Application of vacA Sequencing in Helicobacter pylori for Classification of Specimens from Healthy Persons, and from Hepatobiliary and Gastroduodenal Patients

Helicobacter pylori is an important causative agent of gastrointestinal and hepatobiliary diseases. The aim of this study was to investigate phylogenetic relationships among H. pylori strains from the oral cavities of healthy individuals, gastric biopsies from gastroduodenal (GI) patients and bile samples from hepatobiliary (HB) patients in the northeast of Thailand. The DNA sequences of a portion of the vacuolating cytotoxin A gene (vacA) were investigated. Phylogenetic trees were constructed using the neighborjoining and maximum-likelihood methods. The vacA sequences of H. pylori fell into four main groups on the trees. Strains from healthy persons fell into two widely separated but well-supported groups, while most H. pylori vacA sequences from HB patients were distributed in another two groups. In contrast, the H. pylori strains from GI patients were scattered across the tree, without a clear geographical pattern. In conclusion, the sequence of vacA may be useful to classify the genetic relationship of H. pylori derived from different sources.

Helicobacter pylori is a Gram-negative spiral-shaped bacterium, a causative agent of many gastrointestinal diseases such as dyspepsia, gastritis, peptic ulcer and gastro-esophageal reflux disease, which can lead to gastric adenocarcinoma 1 .This pathogen has also been recognized as a type I carcinogen by the World Health Organization's International Agency for Research on Cancer because it is associated with gastric mucosa-associated lymphoid tissue lymphoma and gastric adenocarcinoma 2 .
Helicobacter pylori was recently proposed to be associated with diseases outside the gastroduodenal tract, particularly the hepatobiliary duct system 3 .Such diseases include cholangiocarcinoma (CCA), which is a primary cancer of the biliary epithelium and is highly endemic in Northeast Thailand 4 .Several studies have demonstrated that the virulence genes in H. pylori, including the vacuolating cytotoxin gene A (vacA) and cytotoxin-associated gene A (cagA), may help to predict the association between H. pylori infection and clinical outcomes 4,5 .However, cagA and other virulence genes (viz.babA, cagE and iceA) are present in only some strains of H. pylori, whereas vacA is present in all H. pylori strains 6 .
Vacuolating cytotoxin induces intracellular vacuoles, leading to epithelial damage of eukaryotic cells 7 by suppressing epithelial proliferation, promoting apoptotic cell death 8 , inducing cytoskeletal changes 8 and suppressing epithelial proliferation 9 .Each allelic variant of vacA contains one of two classes of signal-region variant (s1 or s2) and one of two classes of middle region (m1 or m2) variant 10 .Allelic variation among strains results in variation of vacuolating activity 11 .The s1/m1 H. pylori strain displays a higher vacuolating activity than that does the s1/m2 strain and might be associated with gastritis and gastric adenocarcinoma.In contrast, the vacA s2/m2 strain is rarely associated with such diseases due to the absence of cytotoxic activity 12 .Although, vacA genotype status might help predict the development of diseases, one report has suggested that genotype of vacA was not associated with the clinical outcome of peptic ulcer, duodenal ulcer and gastric cancer 13 .Therefore, in this study, we test the idea that the vacA sequence can be used to distinguish among clinical sources of specimens.
Routes of H. pylori transmission are oraloral 14 , gastro-oral 15 and fecal-oral 16 .Saliva can be a reservoir of H. pylori, potentially infecting or reinfecting the stomach 17 .It might be that the genotypes of H. pylori strains in saliva of healthy persons can be useful for mapping the geographical distribution of H. pylori strains associated with gastroduodenal (GI) and hepatobiliary (HB) diseases.To explore this possibility, we sequenced portions of the vacA gene from specimens from healthy individuals and from patients with various gastroduodenal and hepatobiliary diseases.

Study subjects
Study subjects consisted of healthy persons, patients with GI diseases and patients with hepatobiliary diseases admitted to Srinagarind Hospital, Faculty of Medicine, Khon Kaen University, Thailand.Approval for the study was obtained from the Human Ethics Committee of Khon Kaen University (approval nos.HE571489 and HE581271) and prior written informed consent was obtained from all participants.

