Genome sequence of Helicobacter suis supports its role in gastric pathology

Helicobacter (H.) suis has been associated with chronic gastritis and ulcers of the pars oesophagea in pigs, and with gastritis, peptic ulcer disease and gastric mucosa-associated lymphoid tissue lymphoma in humans. In order to obtain better insight into the genes involved in pathogenicity and in the specific adaptation to the gastric environment of H. suis, a genome analysis was performed of two H. suis strains isolated from the gastric mucosa of swine. Homologs of the vast majority of genes shown to be important for gastric colonization of the human pathogen H. pylori were detected in the H. suis genome. H. suis encodes several putative outer membrane proteins, of which two similar to the H. pylori adhesins HpaA and HorB. H. suis harbours an almost complete comB type IV secretion system and members of the type IV secretion system 3, but lacks most of the genes present in the cag pathogenicity island of H. pylori. Homologs of genes encoding the H. pylori neutrophil-activating protein and γ-glutamyl transpeptidase were identified in H. suis. H. suis also possesses several other presumptive virulence-associated genes, including homologs for mviN, the H. pylori flavodoxin gene, and a homolog of the H. pylori vacuolating cytotoxin A gene. It was concluded that although genes coding for some important virulence factors in H. pylori, such as the cytotoxin-associated protein (CagA), are not detected in the H. suis genome, homologs of other genes associated with colonization and virulence of H. pylori and other bacteria are present.


Introduction
Helicobacter (H.) suis is a very fastidious, spiral-shaped, Gram-negative bacterium requiring a biphasic culture medium at pH 5 enriched with fetal calf serum, and a microaerobic atmosphere for in vitro growth [1]. H. suis colonizes the stomach of more than 60% of slaughter pigs [1,2]. Although the exact role of H. suis in gastric disease in pigs is still unclear, it has been associated with chronic gastritis [3,4] and ulcers of the pars oesophagea of the stomach [5][6][7]. This may result in significant economic losses due to sudden death, decreased feed intake and reduced daily weight gain [8]. A reduction of approximately 20 g/day in weight gain was observed in animals experimentally infected with H. suis, compared to the non-infected control animals [9].
Bacterial gastric disorders in humans are mainly caused by Helicobacter pylori [10]. However, non-Helicobacter pylori helicobacters (NHPH) have also been associated with human gastric disease with a prevalence ranging between 0.2 and 6% [5]. H. suis is the most frequent NHPH species found in humans, where it was originally named "H. heilmannii" type 1 [11]. There are strong indications that pigs may serve as a source of infection for humans [5,12]. In the human host, H. suis has been associated with peptic ulcer disease [13], gastric mucosa-associated lymphoid tissue (MALT) lymphoma [14] and chronic gastritis [15]. In rodent models of human gastric disease, the bacterium causes severe inflammation and MALT lymphoma-like lesions [16].
Up to now, little is known about the pathogenesis of H. suis infections. To improve understanding in the genes playing a role in pathogenicity, gastric colonization and persistence of H. suis, a genome-wide comparison with the well-investigated H. pylori genome was performed. Some virulence factors may indeed be similar for both bacteria. As there may also be differences, ab initio annotations of the H. suis genome were performed as well.

Genome sequencing
A pyrosequencing (454 Life Sciences Corporation, Branford, CT, USA) assay was applied to the genome of the type strain of H. suis (HS1 T = LMG 23995 T = DSM 19735 T ) and H. suis strain 5 (HS5), isolated from the gastric mucosa of two different swine, according to the method described by Baele et al. [1]. Quality filtered sequences were assembled into contigs using a 454 Newbler assembler (Roche, Branford, CT, USA).

Functional annotation
In order to maximize the number of quality gene annotations, two different annotating approaches were followed: cross-mapping with three Helicobacter pylori strains (26695, Shi470, and G27 with NCBI accession numbers NC_000915, NC_010698, and NC_011333, respectively), and ab initio annotation.

