IpaH7.8 Activity Contributes to Bacterial Multiplication.

Because inflammatory cell death in response to bacterial infection is an important part of the host immune defense mechanism (1⇓–3, 26), we wished to investigate whether the secretion of IpaH7.8 during Shigella infection is beneficial to the bacteria. To this end, we infected BALB/c mice with Shigella via the intranasal route, which is a well-established model for evaluating Shigella pathogenesis (41). BALB/c mice were infected intranasally with a sublethal dose of WT Shigella, an IpaH7.8 mutant (ΔipaH7.8), the IpaH7.8-complemented strains ΔipaH7.8/ipaH7.8 or ΔipaH7.8/ipaH7.8CA (in which the Cys residue 357 was replaced with Ala) which lack putative E3 ligase activity, or the T3SS-deficient mutant strain S325. Lung bacterial load, inflammation, and IL-1β production were examined 24 h postinfection (Fig. 1). Infection with either ΔipaH7.8 or ΔipaH7.8/ipaH7.8CA produced lower levels of IL-1β in lung tissue than did infection with either WT or ΔipaH7.8/ipaH7.8 Shigella (Fig. 1A). Although all strains induced an inflammatory response in the lung, the responses to WT and ΔipaH7.8/ipaH7.8 infections were much stronger than the responses to ΔipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 infections (Fig. 1 A–C). Notably, the lung bacterial loads were significantly higher in mice infected with WT or ΔipaH7.8/ipaH7.8 than in mice infected with ΔipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 Shigella (Fig. 1D). Together, these results strongly suggested that IpaH7.8 E3 ligase-dependent intracellular activity benefits the pathogen by promoting infection.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

IpaH7.8 activity contributes to bacterial proliferation. IpaH7.8 augments IL-1β production and bacterial multiplication in a murine pneumonia model. Five-week-old female BALB/c mice were inoculated intranasally with the indicated Shigella strains (5 × 10⁷ cfu). The lungs were removed 24 h postinoculation and assayed for cytokine production (A), lung weight (B), H&E staining (C), and the number of viable bacteria (D). (Magnification: 200×.) n = 20; *P < 0.05. The error bars represent the SD of the measurements.

Shigella IpaH7.8-Mediated Cell Death Is Macrophage Specific.

We infected BALB/c bone marrow-derived macrophages (BMDMs) with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 bacteria and assessed the ability of each bacterial strain to kill macrophages using the LDH assay. We confirmed that Shigella induce rapid macrophage cell swelling and lysis within 2 h through a T3SS-dependent mechanism (19, 39, 40) and determined that cell death occurred in an IpaH7.8-dependent manner (Fig. 2A). Under conditions in which there was a uniform number of bacteria in the cell cytosol (Fig. S1A), IpaH7.8 activity was confined almost exclusively to peritoneal macrophages and to the macrophage cell lines J774A.1 and RAW264.7 (Fig. 2A) but was not observed in HeLa cells, 293T cells, Jurkat T cells, or Raji B cells, suggesting that IpaH7.8-dependent cell death is specific to macrophages. That observation was consistent with previous macrophage studies by Fernandez-Prada et al. (39) and Paetzol et al. (40). Because other Shigella IpaH E3 ligase effectors, e.g., IpaH9.8 and IpaH4.5, had no appreciable ability to induce rapid macrophage cell killing (Fig. S1B), and because the observed IpaH7.8 activity was dependent on its E3 ligase activity (Fig. 2A and Fig. S1B), we concluded that IpaH7.8 is a key player in eliciting macrophage cell death during Shigella infection.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Shigella-induced IpaH7.8-dependent cell death occurs via caspase-1 activation and IL-1β/IL-18 maturation. (A) BMDMs and peritoneal macrophages were isolated from BALB/c mice. Macrophages or the indicated cell lines were infected with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 Shigella (MOI of 10), and LDH was measured in the culture supernatants to determine bacteria-induced cytotoxicity. *P < 0.01 versus LDH values in cells infected with WT Shigella. (B) The active form of caspase-1 was detected by immunoblotting. BMDMs were infected with the indicated Shigella strains at an MOI of 10 or with Salmonella strains at an MOI of 1 at 37 °C for 1 h. Infected cells and all proteins in the medium were collected and subjected to immunoblotting using the anti-caspase1 P20 antibody. (C) Cytokine production (IL-1β, IL-18, IL-6, and MIP-2) was examined in BMDMs infected with the indicated Shigella strains for 2 h. *P < 0.01. (D) BMDMs were isolated from WT or caspase-1–KO mice and differentiated into macrophages. Cells were infected with Shigella (MOI of 10) for 4 h. LDH was measured in the culture supernatants from 30 min to 4 h post Shigella infection. *P < 0.01 versus LDH values in cells infected with WT Shigella. (Also see Fig. S2.) (E) HMDMs were infected with the indicated Shigella strains (MOI of 10) at 37 °C for 30 min and then were examined for LDH release and cytokine production (IL-1β and IL-8). *P < 0.01. The error bars represent the SD of the measurements. (Also see Fig. S1.)

