Identification of the microsporidian Encephalitozoon cuniculi as a new target of the IFNγ-inducible IRG resistance system

Abstract

The IRG system of IFNγ-inducible GTPases constitutes a powerful resistance mechanism in mice against Toxoplasma gondii and two Chlamydia strains but not against many other bacteria and protozoa. Why only T. gondii and Chlamydia? We hypothesized that unusual features of the entry mechanisms and intracellular replicative niches of these two organisms, neither of which resembles a phagosome, might hint at a common principle. We examined another unicellular parasitic organism of mammals, member of an early-diverging group of Fungi, that bypasses the phagocytic mechanism when it enters the host cell: the microsporidian Encephalitozoon cuniculi. Consistent with the known susceptibility of IFNγ-deficient mice to E. cuniculi infection, we found that IFNγ treatment suppresses meront development and spore formation in mouse fibroblasts in vitro, and that this effect is mediated by IRG proteins. The process resembles that previously described in T. gondii and Chlamydia resistance. Effector (GKS subfamily) IRG proteins accumulate at the parasitophorous vacuole of E. cuniculi and the meronts are eliminated. The suppression of E. cuniculi growth by IFNγ is completely reversed in cells lacking regulatory (GMS subfamily) IRG proteins, cells that effectively lack all IRG function. In addition IFNγ-induced cells infected with E. cuniculi die by necrosis as previously shown for IFNγ-induced cells resisting T. gondii infection. Thus the IRG resistance system provides cell-autonomous immunity to specific parasites from three kingdoms of life: protozoa, bacteria and fungi. The phylogenetic divergence of the three organisms whose vacuoles are now known to be involved in IRG-mediated immunity and the non-phagosomal character of the vacuoles themselves strongly suggests that the IRG system is triggered not by the presence of specific parasite components but rather by absence of specific host components on the vacuolar membrane.

Figure 1. IFNγ restricts E. cuniculi growth in mouse embryonic fibroblasts.

(A) Mouse embryonic fibroblasts (MEFs) from C57BL/6 mice were induced with IFNγ for 24 h or left uninduced before infection with E. cuniculi spores. Cells were fixed at the indicated time points and the number of meronts (stained with anti-meront mAb 6G2) per 500 host nuclei (stained with DAPI) was counted. The inhibition in the IFNγ-treated sample compared to the uninduced control sample is presented as mean +/− standard deviation (SD) of 3–7 replicates per time point from at least 2 individual experiments. Significant differences (of 0.5 h, 1 h and 2–3 h compared to 24–26 h) were calculated with a two tailed T-test. (B) MEFs were induced with IFNγ or left uninduced, infected with E. cuniculi spores for 24 h and stained as in A. Single meronts and meronts that divided once (double meront) were counted per 500 host nuclei and shown as percent of total vacuoles of uninduced controls. Numbers indicate the counted number of single or double meronts per 500 host cells. Data from three independent experiments (Exp. 1–3) is presented. (C) MEFs were stimulated with IFNγ and/or infected with E. cuniculi spores for 2 or 5 days or left untreated. Cell lysates were separated by SDS-PAGE and Western Blots were cut into three regions to probe for anti-meront mAB 6G2 as well as anti-spore wall protein 1 pAS SWP1. Calnexin staining served as loading control and Irgb6 staining (mAB B34) to proof IFNγ-induction. The asterisk marks an unknown E. cuniculi-derived protein that is detected by the Calnexin antibody. These Western Blots emerged from one single SDS-PAGE, the 45–70 kDa region was first probed with mouse mAB B34, stripped, and then probed for anti-SWP1 rabbit pAS. Experiments for both time points were performed at least three times.

Figure 2. IRG proteins accumulate at the E. cuniculi parasitophorous vacuolar membrane.

(A–F) MEFs were induced with IFNγ for 24 h and then infected with E. cuniculi spores for 24 h. Fixed cells were stained for anti-meront mAB 6G2 as well as for endogenous Irga6 (165/3 pAS and 10D7 mAB), Irgb6 (A20 pAB), Irgd (2078 pAS) and Irgm2 (H53 pAS). Nuclei were labeled with DAPI. Representative microscopic images of IRG-positive E. cuniculi PVMs are presented. Yellow arrows point at the IRG-loaded PVM which is magnified at the end of each panel (zoom in the following order: upper left: merged image, upper right: phase contrast, lower left: anti-meront, lower right anti-IRG); white arrow: unloaded meront, every scale bar is 10 µm. (G) Quantification of Irga6 and Irgb6 loading onto the E. cuniculi PVM at different time points post infection. 100 vacuoles were evaluated per sample, a black dot indicates that sample was not counted, three independent experiments are shown.

Figure 3. IRG proteins load onto PVM in a cooperative manner.

MEFs were induced with IFNγ for 24 h and then infected with E. cuniculi spores for 12 h (A) or 24 h (B, C). Fixed cells were stained for meronts (mouse mAB 6G2, red) as well as for endogenous Irga6 (rabbit pAS165/3, green) and Irgb6 (goat pAB A20, far-red pseudocolored in magenta). Nuclei were labeled with DAPI. Representative images from 4 independent experiments are shown; white box indicates enlarged area shown below; scale bar: 10 µm. (D) Quantification of cooperative loading after 24 h; Irgb6-single, Irga6-single or Irgb6/Irga6-double (both) positive meronts are shown as % of total 6G2-positive meronts; 100 vacuoles were counted in each independent experiment (Exp. 1–3).

