Distinct dsRNA sensors are activated during infection with Ad5 mutants

We recently showed that infection with splicing-defective Ad5 mutants containing a deletion of the E4 region (∆E4), which encodes viral proteins required for efficient processing of viral late transcripts, leads to accumulation of viral dsRNA, as detected by immunofluorescence staining and RIP-seq using the dsRNA-specific J2 monoclonal antibody [45]. This is accompanied by phosphorylation of dsRNA-activated kinase PKR during ∆E4 infection [44,45]. Additionally, we showed that infection with Ad5 mutants lacking the two VA RNAs (∆VA) does not lead to detectable dsRNA formation with J2 antibody staining, despite prior literature which established that infection by this mutant results in PKR activation [50,51]. To validate these observations further, we performed immunofluorescence using the dsRNA-specific monoclonal antibody 9D5, which is reportedly more sensitive than the widely used antibody J2 [58]. We evaluated dsRNA formation in A549 cells infected with wild-type (WT) Ad5, or the ∆VA and ∆E4 mutants (Fig 1A). Staining revealed faint cytoplasmic puncta representing mitochondrial dsRNA in uninfected cells, whereas large cytoplasmic aggregates were detected in cells transfected with the dsRNA analog poly(I:C) [32,59]. In agreement with our recently reported results [45], robust nuclear staining was visualized in cells infected with the ∆E4 mutant, while no positive signal was detected for infections with Ad5 WT or ∆VA mutant (Fig 1A). Infection with mutants was further validated for the expression of VA RNAs I and II by primer extension analysis and expression of viral early and late proteins by immunoblotting (Fig 1B and 1C). As expected, ΔE4 infection showed strongly reduced late viral protein expression compared with Ad5 WT and ΔVA infection, consistent with the known replication and RNA-processing defects of this mutant [43,45,60–62]. However, comparable expression levels of the early viral protein DBP were achieved, suggesting similar levels on infection. Under these conditions, activation of the PKR pathway was confirmed by phosphorylation of PKR and eIF2α, using poly(I:C) transfection as a positive control (Fig 1C). These data demonstrate that although infection by both ∆VA and ∆E4 mutants induces PKR activation, formation of detectable dsRNA occurs only during ∆E4 infection.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Activation of different dsRNA sensors during infection with Ad5 mutants.

(A) A549 cells were transfected with poly(I:C) at 1 µg/mL for 6 h, mock-infected, or infected with Ad5 WT or ∆VA and ∆E4 mutants at an MOI of 10 for 48 h. Fixed cells were probed with the dsRNA-specific monoclonal antibody 9D5 (green), showing bright cytoplasmic staining upon poly(I:C) transfection and nuclear staining during infection with the ∆E4 mutant. The viral protein DBP (magenta) marks viral replication centers (VRCs). Nuclei were stained with DAPI (blue) and outlined by dashed white lines in merged images. Scale bar = 15 µm. (B) Whole cell RNA was prepared over a time course of infection and analyzed by PAGE. VA RNA expression was detected by primer extension analysis with fluorescently labeled, VA-specific primers. (C) Time-course immunoblot analysis of A549 cells infected with Ad5 WT or ∆VA and ∆E4 mutants. Poly(I:C) transfection was used as a positive control. PKR pathway activation was measured by autophosphorylation of PKR and phosphorylation of its downstream target eIF2α. RLR pathway activation was assessed by phosphorylation of IRF3. Viral proteins were detected with antibodies to early (DBP and E4orf6) and late proteins (Hexon, Penton, and Fiber). GAPDH was used as loading control. (D,E) Alternatively, RLR activation was measured by RT-qPCR for (D) IRF7 and (E) IFNB1 transcripts. A549 cells were infected or treated as in (C) and total RNA was extracted at 40 hpi and reverse transcribed. GAPDH-normalized data from three independent biological replicates are shown. Bars represent mean and error bars indicate standard deviation, with paired replicate values indicated by colored dots. Significance was calculated by one-way ANOVA followed by Dunnett’s multiple comparisons test, with * P < 0.05, ** P < 0.01 and non-significant results not displayed. (F) Total RNA was extracted from A549 cells treated as in (C) and run on BioAnalyzer to assess RNA integrity and RNase L activity. Results show ribosomal RNA cleavage following poly(I:C) transfection or infection with ∆E4, but not with WT or ∆VA mutant. (G) Alternatively, A549 lacking individual OAS isoforms (O1-3) or RNase L (RL) were infected with Ad5 WT or ∆VA and ∆E4 mutants. Total RNA was collected at 40 hpi and run on BioAnalyzer. Ribosomal RNA degradation indicates ∆E4-induced RNA cleavage is dependent on OAS3 and RNase L.

https://doi.org/10.1371/journal.ppat.1014452.g001

We next asked whether other dsRNA sensor pathways are activated during infection with WT or mutant viruses. We assessed RLR pathway activation by immunoblot analysis for IRF3 phosphorylated at serine 386, and RT-qPCR for measurement of IRF7 and IFNB1 transcripts. As expected, poly(I:C) transfection robustly induced IRF3 phosphorylation, consistent with previous reports [7–9]. In comparison, phosphorylated IRF3 was barely detectable in cells infected with either Ad5 WT or mutant viruses (Fig 1C). Similarly, while poly(I:C) transfection strongly increased expression of IRF7 and IFNB1, infection with the ∆VA mutant led to only modest increase in the expression of IFNB1, consistent with PKR activation (Fig 1D and 1E) [63]. The lack of induction of either IRF3 or IRF7 suggests that the RLR pathway is not activated during infection with Ad5 WT or the ∆VA and ∆E4 mutants.

Activation of the OAS1–3/RNase L pathway leads to characteristic degradation of 18S and 28S rRNAs, which can be visualized by bioanalyzer [18,19,64]. To determine whether this degradation occurs during infection with Ad5 mutants, total cellular RNA was extracted from infected cells at different timepoints or following poly(I:C) transfection, and samples were analyzed for rRNA integrity (Fig 1F). In line with prior observations, poly(I:C) transfection induced evident ribosomal RNA degradation [64]. RNA degradation was observed in late-stage ∆E4-infected samples but was absent during WT and ∆VA infections. Additionally, RNA degradation was completely abrogated during ∆E4 infection in A549 cells lacking RNase L or OAS3, but not in OAS1 or OAS2 knockout (KO) cells (Fig 1G). This is consistent with prior findings showing that OAS3 is the primary upstream dsRNA receptor during diverse viral infections [64], despite recent studies suggesting that OAS1 and OAS2 may also promote context-dependent RNase L-mediated antiviral responses against certain viruses [65,66]. Collectively, these results suggest that formation of nuclear dsRNA during infection with ∆E4 mutant activates both PKR and RNase L pathways, while infection with ∆VA mutant does not yield detectable dsRNA and only activates the PKR signaling pathway.