IpaH7.8-Dependent Cell Death Occurs Via Caspase-1 Activation.

We determined that Shigella infection of macrophages caused rapid cell death in an IpaH7.8 E3 ligase-dependent manner. Previous studies showed that Shigella infection of BMDMs causes rapid cell death via inflammasome-mediated caspase-1 activation (19). Therefore, we examined the cell death of BMDMs treated with z-VAD (a pan-caspase inhibitor), z-YVAD (a caspase-1 inhibitor), or z-DEVD (a caspase-3 inhibitor) upon infection with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, or S325 Shigella for up to 3 h. IpaH7.8-dependent cell death decreased with z-VAD and z-YVAD treatment but not with z-DEVD treatment (Fig. S2A), indicating that IpaH7.8-dependent macrophage cell death upon Shigella infection occurred mainly through caspase-1 activation. To confirm these findings, we examined the effects of IpaH7.8 on caspase-1 activation and on IL-1β, IL-6, IL-18, and MIP-2 production by infecting BMDMs with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 Shigella for 1 h. Caspase-1 processing and production of IL-1β and IL-18, but not of IL-6 and IL-8/MIP-2, occurred in an IpaH7.8 E3 ligase-dependent manner (Fig. 2 B and C). To confirm further that caspase-1 activation was involved in this process, BMDMs isolated from caspase-1–KO or WT BALB/c mice were stimulated with LPS and subsequently infected with the various Shigella strains. As shown in Fig. 2D, IpaH7.8 E3 ligase-dependent BMDM cell death required capase-1 activity. Of note, the IpaH7.8-dependent death of BMDMs was markedly prevented by glycine treatment, which can block pyroptosis (Fig. S2B). The effects of IpaH7.8 on IL-1β and IL-18 production also were confirmed in Shigella-infected human monocyte-derived macrophages (HMDMs) (Fig. 2E). Together, these results strongly suggested that pyroptotic cell death occurs in an IpaH7.8 E3 ligase-dependent manner during Shigella infection.

IpaH7.8 E3 Ligase-Dependent Cell Death Occurs Through the Activation of the NLRP3 and NLRC4 Inflammasomes.

