Characterisation of Streptococcus pneumoniae opacity phase variation

Abstract

Streptococcus pneumoniae (the pneumococcus) is a common nasopharyngeal commensal in humans that can invade deeper tissues causing a range of diseases, resulting in significant global morbidity and mortality. Any given strain of S. pneumoniae can undergo phase variation between two different colony phenotypes, termed opaque (O) and transparent (T), as determined by their morphology on clear, solid media. In animal models, the O form is more virulent and commonly isolated from normally sterile sites, such as the blood, lungs and brain, compared to its T counterpart, which is more likely to reside in the nasopharynx. To date, the mechanism involved in the switch between opacity phenotypes is not known and there is discordance in the literature in terms of the precise molecular differences between them. This project set out to comprehensively characterise phenomic (2D-DIGE), transcriptomic (DNA microarray) and genomic (IonTorrent sequencing) differences between opacity variants in three unrelated pneumococcal strains (D39, serotype 2; WCH16, serotype 6A; and WCH43, serotype 4). Proteomic, transcriptomic and genomic analyses revealed strain-specific differences associated with phase variation, but no single difference was consistent across the three strains tested. Nevertheless, there were examples of proteins and genes belonging to the same putative functional groups that were regulated similarly in both the O and T phase between strains. One example was the upregulation of proteins and genes related to genetic competence in the O variants. However, mutagenesis of one such gene, encoding the competence stimulating peptide receptor (comD), did not alter opacity phenotype. There were also inconsistencies between genes identified as differentially expressed at an mRNA and protein level, such as genes involved in cell division, and amino acid biosynthesis and acquisition. These genes were upregulated at an mRNA level in the T variants, but no such upregulation in protein expression was identified by proteomic analysis. At a genomic level, all single nucleotide polymorphisms identified by IonTorent sequencing were independently verified. However, the insertion/deletions, particularly those associated with homopolymeric tracts were all found to be false call-outs, which is a limitation of this sequencing technology. The role of epigenetic changes mediated by genetic rearrangements in a Type I restriction-modification (RM) system on opacity phenotype was also investigated. These rearrangements result in switching between six alternative DNA methylation site specificities impacting on genomic methylation patterns. Examination of six “locked” mutants (SpnD39IIIA-F) with monospecific DNA methylation patterns, indicated that there was an epigenetic impact on colony opacity phenotype and virulence. Importantly, these genetic rearrangements at the Type I RM locus also occurred during experimental infection. However, there were inconsistencies between the opacity phenotypes of locked mutants and the RM allele distribution in wild-type D39O and D39T. Thus, RM allele switching cannot fully account for colony opacity phase variation in pneumococci. This study identified that pneumococcal phase variation is a complex, multifactorial phenomenon and different strains employ alternative mechanisms to attain opacity phenotypes. Furthermore, epigenetic changes impact pneumococcal opacity morphology and pathogenicity. Hence, the role of epigenetic factors in phase variation and pathogenesis should be investigated in future studies.