Nine Lancefield group G SDSE strains were isolated from the Kanazawa University Hospital (Table 1): KNZ01 was isolated from the joint fluid; six strains (KNZ03, KNZ04, KNZ06, KNZ07, KNZ12, and KNZ15) were isolated from patients with bacteremia via blood cultures; KNZ10 and KNZ16 via swabbing open wounds. After Lancefield antigen and hemolysis tests on blood agar plates, the β-hemolytic group G strains were identified as SDSE using the API 20 Strep kit (bio Merieux). Group G SDSE strain ATCC 12394 was obtained from the American Type Culture Collection (ATCC)¹. SDSE strains KNZ01, KNZ03, and ATCC 12394 were cultured overnight in THY medium consisting of Todd-Hewitt broth (Becton Dickinson, Franklin Lakes, NJ, United States) and 0.2% yeast extract (Becton Dickinson) at 37°C under 5% CO₂ atmosphere. Before use in experiments, bacterial cells were cultured overnight, suspended in fresh THY medium, and incubated until the optical density at a wavelength of 600 nm (OD₆₀₀) reached from 0.01 to 0.4–0.5. All experiments were conducted in accordance with the WHO Laboratory biosafety manual and the institutional safety procedure of Kanazawa University.

The human keratinocyte cell line (HaCaT cells; Boukamp et al., 1988) was cultured in Dulbecco’s modified Eagle’s medium (DMEM; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) containing 8% fetal bovine serum (FBS; Biosera, Kansas, MO, United States) and 100 U/ml penicillin/100 μg/ml streptomycin/250 ng/ml amphotericin B (FUJIFILM Wako Pure Chemical Corporation) at 37°C under 5% CO₂ atmosphere.

KNZ01 and KNZ03 strains were sequenced using an MiSeq system. DNA was prepared for Illumina library construction by Nextera XT library kits (Illumina, San Diego, CA, United States) according to the manufacture’s instruction. Approximately 1 million 301 bp pair-end reads were trimmed based on base quality (quality score limit = 0.05) and removed if the reads had more than two ambiguous nucleotides or were less than 15 bp in length followed by de novo assembly to construct contigs using the CLC genomics workbench software program (CLC bio, Aarhus, Denmark) with the parameters; mapping mode of simple contig sequences creation (fast), minimum contig length = 500, automatic bubble sizing, automatic word sizing, scaffold performing, and auto-detection of paired distances. The whole genomes have been registered with DNA Data Bank of Japan database under BIOSAMPLE accession numbers SAMD00191542 (KNZ01), SAMD00191543 (KNZ03), SAMD00191544 (KNZ04), SAMD00191545 (KNZ06), SAMD00191546 (KNZ07), SAMD00191547 (KNZ10), SAMD00191548 (KNZ12), SAMD00191549 (KNZ15), SAMD00191550 (KNZ16). CDS sequences were extracted by DFAST (Tanizawa et al., 2018). emm type was determined by a BLAST search against the database in the Centers for Disease Control and Prevention². MLST profiles were submitted to pubMLST (Jolley and Maiden, 2010; McMillan et al., 2010). Average nucleotide identity (ANI) was calculated using the ANI Calculator (Yoon et al., 2017). Forty-eight SDSE genome sequences were obtained from the National Center for Biotechnology Information website. Single nucleotide polymorphisms (SNPs) of core genomes were extracted by Parsnp webtool (Treangen et al., 2014) using the SDSE 167 core genome as a reference (Watanabe et al., 2013). A phylogenetic tree was constructed using the MEGA5 software with the neighbor-joining method and 100 bootstraps (Tamura et al., 2011).