Infection with ∆VA and ∆E4 mutants induces formation of distinct cytoplasmic RNP granules

Upon activation, PKR promotes the formation of stress granules (SGs), while RNase L activation triggers the assembly of RNase L-dependent bodies (RLBs) [24]. To investigate whether infection with Ad5 mutants induces granule formation following activation of dsRNA sensors, we used immunofluorescence to evaluate protein localization in A549 cells stained for G3BP1 and PABPC1 (Fig 2A). As a control for SG assembly, cells were treated with sodium arsenite (SA), which activates the eIF2α kinase HRI in response to oxidative stress [17]. Alternatively, poly(I:C) was used for RNase L activation and induction of RLBs [23]. No granules were detected in cells infected with Ad5 WT, whereas ∆VA infection induced the assembly of large cytoplasmic granules similar to SGs induced by SA treatment. In contrast, infection with the ∆E4 mutant led to the formation of small, spherical granules resembling RLBs assembled upon poly(I:C) transfection (Fig 2A). In ∆E4 mutant infected cells, granule assembly was accompanied by nuclear accumulation of PABPC1, a hallmark of RNase L-mediated cytoplasmic mRNA decay [22,23]. These observations were consistently reproduced across multiple cell lines (S1 Fig) and confirmed in A549 cells co-stained for G3BP1 and the viral protein DBP (S2A Fig), indicating that infection with the ∆VA mutant produces SG-like granules, whereas the ∆E4 mutant promotes the assembly of RLB-like structures, with no granules being assembled during WT infection.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Infection with Ad5 mutants leads to assembly of granules similar to stress granules (SGs) and RNase L-dependent bodies (RLBs).

(A-C) A549 cells were treated with sodium arsenite (SA, 0.5 or 1.0 mM for 1 h), poly(I:C) (0.5 or 1 µg/mL for 6 h) or infected with Ad5 WT, ∆VA or ∆E4 at an MOI of 10, as indicated. (A) Cells were fixed at 40 hpi and stained for granule markers G3BP1 (green) and PABPC1 (magenta), with nuclei stained in blue (DAPI). No granules were observed in untreated cells, while SA and poly(I:C) treatments induced the assembly of SGs and RLBs, respectively. Granule assembly was not detected upon WT infection, but SGs and RLBs were observed in ∆VA and ∆E4 infected cells, respectively. Scale bar = 15 µm (main panels), 2 µm (cropped panels). (B,C) Histograms show the percentage of cells positive for SGs or RLBs over time following infection or increasing concentrations of SA or poly(I:C). For quantification, a range of 200 to 350 cells imaged in at least six random fields were counted manually for each condition in each replicate. Data are representative of three independent biological replicates, with paired replicate values indicated by colored dots. Bars represent mean, and error bars indicate standard deviation.

https://doi.org/10.1371/journal.ppat.1014452.g002

The formation of cytoplasmic RNP granules was further characterized and quantified over a time course of infection with Ad5 WT and mutants. The percentage of G3BP1 granule-positive cells was quantified by immunofluorescence in A549 cells, with SA or poly(I:C) treatments serving as controls for SG and RLB induction respectively (Fig 2B and 2C). Infection with the ∆VA mutant induced SG formation as early as 16 hours post infection (hpi), peaking at 24 hpi (Fig 2B). Alternatively, assembly of RLBs during ∆E4 infection was only observed in a small proportion of cells by 24 hpi, increasing in frequency by 40 hpi (Fig 2C). These results closely correlate with the kinetics of dsRNA formation previously reported for the ∆E4 mutant [45] and with the timing of PKR and RNase L activation during mutant virus infection in our assays (see Fig 1C and 1F). We noted that a considerable number of cells lacking obvious cytoplasmic G3BP1 foci also displayed nuclear PABPC1 signal during the late stage of WT and ∆VA infections (S2B and S2C Fig) which resembled sites of nuclear speckle reorganization reported in Ad5 infected cells [67–69]. To address whether changes in PABPC1 localization impacted SG assembly, we also quantified the proportion of ∆VA-infected cells showing nuclear PABPC1, comparing cells with or without granules. Similar numbers of granules were observed in both groups, suggesting that nuclear PABPC1 localization does not impact granule assembly during ∆VA infection (S2D Fig). Altogether, these findings demonstrate that ∆VA and ∆E4 mutants differentially induce the assembly of RNP granules with distinct kinetics, consistent with the differential activation of PKR and RNase L during infection.

Virus-induced RNP granules reflect compositional differences between canonical SGs and RLBs

Despite being visually distinct, morphological differences alone are insufficient to distinguish SGs from RLBs, and viral infection may introduce additional context-dependent effects on granule composition. To address these limitations, we first sought to establish comparative reference proteomes for canonical SGs and RLBs under non-infectious conditions. Several studies have characterized SG components by proteomic approaches [57,70,71]. In comparison, the RLB proteome has been examined in only one study, which used differential centrifugation of G3BP1-associated proteins in U2OS cells and suggested that, despite sharing many proteins with SGs, these granules also contain specific components [23]. To explore further differences in protein composition between both condensates, we employed proximity labeling to identify proteins differentially enriched in granules upon SA or poly(I:C) treatments. For this, we used HEK293T cells with APEX2-GFP fused in-frame with the endogenous G3BP1 locus (S3A and S3B Fig) [57,72]. Proteomic datasets showed consistent normalization and reproducibility across conditions and replicates (S3C–S3G Fig). Comparison of SG-enriched proteins identified in our analysis with a previous dataset published using the same system showed significant overlap of identified proteins, validating our experimental approach (S3H Fig). In contrast, comparison with the previous RLB dataset showed limited overlap, which could reflect cell-type specific differences in protein recruitment or methodological biases between fractionation and proximity labeling techniques (S3I Fig).

Analysis of the SG and RLB datasets for differential protein enrichment revealed substantial overlap but also clear differences in granule composition (Fig 3A and 3B and S1 Table), including proteins previously reported to be present in both granules (e.g., UBAP2L, Caprin1) or uniquely enriched in SGs (e.g., FAM120A, TIA1) [23,37]. We validated our proteomics data by immunofluorescence to assess the recruitment of proteins to SGs and RLBs in cells treated with SA or poly(I:C), and examined these differences in granules induced by viral mutants (Figs 3C, 3D, S4A and S4B). We confirmed the accumulation of FXR1, FMRP, Caprin1 and UBAP2L in both granules, whereas FAM120A, EIF4G1, PRRC2C and AGO2 localized exclusively to SGs induced by either SA or ∆VA infection. LSM14A was significantly more enriched in the RLB dataset, showing strong staining in granules induced by poly(I:C) and ∆E4 infection, but only partial relocalization from P-bodies to SGs. To determine whether this effect is specific to LSM14A, we also probed cells for the P-body marker DCP1B. Lack of colocalization between DCP1B and G3BP1-stained RLBs confirmed the specific ability of LSM14A to shuttle between P-bodies, SGs and RLBs (S4C Fig). Collectively, these results reinforce previously documented differences between SGs and RLBs, and they suggest that granules induced during adenovirus mutant infection closely resemble those formed in response to chemical stressors, although additional studies will be required to establish their full proteomic equivalence in infected cells.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Proteomic analysis reveals differential enrichment of proteins in SGs and RLBs.

