Superantigenic Activity of emm3 Streptococcus pyogenes Is Abrogated by a Conserved, Naturally Occurring smeZ Mutation

Streptococcus pyogenes M/emm3 strains have been epidemiologically linked with enhanced infection severity and risk of streptococcal toxic shock syndrome (STSS), a syndrome triggered by superantigenic stimulation of T cells. Comparison of S. pyogenes strains causing STSS demonstrated that emm3 strains were surprisingly less mitogenic than other emm-types (emm1, emm12, emm18, emm28, emm87, emm89) both in vitro and in vivo, indicating poor superantigenic activity. We identified a 13 bp deletion in the superantigen smeZ gene of all emm3 strains tested. The deletion led to a premature stop codon in smeZ, and was not present in other major emm-types tested. Expression of a functional non-M3-smeZ gene successfully enhanced mitogenic activity in emm3 S. pyogenes and also restored mitogenic activity to emm1 and emm89 S. pyogenes strains where the smeZ gene had been disrupted. In contrast, the M3-smeZ gene with the 13 bp deletion could not enhance or restore mitogenicity in any of these S. pyogenes strains, confirming that M3-smeZ is non-functional regardless of strain background. The mutation in M3-smeZ reduced the potential for M3 S. pyogenes to induce cytokines in human tonsil, but not during invasive infection of superantigen-sensitive mice. Notwithstanding epidemiological associations with STSS and disease severity, emm3 strains have inherently poor superantigenicity that is explained by a conserved mutation in smeZ.


Introduction
Streptococcus pyogenes is a major human pathogen, responsible for a wide spectrum of clinical manifestations. These range from non-invasive and self-limiting to severe, potentially lethal invasive infections complicated by streptococcal toxic shock syndrome (STSS). The development of STSS during S. pyogenes invasive infection is hypothesized to result from exposure to secreted streptococcal superantigens that leads to excessive proliferation of T cells, accompanying cytokine release, inflammation and tissue damage [1]. S. pyogenes can express several different superantigens (reviewed in [2]) that can vary in their potency, thus differences in mitogenicity within and between M/emm-types can be influenced by the complement of superantigen genes, as well as differences in expression or degradation [3][4][5][6]. SMEZ is the most potent streptococcal superantigen described, although produced in small amounts compared to other superantigens, [7][8][9][10] and the smeZ gene is present in the majority of S. pyogenes strains, with over 40 different alleles [8,11].
In the UK, which has the highest measured rate of invasive S. pyogenes disease in Europe (3.33 per 100,000), infections with M/ emm3 type S. pyogenes are independently associated with a three-fold increased risk of STSS, compared to patients infected with a reference strain group (M/emm28) that is represented by large numbers, and is associated with an average case fatality ratio [12]. Studies in the USA, Canada, and elsewhere in Europe have also highlighted an increased risk of death associated with M/emm3 S. pyogenes infection [13][14][15][16].
In this study, we compared superantigenicity of different S. pyogenes emm-types, using mitogenicity as a surrogate measure, and explored the hypothesis that superantigenic potency would be associated with STSS. Surprisingly we observed that emm3 strains had almost negligible mitogenicity compared to other emm-types. We sought to determine whether this was due to a recently identified 13 bp deletion within smeZ that yielded a premature stop codon and predicted inactive protein expression [17]. We went on to determine the frequency of the 13 bp deletion and the impact on superantigenic function in emm3 strains.