The hemolytic activity of SDSE strains was measured as described previously (Ogura et al., 2018). SDSE strains were cultured overnight in THY medium supplemented with 1% Tween 80 (Sigma Aldrich, St. Louis, MO, United States), centrifuged at 1500 × g, and resuspended in fresh THY medium at OD₆₀₀ = 1.0. Sheep whole blood (5 ml) in Alsever’s solution (Nippon Bio-Test Laboratories, Inc., Saitama, Japan) was centrifuged at 1200 × g for 20 min, washed twice with sterile phosphate-buffered saline (PBS), and resuspended in 50 ml PBS containing 2.5% (v/v) bovine serum albumin (FUJIFILM Wako Pure Chemical Corporation). After pre-incubation of the red blood cell (RBC) solution at 37°C for 15 min in the presence or absence of 50 mM 2-mercaptoethanol (2-ME; SLO enhancer; Nakalai Tesque, Kyoto, Japan; Van Epps and Andersen, 1971) and 0.01% trypan blue (TB; SLS inhibitor; FUJIFILM Wako Pure Chemical Corporation; Taketo and Taketo, 1982), bacterial solution and sheep RBC solution were incubated at a ratio of 1:10 at 37°C for 2 h under 5% CO₂ atmosphere. The sheep RBC solution incubated with 0.2% Triton X-100 and THY medium only were used as hemolysis positive and negative controls, respectively. After centrifugation at 2000 × g for 5 min, the absorbance of the supernatants was measured at a wavelength of 550 nm.

Streptococcus dysgalactiae subsp. equisimilis strains at mid-exponential phase (OD₆₀₀ = 0.4–0.5) were centrifuged, washed with PBS, and resuspended in DMEM. HaCaT cells were seeded at 5 × 10⁴ per well in 96-well plates in DMEM and cultured at 37°C under 5% CO₂ atmosphere. Confluent HaCaT cells were washed twice with DMEM and inoculated with SDSE strains at a multiplicity of infection (MOI) of 10. HaCaT cells with DMEM only was used as a cytotoxic negative control. After 2 h incubation at 37°C under 5% CO₂ atmosphere, the supernatants were collected and centrifuged at 1500 × g for 10 min. Cytotoxic activity was measured using the CytoTox-Fluor^TM Cytotoxicity Assay kit (Promega Corp., Madison, WI, United States) according to the manufacturer’s instructions (Running et al., 1999).

Streptococcus dysgalactiae subsp. equisimilis strains were cultured overnight in THY medium supplemented with 1% Tween 80, centrifuged at 1500 × g, and resuspended in fresh THY medium at OD₆₀₀ = 0.1 in the absence or presence of ProteoGuard EDTA-Free Protease Inhibitor Cocktail (Takara Bio Inc., Shiga, Japan). After incubation at 37°C for 2 h, the supernatants were collected for Western blot analysis. Anti-SLO primary (#64-001) and anti-rabbit HRP-conjugated secondary (#HAF008) antibodies were purchased from Bio Academia (Suita, Osaka, Japan) and R&D Systems (Minneapolis, MN, United States), respectively. HRP-bound protein bands were detected by Immobilon Western Chemiluminescent HRP Substrate (#WBKLS0500; Merck Millipore, Bedford, MA, United States) and visualized by LuminoGraph I (ATTO). Band intensities were measured using ImageJ software (Schneider et al., 2012).

Streptococcus dysgalactiae subsp. equisimilis adherence to HaCaT cells was quantified as described previously (Ozeri et al., 1998; Pappesch et al., 2017). Briefly, SDSE strains at the mid-exponential phase were centrifuged, washed with PBS, and resuspended in DMEM (Fraction A). Confluent HaCaT cells in 24-well plates were washed twice with PBS and inoculated with SDSE strains at an MOI of 10. For HaCaT cells-free control, SDSE strains were suspended in 24-well plates without HaCaT cells (Fraction B). After 2 h incubation at 37°C under 5% CO₂ atmosphere, the culture supernatants were collected (Fraction C). After washing thrice with PBS, the cell-wash fluids were also collected (Fraction D). To detach the HaCaT cells, trypsin/EDTA (FUJIFILM Wako Pure Chemical Corporation) was added, followed by lysis of HaCaT cells with sterile distilled water (Fraction E). The colony-forming units (CFUs) of Fractions A to E were measured by serial dilution in PBS and plating on THY agar plates. Adherence (%) was defined as follows: [E/(C + D + E)] × 100. Growth (%) was defined as follows: [B or (C + D + E)/A] × 100.

