Distinct Modulation of Helicobacter pylori HtrA Activity by Divalent Transition Metals, Impacting HtrA Stability, Oligomerization and E-Cadherin Shedding

Abstract

The Group-1 carcinogen Helicobacter pylori (H. pylori) secretes the serine protease high-temperature requirement A (HtrA), which is directly involved in the disruption of the epithelial barrier in the stomach. HtrA cleaves the extracellular domains of junctional proteins, including E-cadherin (CDH1), claudin-8, occludin, or desmoglein-2, to open intercellular adhesions, allowing H. pylori to transmigrate to subepithelial regions of the gastric mucosa. In our previous work, we found that Zn²⁺ and Cu²⁺ ions efficiently blocked the HtrA activity. However, the impact of other divalent ions on HtrA activity is rather unknown. In this report, we unexpectedly found a stimulating effect through Mn²⁺, Ni²⁺ and Co²⁺ ions on HtrA oligomerization and activity. In contrast to other tested ions, increasing concentrations of Mn²⁺, Ni²⁺ and Co²⁺ strongly enhanced HtrA multimerization as determined in SDS-PAGE under non-reducing conditions and in casein zymography. Increased proteolytic activity of HtrA was further assessed in in vitro cleavage experiments using casein and CDH1 as substrates. Mechanistically, divalent ions bound to HtrA and induced an active conformation, which strongly increased CDH1 cleavage in vitro. The importance of enhanced HtrA activity was finally underlined by the analysis of CDH1 cleavage in H. pylori infection experiments, showing that Ni²⁺ potentiated HtrA-mediated CDH1 shedding. In summary, this study demonstrates that divalent ions exhibit different effects on HtrA activity and that Ni²⁺ and Co²⁺ enhance proteolytic activity by promoting oligomerization, suggesting that metal availability in the gastric environment affects H. pylori virulence.

1. Introduction

Persistent infections with the human pathogen Helicobacter pylori (H. pylori) represent the primary cause for the development of chronic gastritis, ulceration, lymphomas of the mucosa-associated lymphoid tissue (MALT) and gastric adenocarcinoma [1,2]. Even though the overall incidence of stomach cancer is declining, stomach cancer is still the third leading cause of cancer-related deaths. The development of gastric cancer is a gradual process from normal gastric mucosa to intestinal gastric cancer, initiated by H. pylori-induced chronic gastritis, followed by atrophic gastritis, intestinal metaplasia and dysplasia, which ultimately leads to invasive gastric adenocarcinoma [3,4]. H. pylori expresses an arsenal of virulence factors implicated in complex networks of signal transduction pathways leading to inflammation-driven carcinogenesis [5,6]. In particular, the secreted bacterial chaperone and serine protease high-temperature requirement A (HtrA) represents an essential gene, which is expressed by all natural H. pylori strains and plays a major role in pathogenesis [7].

H. pylori HtrA belongs to a family of evolutionarily conserved serine proteases and chaperones and contains an N-terminal signal peptide important for the periplasmic localization. The protease domain of H. pylori HtrA harbors the catalytic triad histidine (H), aspartic acid (D) and serine (S), which is followed by two C-terminally located PDZ (postsynaptic density protein 95 [Dlg4], Disks large 1 [Dlg1] and Zonula occludens-1 [ZO-1]) domains, which are implicated in substrate recognition and binding [8]. HtrA forms trimers, which are inactive towards protein substrates but can form active higher-ordered 12-, 24- and 30-mers [9,10,11]. The expression of HtrA is crucially important for bacterial survival under stress conditions (e.g., extreme pH values, elevated temperatures, salt concentrations, etc.) [12,13]. However, HtrA secreted into the extracellular environment also mediates the cleavage of host cell surface proteins [14,15,16]. It has been demonstrated that HtrA cleaves the ectodomains of E-cadherin (CDH1) [15], claudin-8 (CLDN8), occludin (OCLN) [16] and desmoglein-2 (DSG-2) [14], which represent key molecules in functional adherens junctions, tight junctions and desmosomes in the intact epithelium. The HtrA-mediated shedding of cell adhesion molecules enables H. pylori to locally open intercellular junctions and to transmigrate across the epithelial barrier [15,16]. HtrA-mediated loss of cell–cell adhesion is an important step in the pathogenesis, since H. pylori needs access to the basolateral and basal compartments to inject the oncoprotein cytotoxin-associated gene A (CagA) via a type-IV secretion system (T4SS) into the cytoplasm of host cells [15,16]. Subsequently, CagA deregulates various cellular signaling pathways that control cell proliferation, differentiation and survival, thereby inducing precancerous signaling [16].

