Converging PMF Calculations of Antibiotic Permeation across an Outer Membrane Porin with Subkilocalorie per Mole Accuracy

The emergence of multidrug-resistant pathogens led toa criticalneed for new antibiotics. A key property of effective antibioticsagainst Gram-negative bacteria is their ability to permeate throughthe bacterial outer membrane via transmembrane porin proteins. Moleculardynamics (MD) simulations are, in principle, capable of modeling antibioticpermeation across outer membrane porins (OMPs). However, owing tosampling problems, it has remained challenging to obtain convergedpotentials of mean force (PMFs) for antibiotic permeation across OMPs.Here, we investigated the convergence of PMFs along a single collectivevariable aimed at quantifying the permeation of the antibiotic fosmidomycinacross the OprO porin. We compared standard umbrella sampling (US)with three advanced flavors of the US technique: (i) Hamiltonian replicaexchange with solute tempering in combination with US, (ii) simulatedtempering-enhanced US, and (iii) replica-exchange US. To quantifythe PMF convergence and to reveal hysteresis problems, we computedseveral independent sets of US simulations starting from pulling simulationsin the outward and inward permeation directions. We find that replica-exchangeUS in combination with well-chosen restraints is highly successfulfor obtaining converged PMFs of fosmidomycin permeation through OprO,reaching PMFs converged to subkilocalorie per mole accuracy.

Automated summary

The emergence of multidrug-resistant pathogens calls for the development of new antibiotics. Effective antibiotics against Gram-negative bacteria must be able to permeate the bacterial outer membrane via transmembrane porin proteins. However, obtaining accurate simulation results for antibiotic permeation across outer membrane porins has been challenging due to sampling problems.