Emergence of a New Highly Successful Acapsular Group A Streptococcus Clade of Genotype emm89 in the United Kingdom

Abstract

Group A Streptococcus (GAS) genotype emm89 is increasingly recognized as a leading cause of disease worldwide, yet factors that underlie the success of this emm type are unknown. Surveillance identified a sustained nationwide increase in emm89 invasive GAS disease in the United Kingdom, prompting longitudinal investigation of this genotype. Whole-genome sequencing revealed a recent dramatic shift in the emm89 population with the emergence of a new clade that increased to dominance over previous emm89 variants. Temporal analysis indicated that the clade arose in the early 1990s but abruptly increased in prevalence in 2008, coinciding with an increased incidence of emm89 infections. Although standard variable typing regions (emm subtype, tee type, sof type, and multilocus sequence typing [MLST]) remained unchanged, uniquely the emergent clade had undergone six distinct regions of homologous recombination across the genome compared to the rest of the sequenced emm89 population. Two of these regions affected known virulence factors, the hyaluronic acid capsule and the toxins NADase and streptolysin O. Unexpectedly, and in contrast to the rest of the sequenced emm89 population, the emergent clade-associated strains were genetically acapsular, rendering them unable to produce the hyaluronic acid capsule. The emergent clade-associated strains had also acquired an NADase/streptolysin O locus nearly identical to that found in emm12 and modern emm1 strains but different from the rest of the sequenced emm89 population. The emergent clade-associated strains had enhanced expression of NADase and streptolysin O. The genome remodeling in the new clade variant and the resultant altered phenotype appear to have conferred a selective advantage over other emm89 variants and may explain the changes observed in emm89 GAS epidemiology.

Importance: Sudden upsurges or epidemic waves are common features of group A streptococcal disease. Although the mechanisms behind such changes are largely unknown, they are often associated with an expansion of a single genotype within the population. Using whole-genome sequencing, we investigated a nationwide increase in invasive disease caused by the genotype emm89 in the United Kingdom. We identified a new clade variant that had recently emerged in the emm89 population after having undergone several core genomic recombination-related changes, two of which affected known virulence factors. An unusual finding of the new variant was the loss of the hyaluronic acid capsule, previously thought to be essential for causing invasive disease. A further genomic adaptation in the NADase/streptolysin O locus resulted in enhanced production of these toxins. Recombination-related genome remodeling is clearly an important mechanism in group A Streptococcus that can give rise to more successful and potentially more pathogenic variants.

FIG 1

United Kingdom epidemiology of emm89 invasive GAS disease (iGAS). (A) Incidence of emm89 iGAS (black solid line, left axis) alongside the total numbers of iGAS cases of all emm types (gray dotted line, right axis). (B) Thirty-day case fatality rates of emm89 iGAS (±95% confidence interval).

FIG 2

Phylogenetic analysis of all 131 emm89 isolates sequenced. (A) Maximum likelihood unrooted tree based on core SNPs compared to the reference strain H293. Eighty-three of 131 isolates clustered into a distinct emergent clade (circled) separate from the 47/131 isolates that clustered with the reference strain H293 (indicated by asterisk). A potential hypermutator strain, H1041, is indicated by a red branch. (B) Schematic of the H293 genome. The outer ring indicates the positions of the six regions of recombination present in emergent clade-associated emm89 strains relative to H293 and all other non-clade-associated emm89 strains (regions 1 to 6 shown in red). The inner two rings represent GC skew and GC content. (C) Mid-rooted phylogenetic tree based on core genome SNPs compared to the reference strain H293 excluding all regions of predicted recombination. The emergent emm89 clade is distinct from the rest of the diverse emm89 population. Strains are color coded by year of isolation. The presence of each of the 11 different superantigen genes is indicated to the right of the tree.

FIG 3

Rise of the emergent clade to dominance in the United Kingdom emm89 population. Between 2004 and 2007, the United Kingdom was represented by a diverse emm89 population (blue line, left axis); however, there was a sudden switch between 2007 and 2008 and the emergent clade-associated emm89 variant (red line, left axis) became the dominating type in the population. The rise of the emergent clade coincided with an increase in the frequency of emm89 iGAS in the United Kingdom (gray dotted line, right axis). Association with the emergent clade was determined by WGS of 126 emm89 strains (excluding pairs of strains isolated from the same patient) isolated in 2004 to 2013 (n = 10 for 2005, 2006, 2007, 2011, and 2013; n = 11 for 2004 and 2008; n = 28 for 2009; n = 17 for 2010; n = 9 for 2012).

FIG 4

Prophage-like element found in the emergent clade-associated strains. In clade-associated strains, the superantigen gene speC (shown in red) and the DNase gene spd1 (shown in orange) were associated with a prophage-like element (ΦM89.1) that shared a high level of identity to a prophage found in the genome of M1 strain SF370 (Φ370.1). Bacteriophage structural genes found in Φ370.1 (20) are absent in ΦM89.1. The figure was drawn using Easyfig (52).

FIG 5

Variation in phenotype between emm89 strains. (A) NADase activities of 34 sequenced emm89 strains were measured in the culture supernatant; the result shown is the highest 2-fold dilution able to hydrolyze NAD⁺. Data represent the median (plus range) from four independent experiments. (B) SLO activities of 34 sequenced emm89 strains were measured in the culture supernatant by lysis of sheep erythrocytes and reported as the percentage of activity relative to the positive control (H₂O). Data represent the means (+ standard deviations) from three independent experiments. (C) Production of hyaluronic acid was measured using an ELISA-based assay for hyaluronan. Representatives of the non-clade-associated emm89 strains tested all had detectable levels of hyaluronic acid. All representatives of emergent clade-associated strains tested had undetectable levels of hyaluronic acid (N.D., not detected). Data represent the means (+ standard deviations) from three independent experiments. Hyaluronic acid capsule was measured as femtograms per 1 × 10³ CFU. (D) Both non-clade-associated and clade-associated emm89 strains survive equally well in whole human blood. There was no overall difference in multiplication factor between the two types of strains, including (as shown) or excluding (not shown) strains with covR/S or rocA mutations. Strains were grown in duplicate using a single donor, and data represent the means. The experiment was repeated in a second donor with similar results. (E) The ability of invasive non-clade-associated strains (n = 23) and emergent clade-associated strains (n = 21) to adhere to and colonize uncoated plastic after 24 h of culture was measured by staining with Gram’s crystal violet. Data represent the means from three experiments. **, P ≤ 0.01, Mann-Whitney comparison. Strains with blue bars or unfilled circles (non-clade-associated emm89 strains) and red bars or unfilled squares (emergent clade-associated emm89 strains) carry mutations in covR/S or rocA known to control expression of nga, slo, and hyaluronic acid capsule; mutations in either covR/S or rocA would enhance nga, slo, and capsule expression.