Stress granules (SGs) are cytoplasmic assemblies of RNA and protein that form when translation is repressed during the integrated stress response. SGs assemble from the combination of RNA–RNA, RNA–protein and protein–protein interactions between messenger ribonucleoprotein complexes (mRNPs). The protein adenosine deaminase acting on RNA 1 (ADAR1, also known as ADAR) recognizes and modifies double-stranded RNAs (dsRNAs) within cells to prevent an aberrant innate immune response. ADAR1 localizes to SGs, and since RNA–RNA interactions contribute to SG assembly and dsRNA induces SGs, we examined how ADAR1 affects SG formation. First, we demonstrate that ADAR1 depletion triggers SGs by allowing endogenous dsRNA to activate the integrated stress response through activation of PKR (also known as EIF2AK2) and translation repression. However, we also show that ADAR1 limits SG formation independently of translation inhibition. ADAR1 repression of SGs is independent of deaminase activity but is dependent on dsRNA-binding activity, suggesting a model where ADAR1 binding limits RNA–RNA and/or RNA–protein interactions necessary for recruitment to SGs. Given that ADAR1 expression is induced during viral infection, these findings have implications for the role of ADAR1 in the antiviral response.

This article has an associated First Person interview with the first author of the paper.

Stress granules are cytoplasmic assemblies of RNA and protein that form when translation initiation is limited (Buchan and Parker, 2009; Kedersha et al., 1999). While the functions of stress granules are poorly understood, they have been proposed to have roles in stress tolerance and RNA stability (Hofmann et al., 2021; Protter and Parker, 2016). Stressors that induce translational shutdown and stress granule formation include oxidative stress, heat shock and osmotic stress (Kedersha and Anderson, 2007). The presence of cytoplasmic double-stranded RNA (dsRNA) during viral replication also induces stress granules through activation of PKR (also known as EIF2AK2), which then represses translation by phosphorylation of eIF2α (EIF2A) (Lemaire et al., 2008; Okonski and Samuel, 2013).

Stress granules form from a pool of non-translating mRNPs and RNA-binding proteins, and their formation is driven by a combination of RNA–RNA, RNA–protein and protein–protein interactions (Guillén-Boixet et al., 2020; Jain et al., 2016; Khong et al., 2017; Sanders et al., 2020; Van Treeck and Parker, 2018; Yang et al., 2020). Stress granule formation can be influenced by trapping RNAs on polysomes, increasing or decreasing the levels of key proteins that regulate stress granule formation (Tourrière et al., 2003), altering post-translational modifications that affect assembly, and/or modulating the availability of RNA–RNA interactions (Buchan and Parker, 2009; Ohn and Anderson, 2010; Tauber et al., 2020). Given the diverse manners in which stress granule assembly can be affected, we hypothesized that the protein ADAR1 might affect stress granule formation in multiple ways.

Adenosine deaminase acting on RNA 1 (ADAR1, also known as ADAR) is a ubiquitously expressed, essential protein that catalyzes the conversion of adenosines to inosines (A-to-I editing) within regions of dsRNA, which introduces I-U mismatches and destabilizes dsRNA duplexes (Bass, 2002; Bass and Weintraub, 1988). This editing functions at least in part to prevent aberrant activation of the innate immune response due to endogenous dsRNAs, which can trigger translational shutdown (Chung et al., 2018; Lamers et al., 2019). ADAR1 may plausibly influence stress granule formation in several ways. First, ADAR1 editing represses PKR activation, which could limit stress granule assembly, since dsRNA-induced stress granules require PKR (Burke et al., 2020). Additionally, ADAR1 may modulate intermolecular RNA–RNA or RNA–protein interactions that contribute to stress granule assembly.

Here, we investigate the mechanisms of ADAR1-mediated stress granule inhibition. We demonstrate that ADAR1 deficiency induces PKR-dependent translational shutoff and stress granule formation, consistent with a previous report (Berchtold et al., 2018). Furthermore, we show that ADAR1 inhibits stress granule formation independently of translational shutoff. Surprisingly, this effect is independent of ADAR1 A-to-I editing activity and suggests that ADAR1 can inhibit stress granule assembly by modulating the composition and/or organization of messenger ribonucleoprotein complexes (mRNPs). One isoform of ADAR1 is induced during viral infection, which suggests that ADAR1 may play a role in modulating stress granule formation during infections.

ADAR1 deficiency induces stress granules

Intracellular dsRNA can trigger translational shutdown by activation of PKR (Lemaire et al., 2008; Okonski and Samuel, 2013). Since ADAR1 functions to destabilize dsRNA duplexes (Bass and Weintraub, 1988), we examined whether ADAR1 deficiency would affect stress granule assembly. Previous results have demonstrated that ADAR1 knockdown (KD) induces stress granules in some cells (Berchtold et al., 2018). To see whether we could reproduce this observation and understand the mechanism, we performed an siRNA-mediated knockdown of ADAR1 in U-2 OS cells. ADAR1 depletion was previously reported to activate RNase L (RL) (Li et al., 2017), which triggers widespread RNA degradation, can prevent stress granule formation and lead to cell death (Burke et al., 2019, 2020; Li et al., 2017). Given this, we performed KD experiments in RL knockout (KO) cells (Burke et al., 2019).

Consistent with previously published results (Berchtold et al., 2018), by staining for the stress granule marker PABPC1, we observed that ∼5% of ADAR1 siRNA-treated U-2 OS cells have PABPC1 foci at any given time under non-stressed conditions (Fig. 1A; Fig. S1A). We also observed the same result by staining for G3BP1 in a wild-type (WT) U-2 OS cell line, suggesting that ADAR1 KD does not activate endogenous RL (Fig. S1D). To demonstrate that this result was not an artifact of the siRNA treatment, we attempted to create ADAR1 KO cells using CRISPR-Cas9. ADAR1 KO is lethal; however, it has been reported to be rescued by KO of RL (Li et al., 2017), so we performed KO of ADAR1 using CRISPR-Cas9 in RL KO U-2 OS cells. However, we only obtained hypomorphic ADAR1+/− cell lines, suggesting that ADAR1 cell lethality is not solely mediated by RL. Consistent with the ADAR1 KD, we also observed that ∼5% of ADAR1+/− U-2 OS cells spontaneously formed PABPC1 foci (Fig. S1B,C). Thus, the reduction of ADAR1 levels leads to formation of PABPC1 foci.

Fig. 1.

ADAR1 deficiency induces stress granule formation in a subset of cells. (A) Immunofluorescence (IF) staining for ADAR1 and PABPC1 in RL KO U-2 OS cells treated with either a non-targeting or ADAR1 siRNA. The proportion of cells with stress granules indicated. Scale bars: 20 μm (B) Single-molecule FISH validation of a stress granule-enriched lncRNA, NORAD, and stress granule-depleted mRNA, GAPDH, in ADAR1-deficient cells (ADAR1+/−). IF for G3BP1 also shown. Deconvolved images are shown, with dashed boxes indicating regions shown in magnified views. Nuclei are stained with DAPI (blue). Scale bars: 20 μm (bottom), 2 μm (top). (C) Puromycin labeling and IF for PABPC1 in ADAR1 siRNA-treated cells. Nuclei are stained with DAPI (blue). Scale bars: 20 μm (D) Quantification of the experiment shown in C (SGs, stress granules; w/, with; w/o, without). Data are mean±s.d., seven images were analyzed. ***P<0.001 (unpaired two-tailed t-test). (E) WT and PKR KO A549 cells treated with an ADAR1 siRNA and stained for stress granule marker G3BP1. The proportion of cells with stress granules is indicated. Scale bars: 10 μm. Data in A–C,E are representative of three or more experiments.

Fig. 1.

