|Home | About | Journals | Submit | Contact Us | Français|
Recently, linear ubiquitin assembly complex (LUBAC)-mediated linear ubiquitination has come into focus due to its emerging role in activation of NF-κB in different biological contexts. However, the role of LUBAC in LMP1 signaling leading to NF-κB and interferon regulatory factor 7 (IRF7) activation has not been investigated. We show here that RNF31, the key component of LUBAC, interacts with LMP1 and IRF7 in Epstein-Barr virus (EBV)-transformed cells and that LUBAC stimulates linear ubiquitination of NEMO and IRF7. Consequently, LUBAC is required for LMP1 signaling to full activation of NF-κB but inhibits LMP1-stimulated IRF7 transcriptional activity. The protein levels of RNF31 and LMP1 are correlated in EBV-transformed cells. Knockdown of RNF31 in EBV-transformed IB4 cells by RNA interference negatively regulates the expression of the genes downstream of LMP1 signaling and results in a decrease of cell proliferation. These lines of evidence indicate that LUBAC-mediated linear ubiquitination plays crucial roles in regulating LMP1 signaling and functions.
IMPORTANCE We show here that LUBAC-mediated linear ubiquitination is required for LMP1 activation of NF-κB but inhibits LMP1-mediated IRF7 activation. Our findings provide novel mechanisms underlying EBV-mediated oncogenesis and may have a broad impact on IRF7-mediated immune responses.
Epstein-Barr virus (EBV) is clearly an oncogenic virus in humans and is quite frequently involved in AIDS-related malignancies. EBV-encoded latent membrane protein 1 (LMP1), a member of the tumor necrosis factor receptor (TNFR) superfamily, is essential for altered cell growth, survival, adhesion, and invasive potential. LMP1 shares many features with CD40 and Toll-like receptors (TLRs) in signaling transduction, including the activation of NF-κB, mitogen-activated protein kinases (MAPKs), and phosphatidylinositol 3-kinase (PI3K)/Akt (1,–5).
Interferon regulatory factor 7 (IRF7) is the “master” regulator of type I interferon (IFN) production in antiviral innate immunity. IRF7 also plays a role in EBV latency maintenance and oncogenesis, among other potential functions (6). IRF7 is overexpressed in EBV-positive AIDS-related central nervous system lymphoma and causes anchorage-independent growth of NIH 3T3 cells that induce tumor formation in nude mice (7). In addition, IRF7 targets the BamHI-A rightward transcript (BART) P1 promoter (8), and BART-derived microRNAs (miRNAs) play a role in EBV oncogenesis (9, 10). Furthermore, we and others have presented vast evidence that IRF7 is induced, as well as activated, by LMP1 (11,–15). Thus, tight regulation of IRF7 activity is of great importance in EBV latent infection, not only for shaping EBV latency by refining type I IFN signaling, but also for EBV oncogenesis.
Ubiquitination is a pervasive mechanism important to the activation of NF-κB and other transcription factors, including IRFs, and is crucial to coordinating complex aspects of many fundamental cellular activities. DNA tumor viruses, especially EBV, are well known to manipulate the host ubiquitin (Ub) machinery to facilitate their latent persistence and oncogenesis (16,–20). We have provided solid evidence showing that IRF7 is activated by LMP1 through a TRAF6/RIP-dependent, K63-linked ubiquitination pathway (11, 12). We have further shown that TNFAIP3 regulates this process as a deubiquitinating enzyme (DUB) (21).
