Sequencing of the variable region of rpsB to discriminate between Streptococcus pneumoniae and other streptococcal species

The vast majority of streptococci colonizing the human upper respiratory tract are commensals, only sporadically implicated in disease. Of these, the most pathogenic is Mitis group member, Streptococcus pneumoniae. Phenotypic and genetic similarities between streptococci can cause difficulties in species identification. Using ribosomal S2-gene sequences extracted from whole-genome sequences published from 501 streptococci, we developed a method to identify streptococcal species. We validated this method on non-pneumococcal isolates cultured from cases of severe streptococcal disease (n = 101) and from carriage (n = 103), and on non-typeable pneumococci from asymptomatic individuals (n = 17) and on whole-genome sequences of 1157 pneumococcal isolates from meningitis in the Netherlands. Following this, we tested 221 streptococcal isolates in molecular assays originally assumed specific for S. pneumoniae, targeting cpsA, lytA, piaB, ply, Spn9802, zmpC and capsule-type-specific genes. Cluster analysis of S2-sequences showed grouping according to species in line with published phylogenies of streptococcal core genomes. S2-typing convincingly distinguished pneumococci from non-pneumococcal species (99.2% sensitivity, 100% specificity). Molecular assays targeting regions of lytA and piaB were 100% specific for S. pneumoniae, whereas assays targeting cpsA, ply, Spn9802, zmpC and selected serotype-specific assays (but not capsular sequence typing) showed a lack of specificity. False positive results were over-represented in species associated with carriage, although no particular confounding signal was unique for carriage isolates.