Whole Genome Sequencing of Aggregatibacter actinomycetemcomitans Cultured from Blood Stream Infections Reveals Three Major Phylogenetic Groups Including a Novel Lineage Expressing Serotype a Membrane O Polysaccharide

Abstract

Twenty-nine strains of Aggregatibacter actinomycetemcomitans cultured from blood stream infections in Denmark were characterised. Serotyping was unremarkable, with almost equal proportions of the three major types plus a single serotype e strain. Whole genome sequencing positioned the serotype e strain outside the species boundary; moreover, one of the serotype a strains was unrelated to other strains of the major serotypes and to deposited sequences in the public databases. We identified five additional strains of this type in our collections. The particularity of the group was corroborated by phylogenetic analysis of concatenated core genes present in all strains of the species, and by uneven distribution of accessory genes only present in a subset of strains. Currently, the most accurate depiction of A. actinomycetemcomitans is a division into three lineages that differ in genomic content and competence for transformation. The clinical relevance of the different lineages is not known, and even strains excluded from the species sensu stricto can cause serious human infections. Serotyping is insufficient for characterisation, and serotypes a and e are not confined to specific lineages.

Keywords: average nucleotide identity; core and accessory genes; principal component analysis; taxonomy.

Figure 1

Neighbour-joining dendrogram of Aggregatibacter actinomycetemcomitans based on 1146 concatenated core genes (1,104,001 nucleotides) of 42 whole genome-sequenced strains. The type strain is designated with a superscript T. Twenty-nine strains were from cases of bacteraemia (designated PN), five oral strains were included to describe lineage III, and seven WGSs were downloaded from Genbank; see Supplementary Table S1 for further description and origin of strains. The outgroup (strains PN_561 and SC1083) reduces the number of core genes and should probably be excluded from the species. Serotypes and phylogenetic lineages are shown; nt, non-typeable by immunodiffusion with antisera [8]. The bar represents 5000 residue substitutions.

Figure 2

Intact and inactivated competence genes from lineage III strains using strain HK_1651 as reference. Red genes have premature stop codons caused by single base substitutions, single base deletions, or double nucleotide insertions. Yellow genes are disrupted by large insertions (11–28 Kb). Green genes have deletions (150–722 bp) of either the first (sxy, pilC, comE) or the last part (comC) of a gene. The first 73 amino acids of gene sxy in strains HK_907 and PN_696 are replaced with an unrelated coding sequence (blue genes). Green and blue arrows mark the transition from competence gene to unrelated sequence, or vice versa.

Figure 3

(A), principal component analysis of presence/absence of accessory homologs in 40 strains of A. actinomycetemcomitans. (B), line plot of the eigenvalues of factors or principal components in the analysis. The two principal components depicted in 3A comprise 30% of the sum of variances of all individual principal components (3B).