Parameters Governing Invasive Disease Propensity of Non-M1 Serotype Group A Streptococci

Abstract

Group A Streptococcus (GAS) causes rare but life-threatening syndromes of necrotizing fasciitis and toxic shock-like syndrome in humans. The GAS serotype M1T1 clone has globally disseminated, and mutations in the control of virulence regulatory sensor kinase (covRS) operon correlate with severe invasive disease. Here, a cohort of non-M1 GAS was screened to determine whether mutation in covRS triggers systemic dissemination in divergent M serotypes. A GAS disease model defining parameters governing invasive propensity of differing M types is proposed. The vast majority of GAS infection is benign. Nonetheless, many divergent M types possess limited capacity to cause invasive infection. M1T1 GAS readily switch to a covRS mutant form that is neutrophil resistant and frequently associated with systemic infection. Whilst non-M1 GAS are shown in this study to less frequently accumulate covRS mutations in vivo, such mutants are isolated from invasive infections and exhibit neutrophil resistance and enhanced virulence. The reduced capacity of non-M1 GAS to switch to the hypervirulent covRS mutant form provides an explanation for the comparatively less frequent isolation of non-M1 serotypes from invasive human infections.

Introduction

Group A Streptococcus (GAS) causes 700 million infections each year, resulting in over 500,000 deaths. Invasive GAS infections account for >600,000 disease episodes, incurring a death rate of approximately 25% [1]. Over the past 30 years, a resurgence of life-threatening invasive GAS pathologies has been documented, in parallel with the emergence of the globally disseminated M1T1 clone [2, 3]. The M1T1 clone remains the most frequently isolated M serotype from cases of invasive GAS infection and also from simple pharyngitis [4]. Nonetheless, other GAS serotypes cause invasive infections in Western populations, and GAS invasive disease in indigenous populations, such as Aboriginal communities of Northern Australia, is often associated with multiple M serotypes [5].

Historically, it is documented that hyperencapsulated GAS isolates are associated with invasive human infection and increased virulence in murine models [6, 7, 8, 9, 10]. Similarly, an inverse correlation has been described between SpeB production and disease severity in both human clinical disease and in murine models [11, 12]. In the GAS M1T1 clone, mutations in the control of virulence regulatory sensor kinase (covRS; alternatively designated csrRS) operon are selected for in vivo, and result in upregulation of capsule, loss of SpeB expression, increased disease severity in murine infection models and are more frequently associated with severe human invasive disease [10, 13, 14, 15].

The GAS M1T1 clone is distinguished from closely related M1 strains by the acquisition of the bacteriophage-encoded DNase Sda1 and superantigen SpeA [16, 17]. Human neutrophil-mediated killing of GAS selects for the neutrophil-resistant covRS mutant form of M1T1. The acquisition of the bacteriophage-encoded sda1 gene provided M1T1 with enhanced capacity to switch to the covRS mutant form [14], as Sda1 mediates escape from neutrophil extracellular traps [18, 19]. The loss of SpeB-mediated proteolytic degradation in vivo, as a result of covRS mutation, preserves expression of Sda1 and other virulence factors [20], allowing GAS to recruit and activate the broad-spectrum human protease plasmin on the bacterial surface, resulting in extensive tissue destruction and triggering systemic dissemination [11, 14].

While significant advances in the understanding of GAS M1T1 invasive disease initiation have been made, parameters governing invasive propensity of other M types have not been elucidated. Recently, analysis of GAS isolates of varying M type documented an association between invasive clinical isolates and mutations in genes encoding GAS global gene regulators (covRS/csrRS and ropB/rgg)[21]. In this study, we have examined a set of non-M1 serotype GAS isolates to determine whether such mutations trigger systemic dissemination in divergent M types. A model describing the invasive potential of differing M types is proposed.

Materials and Methods

GAS Strains and Culture Conditions

Clinical GAS isolates examined in this study have been described previously (table 1). Routine culture of GAS was conducted in stasis at 37°C in Todd-Hewitt broth supplemented with 1% (w/v) yeast extract or on horse-blood agar. GAS cultures for use in microarray experiments were propagated in Todd-Hewitt broth supplemented with 1.5% (w/v) yeast extract.

