UCLA

Phosphoryl transfer from a histidine kinase to a response regulator in two-component signaling systems leads to cellular modifications that enable bacteria to adapt to environmental changes. The response regulator NarL, of the Escherichia coli Nar two-component system, becomes phosphorylated by the histidine kinase NarX and regulates genes involved in nitrate respiration. Phosphorylation at the N-terminal "receiver" domain of NarL exposes distant molecular surfaces, including regions of the C-terminal domain that dimerize upon DNA binding. To investigate other alterations in domain surface interactions upon phosphorylation, NarL and its individual domains were examined by analytical ultracentrifugation. Phosphorylation was demonstrated to induce full-length NarL dimerization and tetramerization, and N-terminal domain dimerization. The C-terminal domain was unable to dimerize alone, suggesting that dimerization of the in-tact protein occurs via receiver domains, while C-terminal domain dimerization is driven by DNA binding. Independent receiver-domain dimerization implicates the widespread alpha4-beta5-alpha5 surface as the dimerization site, since this region is unavailable in full-length NarL, which was shown to be monomeric. Also, receiver domain dimerization and NarL oligomerization may fulfill binding requirements at low-affinity promoter regions.

Two unphosphorylated NarL crystal structures reveal, previously unreported, distinct equilibrium states with variability in residue positions and polar contacts. The conformations of two activation-associated residues in the receiver domain show that the monoclinic NarL crystal structure represents a semi-activated state. Both structures were analyzed with respect to the interdomain interface and active site in order to gain an understanding of the activation mechanism that leads to domain separation. A solvent-accessibility analysis, along with structural comparisons, confirmed that the conformation of Lys196, in the semi-activated NarL structure, is also representative of activation. Mobility ratios and correlation fluctuation calculations showed that vital interface-loop residues exhibit low mobilities, however can affect the motion of other interface residues, including those that bind DNA. Therefore, interface mutations that lead to constitutively active phenotypes may also result from the interdependence of residue motions. In contrast to the interface loops, the active-site loops were characterized by high mobility ratios. Structural modifications of these flexible loops are expected to coincide with the proposed movements of certain active-site residues.