Cortactin Promotes Effective AGS Cell Scattering by Helicobacter pylori CagA, but Not Cellular Vacuolization and Apoptosis Induced by the Vacuolating Cytotoxin VacA

Cortactin is an actin-binding protein and actin-nucleation promoting factor regulating cytoskeletal rearrangements in eukaryotes. Helicobacter pylori is a gastric pathogen that exploits cortactin to its own benefit. During infection of gastric epithelial cells, H. pylori hijacks multiple cellular signaling pathways, leading to the disruption of key cell functions. Two bacterial virulence factors play important roles in this scenario, the vacuolating cytotoxin VacA and the translocated effector protein CagA of the cag type IV secretion system (T4SS). Specifically, by overruling the phosphorylation status of cortactin, H. pylori alternates the activity of molecular interaction partners of this important protein, thereby manipulating the performance of cytoskeletal rearrangements, endosomal trafficking and cell movement. Based on shRNA knockdown and other studies, it was previously reported that VacA utilizes cortactin for its cellular uptake, intracellular travel and induction of apoptosis by a mitochondria-dependent mechanism, while CagA induces cell scattering, motility and elongation. To investigate the role of cortactin in these phenotypes in more detail, we produced a complete knockout mutant of cortactin in the gastric adenocarcinoma cell line AGS by CRISPR-Cas9. These cells were infected with H. pylori wild-type or various isogenic mutant strains. Unexpectedly, cortactin deficiency did not prevent the uptake and formation of VacA-dependent vacuoles, nor the induction of apoptosis by internalized VacA, while the induction of T4SS- and CagA-dependent AGS cell movement and elongation were strongly reduced. Thus, we provide evidence that cortactin is required for the function of internalized CagA, but not VacA.


Introduction
Helicobacter pylori represents one of the most successful pathogenic bacteria in humans; about half of the world's population is infected by this microbe [1]. Though in most cases patients remain asymptomatic, prolonged infection with the bacterium has been associated with various gastric diseases such as chronic inflammation of the stomach, peptic ulcers, or gastric adenocarcinoma in a subset of individuals [2][3][4][5]. Because of the capability to induce the development of stomach cancer, H. pylori has been classified as the first class I bacterial carcinogen in history [6]. Persistent colonization of the stomach is the main goal of H. pylori and proceeds by highly complex and multistep processes. To survive under the harsh conditions in the human stomach, the pathogen neutralizes the surrounding stomach acid to a survivable level by secreting the urease enzyme [7]. To find and penetrate the gastric mucus layer, H. pylori can sense differences in the pH gradient of the stomach [8,9]. Once H. pylori has reached the gastric epithelial cell surface, it employs a multitude of virulence factors, such as the outer membrane adhesins HopQ, BabA, or SabA to adhere to the cells [10][11][12], or the protease HtrA to disrupt tight and adherent junctions in the gastric epithelium, and gain access to the intercellular space via cleavage of E-cadherin, host cell [38]. In the present study, we aimed to take a closer look at the role of cortactin during infection with H. pylori, with specific interest regarding the functions of CagA and VacA. To investigate cortactin activity in more detail, we have recently established a stable cortactin knockout in the AGS gastric epithelial cell line using the CRISPR-Cas9 geneediting tool [49]. Deletion of the cttn gene in AGS cells disrupted H. pylori-induced signaling, in particular, by diminishing the activation levels of FAK, Src and Abl tyrosine kinases. Furthermore, cortactin was found to be vital in the CagA phosphorylation process [49]. For this purpose, we infected AGS wild-type (wt) and AGS∆cttn knockout cells with H. pylori wt and isogenic mutants. We show that cortactin deficiency did not negatively affect the uptake or intracellular function of VacA during infection with H. pylori, or by addition of the purified protein, but did diminish the expression of the CagA-dependent elongation phenotype in AGS cells. The VacA-dependent apoptosis and vacuolization rates during H. pylori infection or treatments with purified VacA were found to be at similar rates in both the wt and cortactin-deficient AGS cells. Taken together, the presented data shows that cortactin is rather involved in CagA-mediated AGS cell scattering signaling, while its role in VacA-mediated vacuolization and apoptosis seems to be dispensable.

