TRIM25 Promotes TNF-α–Induced NF-κB Activation through Potentiating the K63-Linked Ubiquitination of TRAF2

Nuclear factor κB is a pivotal transcription factor that influences cell proliferation, apoptosis, inflammation, and innate immunity. Abnormal regulation of NF-κB signaling has been linked to viral infection, tumorigenesis, autoimmune diseases, and septic shock (1–4). In unstimulated cells, NF-κB is bound to IκBα (the inhibitory protein of NF-κB) and retains in the cytoplasm. Several stimuli, including TNF (TNF-α), IL-1β, and TLRs, can activate NF-κB signaling by regulating downstream molecules, although the regulatory mechanisms are still not well illuminated.

Ubiquitination is an important modified mode of protein that plays critical roles in the regulation of TNF-α–induced NF-κB signaling (5–7). A hallmark following TNF-α stimulation is the activation of NF-κB and the transcription of target genes. The binding of TNF-α to TNFR1 leads to the recruitment of TRADD, and subsequently, TRADD recruits TRAF2, TRAF5, cIAP1, cIAP2, plus RIP1 to form a complex. At this time, these molecules may be targeted by multiple polyubiquitin chains, serving as a scaffold to recruit TAK1, TAB1, and TAB2. The active TAK1 further phosphorylates the MAPKs and IKK complex to initiate MAPK and NF-κB cascades. Next, the IKK complex (composed of IKKα, IKKβ, and NEMO) phosphorylates IκB, which results in the degradation of IκB by proteasome pathway, leading to the nuclear translocation of NF-κB (8, 9). Various types of ubiquitination are involved in the signal transduction. For example, TAB2 and TAB3 prefer to bind to K63-linked ubiquitination of RIP1, which results in the activation of TAK1 (10). TAK1 can be activated by tripartite motif (TRIM) protein 8 through K63-linked ubiquitination (11). The K63-linked ubiquitination of NEMO regulates the assembly of the IKK complex (12, 13). The degradation of IκB derives from its K48-linked ubiquitination. All evidence suggests that ubiquitination plays an important role at different levels of TNF-α–induced NF-κB signaling.

TRIM25, which is an E3-ubiquitin ligase containing a RING finger domain, a B box/coiled-coil domain, and a SPRY domain, is involved in various cellular processes, including cell proliferation and innate immunity (14). Previously, TRIM25 was reported to regulate antiviral innate immunity by delivering K63-linked polyubiquitin chains to RIG-I (15). A recent report found that TRIM25 activated RIG-I signaling relying on the isolated 2CARD of RIG-I, whereas RIPLET could activated full-length RIG-I via dual ubiquitin-dependent and -independent manners, which demonstrated the different mechanisms of RIG-I signaling activation (16). TRIM25 promoted type I IFN production by promoting the degradation of MAVS through the ubiquitin-proteasome system (17). It also has been reported that TRIM25 enhanced K63-linked ubiquitination of TRAF6 to regulate NF-κB signaling through the MDA5–MAVS antiviral signaling axis (18). Another study demonstrated that TRIM25 exerted its function through the TGF-β pathway to influence migration and invasion of cancer cells (19). But the role of TRIM25 is not clear in TNF-α–induced NF-κB signaling pathway and inflammation.

In this study, we identified TRIM25 as a positive regulator in TNF-α–induced NF-κB signaling through a luciferase assay screening system. TRIM25 potentiated the activation of NF-κB by targeting TRAF2 to promote its K63-linked ubiquitination through the RING domain and further enhanced the interaction between TRAF2 and TAK1 or IKKβ. Knockdown of TRIM25 inhibited NF-κB activation and, subsequently, resulted in the weak interaction between TRAF2 and TAK1. To our knowledge, our results have identified an unrecognized role for TRIM25 in the positive regulation of NF-κB signaling by promoting K63-linked ubiquitination of TRAF2 that will be helpful to keep gaining insight into many human diseases, including infection, inflammatory diseases, and cancer.

