Abstract
The Coronaviridae are a family of viruses with large RNA genomes. Seven coronaviruses (CoVs) have been shown to infect humans, including the recently emerged severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the cause of coronavirus disease of 2019 (COVID-19). The host response to CoV infection is complex and regulated, in part, by intracellular antiviral signaling pathways triggered in the first cells that are infected. Emerging evidence suggests that CoVs hijack these antiviral responses to reshape the production of interferons and proinflammatory cytokines. Processing bodies (PBs) are membraneless ribonucleoprotein granules that mediate decay or translational suppression of cellular mRNAs; this is particularly relevant for proinflammatory cytokine mRNA which normally reside in PBs and are repressed. Emerging evidence also suggests that PBs or their components play important direct-acting antiviral roles, providing a compelling reason for their frequent disassembly by many viruses. No information is known about how human CoVs impact PBs. Here, we provide data to show that infection with the human CoV, OC43, causes PB disassembly. Moreover, we show that several SARS-CoV-2 gene products also mediate PB loss and virus-induced PB loss correlates with elevated levels of proinflammatory cytokine mRNAs that would normally be repressed in PBs. Finally, we demonstrate that stimulating PB formation prior to OC43 infection restricts viral replication. These data suggest that SARS-CoV-2 and other CoVs disassemble PBs during infection to support viral replication and evade innate immune responses. As an unintended side effect, the disassembly of PBs enhances translation of proinflammatory cytokine mRNAs which normally reside in PBs, thereby reshaping the subsequent immune response.
Introduction
PBs are ubiquitous, biomolecular condensates that form via liquid-liquid phase separation of proteins with regions of intrinsic disorder, and from RNA-protein and RNA-RNA interactions (1-5). PBs are comprised of the enzymes required for mRNA turnover, including those needed for decapping (Dcp2 and co-factors Dcp1a and Edc4/Hedls) and decay of the RNA body (5’-3’ exonuclease Xrn1 and RNA helicase Rck/DDX6) and some components of the RNA-induced silencing complex (2,6). The RNA found in PBs consists of one third of all coding transcripts which are poorly translated and non-coding RNAs (2,4,7,8). mRNA transcripts, consisting predominantly of grouped regulatory mRNAs with related biological functions, are delivered to PBs by RNA-binding proteins (RBPs) (2,4). One such group bear destabilizing AU-rich elements (AREs) in their 3’-untranslated regions (3’-UTRs) and encode potent regulatory molecules like growth factors, pro-inflammatory cytokines, and angiogenic factors, making their turnover and/or suppression in PBs fundamental to our physiology (9-11). We and others showed that the presence of visible PBs correlates with increased turnover/suppression of ARE-mRNAs (11-15). Conversely, when PBs are lost, constitutive ARE-mRNA suppression is reversed. This provides cells with a means to rapidly respond to stimuli that disassemble PBs to produce proinflammatory cytokines, making PB disassembly an important yet underappreciated regulatory mechanism that tunes the production of potent proinflammatory cytokines that contain AREs, molecules like IL-6, IL-8, IL-1β, and TNF (9).
PBs are constitutive, dynamic ribonucleoprotein (RNP) granules that change in size and number in response to different stimuli. We and others have shown that stressors that activate the p38/MK2 MAP kinase pathway, as well as many virus infections elicit PB disassembly (12,13,15-18). Disassembly can occur by a direct interaction between a viral protein(s) and a PB component that is subsequently re-localized to viral replication and transcription compartments (vRTCs) (19-21) or cleaved by viral proteases (21-23). Viruses can also cause PB disassembly indirectly by activating p38/MK2 signaling (12,13). Despite numerous reports of viral gene products that trigger PB disassembly, corresponding reports of viral gene products that stimulate PB formation are rare, which suggests that PBs possess direct antiviral function and their disassembly may favour viral replication in ways that we fail to grasp (24). Even though other RNPs such as stress granules have emerged as important components of our antiviral defenses that contribute to sensing virus and triggering innate immune responses (25-27), the evidence to support a direct antiviral role for PBs is less well established (24). A direct-acting antiviral role has been defined for several PB-localized enzymes that impede viral replication (e.g. APOBEC3G, MOV10). However, in these cases, the mechanism of viral restriction was attributed to the enzymatic activity of the PB protein(s) and its localization to PBs was not deemed as significant (20,21,23,28-35). PBs also harbour antiviral proteins that are important for innate immune signaling, diverting them and keeping them ready for the inducement of the desired antiviral response (20,36). It remains unclear if organization of particular proteins in PBs, or the higher order condensation of many proteins into the PB, regulates its antiviral activities. Nonetheless, the disassembly of PBs by diverse viruses strongly suggests their importance.
