H. pylori 26695 (ATCC700392) was used for construction of unmarked deletion in this study. H. pylori strains were grown on chocolate agar plates supplemented with 10% sheep's blood, and cultivated at 37°C under microaerobic conditions (85% N₂, 10% CO₂, 5% O₂). Brain Heart Infusion broth supplemented with 10% fetal bovine serum (FBS) was used as liquid medium to test the growth of H. pylori with shaking at 100 rpm under microaerophilic conditions after inoculation. The motility of H. pylori was determined on soft agar (0.4%) as described by Worku and Belogolova (Worku et al., 2004; Belogolova et al., 2013). Escherichia coli strains were routinely cultured at 37°C in Luria-Bertani medium. Antibiotics were used at the following concentrations when needed: 15 mg/l kanamycin (Km) or 10 mg/l chloramphenicol (Cm) for H. pylori; 100 mg/l ampicillin (Ap) or 30 mg/l chloramphenicol (Cm) for E. coli. The human gastric epithelial cell line GES-1 (ATCC, Rockville, MD) was used to study the pathogenicity of H. pylori mutant strains. GES-1 cells were cultured with Dulbecco's Modified Eagle Medium (DMEM) containing 10% FBS. In co-cultured system of H. pylori and GES-1 cells, the value of multiplicity of infection (MOI) was 200:1. Bacterial strains and plasmids used in this study are listed in Table 1, and primers are listed in Table 2.

Plasmid pTSK carrying erythromycin-resistant cassette flanked by two FRT sites was used as the backbone (Wang et al., 2014). As the erythromycin-resistance gene (ermF) on pTSK does not work effectively in H. pylori, a kanamycin-resistance gene (aphA) was used to replace ermF in pTSK. pTSK was digested with BssH II into two fragments, and the larger backbone fragment was recovered followed by treated with alkaline phosphatase. aphA was amplified from pHimarEm1 with primers BssHII-KmF and BssHII-KmR, digested with BssH II, and ligated with the backbone. The resulting plasmid, named pTSKHP, contains a kanamycin-resistant cassette flanked by two FRT sites with the same orientation (Figure 1A).

The FLP expression plasmid pCHF (Wang et al., 2014) in Cytophaga hutchinsonii was used as the backbone to construct the FLP recombinase expression plasmid in H. pylori. The pHel1 fragment (Heuermann and Haas, 1995) containing a H. pylori-specific replicon was amplified by PCR using the primers pHel1-1 and pHel1-2, and inserted into pCHF after digested with Sac I and Sal I, to generate pCHF-p. The chloramphenicol-resistance gene (cat) on pCHF could not be expressed in H. pylori. Therefore, another chloramphenicol-resistance gene (catGC) from pTnMax9 was amplified with primers cm-F and cm-R to replace cat after digested with Sal I and Sph I. The yielding plasmid was named pCHFHP (Figure 1A).

The process of unmarked deletion of hp0788 was illustrated in Figure 1B. Double-crossover recombination plasmid for hp0788 was constructed as follows. An 828 bp fragment spanning the first 203 bp of hp0788 and its upstream sequence was used as the upstream homologous arm (H1). It was amplified with primers 0788H1F and 0788H1R, digested with Sac I and Sal I, and then ligated into the corresponding sites of pTSKHP. Downstream homologous arm (H2) was a 1105 bp fragment containing the last 69 bp of hp0788 and its downstream sequence. The fragment was amplified with primers 0788H2F and 0788H2R. After digested with Kpn I and BamH I, it was also ligated into the corresponding sites of pTSKHP, yielding the plasmid pTSKHP-0788.

pTSKHP-0788 was transformed into H. pylori by electroporation as previously described (Ji et al., 2016). The transformants were selected with 15 mg/l of kanamycin and confirmed by PCR with diagnostic primers (test1/kt2 and kt3/test4). The verified transformant cells were used as recipient cells for transforming pCHFHP by electroporation. After incubation at 37°C on chloramphenicol selective plates for 7–10 days, the transformant colonies were inoculated into fresh medium without antibiotics for the elimination of pCHFHP.

Finally, diagnostic PCR with primers test1 and test4 was preformed to verify the excision of aphA gene in the final transformant cells. The cells were also streaked on serum plates containing kanamycin or chloramphenicol to confirm the loss of the exogenous resistance genes.

Total RNA of wild-type H. pylori and hp0788 mutant was isolated using the RNApure Bacteria Kit (CWBIO, Beijing, China). Traces of DNA in extractive RNA were eliminated and first-strand cDNA was synthesized through HiFiScript cDNA Synthesis Kit (CWBIO, Beijing, China). The cDNA was used as the template to carry out PCR with primers listed in Table 2 (RT0786-1, RT0786-2, RT0787-1, RT0787-2, RT0788-1, RT0788-2, RT0789-1, RT0789-2, RT0790-1, and RT0790-2). In PCR control reactions, RNA without reverse transcription was used as the template to determine whether the RNA was free of genomic DNA.