Table 1.

Characteristics of GAS isolates and mutant strains utilized

Isolate	emm type	Isolate origin	Clinical origin	Reference or source
5448	1.0	USA	invasive; STSS/NF	[20]
5448 AP	1.0	animal passage	generated during murine passage	[20]
NS13	53	Australia	invasive; blood	[37]
NS88.2	98.1	Australia	invasive; blood	[37]
NS179	9.1	Australia	invasive; blood; pustules on foot	[37]
NS210	22	Australia	invasive; diabetic ulcer with fever	[37]
NS223	91	Australia	invasive; blood	[37]
NS452	25	Australia	invasive; cellulitis; wound	[37]
NS455	52	Australia	invasive; blood	[37]
NS501	14	Australia	invasive; blood	[37]
A20	23	Japan	invasive; blood	[41]
NS730	90	Australia	invasive; nf; pus from left hip	[37]
NS733	90	Australia	invasive; nf; wrist aspirate	[37]
NS931	69	Australia	invasive; nf; blood	[37]
NS1133	101	Australia	invasive; blood	[37]
ALAB49	53	USA	superficial; impetigo; skin lesion	[42]
NS10	53	Australia	superficial; throat swab	[37]
NS14	102	Australia	superficial; post-operative wound	[37]
NS32	101	Australia	superficial; wound infection	[37]
NS50.1	108	Australia	superficial; wound infection	[37]
NS53	71	Australia	superficial; fever	[37]
NS59	53	Australia	superficial; wound infection	[37]
NS236	77	Australia	superficial; sore throat; throat swab	[37]
NS253	52	Australia	superficial; wound infection	[37]
NS265	56	Australia	superficial; wound infection	[37]
NS297	44/61	Australia	superficial; skin sore	[37]
NS474	58	Australia	superficial; wound infection	[37]
NS488	12	Australia	superficial; sinusitis; pharyngeal pus	[37]
NS836	ck249	Australia	superficial; wound infection	[37]
NS88.2rep	98.1	NA	sogenic covS repaired NS88.2 strain	this study
NS88.2covS	98.1	NA	reverse complemented NS88.2 covS mutant	this study

Clinical origin classified as invasive if infected tissue is normally sterile in a healthy host. STSS = Streptococcal toxic shock-like syndrome; NF = necrotising fasciitis; NA = not applicable.

SpeB Activity Assays

SpeB cysteine protease activity in cell-free stationary-phase supernatants was determined using the chromogenic substrate N-benzoyl-Pro-Phe-Arg-p-nitroanilide-hydrochloride (Sigma), according to the method of Hytönen et al. [22]. To screen large numbers of GAS colonies recovered following subcutaneous murine passage (n = 1,500), single colonies were transferred to designated grid locations on Columbia agar plates supplemented with 15% (v/v) commercial skim milk (Devondale) and assayed for secreted SpeB activity as described by Ashbaugh et al. [23].

Western Blot Analysis

Stationary-phase supernatant proteins were concentrated 37.5-fold in 100 mM Tris (pH 7.6) by precipitation with 10% trichloroacetic acid. SpeB protein was then detected using Western blot analysis essentially as previously described [11].

Quantification of Hyaluronic Acid Capsule Biomass

Overnight cultures in Todd-Hewitt broth supplemented with 1% (w/v) yeast extract were subinoculated 1:14 and grown to mid-logarithmic phase (OD₆₀₀ 0.6). Capsule extraction and quantification were conducted using the method of Ashbaugh and Wessels [24].

DNA Sequence Analysis of the covRS Locus

The method for mapping mutations in the GAS covRS operon was as described previously [14]. Genomic DNA was isolated using the QIAGEN DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturer's guidelines. For each operon, sequences were assembled in Chromas Pro v1.33 (Technelysium Pty Ltd) and aligned with the intact covRS operon of the M1T1 isolate 5448 (BioEdit v7.0.9.0; Ibis Biosciences).

