# Rational Design and Virtual Screening of Antimicrobial Terpene-Based Leads from Marrubium vulgare Essential Oil: Structure-Based Optimization for Food Preservation and Safety Applications

## TL;DR

This study uses computational methods to design improved antimicrobial compounds from Marrubium vulgare essential oil for food preservation.

## Contribution

A novel computational framework for redesigning terpene scaffolds into selective LasB inhibitors for food safety.

## Key findings

- Thymol–Leu–Trp–phosphinic acid showed strong predicted binding affinity and selectivity over human proteases.
- Molecular dynamics simulations confirmed the stability and zinc coordination of the lead candidate.
- Phosphinic zinc-binding groups, Leu–Trp linkers, and phenolic scaffolds were identified as optimal structural features.

## Abstract

Pseudomonas aeruginosa elastase LasB accelerates refrigerated food spoilage through proteolytic degradation of muscle and milk proteins. While Marrubium vulgare essential oil terpenes exhibit antimicrobial activity, their weak potency and nonspecificity limit direct food preservation applications. This computational study aimed to rationally redesign terpene scaffolds into predicted selective LasB inhibitors. A virtual library of 635 terpene–peptide–phosphinic acid hybrids (expanded to 3940 conformers) was evaluated using consensus molecular docking (Glide/Flare) against LasB (PDB: 3DBK) and three human off-target proteases. Top candidates underwent duplicate 150 ns molecular dynamics simulations with MM/GBSA binding free-energy calculations. Computational screening identified thymol–Leu–Trp–phosphinic acid as the lead candidate with predicted binding affinity of −12.12 kcal/mol, comparable to reference inhibitor phosphoramidon (−11.87 kcal/mol), and predicted selectivity index of +0.12 kcal/mol representing a 2.3 kcal/mol advantage over human proteases. Molecular dynamics simulations indicated exceptional stability (98.7% stable frames, 0.12 Å inter-replica RMSD) with consistent zinc coordination. Structure–activity analysis revealed phosphinic zinc-binding groups (+1.57 kcal/mol), Leu–Trp linkers (+2.47 kcal/mol), and phenolic scaffolds (+1.35 kcal/mol) as predicted optimal structural features. This in silico study provides a computational framework and prioritized candidate set for developing natural product-derived food preservatives. All findings represent computational predictions requiring experimental validation through enzymatic assays, food model studies, and toxicological evaluation.