(A) Volcano plot showing differentially enriched proteins following APEX proximity labeling in HEK293T cells expressing G3BP1-APEX2-GFP fusion protein. Cells were treated with sodium arsenite (SA, 1.0 mM for 1 h) or poly(I:C) (1.0 µg/mL for 4 h) to induce granule formation before labeling. The x-axis shows the log₂ fold-change of the APEX abundances, normalized to input, for SA compared to poly(I:C). The y-axis indicates the –log₁₀ p-value for the comparison of normalized APEX abundances for the three biological replicates of SA and poly(I:C) treatments. Proteins significantly enriched in SGs (left, green) or RLBs (right, magenta) are highlighted. Proteins validated by our proteomic analysis or identified in previously published datasets are labeled in the figure. (B) Venn diagram comparing proteins enriched in SGs and RLBs. Proteins that are unique to SGs (left, green), unique to RLBs (right, magenta), or shared between the granules (intersection) are indicated. Proteins further characterized in this study are highlighted by asterisks. (C,D) Validation of APEX-identified proteins was performed in A549 cells treated with sodium arsenite (SA, 0.5 mM for 1 h), poly(I:C) (0.5 µg/mL for 6 h) or infected with Ad5 WT, ∆VA or ∆E4 viruses at an MOI of 10 for 40 h. G3BP1 staining (green) was used as a marker for granule assembly. (C) Co-staining with FXR1 or UBAP2L (magenta) confirms recruitment of proteins to both SGs and RLBs. (D) Staining for EIF4G1 or FAM120A (magenta) shows proteins recruited to SGs, but absent from RLBs. Nuclei were stained with DAPI (blue). Scale bar = 15 µm (main panels), 2 µm (cropped panels).

https://doi.org/10.1371/journal.ppat.1014452.g003

Unlike stress granules, RLBs have been shown to assemble independently of G3BP1 and G3BP2 proteins [23]. To evaluate whether granules induced during infection share this property, we infected A549 G3BP1/2 KO cells with Ad5 WT or mutants, using SA and poly(I:C) treatments as controls. Cells were stained for FXR1 and PABPC1, confirming the assembly of RLBs in response to either poly(I:C) or ∆E4 infection (S5A Fig). In comparison, stress granules did not form in response to SA or ∆VA infection, consistent with G3BP proteins being required for SG maturation. By comparison, UBAP2L has also been described as an important structural component of SGs, but recent work showed that UBAP2L can localize to P-bodies in cells lacking G3BP proteins under certain stress conditions [73–76]. Since UBAP2L is also recruited to RLBs, we examined differences in granule composition in the G3BP1/2 KO background co-staining cells for UBAP2L with either FXR1 or DCP1B (S5B and S5C Fig). In SA-treated cells, UBAP2L was recruited to P-bodies and showed minimal co-recruitment of FXR1 to these cytoplasmic granules. By contrast, RLBs induced by poly(I:C) or ∆E4 infection displayed robust FXR1 staining and were distinct from P-bodies stained by DCP1B. Together, these findings demonstrate that ∆E4-induced granules are similar to poly(I:C)-induced RLBs, with similar recruitment of protein markers, structural distinction from P-bodies, and assembly in cells lacking G3BP proteins.

PKR and RNase L show distinct requirements for formation of virus-induced RNP granules

To establish whether virus-induced granules form downstream of cellular dsRNA sensor pathways, we evaluated granule assembly upon infection in A549 cells individually knocked out for PKR or RNase L. Immunoblot analysis confirmed the loss of PKR and RNase L expression, and successful infection of each cell line with Ad5 WT and mutant viruses (Fig 4A). We also quantified viral genome accumulation by qPCR as a control for differences in viral replication across KO cell lines (Fig 4B). As expected, ΔVA genome accumulation was modestly but significantly increased in PKR KO cells, whereas ΔE4 genome accumulation remained defective and was not rescued by loss of PKR or RNase L, in agreement with previous literature reports [43,56,77,78]. We then assessed granule formation in response to chemical stressors or viral infections by staining cells for G3BP1 and PABPC1, followed by quantification of granule‑positive cells (Fig 4C–4E). No granules were observed in untreated or WT infected cells. As expected, SG formation during SA treatment was unaffected by the absence of PKR or RNase L expression. By comparison, RLB assembly occurred independently of PKR expression following poly(I:C) transfection in PKR KO cells. In RNase L KO cells, however, poly(I:C) induced the formation of large SGs, consistent with previous findings indicating that poly(I:C) triggers SG assembly in RNase L-deficient cells in response to PKR activation [23]. During ∆VA infection, granules were visualized in parental and RNase L KO cells, but not in PKR KO cells. This suggests SGs form downstream of PKR activation in cells infected by the ∆VA mutant (Fig 4D). Unexpectedly, ∆E4-induced cytoplasmic granules were detected at similar proportions in all cell lines, including those lacking RNase L (Fig 4E). Furthermore, granule assembly was accompanied by nuclear relocalization of PABPC1, as observed in parental cells (Fig 4C). Notably, this occurred despite the lack of RNase L activity, as confirmed by the absence of ribosomal RNA degradation in these cells (see Fig 1G). Similar results were observed in U2OS RNase L KO cells following poly(I:C) transfection or ∆E4 infection (S6A and S6B Fig). However, further granule quantification in A549 cells showed that ∆E4-induced granules were significantly reduced in number in RNase L KO cells, whereas mean granule area was equivalent across all conditions (S6C–S6E Fig). These findings demonstrate that ∆VA infection induces PKR-dependent SGs, whereas ∆E4 infection promotes the assembly of RLB-like granules which are also detected in the absence of RNase L activity, although RNase L might contribute to their abundance.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Requirements for granule formation during mutant AdV infection.

