Homology Model of a Catalytically Competent Bifunctional Rel Protein

Abstract

Bacteria have developed different bet hedging strategies to survive hostile environments and stressful conditions with persistency being maybe the most elegant yet still poorly understood one. Persisters' temporary tolerance to antibiotic treatment hints at their role not only in chronic and recurrent infections but also in the insurgence of resistant strains. Therefore, hampering persisters formation might represent an innovative strategy in the quest for new effective antimicrobial compounds. Among the molecular mechanisms postulated for the persister phenotypic switch, we decided to focus our attention on the stringent response and, in particular, on the upstream triggering step that is the accumulation of guanosine tetra- and pentaphosphate, collectivity called (p)ppGpp. Intracellular levels of (p)ppGpp are regulated by a superfamily of enzymes called RSH (RelA/SpoT homologue) that are able to promote its synthesis via pyrophosphate transfer from an ATP molecule to the 3' position of either GDP or GTP. These enzymes are classified based on the structural domain(s) present (only synthetase, only hydrolase, or both). Here we present our work on Rel _Seq (from S. equisimilis), still the only bifunctional Rel protein for which a GDP-bound "synthetase-ON" structure is available. Analysis of the synthetase site, occupied only by GDP, revealed a partially active state, where the supposed ATP binding region is not conformationally apt to accommodate it. In order to achieve a protein model that gets closer to a fully active state, we generated a chimera structure of Rel _Seq by homology modeling, starting from the crystal structure of the catalytically competent state of RelP, a smaller, single-domain, Rel protein from S. aureus. Molecular dynamics simulations allowed verifying the stability of the generated chimera model. Virtual screening and ligand design studies are underway.

Keywords: (p)ppGpp; RelP; RelSeq; bacterial persisters; chimera; homology modeling; molecular dynamics.

FIGURE 1

Rel_Seq X-ray crystal structure and the (p)ppGpp synthesis reaction it catalyses. (A) Rel_Seq crystal structure (1VJ7. pdb, chain A, residues 5–341 are shown) in complex with GDP and Mn²⁺ (blue sphere). (B) p)ppGpp is obtained by enzymatic transfer of a PP group from ATP to the 3′-OH group of either GDP or GTP.

FIGURE 2

S. aureus RelP small alarmone synthetase and its postulated reaction mechanism (A) RelP pre-catalytic state (6EWZ.pdb) in complex with GTP, AMP-CPP and Mg²⁺ (gray sphere). (B) Proposed reaction mechanism for pppGpp synthesis.

FIGURE 3

Structure superimposition of the SYNTH domains of Rel_Seq and RelP. (A) Superimposition of Rel_Seq SYNTH domain (gray ribbons, 1VJ7. pdb chain A, residues 197–341, with GDP ligand) and RelP (yellow ribbons, 6EWZ.pdb, chain A, residues 29–188). (B) Catalytic residues of both enzymes are shown (D264, E323 of Rel_Seq in gray, D107, E174 of RelP in red). Distances between the side chains are reported as calculated on the C atoms of the carboxylate groups. Mg²⁺ is shown as a red sphere with the two coordinated waters in blue, AMP-CPP is in white tube representation while GTP is omitted for clarity.

FIGURE 4

Chimera models generated. The five chimera models generated by homology modeling are overlaid. Blue: coordinates taken from the template; yellow: coordinates taken from the template with optimized side chains; red: rebuilt residues. GTP, AMP-CPP, and the Mg²⁺ ion (blue sphere) are shown within the SYNTH binding site.

FIGURE 5

Analysis of the chimera protein stability over the MD simulation time. RMSD was calculated on protein backbone atoms (C, Cα, N, O, H) for the three independent runs with respect to the refined chimera model as a function of the simulation time. (A) all residues, (B) HD residues alone (5–159) and (C) SYNTH residues alone (178–341).

FIGURE 6

Chimera fluctuation analysis over the MD simulation time. RMSF calculated on protein backbone atoms (C, Cα, N, O, H) in the three replicas (#1 red, #2 black, #3 blue) as a function of the residue number. HD (residues 5–159), C3HB (residues 135–195) and SYNTH (residues 176–341). Peaks with the highest RMSF values are labeled.

FIGURE 7

Analysis of Rel_Seq crystal-modelled interface. Angle formed between helices α11 and α13, the interface between the unaltered SYNTH (from 1vj7. pdb) and the homology modeled portion, during replica#1 (red), #2 (black) and #3 (blue) calculated from the scalar product of the two vectors that run along the two helices.

FIGURE 8

Protein cluster analysis results for the SYNTH domain. Cα atom of residues 197–334 are shown. (A) replica #1, (B) replica #2, (C) replica #3. The representative structures of the main clusters are superimposed to the most populated cluster. RMSD values calculated relative to the most populated cluster c1 < 2 Å (Supplementary Table 3). Blue: helix α13; red spheres: Cα atoms of residues with higher RMSD values.