The scm2 positive S. canis strains IMT40096 and IMT42870 were isolated by routine diagnostics at Freie Universität Berlin in 2016, as well as scm1 positive S. canis strain IMT40165. Bacteria were routinely grown in Brain Heart Infusion (BHI) or on Brain Heart Agar (BHA) at 37°C without shaking. In preparation for electroporation, IMT40096 was grown overnight in BHI and subcultured into 50 mL of BHI the following day. The resulting culture was harvested at an OD of 0.5 at a wavelength of 600 nm by centrifugation (10 min at 3250 × g), washed three times in 50 mL 20% sterile glycerol solution, and resuspended in 1 mL of 10% glycerol solution. Aliquots of 100 μL were kept frozen at −80°C until needed. Escherichia coli (E. coli) EC101 were cultivated in Luria-Bertani media (LB). For chemical competence, EC101 was subcultured into 50 mL of LB and grown at 37°C 200 rpm until an OD of 0.6 at 600 nm was reached. The resulting suspension was put on ice for 1 h, washed twice with 10 mL of 0.1 M CaCl₂ by centrifugation at 7000 rpm for 7 min, put on ice for 1 h to be then resuspended in 1 mL of 0.1 M CaCl₂. Aliquots of 100 μL were produced with 10% glycerol and stored at −80°C until necessary. Where indicated, antibiotics were supplemented as follows. E. coli: erythromycin (Erm) 200 μg/mL; S. canis: Erm 2 μg/mL.

DNA extraction was conducted with the QIAmp DNA kit according to manufacturer’s instructions. DNA measurements were made with the Qubit^™ 2.0 Fluorometer with the dsDNA HS.

The libraries for whole-genome sequencing (WGS) were prepared using the Nextera XT DNA Library Preparation Kit (Illumina, Inc., San Diego, United States) according to the manufacturer’s recommendations. The 2 × 300 bp paired-end sequencing in 40-fold multiplexes was performed on the Illumina MiSeq platform (Illumina, Inc., San Diego, United States) with MiSeq Reagent Kit v3 (600 cycle).

The raw Illumina reads were quality checked by FastQC (Andrews, 2010) and trimmed by Trim Galore v0.6.6 (RRID:SCR_011847). De novo assembly into contigs was carried out with Unicycler (Wick et al., 2017).

Protein alignment was conducted with the ClustalW webtool (Minh et al., 2020). A Maximum Likelihood tree was then constructed with the IQtree v2.2.0.3 software while performing 1,000 boostraps with UFBoot2 and the best fit model selected by ModelFinder (Sievers et al., 2011; Kalyaanamoorthy et al., 2017; Hoang et al., 2018). Tree visualization was performed with the iTOL: Interactive Tree of Life webtool (Letunic and Bork, 2021).

Human FITC-labeled fibrinogen was purchased from Sigma. Ethidium homodimer-1 and the secondary rabbit anti-mouse IgG Alexa Fluor 488-conjugated antibody (Cat#: A-11059; RRID: AB_2534106) were obtained from Life Technologies. 16% formaldehyde without methanol and CitiFluor^™ CFM3 mounting medium were obtained from Electron Microscopy Science.

Targeted insertional mutagenesis of scm2 was performed using the pGh9:Δscm vector (Maguin et al., 1992) obtained by inverse PCR of the pGh9:ISS1 vector (NCBI: txid481005) with the following primer pair which included NcoI restriction sites: 5′-GGGCCATGGGCTCCTTGGAAGCTGTCAG-3′; 5′-CCCCCATGGGGTACCCAATTCGCCCTATA-3′. The vector was then inserted into E. coli EC101 via chemical competence and grown for 2 days in erythromycin to induce plasmid replication. The scm2 gene with 500 bp upstream and downstream of it was amplified by PCR using the following primer pair which also included NcoI restriction sites: 5′-GGGCCATGGGACATCGTGAATTTTGCCG-3′; 5′-CCCCCATGGCACTTGGATAGGATGCAC-3′. PCR products were cloned into the temperature-sensitive vector pGh9: ΔISS1 by NcoI digestion and alkaline phosphatase to avoid religation products. The resultant plasmid was then amplified by inverse PCR to knock out scm2 while keeping the 500 bp regions around it. This was done with the following primer pair which included EcoRV restriction sites at the 500 bp regions around scm2: 3′-GGGGATATCGTGACAGGCTATCTTTAG-5′; 3′-CCCGATATCAGCCTCGTTACAGACTATCC-5′. The plasmid was then inserted into EC101 to produce pGh9:Δscm2 which was verified by band size measurement after EcoRV and NcoI digestion (1 band at 3500 bp representing the vector and 2 bands at 500 bp representing the 500 bp regions around scm2). One μg of the resultant knockout plasmid was introduced into IMT strain 40096 by electroporation, and Erm resistant transformants were identified at the permissive temperature for plasmid replication (28°C). Single-crossover Campbell-type chromosomal insertion was selected by shifting to the nonpermissive temperature (37°C) while maintaining Erm selection. SCM phenotype was determined on BHA plus Erm at 37°C. Verification of the mutagenesis was conducted via PCR using the scm2 amplification primers described above.