The family Coronaviridae includes seven viruses that infect humans, including the four circulating ‘common cold’ coronaviruses (CoVs), HCoV-OC43, HCoV-229E, HCoV-NL63, and HCoV-HKU1 and three zoonotic viruses that cause severe disease in humans: MERS-CoV, SARS-CoV, and the recently emerged SARS-CoV-2 (37,38). The latter is the infectious cause of coronavirus disease of 2019 (COVID-19). When severe, COVID is characterized by aberrant proinflammatory cytokine production, endothelial cell (EC) dysfunction and multiple organ involvement and has resulted in more than a million deaths worldwide thus far (39-44). Despite previous CoV epidemics, we do not yet appreciate precisely how SARS-CoV-2 infection causes the pathology observed in COVID and urgently need to define these molecular mechanisms to inform novel therapeutic strategies. Although all CoVs encode multiple proteins that hijack antiviral and interferon (IFN) responses (45), SARS-CoV-2 excels in this regard and a mismanaged IFN response is emerging as a major clinical determinant of COVID outcomes (45-48). In support of this, new evidence continues to reveal the multitude of mechanisms used by SARS-CoV-2 to out compete antiviral responses (49-53). For example, four SARS-CoV-2 non-structural proteins (nsp) were recently reported to each interact with specific cellular RNA targets to dramatically diminish IFN-β production and promote viral propagation (52).
To contribute to an enhanced understanding of how SARS-CoV-2 and other CoVs usurp cellular antiviral responses and alter cytokine mRNA expression profiles, we performed an analysis of PBs. There is no published literature on human CoVs and PBs, and only two previous reports mentioned PB dynamics after CoV infection. Murine hepatitis virus (MHV) was reported to increase PBs at early infection times, while transmissible gastroenteritis coronavirus (TGEV) infected cells displayed complete PB loss by 16 hours post infection (24,54-56). We now present the first evidence to show that PBs are targeted for disassembly by a human CoV. We also show that SARS-CoV-2 encodes multiple proteins that when expressed alone are capable of causing PB loss, supporting a coordinated effort by CoVs to disassemble these granules. Finally, we show that prior formation of PBs restricts infection with OC43, delaying the expression of the viral nucleocapsid protein and preventing infectious progeny production. Taken together, these results suggest PBs play central role in the cellular antiviral responses to CoV infection.
Results
A screen of SARS-CoV-2 genes reveals mediators of PB loss
The genome of SARS-CoV-2 is predicted to contain up to 14 open reading frames (ORFs). The two N-terminal ORFs (1a and 1ab) encode two large polyproteins which are processed by viral proteases into 16 non-structural proteins (nsp1-16) essential for viral genome replication and transcription (37). The 3’ end of the SARS-CoV-2 genome is predicted to code for ORFs that are expressed from 9 subgenomic mRNAs (57). Among these are the four structural proteins spike (S), envelope (E), membrane (M) and nucleocapsid (N) and up to 9 potential accessory proteins, not all of which have been validated in infected cells (37). To test the role of SARS-CoV-2 gene products in PB disassembly, we obtained a plasmid library of 27 SARS-CoV-2 genes from the Krogan lab; this library included 14 nsps (excluding nsp3 and nsp16), all structural (S, E, M, N) and candidate accessory genes (ORFs 3a, 3b, 6, 7a, 7b, 8, 9b, 9c, 10) and a catalytically inactive mutant version of the nsp5 3C-like protease (C145A) (57). We individually transfected each plasmid into HeLa cells that express a Dox-inducible GFP-tagged version of the PB-resident protein, Dcp1a (6). When fixed, these GFP-positive puncta co-stain with endogenous PB proteins such as the RNA helicase DDX6/Rck (Fig 1A) and the decapping co-factor hedls/Edc4 (Fig S1). While control cells or those transfected with the envelope (E) protein displayed DDX6-positive