Incidence and Effects of Acquisition of the Phage-Encoded ssa Superantigen Gene in Invasive Group A Streptococcus

The acquisition of the phage-encoded superantigen ssa by scarlet fever-associated group A Streptococcus (Streptococcus pyogenes, GAS) is found in North Asia. Nonetheless, the impact of acquiring ssa by GAS in invasive infections is unclear. This study initially analyzed the prevalence of ssa+ GAS among isolates from sterile tissues and blood. Among 220 isolates in northern Taiwan, the prevalence of ssa+ isolates increased from 1.5% in 2008–2010 to 40% in 2017–2019. Spontaneous mutations in covR/covS, which result in the functional loss of capacity to phosphorylate CovR, are frequently recovered from GAS invasive infection cases. Consistent with this, Phostag western blot results indicated that among the invasive infection isolates studied, 10% of the ssa+ isolates lacked detectable phosphorylated CovR. Transcription of ssa is upregulated in the covS mutant. Furthermore, in emm1 and emm12 covS mutants, ssa deletion significantly reduced their capacity to grow in human whole blood. Finally, this study showed that the ssa gene could be transferred from emm12-type isolates to the emm1-type wild-type strain and covS mutants through phage infection and lysogenic conversion. As the prevalence of ssa+ isolates increased significantly, the role of streptococcal superantigen in GAS pathogenesis, particularly in invasive covR/covS mutants, should be further analyzed.


INTRODUCTION
Streptococcus pyogenes [group A Streptococcus (GAS)] is a gram-positive bacterium that causes diseases like pharyngitis, pyoderma, scarlet fever, necrotizing fasciitis, and toxic shock syndrome. In 2011, a scarlet fever outbreak was reported in Hong Kong (Hsieh and Huang, 2011;Tse et al., 2012). The scarlet fever isolates typically harbor phage-associated superantigen genes ssa and speC, and the DNase gene spd1 (Tse et al., 2012;Ben Zakour et al., 2015). Luk et al. (2012) found insufficient evidence to support the association between increased scarlet fever incidence or severity and a particular emm type, virulence gene profile, or the presence of specific foreign genetic elements. Nonetheless, Davies et al. (2015) showed that ssa is absent in the clade not associated with scarlet fever, but variably present (95% isolates) in scarlet feverassociated clades, suggesting that ssa acquisition could be potentially related to the expansion of scarlet fever-associated emm12-type clones within the Hong Kong population.
An recent increased incidence of scarlet fever has also been reported in South Korea, Singapore, England, and Germany (Turner et al., 2016;Park et al., 2017;Brockmann et al., 2018;Kim and Cheong, 2018;Lamagni et al., 2018;Walker and Brouwer, 2018;You et al., 2018;Yung and Thoon, 2018). Park et al. (2017) showed that emm4, emm28, emm1, and emm3 contributed to scarlet fever in South Korea. Notably, antibiotic resistance was uncommon in these scarlet fever-associated isolates (Park et al., 2017). Interestingly in England, the most prevalent isolates recovered during a national increase in scarlet fever incidence in 2016 were an emergent clone of the emm1 M1T1 genetic lineage, designated M1T1 UK , which was characterized by increased production of the phage-encoded SpeA superantigen (Lynskey et al., 2019). This contrasts with the results in Hong Kong and Beijing where during an increased incidence of scarlet fever cases in 2011. The most prevalent isolates recovered were polyclonal emm12 strains and the vast majority of which were characterized by the presence of the phage-encoded streptococcal superantigen (SSA; Davies et al., 2015). These findings show that emm type, antibiotic-resistant phenotype, and superantigen gene in scarlet fever-associated isolates vary from country to country; therefore, the specific factors leading to scarlet fever resurgence are not completely understood.
Streptococcal superantigen is a 260-residue protein with 60% sequence identify to that of staphylococcal enterotoxin B (SEB; Reda et al., 1994). Brouwer et al. (2020) showed that glutathione from streptolysin O (SLO)-lysed host cells not only enhances SSA production but also activates its superantigen activity. ssa has been detected in the toxic shock syndrome related M3 isolates (Mollick et al., 1993), suggesting it to be a potential virulence factor of GAS. As the prevalence of ssa+ GAS isolates was >70% in certain geographic areas (Li et al., 2020a), the effects of ssa acquisition or ssa+ isolate dissemination on GAS diseases need to be further clarified.
Previous studies have reported that isolates with spontaneous mutations in the covR/covS operon are highly related to severe diseases such as necrotizing fasciitis and toxic shock syndrome (Ikebe et al., 2010;Friaes et al., 2015). CovR/CovS is a two-component regulatory system in GAS, and intracellular CovR is phosphorylated by the sensor kinase CovS (Trevino et al., 2009;Tran-Winkler et al., 2011). Mutations in the covR/covS operon increase the production of virulence factors such as hyaluronic acid capsule, M protein, SLO, and streptokinase (Tran-Winkler et al., 2011;Chiang-Ni et al., 2017;Horstmann et al., 2018). Furthermore, CovR/CovS also regulates the expression of phage-related genes such as DNase sda1 (Walker et al., 2007). Increased Sda1 expression by a covS mutant promoted bacterial escape from immune clearance by degrading the DNA structure of the neutrophil extracellular traps (Walker et al., 2007).
In this study, we found that four ssa+ isolates lacked phosphorylated CovR, and ssa transcription was upregulated in these isolates. In line with the repression in ssa and slo transcription by the CovR/CovS system, ssa deletion in the emm1 and emm12 covS mutants attenuated bacterial growth activity in human blood, suggesting that ssa acquisition by covS mutants or spontaneous mutations in the covR/covS operon of ssa+ isolates could enhance bacterial survival during infection. As the prevalence of ssa+ isolates dramatically increased in northern Taiwan, the impact of the dissemination of ssa+ isolates on our society should be continuously monitored and studied.