Cortactin Knockout Is Associated with Diminished AGS Cell Scattering and Elongation Phenotype
Scattering and elongation, as well as endocytic travelling and apoptosis, are affected in AGS gastric epithelial cells infected by H. pylori [22,27,51]. To study whether cortactin deficiency might impact these AGS cell responses during H. pylori infection, we utilized the AGS∆cttn knockout cell line that we had previously created using the CRISPR-Cas9 approach [49]. We confirmed the knockout of cortactin in three AGS∆cttn cell clones, termed clone 1, 4 and 8 ( Figure 1A). We have chosen the well-described duodenal ulcer H. pylori strain P12, which has a functional T4SS and the active s1/m1 vacA allel [10,14,15,17,21,37,48,49], and produced isogenic knockouts in the cagA and vacA genes, as well as a structural T4SSinactive cagE mutant, which is unable to deliver CagA, with additional vacA mutation. The correct expression or absence of the resulting proteins were confirmed by Western blotting using H. pylori-specific antibodies ( Figure 1B). We then infected AGS wt and three AGS∆cttn clones with P12 wt or its isogenic P12∆cagA, P12∆vacA and P12∆cagE/∆vacA mutants. After 8 h of infection, we immunostained AGS cells with specific α-cortactin antibodies, counterstained filamentous actin with TRITC-conjugated phalloidin, and then analyzed the samples by fluorescence microscopy. Infection with P12 wt and P12∆vacA resulted in strong expression of the elongation phenotype in AGS wt cells as expected, while in all AGS∆cttn clones, the number of elongated cells was reduced ( Figure 2A). Almost no elongated cells could be found in the non-infected mock control or infections using the isogenic P12∆cagA or P12∆cagE/∆vacA mutants in either AGS wt or AGS∆cttn cells. Quantification analysis revealed that the expression of the elongation phenotype in AGS∆cttn cells, though not completely abolished, was significantly downregulated in infections with both P12 wt and P12∆vacA. P12 wt infection revealed a decrease from approximately 80 ± 5% in AGS wt by roughly half in AGS∆cttn cells (38 ± 4%, 33 ± 3%, and 44 ± 3% for the clones 1, 4, and 8, respectively), which held true for P12∆vacA (85 ± 3% in AGS wt to 33 ± 3%, 38 ± 4% and 42 ± 4% in the clones 1, 4, and 8, respectively) ( Figure 2B). Therefore, cortactin appears to be a major signal transducer of H. pylori-induced and CagA-dependent AGS cell scattering and elongation.  [52] and integrin-β1 expression, which is required for the AGS cell elongation phenotype [53], were not compromised in the AGSΔcttn knockout cells. (B) Confirmed knockout of cagA, vacA and cagE genes in the H. pylori mutant strains using specific antibodies against the corresponding proteins. The FlaA blot was used as a loading control. and integrin-β1 expression, which is required for the AGS cell elongation phenotype [53], were not compromised in the AGS∆cttn knockout cells. (B) Confirmed knockout of cagA, vacA and cagE genes in the H. pylori mutant strains using specific antibodies against the corresponding proteins. The FlaA blot was used as a loading control.

