Ubiquitin-mediated response to microsporidia and virus infection in C. elegans

Abstract

Microsporidia comprise a phylum of over 1400 species of obligate intracellular pathogens that can infect almost all animals, but little is known about the host response to these parasites. Here we use the whole-animal host C. elegans to show an in vivo role for ubiquitin-mediated response to the microsporidian species Nematocida parisii, as well to the Orsay virus, another natural intracellular pathogen of C. elegans. We analyze gene expression of C. elegans in response to N. parisii, and find that it is similar to response to viral infection. Notably, we find an upregulation of SCF ubiquitin ligase components, such as the cullin ortholog cul-6, which we show is important for ubiquitin targeting of N. parisii cells in the intestine. We show that ubiquitylation components, the proteasome, and the autophagy pathway are all important for defense against N. parisii infection. We also find that SCF ligase components like cul-6 promote defense against viral infection, where they have a more robust role than against N. parisii infection. This difference may be due to suppression of the host ubiquitylation system by N. parisii: when N. parisii is crippled by anti-microsporidia drugs, the host can more effectively target pathogen cells for ubiquitylation. Intriguingly, inhibition of the ubiquitin-proteasome system (UPS) increases expression of infection-upregulated SCF ligase components, indicating that a trigger for transcriptional response to intracellular infection by N. parisii and virus may be perturbation of the UPS. Altogether, our results demonstrate an in vivo role for ubiquitin-mediated defense against microsporidian and viral infections in C. elegans.

Figure 1. C. elegans gene expression during infection with N. parisii.

A) Diagram of N. parisii infection stages in C. elegans intestinal cells. B) Synchronized populations of fer-15;fem-1 sterile animals were inoculated with N. parisii spores and collected for RNA extraction at timepoints corresponding to specific stages of infection. Uninfected controls were included for each timepoint. C) Number of significantly (FDR<0.05) up- or downregulated C. elegans genes during infection with N. parisii. D) Proportion of intestine- and germline-associated C. elegans genes with significantly altered expression at each timepoint. The “reference” bars indicate intestine or germline associated genes as a percentage of the C. elegans genome (20,404 genes). At each timepoint, 39% to 56% of all highly regulated genes were associated with the intestine, which represents a significant enrichment (chi-squared test, p<1.03E-26, all comparisons), while germline genes were significantly underrepresented (chi-squared test, p<1.51E-05, all comparisons) (Figure 1D, Table S2). E) Correlations between genes regulated by N. parisii infection and genes upregulated by other pathogens, stressors and immunity pathways. Gene sets were compared using the GSEA software (see Table S5 for detailed summary of results) and normalized enrichment scores (NESs) with a relaxed significance threshold (FDR<0.25, p<0.05) are reported in the figure. A positive NES (yellow) indicates a correlation with genes upregulated in response to N. parisii infection, while a negative NES (blue) indicates a correlation with genes downregulated in response to N. parisii infection (see Materials and Methods for analysis details). Black indicates no significant (FDR<0.25, p<0.05) correlation, and an NES with FDR<0.05 is indicated with an asterisk.

Figure 2. The SCF ligases, UPS and autophagy limit the growth of N. parisii in the C. elegans intestine.

A) Fluorescence and bright field images demonstrating FISH staining with a probe against N. parisii rRNA used to quantify pathogen load in the C. elegans intestine following 24 hours of infection with N. parisii. Scale bar  = 100 µm. B–F) Quantification of pathogen load (see Materials and Methods) in nematodes treated with RNAi against SCF ligase components (B), against ubiquitin (ubq-2) and two components of the proteasome (pas-5 and rpn-2) (C), against ubq-2, +/− fumagillin (D), against ubq-2, +/− FUdR (E), against autophagy components (F), and the C. elegans TOR ortholog let-363 (G). Pathogen area occupying each RNAi-treated animal was normalized to mean L4440 control values. The number of animals analyzed for each condition (n) is indicated. Mean +/− SEM is shown for all analyzed animals (data for B, C, F are from three independent experiments, data for D, G are from two independent experiments and data for E are from one experiment). Each independent experiment was comprised of two separate populations of animals. Statistical significance was assessed using a one-way ANOVA with Dunnett's Multiple Comparisons Test for B, C, and F, with Bonferroni Multiple Comparison Test for D and E, and with student's t-test for G (***p<0.001, **p<0.01, *p<0.05).

