Bacterial sortases are cysteine transpeptidases that regulate the covalent linkage of several surface protein virulence factors in Gram-positive bacteria. Virulence factors play significant roles in adhesion, invasion of host tissues, biofilm formation and immune evasion, mediating the bacterial pathogenesis and infectivity. Therefore, sortases are emerging as important targets for the design of new anti-infective agents. We employed a computational study, based on structure derived descriptors and molecular fingerprints, in order to develop simple classification methods which could allow predicting low active or high active SrtA inhibitors. Our results indicate that a highly active SrtA inhibitor has a molecular weight ranging between 180 and 600, contains one up to four nitrogen atoms, up to three oxygen atoms and under 18 hydrogen atoms. Also the hydrogen acceptor number and the molecular flexibility, as assessed by the number of rotatable bounds, have emerged as the most relevant descriptors for SrtA affinity. The Bemis-Murcko scaffolding revealed favoured scaffolds as containing at least two ring structures bonded directly or merged in a condensed cycle. This data represent a valuable tool for identifying new potent SrtA inhibitors, potential anti-virulence agents targeted against Gram-positive bacteria, including multiresistant strains.
Keywords: Bemis-Murcko scaffolding; Gram-positive bacteria; anti-virulence drugs; classification method; cysteine transpeptidases; hydrogen bond acceptors; rotatable bonds.