Erythromycin and Clindamycin Susceptibility Test
The susceptibility of GAS isolates to erythromycin and clindamycin was determined using a disk diffusion assay according to the Clinical and Laboratory Standards Institute Guideline (Clinical and Laboratory Standards Institute, 2018). Inducible clindamycin resistance in GAS isolates was determined using the D test with erythromycin and clindamycin disks. A flattened inhibition zone around the clindamycin disk proximal to the erythromycin disk was considered a positive result.

DNA Manipulations
Genomic DNA was extracted using a previously described method (Chiang-Ni et al., 2019a), and emm typing, PCR amplification, and DNA sequencing were performed according to the protocol from the Centers for Disease Control and Prevention. 1 The ssa and erythromycin resistance genes (mefA, ermB, and ermTR) were screened for by PCR amplification using previously described primers (Table 1; Meisal et al., 2010;Villasenor-Sierra et al., 2012).

Southern Blot Hybridization
Group A Streptococcus chromosomal DNA was digested using HindIII, and the DNA fragments were resolved on 0.9% agarose gel. The PCR product of the chloramphenicol cassette (716 bp; Table 1) was labeled with alkaline phosphatase as the probe, and DNA hybridization was performed according to the manufacturer's instructions (AlkPhos Direct Labeling and Detection System; GE Healthcare UK Limited; Amersham, United Kingdom). The signal was detected using a Gel Doc XR+ system (Bio-Rad; Hercules, CA, United States).
ssa and covS Isogenic Mutant Construction ssa and flanking upstream and downstream regions (1805 bp) were amplified using primers ssa-BamHI-F and ssa-BamHI-R (Table 1), and the PCR product was ligated into a T-A cloning vector (Yeastern Biotech). ssa was removed by PCR with reverse primers ssa-SacII-F and ssa-SacII-R (Table 1) and replaced by the chloramphenicol cassette from Vector78 (Tsou et al., 2010) at the SacII site. The ssa knockout DNA fragment was sub-cloned into the temperature-sensitive vector pCN143 (Chiang-Ni et al., 2016) at the BamHI site (designed as pCN211). The plasmids pCN211 and pCN160 (Chiang-Ni et al., 2019a) were transformed into emm12 SPY128 and emm1 SPY131, respectively, by electroporation for allelic exchange, and the transformants were selected using 3 μg/ml chloramphenicol as described previously (Chiang-Ni et al., 2019b). ssa and covS deletion in the selected transformants was confirmed by Sanger sequencing. To construct covS and ssa double mutants, pCN160 was transformed into ssa mutants. covS deletion in these ssa mutants was selected according to the encapsulated phenotype and confirmed by Sanger sequencing.