VacA-Dependent Vacuole Formation Is Not Hampered in Infected AGS∆cttn Cells
Besides contributing to CagA signaling, cortactin knockdown by shRNA has been previously reported to be implicated in VacA-dependent activities in AGS cells, when incubated with acid-activated VacA [38]. Therefore, we aimed to identify whether the VacA uptake and its intracellular signaling were affected by the complete lack of cortactin during infection with H. pylori. In this experiment, AGS wt exhibited substantial elongation phenotype upon infection with P12 wt and P12∆vacA, which was diminished in AGS∆cttn cells, as seen above. Surprisingly, only minor differences in VacA-dependent vacuole formation were observed in infections regardless of the chosen AGS cell line. Infection with P12 wt resulted in the mild formation of VacA-dependent vacuoles in both AGS wt and AGS∆cttn cells. In contrast, infection with P12∆cagA exhibited stronger vacuole formation, while infections using the P12∆vacA or P12∆cagE/∆vacA mutants displayed no vacuole formation at all, as expected ( Figure 3A, example vacuoles indicated by arrows). As assessed by phase contrast microscopy analysis of at least 100 randomly selected cells per experiment, infection with P12 wt resulted on average in the formation of 1.3 and 1.1 vacuoles per AGS wt and AGS∆cttn cell, respectively ( Figure 3B). In turn, when infected with the P12∆cagA strain, AGS wt and AGS∆cttn cells revealed on average 2.9 and 3.1 vacuoles, respectively. Therefore, different H. pylori mutants revealed varying vacuolating activities, which, however, were independent of the chosen cell line, AGS wt or AGS∆cttn, respectively.    Next, we wanted to analyze if dose-dependent properties could affect vacuole formation during infection. One might take into consideration that the saturation of host cells with VacA due to high bacterial MOIs could result in similar vacuolization quantities between the different AGS cell types. For this purpose, we first infected AGS wt and AGSΔcttn cells with the various H. pylori strains for 24 h at MOI of 50, followed by neutral red uptake assay and polarized light microscopy. Similar to our above results for shorter infection times (8 h) and MOI of 100, robust vacuole formation was observed in both AGS cell lines ( Figure 4A). To investigate this in more detail, we next infected AGS wt and AGSΔcttn cells with P12 wt in parallel at MOIs of 0, 5, 10, 25, 50, and 100. After 24 h, we performed the neutral red uptake assay and normalized the absorbance data to the number of adhered cells, since an increase in MOI normally led to increased cell lifting over time. However, both AGS wt and AGSΔcttn showed increased vacuolization rates with increasing MOIs, but no significant differences were seen between the two cell lines (Figure 4B). We further analyzed vacuole formation in our AGS cell lines upon H. pylori infection by a second approach, the neutral red uptake assay. Similarly, after 24 h of infection, AGS wt and AGS∆cttn cells showed a highly significant increase in vacuolization when infected with P12 wt, and especially P12∆cagA, but not by P12∆vacA or P12∆cagE/∆vacA ( Figure 3C). Thus, in concordance with the results obtained by phase contrast microscopy ( Figure 3B), the neutral red uptake assay also showed no significant difference in vacuolization between infected AGS wt and AGS∆cttn cells.
Next, we wanted to analyze if dose-dependent properties could affect vacuole formation during infection. One might take into consideration that the saturation of host cells with VacA due to high bacterial MOIs could result in similar vacuolization quantities between the different AGS cell types. For this purpose, we first infected AGS wt and AGS∆cttn cells with the various H. pylori strains for 24 h at MOI of 50, followed by neutral red uptake assay and polarized light microscopy. Similar to our above results for shorter infection times (8 h) and MOI of 100, robust vacuole formation was observed in both AGS cell lines ( Figure 4A). To investigate this in more detail, we next infected AGS wt and AGS∆cttn cells with P12 wt in parallel at MOIs of 0, 5, 10, 25, 50, and 100. After 24 h, we performed the neutral red uptake assay and normalized the absorbance data to the number of adhered cells, since an increase in MOI normally led to increased cell lifting over time. However, both AGS wt and AGS∆cttn showed increased vacuolization rates with increasing MOIs, but no significant differences were seen between the two cell lines ( Figure 4B).  or were left untreated. After 24 h infection, cells were subjected to neutral red uptake assay, followed by quantification of absorbance intensity. Mean values ± standard error are shown in graphs; n.s., non-significant.

Cortactin Deficiency Does Not Impact Induction of Apoptosis Provoked by H. pylori
Since we observed that cortactin apparently had no influence on VacA's ability to induce cellular vacuolization, we next wanted to take a closer look regarding its role in apoptosis. VacA localizes intracellularly to the membrane of mitochondria, where it induces the release of cytochrome-c into the cytoplasm and thereby induces an apoptotic cascade [33]. To study whether the absence of cortactin affects the induction of apoptosis in AGS cells by H. pylori, we infected the cells for 24 h. Addition of 200 μM of H2O2 to the