LPS (L4391-1 MG) and IL-1β (SRP3083) were purchased from Sigma-Aldrich. Recombinant Human TNF-α Protein (10602HNAE25) was purchased from Thermo Fisher Scientific.

HRP–anti-Flag (M2) (A8592) and anti–β-actin (A1978) were purchased from Sigma-Aldrich; HRP–anti-hemagglutinin (12013819001) was purchased from Roche Applied Science; anti-c-Myc-HRP (HT101) was purchased from TransGen Biotech; anti–p-IκBα (2859), anti-IκBα (4814), anti-IKKβ (8943), anti–phospho-IKKα/β (2697), and anti-TRIM25 (13773), TAK1 (5206) were purchased from Cell Signaling Technology; anti-rabbit–HRP IgG (sc-2004), anti-mouse IgG–HRP (sc-2005), and TRAF2 (sc-876, sc-136999) were purchased from Santa Cruz Biotechnology.

Empty vector pcDNA3.1 was from Cui Lab at Sun Yat-sen University. TRIM25, TRAF2, TRAF5, TAK1, TAB1, TAB2, IKKβ, and NEMO genes were cloned from A549 cDNA and subcloned into the pcDNA3.1 vector. The plasmids were confirmed by DNA sequencing at Sangon Biotech.

Human embryonic kidney 293T (HEK293T) cells, HeLa cells, TRAF6-knockout (KO) HEK293T, PBMCs, and THP-1 cells were cultured in DMEM (HyClone) or RPMI 1640 (Life Technologies) with 10% FBS (Genestar), and 1% l-glutamine (Life Technologies) at 37°C in 5% CO₂. HEK293T, HeLa, and THP-1 cells were provided from Dr. X.-F. Qin’s laboratory (Sun Yat-sen University). TRAF6-KO HEK293T cells were kindly provided by Dr. J. Cui (Sun Yat-sen University). Expression plasmids were transfected with Lipofectamine 2000 (Invitrogen), according to the manufacturer’s instructions.

HEK293T cells were plated in 24-well plates at a density of 6 × 10⁴ cells per well. After overnight incubation, the cells were cotransfected with plasmids encoding NF-κB luciferase reporter (30 ng per well) and the internal control Renilla luciferase (10 ng per well) together with 10–25 ng of empty vector as the control or plasmids encoding HA-TRIM25, Flag-TRAF2, Flag-TRAF5, Flag-TAK1, HA-TAB1, Flag-IKKβ, Flag-P65, and other plasmids with indicated target genes. After 24 h of transfection, the cells were treated with TNF-α for 6 h, and these cells were harvested in passive lysis buffer (Promega). The cell lysates were measured using the Dual-Luciferase Assay Kit according to the manufacturer’s protocol (Promega). All the assays were performed in triplicate and repeated a minimum of three times.

HEK293T cells were plated in 24-well plates at a density of 6 × 10⁴ cells per well. Cells were transfected with target plasmids using Lipofectamine 2000 according to the manufacturer’s instructions and extracted in 120 μl of low-salt lysis buffer (50 mM HEPES [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl₂, 10% glycerol, 1% Triton X-100) supplemented with 5 mg/ml protease inhibitor (Thermo Fischer Scientific) and Phosphatase Inhibitor Cocktail (Roche), then samples of 20-μl total proteins were subjected to SDS-PAGE.

For immunoprecipitation (IP) experiments, whole cell extracts were prepared after transfection or stimulation with ligands as the method of immunoblot assay, followed by incubation with the anti-Flag agarose gels (Sigma-Aldrich) overnight at 4°C. The beads were washed five times with low-salt lysis buffer, resuspended with 2× SDS Loading Buffer (FD Biotechnology), and boiled for 5 min. The released proteins were subjected to SDS-PAGE and transferred onto PVDF membranes with subsequent blocking using 5% skim milk. Membranes were incubated with specific Abs and detected using chemiluminescence (Millipore).

