Study of a TCS activated by light in Brucella abortus.

SYCZ, G.; GOLBAUM, F.A.; FERNÁNDEZ, I.; RINALDI, J.; KLINKE, S.; OTERO, L.H.

Hyderabad

Congreso; 24th Congress and General Assembly of the International Union of Crystallography,; 2017

IUCR

Two-component signal transduction systems (TCSs) are modules that allow bacteria to rapidly adapt to changing environmental conditions. In the most common case, they are formed by a sensor histidine kinase (HK) which, upon sensing of an external signal, is able to phosphorylate a cognate response regulator (RR). The latter protein undergoes structural changes that are able to modify gene expression by directly binding to DNA, catalyze metabolic reactions or alter protein-protein interactions. The pathogenic bacterium Brucella abortus, the causative agent of the disease brucellosis, bears a particular two-component system formed by a dimeric cytoplasmic three-domain blue-light sensor HK (LOV-PAS-HK) and two monomeric RRs called PhyR and CheY. The activation of this HK has been shown to increase the virulence of this pathogen. Additionally, a mutant in the LOV-PAS-HK gene presents reduced infectivity, suggesting that this system is a virulence factor and an important regulatory system during the infection process. With the goal of understanding at the atomic level the activation and signal transduction events of this system, we aimed to solve the three-dimensional structures of these proteins by means of X-ray diffraction. The core of the blue-light sensor FMN-binding LOV domain (which stands for Light, Oxygen, and Voltage) was crystallized and its structure solved at 1.64 Å resolution in the dark. It adopts the alpha/beta PAS (Per-ARNT-Sim) domain fold and dimerizes through a hydrophobic central beta-scaffold in an antiparallel way [1]. Interestingly, we were also able to determine the structure of a construct comprising the LOV core domain plus an N-terminal capping helix at 2.34 Å resolution (N-LOV), observing now the expected parallel dimerization nature of the protein. This arrangement has been confirmed with the recent determination of the N-LOV-PAS structure at 2.74 Å resolution in the lab, which holds a long connecting ?J? helix between both globular domains. Additionally, we were able to solve the crystal structure of the isolated HK domain at 2.51 Å resolution by sulfur SAD in a challenging procedure, since the closest HK present in the PDB bears only 20% sequence identity, the crystals belong to the low symmetry P2(1) space group and there were four copies of the molecule in the 108-kDa asymmetric unit [2]. Interestingly, the HK structure presents two different dimeric assemblies in the asymmetric unit: one similar to the already described canonical HK parallel homodimers (C) and the other, an antiparallel non-canonical (NC) dimer, each with distinct relative subdomain orientations and dimerization interfaces [3]. To finish, we were also able to determine the structure of the PhyR RR at 2.05 Å resolution, and efforts are underway to obtain the structure of the full LOV-PAS-HK protein as well as HK-RR complexes. All these protein structures, together with spectroscopic, activity and biophysical assays, allowed us a better understanding of this crucial system for the pathogenicity of Brucella.