Concurrent genotyping of Helicobacter pylori virulence genes and human cytokine SNP sites using whole genome amplified DNA derived from minute amounts of gastric biopsy specimen DNA

Background Bacterial and cellular genotyping is becoming increasingly important in the diagnosis of infectious diseases. However, difficulties in obtaining sufficient amount of bacterial and cellular DNA extracted from the same human biopsy specimens is often a limiting factor. In this study, total DNA (host and bacterial DNA) was isolated from minute amounts of gastric biopsy specimens and amplified by means of whole genome amplification using the multiple displacement amplification (MDA) technique. Subsequently, MDA-DNA was used for concurrent Helicobacter pylori and human host cellular DNA genotyping analysis using PCR-based methods. Results Total DNA was isolated from gastric biopsy specimens of 12 subjects with gastritis and 16 control subjects having a normal mucosa. The DNA was amplified using a multiple displacement amplification (MDA) kit. Next, concurrent genotyping was performed using H. pylori-specific virulence gene PCR amplification assays, pyrosequencing of bacterial 16S rDNA and PCR characterisation of various host genes. This includes Interleukin 1-beta (IL1B) and Interferon-gamma receptor (IFNGR1) SNP analysis, and Interleukin-1 receptor antagonist (IL1RN) variable tandem repeats (VNTR) in intron 2. Finally, regions of the vacA-gene were PCR amplified using M13-sequence tagged primers which allowed for direct DNA sequencing, omitting cloning of PCR amplicons. H. pylori specific multiplex PCR assays revealed the presence of H. pylori cagA and vacA genotypic variations in 11 of 12 gastritis biopsy specimens. Using pyrosequencing, 16S rDNA variable V3 region signatures of H. pylori were found in 11 of 12 individuals with gastritis, but in none of the control subjects. Similarly, IL1B and IFNGR1-SNP and IL1RN-VNTR patterns could be established in all individuals. Furthermore, sequencing of M13-sequence tagged vacA-PCR amplicons revealed the presence of highly diverse H. pylori vacA-s/i/m regions. Conclusion The PCR-based molecular typing methods applied, using MDA-amplified DNA derived from small amounts of gastric biopsy specimens, enabled a rapid and concurrent molecular analysis of bacterial and host genes in the same biopsy specimen. The principles and technologies used in this study could also be applied to any situation in which human host and microbial genes of interest in microbial-host interactions would need to be sequenced.


