Variable recombination dynamics during the emergence, transmission and 'disarming' of a multidrug-resistant pneumococcal clone

Abstract

Background: Pneumococcal β-lactam resistance was first detected in Iceland in the late 1980s, and subsequently peaked at almost 25% of clinical isolates in the mid-1990s largely due to the spread of the internationally-disseminated multidrug-resistant PMEN2 (or Spain6B-2) clone of Streptococcus pneumoniae.

Results: Whole genome sequencing of an international collection of 189 isolates estimated that PMEN2 emerged around the late 1960s, developing resistance through multiple homologous recombinations and the acquisition of a Tn5253-type integrative and conjugative element (ICE). Two distinct clades entered Iceland in the 1980s, one of which had acquired a macrolide resistance cassette and was estimated to have risen sharply in its prevalence by coalescent analysis. Transmission within the island appeared to mainly emanate from Reykjavík and the Southern Peninsular, with evolution of the bacteria effectively clonal, mainly due to a prophage disrupting a gene necessary for genetic transformation in many isolates. A subsequent decline in PMEN2's prevalence in Iceland coincided with a nationwide campaign that reduced dispensing of antibiotics to children in an attempt to limit its spread. Specific mutations causing inactivation or loss of ICE-borne resistance genes were identified from the genome sequences of isolates that reverted to drug susceptible phenotypes around this time. Phylogenetic analysis revealed some of these occurred on multiple occasions in parallel, suggesting they may have been at least temporarily advantageous. However, alteration of 'core' sequences associated with resistance was precluded by the absence of any substantial homologous recombination events.

Conclusions: PMEN2's clonal evolution was successful over the short-term in a limited geographical region, but its inability to alter major antigens or 'core' gene sequences associated with resistance may have prevented persistence over longer timespans.

Figure 1

Phylogenomic analysis of the sequenced isolates. (A) Maximum likelihood phylogeny of the isolates based on likely point mutations. The branches of the phylogeny are coloured according to the geographical location of the isolates, as reconstructed through the tree on the basis of maximum parsimony. The PMEN2 and PMEN22 clades are indicated by the black vertical bars; the grey shaded box marks the ‘Outgroup’ clade. Isolates from Iceland are indicated by the orange boxes highlighting clades IC1 and IC2. Isolates from Spain previously providing evidence of transmission from continental Europe are marked with single asterisks; the German isolate most closely related to clade IC1 is marked with a double asterisk. The PMEN22 isolate that has become resistant to β-lactams is marked by the † symbol. (B) Simplified annotation of the reference genome against which sequence reads were mapped. The positions of putative mobile genetic elements (MGEs) is marked, along with the capsule polysaccharide synthesis locus (cps), Pneumococcal Pathogenicity Island 1 (PPI-1), the gene for penicillin-binding protein 2B (pbp2b), and the antigen-encoding genes pspA and pspC. (C) Density of recombination events across the genome: (i) across the entire collection, (ii) within only the PMEN2 clade. (D) Distribution of recombination events. The dashed line defines a panel consisting of a row for each of the sequenced isolates in the collection, with a column for each base in the reference sequence. Putative recombination events are indicated by the coloured blocks: red events are reconstructed as occurring on internal branches and are, therefore, shared by multiple isolates through common descent, whereas blue blocks are recombinations occurring on terminal branches, which are unique to individual isolates.

Figure 2

Comparison of integrative and conjugative elements (ICE) identified in the collection. Eight examples of composite Tn5253-type ICE found in the collection are aligned: ICESp6BST273, found in PMEN22, ICESp6BST90, found in PMEN2 (examples modified by the integration of macrolide resistance cassettes are included), and four ICE representing the diversity found in the outgroup. Red bands indicate BLAT matches between sequences in the same orientation and blue twisted bands indicate BLAT matches between sequences in opposite orientations. In both cases, the intensity of the colour indicates the strength of the match. Protein coding sequences (CDSs) are indicated by pink boxes, except those encoding antibiotic resistance determinants, which are coloured dark blue. The vertical position above or below the central line indicates whether the CDS is encoded on the forward or reverse strand of the sequence. The tetM tetracycline resistance gene, found within the Tn916-type elements, is marked at the top and bottom of the alignment. The ermB macrolide resistance gene is found within Tn917 cassettes (boxed in purple throughout the alignment) or Omega cassettes (boxed in green throughout the alignment). The cat chloramphenicol acetyltransferase is carried on the linearized pC194 plasmid, indicated by blue boxes on the annotation. BLAT, BLAST-like alignment tool.