Emm3 Isolates from STSS Patients Lack Mitogenic Activity Compared with Other STSS Isolates, Both in vitro and in vivo
The mitogenicity of a panel of 63 S. pyogenes isolates from cases of confirmed STSS was evaluated in vitro using human mononuclear cells (MNC) ( Figure 1A). Strains represented the most common emm-types associated with STSS in the UK (emm1, n = 11; emm3, n = 12; emm12 and emm87, n = 11; emm18, n = 5; emm28, n = 7; emm89, n = 6). The complement of superantigens carried by each isolate was determined by PCR (Table S1). Emm3 and emm1 strains were markedly less mitogenic than other emm-types ( Figure 1A) despite their reported association with STSS; this was reproducible even when MNC from alternative donors were used (not shown). In particular, some emm3 strains demonstrated mitogenic activity that was barely above control levels, despite the presence of 4 or 5 superantigen genes (speA, smeZ, speG, ssa, and speK).
To determine whether emm3 S. pyogenes strains might change in phenotype in vivo, during an infection, the superantigenic activity of S. pyogenes strains of different emm-types was compared in vivo using a murine intramuscular infection model; three representative strains were tested for each of the seven STSS emm-types. In this model, secreted bacterial superantigens enter the systemic circulation, hence mouse serum acquires mitogenic properties towards human lymphocytes [18]. Murine sera obtained 24 hours after onset of infection were tested for mitogenic activity with human MNC as a measure of released superantigen ( Figure 1B). Similar to the in vitro results, emm3 and emm1 strains released markedly less mitogen into mouse serum than other emm-types, with the exception of emm89 strains. These differences were not attributable to differences in bacterial counts during infection; despite producing the lowest mitogenic activity in mouse serum, emm3, emm1 and emm89 strains demonstrated the greatest systemic spread from the site of infection (thigh tissue) to the spleen and liver (Table S2).

Emm3 Isolates Carry a Mutation in smeZ
Previously we had determined that five contemporary emm3 S. pyogenes strains carried a smeZ gene containing a 13 bp deletion at nucleotide position 316, that leads to a frame shift and a premature stop codon predicted to truncate SMEZ [17] ( Figure 2). This smeZ allele (smeZ59N) was restricted to emm3 strains and was not detected in 44 other emm-types tested [17]. We detected the 13 bp smeZ deletion mutation by PCR in all 12 emm3 S. pyogenes strains associated with STSS (strains listed in Table S1). The same mutation was detected in 27/27 other emm3 S. pyogenes invasive diseasestrains submitted to the reference laboratory that included three of the most common MLST sequence types for emm3 strains; ST15, ST315, and ST406 (http://spyogenes.mlst. net/). The mutation was also identified in the two existing NCBI sequences for M3 S. pyogenes originating from the USA (MGAS315, GenBank accession number AE014074, [19]) and from Japan (SSI-1, GenBank accession number BA000034, [20]).

Contribution of the 13 bp smeZ Mutation to Low Mitogenicity of emm3 Strains
The mutation within the smeZ-M3 gene suggested that any protein, if expressed, would be non-functional, thus potentially accounting for the low mitogenic activity observed for emm3 strains. To confirm that the SMEZ-M3 protein is non-functional and that the low mitogenic activity of emm3 isolates was not accounted for by enzymatic degradation of SMEZ or interference with T cell mitogenicity, an emm3 isolate, GAS-M3 (with the smeZ-M3 gene) was transformed with a plasmid that conferred over-expression of either a functional SMEZ protein from an M89 strain (GAS-M3 smeZ-M89 ) or the truncated SMEZ-M3 protein (GAS-M3 smeZ-M3 ). Supernatant from strain GAS-M3 smeZ-M89 increased proliferation of human MNCs by 3-fold compared to the parental (untransformed) emm3 isolate, GAS-M3 ( Figure 3A), confirming that the M3 strain does not produce an inhibitor that interferes substantially with mitogen function. Supernatant from GAS-M3 smeZ-M3 failed to enhance proliferation, confirming that SMEZ-M3 is non-functional as a mitogen. Transformation of GAS-M3 with the empty control plasmid (GAS-M3 control ) did not affect proliferation of human MNCs compared to untransformed GAS-M3 (not shown). Real time PCR analysis demonstrated that the transformed isolates expressed equivalent amounts of smeZ transcript from either plasmid (GAS-M3 smeZ-M89 :163.0 copies of smeZ per 10,000 copies of proS and GAS-M3 smeZ-M3 :183.8 copies of smeZ per 10,000 copies of proS); therefore differences in expression could not account for the differences in proliferation.