Cortactin Deficiency Does Not Impact Induction of Apoptosis Provoked by H. pylori
Since we observed that cortactin apparently had no influence on VacA's ability to induce cellular vacuolization, we next wanted to take a closer look regarding its role in apoptosis. VacA localizes intracellularly to the membrane of mitochondria, where it induces the release of cytochrome-c into the cytoplasm and thereby induces an apoptotic cascade [33]. To study whether the absence of cortactin affects the induction of apoptosis in AGS cells by H. pylori, we infected the cells for 24 h. Addition of 200 µM of H 2 O 2 to the cells served as a positive control to trigger apoptosis [54]. After infection, the AGS cell lines were stained with the apoptosis marker Annexin V tagged with fluorescein isothiocyanate (FITC), followed by fluorescence microscopy ( Figure 5). Quantification of these data is shown in Figure 6.
Interestingly, statistical analysis of the Annexin V staining revealed that the P12∆cagA mutant exhibited only a non-significant difference in apoptosis compared to P12 wt infections. In contrast, infection with the P12∆vacA mutant resulted in reduced apoptosis, as expected. Intriguingly, the double mutant P12∆cagE/∆vacA displayed a slight, but significant increase in apoptosis rate compared to P12∆vacA infection ( Figure 6). Thus, it appears that the lack of a functional T4SS and VacA in the P12∆cagE/∆vacA double mutant exhibit anti-apoptotic properties.
To corroborate the above findings further, we analyzed another apoptosis marker during infection. It was previously reported that H. pylori infection activates caspase-3 to trigger apoptosis [55]. Full-length caspase-3 is approximately 32 kDa in size and, upon activation, two auto-cleavage events release the active 17 kDa large subunit ( Figure 7A). Using two commercial α-caspase-3 antibodies recognizing each form, we found that fulllength caspase-3 is expressed in the uninfected mock controls of AGS wt and AGS∆cttn cells at similar level, while the active form is not, as expected. Upon infection with P12 wt or P12∆cagA, the intensity of full-length caspase-3 band decreased substantially, and the active subunit band appeared accordingly at the same time ( Figure 7A). Thus, these two strains exhibited the strongest capacity to induce apoptosis. In agreement with the Annexin V staining assay described above, the P12∆vacA and P12∆cagE/∆vacA mutants revealed significantly reduced band intensities for the 17 kDa active form and lower rates of apoptosis. Interestingly, no caspase-3 activation at all was observed in the H 2 O 2 -treated positive control, suggesting that H 2 O 2 might trigger a caspase-3-independent apoptosis pathway. All quantification data are presented in Figure 7B and show similar rates of apoptosis induction in both infected AGS wt and AGS∆cttn cells.
Next, we performed live cell imaging to investigate apoptosis events and the elongation phenotype in the infected AGS wt and AGS∆cttn mutant cell lines. As observed above, AGS∆cttn cells displayed less elongation than their wt counterparts ( Figure 8). The apoptosis occurred at roughly the same time post-infection (200 min in AGS wt versus 180 min in AGS∆cttn cells) and was associated with the occurrence of typical apoptotic bodies when the infected cells burst (Figure 8, arrows). Thus, the knockout of cortactin has no impact on VacA-induced apoptosis in H. pylori-infected AGS cells. Interestingly, statistical analysis of the Annexin V staining revealed that the P12ΔcagA mutant exhibited only a non-significant difference in apoptosis compared to P12 wt infections. In contrast, infection with the P12ΔvacA mutant resulted in reduced apoptosis, as expected. Intriguingly, the double mutant P12ΔcagE/ΔvacA displayed a slight, but significant increase in apoptosis rate compared to P12ΔvacA infection ( Figure  6). Thus, it appears that the lack of a functional T4SS and VacA in the P12ΔcagE/ΔvacA double mutant exhibit anti-apoptotic properties. To corroborate the above findings further, we analyzed another apoptosis marker during infection. It was previously reported that H. pylori infection activates caspase-3 to trigger apoptosis [55]. Full-length caspase-3 is approximately 32 kDa in size and, upon activation, two auto-cleavage events release the active 17 kDa large subunit ( Figure 7A). Using two commercial α-caspase-3 antibodies recognizing each form, we found that fulllength caspase-3 is expressed in the uninfected mock controls of AGS wt and AGSΔcttn cells at similar level, while the active form is not, as expected. Upon infection with P12 wt or P12ΔcagA, the intensity of full-length caspase-3 band decreased substantially, and the active subunit band appeared accordingly at the same time ( Figure 7A). Thus, these two strains exhibited the strongest capacity to induce apoptosis. In agreement with the Annexin V staining assay described above, the P12ΔvacA and P12ΔcagE/ΔvacA mutants revealed significantly reduced band intensities for the 17 kDa active form and lower rates of apoptosis. Interestingly, no caspase-3 activation at all was observed in the H2O2-treated positive control, suggesting that H2O2 might trigger a caspase-3-independent apoptosis pathway. All quantification data are presented in Figure 7B and show similar rates of apoptosis induction in both infected AGS wt and AGSΔcttn cells.  Figure 5 was assessed and the percentage of apoptotic cells was determined in three independent experiments. Mean values ± standard error are shown in the graph with significant differences corresponding to p ≤ 0.05 (*) and p ≤ 0.001 (***).

