Helicobacter pylori modulates host cell responses by CagT4SS-dependent translocation of an intermediate metabolite of LPS inner core heptose biosynthesis

Abstract

Highly virulent Helicobacter pylori cause proinflammatory signaling inducing the transcriptional activation and secretion of cytokines such as IL-8 in epithelial cells. Responsible in part for this signaling is the cag pathogenicity island (cagPAI) that codetermines the risk for pathological sequelae of an H. pylori infection such as gastric cancer. The Cag type IV secretion system (CagT4SS), encoded on the cagPAI, can translocate various molecules into cells, the effector protein CagA, peptidoglycan metabolites and DNA. Although these transported molecules are known to contribute to cellular responses to some extent, a major part of the cagPAI-induced signaling leading to IL-8 secretion remains unexplained. We report here that biosynthesis of heptose-1,7-bisphosphate (HBP), an important intermediate metabolite of LPS inner heptose core, contributes in a major way to the H. pylori cagPAI-dependent induction of proinflammatory signaling and IL-8 secretion in human epithelial cells. Mutants defective in the genes required for synthesis of HBP exhibited a more than 95% reduction of IL-8 induction and impaired CagT4SS-dependent cellular signaling. The loss of HBP biosynthesis did not abolish the ability to translocate CagA. The human cellular adaptor TIFA, which was described before to mediate HBP-dependent activity in other Gram-negative bacteria, was crucial in the cagPAI- and HBP pathway-induced responses by H. pylori in different cell types. The active metabolite was present in H. pylori lysates but not enriched in bacterial supernatants. These novel results advance our mechanistic understanding of H. pylori cagPAI-dependent signaling mediated by intracellular pattern recognition receptors. They will also allow to better dissect immunomodulatory activities by H. pylori and to improve the possibilities of intervention in cagPAI- and inflammation-driven cancerogenesis.

Fig 1. The H. pylori hldE (core LPS heptose biosynthesis) gene cluster, its proposed biosynthetic pathway, and strain characterization of H. pylori hldE gene cluster mutants.

A) Heptose phosphate biosynthesis gene cluster in H. pylori. Genes HP0861 through HP0857 of reference strain 26695 probably form an operon (see also [110]). B) Biosynthesis pathway, LPS intermediates, metabolites and contributing enzymes of the core heptose pathway predicted in H. pylori. In red color, the biosynthesis functions of the bifunctional hldE gene (HP0858) in analogy to its E. coli ortholog are depicted within the pathway. The metabolite HBP, produced through the action of HldE enzyme, is highlighted by a red square. To the right, a schematic structure of H. pylori LPS is shown, indicating its main structural building blocks. The dashed line indicates the potential intermediate LPS structure in the absence of the GmhA/HldE pathway, which consists solely of the lipid A and the keto-deoxy-octonate (KDO) substructures. Gene designations refer to gene names in strain 26695. For references regarding LPS biosynthesis see main text. C) Growth characteristics of H. pylori strain N6 core heptose phosphate biosynthesis mutants (designated by strain 26695 gene numbers) in comparison to the N6 reference strain. Growth curves of single strains were performed in liquid culture for up to 48 h. D), E) Quantification of IL-8 secretion of AGS cells coincubated for 4 h with live H. pylori strains. Hp N6 = strain N6 shown in panels D, F, G; Hp P12 = strain P12 shown in panels E,F,G. F) silver-stained SDS gel of treated lysate (ETL) preparations (Methods) from H. pylori and its core heptose biosynthesis mutants. Detected glycans include H. pylori LPS chains as annotated on the right: IC (inner core of LPS), LC (longer LPS chains, including outer core and O-antigen chains) [16] G) Determination of cell adherence of H. pylori wild type strains N6 and P12 and their isogenic core heptose biosynthesis mutants to AGS cells (Methods). Strain designations in all figure panels refer to gene names in the gmhA/hldE cluster of strain 26695 as indicated in panel B. comp. = complemented. Statistically significant differences between strains during growth in liquid culture in panel C were calculated by two-way ANOVA, followed by Tukey’s multiple comparison test: **p<0.01, ****p<0.0001, they are depicted next to the graphs for the last time point (48 h); see also S1 Table for full results. Statistical significances in D), E), G) were determined by two-tailed, non-paired Student’s t-test; **p<0.01, ***p<0.001, ****p<0.0001, ns = non-significant.