Cross-mapping annotation
A custom BLAST [17] database was created from the HS1 T and HS5 genomic contigs. The H. pylori proteome and non-coding RNAs were aligned (tblastn program of the BLAST suite, e threshold set to 10 -3 ) to the H. suis database. For each BLAST hit the following additional information was analysed: 1) (secretion) signal peptide cleavage site if present, as assessed by the SignalP 3.0 program [18,19]; 2) specifications of transmembrane helices (number, start and end positions, presumed topology with regard to the cytoplasmic membrane) if present, as assessed by the TMHMM program [20]; 3) an estimate of the ribosome binding strength of the mRNA region preceding the most probable start codon. Ribosome binding strength was estimated by applying two established facts: i) on an mRNA strand, usually within 20 nucleotides before the actual start codon, the reverse complement of 5 to 7 nucleotides near the 16S rRNA 3' end acts as an attractor and positioner for the ribosomal small subunit; this region is known as the Shine-Dalgarno sequence [21,22]; ii) in Gram-negative bacteria an AU-rich mRNA region some 16 nucleotides long and immediately preceding the Shine-Dalgarno sequence may also attract and position ribosomes to help initiate translation of the correct, biologically active gene product [23,24]. For H. suis, the Shine-Dalgarno sequence was determined to be a subsequence of AGGAGGU (which is the reverse complement of the 3' end of the 16S rRNA), and the minimum AU-richness (equivalent to ribosome binding capacity) of the preceding region was arbitrarily set to 10/16. For each theoretical ORF a range of possible start codons was scored; the higher the similarity to the ideal Shine-Dalgarno sequence, or the AU-richer the preceding region, or the better a combination of both, the more likely the potential start codon is to be the actual start codon.

Ab initio annotation
For ab initio annotation, theoretical open reading frames (ORFs) were first determined using the EMBOSS getorf tool (with minimum ORF length set to 90 nucleotides, and taking all alternative start codons into account) [25]. All ORFs were translated subsequently, and BLAST (blastp program) was performed with an e threshold of 10 -15 against the Uniprot-KB universal protein database. The generalist algorithm of getorf yielded roughly a tenfold of the expected natural ORFs, reducing the risk of false negatives. In order to keep the false positive rate low, extra parameters were considered: 1) percentage alignment between query and hit ORFs; 2) percentage similarity or conservation between aligned portions of query and hit ORFs; 3) ribosome binding strength (for more details see above). To determine the presence of one or more conserved domains a rpsblast search (with default parameter values) was carried out for every single theoretical ORF against the compiled Conserved Domain Database which holds protein domain alignments from several other database sources [26].

General features of the H. suis genome
In the HS1 T genome a total of 1 635 292 base pairs and in the HS5 genome 1 669 960 bp were sequenced, both with an average GC content of 40%. In contrast to H. pylori, only one copy of both the 16S and 23S rRNA genes was detected, but like H. pylori, H. suis has three copies of the 5S rRNA gene. Thirty-eight transfer RNAs were identified. On the whole, 1266 ORFs from HS1 T and 1257 from HS5 were detected, of which 194 and 191 encoded hypothetical proteins respectively. In 98 and 92 ORFs a signal peptide cleavage site was detected, demonstrating predicted secreted proteins of HS1 T and HS5 respectively. The TMHMM program predicted 210 and 206 proteins with at least one transmembrane helix for HS1 T and HS5 respectively. The sequence fraction identical for HS1 T and HS5 is henceforward described together as the "H. suis genome".

Genes possibly involved in gastric colonization and persistence
Homologs of H. pylori genes involved in acid acclimation, chemotaxis, adhesion to gastric epithelial cells, oxidative stress resistance (Table 1), and motility were detected in the H. suis genome. The latter were identified as a flagellar biosystem similar to that of H. pylori [27]. Moreover, H. suis contains a fibrinonectin/fibrinogen-binding protein coding gene, but the corresponding protein lacks a transmembrane helix or signal peptide cleavage site according to the bioinformatics tools mentioned earlier.
Homologs coding for CMP-N-acetylneuraminic acid synthetase (NeuA) (HSUHS1_0474, HSUHS5_0481),  Table S1. Genes coding for members of major H. pylori OMP families (Hop, Hor, Hof proteins, iron-regulated and efflux pump OMPs) could be aligned with the H. pylori genome. Both H. suis strains contain the hof genes hofA, C, E, F, the hop genes hopE, G-2 and H, and the hor genes horB, C, D, and J. Additionally, HS1 T contains homologs of the hopW protein precursor and horE, whereas HS5 possesses additional homologs of horA, horF, and horL. No members of the Helicobacter outer membrane (hom) family were detected in H. suis. Besides the major H. pylori OMP family proteins, the H. suis genome contains some predicted OMPs based on their N-terminal pattern of alternating hydrophobic amino acids similar to porins, encompassing omp29 for HS1 T and omp11 and omp29 for HS5. A 491 amino acids membrane-associated homolog of the virulence factor MviN, aligned for 92% with the MviN homolog of H. acinonychis (Hac_1250), was also present in H. suis.

Type IV secretion systems in H. suis
Of the H. pylori type IV secretion systems (T4SS), only two members of the cag pathogenicity island (cagPAI) were identified in the H. suis genome (cag23/E and cagX). Most members of the comB transport apparatus were present. These include comB2, B3, B6, B8 and a number of additional genes not classified as comB: recA, comE, comL and dprA. H. suis possesses genes encoding VirB-and VirD-type ATPases (virB4, B8, B9, B10, B11, and virD2, D4), all designated members of the H. pylori type IV secretion system 3 (tfs3). The HS1 T and HS5 T4SS are presented in Table 2.

