New Insights on Signal Propagation by Sensory Rhodopsin II/Transducer Complex

The complex of two membrane proteins, sensory rhodopsin II (NpSRII) with its cognate transducer (NpHtrII), mediates negative phototaxis in halobacteria N. pharaonis. Upon light activation NpSRII triggers a signal transduction chain homologous to the two-component system in eubacterial chemotaxis. Here we report on crystal structures of the ground and active M-state of the complex in the space group I212121. We demonstrate that the relative orientation of symmetrical parts of the dimer is parallel (“U”-shaped) contrary to the gusset-like (“V”-shaped) form of the previously reported structures of the NpSRII/NpHtrII complex in the space group P21212, although the structures of the monomers taken individually are nearly the same. Computer modeling of the HAMP domain in the obtained “V”- and “U”-shaped structures revealed that only the “U”-shaped conformation allows for tight interactions of the receptor with the HAMP domain. This is in line with existing data and supports biological relevance of the “U” shape in the ground state. We suggest that the “V”-shaped structure may correspond to the active state of the complex and transition from the “U” to the “V”-shape of the receptor-transducer complex can be involved in signal transduction from the receptor to the signaling domain of NpHtrII.

Rfree was calculated for 5% of observed reflections, omitted from the refinement and Rwork calculation and picked randomly within thin resolution shells.

Supplementary Video 1
The movie clip shows the transition in the complex structure according to the normal mode 3 in line with the suggested mechanism of signal propagation.

Protein preparation
The coding regions of the N. pharaonis SRII and C-terminal truncated transducer  genes were cloned into a pET27bmod expression vector in frame with a C-terminal His 7 tag.
Proteins were expressed in E. coli strain BL21(DE3), and purified as described 1 . After imidazole removal by DEAE chromatography, SRII-His and HtrII-157-His were mixed in a 1:1 molar ratio, followed by the reconstitution into purple membrane (the bacteriorhodopsin containing membrane patches of H. salinarum) lipids (protein to lipid ratio 1:35) using BioBeads II for detergent removal.
The reconstituted proteins were pelleted by centrifugation at 100,000g and solubilized with 2% noctyl-b-D -glucopyranoside for 16 h at 4° C in the dark. The solubilized complex was used for crystallization after removal of insoluble material by centrifugation at 100,000g.
Protein was expressed and purified according to Impact kit (New England Biolabs) protocols.

Trapping of the M intermediate
To mW×mm -2 ), then cooled (by unblocking the cryostream) while the blue light was still on. One second after cooling started, the illumination was turned off. X-ray data were collected in the dark. Starting from a polyalanine model of NpSRII/NpHtrII (Protein Data Bank accession code 1H2S) the molecular replacement solution was completed by the automated refinement procedure (ARP/wARP 6 ) using ground state data. The data refinement was performed using Refmac5 7 and phenix.refine 8

Modeling of the HAMP domain
In order to better understand the influence of the crystal packing on the NpHtrII structure, we modeled its possible continuation to the first HAMP domain (residues 83-136), which is not observed in the crystal. It is assumed that the linker between the transmembrane helix TM2 and the HAMP domain adopts an alpha-helical conformation (for more details see previous reports 10 ). The HAMP domain was modeled by homology, based on the NMR structure of the Archeoglobus fulgidis hypothetical protein Af1503 (PDB code 2ASW). The transmembrane part of the NpHtrII with the bound NpSRII receptor were taken from the crystal structure. The combined model was minimized using a symmetry module from SAMSON modeling package 11 in the corresponding symmetry space group, with alpha-helical constraints imposed on NpHtrII residues 78-87. All water molecules, fragments of lipids and cofactors found in the crystal structure were preserved during the structure optimization. Crystal packing was maintained with implicit crystallographic symmetry transforms applied according to the considered space group. Interactions within modeled molecules and its symmetrical replicas were computed using a smooth version of CHARM19 force field 12 . A distance cutoff value of 8 Å was used during the structure optimization. The optimization was performed using the steepest descent method in the dihedral subspace with random jumps from the local minima according to the Metropolis acceptance criterion. The local minimum was defined as a structure with all accelerations on its dihedral degrees of freedom smaller then 1x10-2 Å/fs2.
Finally, the structure with the lowest energy was chosen as a final model. We used PyMOL software to calculate vacuum electrostatics potentials on molecular surfaces, as well as to produce illustrations for this study 13 .