Helicobacter pylori HP0231 Influences Bacterial Virulence and Is Essential for Gastric Colonization

The Dsb protein family is responsible for introducing disulfide bonds into nascent proteins in prokaryotes, stabilizing the structure of many proteins. Helicobacter pylori HP0231 is a Dsb-like protein, shown to catalyze disulfide bond formation and to participate in redox homeostasis. Notably, many H. pylori virulence factors are stabilized by the formation of disulfide bonds. By employing H. pylori HP0231 deficient strains we analyzed the effect of lack of this bacterial protein on the functionality of virulence factors containing putative disulfide bonds. The lack of H. pylori HP0231 impaired CagA translocation into gastric epithelial cells and reduced VacA-induced cellular vacuolation. Moreover, H. pylori HP0231 deficient bacteria were not able to colonize the gastric mucosa of mice, probably due to compromised motility. Together, our data demonstrate an essential function for H. pylori HP0231 in gastric colonization and proper function of bacterial virulence factors related to gastric pathology.


Introduction
Helicobacter pylori infection is one of the most common bacterial infections worldwide. Chronic gastritis induced by H. pylori represents the first step of a pathological cascade that can eventually progress to cancer [1,2]. In order to cope with the acidic conditions H. pylori encounters in the stomach, the bacterium produces urease, thereby hydrolyzing urea and increasing the pH locally [3]. In addition, motility mediated by flagella and cell morphology is essential for successful colonization of the gastric epithelium [4].
The immune response towards the bacterium as well as the degree of gastric pathology have been related to the presence of certain bacterial virulence factors. Of those, the Cag Pathogenicity Island (CagPAI) and VacA are considered the major pathogenic factors, and their presence defines populations at higher risk for gastric cancer development [5]. The CagPAI encodes CagA and several genes composing a type IV secretion system (T4SS), capable of translocating CagA into host epithelial cells. Once translocated, CagA interacts with a number of host proteins and activates signaling cascades that contribute to pathology [6]. Conversely, VacA is a secreted cytotoxin which is internalized, forming vacuoles in the host cells, and causing changes in mitochondrial potential [7] or disrupting autophagy [8].
Disulfide bonds, which in prokaryotes are introduced by proteins of the Dsb family, stabilize the structure of many proteins. Dsb proteins are crucial for correct folding or assembly of different pathogenic factors such as toxins, adhesins or components of type III secretion systems. Hence, loss of disulfide bond formation due to mutations in dsb genes were found to reduce pathogenesis of bacteria and viruses [9][10][11][12]. Dsb proteins have been largely studied in bacteria such as E. coli. In contrast, the disulfide bond formation machinery of H. pylori is still less characterized. H. pylori possesses different putative Dsb homologs: HP0595 (DsbI) was identified as a DsbB homolog, HP0377 as DsbC, and HP0265 as CcdA homolog [13,14]. However, it has been recently shown that HP0377 is a cytochrome c maturation (CcmG) protein playing also a role in Dsb isomerization pathway [15]. Initial structural studies proposed HP0231 as a DsbG homolog [16], however, more recent data showed HP0231 to have DsbA-activity [17,18]. Notably, disulfide formation mediated by HP0231 is important for folding, solubilization, production and secretion of H. pylori cysteine rich protein (HcpE), which is suggested to contribute to bacterial virulence [18]. Interestingly, presence of disulfide bonds has been reported in bacterial virulence factors such as VacA [19] or components of the T4SS [20], suggesting these are potential targets of the H. pylori Dsb system and specifically of HP0231.
In order to elucidate whether HP0231 is required for proper functionality of virulence factors containing disulfide bonds such as VacA and the T4SS, we generated HP0231-deficient H. pylori strains. In addition, we studied the impact of loss of HP0231 in gastric colonization in mice. Our results suggest an important role for HP0231 in the establishment of the infection and in determining bacterial virulence.

