Creation and Initial Characterization of Isogenic Helicobacter pylori CagA EPIYA Variants Reveals Differential Activation of Host Cell Signaling Pathways

The polymorphic CagA toxin is associated with Helicobacter pylori-induced disease. Previous data generated using non-isogenic strains and transfection models suggest that variation surrounding the C-terminal Glu-Pro-Ile-Tyr-Ala (EPIYA) motifs as well as the number of EPIYA motifs influence disease outcome. To investigate potential CagA-mediated effects on host cell signaling, we constructed and characterized a large panel of isogenic H. pylori strains that differ primarily in the CagA EPIYA region. The number of EPIYA-C motifs or the presence of an EPIYA-D motif impacted early changes in host cell elongation; however, the degree of elongation was comparable across all strains at later time points. In contrast, the strain carrying the EPIYA-D motif induced more IL-8 secretion than any other EPIYA type, and a single EPIYA-C motif induced comparable IL-8 secretion as isolates carrying multiple EPIYA-C alleles. Similar levels of ERK1/2 activation were induced by all strains carrying a functional CagA allele. Together, our data suggest that polymorphism in the CagA C-terminus is responsible for differential alterations in some, but not all, host cell signaling pathways. Notably, our results differ from non-isogenic strain studies, thus highlighting the importance of using isogenic strains to study the role of CagA toxin polymorphism in gastric cancer development.

inoculum, two replicates of the cells were lysed, serially diluted, and plated for CFU. (c) Bacterial internalization was monitored by treating the remaining two replicates of cells with 200 g/mL gentamicin (gent) for 3 hrs, lysis, serially dilution, and plating. The percent of adherent H. pylori was determined relative to the starting inoculum. The percent of internalized H. pylori was determined relative to the number of adherent H. pylori. There were no statistically significant differences between strains in adherence to (P=0.9932) and internalization into AGS (P=0.5306) as analyzed by an ordinary one-way ANOVA. The data are presented as the mean + the SEM and represent three independent experiments. Together the data show that CagA EPIYA polymorphism does not alter bacterial growth or adherence to or internalization into host cells. (d). To characterize CagA translocation and phosphorylation following strain construction, AGS cells were infected for 8 hrs at an MOI of 100 with WT G27 or with the ABCC Restorant, EPIYA, and cagA control strains. Lysates were analyzed for total CagA (CagA) and phosphorylated CagA (pTyr). The Western blot images were cropped to show only the region corresponding to CagA (top) and phosphorylated CagA (bottom). As expected, no CagA was detected in the uninfected cells or the cagA control infected lysates. No phosphorylated band for CagA was detected in the EPIYA or cagA infected lysates. 8 1.5 IQR are shown as individual data points. Differences in host cell elongation between strains were determined by an ordinary one-way ANOVA (see Supplementary Table S2). (b) The relative length/breadth ratios were then grouped (0<4, 4<8, ≥8) and the percent of elongated cells was determined to better illustrate changes in elongation across strains. AGS cell elongation was significantly increased for infected cells compared to the uninfected controls, and for cells infected with an intact EPIYA region (G27 and ABCC Restorant) as compared to cagA, EPIYA, and cagA::cat.
There were no appreciable differences in elongation between the cagA, EPIYA, and cagA::cat strains or between the G27 and the ABCC Restorant strains. The data represent three independent experiments. Figure S4. IL-8 secretion is infection and EPIYA dependent. AGS cells were infected with the isogenic control strains at an MOI of 10. A sample of the co-culture supernatant was taken at 12, 24, and 36 hrs post-infection and analyzed by ELISA for total IL-8. At all time points, there was a significant increase in IL-8 secretion in cells infected with H. pylori as compared to uninfected control cells (Groups B and C vs. A; P<0.0001, two-way ANOVA). There was a significant increase in IL-8 secretion from cells infected with H. pylori that expressed CagA with an intact EPIYA motif (Group C) as compared to strains missing or containing a truncated CagA (Group B; P<0.05, two-way ANOVA). There were no differences in IL-8 secretion between cells infected with cagA, EPIYA, and cagA::cat or between WT G27 and the ABCC Restorant. ELISA data are presented as the geometric mean + 95% confidence interval and represent three independent experiments. P values were adjusted for multiple comparisons using the Tukey's multiple comparison test. .