RNA Manipulation and Quantitative PCR Analysis
RNA extraction and reverse transcription were performed as previously described (Chiang-Ni et al., 2009). The bacterial strains were cultured for 6 h (exponential phase) and 8 h (stationary phase), and total RNA was extracted for analysis. Quantitative PCR was performed in a 20 μl mixture containing 1 μl cDNA, 0.8 μl primer (10 μM), and 10 μl Sensifast Lo-ROX premixture (Bioline, Ltd.; London, United Kingdom) according to the manufacturer's instructions. Three biological replicates were performed, and ssa expression level was normalized to that of gyrA and analyzed using the threshold cycle (∆∆C T ) method (Roche LightCycler ® 96 System; Roche Molecular Systems, Inc.; Pleasanton, CA, United States). All values from the control and experimental groups were divided by the mean value of the control samples before statistical analysis (Valcu and Valcu, 2011). Primers used to detect ssa ( Table 1) and gyrA were designed using Primer3 (v.0.4.0) according to the HKU360 sequence (NCBI accession no. CP003901.1).

Phostag Western Blot Hybridization and Western Blot Analysis
Phostag western blotting was performed as previously described (Chiang-Ni et al., 2019a). Briefly, 10 μg bacterial total protein was mixed with 6× loading dye (without boiling) and separated using 10% SDS-PAGE containing 10 μM Phostag (Wako Pure Chemical Industries, Ltd.; Richmond, VA, United States) and 0.5 M MnCl 2 for 120-140 min at 100 V at 4°C. For detecting SLO, 30 μl bacterial culture supernatants were mixed with 6× protein loading dye and separated by 12% SDS-PAGE. The separated proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore; Billerica, CA, United States), which was blocked with 5% skim milk PBST buffer (PBS containing 0.2% Tween 20) at 37°C for 1 h. CovR was detected by the anti-CovR serum (Chiang-Ni et al., 2016), and SLO was detected by the anti-SLO antibody (GeneTex; Irvine, CA, United States). The phosphorylated and nonphosphorylated CovR and SLO were visualized using a previously described method (Chiang-Ni et al., 2020).

Whole Blood Model
The growth of GAS strains in human whole blood was studied according to a previous study with modifications (Brouwer et al., 2020). Freshly drawn heparinized venous blood from healthy adults was aliquoted (360 μl) into wells of a 24-well plate. GAS strains were grown to the exponential phase (OD 600 = 0.6) in TSBY, washed, resuspended in 1× PBS buffer at ~5 × 10 5 colony forming unit (CFU)/ml, and added to whole blood in a final volume of 400 μl (~5 × 10 4 CFU/ml). After incubating for 1.5 h at 37°C, the growth of GAS strains was analyzed by plating serial dilutions on TSBY plates. Experiments were performed with blood from three different donors.

Phage Induction, Infection, and Lysogenic Conversion
Mitomycin C (0.2 μg/ml; Sigma-Aldrich; St. Louis, MO, United States) was added to the culture of donor strain grown to the early exponential phase and incubated at 37°C for another 4 h. Culture supernatants were passed through a 0.45 μm filter (Millipore Ireland Ltd., Co.; Cork, Ireland) to remove bacteria and large fragments, and phage particles were collected from the filtrate by ultra-centrifugation at 112,000 × g for 2 h at 10°C. The recipients were grown to the logarithmic phase in the presence of 5 mM CaCl 2 and then co-incubated with the phage particles at 37°C for 3 h. The bacterial suspension was plated on agar plates supplemented with chloramphenicol (3 μg/ml) and resistant convertants were collected for Southern blotting analysis.

Statistical Analysis
Statistical analysis was performed using Prism software, version 5 (GraphPad; San Diego, CA, United States). Significant differences between multiple groups were determined using the ANOVA. Post-test for ANOVA was analyzed using Tukey's honestly significant difference test. Statistical significance was set at p < 0.05.