Cells were treated with target plasmids or small interfering RNAs (siRNAs) as indicated, then stimulated by ligands in 24-well plates. The total cellular RNA was isolated by TRIzol Reagent (Invitrogen) according to the manufacturer’s protocol, and the first-strand cDNA was synthesized from total RNA using a PrimeScript 1^st strand cDNA Synthesis Kit (Takara Bio). Real-time PCR (RT-PCR) was performed with SYBR Green qPCR Master Mix (Takara Bio). IL-1β, IL-6, and TNF-α induced by NF-κB were analyzed by RT-PCR at 94°C for 5 min, followed by 40 cycles at 94°C for 20 s, at 55°C for 20 s, at 72°C for 20 s, and finally, at 72°C for 10 min. The following specific primers were used for RT-PCR: hGAPDH forward primer, 5′-ACAACTTTGGTATCGTGGAAGG-3′ and hGAPDH reverse primer, 5′-GCCATCACGCCACAGTTTC-3′; hIL-1β forward primer, 5′-ATGATGGCTTATTACAGTGGCAA-3′ and hIL-1β reverse primer, 5′-GTCGGAGATTCGTAGCTGGA-3′; hIL-6 forward primer, 5′-AGAGGCACTGGCAGAAAACAAC-3′ and hIL-6 reverse primer, 5′-AGGCAAGTCTCCTCATTGAATCC-3′; hTNF-α forward primer, 5′-CCAGACCAAGGTCAACCTCC-3′ and hTNF-α reverse primer, 5′-CAGACTCGGCAAAGTCGAGA-3′; TRIM25 forward primer, 5′-AGCAGCTACAACAAGAATACACG-3′ and TRIM25 reverse primer, 5′-GGCTCTGTTCAATCTCCTCCT-3′.

LipoRNAiMax (Invitrogen) was used according to the manufacturer’s protocols for transfection of siRNAs into HEK293T cells, HeLa cells, THP-1 cells, or PBMCs. The sequences of TRIM25 siRNAs are as follows: human TRIM25-specific siRNA 1) sense sequencing (SenseSeq): 5′-AGGUCCACCUGAUGUAUAATT-3′ and antisense sequencing (AntiSeq): 5′-UUAUACAUCAGGUGGACCUTT-3′; 2) SenseSeq: 5′-CCCUGAGGCACAAACUAACTT-3′ and AntiSeq: 5′-GUUAGUUUGUGCCUCAGGGTG-3′; and 3) SenseSeq: 5′-GCAAAUGUUCCCAGCACAATT-3′ and AntiSeq: 5′-UUGUGCUGGGAACAUUUGCTT-3′.

Data were compared between the different test groups using a Student t test and GraphPad Prism 5.0 Software. All the experiments were repeated at least three times independently, and the differences between groups were considered significant when *p < 0.05, **p < 0.01, and ***p < 0.001.

To investigate the roles of TRIM family proteins in NF-κB signaling induced by TNF-α, we screened the function of a panel of TRIM proteins using luciferase assays. We transfected HEK293T cells with an NF-κB luciferase reporter and an internal control Renilla thymidine kinase luciferase, as well as one of the candidate genes encoding TRIM, and stimulated them with TNF-α. Notably, TRIM8, TRIM13, and TRIM38, which were previously reported to positively or negatively modulate NF-κB signaling by targeting TAK1, NEMO, or TAB2/3, were identified as hits in the screening (Fig. 1A), thus validating this experimental approach (11, 20, 21). These data also led to the identification of TRIM25 as a potent activator of NF-κB signaling. To further determine whether TRIM25 functions in TLR- and/or cytokine-mediated NF-κB activation, we transfected HEK293T or HEK293T/TLR4 cells with an NF-κB luciferase reporter and the internal control Renilla luciferase, as well as an increasing amount of TRIM25 plasmid, and then stimulated the cells for 6 h with TNF-α, IL-1β, or LPS. As shown in Fig. 1B–D, TRIM25 potently activated NF-κB activation induced by TNF-α, IL-1β, or LPS in a dose-dependent manner. It is reported that TRIM25 can interact TRAF6 and enhance its K63-linked ubiquitination, so we speculated that TRIM25 might regulate IL-1β– or LPS-mediated NF-κB signaling through TRAF6 (18). Next, we used TNF-α to treat TRAF6 KO cells and found TRIM25 also enhances TNF-α–induced NF-κB signaling (Fig. 1E), indicating that TRIM25 activated TNF-α–induced NF-κB in a TRAF6-independent manner.

