Satellite Program The 47th Lorne Conference on Protein Structure and Function 2022

Harnessing RNA polymerase protein structure to predict drug resistance in tuberculosis and COVID-19 (#35)

Stephanie Portelli 1 2 , Douglas E.V. Pires 1 3 , David B. Ascher 1 2
  1. Computational Biology and Clinical Informatics, Baker Heart and Diabetes Institute, Melbourne, Victoria, Australia
  2. The Department of Biochemistry and Pharmacology, Bio21 Institute, University of Melbourne, Parkville, Victoria, Australia
  3. School of Computing and Information Systems, University of Melbourne, Parkville, Victoria, Australia

The RNA polymerase (RNAP) protein complex is responsible for gene transcription, making it essential for survival. This essentiality, particularly in micro-organisms, has made DNA- and RNA-dependent RNAP good drug targets in infections like tuberculosis, and more recently COVID-19.

In treating tuberculosis, rifampicin, which binds the RNAP β-subunit to sterically block transcription, formed part of the backbone treatment for the past 60 years. This extensive use has selected for resistance-conferring mutations within the protein target, making treatment more challenging. To address this, we computationally characterized the effects of 270 rifampicin resistant, and 49 susceptible mutations, on target stability, dynamics, and molecular interaction affinities. Our analyses highlighted that rather than affecting affinity to rifampicin, resistance mutations reduced the affinity to nucleic acids and other RNAP subunits. We used this mechanistic insight to develop a machine learning-based predictor to pre-emptively identify novel resistance mutations. Our tool, SUSPECT-RIF (StrUctural Susceptibility PrEdiCTion for RIFampicin), has outperformed the WHO-endorsed diagnostic tool by identifying mutation resistance phenotype with 90.9% accuracy, 92.2% sensitivity and 83.6% specificity. Further to that, we have validated the predictive utility of SUSPECT-RIF in M. leprae, S. aureus and P. aeruginosa infections clinically treated by rifampicin.

Based on these observations, we adopted similar techniques to pre-emptively assess the risk of resistance development to molnupiravir, a nucleic acid analogue currently in Phase III clinical trials for treating COVID-19. We analyzed variation data from 180,000 SARS-CoV-2 genomes and identified 8 NSP12 mutations at the molnupivair binding site, of which, 2 significantly lowered drug affinity. Additionally, these mutations were observed to impart a high protein fitness cost, suggesting that, despite their effect on drug affinity, their establishment within a population leading to clinical resistance is unlikely.

Our analyses on RNAP mutations across different organisms highlight the importance of understanding concomitant protein-level mechanisms for effective drug stewardship efforts.