The activity of H. pylori HtrA has been intensively investigated in the last 15 years. It has become increasingly apparent that HtrA-mediated cleavage of specific substrates is affected by divalent cations, either by hampering or boosting protease activity or substrate accessibility [17,18,19]. In particular, calcium ions (Ca²⁺) are required for the stabilization of adherens junctions [20,21]. Structurally, CDH1 consists of five extracellular domains (EC1-EC5), a transmembrane domain (TD) and an intracellular domain [21]. Functional adhesion junctions are formed via Ca²⁺-dependent trans-interactions of the EC1-EC5 domains of neighboring epithelial cells. Ca²⁺ binds to linker regions between the individual EC domains [20,22], containing a [VITA]↓[VITA]-x-x-D-[DN] sequence, which was identified as a preferred signature site for HtrA [17]. However, Ca²⁺ binding to these signature sites efficiently protects CDH1 from cleavage by HtrA [19], while the cleavage site located between the EC5 and the TD is Ca²⁺-independent and allows shedding of the 90 kDa extracellular domain into the supernatant of H. pylori-infected cells [17,18]. While Ca²⁺ did not directly alter the proteolytic activity of HtrA, but merely blocked its accessibility to the CDH1 substrate, Zn²⁺ and Cu²⁺ ions exhibited a direct inhibitory effect on the catalytic activity of HtrA through binding to the active pocket of HtrA and through affecting the stability of HtrA via an allosteric loop comprising the region I161-S169 [18], which is involved in stabilizing HtrA in active or inactive conformations [23].

In this study, we investigated the effects of divalent cations on HtrA function and found that Ni²⁺, Co²⁺ and Mn²⁺ increased HtrA oligomerization and caseinolytic activity. However, while Mn²⁺ blocked and Co²⁺/Ni²⁺ enhanced CDH1 cleavage in vitro, only Ni²⁺ potentiated HtrA-dependent CDH1 shedding during H. pylori infection, implying that divalent ions exhibit activity modulating roles in H. pylori pathogenesis.

2. Materials and Methods

2.1. Recombinant Proteins

Recombinant human CDH1 and casein were obtained from Sino Biological (Vienna, Austria) and Carl Roth (Karlsruhe, Germany), respectively. Purification of recombinant HtrA wt (obtained from the H. pylori strain 26695, UniProtKB entry G2J5T2) and the S221A, S164A, D165A, S166A and D168A HtrA mutants has been previously described [18,24]. Briefly, transformed E. coli BL21 was cultivated in terrific broth (TB) medium (Lactan, Vienna, Austria) and the expression of GST-tagged HtrA proteins was induced by the addition of 100 µM IPTG (Thermo Fisher Scientific, Schwerte, Germany) at 30 °C for 3 h. E. coli was harvested and lysed in phosphate-buffered saline (PBS) by sonication. Cell debris was removed by centrifugation at 14,000× g and GST-HtrA proteins were enriched with glutathione sepharose beads (GE Healthcare Life Sciences, Vienna, Austria). The GST-tag was cleaved-off through incubation with PreScission protease (GE Healthcare Life Sciences, Vienna, Austria) for 16 h at 4 °C and the proteins were finally dialyzed in HEPES-buffered saline (HBS; 50 mM HEPES (pH 7.4), 150 mM NaCl).

