Virtual Screening of Potential RoxS Inhibitors and Evaluation of Their Antimicrobial Activity in Combination with Antibiotics against Clinically Resistant Bacteria

Abstract

Pseudomonas aeruginosa with difficult-to-treat resistance has been designated as an urgent or serious threat by the CDC in the United States; therefore, novel antibacterial drugs and combination strategies are urgently needed. The sensor kinase RoxS is necessary for the aerobic growth of Pseudomonas aeruginosa. This study aimed to screen candidate RoxS inhibitors and evaluate their efficacy in treating multi-drug-resistant and extensively drug-resistant Pseudomonas aeruginosa in combination with meropenem and amikacin to identify promising combination strategies. RoxS protein structures were constructed using homology modeling and potential RoxS inhibitors, including Ezetimibe, Deferasirox, and Posaconazole, were screened from the FDA-approved ZINC drug database using molecular docking and molecular dynamics simulations. MIC and checkerboard assays were used to determine the in vitro antimicrobial efficacy of the three drugs in combination with antibiotics. The results of in vitro experiments showed an additive effect of 100 μg/mL Deferasirox or 16 μg/mL Posaconazole in combination with meropenem and a synergistic effect of 1.5 μg/mL Deferasirox and amikacin. In summary, these three drugs are potential inhibitors of RoxS, and their combination with meropenem or amikacin is expected to reverse the resistance of P. aeruginosa, providing new combination strategies for the treatment of clinically difficult-to-treat Pseudomonas aeruginosa.

Keywords: Pseudomonas aeruginosa; RoxS; checkerboard assay; homology modeling; molecular docking; molecular dynamics simulations.

Figure A1

Structural formula of compounds with TOP 10 docking scores(A–J). The numbering of the compounds was consistent with that of Table 2.

Figure A2

Two-dimensional interaction diagrams of compounds with TOP 10 docking scores (A–J). The purple arrows represent hydrogen bonding interactions and the green lines represent Pi-Pi stacking.

Figure 1

Structures of RoxS protein predicted using Swiss-Model (A–F).

Figure 2

Structures of RoxS protein predicted using I-TASSER (A–E).

Figure 3

Structure of RoxS protein predicted using Alphafold. Model Confidence: Dark blue: Very high (pLDDT > 90), light blue: Confident (90 > pLDDT > 70), yellow: Low (70 > pLDDT > 50); Orange: Very low (pLDDT < 50). AlphaFold produces a per-residue confidence score (pLDDT) between 0 and 100. Some regions with low pLDDT may be unstructured in isolation.

Figure 4

Quality evaluation results of Model 1 provided by I-TASSER: (A) Overall quality factor of ERRAT; (B) PROSA-web Z-scores of all protein chains; (C) PROCHECK analysis of the Ramachandran plot. The Overall quality factor of ERRAT > 95% means the structure has good resolution. The distribution of >90% of the residues in the most favorable region of the Ramachandran plot indicates that the protein structure is reasonable. The energy of the target protein structure is considered reasonable if the Z-score values of the target protein are distributed within the region of the graph represented by the Z-score values of these known proteins in PROSA.

Figure 5

Quality evaluation results of Model provided by Alphafold: (A) Overall quality factor of ERRAT; (B) PROSA-web Z-scores of all protein chains; (C) PROCHECK analysis of the Ramachandran plot. The Overall quality factor of ERRAT > 95% means the structure has good resolution. The distribution of >90% of the residues in the most favorable region of the Ramachandran plot indicates that the protein structure is reasonable. The energy of the target protein structure is considered reasonable if the Z-score values of the target protein are distributed within the region of the graph represented by the Z-score values of these known proteins in PROSA.

Figure 6

Two-dimensional interaction diagram of coenzyme Q1 and residues around pocket 2 (A), coenzyme Q1 and residues around pocket-GGXXNFP (B).

Figure 7

RMSD of protein backbone and ligand: (A) Protein Backbone fit to Backbone, (B) ligand fit to Backbone. EB, Ezetimibe; DSX, Deferasirox; PCZ, Posaconazole; RMSD, Root Mean Square Deviation.

Figure 8

Statistical plot of ligand–protein interaction: (A) Ezetimibe; (B) Deferasirox; (C) Posaconazole.

Figure 9

Combined antibacterial activity of Ezetimibe and meropenem (A), Deferasirox and meropenem (B), Posaconazole and meropenem (C). Legends: The vertical bar on each data point represents the standard error of the mean. The mean value of each group is located above the error bars. EB, Ezetimibe; DSX, Deferasirox; PCZ, Posaconazole; MEM, meropenem. * indicates p  ≤  0.05, ** p  ≤  0.01, *** p  ≤  0.001.

Figure 10

Combined antibacterial activity of Ezetimibe and Amikacin (A), Deferasirox and Amikacin (B), Posaconazole and Amikacin (C). Legends: The vertical bar on each data point represents the standard error of the mean. The mean value of each group is located above the error bars. EB, ezetimibe; DSX, Deferasirox; PCZ, Posaconazole; AMK, Amikacin. * indicates p  ≤  0.05, ** p  ≤  0.01, *** p  ≤  0.001.