Comparative analysis of prophage-like elements in Helicobacter sp. genomes

Data collection and identification of Helicobacter prophages

Eighty-one complete Helicobacter genomes were downloaded from NCBI (the National Center for Biotechnology Information). Helicobacter prophages were identified using a previously reported method (Fan et al., 2014). In the first place, we used PHAST (http://phast.wishartlab.com/index.html) to analyze bacterial genomes to find candidate prophages. Next, we screened integrase gene from prophage genomes to drop false positives results. Finally, based on the presence of significant homology between ORFs (open reading frames) and known phage genes, we obtain Helicobacter prophages.

Genomic and comparative genomic analyses of Helicobacter prophages

Prophage flanking sites attL and attR were identified using DNAMAN. Prophage genes were annotated using Glimmer (Delcher et al., 2007). Dot plot comparisons of Helicobacter prophage genomes were carried out with Geneious software (Kearse et al., 2012). Global genome comparison was performed using BLASTn, at NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi), and results were shown by ACT software. For all software, default settings were used.

Prophages in Helicobacter sp. genomes

Eighty-one complete Helicobacter sp. genomes (Table S1) were retrieved. Thirteen prohages (Table 1) were detected using a previously reported method (Fan et al., 2014), eight of them were novel, and five of them have been described in the literature (Luo et al., 2012). Moreover, seven reported prophages (Table 1) from Helicobacter genomes were not detected in the screen (Arnold et al., 2011; Eppinger et al., 2006; Lehours et al., 2011; Luo et al., 2012). Two of them, contained in the genomes of H. acinonychis str. Sheeba and H. felis ATCC 49179, have not been designated. We named them phiHac_1 and phiHFELIS_1, respectively. It is worth noting that phiHac_1, phiHFELIS_1 and two other prophages from H. pylori str. Egypt, ΦHPE1 and ΦHPE2, all lack sequence information. The original papers where these prophages were identified did not provide the sequence information and we cannot retrieve it from the corresponding genomes using our screening method. We therefore discarded them during follow-up analyses. In general, sixteen prophages are analysed.

The size of all Helicobacter prophage genomes varies between 5.5 kb and 39.3 kb. Based on the presence of predicted prophage proteins and the length of the prophage genomes, nine sequences were designated as full-length prophages, and seven sequences were labeled prophage-like elements.

Comparative genomics of Helicobacter (pro)phages

We carried out a comparative genomics analysis of sixteen Helicobacter prophages with known sequence information using dot plot matrix (Fig. 1). Two Helicobacter phages, KHP30 (Uchiyama et al., 2013) and KHP40 (Uchiyama et al., 2012), were selected as the reference for DotPlots. This revealed that most Helicobacter (pro)phages can be sorted into a common group called a ‘cluster’ (designated ‘Cluster A’) based on the similarities of their genomes. Helicobacter phages of Cluster A can be further divided into subclusters, according to their genomic sequences. These were designated subcluster A1 (containing phiNY40_1, phiK750_1, Sheeba, KHP30, KHP40, 1961P, phiHP33, Cuz20 and India7), subcluster A2 (containing Gambia94/24, phiK747_1, phiK749_1 and phiK748_1), subcluster A3 (B38), and subcluster A4 (F16), respectively. Other Helicobacter phages were grouped into Cluster B (phiHH_1), Cluster C (phiHCD_1), and Cluster D (phiHBZC1_1), as appropriate.

Helicobacter phage Cluster A

Based on the similarities of their genomes, Helicobacter Cluster A phages were divided into four subclusters. Phages belonging to one subcluster are more closely related to each other than to phages in the remaining subclusters (Figs. S1 and S2). Some subcluster A1 phages (phiK750_1, Sheeba, KHP30, KHP40, 1961P, phiHP33, Cuz20 and India7) possess 70.57% identity with each other, as determined by multiple genomic sequence alignments in DNAMAN. In addition, a BLASTn comparison of phiNY40_1 and phiK750_1 revealed one major sequence segment (8,953 bp) with 81% identity and three segments (3,550 bp, 3,039 bp, and 1,997 bp) with identity greater than 76%. Based on the multiple genomic sequence alignments, all subcluster A2 phages displayed 82.79% identity between each other.

Different subclusters in Helicobacter phage Cluster A possess segments of DNA similarity. Phages of subclusters A2, A3, and A4 all shared sequence similarity with subcluster A1 phages (Fig. 2). These are remnant prophage-like elements that have lost sequence segments during evolution. Subcluster A2 prophages retained an upstream region with many virion-associated genes of the subcluster A1 prophages. Subcluster A3 prophage (prophage B38) retained only an incomplete upstream region (5.5 kb) of subclusters A1 and A2 prophages. Subcluster A4 prophage (prophage F16) retained a downstream region containing many DNA metabolism genes of the subcluster A1 prophages. Genome organization of most Cluster A phages has been reported (Luo et al., 2012).

Helicobacter phage Cluster B

Cluster B contains only one Helicobacter prophage, phiHH_1. The genome size of phiHH_1, which lacks the lysin gene, is 16.2 kb. Therefore, phiHH_1 is considered to be a prophage-like element. This prophage is integrated into the H. hepaticus ATCC 51,449 genome, extends from HH_0750 (the integrase gene) to HH_0772 (encoding a carbohydrate-binding protein), and contains twenty-three ORFs (Fig. 3; Table S2). PhiHH_1 prophage is flanked by 15 bp attL and attR sites (Table 1). Twelve ORFs were assigned phage gene status after homologous analysis of protein sequences (Table S2). Based on database searches, nine of these encode specific functions, namely, integrase (HH_0750), DNA transposition protein (HH_0752), host-nuclease inhibitor protein Gam (HH_0754), Rha family transcriptional regulator (HH_0755), DNA-binding protein RdgB (HH_0756), phage Tail Collar Domain family (HH_0761), DNA methyltransferase (HH_0763), tape measure protein (HH_0771), and carbohydrate-binding protein (HH_0772).