(A) Immunoblot analysis of viral late proteins, RNase L expression, and PKR pathway activation in parental (–) and KO A549 cell lines lacking PKR and RNase L (RL). GAPDH was used as a loading control. Cells were infected with Ad5 WT, ∆VA and ∆E4 at an MOI 10 for 40 h. (B) Viral genome accumulation was quantified by qPCR at 40 hpi under the same infection conditions. Genome accumulation for ∆VA infection was modestly increased in PKR KO cells, whereas ΔE4 genome accumulation remained low across cell lines. (C) Panels show merged IF images for granule markers G3BP1 (green) and PABPC1 (magenta) in the indicated A549 cell lines (parental, PKR KO, RNase L KO) following chemical stress with sodium arsenite (SA, 0.5 mM 1 h) and poly(I:C) (0.5 μg/mL 6 h) or viral infection for 40 h. Nuclei are stained with DAPI (blue). Scale bar = 15 µm. (D,E) Histograms show the proportion of cells displaying SGs (D) or RLBs (E) across the different conditions for each cell line. Between 200 to 350 cells were manually counted across six randomly acquired fields for each condition in each replicate. For all graphs, statistical significance was assessed by two-way ANOVA followed by Tukey’s multiple comparisons test, with * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 and ns for non-significant results. Results are representative of three independent biological replicates. Bars represent mean, and error bars indicate standard deviation, with paired replicate values indicated by colored dots.

https://doi.org/10.1371/journal.ppat.1014452.g004

An additional hallmark of RNase L activity is the widespread cleavage of cytoplasmic RNAs, which drives RLB assembly and bulk PABPC1 translocation to the nucleus [79]. To determine whether granule assembly during ∆E4 infection is accompanied by cytoplasmic mRNA turnover, we performed single-molecule fluorescent in situ hybridization (smFISH) for GAPDH transcripts in parental and RNase L KO cells infected with Ad5 WT or the ∆E4 mutant, or transfected with poly(I:C) (Fig 5A–5D). Consistent with previous findings, GAPDH mRNA signal was strongly reduced in response to poly(I:C) [22]. Reduced GAPDH mRNA signal was also observed in ∆E4-infected cells. In contrast, GAPDH signal during WT infection was equivalent to signal quantified in uninfected cells (Fig 5A and 5B), suggesting that the reduction in GAPDH mRNA in ∆E4-infected cells is a result of RNase L activation. Supportive of this interpretation, GAPDH mRNA levels were comparable across all conditions in RNase L KO cells (Fig 5C and 5D). This confirms that the global degradation of cytoplasmic mRNAs requires RNase L activity in infected cells. Collectively, these data indicate that ∆E4-induced granules accompanied by nuclear PABPC1 translocation do not depend on widespread mRNA degradation, since there was no evidence of GAPDH mRNA turnover in infected RNase L KO cells. This suggests the existence of additional mechanisms for granule assembly during infection with the ∆E4 mutant virus.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Widespread mRNA turnover depends on RNase L in ∆E4-infected cells.

(A-D) A549 parental or RNase L KO cells were treated with poly(I:C) (1 µg/mL for 6 h) or infected with Ad5 WT or ∆E4 (MOI 10 for 40 h). Cells were stained by smFISH for GAPDH transcripts (magenta), PABPC1 (green) to indicate granule assembly and nuclei is stained with DAPI (blue). Scale bar = 15 µm. GAPDH smFISH intensity was quantified per cell as corrected total cell fluorescence (CTCF = Area × [Mean intensity − Mean background]). Graphs show individual cell values as small dots, with colors indicating biological replicates. Large dots indicate the median value for each biological replicate (n = 3). Black bars indicate mean and standard deviation of the three biological replicate medians. Statistical significance was measured by one-way ANOVA followed by Dunnett’s multiple comparisons test using biological replicate medians, with ns = not significant and * = P < 0.01. (A,B) In parental cells, GAPDH mRNA signal was significantly reduced upon poly(I:C) transfection and showed reduced levels in ∆E4-infected cells (P = 0.064). Quantification was performed in a total of mock n = 228, poly(I:C) n = 214, Ad5 WT n = 264 and ∆E4 n = 183 cells. (C,D) In RNase L KO cells, GAPDH mRNA signal was observed among all conditions. The total number of cells analyzed was mock n = 301, poly(I:C) n = 271, Ad5 WT n = 261 and ∆E4 n = 197. For poly(I:C) and ∆E4 conditions, only cells displaying granules or nuclear PABPC1 localization were quantified.

https://doi.org/10.1371/journal.ppat.1014452.g005

Cytoplasmic granules induced by ∆E4 infection share properties with canonical RLBs

To characterize further the extent to which ∆E4-induced granules are similar to canonical RLBs, we performed immunofluorescence in RNase L KO cells following poly(I:C) transfection or ∆E4 infection. In these cells, granules assembled in response to ∆E4 infection showed recruitment of UBAP2L and FXR1 in a similar way to poly(I:C)-induced granules (Fig 6A). In contrast, co-staining of G3BP1 with EIF4G1 or DCP1B confirmed ∆E4-induced granules are distinct from SGs and P-bodies (Fig 6B and 6C). Consistent with prior reports showing that RLBs contain polyadenylated RNA, we also showed by poly(A)+ FISH that ∆E4-induced granules stain for this marker, despite lack of widespread mRNA turnover (Figs 5C, 5D and 6D) [23]. Another difference between RLBs and SGs is their requirement for active translation during granule assembly. SG formation requires polysome disassembly, which is blocked by the translation inhibitor cycloheximide (CHX) [80]. In comparison, CHX treatment does not block RLB assembly as RNase L also cleaves mRNAs bound to polysomes [22]. To assess the impact of CHX on RNP granule assembly, parental and RNase L KO cells were transfected with poly(I:C) or infected with ∆E4 and stained for granule markers. Poly(I:C) induction of RLBs in parental cells was not affected by translation arrest, while the assembly of SGs was completely blocked in RNase L KO cells. By contrast, ∆E4 infection induced granules resembling RLBs in all conditions, which was accompanied by nuclear staining for PABPC1 (Fig 6E and 6F). Together, these findings indicate that granules induced during ∆E4 infection in RNase L KO cells are distinct from SGs, while still sharing compositional and biochemical hallmarks of canonical RLBs.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Granules induced by ∆E4 infection in RNase L KO cells share structural and biochemical properties with RLBs.

(A-D) A549 RNase L KO cells treated with poly(I:C) (1 µg/mL for 6 h) or infected with ∆E4 (MOI 10 for 40 h) were stained for different granule markers. (A) UBAP2L (green) and FXR1 (magenta) show colocalization of both proteins to SGs induced by poly(I:C) and RLB-like granules induced during ∆E4 infection. Alternatively, the co-staining of G3BP1 with EIF4G1 (B) or DCP1B (C) confirms ∆E4-induced RLBs do not colocalize with SG-unique proteins or P-bodies. (D) Staining for G3BP1 (green) and detection of polyadenylated RNA by FISH (magenta) shows recruitment of poly(A)+ RNA to both SGs and ∆E4-induced granules. (E) A549 parental and RNase L KO cells (F) were transfected with poly(I:C) (1 µg/mL for 6 h) or infected with ∆E4 (MOI 10 for 40 h). Cells were treated with DMSO or cycloheximide (CHX) 100 µg/mL to block translation and stress granule assembly. Fixed cells were stained for G3BP1 (green) and PABPC1 (magenta), with DAPI staining nuclei (blue). CHX treatment does not block poly(I:C) induced RLBs in parental cells, but blocks SG assembly in RNase L KO cells. ∆E4 infection induced RLB-like granules independently of CHX treatment in both cell lines. Scale bar = 15 µm (upper panels), 2 µm (cropped panels).