ADAR1 deficiency induces stress granule formation in a subset of cells. (A) Immunofluorescence (IF) staining for ADAR1 and PABPC1 in RL KO U-2 OS cells treated with either a non-targeting or ADAR1 siRNA. The proportion of cells with stress granules indicated. Scale bars: 20 μm (B) Single-molecule FISH validation of a stress granule-enriched lncRNA, NORAD, and stress granule-depleted mRNA, GAPDH, in ADAR1-deficient cells (ADAR1+/−). IF for G3BP1 also shown. Deconvolved images are shown, with dashed boxes indicating regions shown in magnified views. Nuclei are stained with DAPI (blue). Scale bars: 20 μm (bottom), 2 μm (top). (C) Puromycin labeling and IF for PABPC1 in ADAR1 siRNA-treated cells. Nuclei are stained with DAPI (blue). Scale bars: 20 μm (D) Quantification of the experiment shown in C (SGs, stress granules; w/, with; w/o, without). Data are mean±s.d., seven images were analyzed. ***P<0.001 (unpaired two-tailed t-test). (E) WT and PKR KO A549 cells treated with an ADAR1 siRNA and stained for stress granule marker G3BP1. The proportion of cells with stress granules is indicated. Scale bars: 10 μm. Data in A–C,E are representative of three or more experiments.

The PABPC1 foci that formed in ADAR1-deficient cells were canonical stress granules by several criteria. First, these granules contained canonical stress granule markers, including G3BP1, FMRP (also known as FMR1), PABPC1, eIF4G1, TIA-1 and poly(A)+ RNA (Fig. 1A; Fig. S1D,E). They were also enriched for the NORAD long non-coding RNA (lncRNA) and were depleted for GAPDH mRNA, indicating that they have an RNA composition similar to that of stress granules (Fig. 1B) (Khong et al., 2017). Additionally, ADAR1-deficient cells with constitutive stress granules had reduced translation compared to that of cells without stress granules, as observed using puromycin labeling (Fig. 1C,D). Taken together, these results suggest that ADAR1 deficiency results in translational repression, leading to canonical stress granule formation in a subset of cells.

Stress granule induction by ADAR1 deficiency is due to PKR activation

A reasonable hypothesis is that ADAR1 deficiency triggers translational repression and stress granule induction by activating the kinase PKR. PKR is known to be activated by dsRNA and phosphorylates eIF2α to shut off translation initiation, which can trigger stress granule formation (Lemaire et al., 2008; Proud, 2005; Samuel, 1979). Thus, we hypothesized that ADAR1 deficiency results in increased dsRNA in the cytoplasm, which results in PKR activation and stress granule formation. To test this hypothesis, we knocked down ADAR1 in WT and PKR KO A549 cells to ask whether the constitutive stress granules seen in ADAR1-deficient cells are dependent on PKR.

An important result was that we did not observe any stress granules following ADAR1 KD in PKR KO cells (Fig. 1E). In contrast, we observed that ∼5% of WT A549 cells treated with the ADAR1 siRNA had constitutive stress granules. We interpret these results to indicate that PKR activation is responsible for the constitutive stress granules seen in ADAR1-deficient cells, which is consistent with previous results (Li et al., 2012; Okonski and Samuel, 2013). If other eIF2α kinases were activated by ADAR1 deficiency, we would have observed stress granules in PKR KO cells treated with the ADAR1 siRNA. Since we did not observe any stress granules in the PKR KO cells, we interpret this to mean that PKR is the primary eIF2α kinase activated by ADAR1 deficiency.

Canonical stress granules disassemble when translation is restored; thus, we sought to ask whether the constitutive stress granules in ADAR1-deficient cells disassemble upon treatment with a translation activator. To answer this, we utilized a stable GFP–G3BP1-expressing U-2 OS cell line to study the constitutive stress granules in ADAR1-deficient cells by live-cell imaging. As expected, a subset of GFP–G3BP1-expressing ADAR1-deficient cells had constitutive stress granules (Fig. S2B). ISRIB is an eIF2B activator that restores translation in response to eIF2α phosphorylation and causes rapid disassembly of sodium arsenite-induced stress granules (Fig. S2A) (Rabouw et al., 2019; Sidrauski et al., 2015). Thus, if PKR activation solely represses translation by the phosphorylation of eIF2α, then ISRIB should reverse the formation of stress granules in ADAR1-deficient cells.

Surprisingly, we observed that the constitutive stress granules in ADAR1-deficient cells were resistant to ISRIB-induced disassembly (Fig. S2B). This is consistent with recent reports of stress granules induced by the synthetic dsRNA analog poly(I:C) being resistant to ISRIB (Burke et al., 2020). This suggests that dsRNA or ADAR1 deficiency induce stress granules by the activation of PKR through at least one mechanism that is independent of ISRIB modulation, and therefore is likely to be independent of eIF2α phosphorylation.

We also examined whether cycloheximide (CHX) would impact ADAR1 deficiency-induced stress granules, since when CHX inhibits translation elongation, it traps mRNAs in polysomes and prevents stress granule assembly (Kedersha et al., 2000; Mollet et al., 2008). We observed that a 1 h treatment with CHX did not resolve stress granules formed in ADAR1-deficient cells (Fig. S2E). However, we also observed that cells treated with 500 μM sodium arsenite and then allowed to recover in CHX-containing medium resolved their stress granules slower than cells allowed to recover in regular medium (Fig. S2C,D). This can be explained by the fact that sufficient levels of translation initiation are required for CHX to trap mRNAs in association with ribosomes. Thus, removing sodium arsenite and allowing cells to recover efficient translation initiation disassembles stress granules faster than the addition of CHX. This is consistent with another report suggesting that CHX efficiently resolves stress granules in cells with moderate levels of translational repression, but not in cells with high translational repression (Mollet et al., 2008).

ADAR1 regulates stress granules independently of translation

One model for stress granule formation (Van Treeck and Parker, 2018) posits that the formation of stress granules occurs through a summation of protein–protein, protein–RNA and RNA–RNA interactions and that defects in one group of interactions can be compensated for by increasing other interactions. For example, deletion of the key stress granule proteins G3BP1 and G3BP2 (referred to collectively as G3BP1/2) prevents stress granule formation in most stresses despite translational repression (Guillén-Boixet et al., 2020; Kedersha et al., 2016; Sanders et al., 2020; Tourrière et al., 2003; Yang et al., 2020), but stress granules can be restored in G3BP1/2 KO cells by inhibiting the action of eIF4A, which limits RNA–RNA interactions contributing to stress granule assembly (Tauber et al., 2020). We hypothesized that if ADAR1 functions to limit trans RNA–RNA interactions, then perhaps the increased available interactions in ADAR1-deficient cells may increase stress granule formation. One way to test this possibility is to see whether ADAR1 deficiency can rescue stress granules in G3BP1/2 KO cells, which normally do not form stress granules in sodium arsenite (Fig. S3A). To address this issue, we performed siRNA knockdowns of ADAR1 in G3BP1/2 KO cells and assessed stress granule formation.

We did not observe any stress granules in unstressed ADAR1-deficient cells (Fig. 2A), indicating that the formation of stress granules in unstressed ADAR1-deficient cells is dependent on G3BP1/2. However, we did observe that ADAR1 KD resulted in the formation of small PABPC1 foci in the cytoplasm of G3BP1/2 KO cells treated with sodium arsenite (Fig. 2B,C). These PABPC1 foci also contained other stress granule markers, such as TIA-1 and poly(A)+ RNA (Fig. S3B). Interestingly, the stress granule marker eIF4E did not appear to be enriched in these PABPC1 foci (Fig. S3B), perhaps due to the loss of key interactions provided by G3BP1/2. We interpret this to mean that ADAR1 deficiency can induce the formation of small, stress granule-like PABPC1 foci in G3BP1/2 KO cells during sodium arsenite treatment.

Fig. 2.