The 13 bp Deletion Abrogates Function of SMEZ in Other Strain Backgrounds
To confirm that SMEZ-M3 cannot function as a mitogen in other emm-types, we used genetically modified S. pyogenes strains lacking functional smeZ and determined whether transformation with plasmids encoding smeZ-M3 or smeZ-M89 could complement the defect in mitogenicity. Disruption of smeZ in emm1 S. pyogenes (GAS-M1DsmeZ) led to a ,40% reduction in proliferation ( Figure 3B); this strain also has genes encoding for other superantigens (speA, speG, speJ). A more dramatic reduction in mitogenicity (,80%) was observed after disruption of smeZ in an emm89 strain (GAS-M89DsmeZ) ( Figure 3C), despite the presence of genes encoding for other superantigens (speG, speH, speJ) consistent with an earlier report [9]. Mitogenic activity was fully restored when GAS-M1DsmeZ and GAS-M89DsmeZ were transformed to over-express functional SMEZ-M89. Transformation of GAS-M1DsmeZ ( Figure 3B) and GAS-M89DsmeZ ( Figure 3C) to over-express the M3 form of SMEZ, however, failed to restore mitogenic potential, again confirming that SMEZ-M3 is not a functional mitogen.
The 13 bp smeZ-M3 Deletion Reduces the Potential of emm3 S. pyogenes to Stimulatecytokine Production in Human Tonsil Cells Acute inflammation mediated by smeZ expression may be important for tonsillopharyngeal colonization of GAS [21]. To determine if the 13 bp mutation in smeZ-M3 might affect inflammation in tonsil, human tonsil cells were cultured with bacterial cell-free supernatants from GAS-M3 control , GAS-M3 smeZ-M89 or GAS-M3 smeZ-M3 . After two and five days of culture, production of TNFa, TNFb, IFNc, IL-10, IL-17, IL -5, and IL-12 were measured in supernatant, focusing on cytokines that have previously been identified as important in the human response to superantigens ( Figure 4). In comparison to GAS-M3 smeZ-M3 , expression of functional SMEZ by GAS-M3 smeZ-M89 enhanced the production of the modulating cytokine IL-10 at both times points, and also increased production of TNFa on day 2 and TNFb and IL-17 on day 5 consistent with an effect on T cells. There was also a non-significant increase in IFNc production by GAS-M3 smeZ-M89 . The levels of each cytokine induced by expression of functional SMEZ were comparable to those induced by GAS-M1 control , which naturally expresses a functional SMEZ ( Figure 4). In contrast, expression of SMEZ-M3 by GAS-M3 smeZ-M3 did not affect cytokine production at any time point. Production of IL-5 and IL-12 were below the limit of detection. Experiments were repeated using a second tonsil donor with similar results (not shown).

The 13 bp Deletion in smeZ-M3 Limits Only Interleukin-5 Production during Invasive Infection of Superantigensensitive Mice
To determine whether the non-functioning SMEZ-M3 affects cytokine production in vivo during GAS-M3 infection, we infected superantigen-sensitive HLA-DQ8 transgenic mice with GAS-M3 smeZ-M3 or GAS-M3 smeZ-M89 with a low intramuscular inoculum and measured both tissue and serum levels of cytokines after 24 h using a bead array.
In contrast to ex-vivo stimulation of human tonsil cells and in contrast to previous studies using GAS-M89 [9], expression of functional SMEZ during intramuscular infection of HLA-DQ8 transgenic mice did not increase local tissue levels of TNFa, IL-10, IL-17 or IFNc. In fact, mice infected with GAS that expressed the non-functional smeZ gene (M3 smeZ-M3 ) had a trend for higher tissue cytokine levels (IL-17, IL-10, IL-1b) compared with mice infected with GAS expressing the functional smeZ gene (M3 smeZ-M89 ) ( Figure 5A). Most tissue cytokines were unaffected by the 13 bp deletion in SMEZ however ( Figure S1).
The only cytokine influenced by the 13 bp deletion in smeZ was IL-5; IL-5 was higher in GAS-M3 smeZ-M89 infection, not only at the Emm3 STSS isolates show emm-type specific differences in overall mitogenicity. A.Human MNC proliferation response to culture supernatants from 63 Streptococcus pyogenes streptococcal toxic shock syndrome (STSS) isolates grouped by emm-type (numbers per group: emm1, n = 11; emm3, n = 12; emm12 and emm87, n = 11; emm18, n = 5; emm28, n = 7; emm89, n = 6). Negative; tissue culture media (RMPI) alone. Outliers are represented as individual circles. Representative of 2 experiments performed on different donors. B. Human MNC proliferation response to sera obtained from CD1 mice 24 hours after being infected with one of three S. pyogenes strains representing each emm-type; two mice were infected per strain. Negative; uninfected mouse serum. Proliferative response is measured as counts per minute (cpm) of tritiated-thymidine uptake. Median and 5th, 25th, 50th, 75th, and 95th centiles shown. doi:10.1371/journal.pone.0046376.g001 site of infection but also in the serum ( Figure 5A and B). Serum levels of other cytokines were no different between the two infected mouse groups ( Figure S2).
To ensure that the two bacterial strains used had exhibited the expected phenotype in vivo, we assessed the mitogenic activity present in the thigh tissue and serum 24 h after infection. As expected, GAS-M3 smeZ-M3 resulted in markedly lower mitogenic activity both at the local site of infection (thigh tissue homogenate, Figure 5C) and in the serum ( Figure 5D) compared with GAS-M3 smeZ-M89 . The differences observed in mitogenicity and cytokine response were not due to measurable differences in bacterial burden as both groups had a similar bacterial load in thigh tissue after 24 hrs infection (GAS-M3 smeZ-M3 , median; 2610 5 CFU/mg range; 1610 5 -3610 5 CFU/mg and GAS-M3 smeZ-M89, median; 1610 5 CFU/mg, range; 5610 4 -5610 5 CFU/mg) and neither group demonstrated systemic spread.

