| Abstract |
While ~30% of the human genome encodes membrane proteins, only a handful of structures of membrane proteins have been resolved to high resolution. Here, we studied the structure of a member of the Cys-loop ligand gated ion channel protein superfamily of receptors, human type A γ2α1β2α1β2 gamma amino butyric acid receptor complex in a lipid bilayer environment. Studying the correlation between the structure and function of the gamma amino butyric acid receptor may enhance our understanding of the molecular basis of ion channel dysfunctions linked with epilepsy, ataxia, migraine, schizophrenia and other neurodegenerative diseases. The structure of human γ2α1β2α1β2 has been modeled based on the X-ray structure of the Caenorhabditis elegans glutamate-gated chloride channel via homology modeling. The template provided the first inhibitory channel structure for the Cys-loop superfamily of ligand-gated ion channels. The only available template structure before this glutamate-gated chloride channel was a cation selective channel which had very low sequence identity with gamma aminobutyric acid receptor. Here, our aim was to study the effect of structural corrections originating from modeling on a more reliable template structure. The homology model was analyzed for structural properties via a 100 ns molecular dynamics (MD) study. Due to the structural shifts and the removal of an open channel potentiator molecule, ivermectin, from the template structure, helical packing changes were observed in the transmembrane segment. Namely removal of ivermectin molecule caused a closure around the Leu 9 position along the ion channel. In terms of the structural shifts, there are three potential disulfide bridges between the M1 and M3 helices of the γ2 and 2 α1 subunits in the model. The effect of these disulfide bridges was investigated via monitoring the differences in root mean square fluctuations (RMSF) of individual amino acids and principal component analysis of the MD trajectory of the two homology models-one with the disulfide bridge and one with protonated Cys residues. In all subunit types, RMSF of the transmembrane domain helices are reduced in the presence of disulfide bridges. Additionally, loop A, loop F and loop C fluctuations were affected in the extracellular domain. In cross-correlation analysis of the trajectory, the two model structures displayed different coupling in between the M2-M3 linker region, protruding from the membrane, and the β1-β2/D loop and cys-loop regions in the extracellular domain. Correlations of the C loop, which collapses directly over the bound ligand molecule, were also affected by differences in the packing of transmembrane helices. Finally, more localized correlations were observed in the transmembrane helices when disulfide bridges were present in the model. The differences observed in this study suggest that dynamic coupling at the interface of extracellular and ion channel domains differs from the coupling introduced by disulfide bridges in the transmembrane region. We hope that this hypothesis will be tested experimentally in the near future.
|