More recently, LUBAC (linear ubiquitin assembly complex)-mediated linear ubiquitination has been shown to be crucial for canonical NF-κB activation in diverse biological contexts (22,–27), including apoptotic and immune stimuli, such as tumor necrosis factor alpha (TNF-α) (28, 29), interleukin 1β (IL-1β) (30), genotoxic stress (31), CD40 (32), TLRs (33), NOD2 (34), and NLRP3 (35). However, whether LUBAC is required for NF-κB activation in the setting of EBV infection has not been examined. LUBAC is a ternary ubiquitin ligase complex composed of HOIP (RNF31), HOIL1L (RNF54), and Sharpin and is constitutively formed under normal physiological conditions (28, 29, 36). RNF31 is likely the central component of the complex (37). LUBAC is abundantly expressed in thymus and spleen, implying its potential role in lymphocytes (25). Originally, LUBAC was shown to target TAB2/3 for degradation and therefore negatively regulates TNF-α- or IL-1β-stimulated NF-κB activation (30, 36). Recent reports clearly show that LUBAC specifically activates the canonical NF-κB pathway, but not the JNK pathway, by conjugating linear polyubiquitin chains onto NEMO and RIP1 (25, 38). More recently, the deubiquitinase Gumby (OTULIN), CYLD, and TNFAIP3 were shown to negatively regulate LUBAC-mediated NF-κB activation through different mechanisms in different biological contexts (34, 39,–42). These intriguing findings have advanced our current understanding of the novel functions of ubiquitination in signal transduction pathways and may provide novel paradigms for the treatment of human cancers and immune diseases (43).
In this study, we provide evidence that LUBAC is involved in LMP1 activation of NF-κB but negatively regulates LMP1-promoted IRF7 transcriptional activity. Since LUBAC activation of NF-κB has been shown in other contexts, we focus on LUBAC regulation of IRF7, which has never been investigated in any context.
LUBAC is crucial for canonical NF-κB activation in diverse biological contexts (22,–27). However, whether LUBAC is required for NF-κB activation by LMP1 has never been reported. Furthermore, the relationship between LUBAC and IRF7 has never been investigated. To assess if LUBAC participates in LMP1 signaling transduction to NF-κB and IRF7 activation, we first determined if LMP1 interacts with RNF31. To this end, hemagglutinin (HA)-LMP1 and Flag-RNF31 expression plasmids were transfected into 293T cells, and the cell lysates were subjected to immunoprecipitation (IP). The results showed that LMP1 and RNF31 clearly interact (Fig. 1A).
To check if the interaction between RNF31 and LMP1 is specific, we used a panel of LMP1 deletion mutants for IP assays (Fig. 1B). Our results indicated that LMP1 CTAR2 specifically interacts with RNF31.
Of special note and novelty, we further show that RNF31 and IRF7 interact when both are transiently expressed in 293T cells and that their interaction is much stronger in the presence of LMP1, indicating that LMP1 promotes their interaction (Fig. 1C).
We further checked their endogenous interaction in the context of EBV latent infection. A panel of EBV-transformed cells were subjected to immunoprecipitation with anit-RNF31 and then to immunoblotting with anti-LMP1 and anti-IRF7. The results showed that endogenous RNF31 interacts with both LMP1 and IRF7 (Fig. 1D). Furthermore, we showed that LMP1 interacts with the other two LUBAC components, Sharpin and HOIL1L, in EBV-transformed cells (Fig. 1E).
Taken together, these results demonstrate that RNF31, as a component of LUBAC (37), physically interacts with IRF7 and LMP1 in latently EBV-infected cells.
LUBAC promotes NF-κB activation mainly by conjugating linear ubiquitin chains to NEMO (25, 38). Thus, we sought to evaluate if LUBAC can promote NEMO linear ubiquitination downstream of LMP1 signaling. On the other hand, we have shown that LMP1 stimulates K63-linked polyubiquitination of IRF7 (11, 12). Here, we asked if LMP1 also stimulates linear ubiquitination of IRF7, as we have already shown that IRF7 and RNF31 interact (Fig. 1). To this end, 293T cells were transfected with expression plasmids for Flag-LMP1, Myc-NEMO (or Myc-IRF7), Ub, and LUBAC, which includes Flag-RNF31, HA-HOIL1L, and Flag-Sharpin. Cell lysates were collected 2 days after transfection and denatured before immunoprecipitation. Our results indicated that LUBAC promotes linear ubiquitination of both NEMO and IRF7 and that LMP1 potentiates their ubiquitination (Fig. 2A). LMP1 specifically promotes IRF7 linear ubiquitination, as LMP1 did not enhance IRF3 ubiquitination (Fig. 2A). However, LUBACcs, which consists of the Flag-RNF31 E3-ligase-dead mutant Flag-RNF31 (C699/702S C871/874S) (36), the HA-HOIL1L E3-ligase-dead mutant HA-HOIL1 (C282/285S C447/450S) (36), and the Flag-Sharpin Ub-binding-dead mutant Flag-Sharpin (T358L F359V) (44), did not stimulate ubiquitination of IRF7 (Fig. 2B) or of NEMO (not shown) under all conditions.