The replicative plasmid pCHFHP was used to complement hp0788 in the mutant. A fragment containing Xba I, digested from the plasmid pSJHK with Kpn I and Sph I, was inserted in the corresponding sites of pCHFHP to add the restriction enzyme site XbaI, generating pCHHPK. The fragment containing hp0788, 240 bp upstream of the start codon, and 28 bp downstream of the stop codon was amplified with primers C0788-1 and C0788-2. The amplicon was digested with Xba I and Kpn I and ligated into the corresponding sites of pCHHPK, yielding pCHHP0788. The plasmid pCHHP0788 was then electroporated into the hp0788 mutant and transformants were selected by chloramphenicol resistance.

The genomic fragment from hp0541 to hp0547 in cag pathogenicity island was selected for gene targeting in this study. The schematic representation of the deletion of this region was shown in Figure 7A. First, one of the terminal genes of this genomic fragment, hp0541, was deleted with the same procedure as described above. Then, the other terminal gene hp0547, was also subjected to unmarked deletion. Finally, the transformants were confirmed by diagnostic PCR with primers T1 and T4.

H. pylori cells were harvested and washed with phosphate-buffered saline (PBS, pH 7.4). The pellets were subjected to ultrasonication. After centrifugation at 16,000 g for 5 min at 4°C, cell debris were removed and the cell lysates were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The proteins in the SDS-PAGE gel were transferred onto a 0.45 μm Immobilon-P PVDF membrane (Millipore, MA, USA). After blocked with skim milk, the membrane was incubated with monoclonal mouse antibody (anti-CagA) (Santa Cruz Biotechnology, CA, USA) and horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Cowin Biotech, Beijing, China) successively. Finally, proteins on the membrane were detected by Immobilon Western Chemiluminescent HRP Substrate (Millipore, MA, USA) according to the instructions.

The GES-1 cells (1 × 10⁵ cells/2 ml/well) were plated in 6-well plates in DMEM medium with 10% FBS at 37°C in a 5% CO₂ incubator overnight. Cells were washed with PBS and co-cultured with H. pylori (2 × 10⁷ colony forming units/ml) in 1 ml of DMEM medium for another 30 min incubation. Subsequently, cells were harvested with 1% trypsin-EDTA and collected by centrifugation. After washed three times with PBS to remove non-adherent bacteria, the remaining bound H. pylori on the cells were measured through plate counting and the urease test. The urease test was carried out as the procedure described by Ki et al. (2010).

GES-1 cells at a concentration of 1 × 10⁵ cells/2 ml/well were plated into 6-well plates overnight in DMEM with 10% FBS. H. pylori cells at mid-exponential phase were collected and added into the GES-1 cells culture at an MOI of 200:1. At the time points of 0, 8, and 16 h in co-culture, morphologic characteristics of GES-1 cells were detected through both crystal violet staining and phalloidin labeling methods. Crystal violet staining was carried out as routine procedure. In phalloidin labeling, cells were firstly fixed by paraformaldehyde for half an hour followed by washed with PBS and labeled by fluor-labeled phalloidin (Invitrogen, Waltham, MA). The images of the stained or labeled cells were finally captured through inverted fluorescence microscope (Olympus, Tokyo, Japan).

GES-1 cells were co-cultured with H. pylori as mentioned above. After 0, 8, or 16 h, GES-1 cells were immediately harvested to detect the cell apoptosis rate by Annexin V-FITC Apoptosis Detection Kit (KeyGEN BioTECH, Nanjing, China). According to the instructions, the assay was performed as follows. 1 × 10⁵ GES-1 cells were firstly suspended with 500 μl Annexin binding buffer containing 5 μl FITC-conjugated Annexin V antibody and 5 μl propidium iodide. Then the mixture was incubated at room temperature for 15 min in the dark. Finally, the relative number of apoptotic cells in the mixture was determined using flow cytometry (Zhang et al., 2016).

Cell Counting Kit-8 (CCK-8) (KeyGEN BioTECH, Nanjing, China) was used to determine the cell viability according to the instructions. GES-1 cells were plated in 96-well plates with an initial density of 10³ cells/well, and H. pylori cells were added at an MOI of 200:1. After co-incubation at 37°C for 0, 8, or 16 h, 10 μl of kit reagent was added for another 3 h incubation. The absorbance at 450 nm of the sample was measured by a microplate reader.

GES-1 cells were infected with H. pylori strains as previously described. The supernatant of the culture was collected at 8 h, and IL-8 was measured by ELISA using an HS Human IL-8 kit (Neobioscience, Shenzhen, China), according to the manufacturer's instructions.

