Hossein Fazelinia
Shuang Yan Tang
Computational design of AraC protein with
novel effector specificity
We aim to develop screening or preferably selection systems to facilitate the directed evolution of enzymes which can synthesize fine chemicals with exquisite regio- and stereoselectivity. These selections will have two major benefits over traditional screening methods. First, selection systems allow for the "analysis" of mutant libraries which are significantly larger than those that are practical in traditional screening systems [22,23]. Second, it is usually very difficult to develop screening assays capable of discriminating between functionally similar or identical product isomers. This limitation is overcome by engineering a regulatory protein that selectively binds a desired molecule (and not its isomers) that is the product of a heterologous enzyme reaction or metabolic pathway. Therefore we must first develop highly specific AraC variants that are capable of discriminating between, for example, cis-verbenol and trans-verbenol, verbenone, and myrtenol.
Experimentally, multiple AraC mutagenesis strategies are being examined in E. coli using previously described reporter systems placed downstream of the ParaBAD promoter (for example see refs [24,25]). The DNA-looping region of the ara operon containing the two distant half-sites as well as the two contiguous half-sites remain intact in our reporter system to ensure tight transcriptional control by AraC. In addition, the native -ara operon is deleted from the host. We are determining how well specific point mutations predicted by our models correspond to improved transcriptional activation, only in the presence of the target molecule (e.g., verbenol). Libraries of the AraC proteins are also being generated and screened using cassette mutagenesis with oligonucleotides of a specified mutagenicity in regions predicted to be critical to substrate binding. Finally, libraries incorporating random point mutations over the entire AraC N-terminal domain are being generated and screened, and the resulting mutants will be compared to those found by computer-guided designs. Appropriate control experiments will be performed to ensure that "improved" AraC variants do not simply have relaxed specificity but rather increased selectivity for the target molecule. Finally, random mutagenesis can be used to "modulate" the sensitivity of improved AraC variants (i.e., increase affinity for the target molecule).