Figure 3. Targeting of N. parisii cells by the autophagy marker GFP::LGG-1 increases upon let-363/TOR RNAi.

GFP::LGG-1-expressing transgenic animals were fixed and stained with a FISH probe against N. parisii rRNA (red) and DAPI for DNA (blue). A) Intestine of an uninfected nematode, and B) an N. parisii-infected nematode, 8 hpi, are shown. N. parisii parasite cells not colocalizing with GFP::LGG-1 (arrowhead) and surrounded by GFP::LGG-1 (boxed in area) are indicated. Scale bars  = 20 µm. C) Enlarged view of boxed in area from panel B, showing cross-section from two other dimensions. Scale bar  = 2 µm. D) Quantification of parasite cell colocalization at 8 hpi with wild-type GFP::LGG-1 following knockdown of let-363 compared to L4440 vector control. Data are normalized for the average level of targeting in each independent experiment. Mean +/− SEM of two independent experiments is shown. Number of individual parasite cells assessed for colocalization is indicated.

Figure 4. N. parisii cells are targeted by host ubiquitin early during infection.

A–C) C. elegans intestines stained with a FISH probe against N. parisii rRNA (red), and DAPI for DNA (blue). A section of the image outlined by a dotted square is enlarged on the right and shows both an N. parisii parasite cell that colocalizes with ubiquitin (arrow), and one that does not (arrowhead). A) Animals were stained with an anti-conjugated-ubiquitin antibody, FK2 (green), and B, C) Transgenic C. elegans intestines expressing wild-type (B) or conjugation-defective mutant (C) GFP::ubiquitin (green). For A–C, scale bar  = 10 µm in main images and 2 µm in enlarged sections. D) Quantification of parasite cell colocalization at 12 hpi (see Materials and Methods for timepoint information) with FK2 antibody in the presence of increasing doses of fumagillin. Difference is significant: chi-squared test, p<6.3E-28, all comparisons. E) Quantification of parasite cell colocalization at 15 hpi with wild-type or mutant GFP::ubiquitin in the presence of increasing doses of fumagillin. Mean +/− SEM of two independent experiments is shown. F) Quantification of parasite cell colocalization at 15 hpi with wild-type GFP::ubiquitin following knockdown of cul-6 RNAi compared to L4440 vector control. Targeting of ubiquitin to parasite cells was less robust and more variable in animals feeding on HT115 RNAi bacteria compared to OP50-1 E. coli, ranging from 10.6% to 2.2% in control animals, and 2.1% to 1.8% in cul-6 RNAi treated animals and the data presented are normalized for the average level of targeting in each independent experiment. Mean +/− SEM of three independent experiments is shown. Number of individual parasite cells assessed for colocalization with ubiquitin is indicated.

Figure 5. N. parisii cells are almost never targeted by ubiquitin later during infection, but ubiquitin forms clusters that accumulate in the C. elegans intestine.