Figure 3

Temporal dynamics of clade IC1 in Iceland. (A) Bayesian skyline plot of the product of effective population size and generation time over the course of the outbreak. The median estimate of this parameter is displayed as the black line, while the 95% credibility interval is indicated by the blue filled area. The solid red line indicates the recorded numbers of penicillin non-susceptible pneumococci of serotype 6B collected from cases of disease in Iceland between 1989 and 1995. The dashed red line indicates the available figures for the number of clinical isolates meeting these criteria collected from the Capital and Southwestern areas from 1995 onwards, which account for more than 90% of the penicillin non-susceptible pneumococci identified in Iceland [8]. (B) Proportion of clade IC1 isolates resistant to tetracycline, erythromycin and chloramphenicol over time. The total number of sequenced isolates from clade IC1 is represented by the black line, relative to the right side vertical axis. The proportions of these isolates resistant to selected antibiotics in each year, excluding those for which no phenotype was available or that were defined as having an ‘intermediate’ resistance phenotype [see Additional file 1: Table S1], are represented by the coloured lines relative to the left side vertical axis. (C) Use of antibiotics in Iceland. Annual defined daily doses (DDD) in Iceland for classes of antibiotics against which IC1 isolates were found to have resistance are displayed. The data in panels (B) and (C) are plotted using the midpoint of each year such that they are displayed on approximately the same timescale as the changes in the effective population size of IC1 in panel (A).

Figure 4

Transmission of clade IC1 within Iceland. (A) A phylodynamic analysis was conducted to reconstruct the transmission of the IC1 clade across Iceland. The tree is coloured according to the most likely location of each branch, as determined through the discrete states phylogeographic model described by [60]. The locations are clustered by region in the key. (B) Each isolate’s resistance profile is represented by red bars indicating full resistance, purple bars indicating intermediate levels of resistance and blue bars indicating sensitivity to the antibiotics annotated at the top of the column; white bars indicate missing data. (C) The isolates are classified according to the phenotypic groups one to six defined in [9] based on resistance phenotype and genotype, which are described in the main text.

Figure 5

Loss of resistance within clade IC1. (A) Loss of erythromycin and tetracycline resistance in group 3 isolates. (i) Plot of coverage calculated by mapping sequence reads from group 3 isolate IC210 to the reference genome Tn916-type element, showing no coverage across the transposon. This suggests it has been lost from the genome. (ii) Comparison of de novo assembly of isolate IC210 with the reference sequence. The red bands between the two sequences indicate BLAT matches between sequences in the same orientation. The annotation of the reference sequence is at the top; protein coding sequences (CDSs) are indicated by pink boxes, except for those that encode resistance determinants, which are coloured blue. The brown boxes indicate the fragments of a CDS that is reformed by the apparent excision of the transposon, as seen in the assembled sequence of isolate IC210 at the bottom of the diagram. (B) Loss of tetracycline resistance in group 2 isolates. The figure shows the structure of the tetM gene, including the positions of the promoter elements, leader peptide and ribosome binding sites (RBS). A 58 bp deletion within the leader peptide, at the bases indicated by the shaded bar, appears to result in the inactivation of the gene. (C) Loss of erythromycin resistance in group 4 isolates. The structure of the ermB gene is displayed as described for (B), with the position of the sequence alteration leading to a frameshift mutation indicated by the shaded bar. (D) Loss of chloramphenicol resistance within some group 5 isolates through deletion of the cat chloramphenicol acetyltransferase gene on the Ωcat(pC194) element. (i) Plot of mapped sequence read coverage and (ii) comparison of de novo assembly with the reference genome for group 5 isolate 0301 + 23540, displayed as in (A). BLAT, BLAST-like alignment tool.