Microarray Design and Production

An oligonucleotide microarray with probes representing M1 core ORFeome in addition to ORFs representing various M1, M3, M18 and Streptococcus dysgalactiae prophages was used in this study. The microarray was an expansion of one described previously [25]. Oligomers (70mers) were obtained from Dr. Kevin McIver and Dr. June Scott, and printed in the Molecular Resource Center, University of Tennessee Health Science Center by the use of MicroGrid II (Genomic Solutions). Additional 70mers, representing M1T1-specific prophages and prophage 3396 of S. dysgalactiae, were designed in batch according to the same design criteria applied for the other oligomers (Oligo Wiz 2.0, http://www.cbs.dtu.dk/services/OligoWiz) and obtained from Integrated DNA Technologies.

DNA-DNA Microarray

DNA-DNA microarray experiments were conducted in dye-flipped biological triplicates for each GAS isolate, and all steps involving Alexa Fluor®dyes were conducted in the dark. RNA-free genomic DNA was extracted from overnight liquid cultures by a modified phenol-chloroform procedure [17] and randomly sheared into <1-kb fragments using a Misonix 3000 cup-horn sonicator and Branson Sonifier®250. Sheared DNA samples were fluorescently labeled with the BioPrime®Total Genomic Labeling System (Invitrogen) as described by the manufacturer. In each hybridization reaction, equal amounts of Alexa Fluor® 3-labeled and Alexa Fluor® 5-labeled samples from different pairs of isolates were combined with hybridization buffer (Genisphere) and applied to the microarray slide. Following incubation at 55°C for 16 h, glass slides were washed, dried via centrifugation and scanned using a GenePix 4000B scanner (Axon Instruments Inc.).

DNA Microarray Data Analysis

The GenePixPro 4.0 software (Axon Instruments Inc.) was used for primary analysis of the scanned GenePix files. The fluorescent intensities were then normalized to the median intensity for each channel. Data from all probes representing the same gene were averaged, and a mean hybridization score was calculated for each gene. An average threshold of 40 median-normalized fluorescence units was selected, under which a gene was called ‘absent’.

Transcriptional Microarray

Overnight GAS cultures were sub-inoculated 1:10 into fresh prewarmed media and grown to mid-logarithmic phase (OD₆₀₀ 0.4). Bacteria were concentrated 20-fold in Buffer RLT (Qiagen) containing β-mercaptoethanol, lysed by mechanical disruption in Lysing matrix B tubes (Q-Biogene) with a FastPrep FP120 Homogenizer (Q-Biogene) and flash frozen for storage at −80°C. Bacterial RNA was extracted using the RNeasy Mini Kit (Qiagen), treated with TURBO DNA-free™DNase to remove contaminating genomic DNA (Ambion), re-concentrated on RNeasy columns (Qiagen) and converted to dendrimer-labeled cDNA with the Genisphere 3DNA Array 900MPX Kit as described by the corresponding manufacturer. Dendrimer-labeled cDNA samples from different pairs of isolates were combined and hybridized to the microarray slide for 16 h at 55°C. Slides were washed to remove unbound cDNA and labeled with dendrimer-targeted Alexa Fluor 546 and Alexa Fluor 647 dyes for 5 h at 55°C. Following a final wash to remove excess fluorescent dye, slides were coated in DyeSaver 2 (Genisphere) to preserve the fluorescent signal, and subsequently polished with toluene/acetone (3:1, v/v) to minimize background fluorescence immediately prior to scanning in a GenePix 4000B scanner (Axon Instruments Inc.). Scanned images were processed with GenePixPro 4.0 software (Axon Instruments Inc.), and all transcriptional and statistical analyses undertaken in silico using GeneSpring GX 10 (Agilent Technologies). All transcriptional microarray data were submitted to the NCBI Gene Expression Omnibus (GEO) according to the MIAME standards (GEO accession No. GSE23825).