2.2. In Vitro Cleavage Experiments

For in vitro cleavage experiments, 10 µg of casein and 250 ng of recombinant HtrA were incubated in 20 µL of 50 mM HBS buffer (pH 7.4). Where indicated, either 1 mM of CaCl₂, MgCl2, MnCl₂, FeCl₂, ZnCl₂, BaCl₂, CuCl₂, NiCl₂, CoCl₂, ethylenediaminetetraacetic acid (EDTA), or ethylene glycol-bis(β-aminoethyl ether)-N, N, N′,N′-tetraacetic acid (EGTA) was added. Reactions were incubated for 16 h at 37 °C, mixed with SDS sample buffer and separated by SDS PAGE. CDH1 in vitro cleavage experiments were performed with 50 ng of recombinant human CDH1 that was incubated with 250 ng HtrA or the proteolytically inactive HtrA S221A mutant (SA) in 20 µL of 50 mM HBS buffer (pH 7.4). Divalent ions were added at concentrations of 50, 100, 250, 500 and 1000 µM and reactions were incubated for 16 h at 37 °C. Full-length CDH1 and cleavage fragments were analyzed by Western blotting.

2.3. Casein Zymography

250 ng of recombinant HtrA were incubated in 50 mM HBS buffer with 1 mM of divalent ions, EDTA, or EGTA, or left untreated for 1 h on ice as previously described. Samples were mixed with non-reducing sample buffer and loaded on 10% polyacrylamide gels supplemented with 0.1% casein. After SDS PAGE, gels were renatured for 1 h in 2.5% Triton X-100. HtrA-dependent casein cleavage was conducted by transferring the gels into developing buffer (50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 5 mM CaCl₂ and 0.02% Brij-35) for 24 h at 37 °C. To detect caseinolytic activity, gels were incubated in 0.5% Coomassie Brilliant Blue R250 (Carl Roth, Karlsruhe, Germany) for 16 h, followed by destaining with 40% methanol and 7% acetic acid for 2 h.

2.4. Thermal Shift Assay

Thermal shift assays were performed as previously described [18]. Briefly, 4 µg of recombinant HtrA wt, or the S164A, D165A, S166A and D168A mutants were mixed with 7× SYPRO Orange in 50 mM HEPES buffer (pH 7.4). Where indicated, increasing concentrations of MnCl₂, CoCl₂ or NiCl₂ were added to the reaction mix and reactions were incubated at 25–95 °C in a 96-well plate using an Applied Bioscience StepOne Plus PCR cycler (Thermo Fisher Scientific, Schwerte, Germany). The fluorescence signal was measured to determine the melting temperature of the protein. All reactions were performed in triplicate.

2.5. SDS-PAGE, Coomassie Blue Staining and Western Blot

For non-reducing SDS-PAGE, HtrA wt was incubated with CaCl₂, MgCl₂, MnCl₂, ZnCl₂, BaCl₂, CuCl₂, FeCl₂, NiCl₂, CoCl₂, EDTA, or EGTA at indicated concentrations in HBS buffer for 60 min at 4 °C. Samples were incubated with non-reducing sample buffer without β-mercaptoethanol and separated according to their molecular weight using 10% polyacrylamide (Carl Roth, Karlsruhe, Germany) gels. For SDS PAGE under reducing conditions, samples were boiled in SDS sample buffer containing β-mercaptoethanol and separated by 10% gels. Subsequently, proteins were analyzed by staining with 0.5% Coomassie brilliant blue G250 (Carl Roth, Karlsruhe, Germany). For Western blot analyses, proteins were blotted on nitrocellulose (Carl Roth, Karlsruhe, Germany) by semidry transfer using 192 mM glycine, 25 mM Tris, 20% methanol buffer (pH 8.3). Membranes were blocked in Rotiblock (Carl Roth, Karlsruhe, Germany) for 1 h at room temperature and incubated with a primary antibody against human CDH1 (ab40772, Abcam, Cambridge, UK) for 16 h at 4 °C on a shaker. The membranes were washed 3× for 5 min in Tris-buffered saline supplemented with 0.1% Tween, followed by incubation with a secondary HRP-linked anti-rabbit antibody for 1 h at RT. After three washing steps, CDH1 was detected with ECL prime detection solution (Amersham) using the ChemiDoc XRS+ imager system (BioRad, Vienna, Austria). HtrA was detected with a specific anti-HtrA antibody as previously described [14].

