Disruption of gastric mucous granule exocytosis by Helicobacter pylori virulence factor CagA

Helicobacter pylori infection is the strongest known risk factor of stomach cancer. Strains harboring the virulence factor CagA (cytotoxin-associated gene A) signicantly stimulate host inammatory response, which increases the risk of ulceration and cancer. However, the mechanisms by which CagA triggers prolonged inammation with mucosal damage remain elusive. Based on a large-scale genetic screen using Drosophila, we identied a novel CagA target Synaptotagmin-like protein 2-a, Slp2-a, an effector of small GTPase Rab27. Using gastric organoid-derived monolayers of polarized mucous cells, we demonstrated that CagA inhibited Slp2-a-mediated docking of mucous granules to the plasma membrane by direct binding to Slp2-a. We further observed aberrant cytoplasmic retention of mucus in human gastric mucosa infected with CagA-expressing strains. These results suggest that CagA could be disrupting the protective mucous barrier by inhibiting Slp2-a-mediated mucous granule exocytosis, which may lead to mucosal damage from luminal acid and pepsin to promote inammation leading to cancer. (Dox [+] CagA WT + GFP, 132 granules in 5 cells from 5 organoids; Dox [+] CagA WT + Slp2-a C2AC2B, 100 granules in 5 cells from 5 organoids). These results indicate that exogenous Slp2-a C2AC2B binds and traps CagA, which liberates endogenous Slp2-a from CagA and eventually allows Slp2-a-associated mucous granules to dock to the apical plasma membrane. CagA therefore inhibits docking of mucous granules to the plasma membrane and disrupts their exocytosis by a direct binding to Slp2-a. mucus retention in the gastric surface mucous cells of CagA-positive H. pylori-infected patients. remains elusive 3 . In this study, we tried an alternative in vivo approach in which we carried out a large-scale genetic screen using a transgenic Drosophila model expressing CagA in the eyes 20 , and discovered a novel CagA target Slp2-a. We demonstrated that CagA disrupted mucous granule exocytosis by inhibiting functions of Slp2-a, using gastric organoid-derived monolayers of polarized mucous cells, an in vitro model that mimics gastric mucosa in vivo 32 . We further conrmed excessive mucus retention in gastric biopsy specimens from CagA-positive H. pylori-infected patients. osmium tetroxide cacodylate dehydrated a of embedded Epon acetate and lead citrate, Japan). ImageJ software Bethesda, MD, USA).

Introduction secretion 24,25 . Secondly, we found that C2AC2B was required and su cient for the interaction with CagA (Fig. 2, b and c). Thirdly, we generated a series of deletion mutants of CagA lacking the EPIYA-repeat region (EPIYA), carboxy-terminal region (C-ter) and CagA-multimerization sequence (CM) within the EPIYArepeat region (ΔEPIYA, ΔC-ter and ΔCM, respectively) and found that ΔEPIYA and ΔCM did not bind to Slp2-a ( Fig. 2d and Supplementary Figs. 1 and 2) 5 . Since ΔCM may disrupt the overall structure of EPIYA ( Supplementary Fig. 2), we next examined the interaction between Slp2-a and CagA ABD, which possesses a single CM in EPIYA, and con rmed that deletion of CM also abolished their interaction (ΔCM, Fig. 2, e and f) 5 . We further found that Leucine 971 (L971) in CM was an indispensable residue for the binding (L971G, substitution of Glycine for L971, Fig. 2e and Supplementary Fig. 3). These results indicate that CagA interacts with Slp2-a C2AC2B via its CM and L971 in CM is critical for the interaction (Fig. 2, c and f).
CagA inhibits interaction between Slp2-a and phospholipids in the inner lea et of the plasma membrane.
The above results prompted us to investigate whether CagA inhibits the interaction between Slp2-a C2AC2B and membrane phospholipids. We rst con rmed a direct interaction of recombinant hexa-histidine-tagged T7-Slp2-a C2AC2B (His-T7-Slp2-a C2AC2B) with PS immobilized on a nitrocellulose membrane ( Fig. 2, g and h, and Supplementary Fig. 4) 26 . We found that the interaction was strongly inhibited by preincubation of His-T7-Slp2-a C2AC2B with a recombinant wild type CagA ABD fused with glutathione-S-transferase (GST-CagA WT), whereas preincubation with GST, GST-CagA ΔCM, or GST-CagA L971G barely affected the interaction (Fig. 2, g and h). These results suggest that CagA inhibits interaction between Slp2-a C2AC2B and phospholipids, which in turn may block the Slp-2-a-mediated docking of mucous granules to the plasma membrane.