It was shown previously that Shigella infection of macrophages stimulates the NLRP3 and NLRC4 inflammasomes and IL-1β production (8, 19). Therefore we examined IpaH7.8 E3 ligase-dependent cell death upon Shigella infection using BMDMs from NLRC4-KO mice on a BALB/c background. Levels of LDH (Fig. S3B) and the active form of caspase-1 (Fig. 2B and Fig. S3C) were lower in BMDMs from NLRC4-KO mice infected with Shigella WT (or with ΔipaH7.8/ipaH7.8), but not with ΔipaH7.8 (ΔipaH7.8/ipaH7.8CA and S325), than in BMDMs from WT mice (Fig. S3 A–C). Because NLRC4-KO mice are limited to the BALB/c background, we examined the effects of shRNA-mediated knockdown of NLRP3, NLRC4, or NLR family pyrin domain-containing 1b (NLRP1b) on IpaH7.8-dependent cell death during Shigella infection. As shown in Fig. 3A and Fig. S3D, knockdown of NLRP3 or NLRC4, but not of NLRP1b, reduced IL-1β production. Furthermore, when NLRP3 was knocked down in BMDMs from NLRC4-KO mice, the IL-1β levels stimulated by IpaH7.8 were decreased further (Fig. 3A and Fig. S3D). We subsequently examined whether IpaH7.8 E3 ligase activity is required to stimulate inflammasome and caspase-1 activation using HEK293T cells transfected with combinations of T7-NLRP3 (or HA-NLRC4), T7-ASC, HA-caspase-1, FLAG-IL-1β, and GFP-IpaH7.8 (or GFP-IpaH7.8CA). The results showed that IpaH7.8, but not IpaH7.8CA, could stimulate both inflammasomes and induce mature IL-1β production in an E3 ligase activity-dependent manner (Fig. 3B). Consistent with these results, macrophages infected with Shigella producing IpaH7.8CA had a reduced ability to induce mature IL-1β production and cell death (Figs. 1A and 2). The diminished effects of IpaH7.8CA-expressing Shigella also were confirmed in a mouse model (Fig. 1). Because recent studies have shown that inflammasome-dependent caspase-1 activation can be inhibited by treating cells with proteasome inhibitors (42⇓–44), we assessed IpaH7.8-dependent macrophage cell death in macrophages treated or not treated with MG132, a ubiquitin-proteasome inhibitor. As shown in Fig. S3E, the IpaH7.8-dependent cell-death response was abolished almost completely in MG132-treated cells (Fig. S3E). Based on these results, we concluded that IpaH7.8 E3 ligase activity is a critical factor in inducing macrophage cell death during Shigella infection.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

IpaH7.8 E3 ubiquitin ligase-dependent cell death occurs via activation of the NLRP3 and NLRC4 inflammasomes. IpaH7.8 induces NLRP3 and NLRC4 activation followed by IL-1β maturation in macrophages. (A) shRNA-mediated knockdown of NLRP3, NLRC4, and NLRP1b. BMDMs from BALB/c WT mice or BALB/c NLRC4-KO mice were infected with pLKO.1 lentivirus expressing shRNA directed against NLRP3, NLRC4, or NLRP1b. A lentivirus carrying the empty pLKO.1 vector was used as a negative control. NLRP3/NLRC4/NLRP1b silencing was assessed (Fig. S3D); then the cells were infected with the indicated Shigella strains (MOI of 10) for 2 h and examined for IL-1β production. *P < 0.05. The error bars represent the SD of the measurements. (B) Reconstitution of the NLRP3 and NLRC4 inflammasomes in 293T cells. The results are representative of three independent experiments. (Also see Fig. S3.)

GLMN Is an IpaH7.8 Target.

These results strongly suggested that Shigella-induced macrophage cell death depends on IpaH7.8 E3 ligase activity. Thus, we hypothesized that the IpaH7.8 E3 ligase targets some host factor(s) that undergoes IpaH7.8 E3 ligase-mediated ubiquitination and participates in the inflammasome–caspase-1 cell-death pathway. To identify these potential targets, we screened for host proteins that interacted with IpaH7.8 (the IpaH7.8 CA mutant was used to minimize targeting of host protein degradation) in a yeast two-hybrid assay using a mouse brain library and identified two potential hits, GLMN (Fig. S4 A and B) and filamin A. Because siRNA-mediated knockdown of filamin A had no effect on macrophage cell death in response to Shigella infection, we investigated GLMN as an IpaH7.8-binding partner and found that GLMN interacted predominantly with the LRR domain of IpaH7.8 (Fig. 4A). GLMN originally was identified as a gene responsible for familial glomuvenous malformations, a disorder that results in localized cutaneous vascular lesions because of loss-of-function mutations in the Glmn gene (45⇓⇓–48). More recently, GLMN was characterized as a Cullin ring ligase inhibitor that blocks interaction with its ubiquitin-conjugating enzyme (E2) (49, 50).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