PBs, PBs were largely absent after transfection of six SARS-CoV-2 genes including the viral nucleocapsid (N) and the accessory gene, ORF7b (Fig 1A). We quantified number of DDX6-positive PBs per cell for each transfection using CellProfiler, as in (58). This quantification was performed in two different ways. In most cases, transfected cells were identified by co-staining for the Strep-tag II fused to each gene, as shown for N, E and ORF7b (Fig 1A). In such cases, we were able to count PBs only in cells with positive staining only and not count PBs in bystander cells (Fig 1B, thresholded). These data identified two SARS-CoV-2 proteins that may cause PB loss in a cell autonomous manner: ORF7b and N (Fig 1B). For the remaining transfections (nsp1, nsp5, nsp6, nsp11, nsp13, nsp14, ORF3b, ORF6, ORF9b, ORF9c) immunostaining for the Strep-tag II was not robust and we were unable to threshold our PB counts using CellProfiler. In these samples, we quantified PBs in all cells (Fig 1B, unthresholded). These data identified four additional SARS-CoV-2 proteins that may cause PB loss: ORF3b, nsp1, nsp6 and nsp11 (Fig 1B). We verified the expression of all constructs, including low expressors (nsp1, nsp5, nsp6, nsp11, nsp13, nsp14, ORF3b, ORF6, ORF9b and ORF9c) by immunoblotting whole cell lysates harvested from parallel transfections (Fig 1C). We were unable to detect nsp4 and ORF10 by immunoblotting; however, we did visualize these proteins in immunostained samples. Consistent with previous work (57), we were unable to detect nsp6 by immunoblotting or by immunostaining, despite our observation that these cells appear to have less DDX6-positive puncta than our controls (Fig 1B, C).
A screen of SARS-CoV-2 genes reveals enhancers of an ARE-containing luciferase reporter
We performed a secondary reporter screen to identify SARS-CoV-2 proteins that enhanced the activity of a firefly luciferase (FLuc) reporter containing an ARE sequence in its 3’ UTR, rendering the mRNA sensitive to constitutive turnover or translational suppression in PBs (59). For this reason, control samples display extremely low levels of FLuc luminescence relative to a non-ARE renilla luciferase (RLuc) transfection control, that reflects the turnover or suppression of the FLuc mRNA (Fig 2A). When reporter constructs are co-transfected with the positive control protein, KapB, which we have previously shown to cause PB disassembly and elevate the FLuc ARE-mRNA reporter (13,59), the relative luminescence is enhanced ∼50-fold (Fig 2B). After co-transfection with the SARS-CoV-2 gene library, we identified three SARS-CoV-2 proteins that significantly elevated relative luminescence as well as or better than KapB: nsp1 (150-fold), nsp14 (80-fold), and ORF7b (40-fold), and two others that elevated luminescence 10-20-fold: nsp13 and ORF6 (Fig 2B). Of these ARE-mRNA regulating SARS-CoV-2 gene products, two overlap with those we showed to cause PB loss, nsp1 and ORF7b (Fig 1).
Validation of SARS-CoV-2 genes that cause PB loss in primary endothelial cells
Endothelial cells (ECs) have emerged as playing a significant role in severe COVID, they are also sources for many of the cytokines elevated in severe disease and are infected by SARS-CoV-2 (48,60-63). We validated top hits from our PB and ARE-containing reporter screen in human umbilical vein endothelial cells (HUVECs) that were transduced with recombinant lentiviruses expressing N, nsp14 or a vector control. Transduced cells were selected and stained for the endogenous PB marker protein DDX6 and for the Strep-tag II on each of SARS-CoV-2 constructs. Compared to the control, we observed PB loss in N-expressing cells and a decrease in PB numbers in the nsp14-transduced cell population (Fig 3). We also observed that in some cells, nsp14 staining, although usually difficult to detect, overlapped with DDX6-positive puncta in nsp14-transfected HeLa cells (Fig 3B, C). The significance of this observation is unclear and we are investigating it further.