A,B) C. elegans intestines stained an anti-conjugated-ubiquitin antibody FK2 (green), and DAPI for DNA (blue): Panel B also includes a FISH probe against N. parisii rRNA (red). A) Sections of uninfected and N. parisii-infected C. elegans intestines. In the infected intestine an N. parisii spore labeled with the anti-conjugated-ubiquitin antibody is enlarged in box inset in upper left, and meront (arrowhead), and host nucleus (N) are indicated. Scale bar  = 10 µm. B) Sections of uninfected and N. parisii-infected C. elegans intestines (30 hpi) shown with ubiquitin cluster (arrow) and ubiquitin staining within an N. parisii meront (arrowhead) indicated. Scale bar  = 10 µm. C) Intestines of uninfected and infected animals expressing an intestinal GFP::ubiquitin transgene (48 hpi, grown at 20°C to prevent construct aggregation) are shown. Small GFP::ubiquitin aggregates are sometimes observed in uninfected animals (arrow). Scale bar  = 20 µm. D) Enlarged portion of box in panel C. Oblong N. parisii meronts are visible through the absence of green (arrowhead) and ubiquitin clusters associating with the meronts (arrow) are indicated. Scale bar  = 10 µm. E) Animals expressing the intestinal GFP::ubiquitin construct were infected with N. parisii and, together with control uninfected animals, fixed at the indicated times. Fixed animals were stained with a FISH probe against N. parisii rRNA to mark the infection and their intestinal cells were inspected for visible GFP::ubiquitin aggregates (30 transgenic animals were inspected per timepoint and condition). F) Animals expressing the intestinal control GFP::ubiquitinΔGG construct were treated and analyzed as in E.

Figure 6. UPS perturbation induces similar gene expression responses to N. parisii infection.

A) Expression of C17H1.6p::gfp and F26F2.1p::gfp in the intestine (pharyngeal myo-2p::mCherry expression is a marker for the presence of the transgene) following infection with N. parisii (8 and 24 hpi). Scale bars  = 100 µm. B) Expression of C17H1.6p::gfp and F26F2.1p::gfp following RNAi against ubq-1, ubq-2, pas-5 and rpn-2 in the absence of infection. C) Expression of endogenous mRNA of C17H1.6 and F26F2.1, as well as the SCF ligase components skr-1, skr-3, skr-4, skr-5 and cul-6 following RNAi against ubq-1, ubq-2, pas-5 and rpn-2 in the absence of infection, as assessed by qRT-PCR. Due to the very large changes in expression of C17H1.6 and F26F2.1 genes, these are presented on a separate graph to allow for expansion of the y-axis and easier observation of expression changes in SCF ligase components. Mean +/- SEM of two to three independent experiments.

Figure 7. Ubiquitin-mediated host response and defense against Orsay viral infection in C. elegans.

A) Viral pathogen load in nematodes treated with RNAi against the SCF ligase components skr-3, skr-4, skr-5, and cul-6 compared to vector control RNAi (L4440), as assessed by qRT-PCR for viral transcript. Mean +/− SEM of three independent experiments shown. B) Viral pathogen load in nematodes treated with RNAi against ubq-2, pas-5 and rpn-2 analyzed as above. Mean +/− SEM of three independent experiments shown. C) Intestines of F26F2.1p::gfp transgenic animals were stained with the FK2 antibody against conjugated ubiquitin (red), and DAPI for DNA (blue). Intestines outlined with white dotted line. Animals infected with virus have increased ubiquitin clustering in some intestinal cells compared to uninfected animals (arrowhead), and also express the GFP reporter. Scale bar  = 20 µm. ** p<0.01.

Figure 8. Model for SCF E3 ligases and ubiquitin-mediated responses to intracellular infection in C. elegans.

Intracellular microsporidia or viral infection triggers the expression of SCF ligase components in C. elegans, including a large number of F-box genes, the cullin cul-6, and Skp1-related genes, skr-3, -4, and -5. Due to the modularity of the SCF ligase complex, many SCF ligases with vast substrate recognition potential may be formed, which could recognize pathogen-derived proteins or host proteins. Ubiquitylation of substrates leads to their degradation by the proteasome or by autophagy, with large substrates such as microsporidia cells, potentially viral particles, and protein aggregates (not shown), targeted by autophagy, and individual pathogen or host proteins by the proteasome. N. parisii parasite cells may be able to suppress or evade ubiquitylation. Both intracellular infection and UPS stress can induce SCF ligase components, and greater demand on the UPS during intracellular infection may contribute to upregulation of SCF ligase components. See Discussion for more details.