https://doi.org/10.1371/journal.ppat.1014452.g006

Infection with ∆E4 mutant leads to translation arrest independent of PKR and RNase L

In addition to RNP granule induction, activation of PKR and RNase L leads to global translational arrest dependent on eIF2α phosphorylation or RNA cleavage, respectively [17,22]. To probe further this phenotype, A549 parental or RNase L KO cells were transfected with poly(I:C) or infected with ∆E4 and briefly pulsed with puromycin prior to fixation. PABPC1 staining was used to mark granule assembly, while puromycin incorporation was used to monitor active translation (S7A and S7B Fig). As expected, robust puromycin signal was observed in mock or granule-negative cells, whereas puromycin incorporation was markedly reduced in poly(I:C) and ∆E4-infected cells displaying granules. Since ∆E4 infection in RNase L KO cells also activates PKR, we asked whether PKR activity is dispensable for induction of ∆E4-dependent granules and translation shutoff in cells lacking RNase L expression. Using A549 PKR/RNase L double-KO cells, we verified that ∆E4 infection could still induce the assembly of cytoplasmic granules similar to RLBs, which was accompanied by nuclear PABPC1 staining (Figs 7A and S7C). In contrast, poly(I:C) failed to induce either SGs or RLBs in cells lacking both dsRNA sensors. Puromycin pulsing in double-KO cells revealed that formation of ∆E4-induced RNP granules is accompanied by a strong reduction in puromycin incorporation in comparison to uninfected or poly(I:C)-transfected cells (Fig 7B). Importantly, this effect is observed despite lack of detectable eIF2α phosphorylation in these cells (Fig 7C). To rule out any effect associated with eIF2α, double-KO cells treated with SA, poly(I:C) or infected with the ∆E4 mutant were treated with ISRIB, which restricts the inhibitory effect of phosphorylated eIF2α on translation initiation [81]. Consistent with previous reports, ISRIB completely prevented the assembly of SGs during SA treatment, restoring translation in these cells (Fig 7D and 7E). In comparison, no rescue of puromycin incorporation was observed in ∆E4-infected cells displaying nuclear PABPC1 or RNP granules, indicating that translation arrest in these cells does not depend on eIF2α phosphorylation. Collectively, these findings demonstrate that ∆E4 infection induces translational repression and assembly of granules independently of PKR, RNase L and eIF2α phosphorylation. This suppports the existence of a non-canonical pathway linking the accumulation of viral nuclear dsRNA to cytoplasmic RNP granule assembly and translational control.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Granule assembly and translational arrest induced by ∆E4 infection is not dependent on RNase L and PKR.

A549 PKR/ RNase L double-KO cells were treated with poly(I:C) (1 µg/mL for 6 h) or infected with ∆E4 (MOI 10 for 40 h). (A) Cells were stained for G3BP1 (green) and PABPC1 (magenta). ∆E4-induced granules are visualized in the cytoplasm, with PABPC1 nuclear staining. No granules were assembled upon poly(I:C) transfection. (B) To assess active translation, cells were treated with puromycin (10 µg/mL) for 10 min before fixation. PABPC1 staining (green) is shown to indicate granule assembly. Puromycin incorporation (magenta) is visualized in mock or poly(I:C)-treated cells, but not in ∆E4-infected cells displaying RNP granules. (C-D) A549 PKR/ RNase L double-KO cells were treated with SA (0.5 mM for 1 h), poly(I:C) (1 µg/mL for 6 h) or infected with ∆E4 (MOI 10 for 40 h). (C) Immunoblotting analysis showing eIF2α phosphorylation upon SA treatment, but not in response to poly(I:C) and ∆E4 infection. (D) Alternatively, cells were further treated with DMSO or ISRIB (200 nM) and incubated with puromycin (10 µg/mL) for 10 min before fixation. ISRIB blocked SA-induced SG assembly and partially restored translation upon stress, but showed no effect in ∆E4-infected cells, as quantified in (E). Graphs show puromycin incorporation per cell (magenta dots) quantified by mean fluorescence intensity (MFI) in two technical replicates. Black bars indicate median and interquartile range. For DMSO, we analyzed mock n = 80, SA n = 89, poly(I:C) n = 66, and ∆E4 n = 59. For ISRIB treatment, we analyzed mock n = 92, SA n = 94, poly(I:C) n = 72, and ∆E4 n = 60. For ∆E4 infection, puromycin signal was quantified in cells displaying granules or nuclear PABPC1 localization. Statistical significance was measured by Kruskal-Wallis test followed by Dunn’s multiple comparisons test, with ns = not significant, **** = P < 0.0001. Nuclei were stained with DAPI (blue). Scale bar = 15 µm (main panels), 2 µm (cropped panels).

https://doi.org/10.1371/journal.ppat.1014452.g007

Cell culture

Parental cell lines were obtained from American Type Culture Collection (ATCC) and cultured at 37 °C and 5% CO₂. A549 cells (CCL-185), HeLa cells (CCL-2), and HEK293 cells (CRL-1573) were maintained in DMEM (Gibco, 11965084) supplemented with 1% v/v sodium pyruvate (Gibco, 11360070) 10% v/v FBS (Avantor, 89510–186) and 1% v/v Pen/Strep (100 U/ml of penicillin, 100 μg/ml of streptomycin, Gibco, 15140–122). W162 cells (Vero cells containing an integrated copy of the AdV E4 region) were a kind gift from G. Ketner [101]. HBEC3-KT cells (CRL-4051) were grown in Airway Epithelial Cell Basal Medium (PCS-300–030) supplemented with Bronchial Epithelial Cell Growth Kit (PCS-300–040) and 1% v/v Pen/Strep. Knockout A549 cell lines lacking individual dsRNA sensors (PKR, OAS1, OAS2, OAS3 or RNase L) were previously described [45,64]. A549 cells with double knockout for PKR and RNase L were generated using the same CRISPR-based approach described for the single knockout lines [45,64]. A549 cells with double knockout for G3BP1 and G3BP2 were previously described [30]. HEK293T cells expressing G3BP1-V5-APEX2-GFP fusion protein were kindly provided by G. Yeo [57,72]. U2OS parental and RNase L knockout cells were a gift from R. Buisson [102]. All cell lines were routinely tested for mycoplasma contamination using the LookOut Mycoplasma PCR Detection Kit (Sigma-Aldrich, MP0035–1KT).