ADAR1 regulates stress granules in a translation-independent manner. (A) Unstressed G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA and stained for ADAR1 and PABPC1. Scale bars: 20 μm. (B) G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA stressed for 1 h with 500 μM NaAsO2 and stained for ADAR1 and PABPC1. Dashed boxes indicate regions shown in zoom images. Scale bars: 20 μm (C) Quantification of percentage of cells with PABPC1 foci in the experiment shown in B. Data are mean±s.d., 20 images were analyzed for each condition from four independent experiments. (D) Puromycin labeling and IF for PABPC1 in unstressed G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA. Scale bars: 20 μm. (E) Puromycin labeling and IF for PABPC1 in G3BP1/2 KO cells stressed for 1 h with 500 μM NaAsO2 and treated with ADAR1 siRNA or non-targeting siRNA. Scale bars: 20 μm. (F) Quantification of the experiment shown in D and E. Data are mean±s.d., 31 unstressed cells and 24 sodium arsenite-treated cells were analyzed for each condition. (G) ADAR1+/− (clone one) and parental RL KO U-2 OS cells under increased hippuristanol concentrations and stained for G3BP1. Scale bars: 10 μm. (H) Quantification of stress granule volume relative to cytoplasmic volume for the hippuristanol (Hipp) experiment shown in G. Data are mean±s.d., four images analyzed for each condition. (I) ADAR1+/− (clone one and two shown) and the parental cell line under increased NaAsO2 concentrations and stained for G3BP1. Scale bars: 10 μm. (J) Quantification of stress granule volume relative to cytoplasm volume for the experiment shown in I. Data are mean±s.d., 4 images analyzed for each condition. *P<0.05; ***P<0.001; n.s., not significant (unpaired two-tailed t-test). Images are representative of three or more experiments.

Fig. 2.

ADAR1 regulates stress granules in a translation-independent manner. (A) Unstressed G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA and stained for ADAR1 and PABPC1. Scale bars: 20 μm. (B) G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA stressed for 1 h with 500 μM NaAsO2 and stained for ADAR1 and PABPC1. Dashed boxes indicate regions shown in zoom images. Scale bars: 20 μm (C) Quantification of percentage of cells with PABPC1 foci in the experiment shown in B. Data are mean±s.d., 20 images were analyzed for each condition from four independent experiments. (D) Puromycin labeling and IF for PABPC1 in unstressed G3BP1/2 KO cells treated with ADAR1 siRNA or non-targeting siRNA. Scale bars: 20 μm. (E) Puromycin labeling and IF for PABPC1 in G3BP1/2 KO cells stressed for 1 h with 500 μM NaAsO2 and treated with ADAR1 siRNA or non-targeting siRNA. Scale bars: 20 μm. (F) Quantification of the experiment shown in D and E. Data are mean±s.d., 31 unstressed cells and 24 sodium arsenite-treated cells were analyzed for each condition. (G) ADAR1+/− (clone one) and parental RL KO U-2 OS cells under increased hippuristanol concentrations and stained for G3BP1. Scale bars: 10 μm. (H) Quantification of stress granule volume relative to cytoplasmic volume for the hippuristanol (Hipp) experiment shown in G. Data are mean±s.d., four images analyzed for each condition. (I) ADAR1+/− (clone one and two shown) and the parental cell line under increased NaAsO2 concentrations and stained for G3BP1. Scale bars: 10 μm. (J) Quantification of stress granule volume relative to cytoplasm volume for the experiment shown in I. Data are mean±s.d., 4 images analyzed for each condition. *P<0.05; ***P<0.001; n.s., not significant (unpaired two-tailed t-test). Images are representative of three or more experiments.

We did not observe any significant difference in the relative translational activity of G3BP1/2 KO cells treated with ADAR1 siRNA compared to that of cells treated with the non-targeting siRNA in either sodium arsenite-treated or -untreated conditions (Fig. 2D–F). We interpret these results to mean that ADAR1 can limit the formation of stress granules even when translation initiation is strongly repressed. Intriguingly, these results also suggest that G3BP1/2 may have a role in the activation of PKR, as we would have expected ∼5% of unstressed G3BP1/2 KO cells to have reduced translation due to PKR activation induced by ADAR1 deficency. This is consistent with results from reports that G3BP1/2 promotes PKR-mediated phosphorylation of eIF2α (Reineke and Lloyd, 2015; Reineke et al., 2012).

The above results suggest that ADAR1 limits stress granule formation in some manner. Although ADAR1 is thought to primarily edit within inverted Alu repeats that form cis dsRNA structures (Athanasiadis et al., 2004; Chung et al., 2018; Kim et al., 2004), we hypothesized that ADAR1 may limit stress granule assembly by disrupting trans RNA–RNA interactions, akin to the recently described role of the eIF4A1 helicase (Tauber et al., 2020). If ADAR1-deficient cells have an increased propensity to form trans RNA–RNA interactions, they should be more sensitive to eIF4A1 helicase inhibition under conditions in which translation repression is moderate. To test this hypothesis, we treated cells with low doses of hippuristanol, a small molecule that inhibits both the helicase activity and the translation initiation function of eIF4A1.

We observed that ADAR1+/− U-2 OS cells formed more stress granules than the parental (RL KO) cell line upon hippuristanol treatment (Fig. 2G,H). However, we did not observe any difference in stress granule assembly when the cells were treated with sodium arsenite, even at low doses (Fig. 2I,J). Both sodium arsenite and hippuristanol inhibit translation initiation, but only hippuristanol inhibits the helicase function of eIF4A. We interpret the finding that ADAR1-deficient cells are increasingly sensitive to only hippuristanol to mean that ADAR1-deficient cells have a higher propensity for RNA–RNA interactions, which contributes to stress granule formation. Thus, one of the functions of ADAR1 is to limit the formation of stress granules.

In addition, we examined whether ADAR1 KD affects stress granule induction upon treatment with poly(I:C). We did not observe a significant difference in stress granule volume between ADAR1 KD cells and cells treated with the non-targeting siRNA upon poly(I:C) stress (Fig. S3C,D). Even though ADAR1-deficient cells are more prone to PKR activation, this observation could be explained if the amount of poly(I:C) was more than sufficient to robustly activate PKR in the control cells as well.

Cytoplasmic ADAR1 overexpression inhibits stress granule formation

We have observed that ADAR1 limits stress granule formation, and that a deficiency in ADAR1 increases the propensity of cells to form stress granules. This led us to a prediction that overexpression of ADAR1 would prevent stress granule formation. There are two isoforms of ADAR1. The p110 isoform is the predominant isoform under normal conditions and is primarily nuclear, whereas the interferon-inducible p150 isoform, which differs only in the N-terminal region, is primarily cytoplasmic (George and Samuel, 1999). The p150 isoform has been reported to localize to stress granules (Ng et al., 2013; Weissbach and Scadden, 2012) and is induced by interferon produced during conditions such as viral infection. Since we have observed that ADAR1 depletion induces stress granules, we sought to ask whether elevating ADAR1 expression levels would have an inverse effect and reduce stress granule formation. To mimic viral infection inducing the p150 isoform, we transiently transfected a plasmid expressing mGFP–ADAR1 p150 into WT U-2 OS cells and stressed the cells with sodium arsenite.

We observed that cells overexpressing mGFP–ADAR1 p150 formed fewer stress granules than the surrounding non-transfected cells (Fig. 3A,B). This was not observed for similarly transfected cells expressing mGFP (Fig. 3A). This result indicates that cytoplasmic ADAR1 limits the formation of stress granules in the presence of sodium arsenite.

Fig. 3.

Cytoplasmic ADAR1 overexpression inhibits stress granule formation. (A) U-2 OS cells transfected with plasmid expressing mGFP or mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (B) Quantification of stress granule volume for the experiment shown in A, comparing mGFP–ADAR1 p150-transfected and non-transfected cells. Data are mean±s.d., 10 images analyzed from three independent experiments. (C) Quantification of stress granule volume of non-transfected and mApple–ADAR1 p150-expressing RL KO U-2 OS cells transfected with 500 ng/ml poly(I:C) for 4 h. 50 images analyzed. (D) Puromycin labeling of sodium arsenite-stressed U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150 stained for PABPC1 and puromycin. Scale bars: 20 μm. (E) Quantification of relative translational activity for the experiment shown in C. Data are mean±s.d., 10 images analyzed. (F) FISH for poly(A)+ RNA in U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm, (G) Quantification of poly(A)+ RNA content for the experiment shown in F. Data are mean±s.d., 6 images analyzed. (H) Single-molecule FISH for GAPDH mRNA in U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Deconvolved images are shown. Scale bars: 20 μm. (I) U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150 E912A, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (J) Quantification of stress granule volume for the experiment shown in I. Data are mean±s.d., 10 images analyzed from three independent experiments. Data in C are from one experiment. Data in D–H are representative of three or more experiments. **P<0.01; ***P<0.001; n.s., not significant (unpaired two-tailed t-test).

Fig. 3.

