Features of protein-protein interactions in two-component signaling deduced from genomic libraries

As more and more sequence data become available, new approaches for extracting information from these data become feasible. This chapter reports on one such method that has been applied to elucidate protein-protein interactions in bacterial two-component signaling pathways. The method identifies residues involved in the interaction through an analysis of over 2500 functionally coupled proteins and a precise determination of the substitutional constraints placed on one protein by its signaling mate. Once identified, a simple log-likelihood scoring procedure is applied to these residues to build a predictive tool for assigning signaling mates. The ability to apply this method is based on a proliferation of related domains within multiple organisms. Paralogous evolution through gene duplication and divergence of two-component systems has commonly resulted in tens of closely related interacting pairs within one organism with a roughly one-to-one correspondence between signal and response. This provides us with roughly an order of magnitude more protein pairs than there are unique, fully sequenced bacterial species. Consequently, this chapter serves as both a detailed exposition of the method that has provided more depth to our knowledge of bacterial signaling and a look ahead to what would be possible on a more widespread scale, that is, to protein-protein interactions that have only one example per genome, as the number of genomes increases by a factor of 10.