Fig 2. H. pylori IL-8 activation is abolished in AGS TIFA k/o cells, while H. pylori CagA translocation is independent of TIFA expression, IL-8 induction, and core heptose biosynthesis pathway intermediates.

A) qRT-PCR detection of TIFA transcript in AGS wt and AGS TIFA k/o (KO) cells. Cells were mock-incubated or cocultured for 2 h with H. pylori strain N6 wild type (wt), isogenic cagY mutant (HP0527) and heptose pathway hldE (HP0858) or rfaD (HP0859) mutants as indicated. The TIFA transcript amounts are given in % of the mock-coincubated AGS parental cell, which was set to 100%. B) IL-8 induction in AGS CRISPR-Cas9 TIFA knock-out (KO) cells. Indicated H. pylori strains (N6 wt and isogenic core heptose mutants) were coincubated for 4 h at an MOI of 25 bacteria per cell with AGS parental cells, TIFA k/o cell single clone and TIFA k/o cell pool. HP0858 comp. is the complemented strain. Significance of differences (p) between N6 wt coincubated wild type and k/o cells were calculated by Student‘s t-test. C) CagA translocation by H. pylori N6 and isogenic cagY (HP0527; CagT4SS functional negative control) and LPS core heptose biosynthesis mutants in AGS cells and AGS TIFA KO cells. AGS cells were coincubated with H. pylori strain N6 wt or isogenic mutant bacteria for 4 h. 20 μg soluble protein (cleared cell lysates) per lane was separated on an SDS gel and blotted to nitrocellulose membrane. Western blots were incubated with antibodies as indicated. HP signifies the antiserum detection of heat-stable H. pylori surface antigens (Dako Cytomation, for antibodies see S6 Table) and was used as a universal control for amounts of invariable H. pylori proteins in the preparations. Actin detection was used as a loading control for amounts of AGS cell proteins. D) densitometric quantitation of CagA and p-CagA (CagA translocation) of Western blot results shown in panel C (see Methods). The intensity values were normalized to human actin and to HP invariable protein for each condition and are depicted in % of the positive control (AGS cells coincubated with H. pylori N6 wild type bacteria), which was set to 100%.

Fig 3. H. pylori core heptose mutants are able to induce the hummingbird phenotype in human gastric AGS cells (wild type and CRISPR/Cas9 TIFA k/o) independently of IL-8 and TIFA.

A) mock-coincubated AGS wild type (wt) cells; B) AGS wt cells coincubated with strain N6 wt bacteria; C) AGS wt cells coincubated with strain N6 HP0858 (hldE) mutant bacteria; D) AGS wt cells coincubated with N6 HP0858 complemented strain; E) mock-coincubated AGS TIFA CRISPR/Cas9 k/o cells; F) AGS TIFA k/o cells coincubated with N6 wt. Bacteria were coincubated with the cells for 4 h after centrifugation, fixed and microscopic images acquired. Black arrowheads designate some hummingbird phenotype cells for clarity. Size bars (white) represent 50 μm. G), H) Quantitation of hummingbird phenotype in bacteria-coincubated AGS wt cells (G) or AGS TIFA k/o cells (H) as shown in panels A) through F). 1,000 cells for each condition were quantitated for cell length using ImageJ (as detailed in Methods). Mutants (of strain N6) coincubated with the cells shown in panels G) and H) are indicated by the respective gene numbers. Statistically significant differences between mock-coincubated and bacteria-coincubated cells are shown above the graphs (*p<0.05, **p<0.01, **** p< 0.001, calculated by non-parametric Kruskal-Wallis test; ns = non-significant).

Fig 4. Results of coincubation of parental, TIFA k/o and TIFA-complemented cell lines with H. pylori and its core heptose LPS biosynthesis mutants and detection of downstream signaling.

