IFN-gamma-inducible Irga6 mediates host resistance against Chlamydia trachomatis via autophagy

Abstract

Chlamydial infection of the host cell induces Gamma interferon (IFNgamma), a central immunoprotector for humans and mice. The primary defense against Chlamydia infection in the mouse involves the IFNgamma-inducible family of IRG proteins; however, the precise mechanisms mediating the pathogen's elimination are unknown. In this study, we identify Irga6 as an important resistance factor against C. trachomatis, but not C. muridarum, infection in IFNgamma-stimulated mouse embryonic fibroblasts (MEFs). We show that Irga6, Irgd, Irgm2 and Irgm3 accumulate at bacterial inclusions in MEFs upon stimulation with IFNgamma, whereas Irgb6 colocalized in the presence or absence of the cytokine. This accumulation triggers a rerouting of bacterial inclusions to autophagosomes that subsequently fuse to lysosomes for elimination. Autophagy-deficient Atg5-/- MEFs and lysosomal acidification impaired cells surrender to infection. Irgm2, Irgm3 and Irgd still localize to inclusions in IFNgamma-induced Atg5-/- cells, but Irga6 localization is disrupted indicating its pivotal role in pathogen resistance. Irga6-deficient (Irga6-/-) MEFs, in which chlamydial growth is enhanced, do not respond to IFNgamma even though Irgb6, Irgd, Irgm2 and Irgm3 still localize to inclusions. Taken together, we identify Irga6 as a necessary factor in conferring host resistance by remodelling a classically nonfusogenic intracellular pathogen to stimulate fusion with autophagosomes, thereby rerouting the intruder to the lysosomal compartment for destruction.

Figure 1. IFNγ-induced inhibition of C. trachomatis growth in WT MEFs.

Host cells were infected for 48 h with C. trachomatis or C. muridarum (MOI 1) and simultaneously treated with 100 U or 200 U/ml IFNγ or left untreated (control). (A) and (C) Immunofluorescence micrographs of MEFs infected with C. trachomatis or C. muridarum, respectively, stained with Chlamydia-IMAGEN kit. Chlamydiae (green), Host cells (red). Cytokine treatment resulted in a low number of detectable small inclusions in C. trachomatis infected cells only. Images taken using the same magnification. (B) and (D) Influence of IFNγ on development of infectious progeny. The yield of C. trachomatis (B), but not C. muridarum (D), infectious progeny decreased considerably upon IFNγ stimulation, Infectivity percentage calculated as follows: IFU/ml estimated for each treated monolayer / IFU/ml of control cells ×100. Infectivity expressed as a percentage of control cells ±standard deviation (SD) from three independent experiments (n = 3). WT, wild type; Ctr, C. trachomatis; Cmur, C. muridarum.

Figure 2. IFNγ triggers accumulation of Irga6, Irgd, Irgm2 and Irgm3 at early inclusions of C. trachomatis.

(A) Double immunofluorescence labelling of IRGs and C. trachomatis in MEFs stimulated for 24 h with 100 U/ml IFNγ and then infected for 3 h with C. trachomatis (MOI 5). IFNγ untreated MEFs were similarly infected. Upon IFNγ induction Irga6, Irgd, Irgm2 and Irgm3 localized to inclusions. Irgb6 colocalized strongly with inclusions in treated as well as untreated cells, whereas Irgm1 localization was minimal. (B) For quantification, around 300 bacterial inclusions were examined for each IRG in cytokine-treated or untreated MEFs. Colocalization expressed as a mean percentage: for each treatment, number of IRG +ve inclusions / total number of inclusions ×100. Error bars ±SD, n = 3. Scale bar represents 4 µm.

Figure 3. IFNγ stimulation induces interaction of lysosomes and autophagosomes with early C. trachomatis inclusions in MEFs.

(A) and (B) MEF monolayers were prestimulated for 24 h with 100 U/ml IFNγ and then infected with C. trachomatis as described in Figure 2. IFNγ untreated control MEFs were similarly infected. (A) Double immunofluorescence labelling 3 h p.i. revealed that IFNγ stimulated the association of the lysosomal marker LAMP1 (red) with C. trachomatis (green) inclusions (compare panel 2 with 1). (B) Percentages of colocalization with LAMP1. (C) Inhibition of lysosomal acidification by 100 nM Baf A1 stopped the IFNγ-mediated inhibition of C. trachomatis inclusion growth. (D) Baf A1 attenuated the IFNγ-induced reduction of C. trachomatis infectivity in MEFs. Infectivity assays were performed as in Figure 1B. Baf A1 led to an increase in the yield of infectious progeny in IFNγ-treated MEFs 48 h p.i. (E) and (F) IFNγ induces localization of autophagosomes to inclusions. MEFs were transfected for 24 h with the autophagosome membrane marker GFP-LC3 and then exposed to 100 U/ml IFNγ for an additional 24 h. Next, cells were infected with C. trachomatis (MOI 5) for 8 h. IFNγ-untreated control cells were similarly infected 48 h post-transfection. LC3 (green) localized to bacterial inclusions (red) in response to IFNγ stimulation (compare panel 2 with 1 in E). (F) Quantification of GFP-LC3-positive C. trachomatis inclusions in the presence or absence of IFNγ. Around 300 bacterial inclusions were examined for LAMP1 or LC3 sequestration. Colocalization expressed as a mean percentage: for each treatment, number of LAMP1 or LC3 +ve inclusions, respectively / total number of inclusions ×100.Scale bars in (A), (C) and (E) represent 2, 20 and 1 µm, respectively. Error bars ±SD, n = 3.