Vacuole Formation and Apoptosis by Purified VacA Is Not Affected in AGS∆cttn Cells
It was previously reported that treatment of AGS wt cells with purified acid-activated VacA alone induced apoptosis, while shRNA knockdown of cortactin significantly reduced the apoptosis rates [38]. Because our above results from infection experiments are in contrast to this study, we purified and acid-activated VacA, followed by treatment of AGS wt and AGS∆cttn cells using the same amounts under the same conditions, as previously described [38]. Our purified VacA fraction was checked by Coomassie brilliant blue staining and confirmed via Western blot using a specific VacA antibody ( Figure 9A,B). However, similar to the infections with live H. pylori, the formation of vacuoles was observed and quantified using the neutral red staining method. We observed a highly significant 2.8-fold increase of neutral red uptake in AGS wt cells and a 2.7-fold increase in AGS∆cttn cells. Statistical analysis confirmed that these differences between the cell lines are not significant ( Figure 9C). Next, we investigated the induction of apoptosis by purified acid-activated VacA. Similar to the vacuolization assay, we found no significant differences between the two cell lines in regard to the induction of apoptosis ( Figure 9D). Therefore, we confirmed that the effect of purified acid-activated VacA both on AGS wt and AGS∆cttn cells is not dependent on the expression of cortactin. Next, we performed live cell imaging to investigate apoptosis events and the elongation phenotype in the infected AGS wt and AGSΔcttn mutant cell lines. As observed above, AGSΔcttn cells displayed less elongation than their wt counterparts (Figure 8). The apoptosis occurred at roughly the same time post-infection (200 min in AGS wt versus 180 min in AGSΔcttn cells) and was associated with the occurrence of typical apoptotic bodies when the infected cells burst (Figure 8, arrows). Thus, the knockout of cortactin has no impact on VacA-induced apoptosis in H. pylori-infected AGS cells.

Vacuole Formation and Apoptosis by Purified VacA Is Not Affected in AGSΔcttn Cells
It was previously reported that treatment of AGS wt cells with purified acid-activated VacA alone induced apoptosis, while shRNA knockdown of cortactin significantly reduced the apoptosis rates [38]. Because our above results from infection experiments are in contrast to this study, we purified and acid-activated VacA, followed by treatment of AGS wt and AGSΔcttn cells using the same amounts under the same conditions, as previously described [38]. Our purified VacA fraction was checked by Coomassie brilliant blue staining and confirmed via Western blot using a specific VacA antibody ( Figure 9A,B). However, similar to the infections with live H. pylori, the formation of vacuoles was observed and quantified using the neutral red staining method. We observed a highly significant 2.8-fold increase of neutral red uptake in AGS wt cells and a 2.7-fold increase in AGSΔcttn cells. Statistical analysis confirmed that these differences between the cell lines are not significant ( Figure 9C). Next, we investigated the induction of apoptosis by purified acid-activated VacA. Similar to the vacuolization assay, we found no significant differences between the two cell lines in regard to the induction of apoptosis ( Figure 9D). Therefore, we confirmed that the effect of purified acid-activated VacA both on AGS wt and AGSΔcttn cells is not dependent on the expression of cortactin.