Cytoplasmic ADAR1 overexpression inhibits stress granule formation. (A) U-2 OS cells transfected with plasmid expressing mGFP or mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (B) Quantification of stress granule volume for the experiment shown in A, comparing mGFP–ADAR1 p150-transfected and non-transfected cells. Data are mean±s.d., 10 images analyzed from three independent experiments. (C) Quantification of stress granule volume of non-transfected and mApple–ADAR1 p150-expressing RL KO U-2 OS cells transfected with 500 ng/ml poly(I:C) for 4 h. 50 images analyzed. (D) Puromycin labeling of sodium arsenite-stressed U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150 stained for PABPC1 and puromycin. Scale bars: 20 μm. (E) Quantification of relative translational activity for the experiment shown in C. Data are mean±s.d., 10 images analyzed. (F) FISH for poly(A)+ RNA in U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm, (G) Quantification of poly(A)+ RNA content for the experiment shown in F. Data are mean±s.d., 6 images analyzed. (H) Single-molecule FISH for GAPDH mRNA in U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Deconvolved images are shown. Scale bars: 20 μm. (I) U-2 OS cells transfected with plasmid expressing mGFP–ADAR1 p150 E912A, stressed for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (J) Quantification of stress granule volume for the experiment shown in I. Data are mean±s.d., 10 images analyzed from three independent experiments. Data in C are from one experiment. Data in D–H are representative of three or more experiments. **P<0.01; ***P<0.001; n.s., not significant (unpaired two-tailed t-test).

We assessed whether overexpression of mApple-tagged ADAR1 p150 also limited stress granule assembly during dsRNA stress in the form of poly(I:C) treatment. Following 4 h of treatment with 500 ng/ml poly(I:C), we observed that ADAR1 p150-overexpressing cells had fewer stress granules than the surrounding non-transfected cells (Fig. 3C). This suggests that ADAR1 p150 may function to limit stress granule assembly during viral infection, a condition where p150 expression is induced.

ADAR1 overexpression does not prevent translational repression

In principle, ADAR1 overexpression could prevent translational repression during sodium arsenite treatment, which could explain the reduction in stress granules. However, puromycin labeling demonstrated that ADAR1-overexpressing cells had similar levels of translation during sodium arsenite treatment to the surrounding cells that form stress granules (Fig. 3D,E). This demonstrates that ADAR1 repression of stress granules in an overexpression context is independent of translational activity.

An alternative explanation for ADAR1 overexpression inhibiting formation of stress granules is that ADAR1 promotes mRNA degradation, thereby reducing the pool of mRNPs available to form stress granules (Burke et al., 2019). To examine this possibility, we examined the levels of bulk poly(A)+ RNA and GAPDH mRNA using fluorescence in situ hybridization (FISH). We observed no difference in the levels of poly(A)+ RNA or GAPDH mRNAs in cells overexpressing ADAR1 (Fig. 3F–H). This demonstrates that ADAR1 does not inhibit stress granule formation by promoting mRNA degradation.

ADAR1 inhibition of stress granules is independent of catalytic activity but requires RNA-binding domains

Another possible mechanism by which ADAR1 could inhibit stress granule formation is to recognize and deaminate dsRNA duplexes, thereby destabilizing intermolecular RNA–RNA interactions that could contribute to stress granule formation. In this model, the deamination activity of ADAR1 would be required for the inhibition of stress granule formation. To examine this possibility, we compared the effects of WT ADAR1 and a catalytically dead ADAR1 p150 E912A mutant (Lai et al., 1995) on stress granule formation following transient transfection.

An important result was that overexpression of ADAR1 p150 E912A repressed stress granule formation similarly to WT ADAR1 (Fig. 3I,J). This argues that ADAR1 inhibition of stress granules is independent of catalytic activity.

To identify the domains of ADAR1 that are responsible for inhibiting stress granule formation, we generated a series of truncation mutants of ADAR1 tagged with the fluorescent protein mApple, which allowed us to identify the transfected cells (Fig. 4A). We generated a construct lacking the C-terminal catalytic deaminase domain (ΔC-term), a construct lacking the three dsRNA binding domains (ΔdsRBD), a construct solely expressing the dsRNA binding domains (dsRBD) and a construct solely expressing the N-terminal Z-DNA binding domains (N-term) (Fig. 4A). Transfection of the constructs did not induce any spontaneous stress granules to form (Fig. 4B). Transfection of plasmids expressing these proteins and examining their effect on stress granule formation after sodium arsenite exposure revealed the following points. First, we observed that overexpression of ADAR1 lacking the deaminase domain (ΔC-term) still inhibited the formation of stress granules in sodium arsenite (Fig. 4B,C). This observation was consistent with our earlier results showing that catalytic activity of ADAR1 p150 is not necessary to inhibit stress granules (Fig. 3H,I). Second, we observed that deletion of all three dsRNA-binding domains (ΔdsRBD) of ADAR1 prevented ADAR1 from inhibiting sodium arsenite-induced stress granule formation (Fig. 4B,C). This demonstrates that the dsRBDs are necessary for stress granule inhibition by ADAR1. However, we observed that the dsRBDs alone did not inhibit stress granule assembly and, in fact, appeared to enhance stress granule assembly (Fig. 4B,C). Thus, the dsRNA-binding domains of ADAR1 are necessary, but not sufficient, to inhibit stress granules in sodium arsenite. Finally, we addressed whether the N-terminal Z-DNA binding domains (N-term) inhibit stress granule assembly. We observed that ADAR1 p150 N-term alone did not inhibit stress granule assembly (Fig. 4B,C). Taken together, these results indicate that both the Z-DNA-binding domains and the dsRNA-binding domains of ADAR1 p150 are necessary to inhibit stress granule assembly. Moreover, we also observed that all of the constructs, with the exception of the dsRBD construct, were able to localize to stress granules, which was most clearly seen with the N-term construct, because it did not reduce the number of stress granules (Fig. 4B). The localization of all ADAR1 constructs with the N-terminal domains to stress granules is consistent with previous reports that Z-DNA-binding domains can target proteins to stress granules (Ng et al., 2013).

Fig. 4.

RNA-binding, not deaminase activity, is required for ADAR1 stress granule inhibition. (A) Schematic representation of the ADAR1 p150 domains and truncation constructs generated. (B) U-2 OS cells transfected with plasmids expressing the various mApple–ADAR1 p150 constructs in unstressed conditions or following stress for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (C) Quantification of stress granule volume for the experiment shown in B. Data are mean±s.d., 50 images from at least three independent experiments analyzed for each construct. ***P<0.001; n.s., not significant (unpaired two-tailed t-test).

Fig. 4.

RNA-binding, not deaminase activity, is required for ADAR1 stress granule inhibition. (A) Schematic representation of the ADAR1 p150 domains and truncation constructs generated. (B) U-2 OS cells transfected with plasmids expressing the various mApple–ADAR1 p150 constructs in unstressed conditions or following stress for 1 h with 500 μM NaAsO2 and stained for PABPC1. Scale bars: 20 μm. (C) Quantification of stress granule volume for the experiment shown in B. Data are mean±s.d., 50 images from at least three independent experiments analyzed for each construct. ***P<0.001; n.s., not significant (unpaired two-tailed t-test).

We have identified two mechanisms by which ADAR1 negatively regulates stress granule formation. First, ADAR1 prevents the activation of PKR by modifying endogenous dsRNAs. The critical observations supporting this conclusion are that the spontaneous stress granules that form in ADAR1-deficient cells require PKR for their formation and show translation repression (Fig. 1C–E). This role is consistent with previous observations that reduced ADAR1 activity enhances the activation of PKR, leading to phosphorylation of eIF2α, translation repression and stress granule formation (Li et al., 2012; Okonski and Samuel, 2013).

A surprising observation is that stress granules induced by the activation of PKR due to ADAR1 defects are not resolved by ISRIB (Fig. S2B). Since ISRIB reverses the effects of eIF2α phosphorylation, this observation argues that PKR activation leads to a second alteration, presumably a phosphorylation event, that also inhibits translation and leads to stress granule persistence. This suggests that attempts to reverse translation inhibition caused by eIF2α phosphorylation using activators of eIF2B will be unsuccessful when the phosphorylation is due to the activation of PKR.

