Overexpression screen of interferon-stimulated genes identifies RARRES3 as a restrictor of Toxoplasma gondii infection

Abstract

Toxoplasma gondii is an important human pathogen infecting an estimated one in three people worldwide. The cytokine interferon gamma (IFNγ) is induced during infection and is critical for restricting T. gondii growth in human cells. Growth restriction is presumed to be due to the induction of interferon-stimulated genes (ISGs) that are upregulated to protect the host from infection. Although there are hundreds of ISGs induced by IFNγ, their individual roles in restricting parasite growth in human cells remain somewhat elusive. To address this deficiency, we screened a library of 414 IFNγ induced ISGs to identify factors that impact T. gondii infection in human cells. In addition to IRF1, which likely acts through the induction of numerous downstream genes, we identified RARRES3 as a single factor that restricts T. gondii infection by inducing premature egress of the parasite in multiple human cell lines. Overall, while we successfully identified a novel IFNγ induced factor restricting T. gondii infection, the limited number of ISGs capable of restricting T. gondii infection when individually expressed suggests that IFNγ-mediated immunity to T. gondii infection is a complex, multifactorial process.

Keywords: Toxoplasma; cell death; egress; growth restriction; infectious disease; interferon gamma; intracellular parasite; microbiology.

Figure 1.. Screen for interferon-stimulated genes (ISGs) impacting Toxoplasma gondii infection.

A549 cells were treated with indicated concentrations of IFNγ for 24 hr and subsequently infected with the type III strain CTG expressing GFP (CTG-GFP) for 36 hr. (A–C) Cells were fixed, stained with anti-GFP and anti-RFP antibodies, and imaged using a Cytation3 Imager. (A–C) Average parasitophorous vacuole (PV) size (A), PVs per field (B), and the percentage of vacuoles containing ≥8 parasites (C) was quantitated for data from 36 hr image sets. (D) Illustration of the method used to conduct the screen presented in (E). (E) A549 cells were transduced with a lentiviral expression cassette co-transcriptionally expressing tagRFP and an ISG of interest in a one gene per well format. After 72 hr, cells were infected with CTG-GFP for 36 hr, fixed, stained with anti-GFP and anti-RFP antibodies, and imaged with a Cytation3 Imager. Statistical significance was determined using a two-way ANOVA with Tukey’s test for post hoc analysis. ISGs enhancing or restricting infection >20% relative to control with p<0.0001 were classified as hits. Hits are shown in red and labeled. (F) WT and IDO1^−/− A549 cells were infected with CTG-GFP for 96 hr. Cells were fixed, stained with anti-SAG1 antibody, and imaged with a Cytation3 Imager. Average total infected area per well is shown. Loss of IDO1 in IDO1^−/− A549 cells was confirmed via western blot. Briefly, cells were treated with or without 1000 U/ml IFNγ for 24 hr before samples were harvested and IDO1 expression was determined. (A–C, F) Data represent the deviation of three biological replicates conducted in technical triplicate. (E) Data represent deviation of two biological replicates conducted in technical duplicate. Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis. ns, not significant; p>0.05, **p<0.01, ****p<0.0001.

Figure 2.. IRF1 and RARRES3 restrict Toxoplasma infection.

(A, B) Wild-type (WT) A549 cells were either not transduced (NT) or transduced with TRIP.RARRES3 or TRIP.FLUC control and split 48 hr later. After 60 hr, cells were stained with Hoechst 33342, SYTOX green, and imaged with a Cytation3 Imager. As a positive control, cells were permeabilized by treatment with methanol (MeOH) for 5 min prior to staining. Average cell number (A) and the percentage of SYTOX staining cells (B) were determined. (C–F) WT A549 cells were transduced with TRIP.RARRES3 or TRIP.FLUC control and infected 72 hr later with CTG-GFP for 36 (C, D) or 96 (E, F) hr. Cells were fixed, stained with anti-GFP and anti-RFP antibodies, and imaged using a Cytation3 Imager. Average PV number per field (C) and PV size (D) were quantitated for 36 hr infections while total area infected per sample (E) and average foci size (F) were quantitated for 96 hr infections. Data in (A) represent four to seven biological replicates conducted in technical triplicate. Data in (B) represent two to four biological replicates conducted in technical triplicate. Data in (C–F) represent three to four biological replicates conducted in technical triplicate. Statistical significance was determined using a Brown-Forsythe and Welch ANOVA (A, B) or a two-way ANOVA with Tukey’s test for post hoc analysis (C–F). *p≤0.05, **p<0.01.

Figure 3.. Comparison of genes induced by IRF1 and IFNγ in A549 cells.