Background
The Gram-negative spiral shaped bacterium Helicobacter pylori [1] is associated with the development of a variety of gastroduodenal diseases such as chronic gastritis, peptic ulcer disease and gastric cancer [2,3]. Virulence factors have been identified, including the production of urease (ureA) [4], a vacuolating cytotoxin (vacA) [5,6], and a cytotoxin-associated antigen (cagA) [7,8]. It has been implied that ureA [9], superoxide dismutase (sod) [10] and heat-shock protein 60 (hsp60) [11], found exclusively within the cytoplasm in other bacteria, are associated with the outer membrane in H. pylori. It is assumed that the variation in disease progression between patients is likely due to differences in bacterial virulence genes.
The H. pylori cytotoxin gene vacA is an important virulence marker. DNA sequence analysis has revealed that the vacA has a mosaic structure comprising allelic variations in the signal and midregion, each having two different alleles (s1/s2, m1/m2) with different biological activities [5,12]. Furthermore, it has been shown that the repeated hydrophilic motif region (RHM) is a potential proteolytic cleavage-site that separates the VacA-94 kDa protein into a 58-kDa and a 37-kDa protein with different biological activities [6,13]. Amino-acid substitutions, insertions and deletions within the RHM-region have been reported [14,15]. It has been speculated that mutations such as truncations, insertions and deletions in the midregion of the vacA gene which lead to in-frame stop codons, are associated with non-toxic H. pylori strains [14]. Recently, it has been shown that the vacA intermediate region (vacA i-region) may provide an important and independent marker of VacA-associated pathogenicity [16].
It is well recognised that the development of gastric disease is strongly influenced by host genetic factors. Cytokines play an important role in H. pylori-induced disease. An association of increased risk of gastric cancer and IL1B polymorphisms [17][18][19] and IL1RN polymorphisms [17,[19][20][21] has been established. The IFNGR1 allele *2 has been associated with high H. pylori-reactive immunoglobulin G antibody concentrations [22] and an increased risk for atrophic gastritis [23]. Thus, for characterisation of bacterium-host interactions, it would appear that a concurrent typing of Helicobacter-specific virulence genes and cytokine polymorphisms in the same DNA, isolated only once from a biopsy specimen, is desirable.
Numerous PCR assays have been developed for the identification of H. pylori-specific virulence genes, reviewed in [15]. To reduce the number of H. pylori-specific virulence gene-based PCR amplification assays needed for such characterisation, we and others have developed multiplex PCR amplification assays which allow a highly sensitive detection and accurate genotyping of H. pylori-DNA [15,[24][25][26]. A recent study described a method along these lines which allowed for concurrent genotyping analysis of H. pylori 16S rDNA, 23S rDNA, cagA genes, and the cellular IL1B gene [27].
However, due to the low amount of bacterial DNA present in human biopsy specimens, the molecular identification of bacteria is often regarded impossible without culturing. To increase the specificity and sensitivity of the analysis of H. pylori infected biopsy specimens, nested PCR amplification in itself [28] or combined with Southern blot analysis of PCR amplicons has been used for molecular typing [29,30]. Identification of PCR amplicons by DNA sequence analysis employing time-consuming and expensive cloning, and subsequent plasmid DNA isolation procedures has also been used. However, these methods are not well suited for large-scale epidemiological screening and in this context, alternative techniques such as direct sequencing of M13-sequence tagged PCR amplicons for molecular typing of i. e. Staphylococcus aureus protein A (SPA-typing) [31], and whole genome amplification (WGA) of DNA using multiple displacement amplification (MDA) have been developed [32,33].
MDA is a PCR independent isothermal amplification technique that relies on the amplification of total genomic DNA, using a chemically modified random hexamer primer at 30°C and a bacteriophage Phi29 proofreading DNA polymerase, completing amplification within a few hours [32,34] In this study, we report on the use of MDA-amplified DNA for the molecular analysis of bacterial and human host genes in minute amounts of gastric biopsy specimens. Moreover, we report on the successful PCR amplicon sequencing, using a single, universal M13 uni (-21) primer targeting all M13-sequence tagged PCR amplicons.

Study subjects, tissue collection and DNA isolation
Fresh frozen (-80°C) archival tissue samples from 28 subjects from a previous study which was approved by the local ethics committee and conducted in accordance with the Helsinki Declaration, including informed written consent, were used [41]. Based on histological findings and rapid urease test from a larger (n = 501) gastroscopic screening study in a randomly selected cohort of the population of Linköping, Sweden [41], a total of 28 individu-als (14 men; median age 58, range 42-73 and 14 women; median age 65, range 43-76) were selected and included in the present study. Among the 12 individuals with gastritis, four had antrum gastritis, five had pangastritis and two subjects had corpus gastritis with H. pylori infection, respectively, and one individual had corpus gastritis without an obvious H. pylori infection. All sixteen control subjects were H. pylori-negative and had a normal gastric mucosa (for details see table 1). DNA from gastric biopsy specimens and bacterial strains Helicobacter pylori 26695 and J99, respectively, were extracted using a BioRobot M-48 Workstation and a MagAttract DNA Mini-48 kit as recommended by the manufacturer (Qiagen, Hilden, Germany).

Generation of MDA-amplified DNA
MDA of the isolated DNA was carried out using a Genom-iPhi DNA amplification kit as recommended by the manufacturer (GE-Healthcare, Uppsala, Sweden). DNA concentrations were determined using a ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). Three separate MDA reactions were performed from each sample. The integrity of the MDA-amplified DNA was analysed by Ethidium bromide stained agarosegel electrophoresis. Appropriate amounts (usually 1 to 2 μl) were used in subsequent PCR amplification assays.

18S rRNA and 16S rRNA PCR amplification
To evaluate the integrity of the prepared MDA-amplified DNA derived from total biopsy DNA, subsequent 18S rDNA and 16S rDNA PCR amplification was performed. In brief, 18S rDNA PCR amplification was carried out using a Quantum 18S rDNA internal standard kit as recommended by the manufacturer (Ambion, Austin, TX, USA) yielding a 489 bp PCR amplicon. 16S rDNA PCR amplification yielding an approximately 450 bp amplicon was carried out using 16S rDNA broad-range primers [42] flanking the variable V3, V4 and V9 regions ( Table 2) and PCR amplification condition No 3 (Table 3) in a final 25 μl reaction volume using thin walled tubes, a HotStarTaq Master-mix kit (Qiagen, Hilden, Germany), and a PTC-100 Thermocycler (MJ-Research; SDS-Biosciences, Falkenberg, Sweden) or Mastercycler gradient (Eppendorf, Hamburg, Germany) as previously described [38].