We present several lines of evidence that ADAR1 also limits stress granules in a manner that is independent of translation repression through PKR. First, ADAR1-deficient cells are more prone to forming stress granules in response to hippuristanol, which inhibits the helicase activity of eIF4A and promotes stress granule assembly (Fig. 2G,H) (Tauber et al., 2020). Second, ADAR1 deficiency results in the formation of stress granule-like PABPC1 foci in G3BP1/2 KO cells during sodium arsenite treatment (Fig. 2B,C; Fig. S2B–D). Finally, overexpression of cytosolic ADAR1 limits the formation of stress granules (Figs 3,4). In all of these cases, alterations in stress granule formation are independent of translation repression (Figs 2E,F and 3D,E).

Several observations suggest that the translation-independent mechanism by which ADAR1 inhibits stress granule formation is through functioning as an RNA-binding protein and not as an RNA deaminase. First, we observed that overexpression of a catalytically dead form of ADAR1, or a deletion variant lacking the deaminase domain, inhibits stress granule formation similar to expression of the WT protein (Figs 3,4). In contrast, deletion variants of ADAR1 lacking either the dsRNA-binding domains or the N-terminal Z-DNA-binding domains no longer inhibit stress granule formation when overexpressed (Fig. 4). This suggests a mechanism in which ADAR1 binding to RNAs prevents their recruitment to stress granules (Fig. 5). In principle, ADAR1 could inhibit stress granule formation by either competing for the binding of RNA binding proteins, such as G3BP1, that promote stress granule formation, or inhibiting intermolecular RNA–RNA interactions either by competing for such interactions, or by stabilizing a more compact form of the mRNP that is less prone to intermolecular interactions (Fig. 5).

Fig. 5.

Model for the role of ADAR1 p150 in inhibition of stress granules during stress. Under typical stress conditions, mRNAs are released from ribosomes and are free to interact with (w/) RNA-binding proteins and other RNAs, leading to the formation of stress granules (SGs). Under conditions of increased ADAR1 p150 expression, such as during viral infection, ADAR1 may compete with other RNA-binding proteins or RNAs for interactions with mRNAs and/or reorganize the organization of mRNPs, inhibiting the formation of stress granules. Created with BioRender.com.

Fig. 5.

Model for the role of ADAR1 p150 in inhibition of stress granules during stress. Under typical stress conditions, mRNAs are released from ribosomes and are free to interact with (w/) RNA-binding proteins and other RNAs, leading to the formation of stress granules (SGs). Under conditions of increased ADAR1 p150 expression, such as during viral infection, ADAR1 may compete with other RNA-binding proteins or RNAs for interactions with mRNAs and/or reorganize the organization of mRNPs, inhibiting the formation of stress granules. Created with BioRender.com.

Since many viruses form RNA-dense replication factories resembling stress granules (Castro et al., 2017; Fernández de Castro et al., 2021), further investigation of these non-canonical mechanisms by which ADAR1 disrupts RNP complexes is warranted in the context of viral replication. ADAR1 may bind to viral dsRNAs and prevent the formation of these replication factories, thereby inhibiting viral replication. The effect of ADAR1 on both PKR and stress granule formation may also have impacts on viral infections. Some viruses modulate stress granule formation because they rely on host translational machinery for their replication, and stress granules may sequester translation initiation factors (Miller, 2011). The induction of the cytosolic p150 isoform of ADAR1 during viral infections (George and Samuel, 1999; Vogel et al., 2020) could thereby inhibit the formation of stress granules in two manners, impacting the viral infection. ADAR1 is thought to have both pro-viral and anti-viral functions, depending on the virus (Samuel, 2011). Whether cytoplasmic ADAR1 inhibition of stress granules has pro-viral or anti-viral effects remains to be determined and may depend on the specific virus being investigated.

Plasmids

pmGFP and pmGFP-ADAR1 p150 were from Addgene (deposited by Kumiko Ui-Tei; Addgene 117926 and 117927, respectively; RRID:Addgene_117926 and RRID:Addgene_117927). pmGFP-ADAR1 p150 E912A was generated from pmGFP-ADAR1 p150 using QuikChange II Mutagenesis (Agilent, 200521) according to the manufacturer's protocol and with the following primers: forward, 5′-ACTGTCAATGACTGCCATGCAGCAATAATCTCCCGGAGAGGCTTC-3′; reverse, 5′-GAAGCCTCTCCGGGAGATTATTGCTGCATGGCAGTCATTGACAGT-3′. pGW1-mApple was a gift from Dr Sami Barmada (Barmada et al., 2014). pGW1-mApple-ADAR1 p150 was generated using In-Fusion HD Cloning (Takara Bio, 638947) with the following primers: forward, 5′-GACGAGCTGTACAAGATGAATCCGCGGCAGGGG-3′; reverse, 5′-GATCCGGTGGATCCCCTATACTGGGCAGAGATAAAAGTTC-3′. pGW1-mApple was linearized using SmaI (NEB, R0141S). pGW1-mApple-ADAR1 p150 ΔC-term was generated from pGW1-mApple-ADAR1 p150 using the following primers: forward, 5′-GACGAGCTGTACAAGATGAATCCGCGGCAGGGG-3′; reverse, 5′-GATCCGGTGGATCCCCTACCCAATCAAGACACG-3′. pGW1-mApple was linearized using SmaI. pGW1-mApple-ADAR1 p150 dsRBD was generated from pGW1-mApple-ADAR1 p150 using the following primers: forward, 5′-GACGAGCTGTACAAGAACCCCATCAGCGGG-3′; reverse, 5′-GATCCGGTGGATCCCCTACCCAATCAAGACACG-3′. pGW1-mApple was linearized using SmaI. pGW1-mApple-ADAR1 p150 Δ dsRBD was generated from pGW1-mApple-ADAR1 p150 using the following primers: forward, 5′-GACGAGCTGTACAAGATGAATCCGCGGCAGGGG-3′; reverse, 5′-TGCCTTCTCGTTCTCCTTCAGCTGGCACTC-3′. pGW1-mApple was linearized using SwaI (NEB, R0604S). pGW1-mApple-ADAR1 p150 N-term was generated from pGW1-mApple-ADAR1 p150 using the following primers: forward, 5′-GACGAGCTGTACAAGAACCCCATCAGCGGG-3′; reverse, 5′-GATCCGGTGGATCCCCTACTTCAGCTGGCACTC-3′. pGW1-mApple was linearized using SmaI. All plasmids were validated by Sanger sequencing from Quintarabio with the following sequencing primers: mApple_Seq_FWD, 5′-ACACCATCGTGGAACAGTACG-3′; mApple_Seq_REV, 5′-CAAATGTGGTATGGCTGAT-3′; p150seq_Fwd1, 5′-GAGGTCTATATAAGCAGAGC-3′; p150seq_Fwd2, 5′-GACTCTGAGGACATGGGTGT-3′; p150seq_Fwd3, 5′-CTACAGTCATGGCTTGCCACG-3′; p150seq_Rev1, 5′-CAGATGGAATGACACAGTC-3′; p150seq_Rev2, 5′-GTCACTGGGGTTACCTCTGTG-3′.

Cell culture and drug treatment

The parental A549 cell line was provided by Dr Christopher Sullivan (University of Texas at Austin, TX, USA; Burke et al., 2016). The PKR KO A549 cell line was generated as described previously (Burke et al., 2019). The WT, GFP–G3BP1 and G3BP1/2 KO U-2 OS cells were provided by Dr Nancy Kedersha and Dr Paul Anderson (Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA; Kedersha et al., 2008, 2016). The RNase L KO U-2 OS cell line was generated as described previously (Burke et al., 2019). All cell lines were maintained at 5% CO2 at 37°C in Dulbecco's Modified Eagle Medium (DMEM; Thermo Fisher Scientific, 12800-082) supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich, F0926, Lot 18L047) and 1% penicillin-streptomycin (Thermo Fisher Scientific, 15140-122). Cells were routinely tested for mycoplasma contamination and were negative.

