The Small t Antigen of JC Virus Antagonizes RIG-I-Mediated Innate Immunity by Inhibiting TRIM25’s RNA Binding Ability

The innate immune response is the first line of defense against viral pathogens, and in turn, many viruses have evolved strategies to evade detection by the host’s innate immune surveillance machinery. Investigation of the interplay between viruses and the innate immune response provides valuable insight into potential therapeutic targets against viral infectious diseases.

these so-called viral IFN antagonists can also inhibit signaling by the type I IFN receptor (IFNAR) to impede induction of ISGs. Interestingly, although RLRs are known mainly for their roles in detecting and restricting RNA viruses, recent studies demonstrated that these RNA sensors, and in particular RIG-I, also sense several DNA viruses, such as herpesviruses, poxviruses, and papillomaviruses, by recognizing viral or host-derived RNAs (27)(28)(29)(30)(31)(32)(33)(34). As a consequence, not only RNA viruses but also DNA viruses have evolved RLR-inhibitory strategies to promote virus replication and/or establish persistent infection (25,35).
Currently, the role of innate immunity and specific PRRs in controlling JCV infection is unknown, as is antagonism of the type I IFN response by JCV-encoded proteins. The ability of JCV to maintain persistent infection suggests that the virus may have evolved mechanisms to evade viral detection by the innate surveillance machinery. However, no such mechanisms have been characterized as of yet. Here, we show that both RIG-I and cGAS are critical for controlling JCV infection in human astrocytes. We further identify the JCV tAg as an IFN antagonist that specifically inhibits RIG-I by interfering with its K63-linked polyubiquitination by TRIM25. Finally, we show that antagonism of RIG-I K63-linked ubiquitination is also conserved in the related polyomavirus BKV tAg.

RESULTS
The RIG-I and cGAS pathways control JCV replication. Work over the past decade demonstrated that most viral pathogens are sensed and restricted by several PRRs of the innate immune system, either in a temporal or in a cell-type specific manner (14,36). Whereas the sensors responsible for the detection of many human viruses have been discovered, it is currently unknown which intracellular surveillance pathways control JCV infection. Therefore, we sought to investigate whether JCV infection triggers innate immune responses and, further, which innate immune sensors are responsible for controlling JCV replication. To test this, we utilized SVGA human glial astrocytes (37), a commonly used and physiologically relevant target cell type for JCV (38,39). Importantly, we confirmed that these cells express key innate immune proteins involved in virus sensing: these are the three major intracellular sensors, cGAS, IFI16, and RIG-I, which are known to detect other DNA virus infections (e.g., herpesviruses, papillomaviruses), their respective adaptor proteins (stimulator of IFN genes [STING] and MAVS), and the common downstream mediators TANK binding kinase 1 (TBK1) and ISG15, the latter being a representative ISG (see Fig. S1A in the supplemental material). As in previous studies (38,40), JCV (Mad-1 strain) productively replicated in SVGA cells over a time course of 14 days, as determined by quantification of JCV VP1 protein or VP1 and TAg transcript expression ( Fig. 1A and B). Further, JCV replication was accompanied by a low-level induction of IFNB1 and OAS1 (which is an ISG) transcripts (less than 5-and 12-fold, respectively). In contrast, the transcripts of the proinflammatory cytokines interleukin 6 (IL-6) and tumor necrosis factor (TNF) were induced more highly upon JCV infection (up to 50-fold or higher), with greatest induction observed at days 10 to 14 postinfection (Fig. 1C).
Next, to identify the intracellular PRRs that sense and restrict JCV infection, we gene-edited SVGA cells using CRISPR-Cas9 technology to abrogate the expression of RIG-I (gene name, DDX58), cGAS (gene name, MB21D1), or IFI16 (Fig. S1B). SVGA cells expressing a nontargeting (NT) control guide RNA served as a control. Knockout (KO) of the sensor proteins was confirmed by immunoblotting (IB), or successful gene-editing was confirmed by sequencing of genomic DNA (Fig. S1B). Furthermore, the loss of functionality of the RIG-I, cGAS, or IFI16 signaling pathway in the respective KO cells was confirmed by measuring chemokine induction in response to specific ligands: rabies virus leader sequence (RABV LE ), which specifically activates RIG-I, or the synthetic dsDNA poly(dA·dT), which is known to trigger cGAS and IFI16 activation (41,42) (Fig. S1C). Quantification of JCV VP1-positive cells showed that JCV replication was greatly enhanced in RIG-I-KO and cGAS-KO SVGA cells compared to that of NT control cells, where cGAS predominantly contributed to JCV restriction early during infection (at 6 days postinfection [dpi]), while both cGAS and RIG-I equally contributed to JCV restriction later during infection (at 12 dpi) ( Fig. 1D and E). In contrast, knockout of IFI16 had minimal or no effect on JCV replication. Consistent with these data, VP1 mRNA expression was enhanced in RIG-I-KO and cGAS-KO SVGA cells, but not in IFI16-KO cells, compared to that in NT control cells (Fig. 1F). Furthermore, JCV replication was significantly increased in MAVS-KO SVGA cells compared to that in NT control cells ( Fig. S1D to F). Together, these data show that cGAS and RIG-I are the major innate immune sensors that control JCV infection in human glial astrocytes.
