Turner et al. Reply to "Emergence of the Same Successful Clade among Distinct Populations of emm89 Streptococcus pyogenes in Multiple Geographic Regions".

REPLY

We are pleased that Friães et al. have elected to examine emm89 strains in Portugal (1); it was our hope that other groups would evaluate emm89 strains from elsewhere, using either whole-genome sequencing (WGS) or the PCR method that we described in our report (2). Subsequent to our publication (2), Zhu et al. used WGS to identify variants of the nga-slo promoter among emm89 isolates in the United States, Finland, and Iceland (3) indicative of the new emergent emm89 clade in these countries. Although they did not use WGS, Friães et al. demonstrated that it is likely that the emergent emm89 clade that we identified in the United Kingdom and elsewhere is also present in Portugal (1).

Friães and colleagues have noted that the focus on different aspects of emm89 strains in the various studies (2–4) makes it hard to determine if the same clade had disseminated in different regions. In our original report, we aimed to provide a single comprehensive description of all six regions of recombination that characterize this emergent emm89 clade (2). Two of these regions, the nga-slo locus (region 2) and the hasABC capsule locus (region 6), were highlighted because of their phenotypic significance; it was beyond the scope of our paper to undertake in-depth analysis of all of the regions. We did, however, clearly demonstrate enhanced NGA-SLO toxin expression and stated that this could be due to a single polymorphism or several polymorphisms within the nga-slo locus and promoter or elsewhere in the genome. Indeed, it may be misleading to focus on a single region or multiple regions of recombination without understanding of the collective impact of all of the changes. In our sampling of the emm89 population, we failed to identify isolates with an intermediate number of recombination regions, raising the possibility that all six remodeled regions promoted the success of this lineage, not simply the two that have been highlighted.

To more clearly illustrate the emergence of the new emm89 clade, we combined WGS data from isolates from the United States, Finland, and Iceland reported by Zhu et al. (3) for which we could confidently determine the sequence type (ST) (n = 737) with our data for United Kingdom and Swiss strains (2) (n = 131). We identified three distinct lineages (Fig. 1A) that are linked to the three previously identified nga-slo promoter variants (3). In agreement with Friães et al., the original United Kingdom emm89 population lineage (with nga-slo promoter “variant 2”) and the emergent acapsular emm89 clade (with nga-slo promoter “variant 3”) are both predominantly ST101; any ST variants within this clade are limited to single-locus variants. Isolates from Finland and Iceland, with a few minor exceptions, clustered either with the original United Kingdom emm89 ST101 lineage or with the new emergent ST101 clade described in our original paper (2). The majority of isolates from the United States, however, clustered either with the third identified lineage (with nga-slo promoter “variant 1”), dominated by ST407 and ST803, or with the new emergent ST101 clade. It appears that, as in our original findings in the United Kingdom, prior to 2007 to 2008, the dominant emm89 lineage in European countries was an ST101-like lineage but that the dominant lineage in the United States was an ST407/ST803-like lineage. Since 2007 to 2008, the new emergent variant, highly likely to have arisen within the ST101 background through recombination, has risen to dominance in both Europe and the United States. All members of this emergent clade have the same six regions of recombination, of which the nga-slo and hasABC regions are just two.

FIG 1 

Phylogenetic analysis of emm89 population. (A) Three lineages were identified in the emm89 population based on the maximum likelihood phylogeny of SNPs identified after mapping to the H293 reference strain of the sequences of 868 emm89 strains from both the United Kingdom study (2) and the United States/Finland/Iceland study (3), where ST could be confidently assigned. We identified 16 sequence types (ST101, n = 621; ST407, n = 160; ST580, n = 3; ST791, n = 2; ST792 to ST800, n = 1 each; ST801, n = 20; ST802, n = 1; ST803, n = 52). Two lineages were equivalent to those in the United Kingdom study (2) which identified an emergent clade type (red) from a previous lineage (blue) where the major ST was ST101 with some single-locus variants (identified in parentheses). These lineages are also associated with the previously termed capsular nga-slo promoter “variant 2” strains (blue) and acapsular nga-slo promoter “variant 3” strains (red). A third lineage (black) was also identified, representing ST407 and ST803 and two other sequence type variants (ST795 and ST799), equivalent to the previously termed capsular nga-slo promoter “variant 1” strains. (B) Regions of recombination in a subset of 68 strains representing all of the sequence types identified compared to the reference strain H293 (ST101) were predicted using Gubbins (5). Branch colors represent lineages and sequence types. Emergent variant ST101 (red) and other single-locus variants (indicated with an asterisk [*]) are shown in the following order from the top of the tree: ST791 (maroon), ST802 (orange), ST796 (yellow), ST801 (pink), ST798 (dark red), ST792 (purple), ST580 (green), ST793 (lilac), ST794 (grey-purple), ST800 (bright pink). Previously dominant “European” variant ST101 (blue) and single-locus variant ST797 (pale blue) are shown, as are previously dominant “United States” variants ST407 (black) and ST803 (grey) and single-locus variants ST799 (pale grey) and ST795 (brown). Regions of recombination identified in each strain are shown as vertical red lines (indicating recombination on internal nodes) or blue lines (indicating recombination on terminal branches), and genome coordinates are given on the bottom line (turquoise). The six previously identified regions of recombination (R1 to R6) were present in the emergent clade strains of ST101 and other single-locus variants within this lineage, along with a variation in the clustered regularly interspaced short palindromic repeat (CRISPR) region (column C) (2). The United States variant ST407/ST803 lineage had several large regions of predicted recombination compared to the H293 reference strain, including one region which encompassed the MLST yqil locus and introduced the sequence type variation.

The ST407/ST803-like lineage sequenced by Zhu et al. (3) was infrequent in European countries and is separated from the two ST101-like lineages by a high level of core single nucleotide polymorphisms (SNPs) (Fig. 1A). In comparison to the emm89 H293 reference (ST101), this lineage contains its own distinct predicted pattern of recombination (Fig. 1B). Given the level of difference between ST101 and ST407/ST803, the success of the ST101 emergent clade in both North America and Europe may therefore reflect a more general competitive advantage of this new lineage. The findings also highlight a propensity of the genetic background of emm89 to undergo recombination.

We agree that the emergent clade has most likely acquired a competitive advantage through enhanced transmission and/or a propensity for superficial infections. In 2009, a year when the incidence of emm89 invasive infections peaked in the United Kingdom (2), emm89 accounted for 16% of all 135 noninvasive Streptococcus pyogenes isolates submitted in April to November to our clinical laboratory (which serves a population of approximately 2 million) and 13% of 125 invasive S. pyogenes isolates submitted in January to November within London to the national reference laboratory. This contrasted with emm1 strains, where a disproportionately high number accounted for invasive isolates (23%) compared with noninvasive isolates (11%) during the same time periods (unpublished data). At present, there are no experimental data that directly compare the invasiveness of the emergent clade with that of previous emm89 lineages. Using population-based mortality data, we found no evidence that the emergent emm89 clade, where associated with invasive infection, caused greater mortality than the original ST101 lineage in the United Kingdom (2).

The potential for certain lineages of S. pyogenes to undergo extensive recombination is of wider significance. It seems likely that future vaccines and/or diagnostic tests will target genes and gene products thought to be crucial for pathogenesis, and yet we have demonstrated that, at least in emm89, some factors can be lost or dramatically altered. The mechanisms explaining why and how these strains undergo recombination-related genome remodeling have not been determined but clearly require recognition and further evaluation.