Protein extraction was performed as described previously (Nerlich et al., 2019). Briefly, scm fragments from different S. canis strains were overexpressed in E. coli via Qiagen’s pQE30 expression vector, followed by purification of the recombinant histidine-tagged proteins as described by the manufacturer (Qiagen ExpressionistTM System).

Survival of S. canis wild-type and the isogenic scm2 mutant in blood was analyzed using whole heparinized canine blood. Handling of dogs and sampling was conducted in strict accordance with the principles of the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes as well as the German Animal Protection Law. Blood was retrieved due to excess in diagnostics procedures.

Briefly, 1 mL canine whole blood was mixed with 1 × 10⁵ CFU of wild-type or mutants. The mixture was incubated on a rotator at 37°C for up to 6 h with sampling each hour. Survival factors were determined by dividing the number of CFU after any incubation period at 37°C by the number of CFU at t = 0 min (CFU were determined by plating of serial dilutions on BHI plates). Three blood samples from different dogs were used.

Immunofluorescence staining and confocal microscopy was performed as described previously (Nerlich et al., 2019). In brief, 100 μL of bacterial overnight culture was allowed to adhere to poly-L-lysinecoated μ-slide 8-well (ibidi) filled with 100 μL PBS for 30 min and subsequently fixed with 3% PFA for 15 min. After washing with PBS samples were incubated with rabbit anti-mouse Alexa Fluor 488-conjugated IgGs in PBS/1% BSA/Tween 20 overnight at 4°C and bacterial RNA/DNA was stained with ethidium homodimer-1. Samples were mounted in CitiFluor^™ CFM3 and examined using an LSM 780 confocal laser-scanning microscope with a PlanApochromat 63×/1.4 NA objective driven by Zen 2012 software (Carl Zeiss). Image stacks with a z-step size of 0.2 μm per plane were acquired with the pinhole set to 1 airy unit in sequential imaging mode to avoid bleed-through of fluorescence emission. Images were deconvolved using Huygens^® Essential 16.10 (Scientific Volume Imaging) and 3D-stacks are displayed as maximum intensity projections and adjusted identically for brightness and contrast in ImageJ/Fiji (Schindelin et al., 2012).

S. canis IMT40096 and IMT40096Δscm2 were grown to mid-exponential phase in TSB medium and washed once in PBS. Bacteria are adjusted to a transmission of 10% equal to 10⁸ bacteria per mL. A total of 1 × 10⁷ bacteria were suspended in 400 μL of PBS containing 0.5% FCS and incubated with 0.5 mg polyclonal rabbit IgG for 30 min at 37°C. After washing in PBS the bacterial pellet was suspended in 100 μL of PBS containing an anti-rabbit ALEXA^® Fluor 488 antibody or FITC-labeled fibrinogen, respectively as described previously and incubated for 30 min at 37°C. Bacteria were washed again in PBS, fixed in PBS containing 3% paraformaldehyde and analyzed by flow cytometry using a FACS LSRII (Becton Dickinson). Cells were analyzed with the FlowJo software v.7.6.5. Cells were gated on Side scatter area (SSC-H) vs. Fluorescin isothiocyanate labeled human fibrinogen from Molecular Innovations (FITC-H). IMT40096 was determined as the positive control and used as the gating reference for the succeeding samples (Supplementary Figure S1). The gating region was set to exclude debris and larger aggregates of bacteria. A total of 10⁴ bacteria were analyzed for fluorescence using log-scale 2 amplification.

The interaction of SCM-1 or SCM-2 with fibrinogen was analyzed by surface plasmon resonance spectroscopy using a Biacore T200 optical biosensor (GE Healthcare). Human fibrinogen (Molecular Innovations) was immobilized (~1,000 response units) on a carboxymethyl dextran sensor chip (CM5) via amine coupling as described previously (Bergmann et al., 2017). SCM proteins were used as analytes in a concentration range of 1.25–20 μg/mL. Binding analysis was performed in PBS containing 0.05% Tween20 at 25°C and a flow rate of 10 μL/min. Data were analyzed using Biacore T200 evaluation software (version 3.2).

