Mechanistic insights revealed by the crystal structure of a histidine kinase with signal transducer and sensor domains

Abstract

Two-component systems (TCSs) are important for the adaptation and survival of bacteria and fungi under stress conditions. A TCS is often composed of a membrane-bound sensor histidine kinase (SK) and a response regulator (RR), which are relayed through sequential phosphorylation steps. However, the mechanism for how an SK is switched on in response to environmental stimuli remains obscure. Here, we report the crystal structure of a complete cytoplasmic portion of an SK, VicK from Streptococcus mutans. The overall structure of VicK is a long-rod dimer that anchors four connected domains: HAMP, Per-ARNT-SIM (PAS), DHp, and catalytic and ATP binding domain (CA). The HAMP, a signal transducer, and the PAS domain, major sensor, adopt canonical folds with dyad symmetry. In contrast, the dimer of the DHp and CA domains is asymmetric because of different helical bends in the DHp domain and spatial positions of the CA domains. Moreover, a conserved proline, which is adjacent to the phosphoryl acceptor histidine, contributes to helical bending, which is essential for the autokinase and phosphatase activities. Together, the elegant architecture of VicK with a signal transducer and sensor domain suggests a model where DHp helical bending and a CA swing movement are likely coordinated for autokinase activation.

Figure 1. Overall structure of VicK.

(A) Domain architecture of the full-length VicK. The numbers below are the breakpoints of functional domains, which are colored differently and labeled on top. (B) Molecular weight of VicK in solution measured by multiangle light scattering (MALS). BSA was used as a control colored in blue and VicK in red. The time range was taken from the HPLC. (C) Overall structure of VicK presented in ribbon. Two monomers are colored in magenta and gold. The corresponding domains are indicated. Helices α1–11 are labeled. Na and Nb indicate the N termini of the monomer A or B while Ca and Cb indicate their C termini. (D) Electrostatic potential surface of the VicK monomer. Red to blue colors represent negative to positive charged areas (−0.75 to +0.75, CCP4mg). VicK monomer B is presented in ribbon with CA domain deleted for better clarity. Contact areas (I–III) of the VicK dimeric interface are marked in yellow arrows.

Figure 2. Structure of VicK HAMP domain.

(A) Sequence alignment of HAMP domains from representative SKs of Streptococci together with A. fulgidus Af1503. The positions of amino acids below are labeled according to S. mutans VicK. The a and d positions defined according to the classical coiled-coil are indicated on the top and highlighted in red. The x and da positions are also labeled according to the unique knobs-to-knobs model defined by Hulko et al. . The red star indicates a conserved Gly in all HAMP domains and highlighted in yellow. (B), (C), and (D) The hydrophobic knobs-to-knobs contacts in the central HAMP domain. The critical residues are shown as blue sticks and labeled in red.

Figure 3. Structure of VicK PAS domain.

(A) A canonical PAS monomer of VicK shown in gold ribbon. Another PAS monomer is fainted for better clarity. β-strands, helices, and loops are labeled according to the overall structure of the VicK molecule. (B) Electrostatic surface of the VicK PAS domain monomer. The color scheme is the same as Figure 1D. Three potential sockets (S1–3) for ligands to bind are labeled. (C) Leucine-zipper interface of the VicK PAS domain dimer with critical residues shown in sticks and labeled on the right.

Figure 4. Helical bending property of the DHp domain.

(A) Structural alignment of two VicK DHp chains; (B) The VicK DHp domain aligned with T. maritima HK853. The HK853 DHp domain is colored in blue. VicK His217 and Pro222 are indicated with arrows. Bottom parts of helices (aa219–255) of the VicK DHp domain were used for alignments. The right is a top view of these aligned structures shown on the left. The blue arrows indicate direction and relative distance of shifts. The red circle represents a central axis parallel to the DHp domain and going into the paper. (C) Conservation and helicity of the DHp domain. The bottom part is a conservation logo using over 40 non-redundant histidine kinases. The top part is the helicity of the DHp domain calculated by Phyre described in Materials and Methods. (D) Autokinase analysis using isotope γ³²P-ATP. The wt VicK and its mutants were analyzed on 15% SDS-PAGE and stained with coomassie blue as shown in the bottom gel. The top gel is an autoradiograph of their autokinase activities of these proteins. A protein marker served as a background control (lane 3).

Figure 5. Proline and threonine are essential for VicK phosphatase activity.

(A) VicR dephosphorylation examined in PMS. Phosphorylated VicR was treated with the same amount of VicK wt or its mutants as shown in the bottom gel. Native and phosphorylated VicR were separated in PMS as labeled on the right of the top two gels, where in one series of reactions additional 5 mM ATP was used (the top 1 gel). (B) VicK phosphatase activity analyzed by HPLC. AcP-treated VicR was incubated with VicK in the reaction buffer with 5 mM ATP (right) and without ATP (left). VicK phosphatase activity was analyzed by the phosphorylation state of VicR on HPLC.

Figure 6. Analyses for autokinase activation.

(A) Structural alignment of two VicK monomers in reference to the DHp domain. The alignment was carried out as described in Figure 4A. The vertical movement is indicated with a black arrow in parallel to the DHp domain while a grey arrow indicates a rotation. (B) The VicK CA domain (magenta) aligned to B. subtilis YycG (light blue) in the presence of ATP (green and gold sticks). His217 is represented by green sticks and several other residues are shown in magenta and gold sticks. The contacts between the distances of 2.5–3.5 Å are highlighted with yellow dashed lines and those between 3.5–4.5 Å with grey dashed lines. (C) Residues critical for the interaction between the DHp and active CA domains. The VicK Cα chain is colored in gray ribbon and residues from the CA domain critical for this interface in cyan sticks. The critical residues from the DHp domain are highlighted in grey sticks. The right panel is a flip-over view of the left. (D) Mutations of key residues in the active interface affected autokinase activity. The amount of wt VicK and mutants used in each reaction were analyzed by 15% SDS-PAGE and stained by coomassie blue as shown in the bottom gel. A protein marker served as a background control (lane 6). The gel containing phosphorylation forms of VicK was exposed to an X-ray film as shown in the top gel.

Figure 7. Model of VicK autokinase activation.

The four cylinders in the middle of each state (I–IV) represent the four helical bundle of the DHp domain. The His217 residue is indicated in sphere and line attached to the four cylinders. The CA domains are shown in an L shape and ATP or its γ-phosphate in a red sphere. The unknown steps for VicK to return to the inactive state are indicated by a black arrow and labeled with a question mark.