A) AGS B) HEK293T parental and CRISPR-Cas9 TIFA k/o cells (pool) were transiently transfected with either an empty vector or a vector expressing human TIFA, by lipofectamine 2000 (HEK) or nucleofection (AGS). On the next day, parental, TIFA k/o and TIFA-complemented cells were coincubated with H. pylori of indicated genotypes (mutants indicated by respective gene numbers) at an MOI of 25 for 4 h. Cell supernatants were analyzed for IL-8 secretion by ELISA. Statistical significance of differences was determined using two-tailed, non-paired Student's t-test (**p<0.01, ***p<0.001, ****p<0.0001). PanelsC) and D) show qRT-PCR data, for il-8 (C) and cxcl-1 (D) transcript amounts for a panel of bacteria-coincubated AGS parental (wt) and TIFA k/o cells (H. pylori strain N6 and isogenic mutants). HP0527 = cagY inactivation mutant of strain N6, which was used as a control condition of a cagPAI functional-negative mutant in cell coincubations. Panel E) shows a Western blot to detect downstream signaling in AGS and HEK293T cells upon H. pylori coculture, dependent on HP0858 (HldE) activity. Cells were coincubated with H. pylori N6 and isogenic mutants for 4 h. Equal amounts of cellular proteins (20 μg) were separated on SDS gels, blotted and probed with antibodies as indicated to detect the cagPAI-dependent downstream activation of p38, TAK1 and IκBa (as reported in [50]). Actin was detected as a loading control. LPS core heptose biosynthesis mutants in all panels are designated with gene names according to the nomenclature of strain 26695 as outlined in Fig 1. F), G) results of RT² Profiler RT-PCR arrays in 96 well format (Innate and Adaptive Immune Responses transcript panel selection; Qiagen.com). PCR arrays were performed on cDNA preparations of AGS cells (parental and TIFA knock-out (KO), coincubated with H. pylori N6 and its isogenic HP0858 mutant. Yellow dots represent over-expressed genes for condition on y-axis. Blue dots represent under-expressed genes for condition on y-axis. Panel F) shows the pairwise comparison of amounts of arrayed transcripts between AGS parental cells, coincubated with H. pylori N6 wild type (wt), versus the same cells coincubated with N6 HP0858 mutant. Panel G) depicts the pairwise comparison of transcript amounts between AGS parental cells coincubated with N6 wt and AGS TIFA k/o cells coincubated with the same strain (see also S8 Fig for extended results; full results are summarized in S2 Table).

Fig 5. H. pylori soluble non-proteinaceous metabolite of the HldE-dependent heptose phosphate reaction activates cells when experimentally transfected.

A), B), C) HEK293T luciferase reporter cell transfection with treated lysate ETL preparations of H. pylori strains reveal HldE-dependent activation in a transient transfection setting using NF-κB luciferase plasmid (pNFκB-luc). A) transiently-transfected HEK293T luciferase reporter cells were coincubated for 3 h with live H. pylori of either wild type strain N6, its isogenic heptose core pathway mutants or the cagPAI function-negative mutant HP0527, as indicated, or parental strain 88–3887, its isogenic cagPAI deletion mutant (dcagPAI), or cagA mutant, respectively. B) transiently transfected HEK293T luciferase reporter cells were coincubated with ETL preparations of H. pylori strains, heptose pathway mutants, or cagPAI mutants added to the cell medium in the absence of transfection agent for 3 h (see silver-stained gel of ETL preparations in Fig 1); C) in the third assay set-up, transiently transfected HEK293T cells were super-transfected with ETL preparations of H. pylori strains and their isogenic mutants. In B) and C, ultrapure H. pylori N6 wt LPS was also coincubated or transfected as an additional control. Cells in A), B), C) were incubated prior to the luciferase measurement for 3 h. All assay conditions were measured in triplicates. Two-tailed, non-paired Student’s t-test indicates significant differences of *p<0.05, **p<0.01, ***p<0.001, ns = not significant. D) time-dependent activation of HEK-Blue Null1 SEAP reporter cells (Invivogen) after transfection with treated lysate preparations (ETL) from H. pylori N6 and its isogenic core heptose biosynthesis mutants. Cell activation by release of secreted alkaline phosphatase into the medium was monitored over a time course of 25 hours post transfection using HEK-Blue real time detection medium (Methods). All conditions were assayed in triplicates. Statistical significance of differences in D) was calculated by Student’s t-test for comparisons between mock and all other conditions (blue symbols) and HPN6 wt and all other conditions (red symbols); ns = non-significant; **p<0.01; ***p<0.001; ****p<0.0001.