The JCV tAg antagonizes the RIG-I-mediated IFN response. Most viruses have evolved effective mechanisms to evade detection by innate immune receptors, which  ensures efficient virus replication and establishment of persistent infection. Since our data showed that both cGAS and RIG-I play important roles in restricting JCV infection, we next asked whether JCV-encoded proteins inhibit the signal transduction mediated by these receptors. The JCV genome encodes six viral proteins (2) which have essential functions in the virus life cycle and also modulate a variety of cellular pathways. However, it is currently unknown whether any of the JCV proteins suppress innate immune signaling. To identify JCV proteins that specifically inhibit cGAS-or RIG-I-mediated innate immunity, we tested the effect of ectopic expression of the six individual JCV proteins on IFN-b promoter activation induced by Sendai virus (SeV), which specifically triggers RIG-I-MAVS signaling ( Fig. 2A and B), or by cGAS and STING coexpression (Fig. S2A) in a luciferase reporter assay. Unexpectedly, none of the JCV proteins had an appreciable effect on cGAS-STING-mediated IFN-b promoter activation under these conditions (Fig. S2A). However, three JCV proteins showed inhibitory effects on SeVinduced innate immune signaling, where JCV tAg had the strongest suppressive effect, which also was dose dependent (Fig. 2B and Fig. S2B).
We also tested the effect of tAg on IFN-b promoter activation triggered by ectopic expression of the related sensor MDA5 or of key downstream proteins of PRR signaling, specifically MAVS, the kinase TBK1, and a constitutively active form of IRF3, IRF3(5D) (43). Whereas JCV tAg effectively blocked the IFN-b promoter activation mediated by RIG-I(2CARD), a constitutively active version of RIG-I, it had no effect on signaling induced by ectopically expressed MDA5, indicating that IFN antagonism by tAg is RIG-I specific. Furthermore, the JCV tAg did not inhibit or only minimally inhibited signaling triggered by ectopic expression of MAVS, TBK1, or IRF3-5D ( Fig. 2C and Fig. S2C), suggesting that JCV tAg acts at the level of the sensor RIG-I or perhaps upstream of it.
In support of these data, JCV tAg also suppressed RIG-I(2CARD)-induced transcript induction of CCL5 and OAS1 with a potency similar to that of the influenza A virus (IAV) NS1 protein, which is a bona fide viral antagonist of RIG-I signaling and thus served as a positive control (44,45) (Fig. 2D). The protein abundances of ISG15 and ISG54 induced by SeV infection were also diminished in the presence of JCV tAg compared to those after empty vector transfection (Fig. S2D). Taken together, these results indicate that the JCV tAg functions as an antagonist of IFN induction by specifically targeting the RIG-I signaling pathway.
TRIM25 is important for control of JCV infection and targeted by JCV tAg. We next sought to identify novel interaction partners of the JCV tAg that may provide clues into the molecular mechanism by which tAg antagonizes RIG-I signaling. Mass spectrometry (MS) analysis of the cellular interactome of affinity-purified FLAG-tagged tAg identified with high peptide-coverage TRIM25, an E3 ubiquitin ligase that is known to be critical for RIG-I activation (Fig. S3A). A coimmunoprecipitation (co-IP) assay confirmed an interaction between JCV tAg and human TRIM25. Of note, the binding of tAg to TRIM25 was comparable to or even stronger than that of IAV NS1 and Nipah virus V (NiV-V) protein, which are known to bind and antagonize TRIM25 (44,46). In contrast to the tAg protein, the JCV Agno, VP1, and VP2 proteins did not bind to TRIM25, strengthening specific TRIM25 targeting by JCV tAg (Fig. 3A). Furthermore, the TAg of JCV, from which tAg is alternatively spliced and shares a conserved DnaJ domain (47), did not bind endogenous TRIM25 either (Fig. 3B). Notably, JCV tAg efficiently bound to exogenous or endogenous TRIM25 in RIG-I-KO cells as efficiently as it did in wild-type (WT) control cells, indicating that the tAg-TRIM25 interaction is not mediated by RIG-I and likely has a direct nature ( Fig. 3C and Fig. S3B).