Neutrophil Killing Assays

The capacity of GAS isolates to survive during co-incubation with human neutrophils in vitro was determined as described by Hollands et al. [25]. Briefly, 2 × 10⁴ colony-forming units (CFU) of mid-logarithmic-phase bacteria were incubated with 2 × 10⁵ neutrophils in RPMI with 2% heat-inactivated plasma for 30 min at 37°C. The final percent survival was calculated following comparison to the same bacterial culture incubated under the same conditions in the absence of neutrophils.

In vivo Phase-Switching

To examine the capacity for a phenotypic phase switch through covRS mutation in vivo, sublethal doses of SpeB-positive GAS isolates, in the order of 10⁷ CFU per dose, were subcutaneously administered in sterile 0.7% (w/v) NaCl to the right flank of C57BL/J6 mice less than 8 months of age (10 animals per isolate). On the third day after infection, mice were sacrificed by CO₂ asphyxiation and the infected cutaneous lesions surgically removed. In vivo passaged bacteria were recovered from murine lesions on horse-blood agar and single colonies assayed for SpeB status as outlined above.

Transgenic Murine Infection Model

Humanized plasminogen transgenic (Tg+)AlbPLG1 mice, heterozygous for the human plasminogen gene [26], served as the animal model for determining GAS invasive potential as previously described [14]. GAS isolates were grown to logarithmic phase (OD₆₀₀ 0.6), washed with sterile 0.7% (w/v) NaCl and appropriately diluted to prepare the inoculum. Final dose of viable bacteria was confirmed using a plate-based serial dilution cultured on horse-blood agar. For each GAS isolate, a cohort of 10 humanized mice were subcutaneously challenged in the right flank and mortality was documented over a 10-day period.

Restoration of covS in NS88.2

The covS gene from NS88.2 was amplified using primers pHYcovSF (5′-gggggatccatggaaaatcagaaacaaaaacag-3′) and pHYcovSR (5′-ggggaattcctaactctctttagactgggcc-3′). The resulting amplicon was cloned into the temperature-sensitive vector pHY304 using BamHI/EcoRI restriction-enzyme digestion and ligation with T4 DNA ligase. Site-directed mutagenesis of the adenine nucleotide at position 581 to guanine was performed according to the method of Sanderson-Smith et al. [27], using primers pHYcovsa581gF (5′-gccaaataactcaacaactagtagccaaaacagcagtcatgc-3′) and pHYcovsa581gR (5′-gcaagactgctgttttggctactagttgttgagttatttggc-3′). The resulting plasmids (pHYcovS and pHYrep) were transformed into Escherichia coli MC1061 using standard electroporation procedures. Allelic replacement mutants were constructed as described previously [28]. The covRS operons of the isogenic mutants NS88.2rep and NS88.2covS were sequenced as outlined above to confirm both the presence of the desired mutation and the integrity of the covRS operon.

GAS Surface Plasmin Activity Assays

GAS were incubated in human plasma as described previously [11]. Overnight growth in Todd-Hewitt broth supplemented with 1% (w/v) yeast extract was diluted to OD₆₀₀ 0.5 and co-incubated with human plasma at 37°C for 3 h. GAS were twice washed with PBS, 0.01% gelatin and 0.01 M EDTA, prior to resuspension in PBS and 0.01% gelatin. Plasmin activity was determined using the chromogenic substrate Spectrozyme PL (American Diagnostica).

Results

Expression of SpeB, Capsule Production and covRS Mutation of Non-M1 GAS

A range of GAS clinical isolates of differing emm types from invasive and benign infections were selected for this study (table 1). The well-characterized M1T1 clinical isolate 5448 and the natural isogenic covS mutant 5448AP [14] were included for comparison. With the exception of 3 strains, 5448AP, NS88.2 and A20, each of the GAS isolates described here expressed and secreted active SpeB at the stationary phase of growth(fig. 1a, b). An inverse correlation between SpeB expression and hyaluronic acid capsule production was observed, with 5448AP, NS88.2 and A20 hyperencapsulated with respect to all SpeB-positive GAS isolates (fig. 1c). GAS M1T1 strain 5448AP contains an adenine insertion at nucleotide position 887 in the ORF of the covS gene, resulting in the premature truncation of the translated CovS protein [14]. Correspondingly, DNA sequence analysis of the SpeB-negative hyperencapsulated strains NS88.2 and A20 revealed a guanine to adenine substitution at position 581 in the covS gene of NS88.2 and a 5-base pair deletion between positions 1,215 and 1,219 of the A20 covS gene (fig. 1d). Each mutation results in premature truncation of the translated CovS protein.