CagA inhibits Slp2-a-mediated docking of mucous granules to the apical plasma membrane and disrupts exocytosis in the gastric surface mucous cells.
To examine the effects of CagA on gastric mucous granule exocytosis, we generated mouse gastric organoids predominantly composed of polarized surface mucous cells in which secretory granules containing a gel-forming mucin, MUC5AC, accumulate beneath the apical plasma membrane ( Supplementary   Fig. 5) 24,27,28 . The organoids carrying tetracycline-regulated V5-CagA ABD expression constructs were selected and its expression was induced with doxycycline (Dox) (Supplementary Fig. 6). We observed excessive mucus accumulation beneath the apical plasma membrane (arrowheads) in the CagA WTexpressing cells (Fig. 3b), whereas CagA ΔCM-and CagA L971G-expressing cells appeared the same as the control Dox (-) cells ( Fig. 3, a, c and d).
CagA therefore inhibits docking of mucous granules to the plasma membrane and disrupts their exocytosis by a direct binding to Slp2-a.
Excessive mucus retention in the gastric surface mucous cells of CagA-positive H. pylori-infected patients.
Finally, we investigated whether mucus retention in the surface mucous cells could be detected in gastric biopsy specimens taken endoscopically from H. pylori-infected patients (14 cases of CagA-positive H. pylori-infected and 12 cases of CagA-negative H. pylori infected) 29,30 . To observe mucus retention, gastric biopsy specimens which do not show marked atrophy (atrophy score 3) or intestinal metaplasia (intestinal metaplasia score equal to or greater than 1) were selected (see Supplementary Table 1 for further information on histological evaluation according to the updated Sydney system) 31 . As shown in Fig. 4 and Supplementary Table 1, surface mucous cells in the majority of CagA-positive H. pylori-infected patients had excessive mucus accumulation (13 in 14 cases), while mucus was observed only in the apical surface of those cells from CagA-negative H. pylori-infected patients (12 in 12 cases). This observation is consistent with our ndings obtained by using gastric organoids shown in Fig. 3. These results suggest that CagA disrupts the mucous barrier by inhibiting Slp2-a-mediated mucous granule exocytosis, which may promote mucosal damage and in ammation.

Discussion
More than twenty binding partners of CagA have been identi ed in cell-based biochemical studies 5,6 , however, in vivo signi cance of the interaction still remains elusive 3 . In this study, we tried an alternative in vivo approach in which we carried out a large-scale genetic screen using a transgenic Drosophila model expressing CagA in the eyes 20 , and discovered a novel CagA target Slp2-a. We demonstrated that CagA disrupted mucous granule exocytosis by inhibiting functions of Slp2-a, using gastric organoid-derived monolayers of polarized mucous cells, an in vitro model that mimics gastric mucosa in vivo 32 . We further con rmed excessive mucus retention in gastric biopsy specimens from CagA-positive H. pylori-infected patients.
Prolonged tissue injury-induced in ammation is widely accepted as a hallmark that promotes cancer development and progression 33,34 . Cell death leads to the release of damage-associated molecular patterns (DAMPs), such as HMGB1 (high-mobility group box-1) and ATP (adenosine triphosphate), and activates immune cells to secrete in ammatory cytokines that promote survival and proliferation of neighboring cells, initiating tumorigenesis. Our nding that CagApositive H. pylori disrupts mucous granule exocytosis, which may lead to barrier dysfunction and subsequent mucosal injury, provides a novel mechanism through which CagA could be promoting in ammation and increases the risk of gastric cancer.
It has been reported that CagA disrupts tight junctions and causes loss of epithelial cell polarity with reduced transepithelial electrical resistance (TEER) in some cells, such as polarized MDCK (Madin-Darby canine kidney) cells 11,13 . Indeed, when we expressed CagA in gastric organoid-derived polarized surface mucous cells, we also detected mislocalization of a tight junction marker ZO-1, however the effect appeared to be much milder compared with that shown in previous reports ( Supplementary Fig. 8). In consistent with our results, Uotani et al. recently demonstrated that H. pylori-infected polarized epithelial cells of human gastroid monolayers largely maintained their normal tight and adherens junctions and TEER, although TEER transiently declined 17 . Furthermore, mucous granule accumulation only occurred beneath the apical plasma membrane and basolateral mis-sorting was not observed in the CagA-expressing cells (Figs. 3, 4, and Supplementary Fig. 9), suggesting that their cell polarity remained mostly intact. We speculate that some compensatory mechanisms may exist in the cells we used and the defect in granule docking in the CagA-expressing cells was caused by the speci c inhibition of the Slp2-a C2AC2B function rather than by disruption of cellular polarity.