Discussion
Invasive infections due to emm3 S. pyogenes are widely associated with increased disease severity and, in the UK, with risk of STSS compared with other emm-types [12][13][14][15]. Intriguingly, emm3 S. pyogenes strains demonstrated very low mitogenic activity compared with other dominant emm-types that cause STSS; indeed, some emm3 strains had little mitogenic activity above background, which was surprising given the complement of 4-5 superantigen genes including speA. The observed deficiency in mitogenicity did not change even when measured in vivo, a setting where expression of superantigens can be enhanced. Emm3 strains carry a 13 bp deletion mutation in the smeZ sequence that changes the reading frame, resulting in a premature stop codon that is predicted to preclude expression of a functional protein [17,22]. Thus, despite the presence of the smeZ gene, emm3 strains have no ability to produce active SMEZ superantigen. To a large extent this explains the failure of emm3 strains to match the mitogenic activity of other strains. The aberrant M3-smeZ was unable to restore mitogenicity when over-expressed in an emm3 strain, and also could not function when transferred to alternative S. pyogenes emm-types (emm1 and emm89). Strong mitogenic activity was associated with a functional smeZ gene and major reductions in mitogenic activity were associated with a non-functioning smeZ gene.
There are over 40 different smeZ alleles across the different emmtypes sharing high pair-wise identity (94-99%) [8,11], however the region surrounding the 13 bp deletion that occurs in the emm3 smeZ allele (smez-59N) is actually highly divergent across all the different alleles from different emm-types. Although no other non-emm3 alleles demonstrated deletions within this region, three smeZ alleles, smez-6, smez-19 and smez-23 from different emm-types had frame-shift single base pair deletions that occur closer to the Nterminus [8]. The 13 bp deletion in smeZ was found in all 39 emm3 GAS strains tested, including strains from three different multilocus sequence types (ST15, ST315 and ST406). Further to this, the 13 bp deletion has also been found in all of 200 UK emm3 strains of ST15, ST315 or ST406, recently sequenced (unpublished). The deletion in smeZ was also present in emm3 genome strains from Japan (SSI-1, GenBank accession number BA000034) [20] and the USA (MGAS315, GenBank accession number AE014074) [19], both ST15. We were also able to identify the deletion in the Canadian emm3 strains, sequenced by Beres et al [23], by mapping the reads available on the short read archive (Project SRP000775) to the smeZ locus from the M1 strain MGAS5005. However, PCR analysis would be required to confirm that the deletion was identical to that found in UK  . emm3 strains have low mitogenicity due to the mutation in smeZ. A. MNC proliferative response to culture supernatants from an emm3 isolate, GAS-M3 carrying the typical M3-smeZ with 13 bp deletion, GAS-M3 over-expressing the functional M89 form of SMEZ (GAS-M3 smeZ-M89 ) and GAS-M3 over-expressing the M3-SMEZ (GAS-M3 smeZ-M3 ). B. Experimental smeZ mutation in emm1 S. pyogenes (GAS-M1) reduced MNC proliferation (GAS-M1DsmeZ). Proliferative response was restored when GAS-M1DsmeZ over-expressed the functional M89 form of SMEZ (GAS-M1 smeZ-M89 ) but not with M3-type form of SMEZ (GAS-M1 smeZ-M3 ). C. A similar result was obtained using an emm89 strain (GAS-M89) with an experimental mutation in smeZ (GAS-M89DsmeZ). Proliferation was restored when GAS-M89DsmeZ over-expressed the functional M89 form of SMEZ (GAS-M89 smeZ-M89 ) but not with M3-type form of SMEZ (GAS-M89 smeZ-M3 ). Negative; media alone.Proliferation was measured as counts per minute emm3 as the region of divergence downstream of the deletion introduced ambiguity in the read mapping. Furthermore, the emm3 smeZ deletion is not a very recent event since we detected the same 13 bp deletion in a 1931 emm3 S. pyogenes puerperal sepsis isolate from Queen Charlottes Hospital, Hammersmith (Lynskey N. et al, unpublished). We postulate that the smeZ deletion is likely to have arisen early in the evolution of the emm3 lineage and the pseudogene conserved over decades.