Statistical analysis was carried out using t-test, and a P-value below 0.05 was considered as significant difference.

In order to apply the FLP-FRT recombination system to achieve genetic manipulation in H. pylori, plasmids pTSKHP carrying FRT sites and pCHFHP expressing FLP recombinase were constructed. The essential gene involved in H. pylori colonization, hp0788 (Kavermann et al., 2003), was selected as a target to construct an unmarked deletion.

pTSKHP contains a kanamycin-resistance gene (aphA) flanked by two FRT sites. For gene targeting, homologous arms were inserted into the multiple cloning sites upstream and downstream of the resistance gene (aphA) on pTSKHP. After gene replacement by recombination, aphA flanked by two FRT sites would be inserted into the targeted site on the genome. Plasmid pCHFHP contains FLP recombinase coding gene (flp), which locates downstream of the chloramphenicol-resistance gene (catGC) directly and will be co-transcribed with catGC. The schematic diagrams of pTSKHP and pCHFHP are shown in Figure 1A.

As shown in Figure 1B, the hp0788 recombinant vector (pTSKHP-0788) was constructed according to the procedure and transformed into H. pylori cells by electroporation. Kanamycin-resistant colonies were selected and verified by PCR. As shown in Figure 1C, two expected bands spanning the neighboring genes and the kanamycin resistance gene were amplified in the mutant from both upstream and downstream, which were absent from the wild-type strain. This result confirmed that hp0788 had been replaced by the kanamycin resistance gene. Then, the kanamycin resistance gene was ejected with the transformation of the flp carrying plasmid pCHFHP. In order to verify the gene deletion between two FRT sites, diagnostic PCR with primers test1/test4 was performed and the results are shown in Figure 1D. A band with an expected size was amplified from each tested transformant, which indicated that the kanamycin resistance gene was lost. Sequencing analysis of the mutants confirmed the deletion of hp0788.

The arrangement of hp0788 and surrounding genes in the genome of H. pylori is shown in Figure 2A. To investigate whether the transcription of the surrounding genes is affected by the deletion of hp0788, RT-PCR was performed as previously described. In the wild type of H. pylori, amplicons corresponding to hp0786, hp0787, hp0788, hp0789, and hp0790 were all present, while the hp0788 amplicon was not obtained in the hp0788 mutant (Figure 2B). This result indicated that hp0788 was deleted. The existence of amplicons corresponding to hp0786, hp0787, hp0789, and hp0790 in the mutant suggests that the deletion of hp0788 do not affect the transcription of surrounding genes. All the amplicons were confirmed by sequencing analysis.

Kavermann et al. reported that hp0788 is an essential gene for H. pylori colonization in the gerbil stomach (Kavermann et al., 2003). However, the exact pathogenic role of this gene is not clear. Therefore, the characteristics of the hp0788-deleted mutant were identified in this paper.

First, bacterial morphology, growth curve, cell motility and adherence to GES-1 cells were determined to analyze the phenotypes of the hp0788-deleted mutant. The hp0788 deficient strain exhibited similar morphology and structure to the wild-type strain according to the observation under transmission electron microscopy (Figure S1). As shown in Figure 3A, the mutant had a similar growth rate with the wild type except that the final cell density of the mutant was a little higher. The results of the motility assay reveal that the deletion of hp0788 do not change the motility of the bacteria either (Figure 3B). However, the adherence of H. pylori to GES-1 cells was changed due to the deletion of hp0788 according to plate counting and the urease test. For the wild type, the mean number of bacteria per cell is 37.7, while the numerical value is 30.2 for the mutant (Figure 3C). The results of urease test also illustrated the reduction in adherence of H. pylori to GES-1 cells (Figure S2).

To verify that the reduction in adherence was caused by the deletion of hp0788, complementation of the mutant was carried out as described in section Materials and Methods. As shown in Figure 3C, the reduction in adherence of H. pylori to GES-1 cells was almost restored in complemented strain. These results imply that HofF acts as an adhesin in the infection process of H. pylori, which is consistent with the previous reports (Kavermann et al., 2003; Cheng et al., 2016).

GES-1 cells were co-cultured with H. pylori as previously described, and morphological changes of GES-1 cells were detected. Figure 4A shows the images of GES-1 cells stained with crystal violet (magnification, 100×). Without infected by H. pylori, GES-1 cells exhibited agglomerated growth at 8 and 16 h. The number of GES-1 cells at 16 h was higher than the number of cells at 8 h, revealing the proliferation of GES-1 cells. After co-cultured with H. pylori (both the wild type and the mutant), GES-1 cells was more dispersed. The numbers of GES-1 cells at 8 and 16 h were lower than the control, which indicated that the proliferation of GES-1 cells was inhibited. Figure 4B shows the images of GES-1 cells labeled by fluor-labeled phalloidin (magnification, 200×). Without infected by H. pylori, GES-1 cells were regular and intact. After co-cultured with H. pylori (both the wild type and the mutant), some cells turned irregular and even became round. No obvious differences were observed between cells cultured with the wild type and with the mutant.

