Coexistence and within-host evolution of diversified lineages of hypermutable Pseudomonas aeruginosa in long-term cystic fibrosis infections

Abstract

The advent of high-throughput sequencing techniques has made it possible to follow the genomic evolution of pathogenic bacteria by comparing longitudinally collected bacteria sampled from human hosts. Such studies in the context of chronic airway infections by Pseudomonas aeruginosa in cystic fibrosis (CF) patients have indicated high bacterial population diversity. Such diversity may be driven by hypermutability resulting from DNA mismatch repair system (MRS) deficiency, a common trait evolved by P. aeruginosa strains in CF infections. No studies to date have utilized whole-genome sequencing to investigate within-host population diversity or long-term evolution of mutators in CF airways. We sequenced the genomes of 13 and 14 isolates of P. aeruginosa mutator populations from an Argentinian and a Danish CF patient, respectively. Our collection of isolates spanned 6 and 20 years of patient infection history, respectively. We sequenced 11 isolates from a single sample from each patient to allow in-depth analysis of population diversity. Each patient was infected by clonal populations of bacteria that were dominated by mutators. The in vivo mutation rate of the populations was ∼100 SNPs/year-∼40-fold higher than rates in normo-mutable populations. Comparison of the genomes of 11 isolates from the same sample showed extensive within-patient genomic diversification; the populations were composed of different sub-lineages that had coexisted for many years since the initial colonization of the patient. Analysis of the mutations identified genes that underwent convergent evolution across lineages and sub-lineages, suggesting that the genes were targeted by mutation to optimize pathogenic fitness. Parallel evolution was observed in reduction of overall catabolic capacity of the populations. These findings are useful for understanding the evolution of pathogen populations and identifying new targets for control of chronic infections.

Figure 1. Isolate sampling points and patient life spans.

P. aeruginosa isolates were collected from two CF patients: CFA and CFD. Hollow symbols: single bacterial isolates. Solid circles: cross-sectional populations of 90 bacterial isolates. *: estimated start of chronic infection. Gray bar: patient life span.

Figure 2. Evolutionary relationships among isolates from CFA and CFD lineages.

Maximum-parsimony phylogenetic trees of CFA (A) and CFD (B) were constructed based on the accumulation of new SNPs relative to ancestors CFA_2004/01 and CFD_1991/01. Alleles of P. aeruginosa reference strain PAO1 were used to root the trees. Lengths of branches are proportional to the number of accumulated SNPs. Branches are designated by capital letters. MRCA: most recent common ancestor.

Figure 3. Minimum spanning trees (MSTs) of genomes among CFA and CFD lineages.

MSTs for CFA (A) and CFD (B) were constructed based on the total number of genes altered by nonsynonymous SNPs and indel mutations in the respective genomes. Links between nodes represents the minimum distance in terms of mutated genes. Numbers above each link indicate the total amount of mutated genes between the two connected nodes (Table S6). For tree construction, ancestors CFA_2004/01 and CFD_1991/01 were considered as origins.

Figure 4. Pathoadaptive genes convergently mutated in CFA and CFD sub-lineages.

The analysis was performed based only on non-synonymous mutated genes that were altered independently in at least half of the 10 evolving sub-lineages CFA I–IV and CFD I–VI.

Figure 5. Mutational spectra and top mutated homopolymeric G∶C SSRs in CFA_2010 and CFD_2011 contemporary isolates.

(A) Percentage of 1–4 bp insertions/deletions located in G∶C and A∶T homopolymeric sequences. (B) The heat map represents individual indels mutations in homopolymeric G∶C SSRs of ≥6 bp, which were mutated in at least half of the coexisting isolates in both CFA and CFD lineages. The color-code indicates the type of mutation. Right: Percentage of MRS-deficient isolates harboring a indel mutation in each analyzed G∶C SSR. (C) Mutations in CFA_2010 and CFD_2011 isolates were analyzed based on the percentage of transitions, transversions, and insertions/deletions.

Figure 6. Average total catabolic function of isolates from CFA and CFD lineages.

Total catabolic function was calculated relative to CFA_2004/01 and CFD_1991/01 as a weighted average across all substrates for each CFA (A) and CFD (B) isolate. Total catabolic function was defined as 1 for the reference levels (CFA_2004/01 and CFD_1991/01). Lower values indicate decay. Isolates CFA_2010/01, CFA_2010/11, and CFA_2010/31 were excluded from the analysis because significant dispersion was observed in the duplicates.