Pyrosequencing analysis
Pyrosequencing template generation was performed using 16S rDNA broad-range PCR amplicons as a template (see above). For that purpose, broad-range 16S rDNA PCR amplicons were purified using a GFX PCR DNA and Gelband purification kit (GE Healthcare, Uppsala, Sweden) and pyrosequencing PCR amplicons were generated using an appropriate amount (usually 1 μl) of 16S rDNA template, a HotStarTaq Master mix kit (Qiagen AB, Solna, Sweden), 5 pmol each of Helicobacter-specific primer 5'biotin HJ-HP-JBS.V3 and broad-range primer B-V3.AS (Table 2), both flanking the 16S rDNA V3 region only, and PCR condition No 1 (Table 3). Pyrosequencing analysis was carried out as described elsewhere [43] using a PSQ 96 MA System (Biotage AB, Uppsala, Sweden) and sequencing primer B-V3.AS (Table 2).

Number
Gastritis classification a H. pylori 16S rDNA b  m2 genotypes [24,26], and PCR condition No 1 were used (Table 3). PCR amplicons were analysed on the automated capillary electrophoresis QIAxcel system and a QIAxcel DNA high resolution kit (Qiagen, Hilden, Germany).  VacA-primers Figure 1 VacA-primers. Schematic representation of the approximate location of the vacA-primers used in the present study ( To identify and establish vacA s/i/m genotype, 2 μl MDAamplified DNA derived from the 11 H. pylori-positive biopsies were PCR amplified with three new vacA-specific primer pairs, using 10 pmol of each primer and PCR condition No 4 ( Table 3). The s-region was amplified with primer M13-SeqS.SE and SeqS.AS, the RHM and i-region with primers M13-seqVac.SE and seqVac.AS, and the mregion with primers M13-seqM.SE and VAG-R ( Fig. 1;   (Table 1). 16S rDNA variable V3 region sequence motifs corresponding to H. pylori 26695 and J99 were detected in six and three samples, respectively. Two specimens had deviating V3 region sequence motifs (Fig. 2) (Fig. 3a). The cagA+/vacA+ combination was present in six MDA-amplified DNA samples (No 9,12,22,23,25,27) and was also found in the control strains. The cagA+/vacA-combination was present in three biopsies (No 6, 14, 18), and the vacA-/cagA-and vacA+/cagA-combinations were found in one biopsy each (No 21 and 28, respectively; Fig. 3a).

Cytokine SNP and IL1RN-VNTR analysis
VacA subtyping by DNA sequence analysis DNA sequencing analysis revealed the presence of different vacA genotypes in all H. pylori strains, indicating that sequencing of M13-sequence tagged PCR amplicons is a more discriminating molecular typing approach than analysis of PCR amplicons by capillary gel electrophoresis as done in this study (Fig. 5, 6, 7, 8; Table 4). The apparent absence of vacA amplicons in the multiplex PCR analysis of sample no 6, 14, 18 and 21, is explained by a 70 bp deletion in the RHM region (Fig. 6). The PCR products for these samples are approximately 600 bp and coincide with the hsp60 amplicon of similar size ( Fig. 3a;  (Fig. 5).
DNA sequence analysis of a 900 bp PCR amplicon, produced using MDA-amplified DNA from biopsy No 18 and primers M13-seqVac.AS and VAG-R (Fig. 1), revealed a large deletion in the midregion of the vacA gene resulting in the absence of several primer target sites (Fig. 6). This explains the absence of a vacA m1/m2 genotype in the vacA-subtyping PCR amplification assay of biopsy No 18 (Fig. 3b).