Collection and processing of saliva samples
Thirty saliva samples were collected from healthy person using the method of Silva et.al (2009), with slight modifications 18 .Two ml of each saliva sample was added to 8 ml of Brucella broth (CRITERION™, USA) and incubated for 3 days at 37 °C with shaking under microaerobic atmosphere (5% O 2 , 10% CO 2 and 85% N 2 ).After centrifugation at 13,000 rpm for 10 minutes at 4 °C, the pellet from each culture was used as a source of DNA for amplification of vacA by nested PCR.

Collection and processing of gastric biopsy and bile specimens
Gastric biopsies were performed on dyspeptic patients undergoing gastro-endoscopic examinations at the Endoscopy Unit, Srinagarind Hospital, Faculty of Medicine, Khon Kaen University.Dyspeptic patients (GI) were characterized as having non-ulcer disease (NUD) (n = 20), peptic ulcer disease (PUD) (n = 3) or gastric cancer (GC) (n = 2).Gastric samples from these patients were positive for the rapid urease test indicating H. pylori infection.Bile samples, positive for H. pylori via PCR detection of ureA gene, were taken from patients with hepatobiliary disorders, namely, CCA (n = 20) and cholelithiasis (n = 9), at the Surgical Unit, Srinagarind Hospital, Faculty of Medicine, Khon Kaen University.The sources of the clinical specimens are summarized in Table 1.

DNA extraction
DNA was extracted from pellets of saliva samples using a modified method based on the Puregene DNA Purification System (Gentra System, USA) 19 .DNA was extracted from bile samples and biopsy specimens using the Gentra System DNA extraction and purification kit (Big Lake, MN, USA), respectively, as previously described 4,20 .

PCR amplification and DNA Sequencing
Amplification of a portion of the vacA gene from each sample was performed by nested PCR using primers and conditions as shown in Table 2.The inner primers (forward: nt.2673-2692, reverse: nt.2929-2948; GenBank GQ331975) amplified a fragment of vacA encoding amino acid positions 892 to 979.The region amplified does not include the s or the m region.Nested PCR assay was performed in a total volume of 25 µl containing 500 ng of DNA template, 0.2 mM dNTPs (Amresco, Ohio, USA), 1X PCR buffer (10 mM Tris-HCl pH 8.3, 50 mM KCl, 0.1 mg/ml bovine serum albumin, 10 mM (NH 4 ) 2 SO 4 , and 1.5 mM MgCl 2 ) (RBC bioscience, Taipei, Taiwan), 0.5 U Taq polymerase (RBC Bioscience, Taipei, Taiwan), and 0.4 µM vacA primers.Thermocycling was conducted in a C1000™ Thermal Cycler (BioRad, USA).Amplicons were purified using spin column PCR clean up kit (DNA sequencing service; BIONEER, Korea) and sequenced using Sanger sequencing method (DNA sequencing service; BIONEER, Korea).

Phylogenetic analysis
The vacA sequences from all samples, including sixteen reference sequences of H. pylori vacA from GenBank (Table 1), were aligned using the MAFFT program (http://www.ebi.ac.uk/Tools/ msa/mafft).The aligned sequences were adjusted and edited using BioEdit v.7.0.5.2.Phylogenetic analyses were conducted in MEGA software version 7.0 21 .The phylogenetic trees were constructed and compared using the neighborjoining and maximum-likelihood methods with 1,000 bootstrap replicates.The Kimura 2-parameter model with Gamma distribution was used as the best model (the model giving the best log-likelihood value.

