Molecular Epidemiology of Mutations in Antimicrobial Resistance Loci of Pseudomonas aeruginosa Isolates from Airways of Cystic Fibrosis Patients

Abstract

The chronic airway infections with Pseudomonas aeruginosa in people with cystic fibrosis (CF) are treated with aerosolized antibiotics, oral fluoroquinolones, and/or intravenous combination therapy with aminoglycosides and β-lactam antibiotics. An international strain collection of 361 P. aeruginosa isolates from 258 CF patients seen at 30 CF clinics was examined for mutations in 17 antimicrobial susceptibility and resistance loci that had been identified as hot spots of mutation by genome sequencing of serial isolates from a single CF clinic. Combinatorial amplicon sequencing of pooled PCR products identified 1,112 sequence variants that were not present in the genomes of representative strains of the 20 most common clones of the global P. aeruginosa population. A high frequency of singular coding variants was seen in spuE, mexA, gyrA, rpoB, fusA1, mexZ, mexY, oprD, ampD, parR, parS, and envZ (amgS), reflecting the pressure upon P. aeruginosa in lungs of CF patients to generate novel protein variants. The proportion of nonneutral amino acid exchanges was high. Of the 17 loci, mexA, mexZ, and pagL were most frequently affected by independent stop mutations. Private and de novo mutations seem to play a pivotal role in the response of P. aeruginosa populations to the antimicrobial load and the individual CF host.

FIG 1

Frequency of sequence variants detected in the 17 target loci of 361 P. aeruginosa CF airway isolates sorted in batches of six bars each. Groups of bars show the total number of SNPs, number of synonymous SNPs, number of nonsynonymous SNPs, and number of premature stop mutations. t, total number of SNPs; yes, present in the reference pangenome; no, absent in the pangenome; u, unique (singular SNPs found in one strain).

FIG 2

Ratio of nonsynonymous to synonymous SNPs in the 17 loci, shown for the total number of SNPs (closed bars) and the proportion of unique SNPs identified in a single strain (open bars). The N/S ratio of 0.25 was determined by genome-wide comparison of SNPs between pairs of clonally unrelated strains (27).

FIG 3

Normalized presentation of the SNP frequency in the 17 loci for SNPs and nonsynonymous SNPs (N-SNPs), shown for all SNPs (upper panel) and unique SNPs identified in a single strain (lower panel).

FIG 4

Schematic diagram of the functions of the 17 targets in the P. aeruginosa cell. ABs, antibiotics; TIIISS, type III secretion system.

FIG 5

Left, AmpD amino acid sequence alignments of the orthologs in P. aeruginosa and Citrobacter freundii. Functionally relevant sequence motifs that account for different conformational states, formation of the stabilizing salt bridge, and the substrate binding site are depicted in orange in the C. freundii sequence. Positions with amino acid substitutions detected in the P. aeruginosa CF strain collection are highlighted in red in the P. aeruginosa sequence, with the changes in the three motifs marked in bold. Right, salt bridge interaction network in inactive (A) and active (B) C. freundii AmpD. Amino acid mutations affecting the salt bridge identified in the P. aeruginosa orthologs of CF strains are indicated in red. (The three-dimensional structure of C. freundii AmpD is republished from reference with permission of the publisher.)

FIG 6

Mapping of amino acid sequence variants found in P. aeruginosa CF airway isolates onto the ribbon diagram of the crystal structure of the dimeric repressor MexZ (51). Monomers are color coded from N (blue) to C (red). The DNA binding domains consist of α-helices 1 to 3 (blue). α-helix 3 is the DNA recognition sequence (circled in red), and α-helices 4 and 6 are in direct contact with the DNA binding domain. α-helices 7 and 9 mediate the contact between the two monomers. (The three-dimensional structure of P. aeruginosa MexZ is republished from reference with permission of the publisher.)