Dynamic properties of dipeptidyl peptidase III from Bacteroides thetaiotaomicron and the structural basis for its substrate specificity – a computational study

Abstract

Dipeptidyl peptidase III (DPP III) from the human gut symbiont Bacteroides thetaiotaomicron (Bt) is the first identified prokaryotic DPP III orthologue. It has low sequence similarity to the thoroughly studied human DPP III, and differently from eukaryotic orthologues it has a cysteine (Cys450) residue in the zinc-binding motif HEXXGH (HECLGH). The recently determined crystal structure of BtDPP III showed that its 3D structure, similar to the structure of the human DPP III, consists of two domains with a wide cleft in between. Although such a striking similarity of the 3D structures of orthologues with low sequence similarity is not surprising, it is no guarantee for similarity of their dynamic properties and the catalytic performance. Here, we report the results of the molecular modelling study of BtDPP III, wild type and its C450S mutant, as well as their complexes with characteristic DPP III substrates Arg-Arg-2-naphthylamide (RRNA) and Lys-Ala-2-naphtylamide (KANA). During several hundred nanoseconds of all-atom MD simulations of the wild type protein, the long range conformational changes, which can be described as protein 'closing', have been traced. We have determined a similar conformational change for the human orthologue as well. However, the amplitude of the change is lower for BtDPP III than for the human DPP III. The MD simulations have been performed using ff03, ff12SB and ff14SB force fields wherein the results of the last two better fit to the experimental results. The hydrogen bond analysis indicates reasons for higher substrate specificity of BtDPP III towards RRNA than KANA as well as for the decrease of the RRNA hydrolysis rate induced by the Cys450 to Ser mutation. The obtained results are in line with the experimental data.

Oxidase or peptidase? A computational insight into a putative aflatoxin oxidase from Armillariella tabescens

“…However, the long range conformational changes have not been traced during these additional 200 ns either (Supporting Information Figure S6). The structure of individual domains is conserved throughout the simulations (Supporting Information Figure S7), as previously reported for human and bacterial orthologues …”

“…Such behavior is inconsistent with all currently known crystallographic structures of AFO, as well as with available crystallographic structures of the other DPP III family members, and suggests an overestimate of electrostatic interactions (Supporting Information Figure S2). On other hand, the tetrahedral model developed by Pang reproduced the active site architecture previously reported within the DPP III family to a certain degree . However, within this model the active site was too rigid and the coordination with more than four ligands was not possible (Supporting Information Figure S2), which is inconsistent with the DPP III reaction mechanism requirements .…”

“…The crystal structure of the AFO (PDB code: 5YFB) was used as the starting point for MD simulations. Residues Met1–Val6, missing in the structure, have not been modeled since we assumed that they belong to the signaling peptide, in analogy with other DPP III orthologues …”

“…Instead the RRNA center of mass was pulled away from the zinc ion from the initial distance of 3.8 Å up to 30 Å using a force constant of 10 kcal mol −1 Å −1 over the course of 200 ns. This time period has been determined long enough for large conformational motions to occur in case of human and Bacteroides thetaiotaomicron orthologues . All other simulation parameters were identical to those in cMD simulations.…”

“…RMSD between the crystal structures of AFO and the closed structure of the human DPP III (PBD code: 3T6B) is about 1.0 Å. Both crystallographic and computational studies on DPPs III prove the existence of another conformation of the enzyme, the so‐called open form, with the two domains separated by a wide cleft . It is important to note that, in the case of the human orthologue, the open form is not catalytically active; both open and closed form exist in the solution, with the balance shifting toward the catalytically active closed form in the presence of a substrate .…”