Next, we used no-tagged or differently tagged TRIM25 to determine the function of TRIM25, and the luciferase assay showed that they had the same function in activating TNF-α–mediated NF-κB–luc activation (Fig. 1F). More importantly, we found that TRIM25 significantly enhanced the degradation of endogenous IκBα protein in the presence of TNF-α (Fig. 1G). In addition, we found that TNF-α treatment resulted in higher expression levels of TNF-α, IL-6, or IL-1β mRNA in HeLa cells with TRIM25-overexpression than in those transfected with empty vector (Fig. 1H). Taken together, these results suggest that TRIM25 activates TNF-α–induced NF-κB activation.

We next detected whether endogenous TRIM25 was involved in the regulation of TNF-α–mediated NF-κB signaling. We generated three pairs of siRNAs specific for TRIM25, all of which efficiently inhibited the expression of endogenous TRIM25 in protein and mRNA levels (Fig. 2A, 2B). As expected, we found that treatment with TNF-α resulted in much lower p-IKK and p-IκBα in HEK293T cells transfected with TRIM25-specific siRNA than in those transfected with control siRNA (Fig. 2C). In RT-PCR assays, we found that TNF-α treatment resulted in lower expression levels of TNF-α, IL-6, or IL-1β mRNA in TRIM25-deficient HeLa cells or PBMCs (Fig. 2D, 2E). Hence, these data suggest knockdown of TRIM25 inhibits NF-κB activity.

To determine the molecular mechanisms by which TRIM25 potentiates TNF-α–induced NF-κB signaling, we transfected HEK293T cells with TRAF2, TRAF5, TAK1-TAB1, IKKβ, or p65 subunit together with TRIM25 plus the NF-κB luciferase reporter. We found that the activation of NF-κB by TRAF2 and TRAF5 was markedly enhanced by TRIM25 (Fig. 3A). In contrast, TRIM25 did not increase TAK1-TAB1–, IKKβ-, or p65-mediated NF-κB activation (Fig. 3A), suggesting that TRIM25 potentiates the NF-κB pathway upstream of TAK1, most likely targeting the TRAF complex. Consistent with these results, we found that knockdown of TRIM25 inhibited NF-κB luciferase activity induced by TRAF2 and TRAF5 but not TAK1-TAB1, IKKβ, or p65 (Fig. 3B). These results suggest that TRIM25 enhances NF-κB signaling upstream of TAK1, at the level of the TRAF complex.