Generation of H. pylori strains by natural transformation
Plasmids to generate HP0231 knockout strains and the corresponding HP0231 complemented strains were previously described [17]. To knock out hp0231, a PCR fragment, which was composed with the upstream region of hp0231 gene followed by a chloramphenicol cassette and the downstream region of hp0231 gene, was cloned into pGEM-T (Promega, USA) easy vector to generate the knockout plasmid pGMTeasy-Δ231. To complement the knockout strains with HP0231, the hp0231 coding region was amplified from H. pylori 26695 genome DNA and was inserted into the shuttle vector pHeL3 to construct the vector pHeL3-231+. The plasmid was then introduced into knockout strains by natural transformation.
To naturally transform H. pylori, 10μg of plasmid DNA were introduced into around 4x10 8 H. pylori bacterial cells during exponential growth phase. After one-day incubation on WC dent plates, bacteria were transferred onto WC dent plates containing antibiotics for selection. Single colonies were picked, grown and analyzed by PCR and western blot (S1A and S1B Fig). No changes in growth were detected between wild type and HP0231 knockout strains (S1C Fig). H. pylori growth curve H. pylori cells growing at exponential phase on WC dent plates were resuspended in PBS and used to inoculate Brucella Broth dent liquid medium containing 10% FCS at an optical density (OD) of 0.5. Cultures were kept shaking under microaerophilic conditions. Samples were taken at different time points and cell density was determined by measuring OD 600 .
For infection experiments, 1x10 5 cells were seeded on a well of a 24-well plate. After 24h, cells were infected with different H. pylori strains at MOI 20. After 6h infection, cells were photographed, supernatants collected and cell lysates obtained for western blot.

CagA translocation (β-lactamase activity)
TEM-1 β-lactamase activity was determined as described elsewhere [32]. Briefly, AGS cells were infected with H. pylori P12 [TEM-1-CagA] for 2.5 hours, and cells were loaded subsequently with the fluorescent substrate CCF4-AM in the respective loading solution (LiveBLAzer-FRET B/G loading kit; Invitrogen) supplemented with 1 mM probenecid (Sigma). Cells were incubated at room temperature in the dark for 90 min. For quantification of translocation, infected cells were measured with a Clariostar plate reader (BMG Labtech) using an excitation wavelength of 405 nm (10 nm bandwidth), and emission at 460 nm (20 nm bandwidth, blue fluorescence) and 530 nm (15 nm bandwidth, green fluorescence). CagA translocation was defined as the emission ratio at 460 nm (sample-blank) divided by 530 nm (sample-blank) (Blue-to-Green ratio). ELISA 1x10 5 AGS cells were infected by different H. pylori strains and the supernatant was collected after 6h infection. The secretion of Interleukin-8 (IL8) was detected using IL-8 Elisa kit, according to manufacturer's instructions (eBioscience, USA).

Vacuolation Assay
Cell vacuolation was monitored by counting vacuole formation in H. pylori-infected AGS cells. Briefly, 5x10 4 AGS cells were seeded into 6-well plates and infected with different H. pylori strains at MOI 100 for 24h. High magnification pictures (400x) were taken at five random positions for each sample. Cells that showed vacuoles were counted and percentage of cells showing vacuoles in total cell number was calculated.

Bacterial Binding Assay
Human gastric cancer cells were seeded in 96-well plates (1.5x10 5 cells per well). Different H. pylori strains were labeled by cell tracer CFDA-SE (Life Technologies, USA) in PBS at 37°C for 30min. After washing three times with PBS, labeled H. pylori were added at MOI 10 to the cells and incubated for 1h at 37°C. Samples were washed with PBS three times and fixed with 100μL 4% paraformaldehyde (PFA) for 5min. Finally, the samples were resuspended with 200μL FACS buffer and analyzed by flow cytometry. Analysis was performed with a FACS CyAn (Beckman Coulter) and the FlowJo software.
Immunofluorescence 5x10 4 AGS cells were seeded on a well of 4-well chamber slide (Thermo Fisher Scientific, USA). After 24h, AGS cells were infected by CFDA-SE labeled H. pylori PMSS1 or PMSS1Δhp0231 at MOI 5 for 1h at 37°C. Samples were washed 5 times with PBS and fixed with 4% PFA for 10min. After washing twice with PBS, samples were stained with pre-warmed deep red/PBS (1:1000 dilution; Thermo Fisher Scientific, USA) for 30min at 37°C. After washing with PBS, samples were mounted with DAPI (Vector laboratories, USA) and fluorescence pictures were taken by confocal microscopy (Olympus Life Science, Japan) (600x magnification).