Although smeZ-M3 encodes a mutated gene that lacks classical T cell mitogenic function, it remains possible that the truncated SMEZ-M3 has an alternative as-yet unrecognized function in GAS-M3 strains, such as, for example, interfering with the function of other superantigens or acting as a non-mitogenic activator of class II-positive cells. Results showed that smez-M3 is transcribed at a similar level to functional smeZ-M89 although we are unable to confirm if truncated SMEZ-M3 polypeptide is actually produced.
Interestingly, emm1 strains were also poorly mitogenic compared with other emm-types, both in vitro and in vivo. This was surprising since emm1 strains all carried speA, speG, and speJ as well as an intact smeZ gene. SPEA production is known to account for only a small proportion of emm1 mitogenic activity in vitro [24], although emm1 strains are able to upregulate production of superantigen during deep tissue infection in vivo [25,26]. SPEJ is a superantigen that selectively expands T cells bearing the Vb2 receptor [27]; since this is one of the largest subfamilies of human T cells, it was expected that emm1 strains might demonstrate heightened mitogenic activity compared with other emm-types. Nonetheless, targeted mutation of smeZ demonstrated clearly that SMEZ is an important component of emm1 mitogenic activity, similar to effects previously observed in an emm89 S. pyogenes strain [9]. We cannot exclude the possibility that emm1 and emm3 strains secrete proteins that are toxic to human T cells, although the ability of smeZ-M89 to confer mitogenicity suggests such an effect is not major. Degradation of superantigens by the streptococcal cysteine protease, SPEB may also contribute to reduced mitogenic activity of S. pyogenes [6] while the enhanced mitogenic properties of other emm-types may relate to synergistic actions of combinations of superantigens.
In addition to inducing T cell-derived cytokines, superantigens such as SMEZ can enhance TLR expression and signaling [9,28]). The 13 bp deletion in smeZ limited pro-inflammatory cytokine production induced by GAS-M3 in tonsil cell suspension. We speculate that this could be of benefit to GAS-M3 as it may limit local inflammation in the pharynx, perhaps delaying the onset of symptoms and the innate or adaptive immune response to infection.
In contrast, the 13 bp deletion in smeZ did not appear to have marked effects on cytokine production elicited by GAS-M3 during in vivo invasive infection, with the exception of IL-5. Indeed the induction of IL-5 appeared to be largely dependent on SMEZ. Superantigen-induced expression of IL-5 in vitro has been previously reported [7,29]. IL-5 is an eosinophilopoietic cytokine able to promote eosinophil maturation and survival; when activated, eosinophils can directly kill and enhance clearance of bacteria [30,31]), thus a reduction in IL-5 may be of benefit to M3 GAS.
It was intriguing that the 13 bp deletion in smeZ did not have a wider impact on cytokines during invasive infection with GAS-M3 and, in certain cases, appeared to result in higher levels of cytokine despite very low levels of mitogenic activity in serum. This indicates that GAS-M3 has potent mechanisms of inducing cytokines during infection that are distinct from superantigen production, although we cannot exclude the additional possibility that a truncated form of SMEZ-M3 may play a role in enhancing inflammation. Emm3 S. pyogenes were recently shown to have acquired a novel prophage conferring the ability to produce a secreted phospholipase A2 which may play a role in lethal sepsis [32]. Clinical criteria that fulfill a diagnosis of septic shock due to S. pyogenes will normally also fulfill criteria required for STSS. Coupled with the many other virulence factors released by S. pyogenes, it is likely that STSS associated with emm3 S. pyogenes infection may not always require superantigen production.