Next, we sought to determine whether TRIM25 could directly interact with the TRAF complex or other signaling proteins within the NF-κB signaling pathway. Coimmunoprecipitation experiments revealed that TRIM25 specifically interacted with TRAF2 rather than the TRAF5 or TAK1 complex (Fig. 4A, 4B). To determine the interaction, we treated HeLa cells with TNF-α and then collected the cell lysates. IP and immunoblot analysis revealed that TRIM25 only weakly interacted with TRAF2 in unstimulated cells. However, TRIM25 strongly interacted with TRAF2 after TNF-α treatment (Fig. 4C). We also found that endogenous TRIM25 interacted with TRAF2 in unstimulated THP-1 cells and PBMCs, and the treatment of TNF-α could enhance their interaction (Fig. 4D, 4E). To identify the domain of TRIM25 responsible for its interaction with TRAF2, we generated three TRIM25 deletion constructs (Fig. 4F), and we found that only ∆RING domain (KD) interacted with TRAF2 (Fig. 4G). It suggests the SPRY domain of TRIM25 is critical for binding to TRAF2. In addition, we found NF-κB activation, as assessed by a luciferase assay, was increased in the cells transfected with TRIM25 (wild type [WT]) but not its deletion constructs (Fig. 4H). It should be noted that both the RING and SPRY domain were critical for TRIM25-mediated enhancement of TNF-α–induced NF-κB signaling.

Because the RING domain of TRIM25 is critical for its function (Fig. 4H), we examined whether the E3 ligase activity of TRIM25 was essential for its function. An enzymatically inactive mutant, TRIM25-CS (C53S), could not facilitate NF-κB reporter activation (Fig. 5A), suggesting that TRIM25 may affect the NF-κB pathways through its E3 ligase activity. Next, we tested whether TRIM25 affected the ubiquitination of TRAF2. We observed that TRIM25 remarkably enhanced the ubiquitination of TRAF2 rather than TRIM25-CS mutant or TRIM25 ΔRING domain (Fig. 5B), and knockdown of TRIM25 limited the ubiquitination of TRAF2 (Fig. 5C). Next, we want to know which kinds of TRAF2 ubiquitination can be affected by TRIM25. Our findings revealed that TRIM25 only enhanced K63-ubiquitination of TRAF2 but not others (Fig. 5D). This finding was further confirmed in the knockdown system (Fig. 5E). Collectively, these results suggest that TRIM25 enhances the K63-linked ubiquitination of TRAF2 using its E3 ligase activity.

After TNF-α treatment, TRAF2 could recruit TAK1 and IKK complex to influence NF-κB signaling, and TRIM25 targeted TRAF2 through enhancement of its ubiquitination. Therefore, we reasoned that TRIM25 may bridge the TRAF2 and TAK1 or the IKK complex. As expected, we found that TRIM25 markedly enhanced the interaction between TRAF2 and TAK1 or IKKβ but did not affect the IKK complex (Fig. 6A–C). TRIM25 deficiency blocked the interaction between TRAF2 and TAK1 (Fig. 6D). Next, we found that knockdown of TRIM25 reduced the endogenous interaction between TRAF2 and TAK1 after TNF-α stimulation (Fig. 6E). We further found that TRIM25-∆RING or -CS mutant failed to bridge TRAF2 to TAK1 (Fig. 6F) or to facilitate TRAF2-mediated TAK1 activation (Fig. 6G). Collectively, these results indicated that the K63-linked ubiquitination of TRAF2 mediated by TRIM25 provided the platform to bridge TRAF2 to TAK1 or IKK complex.

TNF-α, as one of the pleiotropic cytokines that activate NF-κB signaling, leads to the production of several proinflammatory cytokines, such as IL-1β, IL-6, and TNF-α, and further promotes inflammation and immune responses. Increasing evidence indicates that dysregulation of NF-κB leads to several inflammation-associated diseases. Further, proinflammatory cytokines secreted by innate immune cells have been shown to play a pivotal role in tumor development and progression in chronic inflammation conditions, so it is critical for the healthy host to tightly regulate TNF-α–induced NF-κB pathway (1, 3). Our study identified TRIM25 as a positive regulator in TNF-α–induced NF-κB signaling pathway, which might shed light on identifying a potential therapeutic target for inflammation-associated diseases and tumors in the future.