Motility Assays
H. pylori motility was assessed on soft agar plates as previously described [17,33] with slightly modifications. Briefly, the soft agar plates used in this study were composed of 0.35% (w/v) agar, 10% (v/v) FCS, 2.8%(w/v) Brucella Broth base, 5 μg/mL trimethoprim, 5 μg/mL amphotericin B, 10 μg/mL vancomycin and 5 μg/mL cefsulodin. H. pylori were inoculated with a sterile pipette tip into a soft agar plate. Migration of the bacteria on the plate was monitored and pictures were taken after 2 to 3 days of culture.

Mouse experiments
6-8 weeks old C57BL/6 mice were used for H. pylori infection. Mice were orally infected with 4x10 8 bacteria three times every two days and sacrificed after 7 days or 1 month. Stomach pieces were weighted and homogenized in Brain-Heart infusion medium. Gastric homogenate dilutions were plated on WC dent plates containing 200 μg/mL bacitracin, 10 μg/mL nalidixic acid and 3 μg/mL polymyxin B and antibiotics for selection. After 4 to 6 days incubation, H. pylori colonies were counted.

Ethics statement
All animal studies were conducted in compliance with European guidelines for the care and use of laboratory animals and were approved by the Government of Oberbayern (AZ-55.2-1-54-2532-147-12).

HP0231 is essential for CagA translocation
As proteins composing the H. pylori T4SS were suggested to form disulfide bonds to maintain structural integrity, we sought to investigate whether lack of HP0231 could alter T4SS functionality by analyzing CagA translocation into gastric epithelial cells. To this end, we deleted hp0231 in different H. pylori strains (S1A and S1B Fig) and used them for infection of AGS cells. H. pylori wild type bacteria induced a hummingbird phenotype (Fig 1A and S2A Fig). No such phenotypic changes were induced in cells infected with bacteria lacking hp0231 or when using an isogenic strain lacking CagA, whereas complementation of hp0231 reconstituted cell elongation. These data suggested impaired CagA translocation in the absence of HP0231. To further analyze CagA translocation after hp0231 deletion, lysates from infected cells were subjected to western blot to detect CagA phosphorylation. Infection of cells with wild type bacteria resulted in CagA tyrosine phosphorylation (Fig 1B and S2B Fig), while in the absence of hp0231, phosphorylation of CagA could not be detected. As expected, CagA was not phosphorylated after infecting the cells with strains lacking cagA or cagE or infected with H. pylori SS1 (Fig 1B).
Since genetic manipulation of H. pylori often leads to secondary genetic changes resulting in impairment of CagA translocation, cells were infected with bacteria in which the hp0231 gene had been complemented. In contrast to hp0231 deficient bacteria, hp0231-complemented H. pylori strains translocated CagA, as detected by CagA phosphorylation (Fig 1C and S2C Fig), excluding off-target effects after deletion of hp0231.
Defects on CagA translocation were further confirmed by performing CagA translocation assays using a β-lactamase activity reporter assay. β-lactamase activity was observed when using CagA from wild type H. pylori, while no activity was detected when CagA from the hp0231 isogenic knockout strain was analyzed (Fig 1D and S2D Fig).
The presence of a functional T4SS has been related to the activation of different signaling cascades in gastric epithelial cells. Thus, we next analyzed whether lack of hp0231 affected T4SS-mediated signaling and focused on canonical NF-κB. Infection of AGS cells with wild type bacteria induced phosphorylation of p65 (Fig 1E), indicating activation of NF-κB upon infection. Notably, p65 phosphorylation was not detected in cells infected with bacteria lacking hp0231. Similar results were obtained when infecting the cells with bacteria lacking cagE, as expected, while the absence of cagA still led to p65 phosphorylation (Fig 1E). Complementation of the gene in the knockout strain led to activation of NF-κB at similar levels as observed in the wild type situation (Fig 1E), confirming that absence of hp0231 attenuates host signaling events induced by H. pylori.
As activation of NF-κB by H. pylori initiates a pro-inflammatory response, mainly translating into IL-8 expression by gastric epithelial cells, we assessed the effect of deleting hp0231 in IL-8 secretion. Cells infected with wild type H. pylori PMSS1 and the cagA deficient isogenic strain secreted IL-8 in response to the bacteria (Fig 1F). Upon deletion of hp0231 only slight levels of IL-8 were detected, similar to the levels observed after infection with bacteria lacking cagE. Complementation of hp0231 again restored IL-8 secretion (Fig 1F).
The T4SS is a protein complex spanning both bacterial membranes and containing several essential protein components. We have shown previously that two of these proteins, CagI and CagL, are produced at strongly reduced levels when genes encoding several other secretion apparatus proteins are missing, suggesting that production of normal CagI and CagL levels depends on correct assembly of these other components within the secretion apparatus [31]. Interestingly, we found that CagI and CagL are strongly reduced in levels after deletion of hp0231 (S2E and S2F Fig), indicating that hp0231 is also involved in correct assembly of the secretion apparatus.
Together, these results show that lack of hp0231 alters H. pylori T4SS functionality, thereby not only hampering CagA translocation into gastric epithelial cells but also affecting proinflammatory responses towards the bacterium.