GLMN is an IpaH7.8 target. (A, Upper) Schematic representation of the Shigella IpaH7.8 derivatives, including IpaH7.8, IpaH7.8CA, IpaH7.8LRR, and IpaH7.8CTR. (Lower) The GST pull-down assay shows that the LRR region of IpaH7.8 binds to GLMN. (B, Upper) Colocalization of IpaH7.8 and GLMN. J774A.1 cells were nucleofected with T7-tagged IpaH7.8 or the empty vector as a control. The cells were fixed, permeabilized, and probed with a rabbit polyclonal anti-GLMN antibody and a mouse monoclonal anti-T7 antibody at 2.5 h posttransfection. The localization of GLMN and IpaH7.8 in macrophages was analyzed by confocal microscopy. (Scale bars, 10 μm.) (Lower) The graph indicates intensity of the fluorescence signals along the arrow. (C, Upper) Cycloheximide chase assay. (Lower) The remaining endogenous GLMN was quantified. (D) IpaH7.8-mediated GLMN degradation is prevented by treatment with a ubiquitin-proteasome inhibitor. BMDMs were treated for 1 h with the ubiquitin-proteasome inhibitor MG132 (40 μM) or with DMSO as a control and then were infected with the indicated Shigella strains. Cells were harvested at 1 h postinfection, and the amount of endogenous GLMN was assessed by immunoblotting. (E) Ubiquitination of GLMN. Ubiquitination reactions were performed in the presence of purified ubiquitin, ATP, E1, UbcH5B, GLMN, and IpaH7.8 or IpaH7.8CA and then were subjected to immunoblot analysis with anti-GLMN and anti-IpaH antibodies. (Also see Fig. S4.) (F and G) J774A.1 cells were infected with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, ipaH7.8CA, or S325 Shigella for 75 min or were treated with LPS and ATP. The cells were fixed and probed with an anti-GLMN antibody. (F) Imaging analysis indicates that the cytoplasmic GLMN signal decreased as GLMN formed puncta in an IpaH7.8-dependent manner. The arrowheads indicate bacteria. (Scale bar, 5 μm.) (G) The number of cells expressing dispersed GLMN (GLMN⁺ cells) was determined. *P < 0.001. The error bars represent the SD of the measurements.

First, we confirmed the interaction between GLMN and IpaH7.8 (and IpaH7.8CA) in a GST pull-down assay (Fig. 4A). To examine this interaction further, we created a series of truncated GLMN mutants and examined their capacity to interact with IpaH7.8 using a GST-IpaH7.8 pull-down assay. As shown in Fig. S4 C and D, ∼55 amino acids in the GLMN N-terminal domain mediated the interaction with IpaH7.8. Immunohistochemical analyses of macrophage cells expressing T7-IpaH7.8 revealed that IpaH7.8 colocalized with GLMN in the cytoplasm (Fig. 4B). Because IpaH7.8 E3 ligase activity is involved in inflammasome activation during Shigella infection, we investigated whether GLMN underwent IpaH7.8 E3 ligase-dependent degradation. To this end, HEK293T cells were transfected with GFP-IpaH7.8 (or GFP-IpaH7.8CA or GFP mock) and then examined the levels of endogenous GLMN. The GLMN levels were lower in cells transfected with full-length IpaH7.8 than in cells transfected with IpaH7.8CA or the GFP mock control (Fig. S4E). A cycloheximide chase assay using HEK293T cells that were transfected with GFP-IpaH7.8 (or GFP-IpaH7.8CA) confirmed that endogenous GLMN levels decreased in the presence of IpaH7.8, but not IpaH7.8CA, although IpaH7.8 was expressed at lower levels than IpaH7.8CA because of its inherent E3 ligase activity (Fig. 4C). In addition, as shown in Fig. 4D, the decrease in endogenous GLMN in Shigella-infected BMDMs was prevented by MG132 treatment, which blocks proteasome activity. Furthermore, using an in vitro E3 ligase activity assay with purified E1, E2 (UbcH5B), E3 (IpaH7.8 or IpaH7.8CA), and GLMN, we demonstrated that IpaH7.8, but not IpaH7.8CA, ubiquitinates GLMN and IpaH7.8 itself (Fig. 4E). We further examined the cellular distribution of GLMN in macrophages infected with WT, ΔipaH7.8, ΔipaH7.8/ipaH7.8, ΔipaH7.8/ipaH7.8CA, or S325 Shigella. When cells were infected with these bacterial strains, the cytoplasmic GLMN signal decreased as GLMN formed puncta in an IpaH7.8-dependent manner (Fig. 4 F and G). These results strongly suggested that GLMN undergoes proteasome-dependent degradation in an IpaH7.8 E3 ubiquitin ligase-dependent manner.