Isogenic strain construction.
G27 EPIYA. The G27 EPIYA strain was constructed so that the C-terminal EPIYA region of the G27 wild type (WT) strain was replaced with a counter-selectable kanamycin resistance determinant aphA3/sacB (kan-sacB) 2,3 , which encodes kanamycin resistance and sucrose sensitivity.
This strain served as the parental strain for the CagA isogenic strain construction. Briefly, the Nterminal and C-terminal regions flanking the H. pylori 7.13 cagA EPIYA motifs were PCR amplified using 7.13 genomic DNA (gDNA) as a template and the 7.13EPIYA-5'-fp/7.13EPIYA-Mrp and 7.13EPIYA-Mfp/7.13EPIYA-3'-rp primers, respectively. The 7.13EPIYA-Mrp and 7.13EPIYA-Mfp primers were engineered to contain XhoI and SmaI restriction sites. The resulting PCR fragments were fused by SOE PCR using 7.13EPIYA-5'-fp/7.13EPIYA-3'-rp. The SOE product was cloned into pGEM-T Easy to yield pDSM530. EcoR1, XhoI, and SmaI digestion and sequencing with T7/SP6 primers were used to verify proper insertion of the SOE product into the vector. pDSM530 and pKSF-II were next individually digested with XhoI and SmaI and the liberated kan-sacB cassette from pKSF-II was ligated into pDSM530 to yield pDSM531. E. coli transformants were selected for by growth on kanamycin and the resulting plasmid was verified by EcoR1 and Sma1 digestion. H. pylori G27 was then naturally transformed with pDSM531 and transformants were selected for on kanamycin. Proper double homologous recombination and insertion of the kan-sacB cassette into the G27 genome was verified by PCR with the G27EPIYA-5'-fp/7.13EPIYA-Mrp and SacBSCN-F2/Grace2 primer sets. The resulting G27 EPIYA strain was archived as DSM713.  4 . The PCR reactions were as follows: 1). 7.13 gDNA was amplified with 7.13EPIYA-5'-fp/K3-7IS-M1rp for the upstream flanking region; 2) K3 gDNA was amplified with K3-7IS-M1fp/K3-7IS-M2rp for the EPIYA motifs; and 3) 7.13 gDNA was amplified with K3-7IS-M2fp/7.13EPIYA-3'-rp for the downstream flanking region. As previously described, the products were joined by SOE PCR such that the first and second PCR products were joined with 7.13EPIYA-5'fp/K3-7IS-M2rp; this product was then joined with the third PCR product using the 7.13EPIYA-5'fp/7.13EPIYA-3'-rp primer pair. The ABD PCR product was cloned into pGEM-T EASY to yield pDSM533; proper insertion and sequencing with T7/SP6 were performed as described above. In the second phase of construction, site directed mutagenesis was performed on pDSM533 in a series of three PCR reactions to change the alanine in the EPIYA-B motif to a threonine to yield an EPIYT motif (denoted as-B T ). The PCR reactions using pDSM533 were as follows: 1) 7.13EPIYA-5'-fp and K3(A/T)-rp; and 2) K3(A/T)-fp and 7.13EPIYA-3'-rp. Next, these PCR products were purified, mixed and joined by SOE PCR with the 7.13EPIYA-5'-fp/7.13EPIYA-3'-rp primer pair. The SOE product was cloned into pGEM-T Easy to yield pDSM547 and the site directed mutagenesis was verified by sequencing with T7/SP6 primers. G27 EPIYA was naturally transformed with pDSM547 and proper insertion of the EPIYA-AB T D construct was verified by sucrose resistance and kanamycin sensitivity. G27CagAfp and G27CagArp primers, and the subsequent fragment was cloned into pGEM-T Easy (pDSM1389). Proper insertion of the SOE product was verified with EcoRI, XhoI, and SmaI digestion and by sequencing with the T7/SP6 primers. As described for the EPIYA strain, the pDSM1389 and pKSF-II plasmids were digested with SmaI and XhoI in a sequential reaction prior to ligation of pDSM1389 with the liberated kan-sacB cassette. pDSM1390 was then naturally transformed and integrated into G27 by double homologous recombination. Insertion of the kan-sacB cassette into the cagA locus was verified by kanamycin resistance and sucrose sensitivity and by PCR using the G27CagAfp/G27CagArp, SacBSCN-F2/Grace1, and SacBSCN-F2/Grace2 primer pairs for the proper sized products. The resulting strain was archived as DSM1391.

G27 ABCC Restorant.
As a control for genetic manipulation of the strains, the G27 EPIYA strain was restored to the WT ABCC EPIYA sequence. The ABCC EPIYA region was PCR amplified from G27 gDNA with G27EPIYA-F/G27EPIYA-R. The resulting PCR product was cloned into pGEM-T Easy to yield pDSM1419; proper sequence was verified by using the T7/SP6 sequencing primers in combination with primers that aligned to various regions of the cagA EPIYA sequence (CagA2864R, CagA3149F, CagA3493R, and CagA3576F). The G27 EPIYA strain was then transformed with pDSM1419. Transformants were screened for by loss of kanamycin resistance and sucrose sensitivity.
Proper insertion of the EPIYA region in the transformants was verified by PCR amplification of the region using the G27EPIYASeq-F/Grace2 primers to screen for the correct sized product. The PCR product was then sequenced with the G27EPIYASeqF, CagA3149F, G27EPIYA-R, and Grace1 primers to confirm replacement of the kan-sacB cassette with the WT ABCC EPIYA sequence. The resulting strain was archived as DSM1420.  Differences in host cell elongation for each time point were evaluated using an ordinary one-way ANOVA; *P values were adjusted for multiple comparisons using Tukey's multiple comparison test.
Shading represents non-significant P values.