2.6. Cell Culture and Infection Experiments

NCI-N87 cells (ATCC, CRL-5822, Manassas, VA, USA) are gastric epithelial cells isolated from the stomach of a male gastric carcinoma patient. Cells were cultivated in RPMI medium (Sigma Aldrich, Vienna, Austria) substituted with 10% fetal bovine serum (Biowest, Vienna, Austria) and 1% L-glutamine (Biowest, Vienna, Austria) in a humidified atmosphere at 37 °C and 5% CO₂. Cells were seeded in 6-well plates (Greiner, Vienna, Austria) and cultivated for 3 days to reach a density of 1 × 10⁶ cells per well. Before infection, cells were serum-starved and incubated with indicated concentrations of divalent ions. H. pylori strains Hp26695 [25] and N6 wild type [26] were grown on GC agar plates (consisting of 36 g/L GC agar obtained from Sigma Aldrich, Vienna, Austria; 15 g/L Bacto^TM proteose peptone (BD Biosciences, Vienna, Austria)) supplemented with 10% horse serum (Biowest, Nuaillé, France) for 24 h under microaerophilic conditions. For cultivation of the N6 HtrA knockout (N6ΔHtrA) strain [27], H. pylori plates were substituted with 8 µg/mL kanamycin. For infection experiments, H. pylori strains were harvested in PBS. The optical density of the H. pylori suspension was determined by spectrophotometric measurement (OD₆₀₀) using a BioPhotometer Plus device (Eppendorf, Vienna, Austria), and the bacterial concentration was calculated as bacteria/mL using a prepared standard curve. Cells were either mock-treated with PBS or infected with Hp26695, N6 wild type or N6ΔHtrA strains at a multiplicity of infection (MOI) of 20 for 24 h. Supernatants were harvested and cleared from debris by centrifugation at 300× g for 5 min and 15,000× g for 10 min at 4 °C. 100 µL supernatant was mixed with 10 µL of 4× concentrated Laemmli buffer. Samples were boiled at 95 °C with an open lid until supernatants were concentrated to a final volume of 40 µL. Proteins were analyzed by Western blotting.

3. Results

To elucidate how different divalent ions modulate HtrA function, we systematically analyzed their effects on caseinolytic activity, oligomer stability and substrate cleavage. As an initial step, HtrA activity was investigated in in vitro cleavage experiments using casein as a generic protease substrate. EDTA and EGTA served as controls. While HtrA efficiently cleaved casein (Figure 1, lane 3), Zn²⁺ and Cu²⁺ blocked HtrA-mediated casein cleavage (Figure 1, lanes 7 and 9) as reported previously [18]. Ca²⁺, Mg²⁺, Ba²⁺, Fe²⁺, EDTA or EGTA had no effect on HtrA-dependent casein cleavage. Interestingly, among the tested divalent ions, Mn²⁺, Ni²⁺ and Co²⁺ significantly increased casein cleavage (Figure 1A, lanes 6, 11, 12), suggesting that the different divalent ions have specific effects on HtrA activity.

The activity of HtrA is tightly correlated to its ability to form and maintain higher-order oligomeric complexes [28,29]. Therefore, we examined whether divalent cations affect the extent of HtrA oligomerization after incubation of recombinant HtrA with CaCl₂, MgCl₂, MnCl₂, ZnCl₂, BaCl₂, CuCl₂, NiCl₂ or CoCl₂. Monomeric and oligomeric HtrA were then analyzed by SDS PAGE under non-reducing conditions to separate mono- and oligomers, followed by Coomassie blue staining (Figure 1B). In line with our previous observation [18], Zn²⁺ and, to a lesser extent, Cu²⁺ enhanced the multimerization of HtrA compared to untreated HtrA, as evidenced by increased formation of HtrA multimers with a molecular weight above 170 kDa. This observation was underlined by the reduction in monomeric HtrA and auto-processed HtrA (HtrAs) in the range of 40–55 kDa, which are the result of autoproteolytic activities [28] (Figure 1B, compare lane 1 with lanes 5 and 7). In contrast, Ca²⁺, Mg²⁺ and Ba²⁺ did not affect the ratio of HtrA mono- and oligomers (Figure 1B, lanes 2, 3 and 6), while Mn²⁺, Ni²⁺ and Co²⁺ increased HtrA multimerization comparably to Zn²⁺ (Figure 1B, lanes 4, 8 and 9). In order to investigate whether the multimerization induced by the divalent ions is also accompanied by an increase in activity, the proteolytic activity of HtrA was further examined in casein zymography. Therefore, recombinant HtrA was analyzed after preincubation with test ions, and the activity was evaluated through zymography using casein as a substrate. The addition of Mn²⁺, Ni²⁺ and Co²⁺ resulted in increased amounts of multimers, which correlated with enhanced casein degradation (Figure 1C, lanes 4, 8 and 9). Consistent with our previous report, incubation of HtrA with Zn²⁺ and Cu²⁺, which were shown to inhibit HtrA, resulted in a slight increase in HtrA multimer formation. This resulted in an increased HtrA activity in casein zymography, as these ions do not covalently bind to HtrA and are washed out of the gel during zymography preparation [18].

