The major virulence factor of the pneumococcus, and target for conjugate vaccines, is the polysaccharide capsule, which is usually encoded by the highly variable cps locus. Serotype 37 is an unusual pneumococcal type in which the single β-glucosyltransferase gene responsible for serotype capsule production (tts) is located outside of the capsular operon region. Using a previously described automated whole genome sequence (WGS)-based serotyping bioinformatics tool, PneumoCaT, we identified and investigated seven clinical isolates (three from blood cultures) of non-pneumococcal streptococci containing a highly homologous tts and included them in a study panel of 20 isolates which included a 11 further clinical isolates of S. pneumoniae serotype 37, a reference strain of serotype 37 and the S. pseudopneumoniae type strain BAA 960T. The seven non-pneumococcal isolates generated novel alleles at all pneumococcal MLST loci and gave low percentage similarity (<45%) to S. pneumoniae or S. pseudopneumoniae species by comparison of short sequence patterns in genomic data (k-mer analysis). The S. pseudopneumoniae BAA-960T isolate generated two novel alleles in the MLST and gave a high similarity (>99%) to the reference sequence for BAA-960T. Twelve isolates gave high similarity (>77%) to the Streptococcus pneumoniae 5652-06 serotype 19A reference genome sequence and had previously reported MLST alleles. Each of the seven clinical non-pneumococcal strains and all of the 12 S. pneumoniae possessed a β-glycosyltransferase gene (tts) with >95% similarity to the pneumococcal tts reference DNA sequence with 20-22 non-synonymous SNPs. All but two strains in which the tts gene was detected gave positive reactions for serotype 37 in slide agglutination tests with serotype 37 typing sera. Phylogenetic analysis using both SNP and MLST data showed distinct clades corresponding to strains identified as pneumococcus or non-pneumococcus by kmer WGS analysis. Extended k-mer database analysis and ribosomal MLST placed the non-pneumococcal isolates within the S. mitis group. Biochemical and bile solubility assays showed differences between the unusual isolates and S. pneumoniae. All isolates had detectable pneumolysin (ply) genes, but only those that identified as pneumococcus contained the genes for autolysin (lytA) or the ABC transporter lipoprotein A (piaA) with >80% coverage and >95% similarity. Here we report the existence of a novel group of strains distinct from S. pneumoniae, but which can express a pneumococcal serotype 37 capsular polysaccharide which can be associated with clinical disease.
Keywords: Microbial genomics; Microbial identification; Phylogenetics; Streptococci.