Transmembrane signaling of chemotaxis receptor tar: insights from molecular dynamics simulation studies

Abstract

Transmembrane signaling of chemotaxis receptors has long been studied, but how the conformational change induced by ligand binding is transmitted across the bilayer membrane is still elusive at the molecular level. To tackle this problem, we carried out a total of 600-ns comparative molecular dynamics simulations (including model-building simulations) of the chemotaxis aspartate receptor Tar (a part of the periplasmic domain/transmembrane domain/HAMP domain) in explicit lipid bilayers. These simulations reveal valuable insights into the mechanistic picture of Tar transmembrane signaling. The piston-like movement of a transmembrane helix induced by ligand binding on the periplasmic side is transformed into a combination of both longitudinal and transversal movements of the helix on the cytoplasmic side as a result of different protein-lipid interactions in the ligand-off and ligand-on states of the receptor. This conformational change alters the dynamics and conformation of the HAMP domain, which is presumably a mechanism to deliver the signal from the transmembrane domain to the cytoplasmic domain. The current results are consistent with the previously suggested dynamic bundle model in which the HAMP dynamics change is a key to the signaling. The simulations provide further insights into the conformational changes relevant to the HAMP dynamics changes in atomic detail.

Figure 1

Schematic representation of the simulated system. The homodimeric units are colored in white and gray, respectively. Notations for the structural units and the ranges of the residues that compose the units are shown together. The periplasmic and cytoplasmic domains not included in the simulations are indicated with a box and an oval at the top and bottom, respectively. The membrane-water interfaces at the periplasmic side (interface P) and at the cytoplasmic side (interface C) are shown as dashed lines.

Figure 2

Average structure for the AA state obtained from the simulation A is shown from two viewpoints rotated by 90°. DPPC phosphorus atoms are represented as orange spheres to show the water-membrane interfaces. Notations for the helices in the TM domain (TM1, TM1′, TM2, and TM2′) and for those in the HAMP domain (AS1, AS1′, AS2, and AS2′) are shown along with the structure on the left. The control cables that connect TM2 and AS1 are colored in black. The side chains of Trp-209 and Trp-209′ located right above the control cable is explicitly shown.

Figure 3

Average structures of AA (pink) and AH (green) in (a) the lipid-water interface on the periplasmic side (interface P), (b) the TM region, and (c) the lipid-water interface on the cytoplasmic side (interface C), as indicated in the figure on the lower right corner. The spatial restraints from the crystal structure of the periplasmic domain are applied to the periplasmic terminal residues (Leu-165 to Ala-169), the top of helices in a. Downward piston movement of TM2′ in the AH state is observed in both a and c. However, in the TM region shown in B, TM2′ undergoes different motions than simple downward sliding. The conformational change of TM2′ upon ligand binding is coupled with the change at the control cable, as shown with darker colors in b. The TM2′ axis in the AH state (green arrow in b) becomes better aligned to the AS1′ axis than that in the AA state (purple arrow in b). The extents of the TM2′ movements are reported in Table S5 and Figs. S5 and S6.

Figure 4

Contact map differences between AA and AH from the MD average structures. Cyan spots correspond to residue pairs closer in AA than in AH, and magenta spots the opposite. Only the residue pairs within 15 Å in either AA or AH are considered. The regions with large differences are highlighted by the circles with different colors in the figure on the right. (a) The contact difference for the monomer that ligand binds. Prominent changes at the TM1′-TM2′ interface along the inverse-diagonal direction (see colored circles) support discontinuous longitudinal movement of TM2′ against TM1′. (b) The contact difference for the dimer interface. Sliding of AS1′ relative to AS1 is detected, as shown with a green circle.

Figure 5

(a) HAMP domain structures in the AA state (pink) and the AH state (green) show a change in the HAMP orientation. The HAMP structures are compared with the NMR structure of Af1503 HAMP (white) in b and c. (d) C_α RMSF of the HAMP domain, averaged over three independent simulations, are compared in between the AA and AH states for ligand-free monomer (left) and for ligand-bound monomer (right). In this plot, the RMSF within the HAMP domain are obtained by superimposing the HAMP domain structures to exclude the effect of the whole domain motion. In the left panel of d, the RMSF for the AH state becomes smaller at AS1 but larger at AS2 compared to AA, implying that the HAMP helix connected to TM (AS1) becomes less dynamic but the HAMP helix connected to the cytoplasmic domain (AS2) becomes more dynamic upon ligand binding. As a comparison, the RMSF for the ligand-bound monomer shows little difference between the two states, as shown on the right in d.

Figure 6

Proposed transmembrane signaling mechanism induced by attractant stimulus of aspartate binding in Tar is shown schematically. Piston movement in the periplasmic domain is relayed along the TM2′ helix, bringing about bending of TM2′ at the bilayer center and conformational change at the control cable, and finally induces orientation and dynamics change of HAMP AS2 helix, which would further affect conformational change of the cytoplasmic domain.