Sequencing and genetic variation of multidrug resistance plasmids in Klebsiella pneumoniae

Abstract

Background: The development of multidrug resistance is a major problem in the treatment of pathogenic microorganisms by distinct antimicrobial agents. Characterizing the genetic variation among plasmids from different bacterial species or strains is a key step towards understanding the mechanism of virulence and their evolution.

Results: We applied a deep sequencing approach to 206 clinical strains of Klebsiella pneumoniae collected from 2002 to 2008 to understand the genetic variation of multidrug resistance plasmids, and to reveal the dynamic change of drug resistance over time. First, we sequenced three plasmids (70 Kb, 94 Kb, and 147 Kb) from a clonal strain of K. pneumoniae using Sanger sequencing. Using the Illumina sequencing technology, we obtained more than 17 million of short reads from two pooled plasmid samples. We mapped these short reads to the three reference plasmid sequences, and identified a large number of single nucleotide polymorphisms (SNPs) in these pooled plasmids. Many of these SNPs are present in drug-resistance genes. We also found that a significant fraction of short reads could not be mapped to the reference sequences, indicating a high degree of genetic variation among the collection of K. pneumoniae isolates. Moreover, we identified that plasmid conjugative transfer genes and antibiotic resistance genes are more likely to suffer from positive selection, as indicated by the elevated rates of nonsynonymous substitution.

Conclusion: These data represent the first large-scale study of genetic variation in multidrug resistance plasmids and provide insight into the mechanisms of plasmid diversification and the genetic basis of antibiotic resistance.

Figure 1. Antibiotic-resistance profiling of 140 Klebsiella pneumoniae isolates from S1 and S2.

K. pneumoniae resistant to antibiotics is shown in dark; K. pneumoniae sensitive to antibiotics is shown in white; those on the threshold are shown in grey.

Figure 2. Comparison of pKF3 plasmids and two other plasmids (p3 and p4) from K. pneumoniae subsp. pneumoniae MGH 78578.

Mauve alignment of pKF3-140 and p3 (A); pKF3-94 and p4 (B). The colored boxes represent homologous segments completely free of genomic rearrangements. These homologous regions are connected by lines between genomes. Blocks below the center line indicate regions with inverse orientation. Regions outside blocks lack homology between genomes. Within each block there is a similarity profile of the nucleotide sequences, and white regions indicate the sequences specific to a genome.

Figure 3. The mapping of Illumina sequencing reads of S1 and S2 to the three reference plasmids, pKF3-70, pKF3-94, and pKF3-140.

Certain ORFs with elevated copy numbers are labeled.

Figure 4. Detection of SNPs using the Illumina sequencing data of S1.

A fraction of reads (from 2% to 100%) were randomly sampled from S1, and then were used to identify SNPs.

Figure 5. Comparisons of the minimal allele frequency (MAF) of shared SNPs (left) and unique SNPs (right) in S1 and S2.

Figure 6. The coevolving SNPs identified in pKF3-94.

(a) Three possible combinations of adjacent SNPs are illustrated by aligned sequences, in which the combination of two alleles indicates a pair of SNPs should be coevolved. (b) The relative frequency of three combinations of adjacent SNPs in S1. (c) The number of coevolving SNPs identified in S1 and S2.