Structural and computational analysis of the Escherichia coli chaperone protein DmsD

In Gram-negative bacteria, the secretion of proteins that contain redox cofactors is accomplished using the twin arginine translocase (TAT) system, so named because the cofactor containing secretory proteins contain an N-terminal leader peptide with a twin arginine motif. The redox enzyme maturation proteins (REMPs) are molecular chaperones that prevent TAT substrate translocation until the preprotein is folded and its cofactor is incorporated. REMPs then direct the substrate to the TAT translocase. The work presented here explores, from a structural biology perspective, a model REMP: Escherichia coli DmsD. DmsD was crystallized and the structure determined and refined to 2.0 Å resolution. This is the first structure of E. coli DmsD, and contains clear electron density for all 204 amino acid residues in the protein molecule. This was complemented by NMR analysis that characterized the local backbone dynamics of the protein. The dynamic properties of DmsD were also explored by molecular dynamics simulation. These analyses have identified three flexible regions of DmsD, two of which contribute to the putative leader peptide binding site. The third flexible region is located in a patch of residues that were implicated in GTP binding in the homologue TorD. A method for the purification of active TAT leader peptides was devised and used in the generation of a selectively labeled sample for NMR analysis. Chemical shift perturbation analyses are consistent with a hydrophobic groove on the surface of DmsD interacting with the DmsA leader peptide. DmsA is the cofactor containing catalytic subunit of DMSO reductase and the specific substrate of DmsD. The leader peptide binding groove on DmsD overlaps with the previously identified “hot pocket” which is predicted to interact with the twin-arginine motif of the leader peptide, and encompasses regions of the x-ray structure that form crystal contacts with crystallization reagents. Finally, a new model for TAT leader peptide binding is proposed which combines all of the available data.