Genome-wide dissection of globally emergent multi-drug resistant serotype 19A Streptococcus pneumoniae

Abstract

Background: Emergence of multi-drug resistant (MDR) serotype 19A Streptococcus pneumoniae (SPN) is well-documented but causal factors remain unclear. Canadian SPN isolates (1993-2008, n = 11,083) were serotyped and in vitro susceptibility tested. A subset of MDR 19A were multi-locus sequence typed (MLST) and representative isolates' whole genomes sequenced.

Results: MDR 19A increased in the post-PCV7 era while 19F, 6B, and 23F concurrently declined. MLST of MDR 19A (n = 97) revealed that sequence type (ST) 320 predominated. ST320 was unique amongst MDR 19A in that its minimum inhibitory concentration (MIC) values for penicillin, amoxicillin, ceftriaxone, and erythromycin were higher than for other ST present amongst post-PCV7 MDR 19A. DNA sequencing revealed that alleles at key drug resistance loci pbp2a, pbp2x, pbp2b, ermB, mefA/E, and tetM were conserved between pre-PCV7 ST 320 19F and post-PCV7 ST 320 19A most likely due to a capsule switch recombination event. A genome wide comparison of MDR 19A ST320 with MDR 19F ST320 identified 822 unique SNPs in 19A, 61 of which were present in antimicrobial resistance genes and 100 in virulence factors.

Conclusions: Our results suggest a complex genetic picture where high-level drug resistance, vaccine selection pressure, and SPN mutational events have created a "perfect storm" for the emergence of MDR 19A.

Figure 1

Serotype trends amongst multi-drug resistant (MDR) strains obtained from the Canadian Bacterial Surveillance Network between 1993 and 2008 (n = 11,083). MDR 19F (n = 477), 23F (n = 150), and 6B (n = 221) emerged in the pre- PCV 7 introduction era (before 2001) and continued to rise during vaccine introduction (2002-2005), then declined in post-PCV7 introduction era (2006 onwards). MDR 19A (n = 97) was present in the pre-PCV7 at very low levels and began to rise soon after PCV7 was introduced country-wide. Data for 1999 and 2000 were not collected. Data are presented as the percent of all MDR isolates collected for the given year.

Figure 2

Multi-locus sequence typing (MLST) of multi-drug resistant (MDR) serotype 19A isolates (n = 97) obtained from the Canadian Bacterial Diseases Surveillance Network during pre-PCV 7 introduction, vaccine introduction, and post-vaccine introduction eras. Sequence types (ST) are depicted as a percent of all MDR 19 isolates for that period. ST320 has emerged as the singular dominant sequence type amongst MDR 19A isolates in the post-PCV7 era. Novel implies a collection of strains for which no sequence type (ST) was identified within the MLST http://www.mlst.net/ database at this time but submissions have been made and are summarized in Additional file 2. eBURST summary data for MDR 19A, MDR 19F (n = 30) and susceptible 19A (n = 19) controls are appended in Additional file 1http://www.pillailab.com/suppdata/index.html.

Figure 3

Minimum spanning tree of multi-drug resistant (A) (MDR) serotype 19A (n = 97) using BioNumerics software. A categorical clustering was performed based on multi-locus sequence type (MLST). Sequence types sharing the maximum number of single-locus variants were connected first. Each circle represents a sequence type (ST) the size of which is proportional to the number of isolates within that particular ST. Colors within circles indicate the minimum inhibitory concentration (MIC) ranges for penicillin. Relationships between the STs are depicted by the lines connecting the STs and the relative lengths of the branches linking them. Distance coding enumerates the number of differences at a given MLST locus. A distance coding of greater than 2 implies a different clonal complex. Angles of the line connections and the overlapping circles have no significance.

Figure 4

Minimum spanning tree of multi-drug resistant susceptible 19A control isolates (n = 16) using BioNumerics software. A categorical clustering was performed based on multi-locus sequence type (MLST). Sequence types sharing the maximum number of single-locus variants were connected first. Each circle represents a sequence type (ST) the size of which is proportional to the number of isolates within that particular ST. Colors within circles indicate the minimum inhibitory concentration (MIC) ranges for penicillin. Relationships between the STs are depicted by the lines connecting the STs and the relative lengths of the branches linking them. Distance coding enumerates the number of differences at a given MLST locus. A distance coding of greater than 2 implies a different clonal complex. Angles of the line connections and the overlapping circles have no significance.

Figure 5

Minimum spanning tree of multi-drug resistant MDR 19F isolates (n = 29) from the pre-PCV7 era using BioNumerics software. A categorical clustering was performed based on multi-locus sequence type (MLST). Sequence types sharing the maximum number of single-locus variants were connected first. Each circle represents a sequence type (ST) the size of which is proportional to the number of isolates within that particular ST. Colors within circles indicate the minimum inhibitory concentration (MIC) ranges for penicillin. Relationships between the STs are depicted by the lines connecting the STs and the relative lengths of the branches linking them. Distance coding enumerates the number of differences at a given MLST locus. A distance coding of greater than 2 implies a different clonal complex. Angles of the line connections and the overlapping circles have no significance.

Figure 6

The whole genome sequence of a representative CBSN isolate of emergent multi-drug resistant serotype 19A ST320 (Genbank Accession GPID ACNU00000000) was compared to a representative isolate of 19F ST320 from the pre-vaccine era (Genbank Accession GPID ACNV00000000). The Solexa platform (Illumina Inc, San Diego, CA) was used for sequencing with greater than 100X coverage obtained throughout each genome. Depicted here is the whole genome of a representative MDR 19A ST320 in the post-vaccine era. The locations of proteins encoded on the leading and lagging strands are shown on the outer two rings. Gene ontology categories are color-coded. Of the internal ring, the outermost bars indicate SNPs identified in the leading strand and the innermost bars represent SNP identified in the lagging strand relative to MDR 19F ST320. The length of the bars is propostional to the number of SNPs. For detailed gene identification, location of capsule biosynthetic loci, other key alleles, comparison of 19A and 19F SNPs, as well as a mutation profile compared to reference strain R6 (Genbank AE007317), see Additional file 4http://www.pillailab.com/suppdata/index.html.

Figure 7

Pie chart percentage breakdown by gene ontology classification (GenoList, Institut Pasteur, Paris) of single nucleotide polymorphisms (SNPs) belonging to MDR 19A ST320 relative to MDR 19F ST320.

Figure 8

Alignment of the capsular biosynthetic loci between ST320 MDR19A in the post-PCV7 introduction era and pre-PCV7 MDR19F. The capsule locus resides between genes aliA and dexB. The vex operon was present directly adjacent to the capsule locus of 19A but not 19F. Of note are the genes pbp2x and pbp1a, responsible for penicillin resistance, that are present adjacent to the capsule locus. For a full image of the capsule locus and flanking regions please see Additional file 8http://www.pillailab.com/suppdata/index.html.