TRIM25 is characterized by a defined domain structure comprising a central B-box and coiled-coil domain (CCD), which are flanked by an N-terminal RING finger E3 ligase domain (RING) and a C-terminal SPRY (SPla/ryanodine receptor) domain (48). Mapping studies revealed that JCV tAg binds to the TRIM25 SPRY domain, which, despite its low protein expression, efficiently coimmunoprecipitated with JCV tAg. In contrast, the truncation proteins RING-B-box or CCD of TRIM25 did not interact with tAg under the same conditions (Fig. 3D). Importantly, JCV tAg also efficiently bound to endogenous TRIM25 in SVGA cells during native JCV infection (Fig. 3E).
As TRIM25 is a key molecule in the RIG-I signaling pathway (18) and since our data indicated that JCV tAg targets TRIM25, we next determined the relevance of TRIM25 in JCV restriction. To this end, we assessed JCV replication in CRISPR TRIM25-KO SVGA cells (two nonclonal populations) ( Fig. 3F to H and Fig. S3C). Compared to NT control cells, TRIM25-KO SVGA cells showed increased frequency of JCV infection (VP1-positive cells) as well as enhanced viral mRNA expression ( Fig. 3F to H). Since TRIM25 not only is involved in RIG-I signaling but, as recently reported, can also activate the zinc finger antiviral protein (ZAP), which suppresses the replication of certain viruses through translational inhibition and/or viral mRNA degradation (49, 50), we tested whether ZAP also controls JCV infection. We observed no significant difference in JCV infection in CRISPR ZAP-KO SVGA cells from that in NT control cells ( Fig. S3D and S3E), suggesting that TRIM25's anti-JCV activity is independent of ZAP. Collectively, these results indicate that TRIM25 is important for effective control of JCV replication and that, in turn, JCV tAg targets TRIM25 during infection.
JCV tAg impedes the ability of TRIM25 to bind RNA and to modify RIG-I with K63-linked polyubiquitin. We next sought to identify the mechanism by which JCV tAg antagonizes RIG-I signaling. Since tAg effectively binds TRIM25 and because TRIM25 is known to activate RIG-I through catalysis of nondegradative K63-linked polyubiquitination of its N-terminal CARDs, we first tested whether RIG-I ubiquitination was affected by JCV tAg. Ectopic expression of FLAG-tagged JCV tAg profoundly reduced the K63-linked ubiquitination of glutathione S-transferase (GST)-RIG-I(2CARD) (Fig. S4A). Furthermore, whereas JCV tAg strongly reduced the K63-linked ubiquitination of RIG-I (2CARD) in WT cells, it had only a small effect on the very low, residual CARD ubiquitination levels in TRIM25-KO cells (Fig. 4A). These data confirm that TRIM25 is a major E3 ligase for RIG-I CARD ubiquitination and also strengthen the finding that JCV tAg targets the ubiquitination mediated by TRIM25. As only K63-linked ubiquitinated forms of RIG-I, not the unmodified protein, is known to interact with the adaptor protein MAVS (18,19), we tested whether the CARD-CARD interaction between RIG-I and MAVS was affected by tAg. We observed a loss of binding between GST-RIG-I(2CARD) and MAVS-CARD-proline-rich domain (MAVS-CARD-PRD) when JCV tAg-FLAG was overexpressed, consistent with a loss of RIG-I(2CARD) ubiquitination (Fig. 4B). In contrast, both proteins interacted efficiently in vector-transfected cells where RIG-I(2CARD) also showed robust K63-linked ubiquitination (Fig. 4B).