Helicobacter phage Cluster C

Although Helicobacter prophage phiHCD_1 displays some similarity to the subcluster A1 and A4 phages, it is not sufficiently closely related to be included in a common cluster. Therefore, phiHCD_1 is categorized into Cluster C. The genome size of phiHCD_1 is 24.8 kb, which renders it a full-length prophage. Prophage phiHCD_1, inserted between HCD_00885 (thioredoxin-encoding) and HCD_01020 (transposase-encoding) in the genome of Helicobacter cetorum MIT 99-5656, contains twenty-eight ORFs (Fig. 3). The prophage has identical 13 bp attL and attR sites (Table 1). Based on amino acid sequence homology, we identified eighteen ORFs that have sequence similarity to genes of other phages. It was possible to assign function to thirteen of them (Table S3). These are, accordingly: terminase (HCD_00900); phage tail tape measure protein (HCD_00910); phage structure protein (HCD_00920, HCD_00930, and HCD_00960); phage major capsid protein (HCD_00925); UV radiation resistance protein (HCD_00935); phage prohead protease (HCD_00945); phage tail protein (HCD_00955); holin (HCD_00990); portal protein (HCD_01010); transposase (HCD_01015, and HCD_01020).

Helicobacter phage Cluster D

PhiHBZC1_1 is found in Helicobacter bizzozeronii CIII-1. It belongs to Cluster D and does not share any similarities with other Helicobacter phages. As a full-length prophage, the genome size of phiHBZC1_1 is 39.3 kb. There are fifty-eight ORFs in this genome (Fig. 3), spanning a region from HBZC1_17420 (DNA invertase-encoding) to HBZC1_17990 (site-specific recombinase integrase-encoding). The prophage is flanked by two 14 bp attL and attR sites (Table 1). Sequence alignment analysis indicated some level of similarity between thirty ORFs of prophage phiHBZC1_1 and other known phage genes. Of these, twenty-eight ORFs could be assigned biological functionalities (Table S4).

The genome of phiHBZC1_1 can be divided into several different functional modules. The lysis module includes HBZC1_17600 and HBZC1_17620, which encode a holin and a lysozyme protein, respectively. The DNA packaging and virion-associated modules consist of HBZC1_17440, coding for a phage terminase large subunit; HBZC1_17470, encoding a phage tail protein; phage tail tape measure proteins-encoding HBZC1_17480, HBZC1_17490, and HBZC1_17500; phage tail proteins-encoding HBZC1_17530, HBZC1_17540, HBZC1_17550, HBZC1_17630, HBZC1_17640, and HBZC1_17660; HBZC1_17560, encoding a phage tail sheath-like protein; HBZC1_17670, encoding a phage baseplate protein; capsid proteins-encoding HBZC1_17740 and HBZC1_17750; HBZC1_17860, encoding a portal protein; HBZC1_17880, encoding a phage terminase large subunit; and HBZC1_17900, encoding a phage baseplate assembly protein V. The DNA metabolism module comprises of three genes (HBZC1_17420, HBZC1_17570, and HBZC1_17830), whose predicted protein products are phage DNA invertase, DNA methyltransferase, and DNA polymerase, respectively. The transcriptional regulatory module is composed of HBZC1_17460 (encoding a phage late control D family protein), HBZC1_17930 (coding for the repressor LexA), and HBZC1_17970 (encoding a YcfA family protein). The lysogeny module appears to be limited to HBZC1_17990, whose predicted protein product is a phage integrase.

Putative antibiotic resistance genes and virulence factors associated with Helicobacter prophages

Except for phiHBZC1_1, none of the other characterized Helicobacter prophages contain known antibiotic resistance genes. The protein encoded by HBZC1_17700 shows high similarity to multidrug resistance protein D (emrD) of Salmonella enterica subsp. enterica serovar Infantis (Table 2). Multidrug resistance protein D belonging to the major facilitator superfamily facilitates the transport of a variety of antibiotics (Shaheen et al., 2015).

A range of phage-encoded virulence genes was identified within the Helicobacter prophage sequences (Table 2). A DNA methyltransferase-encoding gene was identified in most of the analyzed Helicobacter prophages. DNA methyltransferase is thought to contribute to the specificity of bacterium-host interactions or H. pylori virulence (Vitkute et al., 2001). Furuta and colleagues (2015) found that DNA methyltransferase genes are rapidly evolving in H. pylori genomes, which facilitates H. pylori adaptation to a new host. A protein encoded by phiNY40_1 (NY40_0553) displayed 23% identity with a serine/threonine kinase of Thiorhodococcus drewsii. Phosphorylation of proteins usually occurs during interactions between bacterial cells and host cells and plays a role in bacterial pathogenesis (Cozzone, 2005). Serine/threonine kinases are considered to affect cell survival pathways and contribute to H. pylori pathogenesis (King & Obonyo, 2015). A putative glycosyltransferase is encoded by phiHCD_1. Glycosyltransferases are involved in biosynthesis of LPS (Luke et al., 2010) that can promote proliferation of gastric cancer cells (Tomoda, Kamiya & Suzuki, 2015). An antitoxin component RelB of the addiction toxin-antitoxin (TA) module system RelBE was identified in phiHBZC1_1. The protein plays a role in cell survival (Park, Son & Lee, 2013).