To ensure that LMP1 does stimulate IRF7 linear ubiquitination, we used a pair of mouse B cell lines, A20.2J (Rnf31+/+) and H294.10 (Rnf31−/−). We transfected the cells with LMP1, IRF7, and Ub expression plasmids, and cell lysates were collected after 48 h and subjected to denatured immunoprecipitation with the IRF7 antibody clone G-8 (Santa Cruz) and then to immunoblotting with the linear Ub antibody. As expected, LMP1 promoted IRF7 linear ubiquitination in A20.2J (Rnf31+/+) but not in H294.10 (Rnf31−/−) cells (Fig. 2C).
To further verify LMP1-stimulated IRF7 linear ubiquitination, we performed IP with a linear Ub antibody and then probed with an IRF7 antibody. We clearly detected IRF7 that was pulled down by the linear Ub antibody (Fig. 2D) but did not detect IRF3 under our experimental conditions (data not shown).
Taken together, these results demonstrate that LMP1 promotes linear ubiquitination of NEMO and IRF7 via LUBAC, in which RNF31 is the central component (37).
To check if LUBAC plays a potential role in LMP1 activation of NF-κB, we performed promoter-reporter assays by transfecting 293 cells with combinations of Flag-LMP1 and LUBAC (or LUBACcs) plus NF-κB promoter (NF-κBp)–Luc and Renilla luciferase. The results show that LUBAC, but not LUBACcs, significantly increased LMP1 activation of the NF-κB promoter construct (Fig. 3A).
This finding was further confirmed with the mouse B cell lines A20.2J (Rnf31+/+) and H294.10 (Rnf31−/−). In H294.10 (Rnf31−/−) cells, LMP1-stimulated NF-κB promoter activity was significantly lower than that in the parental A20.2J (Rnf31+/+) cells (Fig. 3B). Further, NF-κB activation was evaluated by IκBα phosphorylation at S32/36. The results showed that a significant increase of IκBα S32/36 phosphorylation was detected in A20.2J (Rnf31+/+) cells, but not in H294.10 (Rnf31−/−) cells, in the presence of LMP1 (Fig. 3C). These results indicate that LUBAC is required for full activation of NF-κB by LMP1.
We then performed a promoter-reporter assay to check the effect of LUBAC on LMP1-stimulated IRF7 transcriptional activity. Surprisingly, our results showed that LUBAC dramatically inhibits LMP1-stimulated IRF7 activity (Fig. 4A) and consequently inhibits IFN-β production mediated by the LMP1/IRF7 pathway. However, LMP1 did not stimulate IRF3 activity, and LUBAC had no significant effect on the basal IRF3-mediated IFN-β production (Fig. 4B). These data indicate that LUBAC-mediated linear ubiquitination specifically inhibits LMP1 activation of IRF7. We also confirmed the finding in A20.2J (Rnf31+/+) and H294.10 (Rnf31−/−) cells, and the results showed that LMP1-promoted IRF7 activity was much higher in H294.10 (Rnf31−/−) cells (Fig. 4C).
To assess the role of LUBAC in regulation of LMP1 target gene expression, we knocked down the endogenous RNF31 in IB4 cells by lentivirus-mediated transfection of RNF31-specific short hairpin RNAs (shRNAs). As shown in Fig. 5A, we reached high knockdown efficiencies by two selected RNF31 shRNA constructs. After selection of the cells with puromycin, we performed immunoblotting for IRF7, IRF4, A20, and SOCS1, all of which are known to be upregulated by LMP1 signaling (7, 45, 46). Our results showed that expression of these genes was significantly downregulated in RNF31 shRNA-expressing cells compared with control shRNA-expressing cells (Fig. 5A). We further confirmed the regulation of IRF4 and SOCS1 and evaluated expression of the IRF7 target genes, Tap2 and ISG56, by quantitative real-time (qRT)-PCR. Our results showed that mRNA levels of Tap2 and ISG56 were elevated in RNF31-deficient IB4 cells (Fig. 5A).
