We studied 18 α-hemolytic S. dysgalactiae subsp. dysgalactiae field isolates of Lancefield group C that had caused bovine subclinical mastitis. Isolates were obtained from 304 milk samples of 248 cows from 8 farms in Portugal that were included in the study. Detailed information regarding isolation methods and identification of field isolates by biochemical methods was described in a study of the subclinical mastitis–associated pathogen S. uberis (4). To confirm identification of S. dysgalactiae subsp. dysgalactiae, the 16S rRNA gene was amplified by PCR and sequenced (5). SmaI/cfr9I-digested DNA banding patterns were obtained by pulsed-field gel electrophoresis for clone identification as described (4).

All genes analyzed by PCR are shown in the Appendix Table. The emm gene subtyping was performed as described (www.cdc.gov/ncidod/biotech/strep/M-ProteinGene_typing.htm). Primers used and conditions for PCR were essentially as described elsewhere (Appendix Table).

Samples without DNA and strains lacking (negative) or carrying (positive) specific genes were used as controls in the PCR. Results were consistent in 2 or 3 PCRs that included these controls. Sequencing of all virulence gene amplicons was performed with the same primers used for amplification (STAB-Vida, Lisbon, Portugal). All sequences were compared with sequences in GenBank by using the BLAST alignment tool (www.ncbi.nlm.nih.gov/BLAST).

Antimicrobial drug resistance against macrolides (erythromycin), lincosamides (pirlimycin), and tetracycline was determined as described (10). Macrolide resistance phenotypes identified were M (resistance to macrolides) and MLS_B (resistance to macrolides, lincosamides and streptogramins B).

We detected bacteriophage-associated virulence genes speM, speK, speC, spd1, and speL. Overall, speM was found in 10 (55.6%) of 18 bovine GCS isolates, speK in 9 (50%), speC and spd1 in 6 (33%), and speL in 4 (22.2%). All but 1 of the PCR products showed expected sizes (Appendix Table). Tn1207.3/Φ10394.4 composite transposon left junction amplicon showed a size of 380 bp instead of 453–6,807 bp as described for GAS (9). No amplification was observed for the right junction of this genetic element.

Figure 1

Figure 1. Dendrogram and pulsed-field gel electrophoresis (PFGE) profiles of group C streptococci (Streptococcus dysgalactiae subsp. dysgalactiae) subclinical mastitis isolates from 8 dairy herds, Portugal. PFGE type-subtype, virulence genotype, antimicrobial drug resistance phenotypes,...

The emm gene encoding the antiphagocytic M surface protein was not amplified in any of the 18 bovine GCS isolates; therefore, no emm types were obtained. Subsets of isolates were erythromycin and pirlimycin resistant (MLS_B phenotype) and contained erm(B) or erm(A) genes (22.2%) or erythromycin susceptible and pirlimycin resistant and contained the linB gene (16.6%). All isolates were tetracycline resistant with a subset (66.7%) carrying tet(M) or tet(O) tetracycline resistance determinants. Distribution of bacteriophage-associated virulence genes and other characteristics of strains are shown in Figure 1.

Figure 2

Figure 2. Alignments of bovine group C streptococci (Streptococcus dysgalactiae subsp. dysgalactiae) alleles of virulence genes from 8 dairy herds, Portugal, with sequences from the National Center for Biotechnology (Bethesda, MD, USA) database...

Sequences of all virulence genes were compared by using the BioEdit sequence alignment editor (www.mbio.ncsu.edu/BioEdit/bioedit.html). One different allele was found for each of the following gene sequences: spd1 (among 6 strains), speC (among 6 strains), and speL (among 4 strains). Two alleles were found for speK (among 9 strains) (speK-1 and speK-2), and 4 alleles were found for speM gene sequences (among 10 strains) (speM-1, speM-2, speM-3, and speM-4). Bovine alleles had sizes of 386 bp (spd1), 222 bp (speC), 444 bp (speL), 232 bp (speK), and 357 bp (speM). Examples of alignments between bovine virulence gene alleles with sequences from GenBank (only most similar ones) are shown in Figure 2.

Using PCR, we determined that bovine GCS S. dysgalactiae subsp. dysgalactiae strains (55.6%) carried >1 GAS-specific bacteriophage virulence-associated genes (spd1, speC, speK, speL, and speM). This finding suggested that bacteriophages may also play a role in the genetic plasticity and virulence of animal GCS.

The speL allele from bovine strains showed higher similarity with the szeM allele (99% maximum identity) from S. equi subsp. zooepidemicus than with the speL allele (97% maximum identity) from S. pyogenes. The szeM gene encodes a superantigen in S. equi subsp. zooepidemicus, which is primarily a pathogen of nonhuman animal species. This organism causes mastitis in cows and mares and is most frequently found in horses (14). We also observed that 3 of the speM alleles found among bovine strains (speM-1, speM-2, and speM-3) also showed higher similarity with superantigen-encoding gene szeL from S. equi subsp. zooepidemicus than with speM gene sequence from S. pyogenes. Another allele (speM-4) showed higher similarity with the sdm gene from S. dysgalactiae subsp. dysgalactiae than with the speM gene from S. pyogenes.

The remaining alleles (spd1, speC, speK-1, and speK-2) from the GCS S. dysgalactiae subsp. dysgalactiae bovine strains showed high similarity with S. pyogenes superantigen genes (98%–99% maximum identity). This finding supports our hypothesis that GAS prophages may play a role in the genetic plasticity of this pathogen. The speC and spd1 genes are known to be localized on the same GAS prophage (15), and both genes were detected in 6 bovine GCS S. dysgalactiae subsp. dysgalactiae isolates in our study.

None of 18 α-hemolytic group C S. dysglacatiae subsp. dysgalactiae bovine isolates in this study were typed by emm-typing because amplification products in the PCR specific for the M surface protein gene emm were not obtained. This result is consistent with those of a report that β-hemolytic, but not α-hemolytic, group C S. dysglacatiae subsp. dysgalactiae isolates of animal origin contained the emm gene (3).

Amplification (380-bp product) of the left junction of the composite transposon in bovine isolates suggests that this mobile element may be inserted in a similar location, the comEC locus, as mapped in S. pyogenes and S. dysglactiae subsp. equisimilis. Absence or unexpected PCR products specific for any of the junctions of this element have been reported in other studies and attributed to possible lack of homology between the target and primers used (9). Detection of the linB gene carried by a large conjugative plasmid (13) in 3 of 18 bovine GCS S. dysgalactiae subsp. dysgalactiae isolates is indicative of horizontal gene transfer.

Our findings indicate that α-hemolytic bovine GCS isolates, which are known to be environmental or contagious pathogens and a cause of bovine mastitis, may be reservoirs of virulence genes encoded by prophages of human-specific GAS. These genes encode exotoxins, superantigens, and streptodornases, which are responsible for GAS virulence and pathogenesis, and may be transferred to other streptococci of human origin by horizontal genetic transfer. Therefore, α-hemolytic isolates should not be disregarded as putative infectious disease agents in humans.