Infection of endothelial cells with human coronavirus causes PB loss
To understand if PBs were lost during infection with human coronaviruses, we established an infection model for the human CoV, OC43, in HUVECs. We tested the ability of OC43 to enter and replicate in these primary cells and confirmed that HUVECs are permissive to OC43 (Fig 4A). We then performed a time-course experiment wherein OC43-infected HUVECs were fixed at various times post infection and stained for DDX6 (PBs) and the viral nucleocapsid (N) protein to denote infected cells. We observed that PBs were lost after infection with OC43 but not lost over the time course of the experiment in our mock-infected control cells (Fig 4A-B). These results showed that PB counts were significantly diminished at 12 and 24 hpi (Fig 4A-B). To determine if the reduced PB counts correlated with changes to cytokine mRNA levels, we harvested total RNA from OC43-infected cells at 24 hpi and performed RT-qPCR for two ARE-containing cytokine transcripts, IL-6 and IL-8 (Fig 4C). Consistent with the observed reduction in PBs, the steady-state levels of IL-6 and IL-8 mRNA both increased ∼20-fold compared to uninfected cells (Fig 4C). We also observed, in some infected cells, DDX6-positive puncta reappeared at 24 hpi, but these puncta did not resemble PBs in terms of their size or shape but were larger, non-spherical aggregates. It is unclear at this time if these DDX6-positive aggregates represent vRTCs or stress granules, another RNP granule that viruses manipulate (25,26). Further experiments will determine the nature of these aggregates.
PBs restrict coronavirus infection
Since SARS-CoV-2 encodes six proteins that induce PB loss and infection with OC43 recapitulates this loss, we hypothesized that PBs may be able to restrict coronavirus infection. To test this, we used lentiviruses to deliver the fluorescently-labeled PB protein, GFP-Dcp1a, or a control GFP protein to HUVECs prior to infecting with OC43. At various times post infection, cells were fixed and stained for the viral nucleocapsid (N) protein (Fig 5A). Although OC43 was able to express N in GFP-positive control cells, we observed viral N gene expression was restricted in cells that expressed GFP-Dcp1a (Fig 5A). The restriction was most pronounced at 6hpi, when we observed only ∼20% of GFP-Dcp1a-positive cells were visibly infected, compared to the GFP-only controls (Fig 5B). At later infection times, more N gene expression was observed in cells also expressing GFP-Dcp1a, suggesting that OC43 infection was able to overcome the restriction imposed by GFP-Dcp1a (Fig 5B). Next, we examined if an increased dose of the GFP-Dcp1a lentivirus, that should elicit the formation of more PBs, would increase the restrictive phenotype (Fig 5C). HUVECs were transduced with 5-fold more GFP-Dcp1a lentivirus than in Fig 5B, infected with OC43, and the infected cell supernatant was collected at 24hpi (Fig 5C). The amount of OC43 infectious particles produced in each condition was determined using a TCID50 assay (64). Infectious OC43 particles were produced from control cell populations transduced with the GFP virus and there was no significant difference in viral particles produced after transduction with a lower dose of GFP-Dcp1a lentivirus (Fig 5C). However, the addition of 5-fold more GFP-Dcp1a lentivirus prior to OC43 infection restricted viral replication and no infectious OC43 particles were detected in the 24hpi-supernatant taken from these cells (Fig 5C). Taken together, these data show that PB fortification prior to OC43 virus infection restricts virus replication, providing a compelling reason why SARS-CoV-2 would coordinate an attack on cellular PBs using multiple viral proteins.
Discussion
In this manuscript, we present data to show that the human CoV, OC43, causes PB loss during infection. The loss of PBs correlates with elevated steady-state levels of two PB-regulated cytokines, IL-6 and IL-8. Using a gene library, we also identify six candidate gene products encoded by SARS-CoV-2 that are capable of inducing PB loss and three candidate gene products that stabilize an ARE-containing luciferase reporter. Moreover, we present evidence that prior fortification of PB granules using overexpression of the PB protein, GFP-Dcp1a, restricts the infectious cycle of OC43, delaying viral nucleocapsid protein production and preventing infectious progeny production. These data support our model that PBs are antiviral granules designed to protect cells from viral invaders.