We then evaluated the ability of LMP1 to induce its targets in the mouse B cell lines A20.2J (Rnf31+/+) and H294.10 (Rnf31−/−). Our results showed that LMP1 can induce expression of IRF4 and IRF7 in A20.2J (Rnf31+/+) cells but not in H294.10 (Rnf31−/−) cells (Fig. 5B). However, IFN-β mRNA, which was shown to be induced by LMP1 through IRF7 upon superinfection (47), was significantly induced in H294.10 (Rnf31−/−) cells, but not in A20.2J (Rnf31+/+) cells, in the presence of LMP1 and IRF7 (Fig. 5C). These results indicated that LUBAC is required for LMP1 induction of IRF7 expression but inhibits LMP1 activation of IRF7.
We next evaluated RNF31 expression in EBV latency by immunoblotting. Our results showed that RNF31 is correlated with LMP1 in EBV latency, that is, its expression is not detectable in SavI cells (latency 1) but is readily detectable in SavIII cells (latency 3). Both P3HR1 and JiJoye cell lines are latency 3. However, P3HR1 lacks LMP1 expression due to lack of the EBNA2 gene. Correspondingly, RNF31 is expressed at a very low level in P3HR1 cells but at a considerably higher level in JiJoye cells (Fig. 6A).
To assess the functional consequences of LUBAC regulation of LMP1 signaling, we evaluated the proliferation of IB4 cells infected with lentivirus expressing RNF31 shRNAs. The results showed that depletion of RNF31 significantly decreased cell proliferation (by 25%) after 48 h (Fig. 6B).
These results indicate that RNF31 is critical for LMP1 oncogenic functions and for maintenance of EBV latency.
In this study, we revealed a close link between RNF31 and EBV latent infection (Fig. 7). We have shown that RNF31 interacts with LMP1 and IRF7 and stimulates linear ubiquitination of NEMO and IRF7. Consequently, LUBAC is required for LMP1 functions by regulating NF-κB and IRF7 activation.
Posttranslational modifications (PTMs) of proteins by ubiquitin and ubiquitin-like proteins are a pervasive theme important to the activation of NF-κB and other transcription factors in many fundamental cellular processes (48, 49). Many pathogens, such as DNA tumor viruses, encode proteins that mimic, block, or redirect the activity of the ubiquitin-proteasome machinery to modify the cellular environment and protect infected cells from host immune attack (50). EBV encodes several proteins that exploit the host ubiquitin system to regulate its latency and persistence in the host cell (20, 50,–53). For example, EBNA1 resistance to ubiquitination-dependent degradation creates a perfect camouflage to prevent recognition by the host immune system (50). EBNA3C, another latent antigen, targets the tumor suppressor retinoblastoma protein (Rb) for proteasome-dependent degradation through the well-known SCFSkp2 ubiquitin ligase in different systems, including B cells (54). The EBV deubiquitinating enzyme BPLF1 targets TRAF6 and inhibits NF-κB activation during lytic replication (55).
On the other hand, EBV exploits the host ubiquitin system to benefit its latency and oncogenesis. For example, EBNA1 competes with p53 to bind to HAUSP and consequently protect the cell from apoptotic challenge by reducing p53 levels (56). LMP1 controls cellular p53 protein levels by modulating its ubiquitination (57, 58). LMP1 also induces expression of a pool of ubiquitin-related proteins, including A20, which we have shown to inhibit LMP1 activation of IRF7 by functioning as a deubiquitinase (21). In this study, we have shown LMP1 utilizes the host LUBAC-mediated linear ubiquitination to regulate NF-κB and IRF7 activation. In fact, a recent report has shown that TRAF1, which participates in LMP1 CTAR1 signaling to NF-κB activation, is a target for LUBAC and that TRAF1 linear ubiquitination contributes to LMP1-mediated cell growth and survival (59). Thus, LMP1 may employ LUBAC for its benefits via different mechanisms. Our results also show that RNF31 is associated with LMP1 in expression levels. Thus, RNF31 is likely a transcriptional target downstream of LMP1 signaling. We will investigate the mechanism underlying RNF31 expression by LMP1.