Discussion
Rapidly and easily acquired information on bacterial and host genes is becoming increasingly important for diagnosis and decision-making when choosing suitable therapies for human infectious diseases. This makes new demands on clinical research and routine laboratories to develop molecularly based methods that can accurately identify and characterise bacterial virulence and host susceptibility and/or resistance genes. The amount of genomic DNA available for such genetic analyses is often limited. Quite often, only minute amounts of bacterial DNA can be found in human biopsies. Moreover, it is desirable to perform concurrent, multiple bacterial and host genotyping analyses from the same, limited amount of biopsy DNA. In such studies, the traditional approach is to purify DNA from cultured bacterial strains isolated from the biopsy specimens [27]. Sample preparation is commonly carried out using time-consuming cartridge or bead-based techniques. These methods do not allow an efficient highthroughput analysis of clinical samples and significant variations of DNA yield and purity can be observed. By contrast, an automated nucleic acid extractor combined with MDA-amplification yields DNA of high purity and integrity that can be used in downstream applications.
Indeed, recent studies have demonstrated that PCR ampli-  In recent years, MDA has been tried out for the amplification of microbial DNA [36,48,49] and total DNA (bacterial and cellular) isolated from human biopsy specimens [38,39]. This study shows the feasibility of using MDAamplified total DNA, isolated from human biopsy specimens, for research and clinical diagnostic analysis of both host and infecting bacterium in the same DNA pool.