Multiple SARS-CoV-2 gene products highlight the diversity of approaches used by one virus to induce PB loss
We screened 27 SARS-CoV-2 gene products by transfection in HeLa cells (57) and identified six candidates that reduce PB numbers including the viral nucleocapsid protein (N), the viral host shutoff factor (nsp1), non-structural proteins nsp6 and nsp11, and accessory proteins ORF7b and ORF3b (Fig 1). The most significant of these candidates was the N protein, which we also showed caused endogenous PB loss in primary ECs (Fig 3). N is a multifunctional RNA-binding protein (RBP) that coats the viral genome and induces phase separation to promote viral particle assembly (37,65-68). N possesses several non-specific RNA-binding regions and is phosphorylated in its central serine-arginine (SR-rich) domain (68,69). SARS-CoV-2 N also contains three putative NLS sequences (70). One possible reason for PB loss may be the indiscriminate RNA binding of N protein which acts as sponge for RNA, pulling it out of cytoplasm, reducing the RNA-protein interactions required for phase separation of PBs (3). We are currently engaged in site-directed and truncation mutagenesis studies to determine the precise region(s) of N that are essential for its effect on PBs (71). Another PB-regulating SARS-CoV-2 protein we discovered is the bifunctional enzyme, nsp14, which possesses an N-terminal exonuclease domain (ExoN) required for proofreading of the CoV polymerase complex and a C-terminal N-methyltransferase (MTase) domain that contributes to viral RNA capping (37,72). Nsp14 caused PB loss in our EC validation studies yet it did not cause PB loss in HeLa cells (Fig 1, 3). Though we are unclear of the precise reason for these discrepant results, one consideration is that GFP-Dcp1a expression in HeLa cells makes PBs more stable and their disassembly more difficult. It was also quite difficult to detect the expression of nsp14 in HeLa cells after their transfection, although nsp14 was detected by immunoblotting; therefore, PBs were counted in all cells, and the results were not specific for cells that expressed nsp14 (Fig 1). In contrast, in our EC validation studies we selected for transduced cells; thus, all ECs contain nsp14 even if the staining is low (Fig 3). This may explain the significant PB loss observed in ECs and not in HeLa cells. Finally, we observed that nsp14 occasionally formed small puncta that overlapped with DDX6-positive PBs in both HeLa cells and ECs (Fig 3A, 3C). There are no previous reports of any other CoV proteins co-localizing to PBs. The significance of this observation remains unclear. We will continue to interrogate nsp14 localization using confocal microscopy and immunoprecipitation assays. Nsp14 was also a significant hit in the luciferase reporter assay (Fig 2), suggesting that nsp14 promotes the stabilization or translation of PB-localized, ARE-containing cytokine mRNA. Consistent with this, alphacoronavirus nsp14 enhanced TNF and IFN-β (both these mRNAs contain AREs) independent of its enzymatic functions, a phenotype that is consistent with a PB-regulating protein (73).
Our initial validation efforts have focused on nsp14 and N (Fig 3); however, our screens did highlight other SARS-CoV-2 proteins that may regulate PBs to enhance translation of ARE-mRNAs. These include the top hit in our luciferase screen, nsp1. All CoVs encode a host shutoff protein called nsp1 that is the first protein expressed after virus entry and an important IFN antagonist that limits IFN and interferon-stimulated gene translation (52,74). SARS-CoV-2 nsp1 binds to 18S rRNA and the blocks the mRNA entry channel in the 40S ribosomal subunit, thereby blocking translation (52,75). Given its role in translation shutoff and RNA decay, we were initially concerned that nsp1-mediated elevation in the ARE-containing FLuc reporter was non-specific because we observed low levels of the RLuc control that could suggest global translation shutoff (52,76-78). However, stress-responsive cellular transcripts are often resistant to the action of nsp1and other viral shutoff proteins suggesting that the sequestration of regulatory transcripts in PBs is a strategy used by cells to withstand viral host shutoff (74,77,79). For this reason, PB-regulated transcripts like the ARE-containing FLuc reporter may be resistant to nsp1-mediated shutoff (Fig 2). We are currently validating if nsp1 is a PB-regulating viral protein in ECs and will explore its effects on cytokine mRNA levels. In addition, ORF7b and the short ORF3b are intriguing hits because both are accessory proteins that have been reported to either activate the kinase p38 (an inducer of PB disassembly) or antagonize IFN responses (53,80). Not picked up in our screen, the viral 3C-like protease, nsp5, is also of particular interest because porcine CoV nsp5 cleaves Dcp1a, an event that would be predicted to cause PB disassembly (81). Although nsp6 and nsp11 were top hits in our unthresholded PB screen (Fig 1), their expression is either toxic or difficult to detect and we are engaged in developing strategies to minimize these issues and study them further.
