Identification and characterisation of Klebsiella pneumoniae and Pseudomonas aeruginosa clinical isolates with atypical β-lactam susceptibility profiles using Orbitrap liquid chromatography-tandem mass spectrometry

There is significant interest in the possibility of predicting antibacterial drug susceptibility directly though the analysis of bacterial DNA or protein. We report the use of Klebsiella pneumoniae, Escherichia coli, Pseudomonas aeruginosa and Acinetobacter baumannii transformants to define baseline predictive rules for the β- lactam susceptibility profiles of β-lactamase positive clinical isolates. We then deployed a robust and reproducible shotgun proteomics methodology to identify β-lactamase positivity and predict β-lactam susceptibility by reference to our baseline predictive rules both in cultured bacteria and in extracts of culture-positive blood. Proteomics and whole genome sequencing then allowed us to characterise K. pneumoniae and P. aeruginosa isolates that differed from the expected β-lactam susceptibility profile, iteratively expanding our predictive rules. Proteomics added considerable value over and above the information generated by whole genome sequencing, allowing for gene expression, not just gene presence to be considered. Specifically, in K. pneumoniae, we identified key differences between acrR and ramR regulatory mutations and compared the effects of OmpK36 Aspartate-Threonine or Glycine-Aspartate dipeptide porin insertions on susceptibility to cefepime and carbapenems. In P. aeruginosa, we identified differences in the gene expression effects of mexR versus nalC mutations and related these to differences in β-lactam MICs against isolates hyper-producing AmpC β-lactamase and or producing a metallo-β-lactamase.