Figure 4. IFNγ cannot induce lysosomal fusion with early C. trachomatis inclusions to suppress bacterial growth in autophagy-lacking (Atg5−/−) MEFs.

(A) Normal development of inclusions (green) was observed in Atg5−/− MEFs (red), despite exposure to either 100 U or 200 U IFNγ/ml. Knockout cells were IFNγ treated, infected, and stained as in Figure 1. (B) Infectivity titration assays onto fresh HeLa cells revealed similar amounts of infectious progeny in IFNγ-treated and IFNγ-untreated Atg5−/− MEFs. Infectivity expressed as a percentage normalized to control (C) C. trachomatis inclusions avoid interaction with lysosomes in autophagy-defective Atg5−/− MEFs despite IFNγ induction. IFNγ treated and untreated Atg5−/− MEFs were infected as in Figure 3A and B. Double immunolabeling demonstrated no recruitment of lysosomes to inclusions [bacteria (green) and LAMP1 (red)] (D) Quantification of LAMP1-positive chlamydial inclusions revealed insignificant colocalization rates. Around 300 bacterial inclusions were examined. Colocalization expressed as a mean percentage: for each treatment, number of LAMP1 inclusions / total number of inclusions ×100. Images in (A) were taken under the same magnification, while scale bars in (C) represent 3 µm. Error bars ±SD, n = 3.

Figure 5. Absence of Irga6 colocalization at C. trachomatis early inclusions in IFNγ stimulated Atg5−/− MEFs.

(A) and (B) IFNγ treated and untreated cells were infected with C. trachomatis as in Figure 2. (A) Confocal microscopic analysis of infected stimulated Atg5-knockout cells shows similar staining patterns for Irgb6, Irgd, Irgm1, Irgm2 and Irgm3 as in stimulated infected WT MEFs (Figure 2). Irga6 does not localize to the inclusion in IFNγ-stimulated and unstimulated Atg5-knockout cells. (B) Quantification of colocalization rates of IRGs with C. trachomatis inclusions. Around 300 bacterial inclusions were examined. Colocalization expressed as a mean percentage: for each treatment, number of IRG +ve inclusions / total number of inclusions ×100. Error bars ±SD, n = 3. Scale bar represents 4 µm.

Figure 6. Absence of Irga6 enhances C. trachomatis replication and resistance against IFNγ-induced growth inhibition.

(A) Deletion of Irga6 promotes intracellular growth of the bacterial inclusion. Unstimulated WT and Irga6-knockout MEFs were infected for 48 h (MOI 1). Micrographs show an increased inclusion size in Irga6-knockout cells (panel 3) compared to WT MEFs (panel 1). IFNγ (100 U/ml) stimulation did not suppress chlamydial growth in Irga6-deficient MEFs (panel 4) as in IFNγ-induced WT MEFs (panel 2). (B) Infectivity of bacteria in Irga6-knockout MEFs is 4-fold higher than in WT MEFs and is unaffected by IFNγ induction. Results depicted as mean percentage normalized to control. (C) Lack of appreciable lysosomal fusion with C. trachomatis inclusions in IFNγ-stimulated Irga6−/− MEFs (panel 4) unlike that in stimulated WT MEFs (panel 2). Cells were stimulated, infected and stained for bacteria (green) and LAMP1 (red) as in Figure 3A. Untreated WT and Irga6-knockout monolayers (panels 1 and 3, respectively) served as controls. (D) IFNγ did not considerably increase rates of LAMP1 localization to C. trachomatis inclusions in Irga6−/− MEFs, compared with that in WT cells. Around 300 bacterial inclusions were examined Colocalization expressed as a mean percentage: for each treatment, number of LAMP1 +ve inclusions / total number of inclusions ×100. Scale bar in (A) and (C) represent 40 and 4 µm, respectively. Error bars ±SD, n = 3.

Figure 7. Irga6 plays a role in the regulation of the host autophagic machinery.

(A–C) Anti-LC3 immunoblot analysis of total lysates from uninfected WT, Atg5−/− and Irga6−/− MEFs or from cultures infected for indicated time points. Some uninfected cell cultures were exposed to 100 nM Rapa for 3 h or to 100 U/ml IFNγ for 32 h. Other monolayers were pretreated with IFNγ for 24 h prior to infection and then infected in the presence of chemicals. 1–7 indicate the different treatments. Host β-actin was used to control equal loading of proteins. (A) Autophagy is induced by Rapa and IFNγ in WT cells, while infection alone does not stimulate autophagy, as indicated by the amount of LC3-II. (B) Defective autophagy in Atg5−/− MEFs was observed, indicated by absence of LC3 processing (C) Abolishment of Irga6 apparently impaired autophagy induction. Only very low amounts of cellular LC3 (LC3-II) can be detected in Irga6-knockout cells, compared to those in WT MEFs. (D)–(E) Quantification of GFP-LC3-vacoules under different conditions in WT (D) and Irga6−/− MEFs (E) treated as indicated for (A)–(C), 24 h after transfection. Around 100 cells were examined by immunofluorescence microscopy. Data normalized against control. Data show average numbers per cell within a set cytoplasmic area. Error bars ±SD, n = 3.