To analyze the effect of ions on oligomerization and activity, we then performed a concentration titration with Mn²⁺, Ni²⁺ and Co²⁺. Based on physiological concentrations for Ca²⁺ [30] and Mg²⁺ [31] in the gastrointestinal tract, HtrA multimerization was tested by applying ion concentrations ranging from 50 to 1000 µM (Figure 2A). The Coomassie-stained gels revealed that Mn²⁺, Ni²⁺ and Co²⁺ increased HtrA multimers already at the lowest tested concentrations of 50 µM (Figure 2A), which was also supported by decreasing monomeric HtrA amounts. Increasing the concentration to 100, 250, 500 and 1000 µM further enhanced HtrA oligomerization, suggesting that Mn²⁺, Ni²⁺ and Co²⁺ stabilize multimerization and activity. These findings were underlined by the detection of HtrA activity in casein zymography. The proteolytic activity of HtrA multimers was enhanced in casein zymography with increasing ion concentrations (Figure 2B), confirming a stabilizing effect of Mn²⁺, Ni²⁺ and Co²⁺ on HtrA.

CDH1 was the first identified biologically relevant substrate for HtrA that is implicated in H. pylori infections [15]. To verify the effects of Mn²⁺, Ni²⁺ and Co²⁺ on HtrA-mediated CDH1 cleavage in vitro, recombinant immunoglobulin-Fc-tagged CDH1 was incubated with the HtrA variants for 24 h with divalent ions as indicated. The CDH1 full-length (CDH1^FL) protein and cleaved fragments were subsequently detected using a specific antibody recognizing the EC5 domain of CDH1 in Western blot analyses. These experiments demonstrated that increasing concentrations of Ni²⁺ and Co²⁺ strongly enhanced the cleavage of CDH1 even at low concentrations of 50 µM, which could be monitored by the loss of CDH1^FL and the increasing amounts of CDH1 cleavage fragments exhibiting molecular weights between 55 and 110 kDa (Figure 3A, middle and right panels). Unexpectedly, Mn²⁺ had an inhibitory effect on CDH1 cleavage activity of HtrA (Figure 3A, left panel), which is in sharp contrast to the previous experiments, indicating that Mn²⁺ increased multimerization and casein cleavage activity. Using casein as a control, it was confirmed that Mn²⁺, Ni²⁺ and Co²⁺ enhanced the caseinolytic activity of HtrA at higher concentrations (Figure 3B).

Previously, we found that Zn²⁺ and Cu²⁺ increased HtrA oligomerization, which involved an allosteric loop in the amino acid stretch I161-S169 [18]. To analyze whether Mn²⁺, Ni²⁺ and Co²⁺ affect HtrA multimerization involving this loop, we examined the thermal stability of HtrA wild type (wt) in the presence of 0.1–1000 µM of MnCl₂, NiCl_2, CoCl₂ and compared the melting temperatures of HtrA wt with HtrA variants harboring point mutations in the allosteric loop. The purity of the recombinant HtrA wild type (wt) and isogenic mutants (S163A, S166A, D165A and D168A) is shown in Figure 4A. In contrast to Mn²⁺ (Figure 4B), both Ni²⁺ (Figure 4C) and Co²⁺ (Figure 4D) increased the stability of HtrA wt at low concentrations of 10–100 µM. However, D165A, S166A and D168A mutations within the HtrA protease strongly inhibited protein stabilization, suggesting that Ni²⁺ and Co²⁺ can directly or indirectly target the allosteric loop to stabilize HtrA multimerization and activity (Figure 4B–D).

