Pathogen-induced ubiquitin-editing enzyme A20 bifunctionally shuts off NF-κB and caspase-8-dependent apoptotic cell death

Abstract

The human pathogen Helicobacter pylori infects more than half of the world's population and is a paradigm for persistent yet asymptomatic infection but increases the risk for chronic gastritis and gastric adenocarcinoma. For successful colonization, H. pylori needs to subvert the host cell death response, which serves to confine pathogen infection by killing infected cells and preventing malignant transformation. Infection of gastric epithelial cells by H. pylori provokes direct and fast activation of the proinflammatory and survival factor NF-κB, which regulates target genes, such as CXCL8, BIRC3 and TNFAIP3. However, it is not known how H. pylori exploits NF-κB activation and suppresses the inflammatory response and host apoptotic cell death, in order to avert the innate immune response and avoid cell loss, and thereby enhance colonization to establish long-term infection. Here we assign for the first time that H. pylori and also Campylobacter jejuni-induced ubiquitin-editing enzyme A20 bifunctionally terminates NF-κB activity and negatively regulates apoptotic cell death. Mechanistically, we show that the deubiquitinylase activity of A20 counteracts cullin3-mediated K63-linked ubiquitinylation of procaspase-8, therefore restricting the activity of caspase-8. Interestingly, another inducible NF-κB target gene, the scaffold protein p62, ameliorates the interaction of A20 with procaspase-8. In conclusion, pathogen-induced de novo synthesis of A20 regulates the shut-off of the survival factor NF-κB but, on the other hand, also impedes caspase-8-dependent apoptotic cell death so as to promote the persistence of pathogens.

Figure 1

A20 upregulation after H. pylori-induced NF-κB activation in gastric epithelial cells. (a) AGS cells were infected with H. pylori for the indicated times. Cell lysates were subjected to IB for analysis of proteins involved in NF-κB activation. (b) Total RNA was isolated after H. pylori infection with and without IKK inhibitor. Changes in A20 transcript (TNFAIP3) expression were investigated by quantitative PCR. Data shown depict the average of triplicate determinations. (c) Cell lysates were harvested after H. pylori infection in the absence or presence of IKK inhibitor for the indicated times and analysed in IB. (d) Cells were infected with different isogenic H. pylori strains wt, virB7 or cagA followed by IB analysis. GAPDH served as a loading control. (e-i) The effect of A20 depletion by siRNA transfection (e) or CRISPR/Cas9-mediated knockout (f-i) on the activity of NF-κB upon H. pylori infection (e, f), C. jejuni infection (g) or TNF stimulation (h) was analysed by IB (e-h) or transactivation assay (i). Representative IBs from at least two independent experiments with similar outcomes are shown. The transactivation assay was performed at least five times. Error bars, S.D., *P<0.02, #P<0.05. (j) Wt and A20^KO_1 cells were infected with H. pylori for 3 h. Total RNA was isolated and changes in the level of NF-κB target genes CXCL8 and BIRC3 were determined by quantitative PCR. Data shown depict the average of triplicate determinations. (k) A20^KO_1 cells were transfected with either pCMV (as empty vector control) or full-length active A20 (pA20) plasmids 24 h prior to H. pylori infection for the times shown. Cell lysates were harvested and analysed by IB