Cell apoptosis assay was performed as described in section Materials and Methods, and the relative number of apoptotic cells was determined using flow cytometry. Figure 5A shows the representative flowcharts, in which apoptosis cells occurred in the second and fourth quadrants. Apoptosis rates of GES-1 cells are calculated from the statistical analysis of apoptotic cells, which is shown in Figure 5B. Cell apoptosis rate of GES-1 cells was very low (about 4.00 ± 3.15%) without infection with H. pylori. After infected by the wild-type H. pylori, the apoptosis rate of GES-1 cells increased to 15.73 ± 7.84% after 8 h and reached 25.26 ± 5.81% after 16 h. For the co-incubation with the mutant, cell apoptosis rate increased to12.46 ± 12.3% and 21.13 ± 10.09% at 8 and 16 h, respectively. Compared with cells infected with the wild type strain, those cells infected with mutant strain showed a reduction in apoptosis. When the complemented strain was used to infect GES-1 cells, the apoptosis rate of GES-1 cells recovered to 15.29 ± 13.1% at 8 h and 25.57 ± 9.46% at 16 h, respectively. These results suggest that the deletion of hp0788 disrupts the action of H. pylori to the GES-1 cells in cell apoptosis.

The viabilities of GES-1 cells after infection with the wild-type H. pylori, the Δ0788 mutant and the complemented strain were also comparatively analyzed. As shown in Figure 6, cell viability of the control without being co-cultured with H. pylori increased with time. In the samples co-cultured with either the wild-type or the mutant strain, cell viability decreased significantly. When infected by the wild type, the value of cell viability decreased to 53.43 or 41.76% relative to the initial value after 8 or 16 h. But with infection by the mutant, the value of cell viability retained 66.07 and 49.84% after 8 and 16 h, respectively. When the complemented strain was used, the value of cell viability decreased again, to 52.49 and 42.68% after 8 and 16 h, which was basically of the same level with wild type strain. These results show that deletion of hp0788 decreases the effect of H. pylori on the viability of infected GES-1 cells. The tendency of cellular damage of infected GES-1 cells is consistent with the results of the cell apoptosis and viability assay.

The cag pathogenicity island (cagPAI) is a very important region for the pathogenicity of H. pylori and the functions of genes within this region are still incompletely clear. So the genomic fragment from hp0541 to hp0547 (about 10 kbp) on cagPAI which contains cagA gene and several genes encoding the type IV secretion system was selected as the target region.

As shown in Figure 7A, the deletion of this region was started with the disruption of hp0541 using pTSKHP-0541, followed by transformation of pCHFHP to obtain the unmarked hp0541 mutant (Δ0541). Then, hp0547 was replaced by the resistant gene on pTSKHP flanked by two FRT sites in Δ0541 mutant. The fragment containing the resistant gene and the other genes in this region will be evicted by the transformation of pCHFHP. The new junction fragments amplified from resultant colonies are shown in Figure 7B. A band with expected size was present in the mutant, while no band was amplified in the wild-type strain, mainly because the undeleted fragment is too long to be amplified using regular PCR procedure. The DNA sequencing of the amplified fragment from the mutant confirmed the deletion. After cultured without antibiotics, the mutant cells showed sensitive to both kanamycin and chloramphenicol (Figure 7C), suggesting the exogenous antibiotic-resistance genes had been eliminated.

Western blot of CagA was performed to examine the deletion on protein level. The result showed that CagA was present in the wild type, but absent from the mutant strain (Figure 7D). This result also indicated that the large fragment from hp0541 to hp0547 had been excised. This mutant provides a useful material for further study on the pathogenic role of the cag pathogenicity island.

It is clear that CagA and cagPAI could induce secretion of interleukin 8 (IL-8) in the infected cells, such as AGS cells and GES-1 cells (Fischer et al., 2001; Odenbreit et al., 2002). To further confirm the deleted mutant, IL-8 production was measured, and the results are shown in Figure 8. Consistent with the results reported by other articles (Fischer et al., 2001; Odenbreit et al., 2002; Belogolova et al., 2013), IL-8 secretion in GES-1 cells was much lower following infection with the mutant (Δ0541-0547) compared with the wild type.

The method of unmarked deletion in H. pylori through FLP-FRT recombination system was submitted in patent application (CN106086054A), and the patent application is under substantive examination presently.