Unveiling Berberine analogues as potential inhibitors of Escherichia coli FtsZ through machine learning molecular docking and molecular dynamics approach

Abstract

The bacterial cell division protein FtsZ, a crucial GTPase, plays a vital role in the formation of the contractile Z-ring, which is essential for bacterial cytokinesis. Consequently, inhibiting FtsZ could prevent the formation of proto-filaments and interfere with the cell division machinery. The remarkable conservation of FtsZ across diverse bacterial species makes it a promising drug target for combating drug resistance. In the present study, 1072 berberine analogues were screened for favorable pharmacokinetic properties. A total of 60 compounds that fulfilled the drug-likeliness criteria and were found to be non-toxic were selected for virtual screening against Escherichia coli FtsZ protein (PDB ID: 8GZY). Molecular docking revealed a strong binding affinity of ZINC000524729297 (- 8.73 kcal/mol) and ZINC000604405393 (and - 8.55 kcal/mol) with FtsZ by strong intermolecular hydrogen bonds and hydrophobic interactions. Subsequently, the docking profiles were validated through a 500 ns MD simulation and MMPBSA analysis of the FtsZ-ligand complexes. The analysis revealed the FtsZ- ZINC524729297 and FtsZ-ZINC000604405393 complexes had the lowest root-mean-square deviation with lowest binding energy and enhanced conformational stability in a dynamic environment. These findings suggest that ZINC524729297 and ZINC000604405393 are the potent lead compound that targets FtsZ and requires further experimental validation.

Keywords: Escherichia coli; Antimicrobial resistance; FtsZ; Machine learning; Molecular docking; Molecular dynamics simulation.

Fig. 1

Flowchart depicting the virtual screening process for identifying effective FtsZ inhibitors.

Fig. 2

Statistical performance of various classifiers for development of machine learning model in the training dataset.

Fig. 3

Molecular docking interaction of ligands with FtsZ (a) FtsZ-ZINC524729297, (b) FtsZ-ZINC000604405393, (c) FtsZ-ZINC000072312902, (d) FtsZ-ZINC000085341281, (e) FtsZ-DB08387, (f) FtsZ-Berberine.

Fig. 4

Molecular dynamics simulation analysis of protein–ligand (a) Root means square deviation (RMSD), (b) Root mean square fluctuation (RMSF) associated with the number of residue, (c) Intermolecular H-bond, (d) Radius of gyration, (e) Interaction energy, (f) Solvent Accessible Surface Area (SASA).

Fig. 5

MM-PBSA analysis of individual residues contributing to total binding free energy (a) FtsZ-ZINC524729297, (b) FtsZ-ZINC000604405393, (c) FtsZ-ZINC000072312902, (d) FtsZ-ZINC000085341281, (e) FtsZ-DB08387, (f) FtsZ-Berberine.

Fig. 6

The 2D conformational projection of Principal Component Analysis (PCA) of FtsZ-ZINC524729297, FtsZ-ZINC000604405393, FtsZ-ZINC000072312902, FtsZ-ZINC000085341281, FtsZ-DB08387, FtsZ-Berberine.

Fig. 7

The graphical depiction of PCA based Free Energy Landscape (FEL) analysis of FtsZ-ligand complexes (a) FtsZ-ZINC524729297, (b) FtsZ-ZINC000604405393, (c) FtsZ-ZINC000072312902, (d) FtsZ-ZINC000085341281, (e) FtsZ-DB08387, (f) FtsZ-Berberine.

Fig. 8

Calculation of dynamics cross-correlation matrix for the FtsZ-ligand complexes based on the Cα-residue (a) FTsZ-ZINC524729297, (b) FTsZ- ZINC000604405393, (c) FTsZ-berberine.

Fig. 9

Functional analysis of molecular properties of finalized compounds and known FtsZ inhibitor.