PBs are antiviral granules that restrict CoV infection
We used the human CoV, OC43, to infect ECs and show that virus infection caused significant PB loss at 12 and 24 hpi (Fig 4). We then fortified PBs using the overexpression of GFP-Dcp1a and showed that PB fortification before infection delayed the visual expression of viral nucleoprotein and restricted the production of infectious viral particles (Fig 5). However, the precise details of the mechanism of restriction remain unclear. We are engaged in ongoing time course experiments to determine if the restrictive phenotype represents a delay of viral progeny production or a complete block. It is also unclear if the prior GFP-Dcp1a overexpression restricts OC43 replication due to a direct antiviral role of the Dcp1a enzyme itself or its ability to promote the formation of larger PBs (82,83). GFP-Dcp1a overexpression is a common approach used to visualize PBs, and we know these GFP-positive puncta co-stain with endogenous PB proteins (Fig 1) (82). However, GFP-Dcp1a puncta are larger and may be more stable than endogenous PBs (56,83). Our ongoing work will utilize other molecular tools that regulate PB dynamics to determine if certain PB resident proteins exert antiviral properties independent from the formation of the PB granule itself. We are currently engaged in experiments to overexpress different PB proteins that have roles in RNA decay (Dcp1a, DDX6 RNA helicase, the decapping cofactor hedls/Edc4 or the exonuclease Xrn1) as well as PB proteins that do not directly mediate RNA decay but play an important condensation role for PB granule formation (Lsm14A, 4E-T) (3,4,84). Each of the proposed proteins will affect PBs differently when overexpressed: Dcp1a, hedls/Edc4 and DDX6 increase PB formation, with Dcp1a expression inducing larger granules than the others, Dcp2 has no effect on granule size/number, and Xrn-1 eliminates PBs because it increases mRNA decay (4,83-86). We predict that other GFP-tagged PB proteins that promote molecular condensation of PBs, such as GFP-DDX6, will also restrict OC43 infection. Conversely, we expect that the overexpression of Xrn1 will promote viral infection (as shown by others (34)) because it prevents PB formation, even though it can degrade viral RNA.
There is increasing evidence in support of our hypothesis that the antiviral role of PB-localized enzymes is promoted by phase separation of molecules into PBs and that the antiviral function of these molecules is lost when PB granules decondense. This has been previously proposed for decapping complexes, the enzymatic activity of which is increased by phase separation and decreased in solution (3-5). However, for PB-localized enzymes that have established antiviral effects, (e.g. APOBEC, MOV10), their ability to restrict virus infection has previously been attributed to their enzymatic activity rather than their localization to PBs (20,21,23,29-35,82). Therefore, it is not yet clear if the antiviral capability of PBs relies on their formation into biomolecular condensates. Here, we consider that the antiviral restriction promoted by PB-localized enzymes requires the granule formation for optimal function. By that definition, factors that promote PB condensation may also be antiviral. Therefore, in part one of our model, we propose that PBs are direct-acting antiviral granules that can restrict virus infection when present as visible condensates; for this reason, they are targeted by disassembly by most viruses.
PBs interact with other innate immune pathways to regulate antiviral signaling
One possibility is that PBs are antiviral because their proteins help the cell respond to signals that activate innate immune pathways (20,22,36,85). It is not clear if PB antiviral activities are connected to established regulators of innate immune signaling such as the RNA sensors RIG-I/MDA5, transcription factors IRF3/NF-kB, or the production of Type I IFNs, though reports suggest this may indeed be the case (36,85,87). The recent demonstration that TRAF6 controls Dcp1a localization within PBs using ubiquitylation suggests that antiviral signaling is more complex than previously appreciated and integrates transcriptional responses with cytokine mRNA suppression in PBs (85,87). Moreover, the PB protein Lsm14A has also been shown to bind to viral RNA/DNA after infection to promote IRF3 activation and IFN-β production (36). Although it remains unclear if organization of particular proteins in PBs, or the higher order condensation of many proteins into the PB, regulates its antiviral activity (82,87), what is clear is that PB disassembly reverses the constitutive decay or translational suppression of cytokine mRNAs that contain AREs that would normally occur there (11-15,88). These data suggest that when viruses coordinate an attack to cause PB loss, this event relieves the cytokine mRNA suppression and acts as a danger signal that signals the immune system for help. In this way, PB disassembly may be a central component of the cellular innate immune response and part of the sequelae of signals that notify the immune system that a cell is infected. In situations where interferon responses are delayed or defective, as is emerging for SARS-CoV-2 and severe COVID (45-53), PB disassembly may be an important contributing factor to pathogenic cytokine responses. Therefore, in part two of our model, we propose that cells view viral PB disassembly as a danger signal and respond by increasing production of proinflammatory cytokines as a call for reinforcements.