Surviving SDSE strains in the presence of human neutrophils were assessed as described previously with some modifications (Pappesch et al., 2017). Human neutrophils were isolated from human whole blood using Polymorphprep (Abbott Diagnostics Technologies AS, Oslo, Norway) according to the manufacturer’s instructions. Fresh human blood was collected from 23- to 36-year-old volunteers. Human neutrophils were suspended in RPMI 1640 (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan). SDSE strains were grown in THY medium overnight and washed with PBS. For opsonization, 1 × 10⁶ to 2 × 10⁶ CFU/ml were incubated for 30 min with 10% human serum at room temperature. Opsonized SDSE strains were incubated with 1 × 10⁶ human neutrophils/ml at an MOI of 1–2 at 37°C for 30 min or 2 h under 5% CO₂ atmosphere. After lysis of neutrophils by sterile diluted water, CFUs were measured. CFUs of SDSE strains before incubation with human neutrophils were used as input controls.

Streptococcus dysgalactiae subsp. equisimilis strains were incubated in RPMI 1640 with 5% fresh human serum at 37°C for 30 min or 2 h under 5% CO₂ atmosphere. CFUs were measured by serial dilution in PBS and plating on THY agar plates.

Subcutaneous infection was performed as described previously (Gera et al., 2014). SDSE strains were grown in THY medium to mid-exponential phase (OD₆₀₀ = 0.5), washed, and resuspended in PBS. Mice were anesthetized by intraperitoneal administration of the mixture of Domitor (0.3 mg/kg; Nippon Zenyaku Kogyo Co., Ltd., Fukushima, Japan), Dormicum (4 mg/kg; Astellas Pharma, Inc., Tokyo, Japan), and Vetorphale (5 mg/kg; Meiji Seika Pharma Co., Ltd., Tokyo, Japan). For subcutaneous infection, 2 cm² areas of the back of 6- to 8-week-old female BALB/c mice (Sankyo Labo Service Corp., Inc., Tokyo, Japan) were shaved followed by subcutaneous injection of 100 μl of air to form air pouch and 200 μl of approximately 3.5 × 10⁹ CFU/ml as verified by plating on THY agar plates. After infection, body weight and severity of cutaneous lesions were monitored every day for 7 days.

For histologic analysis, mice were euthanized by Sevofrane (Maruishi Pharmaceutical Co., Ltd., Osaka, Japan) at 2 days after subcutaneous infection. Cutaneous tissues were fixed with 10% neutral-buffered formalin (FUJIFILM Wako Pure Chemical Corporation). Tissues were stained by Gram Stain Kit (Modified Brown & Brenn; ScyTek Laboratories, Logan, UT, United States) and hematoxylin and eosin (H&E; Muto Pure Chemicals Co., Ltd., Tokyo, Japan) according to the manufacturer’s instructions. Stained tissue sections were examined with an optical microscope (Eclipse E600 with U-III Film Camera System; Nikon Corp., Tokyo, Japan). Images were captured using the imaging software NIS-Elements (Nikon).

Scanning Electron Microscopy analyses were performed as described previously with some modifications (Honda-Ogawa et al., 2017). SDSE strains were grown overnight in THY medium or RPMI 1640 supplemented with 10% fresh human serum at 37°C under 5% CO₂ atmosphere. Bacterial cells were fixed with 2.5% glutaraldehyde (FUJIFILM Wako Pure Chemical Corporation) on aluminum plates coated with Matrigel (Corning, Inc., NY, United States) at room temperature for 1 h and washed twice with PBS. After serial dehydration with 50, 80, and 100% ethanol, the samples were soaked in 100% t-butyl alcohol (FUJIFILM Wako Pure Chemical Corporation), freeze-dried (Freeze Dryer ES-2030, Hitachi Technologies, Tokyo Japan), and coated with 60% Au-Pd alloy (Hitachi E-1030, Hitachi Technologies; sputtering conditions: 15 mA, 7 Pa, 30 s). The SEM images were obtained using JSM-7100F (JEOL Ltd., Tokyo, Japan).

All animal experiments were carried out in accordance with the Declaration of Helsinki. The protocols of the clinical study, the animal experiment, and the human blood experiment were approved by the Ethical Committee of Kanazawa University. In line with institutional guidelines following local legislation, consent was not required. Clinical isolates were isolated, stored, and utilized for studies on according to the guidelines that are open at the homepage of Kanazawa University Hospital. The opportunity for patients to opt-out was always available. None of the patients were under the age of 16. All the genetic recombinant experiments were conducted in accordance with Law Concerning the Conservation and Sustainable Use of Biological Diversity through Regulations on the Use of Living Modified Organisms and approved by The Kanazawa University Safety Committee for genetic recombinant experiments.