Materials and Methods
Bacterial Strains 63 S. pyogenes isolates from patients with STSS meeting the criteria defined by the Working Group on Severe Streptococcal Infections [33], were provided by the Streptococcal Reference Laboratory, Health Protection Agency (HPA, London, UK), and represented emm-types that had been associated with STSS in at least five patients. These included 54 isolates from 2003-2004 (from a total of 167 STSS cases identified in England and Wales) and an additional 9 isolates from 2005-2006. Emm-typing was performed as previously reported [34]. Emm3 isolates from patients with confirmed invasive S. pyogenes infection without STSS were also provided by the HPA.

PCR Analysis
The 13 bp deletion at position 316 of the smeZ nucleotide sequence of emm3 S. pyogenes strains is depicted in Figure 2. PCR primers were designed to amplify and distinguish the mutated smeZ gene, with the 13 bp deletion, from the non-mutated from, using a universal forward smeZ primer, Z1 (CTCCTGAAAAGAGGC-TATTTATG), a universal reverse smeZ primer, Z2 (CATACT-TACTTTTTAGAGGATTC), and an additional forward primer designed to anneal to the mutated (deleted) region, Z3 (AAC-TACTTGTCAGAAGGAATAC). Non-mutated smeZ yielded an amplification product of 796 bp with primers Z1 and Z2, while the mutated smeZ yielded a product of 474 bp with primers Z3 and Z2. Genotyping for other toxin genes was undertaken by multiplex polymerase chain reaction using the method previously reported [35] but analyzed by agarose gel electrophoresis. Primers that amplify the entire smeZ and speJ genes were additionally used, as allelic variation or pseudogenes can yield false negative and false positive results [36] (smeZ-F 59-TAAAGGCTTTTTTGCT TGTTTCA, smeZ-R 59-TTAGGAGTCAATTTCTATATC-TAAATGCCC; speJ-F 59-GATAGTGAAAATATTA, speJ-R 59-TTATTTAGTCCAAAGG).