Alignment of the scm nucleotide sequences of 75 clinical isolates of S. canis (Supplementary Table 1) via Maximum Likelihood revealed the presence of two genetically distinct populations differentiated by the SCM protein (Figure 1A). S. canis strain G361 and IMT40165 used in this study belonged to the same population (hereby referred to as the SCM-1 strains), but had a significantly different scm gene compared to S. canis strains G2, IMT40096 and IMT42870, which formed their own population (SCM-2). Alignment of SCM-1 proteins to SCM-2 showed an identity as low as 35%. However, closer inspection of the sequences showed that the C-terminal region of the two proteins corresponding to the sortase-anchoring motif LPXTG motif is highly conserved (Figure 1B). Given these differences, we wondered if the IgG binding capabilities previously observed in SCM-1 protein expressing S. canis were conserved by the SCM-2 protein.

In a previous study, we identified IgG binding as a function enabled by SCM-1 (Bergmann et al., 2017). Consequently, we investigated whether similar properties are conferred by SCM-2. FACS analysis of SCM-2 S. canis showed significantly reduced IgG binding when compared to G361 SCM-1 reference strain. When quantifying IgG binding, SCM-2 of G2, IMT40096 and IMT42870 displayed a 10-fold lower geometric mean fluorescence intensity (MFI) than G361 SCM-1 (Figure 2A). Correspondingly, SCM-2 strains do not appear to be as efficient in binding IgG.

In accordance with findings by Nerlich et al. (2019), fluorescence microscopy of IMT40165, IMT40096 and IMT42870 showed IgG binding properties and a bacterial aggregation only by S. canis strain IMT40165. In contrast, both SCM-2 strains were negative and did neither bind IgG nor aggregate in its presence (Figure 2B).

Previous studies on M proteins belonging to Streptococcus pyogenes clade Y and Streptococcus equi ssp. zooepidemicus (SEZ) showed that certain M proteins are capable of binding fibrinogen (Sanderson-Smith et al., 2014; Bergmann et al., 2019). Hence, we assessed fibrinogen binding capacity of S. canis in the next step. FACS analysis of SCM-1 S. canis 361 versus SCM-2 S. canis revealed increased human fibrinogen binding in SCM-2 expressing strains. In comparison to the scm1 expressing S. canis G361, the fluorescence of SCM-2 S. canis strain IMT40096 was more than three times higher. In fact, all other SCM-2 strains tested shared an increase in fluorescence corresponding to fibrinogen binding by at least two-fold (Figure 3). Furthermore, deletion of scm2 in IMT40096 background diminished fluorescence to 0.13 compared to the control strain G361. These results suggest that the scm gene is at the center of fibrinogen binding in the tested SCM-2 S. canis strains (Figure 3).

SPR analysis to evaluate the SCM protein’s affinity to fibrinogen showed KDs of around 10⁻⁸ for IMT40096 and 2 × 10⁻⁷ IMT42870, indicating a strong affinity for fibrinogen for the SCM proteins of both strains (Figure 4; Supplementary Figure S2). This is particularly remarkable bearing in mind that both SCM-2 strains originate from different branches of the Maximum Likelihood Tree (Figure 1) and hence display low structural similarities. Nevertheless, functional traits and fibrinogen affinity seems to be conserved. Contrary, fibrinogen binding was not observed for SCM-1 expressing S. canis strain G361 (Supplementary Figure S3). This demonstrates fibrinogen binding capabilities of SCM-2 at protein level.

Finally, we evaluated pathophysiological consequences of SCM-2 deletion in S. canis performing a functional assay. Therefore, survival of IMT40096 and its SCM mutant in canine whole blood was assessed. A significant decrease in survival was observed for the mutant strain compared to the wild-type (Figure 5). At 1 h post-inoculation, the wild-type showed a survival factor of approximately 4, whereas the SCM mutant displayed a survival factor of only 0.5. This difference was even more exacerbated when inoculation was extended to 2 h, where the wild-type reaches survival factors of up to 10 while the SCM mutant stagnates at 0.5. This suggests that, as shown for other M proteins in GAS and S. dysgalactiae spp. equisimilis (Whitnack and Beachey, 1985; Carlsson et al., 2005), SCM-2 and associated fibrinogen binding is critical for the survival of S. canis 2 in canine whole blood.