Hippuristanol (a gift from Dr Jerry Pelletier, McGill University, Montreal, Canada) was diluted in DMSO. Sodium arsenite (Millipore Sigma, S7400) was diluted in UltraPure DNase/RNase-Free Distilled Water (Thermo Fisher Scientific, 10977-023). Hippuristanol and sodium arsenite treatments were performed for 1 h at 37°C. Puromycin (Sigma-Aldrich, P8833-100MG) was diluted in UltraPure DNase/RNase-Free Distilled Water and added to cells at 5 μg/ml for 10 min for puromycin labeling. ISRIB (Fisher Scientific, 52-841-0) was prepared in DMSO, and treatment was done for 15 min. Poly(I:C) (InvivoGen, tlrl-pic) was resuspended in UltraPure DNase/RNase-Free Distilled Water and transfected into cells using Lipofectamine 2000 reagent (Thermo Fisher Scientific, 11668030) according to the manufacturer's protocol. 3 μl Lipofectamine was used for 1 μg poly(I:C). Cycloheximide (Sigma-Aldrich, C7698-1G) was prepared in DMSO and treatment was performed for 1 h.

siRNA knockdowns and plasmid transfections

siRNA knockdowns were performed using Lipofectamine RNAiMAX (Thermo Fisher Scientific, 13778075) according to the manufacturer's protocol for 48 h. The following siRNAs were used: ADAR1 (Thermo Fisher Scientific, 4390824), negative control (Thermo Fisher Scientific, AM4611). Plasmids were transfected using jetPRIME transfection reagent (VWR Scientific, 89129-920) according to the manufacturer's protocol. Cell medium was replaced 2 h post transfection. Cells were fixed for imaging 24 h post transfection.

Generation of ADAR1+/− cell lines

Knockout cell lines were generated as described previously (Burke et al., 2019, 2020). The following sgRNA sequences were used for ADAR1: sense, 5′-CACCGTCTGTCAAATGCCATATGGG-3′; antisense, 5′-AAACCCCATATGGCATTTGACAGAC-3′.

Immunofluorescence

Cells were seeded on glass-bottom 96-well plates (Thermo Fisher Scientific, NC0536760) or glass coverslips (Thomas Scientific, 1217N79; Fisher Scientific, 08-774-383; or Fisher Scientific, 12-541A). Cells were fixed with 4% paraformaldehyde (PFA) for 10 min at room temperature. Cells were then permeabilized with phosphate-buffered saline (PBS) with 0.1% (v/v) Triton X-100 for 5 min at room temperature. Coverslips or glass plates were then blocked for 1 h at room temperature with PBS containing 0.1% (v/v) Triton X-100 and 5% (w/v) bovine serum albumin (BSA) (Millipore Sigma, 26593-100GM). Antibodies were diluted in PBS. The following antibodies were used at 1:500 dilutions: anti-G3BP1 (Abcam, ab56574), anti-PABPC1 (Abcam, ab21060), anti-ADAR1 (Abcam, ab88574), anti-FMRP (Abcam, ab17722), anti-TIA-1 (Abcam, ab40693), anti-eIF4G1 (Proteintech, 15704-1-AP) and anti-eIF4E (Abcam, ab33766) . The anti-puromycin antibody (Millipore Sigma MABE343) was used at 1:1000. The following secondary antibodies were used at 1:1000: FITC-conjugated goat anti-rabbit IgG (Abcam, ab6717), Alexa Fluor 647-conjugated goat anti-rabbit IgG (Abcam, ab150079), Alexa Fluor 647-conjugated goat anti-mouse IgG (Abcam, ab150115), FITC-conjugated goat anti-mouse IgG (Abcam, ab6785) and Alexa Fluor 555-conjugated donkey anti-rabbit IgG (Abcam, ab150062).

Sequential immunofluorescence and FISH

Sequential immunofluorescence and FISH using NORAD, GAPDH and poly d(T) probes was performed as described previously (Khong et al., 2018). Stellaris FISH Probes, human GAPDH with Quasar 570 Dye (Stellaris, SMF-2026-1) was used for GAPDH single-molecule FISH. Custom NORAD probes were as described previously (Khong et al., 2017). Poly d(T) probes were ordered from Integrated DNA Technologies as 30 deoxythymidines with a 5′ Cy5 modification.

SDS–PAGE and western blotting

Cells were lysed in Lysis Buffer 6 (R&D Systems, 895561) supplemented with a cOmplete Protease Inhibitor tablet (Sigma-Aldrich, 11697498001). Lysed cells were freeze thawed to complete lysis and incubated at 95°C for 5 min prior to running on a NuPAGE 4–12% Bis-Tris Gel (Thermo Fisher Scientific, NP0321BOX) and transfer to a nitrocellulose membrane (GE Healthcare, 45004004). Membranes were blocked in Tris-buffered saline with 0.1% Tween-20 (TBS-T) with 5% milk or in PBS with 5% BSA for 1 h at room temperature. Membranes were incubated with primary antibodies for 1 h at room temperature or overnight at 4°C in PBS with 5% BSA (GAPDH) or TBS-T with 5% milk (ADAR1), washed three times with TBS-T, followed by incubation with secondary antibodies for 1 h at room temperature and washing again three times with TBS-T. Membranes were then incubated with the chemiluminescence substrate (Thermo Fisher Scientific, 34076) for 5 min prior to imaging with an ImageQuant LAS 4000 (GE Healthcare). Membranes were then stripped with stripping buffer (Thermo Fisher Scientific, 21059), re-blocked and incubated with antibody as described above. The ADAR1 antibody (Cell Signaling Technology, 14175S) and GAPDH–HRP antibody (Santa Cruz, sc-47724 HRP) were used at 1:1000. The anti-rabbit IgG HRP-linked secondary antibody (Cell Signaling Technology, 7074S) and anti-mouse IgG HRP-linked secondary antibody (Cell Signaling Technology, 7076S) were used at 1:2000.

Microscopy and image analysis

Live-cell imaging was performed on a Nikon Spinning-Disk Confocal microscope with a 2× Andor Ultra 888 EMCCD camera and a 40× air objective. Fixed cells were imaged on a widefield DeltaVision Elite microscope with a 100× oil objective and PCO Edge sCMOS camera. 10–20 z slices were imaged at 0.2 μm distance between slices. All images shown are of a single z plane unless otherwise specified in the figure legend. Deconvolved images are indicated in the figure legend where shown.

Puromycin labeling and poly(A)+ RNA content quantification was performed using ImageJ (NIH). A line was drawn through the cytoplasm of each cell, and the average intensity of the line was obtained. Relative translation activity or poly(A)+ content was determined by comparison of the relative average signal intensities.

Stress granule quantification was performed using Bitplane Imaris image analysis software using the following steps (also described in Khong et al., 2018): (1) Open image in Imaris. (2) Start surface creation wizard. (3) Select parameters to define nuclei (DAPI stain) as a surface. (4) Mask all channels – set voxels inside surface to 0. (5) Start cell creation wizard. (6) Select parameters to define stress granules (PABPC1 or G3BP1 channel). (7) Export all statistics to Microsoft Excel file. (8) Start cell creation wizard. (9) Select parameters to define cytoplasm (PABPC1 or G3BP1 channel). (10) Export all statistics to Excel file. (11) Quantify stress granule volume by taking the ratio of total stress granule volume to cytoplasm volume in each image.

For ADAR1 overexpression experiments, the following steps were also performed: (1) Start surface creation wizard. (2) Select parameters to define ADAR1-overexpressing cells (mGFP or mApple channel) as a surface. (3) Mask all channels – set voxels outside surface to 0. (3) Start cell creation wizard. (4) Select parameters to define stress granules in the masked channel using same parameters as for previous steps (PABPC1 or G3BP1 channel). (5) Export all statistics to Excel file. (6) Start cell creation wizard. (7) Select parameters to define cytoplasm in the masked channel using same parameters as for previous steps (PABPC1 or G3BP1 channel). (8) Export all statistics to Excel file. (9) Subtract ADAR1-overexpressing cytoplasm volume from total cytoplasm volume, do the same for stress granule volume. (10) Obtain relative stress granule volume using the formula:
formula
where SGADAR1 is the stress granule volume of the ADAR1-expressing cells, CADAR1 is the cytoplasm volume of the ADAR1-expressing cells, SGtotal is the total stress granule volume and Ctotal is the total cytoplasm volume.

