Helicobacter cinaedi is a human-adapted lineage in the Helicobacter cinaedi/canicola/'magdeburgensis' complex

Abstract

Helicobacter cinaedi is an enterohepatic Helicobacter that causes bacteremia and other diseases in humans. While H. cinaedi-like strains are isolated from animals, including dog isolates belonging to a recently proposed H. canicola, little is known about the genetic differences between H. cinaedi and these animal isolates. Here, we sequenced 43 H. cinaedi- or H. canicola-like strains isolated from humans, hamsters, rats and dogs and collected 81 genome sequences of H. cinaedi, H. canicola and other enterohepatic Helicobacter strains from public databases. Genomic comparison of these strains identified four distinct clades (clades I-IV) in H. cinaedi/canicola/'magderbugensis' (HCCM) complex. Among these, clade I corresponds to H. cinaedi sensu stricto and represents a human-adapted lineage in the complex. We identified several genomic features unique to clade I. They include the accumulation of antimicrobial resistance-related mutations that reflects the human association of clade I and the larger genome size and the presence of a CRISPR-Cas system and multiple toxin-antitoxin and restriction-modification systems, both of which indicate the contribution of horizontal gene transfer to the evolution of clade I. In addition, nearly all clade I strains but only a few strains belonging to one minor clade contained a highly variable genomic region encoding a type VI secretion system (T6SS), which could play important roles in gut colonization by killing competitors or inhibiting their growth. We also developed a method to systematically search for H. cinaedi sequences in large metagenome data sets based on the results of genome comparison. Using this method, we successfully identified multiple HCCM complex-containing human faecal metagenome samples and obtained the sequence information covering almost the entire genome of each strain. Importantly, all were clade I strains, supporting our conclusion that H. cinaedi sensu stricto is a human-adapted lineage in the HCCM complex.

Keywords: Helicobacter canicola; Helicobacter cinaedi; Type VI secretion system; enterohepatic Helicobacter; genome comparison; human adaptation; metagenome data search.

Fig. 1.

Pairwise ANI analysis of the 67 HCCM complex strains. ANI values were represented in a heatmap. Dendrograms were generated based on the results of average-linkage hierarchical clustering of the ANI values. Clades were defined using 96 % ANI as a cutoff. The isolation sources of each strain are shown.

Fig. 2.

Phylogenetic relationships of the 67 HCCM complex strains. The core-gene-based ML tree of the 67 strains was constructed using the GTR+G4 model and 100 bootstrap replicates. The root of the HCCM complex strains, shown in a dashed box, was inferred from the phylogenetic relationship of the 67 strains with H. labetoulli, an enterohepatic Helicobacter most closely related to the HCCM complex (Fig. S1). The isolation source, clade, country, isolation year, genome size (Mb), GC content (%), and the presence of plasmids (L:≥5 kb, and S:<5 kb) are shown for each strain. Plasmid sequences that were integrated into chromosomes are indicated by asterisks. The presence of AMR-related mutations and T6SS genes are indicated by black squares. Capital letters (R and G) in black squares indicate substitutions by amino acid residues different from those in the known mutations. The presence of tssI genes was confirmed by tblastn search and is indicated by grey squares. The tssH gene (indicated by open circles) is located outside the HVRs shown in Fig. 6. Ten strains whose complete genome sequences are available are also indicated.

Fig. 3.

Differences in genome size (a) and GC content (b) between four clades. Student’s t-test was performed for statistical analysis (P<0.01; **, P<0.001; ***).

Fig. 4.

Accumulation of AMR-related point mutations. (a) Differences in the numbers of mutations between clades are shown. Student’s t-test was performed for statistical analysis (P<0.05; *, P<0.001; ***). (b) Correlation of the numbers of mutations and isolation years in clade I and the other clades (II/III/IV) is shown. The linear regression line, R ² and 95 % confidence interval are indicated.

Fig. 5.

Dot plot matrix analysis of the chromosome sequences of completely sequenced HCCM complex strains. Highly variable regions (HVRs) are indicated by light grey shading and bold lines on the top and right-hand sides. (a) Comparison of eight clade I strains. (b) Comparison of a clade II strain (t36), a clade III (94105) strain and a clade I strain (ATCC BAA-847^T).

Fig. 6.

Comparison of highly variable regions (HVRs) containing T6SS-related genes between eight completely sequenced clade I strains. The gene organization in each HVR and amino acid sequence homologies between genes are shown. T6SS-related genes were divided into three groups and indicated by different colours. (a) Comparison of the HVRs of eight clade I strains. (b) Self-to-self dot plot analysis of the HVRs containing two sets of T6SS-related genes. The data of two strains (P06D7098 and P03D0629) are shown as examples. See Fig. S4 for the data of the remaining three strains. (c) Comparison of a genomic region of a clade II strain (T36) corresponding to the HVR in clade I and the HVR of a clade I strain (P06D0798).

Fig. 7.

Genome sequences of clade I (H. cinaedi sensu stricto) strains detected in metagenome data from three healthy individuals (a) and three colorectal cancer patients (b). The results of read mapping to reference chromosome sequences are shown. The names of metagenome samples and references are indicated at the top and bottom of each panel, respectively. The HVRs are indicated by red lines.