Finally, our nding that the exogenously expressed C2AC2B domain of Slp2-a restores the defect in mucous granule exocytosis shown in Fig. 3 is of great interest. Targeting protein-protein interactions with small molecules has been thought to be di cult, since proteins generally interact with their partner proteins using relatively large and at contact surfaces. However, recent progress in structure-based drug design and screening technology has allowed us to make protein-protein interactions potential therapeutic targets 35,36 . Given that the increasing rate of antibiotic resistance in H. pylori is a clinical challenge in the treatment of the infection in many countries 37,38 , small molecules inhibiting the interaction between CagA and Slp2-a may have potential as a novel drug candidate against mucosal damage and in ammation caused by CagA-positive H. pylori infection.
The surfaces of y eyes were observed under a Hitachi S-4800 scanning electron microscope.
Total RNA Isolation and real-time quantitative RT-PCR.
Total RNA from 50 adult Drosophila heads from a GMR-Gal4/+; +/P{GS}17878 line or an yw line as a wild type control was puri ed with NucleoSpin RNA (Macherey-Nagel, Düren, Germany). cDNA was synthesized using a Cloned AMV First-Strand cDNA Synthesis Kit (Invitrogen, Carlsbad, CA, USA). PCR was performed on a LightCycler 96 system (Roche, Mannheim, Germany) using LightCycler FastStart DNA Master SYBR Green I reaction mix (Roche) according to manufacturer's instructions. The btsz primers were designed to amplify the common coding sequences in the two btsz transcripts, btsz-2 (isoform C) and btsz-3 (isoform D) (Fig. 1c). The rpL32 gene was chosen as an endogenous control and its mRNA quanti cation was used for normalization. Experiments were performed in triplicate. The primers used for quantitative RT-PCR were as follows.
Plasmids and Mutagenesis.
To remove the unique BglII site (AGATCT) in the coding region of Slp2-a, a silent mutation was introduced using the KOD-Plus Mutagenesis Kit (TOYOBO, Osaka, Japan). The sequence of primers used for mutagenesis were as follows.
Using pEF-T7-Slp2-a (BglII-mutated) as a template, a Slp2-a fragment, Slp2-a ΔSHD (amino acids 82-950) was ampli ed by PCR using the following primers and cloned into the pEF-T7 vector with BamH I and Sal I sites. The resultant plasmid was designated pEF-T7-Slp2-a ΔSHD. The open reading frame of ABD-type East Asian CagA derived from H. pylori strain TN2 (GenBank accession No. LC007103) was ampli ed by PCR using following primers and cloned into an expression vector pEF6/V5-His B with Spe I and Eco RI sites. The resultant plasmid was designated pEF6/V5-CagA ABD WT (wild type).

GST-CagA ABD C-ter (reverse), 5'-ACTCTCGAGCTATTTCTGGAAACCACTTTT-3'
To make the hexa-histidine (6 x His)-tagged T7-Slp2-a C2AC2B (amino acids 641-902) protein ( Supplementary Fig. 4a), a DNA fragment coding for Slp2-a C2AC2B was ampli ed by PCR using following primers and cloned into pET-28b with BamH I and Eco RI sites. DNA fragments coding for V5-tagged CagA ABD-WT and -L971G were excised from pEF6/V5-His B CagA ABD WT and L971G respectively by digestion at the BamH I and Pme I sites, and cloned into pPB-TetON-FLIG-N with BamH I and SnaB I sites. The resultant plasmids were designated pPB-TetON-V5-CagA ABD WT and L971G. pPB-TetON-V5-CagA ABD ΔCM was made by inverse PCR using the primers CagA ABD ΔCM-A and -B as described above. The pPB-TetON-V5-CagA ABD WT, ΔCM and L971G were used for the Dox-dependent expression of V5-CagA ABDs in gastric organoids.
The Slp2-a fragment, Slp2-a C2AC2B (amino acids 638-905) was generated by PCR using the following primers and cloned into the D-T7-pRc/CMV expression vector with Not I and Apa I sites.

Isolation of fundic glands.