Statistical analysis was performed using an unpaired two-tailed t-test. For P>0.05, analyses were designated as not significant. The number of images used for each analysis is indicated in the figure legend. Images represent distinct fields of view from the same experiment unless specified in the figure legend.

We thank Dr Nancy Kedersha (Brigham and Women's Hospital, Harvard Medical School) and Dr Paul Anderson (Brigham and Women's Hospital, Harvard Medical School) for the parental, G3BP1/2-KO, and GFP–G3BP1 U-2 OS cell lines. We thank Dr Jerry Pelletier (McGill University) for the hippuristanol. The imaging work was performed at the BioFrontiers Institute Advanced Light Microscopy Core (RRID: SCR_018302). Spinning-disk confocal microscopy was performed on a Nikon Ti-E microscope supported by the BioFrontiers Institute and the Howard Hughes Medical Institute.

Author contributions

Conceptualization: G.A.C., J.M.B., R.P.; Methodology: G.A.C., J.M.B.; Validation: G.A.C.; Formal analysis: G.A.C.; Investigation: G.A.C.; Resources: R.P.; Writing - original draft: G.A.C.; Writing - review & editing: G.A.C., J.M.B., R.P.; Visualization: G.A.C.; Supervision: R.P.; Project administration: G.A.C.; Funding acquisition: R.P.

Funding

This work was supported by funds from the National Institutes of Health (GM045443) and Howard Hughes Medical Institute. Deposited in PMC for release after 12 months.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258783