Viral infection

Adenovirus serotype 5 (Ad5) was originally purchased from ATCC. Ad5 ∆E4 mutant dl1004 was previously described and obtained from G. Ketner [56]. Ad5 ∆VA I/II (dl-sub720) was previously described and obtained from C. Sullivan [49]. ∆E4 virus was expanded and titered on complementing W162 cells, a Vero-derived cell line containing the Ad5 E4 coding region [101]. ∆VA virus was expanded and titered on A549 PKR KO cells. All viruses were purified using two sequential rounds of ultracentrifugation in cesium chloride gradients and stored in virus dilution solution (10 mM Tris-HCl pH 8.0, 100 mM NaCl, 0.1% w/v BSA, 40% v/v glycerol) at -20 °C (short term) or -80 °C (long term). Ad5 WT virus was titered in HEK293 cells and was also titered in the corresponding permissive cell lines used for mutant titering to normalize PFU values across infection conditions. Viral stock titers were determined by conventional plaque assay and all subsequent infections were performed at a multiplicity of infection (MOI) of 10 PFU/cell and harvested at indicated hours post infection (hpi). Cells were infected at 80–90% confluency by incubation with diluted virus in a minimal volume of low serum growth media (2%) for 2 h. After infection viral inoculum was removed and full serum growth media was replaced for the duration of the experiment.

Drugs and treatments

Poly(I:C) (HMW, InvivoGen, tlrl-pic) were transfected at 0.5 µg/mL or 1.0 µg/mL, as described. For transfection, poly(I:C) was complexed with lipofectamine 2000 (Thermo Scientific, 11668027) in Opti-MEM (Gibco, 31985070) following manufacturer’s instructions. Sodium arsenite (Sigma-Aldrich, S7400) was used at 0.5 mM or 1.0 mM, as indicated. For the cycloheximide (CHX, Sigma-Aldrich, C4859) treatment, CHX was added at 100 µg/mL following transfection with poly(I:C), or at 24 hpi following ∆E4 infection, with cells fixed at 40 hpi. ISRIB (MedChem Express, HY-12495A) 200 nM was added to cells simultaneously to SA and poly(I:C) treatments, or at 24 hpi following ∆E4 infection, with cells fixed at 40 hpi. Puromycin (Gibco, A1113802) treatment was performed at 10 µg/mL for 10 min to assess ongoing translation.

Antibodies and inhibitors

The following primary antibodies were used for cellular proteins: Total PKR (Abcam ab32506, WB: 1:1,000), pT446 PKR (Abcam ab32036, WB: 1:1,000), total eIF2α (Cell Signaling 5324S, WB 1:1,000), pS51 eIF2α (Sigma-Aldrich SAB5701788, WB 1:500), GAPDH (GeneTex GTX100118, WB 1:10,000), IRF3 (Thermo Fisher 51–3200, WB 1:1,000), pS386 IRF3 (GeneTex GTX62132, 1:500), RNase L (Santa Cruz sc-74405, 1:500), G3BP1 mouse (BD 611126, IF 1:400), G3BP1 rabbit (Proteintech 13057–2-AP, IF 1:400), PABPC1 (Abcam ab21060, IF 1:600), UBAP2L rabbit (Bethyl Laboratories A300-534A-T, IF 1:100), UBAP2L mouse (Proteintech 67588–1-Ig, IF 1:50), FXR1 (EMD Millipore 05–1529, IF 1:100), EIF4G1 (Abcam ab2609, IF 1:100), FAM120A (Sigma-Aldrich HPA019734, IF 1:500), FMRP (EMD Millipore 05–1235, IF 1:250), Caprin1 (Proteintech 15112–1-AP, IF 1:200), AGO2 (Abnova H00027161-M01, IF 1:100), PRRC2C (Thermo Scientific PA5–55936, IF 1:100), LSM14A (Proteintech 18336–1-AP, IF 1:100), DCP1B (Cell Signaling 13233S, IF 1:100). The following antibodies against dsRNA (anti-dsRNA 9D5 mouse, EMD Millipore 3361, IF 1:2) and puromycin (Sigma-Aldrich MABE343, IF 1:500) were used. For detection of viral proteins, we used anti-adenovirus 5 (against late capsid proteins, e.g., Hexon, Penton, and Fiber) (Abcam ab6982, WB 1:10,000), anti-DBP mouse (gift from D. Ornelles, Clone: B6-8, WB 1:500, IF 1:200), anti-DBP rabbit (Cusabio CSB-PA365892ZA01HIL, IF 1:500), E4orf6 mouse RSA3 (gift from P. Hearing, WB 1:500).

RNA isolation, qRT-PCR and Bioanalyzer

Total RNA was isolated from cells by either TRIzol extraction (Thermo Fisher) or RNeasy Micro kit (Qiagen, 74106), following manufacturer protocols. RNA was treated with RNase-free DNase I (Qiagen, 79256), either on-column or after ethanol precipitation. RNA was converted to complementary DNA (cDNA) using 1 μg of input RNA in the High-Capacity RNA-to-cDNA kit (Thermo Fisher, 4387406). Quantitative PCR was performed using the standard protocol for SYBR Green reagents (Thermo Fisher, 4367659) in a QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems). All primers were used at 10 μM. All values were normalized by the ∆∆Ct method by normalizing first to GAPDH transcripts. Ribosomal RNA integrity was assessed by running total RNA extracts on RNA chips (Agilent RNA 6000 Nano Kit, 5067–1511) using an Agilent 2100 BioAnalyzer. The following primers were used (5’-3’):

1. IFNB1 (Fp: CAGCATCTGCTGGTTGAAGA, Rp: CATTACCTGAAGGCCAAGGA);
2. IRF7 (Fp: GATCCAGTCCCAACCAAGG, Rp: TCTACTGCCCACCCGTACA);
3. GAPDH (Fp: TGCACCACCAACTGCTTAGC, Rp: GGCATGGACTGTGGTCATGAG).

Viral genome accumulation by qPCR

Viral genome accumulation was measured in A549 parental, PKR KO, and RNase L KO cells. Cells were infected with Ad5 WT, ∆VA or ∆VA at MOI 10 for 2 h. After infection, viral inoculum was removed and replaced with complete growth medium. Cells were harvested by trypsinization at 4 hpi or 40 hpi and whole-cell DNA was extracted using the PureLINK Genomic DNA Mini Kit (Invitrogen, K182002). Viral genomes were quantified by qPCR using the standard protocol for SYBR Green reagents, as described above. Viral genomes were assessed with primers targeting the viral genomic DBP region, and total DNA was normalized to genomic tubulin. Viral genomes were normalized for total DNA and total viral genome input at 4 hpi for each condition. The following primers were used (5’-3’):

1. DBP (Fp: GCCATTGCGCCCAAGAAGAA, Rp: CTGTCCACGATTACCTCTGGTGAT);
2. Tubulin (Fp: CCAGATGCCAAGTGACAAGAC, Rp: GAGTGAGTGACAAGAGAAGCC).