Absence of HP0231 reduces H. pylori-induced vacuolation in host cells
VacA is another major virulence determinant of H. pylori potentially forming disulfide bonds. Therefore, we analyzed whether lack of hp0231 influenced VacA functionality in vitro by comparing cellular vacuolation upon infection with wild type and knockout bacteria. Infection with wild type bacteria resulted in vacuole formation in host cells (Fig 2A and 2B) that was reduced in the absence of hp0231 to a similar extent as when infecting cells with VacA deficient bacteria. Vacuolation was restored when hp0231 was complemented in the knockout strain (Fig 2A and  2B). Notably, the bacteria lacking hp0231 expressed VacA at similar levels as wild type bacteria ( Fig 2C). However, lack of hp0231 highly impaired VacA secretion (Fig 2D), indicating that defects on HP0231 alter VacA secretion but not its expression.
To exclude unspecific effects of deletion of hp0231, we analyzed whether lack of hp0231 also altered the function of other virulence factors. Specifically, we focused on gGT, since no disulfide bonds have been reported in its structure [34]. No changes in gGT activity were detected when hp0231 was depleted (Fig 2E), confirming that the effects induced by HP0231 are specific for bacterial factors forming disulfide bonds.

HP0231 is required for gastric colonization
Next, we examined the effects of deleting hp0231 in an infection model in vivo. C57BL/6 mice were infected for 7 and 30 days and gastric colonization was assessed by counting colony formation from stomach homogenates. Bacteria lacking hp0231 were not able to colonize the stomach of mice at either time point of analysis (Fig 3 and S3A Fig), while complementation of the gene reverted the phenotype observed. These results indicate that HP0231 is essential for gastric colonization.

Lack of HP0231 does not affect bacterial binding but impairs motility
To colonize the gastric epithelium, H. pylori needs to bind to epithelial cells using different adhesins. As mice infected with hp0231 deficient bacteria were not colonized, we first analyzed whether lack of hp0231 induced defects in H. pylori binding to epithelial cells. Thus, human gastric epithelial cells were incubated with fluorescence-labeled bacteria and binding was determined by flow cytometry and immunofluorescence. Lack of hp0231 did not affect binding of H. pylori to gastric epithelial cells, since no differences in binding were observed between wild type and hp0231 knockout bacteria (Fig 4A and 4B), while lack of the adhesins BabA and SabA slightly reduced bacterial binding (S3B Fig). As no differences in bacterial binding to host cells were detected, we then analyzed whether defects in motility could explain the lack of gastric colonization observed for the hp0231 knockout strains. Indeed, in the absence of hp0231 H. pylori showed reduced motility, as previously reported in other bacterial backgrounds [17] (Fig 4C and S3C Fig), which may explain the defects in colonization detected.