Athanasiadis
,
A.
,
Rich
,
A.
and
Maas
,
S.
(
2004
).
Widespread A-to-I RNA editing of alu-containing mRNAs in the human transcriptome
.
PLoS Biol.
2
,
e391
.
Barmada
,
S. J.
,
Serio
,
A.
,
Arjun
,
A.
,
Bilican
,
B.
,
Daub
,
A.
,
Ando
,
D. M.
,
Tsvetkov
,
A.
,
Pleiss
,
M.
,
Li
,
X.
,
Peisach
,
D.
et al. 
(
2014
).
Autophagy induction enhances TDP43 turnover and survival in neuronal ALS models
.
Nat. Chem. Biol.
10
,
677
-
685
.
Bass
,
B. L.
(
2002
).
RNA editing by adenosine deaminases that act on RNA
.
Annu. Rev. Biochem.
71
,
817
-
846
.
Bass
,
B. L.
and
Weintraub
,
H.
(
1988
).
An unwinding activity that covalently modifies its double-stranded RNA substrate
.
Cell
55
,
1089
-
1098
.
Berchtold
,
D.
,
Battich
,
N.
and
Pelkmans
,
L.
(
2018
).
A systems-level study reveals regulators of membrane-less organelles in human cells
.
Mol. Cell
72
,
1035
-
1049.e5
.
Buchan
,
J. R.
and
Parker
,
R.
(
2009
).
Eukaryotic stress granules: the ins and outs of translation
.
Mol. Cell
36
,
932
-
941
.
Burke
,
J. M.
,
Kincaid
,
R. P.
,
Nottingham
,
R. M.
,
Lambowitz
,
A. M.
and
Sullivan
,
C. S.
(
2016
).
DUSP11 activity on triphosphorylated transcripts promotes Argonaute association with noncanonical viral microRNAs and regulates steady-state levels of cellular noncoding RNAs
.
Genes Dev.
30
,
2076
-
2092
.
Burke
,
J. M.
,
Moon
,
S. L.
,
Matheny
,
T.
and
Parker
,
R.
(
2019
).
RNase L reprograms translation by widespread mRNA turnover escaped by antiviral mRNAs
.
Mol. Cell
75
,
1203
-
1217.e5
.
Burke
,
J. M.
,
Lester
,
E. T.
,
Tauber
,
D.
and
Parker
,
R.
(
2020
).
RNase L promotes the formation of unique ribonucleoprotein granules distinct from stress granules
.
J. Biol. Chem.
295
,
1426
-
1438
.
de Castro
,
I. F.
,
Fernández
,
J. J.
,
Barajas
,
D.
,
Nagy
,
P. D.
and
Risco
,
C.
(
2017
).
Three-dimensional imaging of the intracellular assembly of a functional viral RNA replicase complex
.
J. Cell Sci.
130
,
260
-
268
.
Chung
,
H.
,
Calis
,
J. J. A.
,
Wu
,
X.
,
Sun
,
T.
,
Yu
,
Y.
,
Sarbanes
,
S. L.
,
Dao Thi
,
V. L.
,
Shilvock
,
A. R.
,
Hoffmann
,
H.-H.
,
Rosenberg
,
B. R.
et al. 
(
2018
).
Human ADAR1 prevents endogenous RNA from triggering translational shutdown
.
Cell
172
,
811
-
824.e14
.
Fernández de Castro
,
I.
,
Tenorio
,
R.
and
Risco
,
C.
(
2021
).
Virus factories
.
Encycl. Virol.
1
,
495
-
500
.
George
,
C. X.
and
Samuel
,
C. E.
(
1999
).
Human RNA-specific adenosine deaminase ADAR1 transcripts possess alternative exon 1 structures that initiate from different promoters, one constitutively active and the other interferon inducible
.
Proc. Natl. Acad. Sci. USA
96
,
4621
-
4626
.
Guillén-Boixet
,
J.
,
Kopach
,
A.
,
Holehouse
,
A. S.
,
Wittmann
,
S.
,
Jahnel
,
M.
,
Schlüßler
,
R.
,
Kim
,
K.
,
Trussina
,
I. R. E. A.
,
Wang
,
J.
,
Mateju
,
D.
et al. 
(
2020
).
RNA-induced conformational switching and clustering of G3BP drive stress granule assembly by condensation
.
Cell
181
,
346
-
361.e17
.
Hofmann
,
S.
,
Kedersha
,
N.
,
Anderson
,
P.
and
Ivanov
,
P.
(
2021
).
Molecular mechanisms of stress granule assembly and disassembly
.
Biochim. Biophys. Acta (BBA) Mol. Cell Res.
1868
,
118876
.
Jain
,
S.
,
Wheeler
,
J. R.
,
Walters
,
R. W.
,
Agrawal
,
A.
,
Barsic
,
A.
and
Parker
,
R.
(
2016
).
ATPase-modulated stress granules contain a diverse proteome and substructure
.
Cell
164
,
487
-
498
.
Kedersha
,
N.
and
Anderson
,
P.
(
2007
).
Mammalian stress granules and processing bodies
.
Methods Enzymol.
431
,
61
-
81
.
Kedersha
,
N. L.
,
Gupta
,
M.
,
Li
,
W.
,
Miller
,
I.
and
Anderson
,
P.
(
1999
).
RNA-binding proteins Tia-1 and tiar link the phosphorylation of Eif-2α to the assembly of mammalian stress granules
.
J. Cell Biol.
147
,
1431
-
1442
.
Kedersha
,
N.
,
Cho
,
M. R.
,
Li
,
W.
,
Yacono
,
P. W.
,
Chen
,
S.
,
Gilks
,
N.
,
Golan
,
D. E.
and
Anderson
,
P.
(
2000
).
Dynamic shuttling of Tia-1 accompanies the recruitment of mRNA to mammalian stress granules
.
J. Cell Biol.
151
,
1257
-
1268
.
Kedersha
,
N.
,
Tisdale
,
S.
,
Hickman
,
T.
and
Anderson
,
P.
(
2008
).
Real-time and quantitative imaging of mammalian stress granules and processing bodies
.
Methods Enzymol.
448
,
521
-
552
.
Kedersha
,
N.
,
Panas
,
M. D.
,
Achorn
,
C. A.
,
Lyons
,
S.
,
Tisdale
,
S.
,
Hickman
,
T.
,
Thomas
,
M.
,
Lieberman
,
J.
,
McInerney
,
G. M.
,
Ivanov
,
P.
et al. 
(
2016
).
G3BP-Caprin1-USP10 complexes mediate stress granule condensation and associate with 40S subunits
.
J. Cell Biol.
212
,
e201508028
.
Khong
,
A.
,
Matheny
,
T.
,
Jain
,
S.
,
Mitchell
,
S. F.
,
Wheeler
,
J. R.
and
Parker
,
R.
(
2017
).
The stress granule transcriptome reveals principles of mRNA accumulation in stress granules
.
Mol. Cell
68
,
808
-
820.e5
.
Khong
,
A.
,
Jain
,
S.
,
Matheny
,
T.
,
Wheeler
,
J. R.
and
Parker
,
R.
(
2018
).
Isolation of mammalian stress granule cores for RNA-Seq analysis
.
Methods
137
,
49
-
54
.
Kim
,
D. D. Y.
,
Kim
,
T. T. Y.
,
Walsh
,
T.
,
Kobayashi
,
Y.
,
Matise
,
T. C.
,
Buyske
,
S.
and
Gabriel
,
A.
(
2004
).
Widespread RNA editing of embedded alu elements in the human transcriptome
.
Genome Res.
14
,
1719
-
1725
.
Lai
,
F.
,
Drakas
,
R.
and
Nishikura
,
K.
(
1995
).
Mutagenic analysis of double-stranded RNA adenosine deaminase, a candidate enzyme for RNA editing of glutamate-gated Ion channel transcripts
.
J. Biol. Chem.
270
,
17098
-
17105
.
Lamers
,
M. M.
,
van den Hoogen
,
B. G.
and
Haagmans
,
B. L.
(
2019
).
ADAR1: “Editor-in-Chief” of cytoplasmic innate immunity
.
Front. Immunol.
10
,
1763
.
Lemaire
,
P. A.
,
Anderson
,
E.
,
Lary
,
J.
and
Cole
,
J. L.
(
2008
).
Mechanism of PKR Activation by dsRNA
.
J. Mol. Biol.
381
,
351
-
360
.
Li
,
Z.
,
Okonski
,
K. M.
and
Samuel
,
C. E.
(
2012
).
Adenosine Deaminase Acting on RNA 1 (ADAR1) suppresses the induction of interferon by measles virus
.
J. Virol.
86
,
3787
-
3794
.
Li
,
Y.
,
Banerjee
,
S.
,
Goldstein
,
S. A.
,
Dong
,
B.
,
Gaughan
,
C.
,
Rath
,
S.
,
Donovan
,
J.
,
Korennykh
,
A.
,
Silverman
,
R. H.
and
Weiss
,
S. R.
(
2017
).
Ribonuclease L mediates the cell-lethal phenotype of double-stranded RNA editing enzyme ADAR1 deficiency in a human cell line
.
eLife
6
,
e25687
.
Miller
,
C. L.
(
2011
).
Stress granules and virus replication
.
Future Virol.
6
,
1329
-
1338
.
Mollet
,
S.
,
Cougot
,
N.
,
Wilczynska
,
A.
,
Dautry
,
F.
,
Kress
,
M.
,
Bertrand
,
E.
and
Weil
,
D.
(
2008
).
Translationally repressed mRNA transiently cycles through stress granules during stress
.
Mol. Biol. Cell
19
,
4469
-
4479
.
Ng
,
S. K.
,
Weissbach
,
R.
,
Ronson
,
G. E.
and
Scadden
,
A. D. J.
(
2013
).
Proteins that contain a functional Z-DNA-binding domain localize to cytoplasmic stress granules
.
Nucleic Acids Res.
41
,
9786
-
9799
.
Ohn
,
T.
and
Anderson
,
P.
(
2010
).
The role of posttranslational modifications in the assembly of stress granules
.
Wiley Interdiscip. Rev. RNA
1
,
486
-
493
.
Okonski
,
K. M.
and
Samuel
,
C. E.
(
2013
).
Stress granule formation induced by measles virus is protein kinase PKR dependent and impaired by RNA adenosine deaminase ADAR1
.
J. Virol.
87
,
756
-
766
.
Protter
,
D. S. W.
and
Parker
,
R.
(
2016
).
Principles and properties of stress granules
.
Trends Cell Biol.
26
,
668
-
679
.
Proud
,
C. G.
(
2005
).
eIF2 and the control of cell physiology
.
Semin. Cell Dev. Biol.
16
,
3
-
12
.
Rabouw
,
H. H.
,
Langereis
,
M. A.
,
Anand
,
A. A.
,
Visser
,
L. J.
,
de Groot
,
R. J.
,
Walter
,
P.
and
van Kuppeveld
,
F. J. M.
(
2019
).
Small molecule ISRIB suppresses the integrated stress response within a defined window of activation
.
Proc. Natl. Acad. Sci. USA
116
,
2097
-
2102
.
Reineke
,
L. C.
and
Lloyd
,
R. E.
(
2015
).
The stress granule protein G3BP1 recruits protein kinase R To promote multiple innate immune antiviral responses
.
J. Virol.
89
,
2575
-
2589
.
Reineke
,
L. C.
,
Dougherty
,
J. D.
,
Pierre
,
P.
and
Lloyd
,
R. E.
(
2012
).
Large G3BP-induced granules trigger eIF2α phosphorylation
.
Mol. Biol. Cell
23
,
3499
-
3510
.
Samuel
,
C. E.
(
1979
).
Mechanism of interferon action: phosphorylation of protein synthesis initiation factor eIF-2 in interferon-treated human cells by a ribosome-associated kinase processing site specificity similar to hemin-regulated rabbit reticulocyte kinase
.
Proc. Natl. Acad. Sci. USA
76
,
600
-
604
.
Samuel
,
C. E.
(
2011
).
Adenosine Deaminases Acting on RNA (ADARs) are both antiviral and proviral
.
Virology
411
,
180
-
193
.
Sanders
,
D. W.
,
Kedersha
,
N.
,
Lee
,
D. S. W.
,
Strom
,
A. R.
,
Drake
,
V.
,
Riback
,
J. A.
,
Bracha
,
D.
,
Eeftens
,
J. M.
,
Iwanicki
,
A.
,
Wang
,
A.
et al. 
(
2020
).
Competing protein-RNA interaction networks control multiphase intracellular organization
.
Cell
181
,
306
-
324.e28
.
Sidrauski
,
C.
,
McGeachy
,
A. M.
,
Ingolia
,
N. T.
and
Walter
,
P.
(
2015
).
The small molecule ISRIB reverses the effects of eIF2α phosphorylation on translation and stress granule assembly
.
eLife
4
,
e05033
.
Tauber
,
D.
,
Tauber
,
G.
,
Khong
,
A.
,
Van Treeck
,
B.
,
Pelletier
,
J.
and
Parker
,
R.
(
2020
).
Modulation of RNA condensation by the DEAD-box protein eIF4A
.
Cell
180
,
411
-
426.e16
.
Tourrière
,
H.
,
Chebli
,
K.
,
Zekri
,
L.
,
Courselaud
,
B.
,
Blanchard
,
J. M.
,
Bertrand
,
E.
and
Tazi
,
J.
(
2003
).
The RasGAP-associated endoribonuclease G3BP assembles stress granules
.
J. Cell Biol.
160
,
823
-
831
.
Van Treeck
,
B.
and
Parker
,
R.
(
2018
).
Emerging roles for intermolecular RNA-RNA interactions in RNP assemblies
.
Cell
174
,
791
-
802
.
Vogel
,
O. A.
,
Han
,
J.
,
Liang
,
C.-Y.
,
Manicassamy
,
S.
,
Perez
,
J. T.
and
Manicassamy
,
B.
(
2020
).
The p150 isoform of ADAR1 blocks sustained RLR signaling and Apoptosis during influenza virus infection
.
PLoS Pathog.
16
,
e1008842
.
Weissbach
,
R.
and
Scadden
,
A. D. J.
(
2012
).
Tudor-SN and ADAR1 are components of cytoplasmic stress granules
.
RNA
18
,
462
-
471
.
Yang
,
P.
,
Mathieu
,
C.
,
Kolaitis
,
R.-M.
,
Zhang
,
P.
,
Messing
,
J.
,
Yurtsever
,
U.
,
Yang
,
Z.
,
Wu
,
J.
,
Li
,
Y.
,
Pan
,
Q.
et al. 
(
2020
).
G3BP1 is a tunable switch that triggers phase separation to assemble stress granules
.
Cell
181
,
325
-
345.e28
.

Competing interests

R.P. is a founder and consultant for Faze Medicines.

Supplementary information