To rationalize the observed metal-dependent effects on HtrA oligomerization, thermal stabilization, and casein and E-cadherin cleavage, we used AlphaFold 3 (AF3) to model its metal binding sites [32]. AF3 predictions did not indicate reliable binding sites for alkaline earth metals such as Ca²⁺, Mg²⁺ and Ba²⁺. By contrast, divalent transition metals revealed a consistent binding site for the trimeric HtrA (HtrA₃) on a special position, where the transition metal was six-fold coordinated by H247 and T243 of each HtrA molecule. This centrally positioned transition metal is shown for Zn²⁺ and Co²⁺ in Figure 5A and Figure 5C, respectively. This binding site was consistently found for all transition metals, i.e., Co²⁺, Ni²⁺, Cu²⁺ and Zn²⁺. However, when we used AF3 to predict additional transition metal sites in HtrA₃, we found important differences for the tested transition metals. When we used AF3 to model complex structures of HtrA3 with Zn²⁺ and Cu²⁺, we found that the second most preferred binding site was coordinated by the active site residues S221 and H116 (see Figure 5A,B). By contrast, for Co²⁺ and Ni²⁺, the second most preferred binding site was found to be on the three-fold symmetry axis, coordinated by residues D213 and S215 from each of the three HtrA monomers (Figure 5C,D). The AF3 models reflect statistical likelihoods, not validated structures. The AF3 models can, therefore, only hypothesize why the active site-blocked Zn²⁺ and Cu²⁺ AF3 models should render the protease inactive. However, this hypothesis awaits future experimental validation.

The impact of Ni²⁺ and Co²⁺ on HtrA activity and HtrA-mediated E-cadherin shedding was verified in infection experiments. Since Mn²⁺ strongly inhibited CDH1 cleavage, Mn²⁺ was not further included in infection experiments. Selecting ion concentrations sufficient to increase HtrA multimerization in the thermal shift assays, NCI-N87 cells were treated with 20 µM of Ni²⁺ or Co²⁺. Subsequently, the cells were either mock-treated or infected with H. pylori strain Hp26695 to induce HtrA-mediated E-cadherin cleavage. The shed N-terminal CDH1 (CDH1^NT) fragments in the supernatant were detected by Western blot analysis using a specific antibody against the EC5 domain (Figure 6A). None of the ion-only controls, neither Ni²⁺ nor Co²⁺ treatment, increased CDH1^NT amounts in supernatants of mock-treated cells (Figure 6A, upper panel, lanes 1, 3 and 5). However, H. pylori-induced CDH1 shedding was increased after the addition of Ni²⁺ (Figure 6A, compares lanes 2 and 4), while Co²⁺ had no significant influence (Figure 6A, compares lanes 2 and 6). To demonstrate that Ni²⁺ increased CDH1 cleavage in an HtrA-dependent manner, we included an isogenic HtrA knockout mutant. We found that Ni²⁺ enhanced CDH1 shedding only in infections with the HtrA-positive H. pylori strain, but not after infection with an HtrA-deficient H. pylori mutant (Figure 6B). These data demonstrate that Mn²⁺, Co²⁺ and Ni²⁺ increase the activity of HtrA in vitro, while Ni²⁺ also promotes CDH1 shedding in response to HtrA-positive H. pylori infections.

4. Discussion

The importance of HtrA for H. pylori pathogenesis has been consistently demonstrated in recent years. H. pylori releases HtrA into the extracellular environment, where it cleaves the extracellular domains of the intercellular adhesion proteins CDH1, CLDN8, OCLN and DSG-2 [14,15,16]. Functionally, extracellular HtrA activity leads to the opening of intercellular adhesions between polarized epithelial cells, which is apparently a major function of extracellular HtrA in H. pylori pathogenesis, particularly since HtrA does not target surface proteins nonspecifically [14]. Consequently, HtrA-mediated loss of intercellular adhesion of the epithelium permits H. pylori’s access to basally located β1-integrins, allowing the translocation of the H. pylori oncoprotein CagA into the cytoplasm of infected cells [15,16].