GLMN Is Functionally Involved in Inflammasome Activation.

To ensure the functional involvement of GLMN in IpaH7.8-mediated inflammasome activation, we investigated the effects of GLMN on NLRP3 inflammasome activation in macrophages treated with LPS and ATP, well-known NLRP3 agonists (1, 2, 7). As shown in Fig. 5A and Fig. S5, in quiescent cells, endogenous GLMN was dispersed throughout the macrophage cytoplasm and nuclei. Upon stimulation of the inflammasome with LPS/ATP, GLMN relocalized into punctate structures (Fig. 5A and Fig. S5). Importantly, Shigella infection also induced the formation of GLMN-containing puncta in an IpaH7.8 E3 ligase-dependent manner (Fig. 4 F and G). Under these conditions, GLMN puncta (red in Fig. 5A) occasionally colocalized with the active form of caspase-1 (green in Fig. 5A). To determine whether GLMN levels could affect IL-1β production, siRNA-mediated knockdown of GLMN was performed in HEK293T cells transfected with T7-NLRP3 (or HA-NLRC4), T7-ASC, HA-caspase-1, or FLAG-IL-1β for 48 h. As shown in Fig. 5B, IL-1β levels were higher in cells treated with anti-GLMN siRNA than in cells treated with the control (Fig. 5B). Together, these data suggested that GLMN puncta associate with the inflammasome. To test this notion, we investigated the effects of GLMN overexpression on inflammasome activation, cell death, and IL-1β maturation. Upon GLMN overexpression, the degree of macrophage cell death after 2 h of LPS/ATP treatment was less than half of that detected in the control, as determined by the LDH assay (Fig. 5C). Similarly, the levels of mature IL-1β in GLMN-overexpressing cells after 1.5 or 2 h of LPS/ATP treatment were ∼40% lower than in the control (Fig. 5C and Fig. S6A). Consistent with these results, GLMN showed striking cytoprotective and anti-inflammatory effects when macrophages overexpressing GLMN were infected with WT Shigella, but not with the ΔipaH7.8 mutant (Fig. 5D). Together, these results strongly suggested that GLMN functionally correlates with inflammasome activation in macrophages during Shigella infection.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