IL1RN-VNTR genotyping results
Pyrosequencing analysis of the16S rDNA variable V3 region revealed the presence of different H. pylori subspecies in the different biopsy specimens which is in agreement with previous reports showing that subtle DNA sequence variations occur in the 16S rDNA variable V3 region of H. pylori, providing a consistent system for subtyping [30,43]. The taxonomy of these H. pylori strains may be a matter of debate. Subdivision of the species H. pylori into subspecies, based on biotypes, pathotypes or serotypes for taxonomic as well as clinical reasons has been suggested [50,51].
Pro-inflammatory IL1B and IL1RN polymorphisms are associated with increased risk of gastric carcinoma in Caucasian populations [52]. Carriers of these pro-inflammatory polymorphisms revealed an increased IL1B gene expression pattern [19,53] in the mucosa and increased prevalence of intestinal metaplasia and atrophic gastritis [19]. Similarly, a genome-wide linkage analysis identified SNPs in IFNGR1 affecting H. pylori infection [22]. Due to the limited number of biopsies analysed, we were not able to draw any statistically significant conclusion regarding allele frequencies in H. pylori-infected and histologically normal individuals. However, the primary goal of the present study was not to perform a clinical study at large but rather to establish new methodological approaches. In analogy with these findings we have recently shown that MDA-DNA derived from minute amounts of archival plasma/serum DNA allowed us to identify cytokine polymorphic SNP-sites by means of pyrosequencing analysis [40]. The use of IL1B-SNP analysis by means of PCRrestriction-fragment-length polymorphism where IL1B-511, IL1B-31 and IL1B+3954 PCR amplicons are digested with restriction enzyme AvaI, TaqI and AluI, respectively, is a widely accepted approach [19][20][21]. However, restriction enzyme digestion, followed by agarose gelelectrophoresis is a time consuming and, in a clinical routine laboratory context, a tedious process requiring up to 100 ng DNA in each assay [20]. Thus, we conclude that MDA-DNA derived from human biopsy specimens provides a reliable source for cytokine-SNP analysis and, therefore, the same approach may be applied for the characterisation of other host genetic factors in population studies at large VacA s-region nucleotide sequence alignment VacA i-and RHM-region nucleotide sequence alignment Figure 6 VacA i-and RHM-region nucleotide sequence alignment. DNA sequence alignment of the obtained sequences containing intermediate A, B and C, and RHM regions, which are shaded in grey. Reference sequence [GenBank:U05675] (i1) and [GenBank:U29401] (i2) are obtained from Rhead et al. [16]. Dots indicate identical nucleotide sequences and gaps indicate nucleotide deletions. Primer binding sites are indicated with boxes.
VacA m-region protein sequence alignment Figure 7 VacA m-region protein sequence alignment. Alignment of deduced amino acid sequence, translated from the obtained vacA m-region DNA sequences. Samples with m2 genotype are shaded in grey. Reference sequence [GenBank:U05675] (m1) and [GenBank:U29401] (m2) are obtained from Rhead et al. [16]. Dots indicate identical nucleotide sequences and gaps indicate nucleotide deletions. 22  VacA i-region protein sequence alignment Figure 8 VacA i-region protein sequence alignment. Amino acid sequence alignment from the vacA variable i-regions. Shaded areas indicate the important B and C regions described by Rhead et al. [16]. U29401-i2  21  6  14  U05676-i1  J99  26695  12  22  23  25  9  28  27  18   i2   i2-i1   i1 i2-i1 Incorporation of universal M13-sequencetags at the 5'end of PCR primers facilitated straightforward sequencing of amplicons, which makes culturing of bacteria from human biopsy specimens and cloning of PCR amplicons prior to DNA sequencing unnecessary. This makes it possible for clinical routine laboratories using this technique to rapidly produce sequencing results. Sequencing of M13-tagged amplicons was first described for SPA-typing [31]. The technique has successfully been applied in our laboratory for direct sequencing of PCR amplicons in gene expression studies of vasopressin receptor mRNA splicevariants expressed in the human gastro-intestinal tract and surrounding tissues [54]. In the present study we have used M13 sequence-tags for successful partial sequencing of the H. pylori virulence vacA gene (Fig. 6). Multiple H. pylori strain infection was not seen in this study, although the number of samples included is small. Primer design is of utter importance since species-specific primers are essential to gain pure and reliable sequencing results directly from DNA isolated from a mixed flora.
According to a recent study, the intermediate region (iregion) of the H. pylori vacA gene is an important and independent marker of VacA-associated pathogenicity [16]. All i1-type but no i2-type strains induce vacuolation. Chimeric versions of the i-region (i1-i2) can induce low level of vacuolation in vitro. According to this criteria, our results show two tentatively non-toxic strains (s2/i2/m2-No 21 and s1/i2/m2-No 6) and one with reduced toxicity (s1/i1-i2/m2 No 14; No 18 is discussed below). The remaining biopsies contain toxic H. pylori s1/i1/m1 strains (Table 4).
False negative PCR results from the multiplex-PCR (Table  4) were revealed when using new primers, targeting conserved regions of the vacA gene. The negative results were caused by deletions in the RHM region, yielding a smaller amplicon than expected by previous analysis (Fig. 6). One sample (No 18) had large deletions in the RHM region of the gene, leaving out vacA-primer binding sites commonly used in vacA PCR amplification assays ( Fig. 1 and 6). The sequence contained several premature stop codons downstream of the deletion (data not shown), indicating a possible inactive form of VacA. However, further analysis is needed in order to establish if this strain can produce an active protein.
The conclusions to be drawn from the present data is that the choice of primers in vacA PCR amplification assays influences the apparent prevalence vacA-positive H. pylori strains and, hence, precaution has to be taken in the interpretation of vacA-negative PCR amplification results. Thus, our results are in good agreement with previous reports that revealed a high level of vacA genotype variation such as single nucleotide mutations, deletion and in frame stop codons in the vacA alleles of non-toxic H. pylori strains [14,15], but it seems questionable to what extent published vacA PCR amplification results can be compared between studies if DNA sequencing is not performed. From our and other studies it seems evident that it is necessary to establish partial DNA sequences from either vacA PCR amplicons (Fig. 3, 5, 6) or from fulllength vacA open reading frames. Thus, high-throughput sequencing of M13 sequence-tagged PCR amplicons appears to be a more promising approach, both in research and clinical routine laboratories. This principle could be applied to any situation in which sequencing of PCR amplicons is desirable and thereby, one could abstain from tedious amplicon cloning and plasmid preparation procedures prior to DNA sequence analysis.
Histology is considered a sensitive method for detection of H. pylori in gastric biopsies. However, a recent study has demonstrated that PCR amplification using H. pylori-specific primers detected H. pylori-DNA in histological-negative gastric biopsies, indicating the clinical relevance of H. pylori detection by PCR amplification in biopsies with characteristic inflammatory changes [55]. Similarly, Kisa and co-workers [56] evaluated different diagnostic methods for the detection of H. pylori in gastric biopsy specimens. In their study the authors concluded that nested PCR amplification assays are necessary to detect H. pylori-DNA in gastric biopsy specimens. From this study we conclude that MDA-amplified DNA derived from minute amounts of biopsy specimen DNA is well suited for PCRamplification and subsequent sequencing of M13 sequence-tagged amplicons. Thus, it is possible to avoid nested PCR amplification assays which often have to be combined with Southern blot analysis to increase the sensitivity and specificity [28,30].

Conclusion
The highly sensitive PCR-based molecular typing methods described here enables reliable and concurrent genotyping analyses of bacterial and host cellular DNA from the same biopsy specimen at a reasonable cost and time.
Although the power of this approach was demonstrated with concurrent H. pylori and cytokine SNP analysis, the principles and technologies could be applied to any situation in which microbial populations are analysed, and in particular for addressing questions concerning microbialhost interactions associated with human health in general.
the version to be published. AR and HJM were responsible for the acquisition, analysis and interpretation of data. KB selected the biopsy specimens in the study. All authors have read and approved the final manuscript.