Discussion:
Cortactin, with its important function in the actin-cytoskeletal organization, has been identified as a prominent target of various pathogenic microbes such as Shigella flexneri [57], Cryptosporidium parvum [58], Helicobacter pylori [37], Trypanosoma cruzi [59] and others [42]. In fact, cortactin represents such an attractive target protein to hijack the actin cytoskeleton that it has therefore been called an "Achilles' heel" of the host cell [42]. Cortactin functions as an interaction partner of multiple signaling proteins via its SH3 domain, en-

Discussion
Cortactin, with its important function in the actin-cytoskeletal organization, has been identified as a prominent target of various pathogenic microbes such as Shigella flexneri [57], Cryptosporidium parvum [58], Helicobacter pylori [37], Trypanosoma cruzi [59] and others [42]. In fact, cortactin represents such an attractive target protein to hijack the actin cytoskeleton that it has therefore been called an "Achilles' heel" of the host cell [42]. Cortactin functions as an interaction partner of multiple signaling proteins via its SH3 domain, enabling it to influence various cellular properties such as cell-to-cell junctions [45], migration [60], or endocytosis [61]. Therefore, it comes as no surprise that H. pylori is also capable to manipulate the host cell via cortactin. Specifically, translocated CagA triggers the tyrosine dephosphorylation of cortactin via a yet unknown phosphatase [26,48] and induces the serine phosphorylation of cortactin [37] through activation of the ERK1/2 kinase signaling pathway [37,62]. Afterwards, serine-phosphorylated cortactin induces the auto-phosphorylation of FAK at tyrosine 397, which triggers a signaling pathway involving the activation of Src and Abl kinases, which results in the pronounced phosphorylation of injected CagA [49]. Furthermore, it was shown that H. pylori triggers cortactin overexpression at the protein level in a CagA-and JNK-dependent manner, which can be associated with gastric carcinogenesis [63]. However, other H. pylori virulence factors such as VacA have also been reported to target cortactin. Previously, Chang and co-workers [38] developed DNA constructs to either overexpress or knockdown cortactin in AGS cells. Based on such experiments, in combination with adding purified, acid-activated VacA, they reported that cortactin is involved in the intracellular trafficking and function of VacA [38]. However, similar studies using clean cortactin knockout cells are not yet available.
We performed H. pylori infection experiments using AGS wt and AGS∆cttn knockout cells under identical conditions and found that the characteristic phospho-CagA-dependent AGS elongation phenotype was strongly diminished in the cortactin knockout cells. This is in agreement with our previous infection studies, which showed a significant reduction of elongated cells by cortactin siRNA [37] and reduced CagA phosphorylation levels in cortactin-deficient AGS cells compared to control cells [49]. In contrast to the report by Chang and co-workers [38], we, surprisingly, could clearly detect VacA-dependent vacuole formation in AGS∆cttn cells at similar level compared to AGS wt cells, suggesting that the internalization and intracellular function of VacA was not hampered by the complete absence of cortactin. Thus, the intracellular transport of VacA by the endocytic machinery proceeds in a cortactin-independent fashion. In fact, cortactin was shown to be involved in clathrin-mediated endocytosis through direct interaction with the GTPase dynamin [61]. However, there is some evidence that VacA exploits the endocytic pathway independent of clathrin or dynamin [50]. Our observation that VacA's vacuolating function was not diminished encouraged us to investigate if cortactin deficiency might affect pro-or antiapoptotic properties of H. pylori. For this purpose, we infected AGS wt and AGS∆cttn cells with H. pylori wt and isogenic deletion mutants, namely ∆cagA, ∆vacA, and ∆cagE/∆vacA followed by staining with the apoptosis marker Annexin V. In our experiments, AGS cells with the cttn gene knockout demonstrated no statistically significant difference in H. pylori-induced apoptosis compared to AGS wt cells. In agreement with the previously reported apoptotic activity of VacA [33,34], we found that infection with the P12∆vacA mutant resulted in a significant reduction of the number of apoptotic AGS cells compared to infection with P12 wt. However, the P12∆cagE/∆vacA double mutant induced a slightly increased number of apoptotic cells compared to P12∆vacA, which is in agreement with CagA's proposed anti-apoptotic properties [26]. In fact, the cagE gene of H. pylori encodes one of the proteins involved in the structural formation of the T4SS, which injects CagA into the host cell [18]. However, we found only a non-significant difference in the number of apoptotic AGS cells upon infection with the P12∆cagA mutant compared to wt infections. These observations were further confirmed when we assessed the activation of executioner caspase-3. Since the double mutant P12∆cagE/∆vacA gene induced higher apoptosis rates of AGS cells compared to P12∆vacA, we assume that a T4SS effector other than CagA might be the primary anti-apoptotic determinant of H. pylori. Injected ADP-Heptose [64] is such a potential candidate. Together, it seems that other H. pylori factors might enhance or antagonize the pro-apoptotic function of VacA during infection. However, these results do not explain the differences in the study by Chang and co-workers [38]. We assume that the authors might have seen minor differences, which were over-interpreted in their discussion.