Genes possibly involved in induction of gastric lesions
Homologs of H. pylori genes involved in induction of gastric lesions in the H. suis genome are summarized in Table 3. Homology searches with the H. pylori vacuolating cytotoxin A gene (vacA) identified HSUHS1_0989 in HS1 T . The corresponding protein, which is exceptional in that it is one of the longest in the world of prokaryotes, possesses three small conserved VacA regions (residues 490-545, 941-995, and 1043-1351), followed by an autotransporter region (residues 2730-2983). The amino acid sequence of the HS5 homolog (HSUHS5_0761) could be aligned for 22% with the H. pylori strain HPAG1 sequence, and possesses only one conserved VacA region (residues 242-298), followed by an autotransporter region (1258-1510). In both vacA homologs, no signal sequence was determined. Additionally, an ulcer-associated adenine-specific DNA methyltransferase (HSUHS1_0375, HSUHS5_0957) coding sequence was identified, whereas a molecular homolog of the ulcer-associated restriction endonuclease (iceA) could not be discovered in H. suis. H. suis contains homologs of pgbA and pgbB encoding plasminogen-binding proteins, though both lacking a transmembrane helix or signal peptide cleavage site according to the bioinformatics tools mentioned earlier.
H. suis harbours homologs of genes coding for the H. pylori neutrophil-activating protein (HP-NapA) and γ-glutamyl transpeptidase (HP-GGT). Homologs encoding the H. pylori flavodoxin fldA and the pyruvate-oxidoreductase complex (POR) members porA, porB, porC, and porD were also identified in H. suis.

Genes possibly involved in gastric colonization and persistence
The results of the present study demonstrate that several H. pylori genes involved in acid acclimation, chemotaxis and motility, have counterparts in the H. suis genome. These genes are known to be essential for colonization of the human gastric mucosa [27][28][29][30][31][32]. Several OMP coding sequences were identified by comparative analyses with H. pylori and other bacterial species. H. suis contains some similar members of the major OMP families described in H. pylori [33]. Some of these OMPs have been described to be involved in adhesion of H. pylori to the gastric mucosa, which is widely assumed to play an important role in the initial colonization and long-term persistence in the human stomach. These include the gastric epithelial cell adhesin HorB [34] and the surface lipoprotein, H. pylori adhesin A (HpaA). HpaA, also annotated as neuraminyllactose-binding hemagglutinin, is found exclusively in Helicobacter and binds to sialic acid-rich macromolecules present on the gastric epithelium [35]. On the other hand, H. suis lacks homologs of several other H. pylori adhesion factors, including genes coding for the blood group antigen binding adhesins babA (hopS) and babB (hopT), the sialic acid binding adhesins sabA (hopP) and sabB (hopO), and the adherence-associated lipoproteins alpA (hopC) and alpB (hopB) [36].
H. suis contains a fibrinonectin/fibrinogen-binding protein coding gene, which may enhance its adherence to injured gastric tissue. Damage to host epithelial cells may indeed expose fibronectin and other extracellular matrix components. Strong homology was found with fibronectin-binding proteins of H. felis (YP_004072974), H. canadensis (ZP_048703091) and Wolinella succinogenes (NP_907753). To our knowledge, no exact function has been given to these proteins in these species. In Campylobacter jejuni, however, fibronectin-binding proteins CadF and FlpA have been shown to be involved in adherence to and/or invasion of host's intestinal epithelial cells [37,38]. According to the bioinformatics tools used here, the H. suis fibronectin-binding protein lacks a transmembrane helix or signal peptidase cleavage site, indicating that it is not surface exposed or secreted. Its real role in colonization therefore remains to be elucidated. Three genes involved in sialic acid biosynthesis (neuA, neuB, and wecB) were annotated in the H. suis genome, indicating that this bacterium may decorate its surface with sialic acid. The presence of surface sialylation has been studied extensively in pathogenic bacteria, where it contributes to evasion of the host complement defense system [39].
Additionally, H. suis possesses genes encoding enzymes involved in oxidative-stress resistance (napA, sodB, katA, mutS, mdaB, and peroxiredoxin coding sequence). This indicates that H. suis may harbour a defense mechanism against the host inflammatory response, contributing to the ability of chronic gastric colonization by this bacterium [40].