Primer extension assay and RNA gel electrophoresis

Total RNA was isolated from cells by TRIzol extraction (Thermo Fisher) following manufacturer protocols. First strand synthesis buffer (Thermo Scientific) and 10 µM of 5’-IRDye800 labeled VA I or II-specific primers were added to equal amounts of RNA for each sample and denatured at 80 °C for 10 min, followed by annealing at 56 °C for 2 h. Elongation mix (100 U SuperScript III (Thermo Scientific), RNase inhibitor (NEB), 2 mM DTT (Thermo Scientific), 1 mM dNTPs (Fischer Scientific), and 10 ng/µl actinomycin D (Cayman Chemical Company) in 1X First Strand Synthesis buffer) was added to each sample and incubated at 55°C for 1 h. RNA was ethanol precipitated, resuspended in 1X RNA Loading Dye, separated on 10% acrylamide, 8 M Urea/TBE gels, and imaged on a LiCOR CLx Odyssey scanner at 800 nm. The following primers were used (5’-3’, 5’-IRDye800 labeled):

1. VA RNA I (TTGTCTGACGTCGCACACCTG);
2. VA RNA II (TCCGGAGGAATTTGCAAGCGG).

Indirect immunofluorescence assays and analysis

Cells were grown on glass coverslips in 24-well plates. After indicated treatments or viral infection, cells were washed with PBS and then fixed in 4% w/v paraformaldehyde (PFA) for 10 min. Cells were permeabilized with 0.5% v/v Triton X-100 in PBS for 10 min and blocked in blocking solution (5% w/v BSA in PBS, 0.1% v/v Tween 20, 0.05% w/v sodium azide) for 1 h. Primary and secondary antibody dilutions were added to coverslips in blocking solution for 1 h, with three PBS washes in between. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) together with secondary antibodies. Secondary antibodies conjugated to Alexa Fluor 488, Alexa Fluor 568, or Alexa Fluor 647 (Invitrogen) against mouse or rabbit were used at a 1:500 dilution. Coverslips were mounted onto glass slides using ProLong Glass Antifade Mountant (Thermo Scientific P36984). Immunofluorescence was visualized using a Zeiss LSM 980 Confocal microscope (Cell and Developmental Microscopy Core at UPenn) and ZEN Blue v. 3.5 software. Images were processed in FIJI and assembled in Figma. Quantification of fluorescence intensity or granule-positive cells was performed from at least six randomly acquired fields and analyzed manually using FIJI. Quantification of RLB number and area was performed in FIJI setting a common threshold value in the G3BP1 channel for granule visualization and analyzing particles (size 0.02 – 1.00 µm²) for each individual cell. Translational activity was measured by quantifying puromycin fluorescence in FIJI and plotted as the mean fluorescence intensity (MFI) for the whole cell area of individual cells. For ∆E4 infection, puromycin signal was quantified in cells displaying granules or nuclear PABPC1 localization. Background was calculated from cell-free regions and subtracted from intensity values.

Fluorescence in situ hybridization

Cells were grown on glass coverslips followed by poly(I:C) transfection or viral infection. At indicated time points, cells were fixed in ice-cold 70% ethanol overnight at 4 °C. After fixation, coverslips were washed with PBS and treated with RQ1 RNase-free DNase (Promega, M6101) for 1 h at 37 °C, followed by two washes in PBS containing 0.1 M EDTA. Primary and secondary antibodies were incubated in PBS for 1 h skipping blocking, with three PBS washes in between. Following secondary staining, cells were fixed in 4% PFA for 10 min and washed twice in PBS. Fluorescence in situ hybridization (FISH) or single molecule FISH were performed for polyadenylated RNA and GAPDH transcripts using Stellaris ShipReady probes (Biosearch Technologies, T30-ATTO647N-1 and SMF-2019–1) and buffers. In brief, coverslips were incubated in Buffer A (1X Buffer A supplemented with formamide, SMF-WA1–60) for 5 min and then incubated overnight at 37 °C with probes diluted in hybridization buffer (SMF-HB1–10). Coverslips were then washed twice in Buffer A for 30 min (with DAPI added to the last wash), rinsed in Buffer B (SMF-WB1–20) for 5 min, and mounted on glass slides. The level of GAPDH transcripts was quantified as total fluorescence per cell and reported as Corrected Total Cell Fluorescence (CTCF = area × [mean intensity − mean background]). For poly(I:C) and ∆E4 conditions, GAPDH mRNA signal was quantified in cells displaying granules or nuclear PABPC1 localization.

SDS-PAGE and immunoblotting

For immunoblotting analysis, cells were lysed in RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with 1% v/v Halt protease and phosphatase inhibitor cocktail (Thermo Scientific, 78446). Protein concentration was determined using a BCA assay kit (Thermo Scientific, 23227). Lysates were mixed with NuPAGE LDS Sample Buffer (4X) containing 1% v/v β-mercaptoethanol and boiled for 10 min, at 95 °C. Proteins were resolved by SDS–PAGE on NuPAGE 4–12% Bis-Tris gels (Invitrogen) using MOPS SDS running buffer (Invitrogen, NP0001). Proteins were transferred to 0.2 μm nitrocellulose membranes (Cytiva Amersham, 10600004) in transfer buffer (25 mM Tris, 192 mM glycine, 10% methanol). Membranes were blocked in 5% w/v non-fat dry milk in TBST (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween-20) supplemented with 0.05% sodium azide. Primary antibodies were incubated overnight at 4 °C in blocking buffer. HRP-conjugated secondary antibodies were incubated for 1h at room temperature, with three TBST washes in between. Immunoblots were developed using SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Scientific, 34580) and imaged using a Syngene G-Box.