A new model for the role of PBs as a nexus of intracellular antiviral responses
We now propose a model for how PBs may regulate cellular innate responses to CoV infection (Fig 6). Our model places PBs at a nexus point, a connection between intracellular direct-acting antiviral responses and proinflammatory cytokine responses. We propose that PBs function as direct-acting antiviral granules that can restrict virus infection if robustly induced into phase-separated molecular condensates; for this reason, they are targeted by disassembly by most viruses. Second, when viral invaders cause PB loss, the cell responds to the inactivation of PBs with a call for reinforcements. This occurs because PBs house suppressed cytokine transcripts which are relieved of their suppression by viral PB disassembly. This model places the PB as a central player in the antiviral response that coordinates the immune reshaping that occurs after CoV infection.
Materials and Methods
Cell culture
All cells were maintained at 37 °C with 5% CO2 and 20% O2. Vero E6 (ATCC), HEK293T cells (ATCC), HeLa Tet-Off cells (Clontech) and HeLa Flp-In TREx GFP-Dcp1a cells (a generous gift from Anne-Claude Gingras (6)) were cultured in DMEM (Thermo Fisher) supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin, 2 mM L-glutamine (Thermo Fisher) and 10% FBS (Thermo Fisher). HUVECs (Lonza) were cultured in endothelial cell growth medium (EGM-2) (Lonza). HUVECs were seeded onto gelatin-coated tissue culture plates or glass coverslips.
Plasmids and Cloning
pLenti-IRES-Puro SARS-CoV2 plasmids were a generous gift from the Krogan Lab (89). pLJM1-GFP-Dcp1a was generated by cloning pT7-EGFP-C1-HsDCP1a (Addgene 25030) into pLJM1-BSD (90)using AgeI and SmaI restriction sites (NEB). pLJM1-KapB-BSD was generated by cloning pBMNIP-KapB (13) into pLJM1-BSD using EcoRI and BamHI restriction sites (NEB).
Transient Transfections
Transient transfections were performed using Fugene (Promega) according to manufacturer’s guidelines. Briefly, HeLa Flp-In TREx GFP-Dcp1a cells were seeded in 12-well plates at 150,000 cells/well in antibiotic-free DMEM on coverslips for immunofluorescence or directly in wells for lysates. Cells were transfected with 1 μg of DNA and 3 μL of Fugene for 48 h before processing.
Lentivirus generation
All lentiviruses were generated using a second-generation system. Briefly, HEK293T cells were transfected with pSPAX2, MD2G, and the plasmid containing a gene of interest using polyethylimine (PEI, Polysciences). Viral supernatants were harvested 48 h post-transfection and frozen at -80°C until use. For transduction, lentiviruses were thawed at 37°C and added to target cells in complete media containing 5 µg/mL polybrene (Sigma) for 24 h. The media was changed to selection media containing 1 µg/mL puromycin or 5 µg/mL blasticidin (Thermo Fisher) and cells were selected for 48 hours before proceeding with experiments.
Immunofluorescence
Cells were seeded onto coverslips for immunofluorescence experiments. Following treatment, cells were fixed for 10 mins at 37 °C in 4% (v/v) paraformaldehyde (Electron Microscopy Sciences). Samples were permeabilized with 0.1% (v/v) Triton X-100 (Sigma-Aldrich) for 10 min at room temperature and blocked in 1% human AB serum (Sigma-Aldrich) 1 h at room temperature. Primary and secondary antibodies were diluted in 1% human AB serum and used at the concentrations in Table 1. Nuclei were stained with DAPI. Samples were mounted with Prolong Gold AntiFade mounting media (Thermo).