Type IV secretion systems in H. suis
Two partial T4SS were predicted in the H. suis genome, namely the comB cluster and the tfs3 system. The H. suis comB system probably plays a role in genetic transformation [41,42]. Transformation of DNA can be responsible for the high degree of diversity among H. suis strains as has been recently demonstrated by multilocus sequence typing of available H. suis strains [43]. The role of the H. pylori tfs3 secretion system in pathogenesis is not exactly known. Seven genes of the tfs3 cluster are homologs of genes involved in type IV secretion: virB4, virB11, and virD4 code for ATPases which move substrates to and through the pore. The latter is coded by transmembrane pore genes virB7, virB8, virB9, and virB10 [44]. All these genes, except virB7 were identified in H. suis, indicating that the H. suis tfs3 can be important in transmembrane transport of substrates in H. suis.
The H. pylori cag pathogenicity island (cagPAI) region encodes a T4SS allowing H. pylori to insert the cytotoxin-associated antigen A (CagA) into the host cell. This process results in altered host cell structure, an increased inflammatory response, and a higher risk for gastric adenocarcinoma [45]. Although H. suis possesses two members of the H. pylori cagPAI (cag23/E and cagX), the majority of genes, including the gene coding for pathology-causing protein (CagA), were not identified. This indicates that HS1 T and HS5 lack a functional cag protein transporter secretion system.

Genes possibly involved in induction of gastric lesions
Genomic comparison of H. suis with H. pylori resulted in the identification of additional genes possibly associated with virulence in H. suis. A H. suis homolog of the H. pylori vacA was detected. VacA is both a cytotoxin of the gastric epithelial cell layer, and an immunomodulatory toxin of H. pylori [46]. H. pylori contains either a functional or non-functional vacA. The H. suis vacA homolog exhibits no vacA signal sequence, indicating that it might encode a non-functional cytotoxin [47]. In vitro and in vivo studies with a knockout mutant of the H. suis vacA could clarify the functionality of the vacA homolog in this Helicobacter species. Strong homology was found with two H. pylori virulence-associated genes namely napA, encoding the HP-NapA and ggt, encoding HP-GGT. The H. pylori GGT has been identified as an apoptosis-inducing protein [48,49]. The HP-NapA protein is designated as a proinflammatory and immunodominant protein by stimulating production of oxygen radicals and IL-12 from neutrophils and recruiting leukocytes in vivo [50,51]. Moreover, HP-NapA also plays a role in protecting H. pylori from oxidative stress by binding free iron [52]. H. suis contains homologs of two H. pylori genes coding for plasminogen-binding proteins, pgbA and pgbB. The corresponding proteins, PgbA and PgbB bind host plasminogen, which subsequently can be activated to plasmin and may contribute to obstructing the natural healing process of gastric ulcers [53,54]. The biological role of the H. suis pgbA and B homologs in chronicity of gastric ulceration is uncertain, as no exact membrane association was found in the corresponding proteins.
The risk to develop MALT lymphomas in H. suis infected human patients is higher than after infection with H. pylori [5,14]. Homologs encoding the H. pylori flavodoxin (fldA) and its electron donor, the POR enzyme complex (porA to D) were found in H. suis. The H. pylori flavodoxin protein (FldA) has been proposed to play a role in the pathogenesis of H. pylori-associated MALT lymphoma, as antibodies against the H. pylori FldA protein were more prevalent in patients with MALT lymphomas compared to patients with other H. pylori-related diseases [55]. Besides, insertion mutagenesis of the fldA and the por complex has shown that these genes are essential for the survival of H. pylori [56]. These observations indicate that fldA and its por complex may play a role in gastric colonization of H. suis and MALT lymphoma development in H. suis infected people.
Recently, the genomes of the carcinogenic H. pylori strain B38 and the carcinogenic and ulcerogenic Helicobacter mustelae have been sequenced [57,58]. Both helicobacters lack homologs of major H. pylori virulence genes (e.g. cagA, babA/B, sabA/B), which are also absent in the H. suis genome. Additionally, H. mustelae lacks a vacA homolog. Despite this absence, infection with H. pylori strain B38 and H. mustelae has been associated with gastric MALT lymphomas and other gastric disorders. Whole genome sequencing data are also available from H. acinonychis strain Sheeba, a gastric pathogen of large felines. Similar to H. suis, H. acinonychis lacks a cagPAI as well as genes encoding BabA/B and SabA/B. Both species contain a vacA homolog, which for H. acinonychis has been described to be fragmented [59,60].
H. suis contains a mviN homolog. This gene has been described to be a virulence factor of several bacterial species, such as Burkholderia pseudomallei and Vibrio alginolyticus [61,62]. In addition to virulence, MviN has been described to be essential for in vitro growth of these and other bacteria [61][62][63]. The biological significance of mviN in the Helicobacter genus, however, remains to be elucidated.