Figure 4. IFNγ suppressive effect on E. cuniculi growth is impaired in GMS-IRG knock-out cells.

(A) Wildtype (wt) or Irgm1/Irgm3 knock-out (KO) MEFs were induced with 200 U/ml IFNγ for 24 h and then infected with E. cuniculi spores for 24 h or left untreated. Cells were fixed and stained for meronts using 6G2 mAB (red) and host nuclei with DAPI (pseudocolored in cyan). Representative fluorescence microscopic images are shown. (B) Quantification of A, representative of two independent experiments. (C/D) Transformed wildtype or transformed IRG knock-out MEFs were induced with IFNγ for 24 h and then infected with E. cuniculi spores or left untreated. Cells were harvested after 2 days (in D) and 5 days (in E) post-infection. Cell lysates were separated by SDS-PAGE and Western blots were cut into three regions and simultaneously probed for anti-meront mAB 6G2, anti-Calnexin pAB, which served as loading control and anti-Irgb6 (mAB B34) or anti-Irga6 (mAB 10E7 for Irgm1 KO and Irgm3 KO MEFs both at 5 d post infection; 165/3 pAS for Irgm1/Irgm3KO MEFs at 5 d post infection) as IFNγ-induction control. The black arrows highlight a 6G2-positive protein band indicating E. cuniculi growth despite presence of IFNγ, which is inhibited in wt cells (grew arrows). The asterisk marks an unknown E. cuniculi-derived protein that is detected by the Calnexin antibody. The white asterisk marks unspecific bands. The four samples of one cell line per time point were analyzed together by one single SDS-PAGE and Western Blot, except for the Irgm1KO MEFs at 2 d post infection. The data represents at least three independent experiments.

Figure 5. E. cuniculi infection triggers IFNγ-dependent host cell death.

(A) Wt MEFs were induced with IFNγ for 24 h and then infected with E. cuniculi spores for 24 h or left untreated. Without fixation, the cells were stained with Hoechst and Propidium iodide dye, photographed under live-cell conditions and automatically enumerated. (B) Quantification of A, graph represents mean values +/− SD from five independent experiments. (C) Wt MEFs were seeded in 96-wells and induced with IFNγ for 24 h (black bars) or left untreated (white bars). Cells were infected with E. cuniculi spores at different MOIs (MOI = 5–20) or with T. gondii Me49 tachyzoites (MOI = 5) as positive control. Cell viability was measured with a formazan-based colorimetric assay 24 h or 48 h post infection and expressed as percentages of uninduced uninfected control cells. Graph represents mean value +/− SD of triplicates of one representative experiment. Three independent experiments were performed; Significance was calculated with two-tailed T-Test: ** p>0.005, *** p>0.0005.

Figure 6. Tryptophan supplementation cannot reverse the IFNγ-mediated E. cuniculi restriction.

C57/BL/6 MEFs (A) or mouse enterocytic CMT-93 cells (B) were treated with IFNγ for 24 h or left uninduced. Indicated doses of L-Tryptophan (W) were added to the medium 30 minutes prior infection with E. cuniculi spores. Cell lysates were prepared 5 days post infection and separated by one SDS-PAGE. Western blots were cut into three regions probed simultaneously with for anti-meront mAB 6G2, anti-Calnexin pAB as loading control and anti-Irga6 (10E7 mAB) as IFNγ-induction control. The data are representative of three independent experiments.

Figure 7. Scheme of different dynamics of IRG action.

Upon E. cuniculi infection, a low but steady number of IRG-positive vacuoles can be detected over the first 24 h post infection accompanied by a continuous loss of viable meronts. Moreover, host cell death is triggered by the combination of E. cuniculi infection and IFNγ induction. Initiation of IRG loading might be a stochastic and asynchronous event that is followed by a rapid elimination of the pathogen. In contrast, IRG loading on PVs of avirulent T. gondii strains starts immediately after parasite invasion. IRG-positive vacuoles seem to accumulate reaching a maximum at about 2 h post infection. When fully loaded, the vacuoles disrupt followed by T. gondii death and host cell death in an invariant order.

Figure 8. Scheme of the IRG resistance system and its target organisms.

In IFNγ-stimulated mouse cells, GMS proteins localise mainly to endomembranes such as the ER and keep membrane-bound or cytosolic GKS proteins in a GDP-bound inactive state. Our current view is that the plasma membrane is protected by a hypothetical unknown factor that inhibits GKS protein-mediated damage. During host cell infection by T. gondii or E. cuniculi, invagination of the plasma membrane creates a parasitophorous vacuole that excludes the hypothetical factor and also does not carry GMS proteins. This “missing-self” allows GKS proteins to activate and accumulate on the PVM leading to the PV disruption, pathogen elimination and ultimately host cell death. However, bacteria entering via phagocytic mechanisms do not actively exclude the hypothetical factor and are therefore targeted for endolysosomal degradation.