While numerous studies have underscored the role of HtrA in bacterial infections, less is known about the environmental influence on the activity of HtrA oligomers. Here, we observed that divalent ions exhibit different effects on the stability and activity of oligomeric HtrA. Among tested divalent ions, Ca²⁺, Mg²⁺, Ba²⁺ and Fe²⁺ neither influenced HtrA oligomerization nor activity. However, for Zn²⁺, Cu²⁺, Mn²⁺, Co²⁺ and Ni²⁺, we detected different implications for oligomeric stability and activity. As expected, Zn²⁺ and Cu²⁺ efficiently blocked HtrA activity, which is consistent with our previous results [18]. Zn²⁺ has been repeatedly described as a potent inhibitor of serine proteases, often through coordination with active sites or nearby histidine/aspartate residues, thereby disrupting the geometry and electrostatics required for catalysis [34,35]. In addition, Zn²⁺ can disrupt substrate binding or stabilize inactive conformations, resulting in reduced proteolytic activity [36]. In in vitro casein cleavage assays, Zn²⁺ and Cu²⁺ inhibited HtrA activity. The observation of active HtrA oligomers in zymography analyses is attributed to the fact that the non-covalently bound Zn²⁺ and Cu²⁺ ions are washed out during zymography and that the remaining HtrA can degrade the substrate casein polymerized in the gel. While Zn²⁺ and Cu²⁺ increased HtrA oligomer stability and efficiently inhibited HtrA activity [18], we found that Ni²⁺ and Co²⁺ strongly increased HtrA oligomer stability and cleavage activity. These observations are consistent with the AF3 predictions of transition metal complexes with trimeric HtrA₃. In the AF3 models, all tested transition metals, i.e., Zn²⁺, Cu²⁺, Co²⁺ and Ni²⁺, shared a preferred binding site in the HtrA trimer, which exploited H247 of each HtrA monomer in an ideal geometry, additionally coordinated by the hydroxyl of T243. While this binding site was not experimentally validated, the AF3-proposed central metal binding site is consistent with the observed stabilization of oligomer formation induced by these metals. AF3 also predicted additional but different binding sites for Zn²⁺ and Cu²⁺ (S221 and H116 at the active site) and Mn²⁺, Co²⁺ and Ni²⁺ (D213 and S215)₃ at the central three-fold axis, as summarized in

Table 1. Occupancy of AF3-proposed divalent metal binding sites.

Binding Site		Zn2+	Cu2+	Mn2+	Co2+	Ni2+
(His247, Thr243)3	central 3-fold axis	x	x	x	x	x
(His116, Ser221)	active site	x	x
(Asp213, Ser215)3	central 3-fold axis			x	x	x

.

These AF3-proposed binding sites require experimental validation. If these sites should be confirmed, the remarkable differences in the transition metal complexes may well be explained by the hard-soft acid-base concept. Copper and zinc are more strongly polarizable (“soft”) than manganese, cobalt and nickel, and therefore prefer imidazole (histidine) over carboxylates (aspartate) [37]. Given multiple binding sites, it is also possible that a bi- or multi-phasic concentration effect may exist with activity-stimulating effects at low concentrations and activity-inhibiting effects at high concentrations. Therefore, we assume that the binding of the various amino acid residues causes the oligomeric HtrA to adopt an active or inactive conformation, which also significantly influences the allosteric HtrA loop. In H. pylori HtrA, the loop spans the amino acid region I161-S169, which has been identified as an important regulatory element in the allosteric coupling between substrate binding, oligomerization and catalytic activation [18]. This region is located on a surface-exposed region of HtrA and can adopt different conformations that influence the intermolecular contacts between the HtrA subunits. By modulating these protein-protein interfaces, the I161-S169 segment presumably contributes to the control of the assembly state, thus linking the dynamics of the quaternary structure to enzymatic function and substrate processing in H. pylori HtrA [18]. Recently, a cancer-associated 171S/L single nucleotide polymorphism (SNP) close to the allosteric loop was identified that promotes HtrA trimer formation, proteolytic activity and cleavage of epithelial junction proteins [38,39], underlining the importance of the allosteric loop for HtrA functions. Since Ni²⁺ enhances the oligomerization and proteolytic activity of HtrA, increased Ni²⁺ levels in the gastric lumen could act as an environmental cofactor to increase HtrA-mediated cleavage of junctional proteins, thereby decreasing the epithelial barrier integrity and facilitating the entry of H. pylori and inflammatory mediators into subepithelial regions. Notably, the 171S/L SNP is located near the allosteric loop, which is critically involved in oligomer formation [38,39] and could further increase Ni²⁺-dependent activation by stabilizing an oligomerizable, proteolytically active HtrA state. This new model would provide a mechanistic link to epidemiological observations that strains encoding more proteolytically active HtrA variants carry a higher risk of gastric adenocarcinoma, as increased barrier disruption is likely to intensify chronic inflammation and pro-oncogenic epithelial remodeling.