Figure 2

A20 impedes apoptotic cell death in pathogen-infected cells. (a) AGS cells were infected with H. pylori or treated with different stimuli for the times indicated followed by IB analysis of procaspase-8 and -3 cleavage. (b) Procaspase-3 activation was determined by imaging flow cytometry of immunostained cleaved caspase-3 (n=3 for H. pylori infection). Treatment with staurosporine (ST, final concentration 1 μM) served as a positive control. (c) MKN-45 cells were infected with H. pylori and analysed by Annexin V/PI staining using imaging flow cytometry. Error bar, S.D.; *P<0.005; **P<0.0005. (d) Wt and A20^KO_1 cells were infected with H. pylori for the indicated time points before analysis by IB. (e) A20^KO_1 cells were transfected with either pCMV (as empty vector control) or full-length active A20 (pA20) plasmids 24 h prior to H. pylori infection for a further 24 h. Cell lysates were harvested and analysed by IB. Wt and A20^KO_1 cells were pretreated with Z-IETD-FMK for 15 min followed by H. pylori infection for 24 h before analysis by IB (f) or Annexin V/PI staining using imaging flow cytometry (g). Representative images and dot plots for the cellular staining are shown. Percentage of apoptotic cells represents the sum of annexin V-positive and double positive cells. Error bar, S.D., *P<0.005; **P<0.0005. The concentrations of Z-IETD-FMK used were 10 μM and 20 μM for panel (f) and 20 μM for panel (g). (h) Wt and A20^KO_1 cells were infected with C. jejuni followed by IB analysis of procaspase-8 and -3 cleavage. Representative IBs from two independent experiments are shown

Figure 3

The DUB activity of A20 limits procaspase-8 activation by removing K63-linked polyubiquitin upon H. pylori infection. (a and b) AGS cells were infected with H. pylori and a caspase-8 co-IP was performed (a) or in vitro translated FLAG-A20 and HA-caspase-8 were co-incubated followed by IPs with the respective anti-FLAG or anti-HA antibody (b) and analysed by IB. In IgG, an isotype-matched antibody was used. (c–e) Wt and A20^KO_1 cells (c) or wt cells treated with 10 μM Z-IETD-FMK (d) or A20^KO_1 cells complemented with pCMV (empty vector control), pC103A (A20 DUB mutant), pC624/627A (A20 E3 ligase mutant) or pA20 (full-length active A20) by transfection (e) were infected with H. pylori and caspase-8 IP under denaturing conditions was performed and analysed in IB for endogenous K63-linked polyubiquitin (c–e). Representative IBs from two independent experiments are shown

Figure 4

Depletion of Cul3 affects the apoptotic cell death of cells infected with H. pylori. (a) AGS cells were infected with H. pylori and subjected to a Cul3 co-IP followed by IB analysis of interacting proteins. (b) Cul3-depleted cells by siRNA transfection were infected with H. pylori. Caspase-8 IP under denaturing conditions was performed and the K63-linked ubiquitinylation status of procaspase-8 was analysed in IB. Representative IBs from two experiments are shown. (c) The impact of Cul3 depletion on the percentage of apoptotic cells was evaluated using Annexin V/PI staining followed by imaging flow cytometry. Representative images and dot plots for the cellular staining are shown. Percentage of apoptotic cells represents the sum of annexin V-positive and double positive cells. Error bar, S.D., *P<0.01; **P<0.005

Figure 5

p62 enhances interaction between A20 and procaspase-8 for effective editing of K63-linked polyubiquitinylation of procaspase-8. AGS cells were infected with H. pylori followed by IB analysis of p62 expression (a) or co-IP with anti-A20 antibody was performed and analysed by IB for interacting proteins (b). (c) p62-depleted cells by siRNA transfection were infected with H. pylori. Procaspase-8 co-IP was performed and analysed by IB for interacting proteins. (d) Caspase-8-depleted cells by siRNA transfection were infected with H. pylori. A20 co-IP was performed and analysed by IB for interacting proteins. (e and f) The effect of p62 depletion by siRNA transfection on procaspase-8 processing (e) and K63-linked ubiquitinylation of procaspase-8 after caspase-8 IP under denaturing conditions (f) was analysed by IB. Representative IBs from two independent experiments are shown. (g) Schematic representation of the findings in this study. Infection with H. pylori leads to the activation of NF-κB (1), which in turn upregulates A20 (2). A20 fulfills two opposing functions, one that inhibits further activation of NF-κB (3), thus providing a negative feedback loop and the other, supported by p62, removes K63-linked polyubiquitin from procaspase-8 (4) to negatively regulate the apoptotic response of the host cell