Since our mapping studies showed that JCV tAg interacts with the C-terminal SPRY domain of TRIM25, we hypothesized that tAg might antagonize specific functions mediated by this domain. The SPRY domain of TRIM25 allows for binding to the RIG-I CARDs (18) and also has recently been shown to have RNA-binding capacity, which ultimately promotes TRIM25 E3 ligase activity to catalyze K63-linked ubiquitination of RIG-I and to induce antiviral signaling (51)(52)(53). Although bona fide RNAs that bind and activate TRIM25 in human cells are currently unknown, it was recently discovered that the mouse-specific long noncoding RNA Lnczc3h7a binds to TRIM25, promoting the K63 ubiquitination of RIG-I and thereby innate immunity (52). We therefore tested the ability of JCV tAg to inhibit the RIG-I-TRIM25 interaction and observed only a minor effect of JCV tAg on the binding of GST-RIG-I(2CARD) to TRIM25 (Fig. S4B), suggesting that JCV tAg antagonizes a different function of the TRIM25 SPRY domain. RNA pulldown assay showed that in vitro-transcribed Lnczc3h7a efficiently bound to TRIM25 in the absence of JCV tAg, as previously shown (52), while cotransfected FLAG-tAg abrogated Lnczc3h7a binding to TRIM25 (Fig. 4C). Unlike tAg, FLAG-tagged TAg of JCV, which did not bind to TRIM25 and thus served as a control (Fig. 3B), did not inhibit the Lnczc3h7a-TRIM25 interaction (Fig. 4C). Furthermore, JCV tAg inhibited Lnczc3h7a binding to TRIM25 in a dose-dependent manner (Fig. 4D), suggesting that the viral protein competes with Lnczc3h7a for binding to TRIM25-SPRY. Although Lnczc3h7a is a bona fide TRIM25 RNA agonist, it is mouse specific and not present in human cells. Therefore, we next investigated whether RNA from JCV-infected human cells interacts with TRIM25 and whether JCV tAg inhibits TRIM25's ability to bind this RNA. Biotinylated total RNA purified from JCVinfected SVGA cells efficiently bound to TRIM25-V5 in vitro, while the presence of FLAG-tagged tAg, but not TAg, strongly reduced RNA binding of TRIM25 (Fig. 4E). Taken together, these data indicate that JCV tAg acts as a RIG-I antagonist by binding to its upstream activator, TRIM25. Binding of JCV tAg to the SPRY domain of TRIM25 inhibits its ability to bind RNA, which ultimately prevents TRIM25 from inducing K63 polyubiquitin-mediated RIG-I activation (Fig. 4F).
BKV tAg blocks RIG-I K63-linked ubiquitination and signaling in a TRIM25independent manner. Other viruses in the Polyomaviridae family also encode tAg proteins which share significant sequence homology with the JCV tAg. The BKV tAg shares 77% amino acid sequence homology with JCV tAg, while SV40 and MCV tAgs are more distantly related to the tAg of JCV, with 66% and 34% homology, respectively (Fig. S5A). To determine whether RIG-I antagonism is also conserved in tAgs from other polyomaviruses, we compared the effect of FLAG-tagged tAg from BKV, SV40, or MCV on RIG-I(2CARD)-mediated antiviral/proinflammatory gene induction to that of JCV tAg (Fig. 5A to D). In this assay, IAV NS1 was included as an additional control. Similarly to IAV NS1, JCV and BKV tAgs significantly inhibited CCL5, MX1, and OAS1 transcript induction by RIG-I(2CARD) compared to that after empty vector cotransfection. In contrast, SV40 and MCV tAgs did not inhibit RIG-I(2CARD)-triggered gene expression. In accord, IFN-b luciferase activity was markedly dampened by ectopic expression of FLAGtagged JCV and BKV tAg to similar degrees (Fig. 5E). Of note, like JCV tAg, the BKV tAg did not suppress transcriptional responses triggered by ectopic expression of IRF3(5D) (Fig. S5B), ruling out the possibility that BKV tAg acts downstream of RIG-I.
BKV tAg also robustly inhibited RIG-I(2CARD) ubiquitination, similarly to JCV tAg, while SV40 and MCV tAgs had no appreciable effect on 2CARD ubiquitination (Fig. 5F). Of note, although BKV tAg was expressed to much lower levels, it inhibited RIG-I (2CARD) ubiquitination as potently as the JCV tAg, hinting at a distinct mechanism employed by BKV tAg. Indeed, we did not observe an interaction between BKV tAg and endogenous TRIM25 (Fig. S5C), while JCV tAg efficiently bound to TRIM25 under the same conditions, suggesting that BKV tAg inhibits RIG-I signaling in a TRIM25-independent manner. All together, these data show that the tAg of JCV and the closely related BKV suppress RIG-I signaling by blocking the K63-linked ubiquitination of RIG-I; in contrast, SV40 and MCV tAg do not impede RIG-I ubiquitination and signaling. Our data further suggest that BKV tAg inhibits RIG-I CARD ubiquitination using a distinct, TRIM25-independent mechanism.