Here, we have shown novel and exciting findings that LUBAC interacts with IRF7 and inhibits IRF7 transcriptional activity, although further investigation is needed to elucidate the mechanism. IRF7 is the “master” regulator of type I IFN production in innate immune responses, required not only for early-stage IFN priming, but also for late-stage IFN amplification. Thus, our findings suggest a potential role for LUBAC in modulation of IRF7-mediated IFN response in innate immunity. In fact, in addition to NF-κB activation downstream of pathogen-associated recognition receptors, LUBAC negatively regulates RIG-I-mediated innate immune responses by targeting RIG-I, TRIM25, and IRF3 for degradation (60, 61) and by disrupting the TRAF3-MAVS complex (62). LUBAC also promotes IRF3 linear ubiquitination at two specific sites to activate the RLR-induced IRF-3-mediated pathway of apoptosis (63). Thus, LUBAC seems to be an important player in modulation of innate immune responses through different mechanisms. We will investigate the role of LUBAC in regulating IRF7-mediated IFN responses.
Our results have defined key roles for LUBAC in controlling LMP1 signaling and functions. The findings broaden our knowledge of the interaction between EBV and the host ubiquitin system in shaping its latency and oncogenesis. The role of linear chains in regulating numerous biological functions will advance our current understanding of the novel functions of ubiquitination in signal transduction and may provide novel paradigms for the treatment of human cancers.
LUBAC constructs and mutants were gifts from Kazuhiro Iwai (28, 36). LUBAC is a complex that includes Flag-RNF31, HA-HOIL1L, and Flag-Sharpin. LUBACcs is a complex that includes the Flag-RNF31 E3-ligase-dead mutant Flag-RNF31 (C699/702S C871/874S), the HA-HOIL1L E3-ligase-dead mutant HA-HOIL1 (C282/285S C447/450S), and the Flag-Sharpin Ub-binding-dead mutant Flag-Sharpin (T358L F359V). The Flag-IRF7 expression construct and its mutants (64, 65) and pGL3/IFNA4p-Luc (66) were described previously. 3×Flag-tagged LMP1 and its mutants were described in our recent paper (67). Other constructs with different tags and deletion and point mutants were generated by subcloning or site-directed mutagenesis (Stratagene) and verified by sequencing. Rabbit linear-Ub monoclonal antibody (clone 1E3) was purchased from Millipore. Mouse anti-IRF7 (clone G-8) for immunoprecipitation, rabbit anti-IRF7 (clone H-246) for immunoblotting, goat anti-mouse IgG-horseradish peroxidase (HRP), mouse anti-rabbit IgG-HRP, and mouse anti-goat IgG-HRP were purchased from Santa Cruz. Goat and rabbit polyclonal anti-RNF31 antibodies from Abcam were used for immunoblotting and immunoprecipitation, respectively. Mouse anti-Sharpin, rabbit anti-HOIL1L, and rabbit anti-RNF31 from Sigma were used for immunoblotting. Flag (clone M2) and Myc (clone 9E10) antibodies were from Sigma and Roche, respectively.
The mouse B cell lines A20.2J (Rnf31+/+) and H294.10 (Rnf31−/−) were gifts from Bruce S. Hostager and were described previously (32). 293 and 293T are human kidney epithelial cell lines. IB4, SavIII, JiJoye, and LCL00045 are human B cell lines transformed with EBV. Epithelial cells were cultured with Dulbecco's modified Eagle's medium (DMEM) plus 10% fetal bovine serum (FBS) and antibiotics, and B cells were cultured with RPMI 1640 medium plus 10% FBS and antibiotics. All cell culture supplies were purchased from Life Technologies.
For knockdown of RNF31 in B cell lines, a set of RNF31 shRNAs (4 in total) cloned in pGIPZ/GFP (green fluorescent protein) were purchased from Open Biosystems. We chose two of them that had the highest knockdown efficiencies for our experiments. Their insert sequences were as follows: 5′-TAAACTTGACACCACGCCA-3′ (shRnf31-1) and 5′-GACACATCACCTCCGTGCT-3′ (shRnf31-2). Lentiviral packing, preparation, infection, and selection of stable cells with puromycin were performed as detailed in our previous publication (68). For other transfections of B cells, the Nucleofector kit for human B cells (Lonza) was used. 293 and 293T cells were transfected with Effectene (Qiagen).