RESULTS
The phylogenetic trees (Figs. 1 and 2) revealed that most sequences from Thai samples fell into distinct clades and were generally also distinct from sequences from other countries.The alignment of 101 partial sequences of the vacA gene was 276 bp in length, and translated to 88 amino acids (Fig. 3).
The neighbor-joining tree (Fig. 1) placed the thirty sequences from healthy Thai people into two very distinct groups with high bootstrap support: group 1 (82% support) consisted of 23 sequences and group 4 (97% support) included 6 sequences.Sequences from Thai hepatobiliary patients mostly fell into two groups, group 2 consisting of 10 sequences from CCA patients together with 5 sequences from Thai gall-stone patients (58% bootstrap support), and group 3 consisting of 8 sequences from CCA patients and 1 sequence from patients with other hepatobiliary  diseases (98% support).In contrast, most vacA sequences from Thai GI patients were not close to those from healthy individuals or HB patients, but were scattered across the tree and often among vacA sequences of GI patients from various countries.Similarly, H. pylori sequences from saliva samples of healthy persons, gastric biopsies of gastrointestinal tract patients (GI) and bile samples of hepatobiliary patients were grouped in the maximum likelihood tree (Fig. 2) in a manner consistent with the neighbor joining tree.Bootstrap values differed slightly between the trees and the 4 groups, while still strongly supported, fell in somewhat different parts of the two trees.
In this study, we also determined the alleles of vacA (s and m) present in the bile samples of hepatobiliary patients which are nearly all of were s1 and m1 (data not shown).Therefore, these alleles can't be used to identify the different clinical origins of samples.
Translated amino acid sequences were compared among the four groups of H. pylori (indicated in Fig. 1 and 2) and found to differ at many positions (Fig. 3).

DISCUSSION
Our study is the first investigation on the genetic diversity and relationships of vacA gene sequences of H. pylori from different sources; saliva from healthy persons, gastric biopsies from gastrointestinal tract patients (GI) and bile samples from hepatobiliary patients (CCA and cholelithiasis) in Thailand.Partial vacA gene sequences from healthy persons were obviously very different from those of HB patients.H. pylori strains from saliva samples of healthy persons fell into two well-supported but widely separated clades (clades 1 and 4).Strains isolated from HB patients, especially those with CCA, mostly also fell into distinct clades (clades 2 and 3).These results indicate that particular genotypes of vacA may occur in strains of H. pylori specific to particular sources (disease status of the host and geography).This applied strongly to strains from CCA patients: this disease is highly endemic in Thailand.Interestingly, most H. pylori strains from GI tracts of Thai patients were distributed among other strains of GI patients from different countries, but not among the strains from healthy persons or HB patients.It seems that most H. pylori strains in saliva of healthy persons in Northeast Thailand may not be derived from the strains isolated from GI and HB patients.The partial vacA gene sequence developed in this study has the potential to distinguish among H. pylori strains from different origins and sources.In this study, genetic diversity and relationships of H. pylori from different sources were determined using neighborjoining and maximum-likelihood methods.Results from the two methods were in broad agreement, suggesting we were close to the true phylogenetic tree.An advantage of the neighbor-joining method is more computational efficiency and speed than maximum-likelihood method, making it suitable for analysis of large data sets.However, this method is unsuitable if very divergent sequences are included 22 .The maximum-likelihood (ML) approach is character-based and is now widely used owing to the development of increasingly realistic and explicit models of sequence evolution.However, likelihood calculations and tree searches under the likelihood criterion are computationally   demanding, making this approach very slow, especially if many sequences are included in the alignment 22 .Researchers can choose the method that best suits their data 24 .
Differences were apparent between the four clades of sequences at both the DNA and the amino acid level.Presumably, this is a consequence of adaptation to different environmental sources, hosts and geography 13,23 .However, the relationship between vacA diversity, origin, and disease state in the host should be further investigated.
In conclusion, these findings support that the partial vacA sequence may be useful as a marker to distinguish H. pylori from difference sources, such as from healthy persons or patients with gastrointestinal and hepatobiliary diseases, and as a marker to investigate genetic relationships among the strains isolated from different sources.

Fig. 1 .
Fig. 1.Neighbor joining tree with 1,000 bootstraps from a dataset of partial vacA gene sequences.Numbers on the branches are bootstrap values (%).Disease conditions are represented by NUD; nonpeptic ulcer diseases, PUD; peptic ulcer diseases and CCA; cholangiocarcinoma

Fig. 2 .
Fig. 2. Maximum likelihood tree with 1,000 bootstraps from a dataset of partial vacA gene sequences.Numbers on the branches are bootstrap values (%).Disease conditions are represented by NUD; nonpeptic ulcer diseases, PUD; peptic ulcer diseases and CCA; cholangiocarcinoma

Table 1 .
Sources of samples used in this study * = All reference strains taken from Genbank were isolated from GI patients

Table 2 .
Primer sequences and PCR condition for detection of H. pylori