Discussion
Disulfide bond formation is an essential post-translational modification to maintain the proper conformational structure of many proteins. In bacteria, the Dsb system is mainly responsible for introducing disulfide bonds, however, not much data on Dsb proteins of H. pylori have been reported to date. Recently, H. pylori HP0231 was reported as a dimeric oxidoreductase that catalyzes disulfide bond formation in the periplasm [17]. In addition, it was found that HP0231 is important for folding and secretion of HcpE, a protein suggested to play a role in virulence [18]. As other virulence factors of H. pylori were suggested to form disulfide bonds,   we analyzed whether HP0231 was important to maintain their functionality. Thus, we first focused on the T4SS. We observed that CagA translocation was impaired in the absence of hp0231, suggesting that lack of hp0231 may destabilize the structure of the T4SS and therefore, virulence factors as CagA are prevented from being injected into host cells.
Two cysteine residues (C771 and C782) located near the C terminus of the p55 subunit of VacA were shown to form a disulfide bond [19], while mutation of either of these cysteines to serine resulted in decreased secretion of VacA [35]. Upon deletion of hp0231, we observed impaired vacuolating activity, which may be attributed to the loss of the disulfide bond stabilizing both cysteines important for the secretion of the toxin. Indeed, impaired VacA secretion was detected in bacteria lacking hp0231.
Our observations not only would strengthen the crucial role previously suggested for HP0231 in bacterial virulence, but also indicate that the functionality of two of the most important H. pylori virulence determinants is highly dependent on their correct folding controlled by HP0231. Nevertheless, structural studies would be necessary in order to identify specific conformational changes in bacterial proteins upon deletion of hp0231. Although H. pylori might possess other proteins involved in disulfide bond formation, our data indicate that they cannot compensate for the lack of hp0231 concerning functionality of the T4SS and VacA.
Studies examining the role of another Dsb protein of H. pylori (DsbI) showed that impairment of disulfide bond formation highly reduced the ability of H. pylori to colonize the stomach of mice [36]. In line, we observed that bacteria lacking hp0231 were not able to colonize the gastric mucosa of mice. Impaired colonization was not due to reduced binding of H. pylori to gastric epithelial cells, but most likely to compromised motility of the hp0231 deficient bacteria. Defects in motility were already reported for the H. pylori strain N6 when hp0231 was deleted [17]. Our results indicate that this is not a strain specific effect, but that HP0231 is important to maintain bacterial motility in general. Notably, the defects on motility observed for the hp0231 deficient bacteria were not attributed to defects in the structure or distribution of the flagella but to changes in the morphology of the cells [17]. These morphological changes concomitant to a lack of motility would highly hamper colonization, and thus the bacteria would be rapidly cleared.
H. pylori infection is still a highly prevalent infection worldwide and current therapeutic strategies based on antibiotic treatment have failed mostly due to an increased rate of antibiotic resistance. Thus, the search of novel therapies based on specific inhibitors represents an interesting alternative to antibiotics. Our data suggest that HP0231 might potentially be a good target to develop inhibitors for the treatment of H. pylori infection. Blocking HP0231 in bacteria might induce morphological and structural changes, which, in the one hand, would reduce bacterial virulence once the bacterium has colonized the stomach and, on the other hand, would contribute to bacterial clearance.
Together, we provide evidence of a novel and important role for H. pylori HP0231 in bacterial virulence and colonization of the gastric mucosa. Moreover, our results should encourage the development of specific Dsb inhibitors to treat this widespread infection.