Cells were transduced with TRIP.IRF1 or TRIP.FLUC control lentivirus. Cells transduced with FLUC were further treated 72 hr later with or without 1000 U/ml IFNγ for 24 hr. All cell populations were subsequently harvested and analyzed by RNA-Seq. (A, B) Changes in gene expression relative to FLUC control expressing cells for cells treated with IFNγ (A) or ectopically expressing IRF1 (B). Genes upregulated in both IRF1 expressing and IFNγ treated cells are defined as ‘Shared’ while genes only upregulated in one of these two cell populations are defined as ‘Unique.’ (C) Comparison of genes induced ≥2-fold with a false discovery rate cutoff of 0.05 by each condition and their overlap with the ISG library used in the screen described in Figure 1. (D–E) Lists of induced genes were analyzed with PANTHER gene ontology analysis. The top 10 most enriched processes amongst genes induced by IFNγ (D) and IRF1 (E) are shown. Redundant terms were excluded from these lists with only the most enriched version of each term remaining.

Figure 4.. RARRES3 restricts Toxoplasma infection in a STAT1 independent manner.

To determine if restriction of T. gondii growth was STAT1 dependent, STAT1^−/− A549 cells were generated. To confirm complete insensitivity to interferon treatment, WT or STAT1^−/− A549 cells were treated with or without 4000 U/ml IFNγ for 6 hr, fixed, stained with anti-IRF1 antibodies, and imaged with a Cytation3 Imager. (A) Representative images and (B) quantitation are shown. Scale bar=50 µm. (C–F) STAT1^−/− A549 cells were transduced with TRIP.RARRES3 or TRIP.FLUC control and infected 72 hr later with CTG-GFP for 36 (C, D) or 96 (E, F) hr. Cells were fixed, stained with anti-GFP and anti-RFP antibodies, and imaged using a Cytation3 Imager. Average PV number per field (C) and PV size (D) were quantitated for 36 hr infections while total area infected per sample (E) and average foci size (F) were quantitated for 96 hr infections. HeLa reporter cell lines expressing GAS-LUC (G), kB-LUC (H), ISRE-GLUC (I), and GFP-LUC (J) were either not transduced (NT) or transduced with TRIP.RARRES3, TRIP.FLUC, or TRIP.GFP. After 72 hr, cells were mock treated or treated with 100 U/ml IFNβ or IFNγ as indicated and infected with CTG-GFP for 36 hr. Cells were harvested for luciferase assay. Data in (B) represent means ± SD of four biological replicates conducted in technical duplicate. Data in (C–F) represent means ± standard deviation of four biological replicates conducted in technical triplicate. Data represent means ± SD of two (G) or three (H–I) biological replicates conducted in technical duplicate. Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis except for (D) where Mann-Whitney’s U-test was used. *p≤0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 5.. RARRES3 deficiency partially reverses IFNγ-mediated restriction of Toxoplasma infection.

(A, B) A549s were transduced with V5 tagged WT RARRES3 or the catalytically inactive mutants C113A or C133S for 72 hr. (A) Cells were infected with CTG-GFP for 96 hr. Cells were harvested, stained with anti-GFP and anti-RFP antibodies, and imaged with a Cytation3 Imager. The total area infected per sample is shown. (B) Western blot from lysates prepared with above RARRES3 expressing cells. Membranes were probed with mouse anti-actin and mouse anti-V5 antibodies overnight followed by goat anti-mouse 680RD and imaged with a LI-COR Odyssey scanner. RARRES3^−/− A549s or wild-type cells transduced with a nontargeting CRISPR/Cas9 sgRNA were transduced with Cas9 resistant TRIP.RARRES3 or TRIP.FLUC as indicated. After 72 hr, cells were treated with or without 100 U/ml IFNγ for 24 hr as indicated and subsequently infected with CTG-GFP for 96 hr. Cells were harvested, stained with anti-GFP and anti-RFP antibodies, and imaged with a Cytation3 Imager. (C) Representative images and (D) quantitation are shown. Scale bar=500 µm. (E) RARRES3^−/− or RARRES3^+/+ A549 cells transduced with a nontargeting CRISPR/Cas9 control sgRNA were transduced with TRIP.FLUC or TRIP.IRF1 derived lentivirus. After 72 hr, cells were infected with CTG-GFP for 96 hr, harvested, stained with anti-GFP and anti-RFP antibodies, and imaged with a Cytation3 Imager. Average total infected area per well is shown. Data represent means ± standard deviation of three (A) or four (D, E) biological replicates conducted in technical triplicate. Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis. ns, not significant; *p>0.05, **p<0.01, ****p<0.0001.