GLMN acts as a negative regulator of NLRP3 and NLRC4 inflammasomes. (A, Upper) Cells were treated with LPS/ATP for 75 min and then were labeled with FAM-YVAD-FMK and an anti-GLMN antibody. Arrows indicate the colocalization of GLMN (red) with the cleaved form of caspase-1, as determined by fluorochrome-labeled inhibitors of caspases (FLICKA) staining (green). (Scale bars, 10 μm.) (Lower) The number of GLMN puncta per cell and percentages of FLICA⁺ GLMN puncta/total GLMN puncta are shown in the graphs (n = 10). *P < 0.001. (B) siRNA-mediated knockdown of GLMN. 293T cells were transfected with or without HA-tagged NLRP3 or NLRC4, T7-tagged ASC, HA-tagged caspase-1, and FLAG-tagged IL-1β for 48 h. These cells were transfected simultaneously with GLMN siRNA or control GFP siRNA for 40 h. The presence of the cleaved form of IL-1β in cell extracts was detected by immunoblotting. (C) J774A.1 cells expressing GFP or GLMN were treated with LPS for 6 h followed by 1 mM ATP. Cytotoxicity and IL-1β cytokine production in the culture supernatants were evaluated at the indicated times. *P < 0.01. (D) GFP- or GLMN-expressing J774A.1 cells were infected with WT or ΔipaH7.8 Shigella at an MOI of 10. LDH and IL-1β production were measured in the culture supernatants. *P < 0.01. The error bars represent the SD of the measurements. (Also see Fig. S5.)

GLMN Dampens Inflammasome Activation and Macrophage Pyroptosis.

In the above context, we wished to establish the biological involvement of GLMN in modulating inflammasome activity. To this end, we created GLMN-null or GLMN-heterozygous mice on a BALB/c background (Fig. S6 B and C). Consistent with findings in other studies (50), GLMN^+/− mice were healthy, fertile, and had no visible disorders, whereas GLMN^−/− embryos at embryonic day (E)7.5–E10.5 were fleshy masses that did not undergo further development. Therefore, we examined Shigella-induced cytotoxicity and IL-1β secretion in BMDMs from GLMN^+/− and GLMN^+/+ mice. An intranasal Shigella infection study in GLMN^+/+ and GLMN^+/− mice indicated that GLMN^+/− animals produced higher levels of tissue IL-1β (Fig. 6A) and produced a stronger inflammatory response than GLMN^+/+ animals (Fig. 6B). Surprisingly, bacterial loads in the lung were remarkably higher in GLMN^+/− mice (Fig. 6C). As shown in Fig. 6 D and E, upon LPS/ATP stimulation or Shigella infection, the levels of cytotoxicity and IL-1β secretion increased up to twofold in BMDMs from GLMN^+/− mice compared with those from GLMN^+/+ mice. Furthermore, siRNA-mediated knockdown of GLMN in GLMN^+/− BMDMs promoted cell cytotoxicity and IL-1β secretion (Fig. 6 F and G and Fig. S6D). These results were consistent with the hypothesis that the level of GLMN in macrophages influences inflammasome activation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

GLMN dampens inflammasome activation and macrophage pyroptosis. (A–C) Six-week-old female BALB/c background GLMN^+/+ and GLMN^+/− mice were inoculated intranasally with the indicated Shigella strains (5 × 10⁷ cfu). The lungs were removed 24 h postinoculation and assayed for cytokine production (A), H&E staining (B), and the number of viable bacteria (C). n = 8; *P < 0.05. (D) BMDMs from GLMN^+/− or GLMN^+/+ mice were treated with LPS for 6 h followed by 1 mM ATP. Cytotoxicity and cytokine production in the culture supernatants were evaluated at the indicated times. *P < 0.05. (E) BMDMs from GLMN^+/− or GLMN^+/+ mice were infected with Shigella at an MOI of 10 for 2 h. Cytotoxicity and cytokine production in the culture supernatants were evaluated. *P < 0.05. (F and G) siRNA-mediated knockdown of GLMN in GLMN^+/− BMDMs. BMDMs from GLMN^+/− mice were transfected with GLMN siRNA or a nontargeting siRNA control for 48 h. Then cells were treated with LPS/ATP (F) or were infected with the indicated Shigella strains at an MOI of 10 (G). LDH and IL-1β production were measured in the culture supernatants. *P < 0.01. The error bars represent the SD of the measurements. (Magnification: 200×) (Also see Fig. S6.)