CovR/CovS Negatively Regulates ssa Transcription
To investigate the effect of CovR phosphorylation inactivation on the ssa expression, isogenic covS mutants of emm1-type (SPY131) and emm12-type (SPY128) isolates were constructed, and ssa transcription was analyzed by quantitative PCR. The results showed that ssa expression in the covS mutants was 7-12-fold higher than that in the parental strains ( Figure 2C). To further verify that ssa expression is repressed by phosphorylated CovR, the covR/covS trans-complementary strains of the covS mutants were constructed, and the phosphorylated CovR and ssa transcription in the covR/covS trans-complementary strains were analyzed. Phostag western blot analysis showed that phosphorylated CovR was detected in the trans-complementary strains, but not in the vector-control strains and covS mutants ( Figure 2D). In addition, ssa transcription was repressed in the covR/covS trans-complementary strains compared to that in the covS mutants ( Figure 2D). These results indicate that ssa transcription is repressed by phosphorylated CovR. Brouwer et al. (2020) showed that the wild-type strain and ssa mutant had similar growth in human whole blood. ssa transcription and SLO production were upregulated in the covS mutants compared to the wild-type strains (Figures 2B,C); therefore, in this study, the role of SSA in covS mutant survival in human whole blood was further investigated. The ssa and covS double mutants of emm1-type SPY131 and emm12-type SPY128 were constructed. The growth of the wild-type strains, covS, and ssa mutants in the culture broth were similar (data not shown). covS mutants were more resistant to phagocytic killing (Sumby et al., 2006); in agreement with this, we observed that the covS mutants of SPY131 and SPY128 had better growth in human blood than the wild-type strains and ssa mutants (Figure 3). Furthermore, the results showed that ssa deletion in the covS mutants significantly attenuated bacterial growth in blood compared to the covS mutants (Figure 3), suggesting that SSA contributes to the survival of covS mutants in human whole blood.

Transfer of ssa by Phage
The ssa+ isolates that did not express the phosphorylated CovR protein were identified (Figure 2A). These isolates could acquire inactivating spontaneous mutations in covR/covS during infection. Nonetheless, the ssa gene is carried by phage and the possibility of covR/covS mutants directly acquired ssa by phage infection could not be excluded. Therefore, whether covS mutants could acquire ssa through phage infection and lysogenic conversion was further investigated. To select the lysogenic convertants after phage infection, ssa was replaced with the chloramphenicol (cm) cassette in the emm12-type SPY128 (SCN279), which was utilized as the ssa-encoding phage donor. After mitomycin C induction, the cm cassette, but not chromosomal speB, was detected in the filtered supernatant from SCN279 (Figure 4A), indicating that the target phage was released from SCN279. Next, the recipients, including the emm1-type A20 strain, its covS mutant AP3, and the CovS kinase-inactivated CovS H280A mutant (Chiang-Ni et al., 2016, 2019b, were incubated with the collected (D) Phosphorylated CovR expression and ssa transcription in vector-control (∆covS+Vec) and the covR/covS trans-complementary (∆covS+PcovR/S) strains. RNA was extracted for reverse transcription-PCR (RT-qPCR) analysis. *p < 0.05. CovR~P, phosphorylated CovR; CovR, nonphosphorylated CovR. Total protein is served as the internal loading control.
FIGURE 3 | Growth activity of SPY131 (emm1), SPY128 (emm12), their ssa mutants (∆ssa), covS mutants (∆covS), and covS and ssa double mutants (∆covS∆ssa) in human whole blood. Group A Streptococcus (GAS) strains were incubated with whole blood from donors for 1.5 h. The number of surviving bacteria in human blood was determined by plating and enumerating the colony forming units (CFUs), which determined growth relative to the initial inoculum. *p < 0.05.
Frontiers in Microbiology | www.frontiersin.org phage particles. We found that lysogenic convertants can be obtained. The cm cassette in the lysogenic convertants was integrated to the same insertion site of the chromosome and located on the same phage in donor strain SCN279 ( Figure 4B).
These results indicate that not only emm1 wild-type strain but also its covS mutants could acquire ssa by lysogenic conversion.