The genomic features of the nine SDSE strains, including clinical specimen, emm type, sequence type (ST), and clonal complex (CC), are shown in Table 1; the most prevalent emm types are stG6792 and stG245, followed by stG480, stG653, and stG166. Five of the nine strains were categorized as CC17 with one novel ST, strain KNZ12; one strain as CC15; the other three strains, as CC25. We chose one stG6792/CC17 strain KNZ01 and one stG245/CC25 strain KNZ03. Here, we also utilized strain ATCC 12394 (stG166/CC25), whose transfer information is available from ATCC, for pathogenicity comparisons; KNZ03 possessed ParC S79YF substitution associated with fluoroquinolone resistance (González et al., 1998). Next, we performed a phylogenetic analysis using concatenated SNPs (Figure 1). The phylogenetic distance between KNZ03 and ATCC 12394 was close; KNZ03 showed 99.7% ANI with ATCC 12394. In contrast, KNZ01 was very close to the other stG6792/CC17 SDSE strains, KNZ07 and RE378, with 99.9 and 99.8%, sequence similarity, respectively.

To examine the hemolytic activities, the three SDSE strains were incubated with 5% sheep RBCs for 2 h (Figure 2). Incubation of KNZ03 with sheep RBCs showed significantly higher hemolysis compared with KNZ01. Treatment with 2-mercaptoethanol (+2-ME), an SLO enhancer, resulted in significant enhancement of hemolysis in KNZ03-treated sheep RBCs, indicating that KNZ03 expressed high SLO activities. However, these hemolytic activities were substantially suppressed by trypan blue (+2-ME/TB), an SLS inhibitor. Western blot analysis showed that SLO was not found in the culture supernatant of KNZ03 (Figure 2B). Also, the protease inhibitor cocktail did not affect the amount of SLO, showing that the absence of SLO was not due to protein degradation. These results suggested that KNZ03 secretes high SLS but not SLO. Western blot analysis also showed that ATCC 12394 released significantly higher SLO protein levels than KNZ01 (Figures 2B,C). SLO is known to exhibit cytotoxic activity to mammalian cells (Bhakdi et al., 1985; Palmer, 2001); thus, we examined the cytotoxicity of the SDSE strains to human keratinocytes (HaCaT cells). As shown in Figure 2D, the incubation of HaCaT cells with ATCC 12394 resulted in significant cell death compared with that of KNZ01 and KNZ03; this is consistent with the amounts of SLO protein (Figure 2C). These results indicated that ATCC 12394 expressed the highest cytotoxic activity to human keratinocytes by producing large amounts of SLO protein.

Adherence and cytotoxicity to keratinocytes are related to severity of skin infections such as necrotizing soft-tissue infections (Humar et al., 2002; Siemens et al., 2015). To investigate adherence of the SDSE strains to human keratinocytes, we counted the number of bacteria attached to HaCaT cells. After 2 h incubation of the SDSE strains with HaCaT cells, the percentage of adherence of ATCC 12394, KNZ01, and KNZ03 was 65, 85, and 75%, respectively (Figure 3A). KNZ01 and KNZ03 showed significantly higher adherence to human keratinocytes than ATCC 12394.

To examine bacterial growth in the presence of human keratinocytes, the SDSE strains were incubated with DMEM supplemented with 8% inactivated FBS (DMEM/FBS) in the absence or presence of HaCaT cells. We counted the CFUs of inoculated bacteria (input controls), bacteria incubated in the absence of HaCaT cells (DMEM/FBS only) for 2 h, and bacteria incubated in the presence of the cells for 2 h (Figure 3B).