DISCUSSION
The role of the innate immune response in controlling persistent JCV infection has not been well studied, nor has the ability of JCV to antagonize specific PRRs and the ensuing type I IFN and ISG induction. JCV infection is known to upregulate IFNs and ISGs in primary human fetal glial cells (54,55). Furthermore, JCV replication is known to be blocked by type I IFNs in both primary human fetal glial and renal epithelial cells (38,56). However, it has remained unknown as to which PRR(s) is responsible for the induction of an antiviral response to JCV infection. In this study, we showed, for the first time, that JCV replication in human astrocytes is restricted by the sensors RIG-I and cGAS, supporting the idea that both intracellular RNA and DNA surveillance pathways contribute to the effective control of JCV infection. These findings are in line with recent reports that demonstrated that DNA viruses are sensed by multiple sensors -both DNA sensors (e.g., cGAS and TLR9) and RNA sensors (in particular RIG-I)-that often act in a temporal manner, which is dependent on the emergence of the respective pathogen-associated molecular pattern (PAMP) that they recognize (28,(30)(31)(32)57). For example, herpes simplex virus 1 (HSV-1) is detected in human fibroblasts by cGAS early during infection, while both RIG-I and cGAS equally contribute to innate sensing of HSV-1 late during infection (28). Furthermore, TLR9 detects HSV-1 in bone marrow-derived IFN-producing cells and dendritic cells (58)(59)(60). While it is well established that genomic DNA of HSV-1 (and also other DNA viruses) is detected by cGAS, some of the RNA species that are recognized by RIG-I during DNA virus infection have been identified just recently.
Conceptually, during DNA virus infection, RIG-I can recognize virus-derived RNAs or host RNA species whose life cycles are perturbed during DNA virus infection. For  instance, during HSV-1 infection, the cellular noncoding RNA RNA5SP141 acts as a RIG-I agonist, and activation of RIG-I by RNA5SP141 likely requires remodeling of the nuclear envelope caused by herpesvirus replication in order to relocalize this pseudogene transcript to the cytoplasm, and also host shutoff of protein synthesis. Specifically, virus-induced host shutoff mechanisms prevent the synthesis of specific RNA5SP141-binding proteins, which normally shield these noncoding RNAs in the cytoplasm. Downregulation of these shielding proteins during HSV-1 infection then liberates RNA5SP141, leading to RIG-I activation (28). For Epstein-Barr virus (EBV) infection, both RNA5SP141 and viral RNAs (i.e., EBV-encoded small RNAs [EBERs], which are abundantly expressed, small noncoding RNAs) were reported to trigger RIG-I signaling (27,29,61). Similarly, during KSHV infection, both viral RNAs and the noncoding polymerase III (Pol III) transcript vault RNAs (vtRNAs) activate the RIG-Imediated antiviral response (32,62). Currently, the physiological PAMPs that trigger RIG-I activation during JCV infection are unknown. It remains to be determined whether RIG-I detects viral RNAs produced during JCV infection or whether host RNAs that are unmasked or mislocalized during JCV infection are sensed by RIG-I. Since polyomaviruses are known to employ effective host shutoff mechanisms, which are most well characterized for SV40, whose tAg has been shown to block host mRNA translation through hypophosphorylation of the eIF4e-binding protein 1 (63), it is tempting to speculate that host shutoff strategies employed by JCV may also induce unmasking of RNA5SP141 or other cellular RNAs, which remains to be determined. Moreover, it is conceivable that viral microRNAs (miRNAs), which are known to be produced during polyomavirus infection (64,65), serve as potential RIG-I ligands.
Practically all viral pathogens are equipped with effective strategies to block IFN induction or the antiviral mechanisms exerted by type I IFNs, and IFN-antagonistic proteins have been identified for many viruses that cause disease in humans. In striking contrast, to date, no type I IFN antagonist has been identified for JCV. Our work identified the JCV tAg as a type I IFN antagonist that specifically targets the RIG-I signaling pathway. Since polyomavirus tAgs have been most well characterized in their ability to bind to the cellular protein phosphatase 2A (66,67) or for their oncogenic properties (however, not in the case of JCV tAg) (68,69), our study unveiled a novel function for polyomaviral tAg proteins that may play a major role for virus persistence and disease outcomes.