Cultivation of Eukaryotic Cells
The AGS cell line derived from the human gastric adenocarcinoma (ATCC CRL-1739) was routinely cultivated at 37 • C and 5% CO 2 in the RPMI 1640 medium (Gibco, Darmstadt, Germany) [65]. The growth media contained 10% fetal calf serum (FCS, Gibco), 1% penicillin/streptomycin (Sigma-Aldrich, Steinheim, Germany) and 0.2% normocin (InvivoGen, Toulouse, France). AGS∆cttn cells have been generated previously [49] and were grown under the same conditions as AGS wt, except for an additional 2 µg/mL of puromycin in the media. The cells were washed two times with PBS and given fresh media without antibiotics before H. pylori infection experiments.

Cultivation of H. pylori Strains
H. pylori strain P12 was grown from stock, stored at −80 • C in BHI media with 20% glycerol, onto GC-Agar plates containing 10% horse serum, 10 µg/mL vancomycin and 4 µg/mL amphotericin [66]. Microaerophilic conditions for bacterial growth were provided using CampyGen (Oxoid) gas packages in 2.5 L anaerobic jars (Oxoid, Wesel, Germany). H. pylori were grown for approximately 48 h at 37 • C, resuspended in BHI media, replated onto a fresh agar plate and incubated for additional 24 h at 37 • C under microaerophilic conditions before they were used for infection experiments. H. pylori deletion mutants P12∆cagA, P12∆vacA and P12∆cagE/∆vacA have been routinely generated by standard gene disruption procedures [67,68], and were grown the same way as the wt strain P12, except for adding 8 µg/mL kanamycin (P12∆cagA and P12∆cagE/∆vacA) or 4 µg/mL chloramphenicol (P12∆vacA) into the agar plates.

Infection of Eukaryotic Cells with H. pylori and Quantification of the Elongation Phenotype
AGS wt and AGS∆cttn cells were grown in 6-or 12-well plates as stated above. Cells were starved overnight in plain RPMI medium, before infection. H. pylori were cultivated as indicated above, resuspended in BHI media and added to the AGS cells at a calculated multiplicity of infection (MOI), as indicated in each figure legend. Infected cells were incubated at 37 • C and 5% CO 2 for 4 h to 24 h depending on the experiment. Quantification of the elongation phenotype was done as described in detail [51]. Briefly, cells were considered to exhibit the elongation phenotype if they displayed thin, needle-like protrusions of a length of 20 µm to 70 µm, as well as a general elongated shape. For quantification purposes, 100 randomly selected cells per experiment were counted and evaluated. The resulting data are from three independent experiments.