Mn²⁺ had no effect on the melting temperature of HtrA, which contradicts the AF3 prediction and observed increased oligomerization detected in zymography and SDS gels under non-reducing conditions. A plausible explanation for the Mn²⁺-dependent increase in HtrA activity and oligomerization despite the absence of a thermal shift is that Mn²⁺ primarily alters the assembly equilibrium rather than the overall folding stability reported by thermal shift assays. In this scenario, Mn²⁺ would preferentially favor an active, oligomerizable conformation without measurably altering the overall unfolding transition. Consistent with this, metal interactions with low affinity or high dynamics can produce strong functional effects but remain largely unnoticed in thermal denaturation measurements. However, HtrA-mediated CDH1 cleavage was inhibited by Mn²⁺. Even though this is in line with the findings by Russell and colleagues, showing that HtrA from Borrelia burgdorferi is inhibited by high Mn²⁺ concentrations [35], we conclude from our experiments that Mn²⁺ exhibits an activity-stimulating effect on HtrA. Finally, further experiments must be conducted on a possible ion influence on the substrate CDH1, as Mn²⁺, similar to Ca²⁺, could also interfere with the substrates and possibly mask HtrA cleavage sites.

In particular, Ni²⁺ ions enhanced HtrA-mediated CDH1 shedding from H. pylori-infected gastric epithelial cells, implying that H. pylori pathogenesis might benefit from Ni²⁺. Generally, trace elements have an important function in physiological processes and are mostly taken up with food. It is already known that specific diets can support or hamper H. pylori survival and pathogenesis, which could also include the uptake of divalent ions in different ways [40,41]. It has been shown that Ni²⁺ is important for the survival of H. pylori in the stomach, which was correlated with its role as a cofactor of urease, a bacterial enzyme that catalyzes urea degradation and neutralizes the local pH [42,43]. H. pylori has developed specialized membrane transport systems to regulate metal ion uptake, particularly for nickel [41]. Additionally, Co²⁺ is utilized by H. pylori as a cofactor for arginase, an important enzyme in the pathogenesis [44]. The fact that H. pylori utilizes these trace elements for enzymatic reactions highlights their implication in colonization and pathogenesis. The regulatory effect on HtrA adds an important aspect to the functional spectrum of trace elements. It is tempting to speculate whether Ni²⁺/Co²⁺-increased HtrA activity might contribute to H. pylori survival in the harsh environment of the stomach, since HtrA plays a crucially important role in bacterial physiology [7,27].

5. Conclusions

Although persistent H. pylori infections were identified as the causative agent for gastric cancer development, genetic susceptibility and environmental factors, such as diet, alcoholism and smoking, can contribute to the progression of this disease. This study demonstrates that the divalent metal ions, such as Ni²⁺ and Co^2+, can enhance H. pylori HtrA multimerization and proteolytic activity through interaction with a regulatory loop, thereby promoting E-cadherin cleavage and epithelial barrier disruption. These findings reveal a metal ion-dependent mechanism of virulence regulation and suggest that environmental or dietary metal exposure may directly influence H. pylori pathogenicity and gastric disease progression.