Bacterial Strain Typing Using Highly Variable Genomic Sequences

Wenjun Li

2008-01-01

Abstract:With the rapid accumulation of bacterial genome sequences and the identification of an increasing amount of strain-specific characteristics of bacteria, i. E. , host-specific adaptation, virulence and antibiotic resistance, the intraspecies genetic diversity of bacteria was found highly underestimated. This, in turn, highlighted the need for reliable, high resolution, and cost-effective genotyping systems for bacteria at the strain level. In this context, the objective of the work described in this thesis was to develop novel genotyping systems using highly variable genomic sequences rationally selected by genome comparison. We focused our research on four human bacterial pathogens: Bartonella henselae, Tropheryma whipplei, Neisseria meningitidis and Francisella tularensis, for which such tools were not available. For B. Henselae, N. Meningitidis and F. Tularensis, we searched the most variable intergenic spacer sequences that we combined using the multispacer typing (MST) strategy. For T. Whipplei, we identified the most variable genomic fragments, protein-coding or not. For B. Henselae, variations in nine highly variable intergenic spacers identified 39 and 16 genotypes among 126 cat and 75 human strains, respectively, which demonstrated a similar genetic diversity among cat and human strains. The phylogenetic analysis revealed four genetic lineages within B. Henselae and suggested that lineage 4 may be less pathogenic to humans. For N. Meningitidis, MST using three intergenic spacers demonstrated a resolution comparable to multilocus sequence typing (MLST) based on seven housekeeping genes by identifying 27 genotypes among 39 strains. However, MST is cost effective because it requires only three PCR-sequencing reactions in contrast to seven reactions in MLST. In F. Tularensis, following rapid genome sequencing of a clinical isolate by 454 technology, we compared three genome sequences and identified eight highly variable intergenic spacers to construct MST. Our results demonstrated that MST using four highly variable intergenic spacers was valid for individual discrimination and phylogenetic classification of F. Tularensis strains, exhibited higher resolution and more reasonable phylogenetic classification than multilocus variable-number tandem-repeat analysis (MLVA). For subtyping T. Whipplei, 4 highly variable genomic sequences (HVGSs) selected by genomic comparison of two strains were used as genetic markers to identify 24 genotypes among 39 T. Whipplei DNA samples from patients and 10 T. Whipplei DNA samples from asymptomatic carriers. By genotypic and phylogenetic analyses, no significant correlation between HVGS genotypes and clinical manifestations of Whipple’s disease, or asymptomatic carriers, was found for the 49 samples tested. Our observations revealed a high genetic diversity of T. Whipplei strains. In summary, our results demonstrate that rational selection of the most variable genomic fragments, and their combined use for genotyping, is a powerful tool for subtyping bacterial species at the strain level. MST is comparatively more discriminatory than other genotyping methods and thus is especially useful for tracking outbreaks of bacterial infection.