins in a single experiment is an important step for increasing the throughput of Genomic-SELEX. In the course of making these improvements, the investigator's group will identify specific nucleic acids that bind tightly to a large number of yeast and human proteins. The proposal would entail the beginnings of a """"""""protein:nucleic acid interaction map"""""""", analogous to the protein:protein interaction maps being developed by Stanley Fields and colleagues. The investigator's group will sequence any specific interacting nucleic acids that are identified, and hope to be able to glean some biological hints as to why the sequence interacts and what it does when it interacts. Towards this latter question, experiments will be done in the cases of those proteins with known activities to determine whether the nucleic acid affects that activity in any way. Another approach that will be taken to try to understand the biological basis for the interactions, especially for human proteins, will be to collaborate with several researchers who specialize in computational methods for identifying genes; these studies will help assure that the human nucleic acid sequences that are found are placed in their genomic context as quickly as possible, which is important because the human sequence will not be completely known for some years yet. Much of the biological emphasis of this study will be on problems in human biology related to cancer and angiogenesis.
