The complete genome sequence of the carcinogenic bacterium Helicobacter hepaticus

Abstract

Helicobacter hepaticus causes chronic hepatitis and liver cancer in mice. It is the prototype enterohepatic Helicobacter species and a close relative of Helicobacter pylori, also a recognized carcinogen. Here we report the complete genome sequence of H. hepaticus ATCC51449. H. hepaticus has a circular chromosome of 1,799,146 base pairs, predicted to encode 1,875 proteins. A total of 938, 953, and 821 proteins have orthologs in H. pylori, Campylobacter jejuni, and both pathogens, respectively. H. hepaticus lacks orthologs of most known H. pylori virulence factors, including adhesins, the VacA cytotoxin, and almost all cag pathogenicity island proteins, but has orthologs of the C. jejuni adhesin PEB1 and the cytolethal distending toxin (CDT). The genome contains a 71-kb genomic island (HHGI1) and several genomic islets whose G+C content differs from the rest of the genome. HHGI1 encodes three basic components of a type IV secretion system and other virulence protein homologs, suggesting a role of HHGI1 in pathogenicity. The genomic variability of H. hepaticus was assessed by comparing the genomes of 12 H. hepaticus strains with the sequenced genome by microarray hybridization. Although five strains, including all those known to have caused liver disease, were indistinguishable from ATCC51449, other strains lacked between 85 and 229 genes, including large parts of HHGI1, demonstrating extensive variation of genome content within the species.

Fig. 1.

Circular representation of the H. hepaticus genome. From the outside to the inside, the first two circles represents the positions of coding sequences transcribed in clockwise and anticlockwise direction, respectively. The colors represent the functional categories of the encoded proteins, as shown in the color legend. The third circle depicts areas of the chromosome where the G+C content is >5% higher (pink) or lower (blue) than the average G+C content (window size = 2000). The fourth and fifth circles represent genes with orthologs in H. pylori (black) and C. jejuni (red). The length of the lines representing the orthologs is proportional to the percentage of amino acid identity. The position of the HHGI1 genomic island is marked in the innermost circle.

Fig. 2.

Similarity of H. hepaticus proteins that have orthologs with known function in both H. pylori and C. jejuni with their H. pylori 26695 and C. jejuni orthologs. The color coding is as in Fig. 1. Each of the axes represents the percentages of amino acid identity with the H. pylori and C. jejuni orthologs, respectively.

Fig. 3.

(A) The large genomic island of H. hepaticus (HHGI1). Red arrows represent genes encoding putative membrane-associated proteins, green arrows represent genes encoding proteins with a leader peptide, and blue arrows indicate genes coding for apparent homologs of other bacterial proteins (light blue, V. cholerae VCA0107–VCA0115; violet, V. cholerae Hcp; dark blue, proteins from other bacteria). Turquoise arrows indicate the three genes that encode proteins (HH0252, HH0259, and HH0275) with homology to VirB10, VirD4, and VirB4. Some smaller ORFs transcribed in the same orientation are not depicted individually but shown as open arrows representing a block of genes. The lower part of the figure shows the same genomic region in one strain, 96-1809, where the complete island is lacking and where only the flanking sequences, HH0232 and HH0303, are present. (B) Genomic variation in H. hepaticus. Twelve H. hepaticus strains were tested for hybridization with a whole-genome DNA microarray. Five strains (ATCC51448, ATCC51450, 95-225, 95-557, and 94-739) contained all genes detected by the array. The other seven strains did not hybridize with 85-229 probes, and the positions of these missing genes in the genome of the sequenced strain ATCC51449 are indicated by red lines. The location of the genomic island HHGI1 that is totally or partially deleted in all seven strains is indicated by the blue rectangle. The array experiments identified six more clusters of at least five genes that were not detected in at least one of the strains.