Genetic basis of coaggregation receptor polysaccharide biosynthesis in Streptococcus sanguinis and related species

Abstract

Interbacterial adhesion between streptococci and actinomyces promotes early dental plaque biofilm development. Recognition of coaggregation receptor polysaccharides (RPS) on strains of Streptococcus sanguinis, Streptococcus gordonii and Streptococcus oralis by Actinomyces spp. type 2 fimbriae is the principal mechanism of these interactions. Previous studies of genetic loci for synthesis of RPS (rps) and RPS precursors (rml, galE1 and galE2) in S. gordonii 38 and S. oralis 34 revealed differences between these strains. To determine whether these differences are strain-specific or species-specific, we identified and compared loci for polysaccharide biosynthesis in additional strains of these species and in several strains of the previously unstudied species, S. sanguinis. Genes for synthesis of RPS precursors distinguished the rps loci of different streptococci. Hence, rml genes for synthesis of TDP-L-Rha were in rps loci of S. oralis strains but at other loci in S. gordonii and S. sanguinis. Genes for two distinct galactose epimerases were also distributed differently. Hence, galE1 for epimerization of UDP-Glc and UDP-Gal was in galactose operons of S. gordonii and S. sanguinis strains but surprisingly, this gene was not present in S. oralis. Moreover, galE2 for epimerization of both UDP-Glc and UDP-Gal and UDP-GlcNAc and UDP-GalNAc was at a different locus in each species, including rps operons of S. sanguinis. The findings provide insight into cell surface properties that distinguish different RPS-producing streptococci and open an approach for identifying these bacteria based on the arrangement of genes for synthesis of polysaccharide precursors.

Keywords: Streptococcus sanguinis; biofilm; coaggregation; galactose epimerase; interbacterial adhesion; receptor polysaccharide; rps locus.

Figure 1

Comparison of rps locus of S. sanguinis SK45 (GenBank HM046485.1) with those of S. gordonii 38 (GenBank AY147914.1) and S. oralis 34 (GenBank AB181234.2) showing percent identity between encoded regulatory proteins (red), glycosyl transferases (blue), polysaccharide polymerases (white), flippases (black), enzymes for synthesis of nucleotide-linked sugars (green) and flanking genes (gray). Synthesis of TDP-L-Rha and UDP-GalNAc depend on additional loci in S. gordonii 38 (GenBank AY147913.1), S. sanguinis SK45 (GenBank JX407104.1, JX407105.1) and S. oralis 34 (GenBank KC620451). The presence of wefB and wefC in the rps2Gn operon of S. gordonii 38 verses wefH in the rps1Gn operons of S. sanguinis SK45 and S. oralis 34 is associated with the structural difference between RPS2Gn and RPS1Gn.

Figure 2

Detection of cell surface RPS on S. gordonii 38, S. gordonii XC8 and S. gordonii XC8(pJY-45) by dot immunoblotting with rabbit anti-RPS2Gn antibody. Expression of S. sanguinis SK45 galE2 from pJY45 restored RPS production in S. gordonii XC8, a galE2 mutant of wild type strain 38.

Figure 3

Molecular phylogenetic analysis of full length GalE protein sequences as deposited in GenBank (Table S1) from strains of S. gordonii (Sg), S. oralis (So) and S. sanguinis (Ss) showing separation of GalE1 and GalE2 in an unrooted tree constructed by Maximum Likelihood method using MEGA5 (Tamura et al., 2011). Bootstrap values (%) are based on 1500 replications; the scale bar refers to genetic divergence as calculated by the MEGA software. The ability of human GalE and inability of E. coli (Ec) GalE to act on acetylated substrates depends in part on C309 in the former protein and Y299 in the latter. C or S occurs at the equivalent position in all streptococcal GalE2 sequences. In contrast, L occurs in all GalE1 sequences and in T. brucei (Tb) GalE (L342), which has substrate specificity like that of E. coli GalE.