Light and Fluorescence Microscopy
AGS wt and AGS∆cttn cells were grown and infected as described above. After 8 h of infection, the cells were fixed using 4% PFA for 10 min at 20 • C. After fixation, the cells were permeabilized with 0.25% Triton-X-100 for 10 min and stained with mouse α-cortactin antibodies (Merck-Millipore, #05-180) followed by adding the secondary, FITC-conjugated goat-anti-mouse antibody. In addition, cells were counterstained with TRITC-conjugated phalloidin (Thermo Fisher Scientific, Darmstadt, Germany, #R415). Samples were analysed using a Leica DMI4000B fluorescence microscope (Leica Microsystems, Wetzlar, Germany). In order to analyze H. pylori-induced vacuoles, infected AGS cells (24 h) were stained with neutral red (see below) and imaged under light microscopy with polarizing contrast. Images were processed using the LAS AF computer software (Leica Microsystems).

Annexin V Apoptosis Assay
AGS wt and AGS∆cttn were grown to approximately 50% confluency and starved overnight in RPMI plain media before being infected with H. pylori. Infected cells were incubated for 24 h at 37 • C and 5% CO 2 . Following infection, the cells were washed two times with HEPES-buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl 2 .) and then incubated at 20 • C in the dark with 0.75 µg/mL Annexin V conjugated with FITC for 30 min. Afterwards, the cells were washed once with HEPES-buffer and were analyzed via fluorescence microscopy. The number of fluorescence positive/apoptotic cells was assessed and the percentage of apoptotic cells compared to the total cell number was determined. Two hundred µM H 2 O 2 was used as a positive control.

Quantification of Vacuolating Activity
AGS wt and AGS∆cttn cells were grown to 70% confluency in a 6-well plate and infected with H. pylori P12 wt, P12∆cagA, P12∆vacA or P12∆cagE/∆vacA and incubated for 8 h at 37 • C and 5% CO 2 . The samples were analyzed by phase contrast microscopy following the infection and the number of formed vacuoles were counted in 100 cells per experiment in triplicate.

Neutral Red Staining
AGS wt and AGS∆cttn cells were grown to approximately 70% confluency and then infected with H. pylori strain P12 wt and its isogenic deletion mutants as described above. The infected cells were incubated for 24 h at 37 • C and 5% CO 2 . Following the infections, vacuolization of AGS wt and AGS∆cttn cells was determined via staining of the vacuoles by neutral red (Nutritional Biochemicals Corporation, Cleveland, OH, USA, #12-935). For this purpose, a fresh solution of 0.5% neutral red in PBS with 0.3% BSA was prepared and filtered through a Whatman paper. Cells were washed once in PBS with BSA and then incubated with the neutral red solution for 5 min at 20 • C. Afterwards, the cells were washed again three times with PBS containing 0.3% BSA. Staining of vacuoles was confirmed by light microscopy with polarizing contrast. For quantitative analysis, neutral red was removed from the cells with acidified 70% ethanol and optical density (OD) was measured at 550 nm [73]. Results were calculated as fold change in infected cells compared to mock control. All measurements were performed in triplicate. Results were normalized to the number of adherent cells. The number of adherent cells was determined as the average number of cells in five different 1 mm 2 squares of each infected well.

Live Cell Imaging
To determine the time of morphological changes and apoptosis AGS wt and AGS∆cttn were grown in 6-well plates to 70% confluency and infected with P12 wt for 4 h at MOI of 50. Images were captured using the JuLI Live Cell Movie Analyzer (NanoEnTek Inc., Waltham, MA, USA) in intervals of 5 min.

In Vitro VacA Preparation and Acid-Activation
VacA was purified from H. pylori culture supernatants as described [73] and tested via SDS-PAGE, either stained with Coomassie brilliant blue R-250 (Bio-Rad Laboratories, Feldkirchen, Germany, #161-0400) [74] or Western blot with a specific VacA antibody (see above). Purified VacA protein was acid-activated as described in [38]. In brief, VacA was diluted in RPMI plain medium adjusted to pH = 2 by addition of HCl and NH4Cl and incubated at 37 • C for 15 min. Afterwards the pH was neutralized back to pH = 7 via addition of NaOH. AGS wt and AGS∆cttn cells were co-incubated with a final concentration of 5 ng/mL acid-activated VacA for 24 h. Following co-incubation, neutral red and Annexin V stainings were performed, to quantify vacuolization and apoptosis respectively, as described above.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author, N.T., upon reasonable request.