293 cells were transfected with expression plasmids as indicated, together with NF-κB–Luc (or IFN-β–Luc) and Renilla luciferase as an internal transfection control. Empty vector was used to equalize the total amounts of DNA in all transfections. Cells were collected 24 h after transfection. Luciferase activity was measured with equal amounts (10% of the total for each sample) of protein lysates with the use of a Dual Luciferase Assay kit (Promega) on a multimode microplate reader (Turner Biosystems). The results are shown as the mean ± standard error (SE) of duplicates for each sample. At least three consistent results were obtained from independent experiments, and representative results are shown. The ability of the empty-vector controls to activate the promoter constructs was set to 1.
For endogenous protein interaction, 1 × 107 cells were used for each IP. For interaction between transiently expressed proteins, 293T cells in 60-mm dishes were collected 48 h after transfection. The cells were lysed with NP-40 lysis buffer (150 mM NaCl, 1% NP-40, 50 mM Tris, pH 8.0, plus protease inhibitors), and cell lysates were subjected to immunoprecipitation with 1.5 μg of the indicated antibodies overnight and then incubated with 40 μl protein A/G beads (Santa Cruz) for 1 h. For ubiquitination detection, a second immunoprecipitation was performed with anti-Myc after denaturing the immunoprecipitated proteins in 1% SDS at 95°C for 5 min. After three washes, proteins on beads were denatured before separation by SDS-PAGE. Immunoblotting was carried out with the indicated antibodies, and signals were detected with an enhanced chemiluminescence (ECL) kit following the manufacturer's protocol (Amersham Pharmacia Biotech).
Quantitative PCR (qPCR) was performed with the use of SYBR green (Applied Biosystems) on a CFX96 real-time PCR detection system (Bio-Rad Laboratories, Inc.). All reactions were run in duplicate. Mean cycle threshold (CT) values were normalized to GAPDH (glyceraldehyde-3-phosphate dehydrogenase), yielding a normalized CT (ΔCT). The ΔΔCT value was calculated by subtracting the respective control from the ΔCT, and the expression level was then calculated by 2 raised to the power of the respective −ΔΔCT value. The results are the average ± SE of duplicates for each sample.
The 3-(4,5-methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) growth assay was used to measure the cell proliferation rate using a Cell Titer 96 nonradioactive cell proliferation assay kit (R&D Systems) according to the manufacturer's recommendations. Basically, 5 × 104 live cells were placed into 85 μl RPMI medium plus 10% FBS in a 96-well plate and incubated with retrovirus expressing RNF31 shRNA or an shRNA control cloned in GIPz/GFP for 0 to 4 days; 15 μl dye solution was added to each well, and the cells were further incubated at 37°C for 4 h. Solubilization solution/stop mix (100 μl) was added, and the cells were incubated for another hour. The absorbance at a wavelength of 570 nm was recorded using a 96-well plate reader (Turner Biosystems). The results are the averages ± SE of duplicates for each sample. The results obtained consistently from at least three independent experiments are shown.
Two-tailed t tests were executed using GraphPad Prism (version 5). A P value of <0.01 was considered significant, and a P value of <0.001 was considered very significant. The data are expressed as means ± SE of duplicate samples, and representative results from at least three independent experiments are shown.
This work was supported by an NIH NIDDK grant to Z.Q.Y. and J.P.M. (R01DK093526), an NIH NIAID grant to Z.Q.Y. and J.P.M. (R01AI114748), an American Society of Hematology Scholar Award to S.N., and in part by NIH grant C06RR0306551.
We thank Bruce S. Hostager for the mouse B cell lines. This publication is the result of work supported with resources of and the use of facilities at the James H. Quillen Veterans Affairs Medical Center.
The contents of this publication do not represent the views of the Department of Veterans Affairs or the United States Government.
We declare that we have no competing interests.