Genomic Sequencing of Bordetella pertussis for Epidemiology and Global Surveillance of Whooping Cough

Abstract

Bordetella pertussis causes whooping cough, a highly contagious respiratory disease that is reemerging in many world regions. The spread of antigen-deficient strains may threaten acellular vaccine efficacy. Dynamics of strain transmission are poorly defined because of shortcomings in current strain genotyping methods. Our objective was to develop a whole-genome genotyping strategy with sufficient resolution for local epidemiologic questions and sufficient reproducibility to enable international comparisons of clinical isolates. We defined a core genome multilocus sequence typing scheme comprising 2,038 loci and demonstrated its congruence with whole-genome single-nucleotide polymorphism variation. Most cases of intrafamilial groups of isolates or of multiple isolates recovered from the same patient were distinguished from temporally and geographically cocirculating isolates. However, epidemiologically unrelated isolates were sometimes nearly undistinguishable. We set up a publicly accessible core genome multilocus sequence typing database to enable global comparisons of B. pertussis isolates, opening the way for internationally coordinated surveillance.

Keywords: Bordetella pertussis; bacteria; epidemiology; multilocus sequence typing; nomenclature database; pertussis; phylogeny; surveillance; whole-genome sequencing; whooping cough.

Figure 1

Maximum-likelihood phylogenetic tree for Bordetella pertussis isolates based on the concatenated multiple sequence alignments of 2,038 core genome multilocus sequence typing loci. The tree was rooted on the Tohama reference isolate (GenBank accession no. NC_002929). Only branch support values >50 are labeled (bootstrap/aLRT-SH). Column 1 to the right of the isolates’ names shows colors indicating the intrafamilial groups or groups of multiple isolates from single patients (corresponding to colors shown in inner circle of Figure 2). Column 2 shows pulsed-field gel electrophoresis types (French nomenclature; white, IVα; gray, IVβ; black, IVγ). Column 3 shows multilocus variable-number tandem-repeat analysis types. Scale bar indicates nucleotide substitutions per site.

Figure 2

Maximum-likelihood phylogenetic tree for Bordetella pertussis based on the concatenated alignments of the 2,038 cgMLST loci sequences of isolates from France (this study) and isolates from outbreaks in the United States and the United Kingdom. The tree was rooted on the Tohama reference isolate (GenBank accession no. NC_002929). Black tree branches indicate fim3-1 clade and green tree branches indicate fim3-2 clade. Intrafamilial groups of isolates and multiple groups of isolates recovered from the same patient are represented by sectors of the internal circle surrounding the tree (corresponding to colors in column 1 of Figure 1). The external circle indicates the geographic origin of isolates (blue, France; red, Vermont, USA; orange, California, USA; light pink, other United States; green, United Kingdom). Scale bar indicates nucleotide substitutions per site.