DISCUSSION
This study showed that the prevalence of ssa+ GAS isolates increased from 1.5% (1/65) in 2008-2010 to 40% (24/60) in 2017-2019 in northern Taiwan. A total of 29 isolates could not produce phosphorylated CovR protein, and four of these isolates were ssa+. We found that ssa transcription is repressed by phosphorylated CovR. Moreover, ssa gene deletion attenuated the growth activity of the covS mutants in human blood, suggesting that the acquisition of ssa by covR/covS mutants or spontaneous mutations in the covR/covS operon in the ssa+ isolates could be related to the increase in bacterial survival during infection. In this study, 10% ssa+ isolates did not produce phosphorylated CovR under the conditions tested. Spontaneous inactivating mutations in the covR/covS operon increases SLO expression in GAS (Sumby et al., 2006;Chiang-Ni et al., 2019a). In addition, some clinical isolates with spontaneous mutations or a truncated allele of rocA (regulator of cov; upstream regulator of CovR/CovS) also increased the SLO expression (Feng et al., 2017;Jain et al., 2017;Horstmann et al., 2018;Chiang-Ni et al., 2020). SSA is a thio-activated superantigen, and the GAS secreted SLO, which triggers glutathione release from host cells to activate SSA in vivo (Brouwer et al., 2020). Brouwer et al. (2020) showed that ssa deletion did not attenuate bacterial survival in human blood; however, our results showed that the covS and ssa double mutants had significantly decreased growth in human blood compared to the covS mutants (Figure 3). As expected (Sumby et al., 2006), the encapsulated covS mutant was resistant to phagocytosis (Figure 3), thus not cleared despite high expression (promoted by SLO) of the superantigen SSA, which might activate T cells and phagocytic cells. Okamoto et al. (2001) showed that macrophages in SEB-pretreated mice were less phagocytic than those in non-pretreated mice. SSA sequence is 60% identity to SEB. Based on these observations, although covS mutant and covS and ssa double mutant were both encapsulated, the covS and ssa double mutant might encounter higher pressure from macrophages than that of covS mutant. Nonetheless, these mutants were co-cultured with human whole blood for 1.5 h; whether SSA could act to macrophages like SEB during this short incubation period is not clear. Therefore, the role of SSA in GAS survival in human blood needs to be further elucidated. Whether the ssa transcription is directly regulated by phosphorylated CovR is not clear; however, our results suggest that SSA could potentially contribute to the pathogenesis of invasive covR/covS mutants. Lynskey et al. (2019) showed that the toxigenic M1T1 clone (M1T1 UK ) is related to the increased incidence of invasive GAS disease in the United Kingdom. Thereafter, the M1T1 UK strain was identified in Canada, the Netherlands, and the United States (Demczuk et al., 2019;Li et al., 2020b;Rumke et al., 2020), demonstrating the possibility of clonal expansion of the M1T1 UK clone. In Taiwan, Yan et al. (2000) showed that 64.3% erythromycin-resistant isolates were clindamycin-susceptible and harbored mefA before 1998. From 2001 to 2010, the erythromycin resistance rate decreased from 53.1 to 0% in southern Taiwan (Chuang et al., 2015). In this study, we found that the erythromycin resistance rate in GAS increased from 3.1% in 2008-2010 to 25% in 2017-2019. Notably, only one isolate harbored mefA (1/28). The changes in phenotype and genotype of erythromycin-resistant GAS suggested that the expansion of erythromycin-resistant (ermB and ermTR) and ssa+ emm12 clones could occur in Taiwan. In the phage infection assay, ssa was replaced with a chloramphenicol cassette as the selection marker, and an identical genetic element was found in the emm12-type donor and emm1-type recipient strains using Southern blot (Figure 4B and data not shown), suggesting that ssa could be transferred by phage. Furthermore, we also found that the chloramphenicol cassette could be transferred between emm12-type and emm73-type isolates (data not shown). These results suggest that clonal expansion and phage infection could be involved in the increased prevalence of ssa+ isolates in Taiwan. The emerged clones in the United States (M1T1 clone) and in Asia (Hong Kong and Taiwan, emm12 clone) are different, but both possess superantigen genes such as speA and ssa. Nonetheless, the exact role of superantigens in the survival fitness of GAS during infection needs to be addressed.
In summary, this study showed an increased prevalence of ssa+ and erythromycin-resistant GAS isolates in northern Taiwan, and demonstrated that ssa acquisition by invasive covS mutants or spontaneous mutations in the covR/covS operon of ssapositive isolates could enhance bacterial growth in human blood. The role of SSA in GAS pathogenesis, particularly in invasive covR/covS mutants, should be further investigated.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Institutional Review Board (201900274B0 and 202000479B0) of the Chang Gung Memorial Hospital, Taiwan. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
CC-N, Y-SL, and C-YL contributed to the conceptualization, methodology, formal analysis, and writing the original draft. Y-SL, C-YL, C-YH, and Y-AS contributed to the methodology and investigation. Y-YC and C-HL contributed to the conceptualization and methodology. CC-N contributed to the writing -review and editing. CC-N and C-HC contributed to the funding acquisition. All authors contributed to the manuscript revision, read, and approved the submitted version.

FUNDING
This work was supported by parts of grants from the Chang Gung Memorial Hospital, Linkou, Taiwan (CMRPD1J0031-3 and CMRPD1K0331) and Ministry of Science and Technology, Taiwan (MOST 109-2320-B-182-036).

ACKNOWLEDGMENTS
We are grateful to the Bacterial Bank of Chang Gung Memorial Hospital, Linkou (Taiwan) for providing clinical GAS isolates.