Biophysical Society Thematic Meeting | Stockholm 2022

Physical and Quantitative Approaches to Overcome Antibiotic Resistance

Wednesday Speaker Abstracts

PREDICTING PERMEATION OF COMPOUNDS ACROSS THE OUTER MEMBRANE OF PSEUDOMONAS AERUGINOSA Gnana Gnanakaran 1 ; 1 Los Alamos National Laboratory, Theoretical Biology and Biophysics, Santa Fe, NM, USA One of the major obstacles in the antibiotic discovery pipeline is the lack of understanding on how to breach antibiotic permeability barriers of Gram-negative pathogens. We have combined mechanistic and machine learning approaches to predict outer membrane (OM) permeation in P. aeruginosa. In the first study, we rationally identify a “chemical vocabulary” specifically related to OM permeability without employing known rules. Specifically, we compute a fragment-based representation of compounds and use a combination of sparse regression and a hierarchical cleansing proceed to select a subset of relevant fragments, which are responsible for OM permeation. By synergizing theory, computation, and experiment, we are able to validate our predictions and to explain the molecular mechanism behind identified fragments promoting compound entry and select candidate compounds from an external library that permeate across OM. In the second study, we generate predictive models that identify specific molecular descriptors that can predict the likelihood that a given compound is capable of OM permeation. While part of these descriptors is computed using traditional approaches based on the physicochemical properties intrinsic to the compounds, all-atom molecular dynamics simulations are used to derive additional bacterium-specific biophysical descriptors. Specifically, properties related to enthalpy, entropy and diffusion are calculated in different subregions of the OM model of P. aeruginosa. A statistical analysis based on hierarchical clustering, rank correlations, and a random forest classifier, finds a set of biophysical descriptors with prediction accuracy of 96%. Our results show the potential to predict small molecule permeation across the OM of P. aeruginosa with high precision and accuracy.

PNEUMOCOCCAL INTERACTIONS WITH THE HOST AS A TARGET FOR THERAPY

Birgitta Henriques-Normark Karolinska Institute, Sweden No Abstract

23