APEX2 proximity labeling

APEX2-mediated proximity labeling was carried out essentially as described by the Yeo Lab [57,72]. Briefly, HEK293T G3BP1-V5-APEX2-GFP cells were seeded in 10 cm plates to be confluent the following day and were either left untreated, treated with sodium arsenite (SA, 0.5 mM for 1 h), or transfected with poly(I:C) (1.0 µg/mL for 4 h) to induce granule formation. One hour prior to labeling (or at the time of SA treatment), 500 µM Biotin Tyramide (R&D Systems, 6241/25) was added to the culture media. Labeling was initiated by adding hydrogen peroxide (Sigma-Aldrich H1009) to a final concentration of 1 mM for 60 s before quenching the biotinylation reaction with Trolox (Cayman Chemical Company, 10011659) and sodium L-ascorbate (Sigma-Aldrich, A7631) at final concentrations of 5 and 10 mM, respectively. Cells were then washed once with ice-cold quenching solution and lysed in plate, on ice, in 500 µl ice-cold 8 M Urea Lysis Buffer (8 M Urea, 150 mM NaCl, 20 mM Tris-HCl pH 8.0, 5 mM Trolox, 10 mM Sodium Ascorbate, 10 mM Sodium Azide, 1X Halt Protease/Phosphatase inhibitor (Themo Scientific, 78445)). Cell lysates were sonicated and cleared by centrifugation at 13,500 x g for 10 min at 4 °C. Protein concentration was determined using the Pierce 660 nm Protein Assay (Themo Scientific, 22660), and equal amounts of protein were reduced with 10 mM TCEP (Thermo Scientific, 77720), alkylated with 15 mM iodoacetamide (Sigma-Aldrich, I1149) in the dark at room temperature for 45 min and inactivated with 10 mM DTT in the dark at room temperature for 15 min. Samples were diluted to 2 M urea by adding 3 volumes of 150 mM NaCl, 20 mM Tris-HCl pH 8.0 with protease inhibitors and quenchers. For affinity purification, 100 µl of streptavidin magnetic beads (Fischer Scientific PI88817) per sample were equilibrated with IP Wash Buffer (2 M Urea, 150 mM NaCl, 20 mM Tris-HCl pH 8.0). Samples were added to the equilibrated beads, incubated with rotation at room temperature for 2 h, and washed 6 times with IP Wash Buffer. After aspirating the final wash, the beads were resuspended in 50 mM triethylammonium bicarbonate (TEAB, Sigma-Aldrich, T7408), frozen on dry ice, and stored at -80 °C until subsequent on-bead digestion. The IP samples were subjected to on-bead digestion using endoproteinase Lys-C (Wako) at a 1:100 (w/w) enzyme-to-substrate ratio. Samples were incubated for 1 hour at 37 °C. Following Lys-C digestion, CaCl₂ was added to a final concentration of 1 mM, and sequencing-grade trypsin (Promega) was added at 500 ng per sample. Corresponding input samples for each IP were diluted to a final Urea concentration of 1 M in 50 mM TEAB. Lys-C digestion was performed as described above, followed by trypsin digestion at a 1:50 (enzyme-to-protein) ratio at 37 °C overnight with agitation. Digestions were quenched by the addition of formic acid (FA) to a final concentration of 5%, achieving a pH ≤ 3. Peptides from both IP and input samples were desalted using Poros Oligo R3 reverse-phase columns (Applied Biosystems, Thermo Scientific) packed into P200 stage tips with C18 3M plugs (3M Bioanalytical Technologies). Purified peptides were dried by lyophilization and stored at −20 °C until further analysis.

Nanoflow liquid chromatography–tandem mass spectrometry (nLC–MS/MS)

The peptide mixture was separated using a Dionex Ultimate 3000 high-performance liquid chromatography (HPLC) system (Thermo Scientific) equipped with a two-column setup, consisting of a reversed-phase trap column (Acclaim PepMap100 C18, 5 μm, 100 Å, 300 μm i.d. × 5 mm, Thermo Scientific) and a reversed-phase analytical column (35 cm, 75 μm i.d., 360 μm o.d., packed with Pur C18AQ, 2.4 μm; Dr. Maisch). Loading buffer was 0.1% trifluoroacetic acid (Merck Millipore) in water. Buffer A was 0.1% formic acid, and Buffer B was 80% acetonitrile (ACN) + 0.1% formic acid. The HPLC was coupled online with a Q-Exactive-HF mass spectrometer (Thermo Scientific). Peptides were eluted using a 140 min ACN gradient (90 minute 5%–25% ACN gradient, followed by a 30 minute 25%–45% ACN gradient, a 10 minute 45%-95% gradient, with a final 10-minute isocratic step at 5% ACN) at a flow rate of 300 nl/min for IP and input samples. The MS instrument was controlled by Xcalibur software (Thermo Fisher Scientific). Samples were batch-randomized to account for instrument variation. The data dependent acquisition (DDA) MS method was designed with the MS1 having a window of 400–1000 m/z, AGC target of 1e6 and maximum inject time (MIT) of 75 ms with the MS2 having automated windows, AGC target of 100% and MIT of 75 ms. Fragmentation was performed with high-energy collisional dissociation (HCD) using normalized collision energies (NCE) of 27%. The selection for ions were charges 2–8, minimum peak intensity of 1e4, and a 3 s maximum cycle time.

Proteomics enrichment analysis

The raw mass spectrometer files were processed for protein identification using the Proteome Discoverer (v2.4, Thermo Scientific) and the Sequest HT algorithm with a peptide mass tolerance of 10 ppm, fragment m/z tolerance of 0.02 Da, and a false discovery rate (FDR) of 1% for proteins and peptides. Quantification was performed using a label-free approach using the “Precursor ions quantifier” node, and peptide abundances were rolled up into protein abundance using the summed abundance algorithm using only unique or razor peptides. All peak lists were searched against the UniProtKB/Swiss‐Prot database of Human sequences (9606; downloaded November 2022) using the parameters as follows: enzyme, trypsin; maximum missed cleavages, 2; fixed modification, carbamidomethylation (C); variable modifications, oxidation (M), protein N‐terminus acetylation. All subsequent protein-level analysis of streptavidin-enriched APEX2-labeled proteins (APEX) and total lysate control (Input) abundance quantification data was performed using custom R scripts. Proteins were filtered to include only those identified by at least 1 unique peptide and with peptide q-value < 0.01. APEX and Input abundance quantification values within each sample were transformed to log2 values and normalized by the sample median to account for technical variation in the abundances across samples. APEX and Input abundance means for each protein, in each condition, were obtained by calculating the arithmetic average of the abundance quantification values across the three biological replicates of each condition. Biological replicates for which abundance quantification values were not obtained were removed from the calculation of the mean. Relative protein APEX and Input abundances within each condition were obtained by calculating the statistical z-scores based on the average and standard deviation within the average abundance quantifications of the respective condition. APEX abundances were normalized by the respective input abundances for each replicate of all infections and treatments. Log2 fold changes of APEX abundances for compared conditions were obtained by comparing the average normalized quantification of the respective conditions. Log2 fold changes were imputed for cases in which the protein was identified in 3/3 replicates in one condition and 0/3 replicates in the compared condition. In these cases, the log2 fold change was defined as the minimum or maximum of the fold changes within the respective condition. Statistical p-values associated with the APEX abundance log2 fold changes were calculated using unpaired, two-sided, student’s t-tests comparing the three replicate normalized quantifications of each compared condition. P-values were calculated only for comparisons in which each condition had quantifications in at least 2 of 3 biological replicates. When comparing APEX abundances across conditions, statistically enriched proteins were defined as those proteins that were identified in 3/3 biological replicates in one condition and 0/3 replicates in the compared condition, or identified in 3/3 biological replicates in one condition and exhibiting a log2 fold change > 1 and p-value < 0.05 versus the compared condition.

Reproducibility and statistics

Experimental observations were confirmed in at least three independent biological replicates. Statistical analyses were performed using GraphPad Prism (version 10.6.0). Number of cells analyzed, statistical tests, error bars, n values, and p-values are specified in the corresponding figure legends.