Three strategically placed hydrogen-bonding residues convert a proton pump into a sensory receptor.

In haloarchaea, light-driven ion transporters have been modified by evolution to produce sensory receptors that relay light signals to transducer proteins controlling motility behavior. The proton pump bacteriorhodopsin and the phototaxis receptor sensory rhodopsin II (SRII) differ by 74% of their residues, with nearly all conserved residues within the photoreactive retinal-binding pocket in the membrane-embedded center of the proteins. Here, we show that three residues in bacteriorhodopsin replaced by the corresponding residues in SRII enable bacteriorhodopsin to efficiently relay the retinal photoisomerization signal to the SRII integral membrane transducer (HtrII) and induce robust phototaxis responses. A single replacement (Ala-215-Thr), bridging the retinal and the membrane-embedded surface, confers weak phototaxis signaling activity, and the additional two (surface substitutions Pro-200-Thr and Val-210-Tyr), expected to align bacteriorhodopsin and HtrII in similar juxtaposition as SRII and HtrII, greatly enhance the signaling. In SRII, the three residues form a chain of hydrogen bonds from the retinal's photoisomerized C(13)=C(14) double bond to residues in the membrane-embedded alpha-helices of HtrII. The results suggest a chemical mechanism for signaling that entails initial storage of energy of photoisomerization in SRII's hydrogen bond between Tyr-174, which is in contact with the retinal, and Thr-204, which borders residues on the SRII surface in contact with HtrII, followed by transfer of this chemical energy to drive structural transitions in the transducer helices. The results demonstrate that evolution accomplished an elegant but simple conversion: The essential differences between transport and signaling proteins in the rhodopsin family are far less than previously imagined.

Systems Biology Approaches to Probe Gene Regulation in Bacteria

Mechanisms that allow pathogens to colonize the host are not the product of isolated genes, but instead emerge from the concerted operation of regulatory networks. Therefore, identifying components and the systemic behavior of networks is necessary to a better understanding of gene regulation and pathogenesis. To this end, I have developed systems biology approaches to study transcriptional and post-transcriptional gene regulation in bacteria, with an emphasis in the human pathogen Mycobacterium tuberculosis (Mtb). First, I developed a network response method to identify parts of the Mtb global transcriptional regulatory network utilized by the pathogen to counteract phagosomal stresses and survive within resting macrophages. As a result, the method unveiled transcriptional regulators and associated regulons utilized by Mtb to establish a successful infection of macrophages throughout the first 14 days of infection. Additionally, this network-based analysis identified the production of Fe-S proteins coupled to lipid metabolism through the alkane hydroxylase complex as a possible strategy employed by Mtb to survive in the host. Second, I developed a network inference method to infer the small non-coding RNA (sRNA) regulatory network in Mtb. The method identifies sRNA-mRNA interactions by integrating a priori knowledge of possible binding sites with structure-driven identification of binding sites. The reconstructed network was useful to predict functional roles for the multitude of sRNAs recently discovered in the pathogen, being that several sRNAs were postulated to be involved in virulence-related processes. Finally, I applied a combined experimental and computational approach to study post-transcriptional repression mediated by small non-coding RNAs in bacteria. Specifically, a probabilistic ranking methodology termed rank-conciliation was developed to infer sRNA-mRNA interactions based on multiple types of data. The method was shown to improve target prediction in Escherichia coli, and therefore is useful to prioritize candidate targets for experimental validation.