Given the important role of TRIM25 in the RIG-I antiviral response, several viruses have evolved to inhibit TRIM25, although through different mechanisms. The NS1 proteins of several IAV strains bind to the CCD of TRIM25, which can improperly position the SPRY domain and interferes with its E3 ligase ability (44,70). The N protein of severe acute respiratory syndrome coronavirus (SARS-CoV) binds to the SPRY domain of TRIM25 and competitively inhibits binding of the E3 ligase to RIG-I (71). In contrast, paramyxovirus V proteins are able to bind the SPRY domain of TRIM25 as well as the CARD domains of RIG-I and block RIG-I ubiquitination in a noncompetitive manner (46). EBV encodes a viral deconjugase, BPLF1, which binds to the TRIM25-RIG-I complex and counteracts the E3 ligase activity of TRIM25 by removing conjugated ubiquitin as opposed to disrupting TRIM25 function (72). Our work discovered that JCV tAg employs a novel mechanism to antagonize TRIM25, that is, by preventing its ability to bind cellular RNA, which has recently been demonstrated to be important for TRIM25 E3 ligase activity and thereby RIG-I activation (51)(52)(53). Consistent with this concept, our data showed that JCV tAg, via binding to the SPRY domain, blocks TRIM25's RNA binding, which ultimately suppresses RIG-I K63-linked ubiquitination and antiviral signaling. Our data demonstrate that JCV tAg targets TRIM25; however, whether it also affects the activity of other E3 ligases that ubiquitinate RIG-I remains to be elucidated.
As Lnczc3h7a, which was recently shown to bind to TRIM25 and to promote RIG-Imediated antiviral signaling, is a mouse-specific long noncoding RNA (lncRNA) (52), we postulate that a specific human lncRNA exists whose binding to TRIM25 is inhibited by JCV tAg to efficiently antagonize type I IFN responses. Although we show that RNA from JCV-infected astrocytes binds to TRIM25, further studies will be required to identify and characterize the specific RNA species that bind(s) to and activate(s) TRIM25 during JCV infection. Whether this RNA is host or virus derived also remains to be determined. Moreover, we expect that our work will stimulate future studies to determine the relevance of the type I IFN response and viral antagonism thereof for the ability of JCV to establish persistent infection and cause disease in appropriate mouse models and in humans. Along these lines, recently developed humanized mouse models in which mice are engrafted with human thymus or human glial cells (39,73) may present opportunities to study the role of JCV tAg-mediated RIG-I inhibition in the maintenance of persistent infection in the kidney, JCV neurotropism, and pathogenesis of PML.
We also found that tAg of the most closely related polyomavirus tested, BKV, also suppressed the K63-linked ubiquitination of RIG-I and RIG-I-mediated upregulation of cytokines, chemokines, and ISGs, while SV40 and MCV tAgs had no discernible effect on RIG-I signaling. Interestingly, our data suggest that although antagonism of RIG-I K63 polyubiquitination is conserved in BKV tAg, the mechanism of inhibition is not. In contrast to JCV tAg, binding of BKV tAg to TRIM25 was not detectable by co-IP, suggesting that BKV tAg blocks RIG-I ubiquitination through a yet-to-be-determined mechanism. One possibility is that BKV tAg recruits a host deubiquitinating enzyme (DUB) to inhibit RIG-I K63-linked ubiquitination and its antiviral signaling. The K63-polyubiquitin-dependent activation of RIG-I is reversed by at least three deubiquitinating enzymes, in particular USP3, USP21, and CYLD (74)(75)(76)(77), to dampen RIG-I signaling as part of a regulatory circuit. As of yet, no virus has been shown to directly recruit any of these DUBs to antagonize RIG-I signaling. Alternatively, miRNAs produced during BKV infection may regulate the expression of TRIM25 or DUBs, which may ultimately suppress RIG-I ubiquitination. Interestingly, such a mechanism has been described for the 3C protein of enterovirus 71 (EV71), which modulates the expression of CYLD by downregulating host miRNA miR-526a (78). Lastly, another possibility is that BKV tAg may inhibit another E3 ligase (e.g., TRIM4, Riplet, or MEX3C) that reportedly can also catalyze RIG-I CARD ubiquitination (17).