Figure 5—figure supplement 1.. RARRES3 does not restrict infection of type I or II strains of Toxoplasma gondii.

A549 cells were transduced with TRIP.RARRES3 and infected 72 hr later with a type I strain expressing GFP (RH88-GFP) (A, B) or a type II strain expressing GFP (Me49-GFP) (C, D) for 4 or 6 days, respectively. Cells were fixed, stained with anti-GFP and anti-RFP antibodies, and imaged using a Cytation3 Imager. Average total infected area per well (A, C) and average area of infection foci (B, D) are shown. Data represent means ± standard deviation of two to three biological replicates conducted in technical triplicate. Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis.

Figure 6.. RARRES3 promotes premature egress of Toxoplasma gondii.

A549 cells were transduced with TRIP.RARRES3 or TRIP.FLUC control and infected 72 hr later with CTG-GFP for 36 (A, B, D, F), 50 (C), or 72 (E) hr. Cells were treated with the cell death inhibitors Z-VAD-FMK (50 µM), GSK’963 (1 µM), GSK’872 (5 µM), NSA (10 µM), and Z-YVAD-FMK (10 µM) or the parasite egress inhibitor compound 1 (5 µM) as indicated during infection. (A, B, E, F) Cell supernatant was collected after infection and lactate dehydrogenase (LDH) activity was determined to measure cell lysis. As a control to measure maximal LDH release, cells were lysed before supernatant collection. (C) Live infection was imaged every 15 min starting 2 hr postinfection until 50 hr postinfection. Time of parasite egress was recorded for at least 100 PVs per condition per replicate. Percentage of total parasites egressed by the end of each hour is indicated. (D) Cells were fixed, stained with anti-RFP and anti-GFP antibodies, and imaged with a Cytation3 Imager. Average cells per field are shown. Data represent the means ± standard deviation (A, B, D–F) or standard error of the mean (C) of three to five biological replicates conducted in technical duplicate (A, B, D–F), or singlet (C). Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis. ns, not significant; P>0.05, *p≤0.05, ***p<0.001, ****p<0.0001.

Figure 6—figure supplement 1.. Compound 1 prevents host cell death during infection.

(A–C) A549 cells were transduced with TRIP.RARRES3 or TRIP.FLUC control and infected 72 hr later with CTG-GFP for 36 hr (A) or 72 hr (B, C). Cells were stained with Hoechst 33342, propidium iodide, and imaged with a Cytation3 imager. As a positive control, cells were permeabilized by treatment with methanol (MeOH) for 5 min prior to staining. The percentage of propidium iodide staining cells (A, B) and average cell number per field (C) were determined. (D–F) HFF cells were pretreated with or without IFNγ for 24 hr and subsequently infected with CTG-GFP for 36 hr. Cells were stained with Hoechst 33342, propidium iodide, and imaged with a Cytation3 imager. As a positive control, cells were permeabilized by treatment with methanol (MeOH) for 5 min prior to staining. (D) Representative images are shown. Scale bar=50 µm. Average cell number (E) and the percentage of propidium iodide staining cells (F) were determined. Data represent means ± standard deviation of two or three biological replicates conducted in technical duplicate. Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis. ***p<0.001, ****p<0.0001.

Figure 7.. IFNγ-dependent host cell death during infection in HFFs is partially RARRES3 dependent.

(A, B) HFF cells were transduced with TRIP.RARRES3 or TRIP.FLUC control and infected 72 hr later with CTG-GFP for 96 hr. Cells were fixed, stained with anti-GFP and anti-RFP antibodies, and imaged using a Cytation3 Imager. Total infected area per well (A) and average number of infection foci (B) are shown. (C–D) WT HFFs expressing a nontargeting sgRNA control or RARRES3 deficient HFFs were pretreated with or without 1000 U/ml IFNγ for 24 hr. (C) Cells were infected with CTG-GFP for 36 hr. Supernatant was collected and lactate dehydrogenase (LDH) activity was determined. As a control to measure maximal LDH release, cells were lysed before supernatant collection. (D) Cells were infected with CTG-GFP for 96 hr. Samples were treated as in (A). Total infected area per sample is shown. (E) HFFs were infected with CTG-GFP for 36 hr in the presence or absence of 5 µM Compound 1. Supernatant was collected and LDH activity was determined. Data represent means ± standard deviation of three (A, B, D, E) or four (C) biological replicates conducted in technical duplicate (A, B, D, E) or singlet (C). Statistical significance was determined using two-way ANOVA with Tukey’s test for post hoc analysis. *p≤0.05, ***p<0.001, ****p<0.0001.