All experiments using mice were approved by the Oita University Animal Ethics Committee and performed according to the Committee's guideline. The stomach extracted from BALB/c mice (6 weeks old) was opened along the greater curvature, washed with ice-cold PBS. The fundus was isolated, cut in small fragments (< 5 mm pieces) and placed into a 15 mL conical tube with ice-cold PBS. After washing with ice-cold PBS until the supernatant was clear, the fragments were treated with TrypLE Express (Gibco) at 37 o C for 30 min with shaking. After removal of the TrypLE Express, the fragments were suspended in dissociation buffer (1% D-solbitol and 1.5% Sucrose in PBS) and the tube was orientated perpendicular to the ground and shaken in hand for 2 min at 2 cycles per sec. After washing with ice-cold PBS, the fragments were resuspended in 10 mL of ice-cold PBS and passed through a 70 µm Falcon cell strainer to remove cell debris and isolate fundic glands.
Gene transfer into fundic glands, generation of fundic organoids predominantly composed of polarized surface mucous cells and induction of transgene expression.
The samples were immersed in QY-1 (n-butyl glycidyl ether) (Nisshin EM, Tokyo, Japan), embedded in Epon 812, and sectioned using an ultramicrotome. The sections were stained with uranyl acetate and lead citrate, and then examined with a transmission electron microscope HITACHI H-7650 (Hitachi, Tokyo, Japan). Morphometric analyses were performed using ImageJ software (NIH, Bethesda, MD, USA).
Gastric biopsy specimens and ethical approval.
To compare mucus accumulation in gastric surface mucous cells in CagA-positive H. pylori-infected patients with that in CagA-negative H. pylori-infected patients, gastric biopsy specimens obtained from Bangladesh and Thailand were selected, since the frequencies of CagA-negative H. pylori-infected patients found in these countries were much higher than those in East Asian countries such as Japan where almost all of H. pylori isolated from patients were CagApositive 29,30 . Biopsy materials were xed with 10% neutral buffered formalin, embedded in a para n block, and sectioned for histological analysis. Gastric mucosa was stained with hematoxylin and eosin (H&E) and May-Giemsa, and degree of monocyte in ltration, neutrophil in ltration, atrophy, intestinal metaplasia and bacterial density were pathologically classi ed into four grades according to the updated Sydney system 31 . The sections were also subjected to the immunochemical analyses using anti-H. pylori rabbit polyclonal IgG (B0471; DAKO Japan, Tokyo, Japan) and anti-CagA rabbit polyclonal IgG (b-300; Santa Cruz). Genotyping of H. pylori isolated from biopsy specimens was performed as described 29,30 . Brie y, the CagA genotype (EPIYA-repeat region) was determined by PCR-based direct sequencing. The study protocols above were approved by the Ethics Committee of Bangladesh Medical Research Council (Dhaka, Bangladesh), the Ethics and Research Committee of Chulalongkorn University Faculty of Medicine (Bangkok, Thailand), and the Ethics Committee of Oita University Faculty of Medicine.
Immuno uorescence staining of gastric biopsy specimens.
To observe mucus retention, gastric biopsy specimens which do not show marked atrophy (atrophy score 3) or intestinal metaplasia (intestinal metaplasia score equal to or greater than 1) were selected (CagA-positive H. pylori-infected 7, CagA-negative H. pylori-infected 7 from Bangladesh, CagA-positive H. pyloriinfected 7, CagA-negative H. pylori-infected 5 from Thailand). The selected biopsy sections were treated for antigen retrieval in a citrate buffer (pH 6.0), permeabilized with 0.1% Triton X-100 in PBS, and blocked with Block Ace, followed by incubation with anti-Mucin 5AC mouse monoclonal IgG (1:500) (45M1; Abcam) and anti-H. pylori rabbit polyclonal IgG (1:200) (B0471; DAKO Japan). Alexa Fluor 488-labelled goat anti-rabbit IgG (1:1,000) (A-11034; Invitrogen) and Alexa Fluor 546-labelled goat anti-mouse IgG (1:1,000) (A-11030; Invitrogen) were used for double staining. Mucus accumulation in the surface mucous cells was classi ed into two categories, negative (-), 'low levels of mucus can be observed in the apical surface of the cells'; positive (+), 'high levels of mucus accumulation can be observed in the cytoplasm of the cells'. The evaluation was independently conducted by four researchers including an expert pathologist (T. U.) in a blinded manner.

Declarations
Data availability: The data that support the ndings of this study are available from the corresponding author on reasonable request.