Taking these findings together, we discovered that JCV infection is effectively sensed by the RNA sensor RIG-I; however, JCV has evolved to subvert RIG-I-mediated innate immunity. Our work identified the tAg-TRIM25 interaction as a potential molecular target to impede persistent JCV infection and JCV-associated severe disease (i.e., PML) for which there is currently no specific treatment. Our discovery of the role of cGAS and RIG-I in controlling JCV infection may also stimulate investigations to test specific agonists of these two sensors as potential therapeutics to diminish JCV reactivation or active replication.
Viruses were added to the cell suspension and incubated overnight in a water bath at 37°C. The cell suspension was centrifuged at 12,000 Â g for 30 min at 4°C, and the supernatant was combined with the supernatant from the initial centrifugation. Samples were ultracentrifuged over a 40% sucrose layer for 3 h at a 100,000 relative centrifugal force (rcf) at 4°C. Pellets were resuspended in phosphate-buffered saline (PBS), and the virus was tested for infectivity by immunostaining for tAg and flow cytometry.
Pulldowns, coimmunoprecipitations, and immunoblotting. Procedures for GST and FLAG pulldown assays were performed as previously described (18,33) Immunoblot analysis was performed as previously described (33). Briefly, proteins were resolved by Bis-Tris polyacrylamide gel electrophoresis (PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad). Membranes were blocked with 5% (wt/vol) nonfat dry milk in PBS-Tween 20 (PBS-T) for 1 h, followed by probing with primary antibodies at room temperature for 1 h or at 4°C overnight.
Mass spectrometry. For the identification of cellular interacting proteins of the JCV tAg, large-scale FLAG pulldown was performed by transfecting 10-by 10-cm dishes of HEK 293T cells with 20 mg of DNA per dish. Cells were harvested and resuspended in ice-cold PBS, followed by lysis in NP-40 buffer (50 mM HEPES [pH 7.4], 150 mM NaCl, 1% [vol/vol] NP-40, and protease inhibitor cocktail [Sigma; P2714]) and preclearing of cell lysates using Sepharose beads for 2 h at 4°C. Precipitates were extensively washed with 1% NP-40 lysis buffer, and proteins bound to M2 FLAG beads were separated on NuPAGE 4 to 12% Bis-Tris gradient gel (Invitrogen). The gel was silver stained according to the manufacturer's instructions (SilverQuest kit; Invitrogen), and bands specific to tAg-transfected lanes, but not the vector control, were excised and analyzed by ion trap mass spectrometry at the Harvard Taplin Biological Mass Spectrometry Facility in Boston, MA.
IFN-b luciferase reporter assays. HEK 293T cells, seeded into 12-well plates (;1 Â 10 5 cells per well), were transfected with 200 ng of the IFN-b luciferase construct and 300 ng of the b-galactosidase (b-gal)-expressing pGK-b-gal plasmid using Lipofectamine and Plus reagent (Life Technologies) as described in reference 33. Either cells were infected as indicated in the figures with SeV (10 hemagglutination units [HAU]/ml) to activate endogenous RIG-I or they were cotransfected with the following amounts of plasmids to induce activation of innate immunity: 1 ng of plasmid encoding GST-RIG-I (2CARD) or GST (negative control), 100 ng FLAG-MDA5, 100 ng cGAS-3Â FLAG, and 25 ng STING-FLAG, 10 ng TBK1-HA, or 300 ng IRF3-5D-FLAG. Forty hours after transfection, cells were lysed and subjected to a luciferase assay (Promega) according to the manufacturer's instructions. Luminescence and absorbance were measured using a BioTek Synergy HT microplate reader, and luciferase values were normalized to b-gal activity to control for transfection efficiency.
Quantitative RT-PCR analysis. HEK 293T cells were transfected with either an empty vector, plasmids encoding FLAG-tagged polyomavirus tAgs, or IAV NS1 (positive control), together with GST-RIG-I (2CARD) (1 ng) or IRF3-5D-FLAG (100 ng). Twenty-four hours posttransfection, total cellular RNA was extracted with the E.Z.N.A. HP total RNA kit (Omega Bio-tek). RNA quality was assessed using a NanoDrop Lite spectrophotometer. Usually, 50 to 500 ng of RNA was used for a one-step quantitative reverse transcription-PCR (qRT-PCR) using the SuperScript III Platinum one-step qRT-PCR kit with ROX dye (Life Technologies) with commercially available FAM (6-carboxyfluorescein) reporter dye primers (Integrated DNA Technologies) specific for the genes OAS1, CCL5, IFIT2, IFNB1, IL6, MX1, and TNF. Gene expression levels were normalized to that of cellular GAPDH. The fold induction of each target gene relative to values from vector-transfected cells was calculated using the comparative threshold cycle method (DDC T method). qRT-PCR assays were performed using a model 7500 Fast RT-PCR system (Applied Biosystems).