Mitogenicity Assay
Total mitogenic activity of cell-free S. pyogenes supernatants was measured using a standard 72 hour human blood mononuclear cell (MNC) proliferation assay, as previously described [9]. Isolates were grown at 37uC to stationary phase in antibiotic-free tissue culture medium (RPMI 1640, Invitrogen, UK) containing fetal calf serum and L-glutamine [9], and bacterial cells were removed by (cpm) of tritiated-thymidine uptake and the percentage proliferation for each strain was calculated relative to the wild type strain (GAS-M3, GAS-M1, GAS-M89 respectively). Data are mean (+standard deviation) of three measurements. Representative of two experiments performed using different donor MNC. doi:10.1371/journal.pone.0046376.g003 Figure 4. Functional SMEZ is required to stimulate production of cytokines from human tonsil cells. Human tonsil cell suspensions were cultured with bacterial cell-free culture supernatants from GAS-M3 control (white bars), GAS-M3 smeZ-M89 (black bars) and GAS-M3 smeZ-M3 (gray bars). After 2 and 5 days incubation cell-free media were obtained from cultures and production of TNFa, TNFb, IL-10, IL-17 and IFNc were measured by ELISA. Horizontal dotted lines represent the mean level of cytokines produced after co-culture with bacterial cell-free culture supernatants from GAS-M1 control on day 2 and day 5. Mean (+ standard deviation) of three replicates measured in duplicate. Representative of two experiments performed on different donors. Statistical analysis was performed using ANOVA with Bonferroni multiple comparison. doi:10.1371/journal.pone.0046376.g004 centrifugation and 0.2 mM filtration. There were no significant differences in growth between emm-types. Proliferation of human MNCs was measured by tritiated-thymidine uptake after 72 hour co-incubation with 1:100 dilution of cell-free, filter-sterilized bacterial supernatant in the absence of human serum as previously described [9]. Human MNCs were obtained from at least two Figure 5. Functional SMEZ is not required for production of inflammatory cytokines in superantigen-sensitive mice. Groups of five superantigen-sensitive HLA-DQ8 female mice were infected for 24 hours intramuscularly (thigh) with 8610 7 CFU emm3 S. pyogenes strains; GAS-M3 smeZ-M89 and GAS-M3 smeZ-M3 .A. Thigh tissue homogenate from mice infected with GAS-M3 smeZ-M3 , over-expressing SMEZ-M3 demonstrated significantly higher levels of IL-17 and IL-1b compared to GAS-M3 smeZ-M89 infected thigh homogenate. Tissue from GAS-M3 smeZ-M3 infection also demonstrated non-significant increases in IL-10 (p = 0.059), IL-1a (p = 0.095) and IL-6 (p = 0.15). In contrast, GAS-M3 smeZ-M89 had higher levels of IL-5 (p = 0.056). TNFa and IFNc were similar in both groups. B. Mice infected with GAS-M3 smeZ-M89 also had significantly higher levels of IL-5 in the circulating serum compared to GAS-M3 smeZ-M3 infected mice. Horizontal dotted lines; lowest detectable level of each cytokine. Mitogenic activity was also measured in the thigh tissue homogenate (C) and serum (D) using human MNCs. GAS-M3 smeZ-M89 demonstrated consistently higher mitogenic activity due to the over-expression of functional SMEZ compared to GAS-M3 smeZ-M3 . Horizontal dotted lines; proliferation level of uninfected control mouse thigh homogenate (C) or serum (D). Median and 5th, 25th, 50th, 75th, and 95th centiles are shown. For analysis, samples with unmeasurable levels of cytokine were assigned a value half the lowest detectable value. Proliferation was measured as counts per minute (cpm) of tritiatedthymidine uptake. Statistical analysis was performed using Mann-Whitney. doi:10.1371/journal.pone.0046376.g005 different healthy donors. In addition to bacterial supernatant, the mitogenic activity of murine sera and thigh tissue homogenate were also tested at a 1:100 dilution and co-incubated with human MNCs for 72 hours.

Genetic Manipulation of emm1, emm3 and emm89 S. pyogenes Isolates
To determine whether mitogenic activity could be restored to an emm3 S. pyogenes strain (GAS-M3) with a functional copy of smeZ, the strain was transformed with the shuttle plasmid pDL278 carrying a copy of a functional smeZ gene from an emm89 strain with native promoter (pDLsmeZ-M89) to generate strain GAS-M3 smeZ-M89 , using methods described previously [37]). As a control, GAS-M3 was also transformed with pDL278 carrying a copy of the smeZ locus with native promoter from emm3 S. pyogenes that contains the 13 bp deletion and premature stop codon (pDLsmeZ-M3), to generate strain GAS-M3 smeZ-M3 . GAS-M3 also has genes for superantigens speA, speG, speK and ssa. To assess whether smeZ-M3 could function in an alternative emm-type S. pyogenes strain, the smeZ gene was firstly disrupted in emm1 S. pyogenes (GAS-M1) using methods described previously [9] to generate GAS-M1DsmeZ. GAS-M1DsmeZ was then transformed with either pDLsmeZ-WT or pDLsmeZ-M3. Additionally, these two plasmids were also used to transform a previously described emm89 strain (GAS-M89) with a disrupted smeZ gene (GAS-M89DsmeZ) [9]. Where additional control strains were needed, parental GAS-M3, GAS-M1 and GAS-M89 strains were also transformed with the empty shuttle plasmid pDL278 to give GAS-M3 control , GAS-M1 control and GAS-M89 control . Strains are listed in Table 1.