After 2 h incubation in DMEM/FBS only (black bars), CFUs of ATCC 12394 and KNZ01 increased 8 and 18 times, respectively, whereas that of KNZ03 increased only 1.5 times compared with that of input controls. In contrast, in the presence of HaCaT cells (gray bars), ATCC 12394 did not increase. Although KNZ01 increased seven times in the presence of HaCaT cells compared with that of input controls, the growth was significantly lower than that in DMEM/FBS only. Interestingly, contrary to ATCC 12394 and KNZ01, we found that the growth of KNZ03 incubated in the presence of HaCaT cells was significantly 2.6 times higher than that in DMEM/FBS only. These results indicated that the three SDSE strains possessed different mechanisms of growth on host keratinocytes. Although some triggers of KNZ03 in response to attachment to epithetical cell are postulated, no supporting data were obtained. It remains unknown what triggers induce the rapid growth of KNZ03 in the presence of keratinocytes.

To examine anti-phagocytic activity, the SDSE strains were incubated with human neutrophils in RPMI 1640 containing 10% fresh human serum for 30 min or 2 h followed by counting of CFUs (Figure 4). Compared with CFUs of input controls (100%), CFUs of ATCC 12394, KNZ01, and KNZ03 decreased to 45, 60, and 80%, respectively, after incubation with human neutrophils for 30 min (Figure 4A, black bars). This result indicated that KNZ03 had the highest anti-phagocytic activity against human neutrophils in a short time. After 2 h incubation (Figure 4B, black bars), CFU of KNZ01 increased significantly to 170% compared with that of input control, indicating that KNZ01 showed significantly high growth after avoidance of phagocytosis by human neutrophils. In addition, we found that KNZ01 showed a significant growth in the presence of human serum. As shown in Figure 4B (white bars), after 2 h incubation in human serum without neutrophils, CFU of KNZ01 increased to 600%, whereas those of ATCC 12394 and KNZ03 increased to 250 and 150%, respectively. These results indicated that KNZ01 had the highest growth activity in human serum among the three SDSE strains, which was involved in the highest survival rate of KNZ01 after 2 h incubation with human neutrophils.

To examine the correlation between our in vitro observations and the severity of skin infections, we subcutaneously infected the SDSE strains into mice (Figure 5). On day 1 or 2 after infection, mice infected with all strains developed skin lesions with erythema and purulence (Figure 5A). On day 6 or 7 after infection, mice infected with all strains formed ulcerative lesions with epidermal hyperplasia around the sites of infection. These results indicated that ATCC 12394, KNZ01, and KNZ03 showed pathogenicity to mice to cause severe subcutaneous infections with ulcerative lesions. On day 2 after infection, the body weight of mice infected with KNZ01 decreased significantly compared with that of mice infected with KNZ03 (Figure 5B). In addition, erythema and purulence of mice infected with KNZ01 looked most severe among the three SDSE strains, although there was no significant difference in lesion sizes (Figure 5A). Gram stain analysis (Figure 6A) showed that, on day 2 after infection, only KNZ01 remained in the subcutaneous tissues (dyed purple), whereas ATCC 12394 and KNZ03 were eliminated. Nevertheless, H&E-stained pathological images showed no significant differences among the three SDSE strains (Figure 6B). These results indicated that KNZ01 could remain in the subcutaneous tissues after infection, which resulted in a significant decrease in body weight of mice and severe erythema on day 2 after infection.

The cell surface structures of the SDSE strains were observed via SEM. To examine whether the cell surface structures are affected by human-derived source (serum), we compared THY medium-treated cells and those treated with RPMI 1640 supplemented with 10% fresh human serum (RPMI/hSerum).

After overnight culture in THY medium, ATCC 12394 exhibited a smooth cell surface structure, whereas KNZ01 and KNZ03 had rough surfaces (Figure 7A, top); however, the cell surface of ATCC 12394 was rough after overnight culture in RPMI/hSerum (Figure 7B, top); also, KNZ01 constructed a biofilm with a large number of extracellular substances after overnight culture in RPMI/hSerum. These results showed that ATCC 12394 and KNZ01 construct different cell surface structures in response to culture conditions. KNZ01 specifically formed significant biofilms in the presence of human serum. Additionally, we found that the biofilm formation of KNZ01 partially decreased by the deletion of a gene encoding a cell wall-anchoring protein (cwap5 gene in Supplementary Table S2), which indicated that the biofilm was partially composed of cell wall-anchoring proteins. No significant change was observed in the KNZ03 between THY medium and RPMI/hSerum, suggesting that KNZ03 lacks the response of the construction of cell surface structures to culture conditions.