JCV infection studies. SVGA cells were infected with JCV (Mad-1) in infection medium (MEM containing 2 mM L-glutamine and 1% [vol/vol] penicillin-streptomycin) with rocking every 15 min. After 2 h, medium was supplemented with complete MEM that contained 2% (vol/vol) heat-inactivated fetal bovine serum, 2 mM L-glutamine, and 1% (vol/vol) penicillin-streptomycin, and cells were incubated under standard culture conditions for the times indicated in the figures.
To determine the frequency of JCV-infected cells, infected SVGA cells were rinsed in PBS and then fixed using 4% (wt/vol) paraformaldehyde (Santa Cruz Biotechnology) for 30 min. Cells were then permeabilized and stained with 1 mg/ml monoclonal anti-VP1 (PAB597, raised to SV40, kindly provided by Ellen Cahir-McFarland [Biogen, Cambridge, MA]) in dilution buffer (1% BSA and 0.3% Triton X-100 in PBS) at room temperature for 1 to 2 h or at 4°C overnight. Cells were rinsed with PBS and then stained with goat anti-mouse Alexa Fluor 488 secondary antibody (Life Technologies; catalog no. A-11001) at a dilution of 1:2,000 at room temperature for 1 to 2 h. VP1-positive cells were quantified by flow cytometric analysis on an LSR-II machine, followed by analysis using the FlowJo software.
Alternatively, JCV-infected SVGA cells were harvested at the times indicated in the figures, and mRNA levels of the viral genes VP1 and TAg were measured by qRT-PCR as described in detail above.
Confocal microscopy. SVGA cells were grown on coverslips in 12-well plates (;3,000 cells per well), followed by infection with JCV (Mad-1) for the number of days indicated in the figures. Cells were fixed on coverslips in 4% (wt/vol) paraformaldehyde (in PBS) for 30 min, permeabilized, and then stained with anti-VP1 as described for flow cytometry analysis. Cells were rinsed with PBS, stained with goat antimouse Alexa Fluor 488 secondary antibody (Life Technologies; catalog no. A-11001) and again rinsed with PBS before the coverslips were mounted in Mowiol mounting medium containing DAPI (49,6-diamidino-2-phenylindole). The slides were analyzed on an Olympus IX81 inverted widefield microscope by automatic tiling, and images shown are representative of the percentage of infected (VP1-positive) cells for a given coverslip of individual experiments.
RNA pulldown assay. Lnczc3h7a was amplified as a PCR product using the primers 59-TAATACGACTCACTATAGGGAGAATACGTAATTTTTAAGC-39 and 59-GTTTGTTTGTTTGTTTGTTTTTCCAGACA GGGTTTCTCTGTATAG-39 and then directly in vitro transcribed using the MEGAscript kit (Life Technologies). In vitro-transcribed Lnczc3h7a or total RNA isolated from JCV-infected SVGA cells at day 8 postinfection (termed Bio-RNA [JCV-infected]) was biotinylated using the Pierce RNA 39 end biotinylation kit (ThermoFisher) according to the manufacturer's instructions. HEK 293T cells were transfected with TRIM25-V5 and tAg-FLAG or TAg-FLAG for 24 h. Cells were harvested and lysed with NT2 buffer (50 mM Tris-HCl, pH 7.0, 150 mM NaCl, 1 mM MgCl 2 , 0.05% [vol/vol] NP-40, SUPERase·In [Thermo Fisher], protease inhibitor, and Ser/Thr phosphatase inhibitor). Cell lysates were incubated with in vitro-transcribed biotinlabeled Lnczc3h7a or Bio-RNA (JCV infected) for 3 h at 4°C, followed by incubation with Dynabeads M-280 streptavidin (ThermoFisher) for another 3 h at 4°C. Beads were washed several times with NT2 buffer, boiled for 10 min at 95°C, and then subjected to immunoblot analysis as indicated.
Statistical analysis. Statistical analyses were performed by two-tailed Student's t test. A P of ,0.05 was considered statistically significant.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.