Immunoblotting
Cells were lysed in 2X Laemmli buffer and stored at -20°C until use. The DC Protein Assay (Bio-Rad) was used to quantify protein concentration as per the manufacturer’s instructions. 10-15 µg of protein lysate was resolved by SDS-PAGE on 4-15% gradient TGX Stain-Free acrylamide gels (BioRad). Total protein images were acquired from the PVDF membranes after transfer on the ChemiDoc Touch Imaging system (BioRad). Membranes were blocked in 5% BSA in TBS-T. Primary and secondary antibodies were diluted in 2.5% BSA, and used at the dilutions found in Table 1. Membranes were visualized using Clarity Western ECL substrate and the ChemiDoc Touch Imaging system (BioRad).
Luciferase Assays
HeLa Tet-Off cells were transfected according to Corcoran J.A. et al. Methods (2011) with pTRE2-Firefly Luciferase-ARE, pTRE2-Renilla Luciferase, and SARS-CoV2 plasmids using Fugene HD (Promega). Firefly and Renilla luciferase activity were quantified using the Dual Luciferase Assay Kit (Promega) and read on a Modulus Microplate luminometer (Promega).
Quantitative PCR
RNA was collected using an RNeasy Plus Mini Kit (Qiagen) according to the manufacturer’s instructions and stored at -80°C until use. RNA concentration was determined and was reverse transcribed using qScript XLT cDNA SuperMix (QuantaBio) using a combination of random hexamer and oligo dT primers, according to the manufacturer’s instructions. Depending on starting concentration, cDNA was diluted between 1:10 and 1:20 for qPCR experiments and SsoFast EvaGreen Mastermix (Biorad) was used to amplify cDNA. The ΔΔquantitation cycle (Cq) method was used to determine the fold change in expression of target transcripts. qPCR primer sequences can be found in Table 2
OC43-CoV Propagation and Infection
Stocks of hCoV-OC43 (ATCC) were propagated in Vero E6 cells. To produce viral stocks, Vero E6 cells were infected at an MOI of 0.01 for 1 h in serum-free DMEM at 33°C. Following infection, the viral inoculum was removed and replaced with DMEM supplemented with 2% heat-inactivated FBS and 100 units/mL penicillin/streptomycin/glutamine. After 6 days, the supernatant was harvested and centrifuged at 2000 RPM for 5 mins to remove cellular debris. Virus stocks were aliquoted and stored at -80°C for single use. Viral titers were enumerated using Reed and Muench tissue-culture infectious dose 50% (TCID50) in Vero E6 cells (64). For infection, cells were seeded onto coverslips to achieve ∼70% confluency after 24 hours. The following day, the growth media was removed and replaced with 100 μl of human CoV-OC43 inoculum and incubated at 37°C for one hour, rocking the plate every 10 minutes to distribute viral inoculum. Following incubation, the virus inoculum was removed and replaced with EGM-2.
Processing Body Quantification
Processing bodies were quantified using an unbiased image analysis pipeline generated in the freeware CellProfiler (cellprofiler.org) (91). First, detection of nuclei in the DAPI channel image was performed by applying a binary threshold and executing primary object detection between 50 and 250 pixels. From each identified nuclear object, the “Propagation” function was performed on the respective CoV2-ORF channel image to define cell borders. The identified cell borders were masked with the identified nuclei to define a cytoplasm mask. The cytoplasm mask was then applied to the processing body puncta channel image to ensure only cytoplasmic puncta were quantified. Background staining was reduced in the cytoplasmic puncta channel using the “Enhance Speckles” function. Using “global thresholding with robust background adjustments”, puncta within a defined size and intensity range were quantified. Size and intensity thresholds were unchanged between experiments with identical staining parameters. Intensity measurements of puncta and CoV2-ORF staining were quantified. Quantification data was exported and RStudio was used for data analysis.
Statistics
All statistical analyses were performed using GraphPad Prism 8.0. Significance was determined using the tests indicated in each of the figure legends.
Figure Legends
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.
- 11.↵
- 12.↵
- 13.↵
- 14.
- 15.↵
- 16.
- 17.
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.
- 31.
- 32.
- 33.
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.
- 41.
- 42.
- 43.
- 44.↵
- 45.↵
- 46.
- 47.
- 48.↵
- 49.↵
- 50.
- 51.
- 52.↵
- 53.↵
- 54.↵
- 55.
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.
- 62.
- 63.↵
- 64.↵
- 65.↵
- 66.
- 67.
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