Fig. 1.

Molecular and phenotypic analyses of GAS isolates representing distinct M serotypes. a SpeB activity in cell-free stationary-phase supernatants of invasive (filled bars) and uncomplicated infection (open bars) GAS isolates. Isolate 5448 represents the globally disseminated M1T1 clone, while 5448AP is a hypervirulent animal passaged variant of 5448 [14]. b Western blot detection of SpeB in stationary-phase GAS supernatants. The 28-kDa mature SpeB protease is indicated with an arrowhead. c Hyaluronic acid capsule biomass of mid-logarithmic-phase invasive (filled bars) and uncomplicated infection (open bars) GAS isolates. d Schematic representation of the covRS operon. DNA sequence analysis confirmed the presence of inactivating covS mutations in the SpeB-deficient isolates A20, NS88.2 and 5448AP. The nature and nucleotide positions of the mutations in each isolate are indicated by the corresponding arrowheads.

DNA Microarray and Transcriptomic Analyses

Three GAS strains with intact covRS (5448, ALAB49 and NS730) and 3 covS mutant forms (5448AP, NS88.2 and A20) were subjected to DNA microarray and transcriptomic analyses. DNA microarray identified 1,580 genes representing the core genome of the 6 GAS strains under examination (online suppl. table 1 and online suppl. fig. 1, www.karger.com/doi/10.1159/000317640). The bacteriophage-encoded sda1 gene, which confers on M1T1 GAS the capacity to switch to the covRS mutant form at high frequency [14], was not present in the genomes of the non-M1 GAS strains (online suppl. table 1 and online suppl. fig. 1). Only ubiquitous genes were included in subsequent transcriptomic analyses undertaken on GAS strains grown to mid-logarithmic phase. The virulence-related genes found to be strongly upregulated in the covS mutant strains, in comparison to the covRS intact strain 5448, include genes of the has operon (hasA, hasB and hasC; capsule biosynthesis), slo (streptolysin O) and the spyA exotoxin. Corroborating studies on M1T1 covS mutation [13], the positive regulators of SpeB activity and expression ropA and ropB [29, 30], in addition to genes of the sag operon (sagB, sagC and sagH), involved in streptolysin S production [31], were found to be downregulated in the covS mutant strains studied here (fig. 2a). No significant SpeB expression was detected at the mid-logarithmic phase of growth, in accordance with previous studies [32]. Principal component analysis of these transcriptomic data revealed 2 distinct clusters of covRS intact strains and covS mutant forms (fig. 2b). These data suggest that differing M types harboring covS mutations express related transcriptomic profiles.

Fig. 2.

In vitro mid-logarithmic phase transcriptional microarray analysis of covRS intact and covRS mutant non-M1 GAS. a log₁₀-fold differential expression of virulence-associated and regulatory genes of the covS mutants 5448AP (filled circles), A20 (open circles) and NS88.2 (shaded circles) compared to the covRS intact M1T1 strain 5448. Selected genes are significantly differentially expressed in 5448AP with respect to 5448 (p < 0.05). b Principal component analysis on the non-M1 isolates ALAB49 (M53) and NS730 (M90), NS88.2 (M98.1) and A20 (M23), in addition to the M1T1 reference strains 5448 and 5448AP, revealed 2 distinct expression profiles in this 6-isolate strain set. Solid-line boxes placed in the plot area highlight strain clusters. The dashed-line box highlights the M1T1 strains 5448 and 5448AP, which, apart from a single base insertion in the 5448AP covS gene, harbor identical genomes.

