The PitA protein contributes to colistin susceptibility in Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa is an opportunistic pathogen that causes a wide range of problematic infections in individuals with predisposing conditions. Infections can be treated with colistin but some isolates are resistant to this antibiotic. To better understand the genetic basis of resistance, we experimentally evolved 19 independent resistant mutants from the susceptible laboratory strain PAO1. Whole genome sequencing identified mutations in multiple genes including phoQ and pmrB that have previously been associated with resistance, pitA that encodes a phosphate transporter, and carB and eno that encode enzymes of metabolism. Individual mutations were engineered into the genome of strain PAO1. Mutations in pitA, pmrB and phoQ increased the minimum inhibitory concentration (MIC) for colistin 8-fold, making the bacteria resistant. Engineered pitA/phoQ and pitA/pmrB double mutants had higher MICs than single mutants, demonstrating additive effects on colistin susceptibility. Single carB and eno mutations did not increase the MIC suggesting that their effect is dependent on the presence of other mutations. Many of the resistant mutants had increased susceptibility to β-lactams and lower growth rates than the parental strain demonstrating that colistin resistance can impose a fitness cost. Two hundred and fourteen P. aeruginosa isolates from a range of sources were tested and 18 (7.8%) were colistin resistant. Sequence variants in genes identified by experimental evolution were present in the 18 resistant isolates and may contribute to resistance. Overall our results identify pitA mutations as novel contributors to colistin resistance and demonstrate that resistance can reduce fitness of the bacteria.

Fig 1. Colistin MICs and growth of experimentally evolved mutants. A.

MIC₅₀ of the experimentally evolved mutants. Biological replicates are displayed as individual data points, with black lines indicating the median values. B. Growth compared to MIC₅₀ for each mutant. Growth (area under the curve) was the average of three biological replicates and is shown as a percentage of growth of parental strain PAO1.

Fig 2. Colistin MICs of transposon and engineered mutants. A.

MIC₅₀ of mutants with transposon insertions. The transposon insertion sites are shown. B. MIC₅₀ of mutants with mutations engineered in single genes and of evolved mutants with single mutations in pmrB. C. MIC₅₀ of multi-gene mutants. Biological replicates are displayed as individual data points, with black lines indicating the median values.

Fig 3. Schematic model for changes to LPS that reduce susceptibility to colistin. A.

In wild-type bacteria, negatively charged carbamoyl and phosphate groups interact with colistin, concentrating the antibiotic at the cell surface. CarB catalyses synthesis of carbamoyl phosphate, the substrate for the PA5005 (WapO) carbamoyl transferase enzyme that adds a carbamoyl group to a heptose residue in the core oligosaccharide. B. Mutations in phoQ or pmrAB result in increased synthesis of Arn proteins that catalyse addition of L-Ara4N to phosphate groups of lipid A, masking the negative charge and reducing affinity for colistin. C. Mutations in pitA reduce the availability of phosphate in lipid A and for decoration of heptose sugars in the inner core of lipopolysaccharide, reducing affinity for colistin. D. Mutations in carB or PA5005 prevent incorporation of carbamoyl groups to the inner core, reducing affinity for colistin.