Real Time PCR Analysis
S. pyogenes cells were obtained following supernatant collection for mitogenicity assay. RNA was extracted and real time PCR was performed as previously described [38] using SYBR Green Jumpstart Taq Readymix (Sigma-Aldrich, UK). Primers Z4 59-TCCCTTCTAAGGAATATCTATAGTACGATTG and Z5 59-TTCCAATCAAATGGGACGG were designed to amplify a 209 bp target region of smeZ common to all known alleles. The housekeeping gene proS was also amplified alongside smeZ, using primers proS-F 59-TGAGTTTATTATGAAAGACGGCTA-TAGTTTC and proS-R 59-AATAGCTTCGTAAGCTTGAC-GATAATC to generate a 93 bp product [39]. Copies of smeZ and proS transcripts were calculated against a standard plasmid containing the target region of the smeZ gene and the proS gene. Copies of smeZ transcript were then normalized against copies of proS transcript.

Tonsil Cell Stimulation
Human tonsils were collected from adults undergoing routine tonsillectomy by the Imperial College Healthcare Trust Tissue Bank. Healthy tissue was cut into small sections which were then filtered through a 70 mM cell sieve with tonsil media (RPMI 1640, 10% fetal calf serum, 100 mM L-Glutamine, 250 iu/ml Penicillin, 250 mg/ml Streptomycin, 100 mg/ml Kanamycin, 2.5 mg/ml Amphotericin B; Invitrogen, UK) (method adapted from [40]). Cells were then washed twice before resuspending to 2610 6 cells per ml in tonsil media. In a 48 well plate format 1 ml of cells were incubated with 10% bacterial culture supernatant as for mitogenicity assays. After 48 hours (day 2) of media 500 ml was removed and stored at 220uC for cytokine analysis, and replaced with 500 ml fresh tonsil media. Cells were further cultured until day 5 when cells were removed by centrifugation and supernatant was stored at 220uC.

Bacterial Infection
Female outbred CD1 mice were infected intramuscularly (thigh) with 10 6 -10 7 CFU of S. pyogenes; three strains were used from each STSS-associated emm-type (emm1, 3, 12, 18, 28, 87 and 89) and two mice were infected per strain. After 24 hours of infection, mice were euthanized and blood was taken by cardiac puncture for CFU quantification and serum. Spleen, liver and infected thigh muscle were excised, individually weighed and homogenized in sterile PBS before plating on Columbia horse blood agar to quantify viable CFU. Remaining blood was centrifuged and serum was collected and stored at 220uC for testing mitogenic activity. Female humanized transgenic C57/BL/10HLA-DQ8 mice (kindly supplied by Daniel Altmann, Imperial College) were used as they are sensitive to superantigens including SMEZ [9]. Groups of 5 mice were infected intramuscularly (thigh) with 8610 7 CFU Table 1. S. pyogenes strains used in this study.

Ethics
The use of anonymized human tonsil tissue for research purposes was conducted within the remit of the Imperial HTA licence covering hospitals within ICHT and patients gave informed consent to use of tissues that would otherwise be discarded. Mice were used in accordance with UK Home Office guidance and subject to protocols set out in PPL 70/7379 approved by the Imperial College Ethical Review Process (ERP) panel.

Statistical Analysis
All statistics were performed using non-parametric analysis with GraphPad Prism version 5.0 for Windows, GraphPad Software, San Diego California USA. *; p,0.05. Figure S1 Cytokine levels in thigh tissue homogenate from superantigen-sensitive HLA-DQ8 mice infected with GAS-M3 smeZ-M89 (White box-whisker) or GAS-M3 smeZ-M3 (Grey box-whisker). Five mice per group were infected intramuscularly and after 24 hours infected tissue was removed and homogenated in sterile PBS. Cytokines were measured using LuminexH. Dotted horizontal line; lowest detectable level of each cytokine. For analysis, samples with undetectable levels of cytokine were assigned a value half the lowest detectable value. (TIF) Figure S2 Cytokine levels in serum from superantigensensitive DQ8 mice infected with GAS-M3 smeZ-M89 (White box-whisker) or GAS-M3 smeZ-M3 (Grey box-whisker). Five mice per group were infected intramuscularly and after 24 hours blood was removed by cardiac puncture. Cytokines were measured using LuminexH. Dotted horizontal line; lowest detectable level of each cytokine. For analysis, samples with undetectable levels of cytokine were assigned a value half the lowest detectable value. (TIF)