Neutrophil Resistance, SpeB Switching and Virulence

In comparison with the covRS intact strains 5448, NS730 and ALAB49, the hyperencapsulated covS mutant strains 5448AP, NS88.2 and A20 displayed enhanced resistance to human neutrophil killing (fig. 3a). Of the covRS intact strains, the M1T1 strain 5448 readily switched to the SpeB-negative covRS mutant form in vivo (fig. 3b), in accordance with previous studies [11, 14, 20]. However, the covRS intact non-M1 strains NS730 and ALAB49 only infrequently switch to a covRS mutant form in vivo (fig. 3b). Utilizing the humanized plasminogen transgenic mouse line AlbPLG1, we assessed the virulence of both covRS intact and mutant GAS strains. The M1T1 strain 5448, which has the capacity to switch at a high frequency to the covRS mutant form in vivo, was virulent in this mouse model (fig. 3c), corroborating previous work [11, 14]. The covRS intact strains NS730 and ALAB49, which infrequently switch to the covRS mutant form (fig. 3b), failed to establish a lethal infection (fig. 3d, e). Each of the covS mutant strains, 5448AP, NS88.2 and A20, were highly virulent (fig. 3f–h). These data suggest that while covRS mutation may occur only infrequently in non-M1 GAS, such covRS mutant forms are hypervirulent.

Fig. 3.

Characterization of the covRS intact isolates ALAB49 (M53) and NS730 (M90), and the covS mutant isolates NS88.2 (M98.1) and A20 (M23), in comparison with the M1T1 reference strains 5448 and 5448AP. a The percent survival of GAS isolates during co-culture with human neutrophils in vitro. b The capacity of SpeB-positive isolates to phase-switch to a SpeB-negative phenotype was assessed following a 3-day subcutaneous passage in C57BL/J6 mice. Each data point represents the percent of SpeBnegative covRS mutants recovered from the infection site of each mouse (n = 10 mice per strain). Subcutaneous infection of humanized plasminogen transgenic AlbPLG1 (n = 10) 5448 (3.9 × 10⁷ CFU/dose) (c), ALAB49 (3.7 × 10⁸ CFU/dose) (d), NS730 (2.2 × 10⁸ CFU/dose) (e), 5448AP (5.1 × 10⁷ CFU/dose) (f), NS88.2 (2.0 × 10⁷ CFU/dose) (g) and A20 (1.2 × 10⁸ CFU/dose) (h).

Repair of the covS Mutation in GAS Strain NS88.2 and Phenotypic Characterization

In order to investigate whether low-frequency switching of non-M1 GAS results in the hypervirulent covRS mutant form, the non-M1 GAS strain NS88.2 harboring a covS mutation was chosen. The adenine nucleotide point mutation at position 581 in the NS88.2 covS gene (fig. 1d) was converted to guanine by allelic replacement mutagenesis to construct strain NS88.2rep, with an intact or ‘repaired’ covS gene. Then, in order to fulfill Koch's molecular postulates [33], allelic replacement mutagenesis was undertaken on NS88.2rep to restore the original adenine nucleotide point mutation, resulting in strain NS88.2covS.

Repair of the covS mutation restored SpeB expression and activity in NS88.2rep, while NS88.2 and NS88.2covS remained SpeB-negative (fig. 4a, b). NS88.2 and NS88.2covS were hyperencapsulated in comparison to NS88.2rep (fig. 4c), and acquired substantial surface plasmin activity following incubation in human plasma (fig. 4d). Both NS88.2 and NS88.2covS also displayed enhanced resistance to killing by human neutrophils (fig. 4e). The covS intact NS88.2rep displayed only limited capacity to switch to the more virulent covRS mutant form (fig. 4f) and was not virulent in comparison to NS88.2 and NS88.2covS in AlbPLG1 mice (fig. 4g). These data support the hypothesis that non-M1 GAS serotypes switch less frequently to the hypervirulent covRS mutant form, providing an explanation for the comparatively less frequent isolation of non-M1 serotypes from invasive human infections.

Fig. 4.

Characterization of NS88.2 covS mutant (NS88.2 and NS88.2 covS) and intact (NS88.2 rep) isogenic strains. a SpeB activity in stationary-phase GAS supernatants. b Western blot detection of SpeB in stationary-phase supernatants. The mature SpeB protease is indicated with an arrowhead. c Hyaluronic acid capsule biomass of mid-logarithmic phase isogenic GAS strains. d Acquired surface plasmin activity following incubation in human plasma. e Percent survival following co-culture with human neutrophils in vitro. f The capacity of the SpeB-positive NS88.2 rep to phase-switch to a SpeB-negative phenotype following a 3-day subcutaneous passage in C57BL/J6 mice. g Subcutaneous infection of humanised plasminogen transgenic A lbPLG1 mice (n = 10 mice per strain) with NS88.2 (1.0 × 10⁷ CFU/dose), NS88.2 rep (1.3 × 10⁷ CFU/dose) and NS88.2 covS (9.8 × 10⁶ CFU/dose).

Discussion

Severe group A streptococcal invasive disease progresses rapidly and results in high patient morbidity, with approximately one quarter of cases being fatal despite the susceptibility of the pathogen to antibiotic treatment [1, 34]. Host genetic factors [35] and the human fibrinolytic protease plasmin [26, 36] have both been documented as contributing to GAS invasive disease potential. Whilst progression of GAS disease from benign mucosal infections to invasive disease occurs infrequently, the M1T1 clone is clinically and epidemiologically associated with deep tissue infections in Western countries [4]. Mutations in the M1T1 covRS locus result in a hyperencapsulated, SpeB-negative, hypervirulent phenotype, and are correlated with invasive diagnosis in patients [10, 12, 13]. The capacity of M1T1 to switch to the covRS mutant form at high frequency may result from acquisition of the phage borne sda1 gene by M1T1 which confers resistance to neutrophil killing [14].

The contribution of similar mutations in covRS to invasive disease potential of non-M1 GAS remains largely unknown. Many differing M types have been associated with invasive infections [21, 34, 37, 38]. A naturally occurring mutation in the covR gene of an M3 isolate from a case of streptococcal toxic shock-like syndrome has been described. This covR mutation was associated with increased capsule expression and enhanced virulence [39]. Mutation in covR, associated with enhanced expression of the interleukin-8 cleaving protease SpyCEP, has also been observed in an M81 serotype GAS isolated from a lethal case of bacteremia and necrotizing fasciitis [40]. A correlation between mutation in global gene regulators (covRS and ropB) and invasive pathology has been documented in a range of M types. In comparison to mutations in ropB, inactivation of covS is clinically predominant and results in greater virulence in murine infection models [21].

In this study, a variety of non-M1 GAS strains from invasive disease episodes were examined, including isolates representing the emm sequence types14,22,23,25,52,53,69,90,91,98.1 and101. These included SpeB-positive isolates and SpeB-negative covS mutant forms. We propose that divergent M types with intact covRS possess an underlying capacity to cause invasive infection. In each serotype background, the covRS mutant form represents a more virulent state, which has greater propensity to cause invasive infection.

Using allelic replacement mutagenesis, we demonstrate that repair of the covS defect in the invasive emm98.1 GAS strain NS88.2 (covS mutant form) renders the isogenic covS intact strain NS88.2rep SpeB-positive, susceptible to neutrophil killing and less able to accumulate surface plasmin activity following growth in human plasma. The covS intact NS88.2rep strain was also found to be avirulent in the humanized plasminogen transgenic mouse model. In comparison to the M1T1 GAS covS intact strain 5448, we propose that the lack of virulence of NS88.2rep is due to the reduced capacity of this strain to switch to the invasive covRS mutant form in vivo. The lower frequency of switching to the covS mutant form limits the number of these invasive variants at the site of local infection. This lower frequency of switching may reflect the absence of the sda1 gene in this genetic background. The reduced capacity of non-M1 GAS serotypes to switch to the hypervirulent covRS mutant form may provide an explanation for the comparatively less frequent isolation of non-M1 serotypes from invasive human infections.