The posttranslation modification of target proteins is important in the regulation of the NF-κB pathway. Increasing numbers of proteins that contain many members of TRIM family were reported to participate in the regulation of the TNF-α–induced NF-κB pathway. Just as reported, TRIM8 associates with TAK1, enhances the binding of K63-linked polyubiquitination to TAK1, and positively regulates TNF-α– and IL-1β–triggered NF-κB activation (11). TRIM13 regulates ubiquitination and turnover of NEMO at the level of IKK complex to suppress the TNF-α–induced NF-κB pathway and also potentiates TLR2-mediated NF-κB activation via K29-linked polyubiquitination of TRAF2 (20, 22). TRIM12c, as a mouse homologue of TRIM5, forms a complex with TRAF6 to enhance the activation of the IFN and NF-κB pathways (23). After stimulation by TNF-α, TRIM22, TRIM38, and TRIM39 inhibit NF-κB pathways by targeting different proteins through different manners of ubiquitination (21, 24, 25). There are still many proteins, such as TRIM45, TRIM52, RACK1, and so on, that have regulated functions in TNF-α–induced NF-κB pathway (26–28). The involvement of so many molecules in the regulation of NF-κB signaling suggests the importance of maintaining its fine regulation. That is to say, the complicated regulatory network of NF-κB signaling is required to ensure an appropriate immune response.

TRIM25, a member of the TRIM proteins family, is involved in cell proliferation, innate immunity, and cancer. It was reported that TRIM25 affected RIG-I–mediated antiviral activity by delivering the K63-linked ubiquitin moiety to the N-terminal CARDs of RIG-I or releasing NEMO and TBK1 complex into the cytosol through the proteasome degradation of scaffold protein MAVS (17, 18). However, there was no report about TRIM25 directly regulating TNF-α–induced NF-κB signaling. Unlike TRIM25 regulates NF-κB signaling by TRAF6 through MDA5-MAVS antiviral signaling axis (18), we found TRIM25 facilitated NF-κB activity induced by TNF-α, TRAF2, and TRAF5 but not TAK1-TAB1, IKKβ, or p65. Also, it enhanced the degradation of IκBα and the transcription of proinflammatory cytokines, such as IL-1β, IL-6, and TNF-α, in response to TNF-α treatment. Immunoprecipitation analyses showed that TRIM25 directly interacted with TRAF2 but not with TRAF5 or TAK1 complex and increased the K63-linked ubiquitination of TRAF2. Conversely, knockdown of TRIM25 reduced the ubiquitination reaction to TRAF2. We also confirmed the endogenous interaction of TRAF2 and TRIM25 by IP analyses in THP-1 cells and PBMCs, and the interaction was further enhanced with the treatment of TNF-α. To identify the detailed mechanism that TRIM25 potentiates TNF-α–induced NF-κB signaling by targeting TRAF2, we confirmed TRIM25 reinforced the interaction between TRAF2 and TAK1 or IKKβ; it merely had no function to the interaction between IKKβ and NEMO. The enhanced interaction potentiated the activation of TAK1 and IKKβ to heighten TNF-α–induced NF-κB signaling, whereas TRIM25-∆RING or TRIM25-CS mutant could affect the interaction of TRAF2 and TAK1, and knockdown of TRIM25 reduced the endogenous interaction between TRAF2 and TAK1 after TNF-α stimulation. In this context, it has been reported that TRIM8 upregulates TNF-α– and IL-1β–triggered NF-κB signaling by targeting TAK1. USP18 inhibits NF-κB signaling by targeting TAK1 and NEMO (11, 29). From this study, we also found that TRIM25 had the interaction with itself by IP, which might be helpful to its stabilization, and then strengthened immune reaction. Previous studies showed that TRAF2 participated in the apoptosis pathway, just as TRAF2 is a negative regulator of TNF-α–induced apoptosis (30), RIPK1 inhibits LPS/TNF–mediated apoptosis through a TRAF2- and caspase-8–dependent pathway (31). Hence TRIM25 may influence the apoptosis pathway by targeting TRAF2, which may link the cross-talk between the NF-κB and apoptosis signaling pathway for a further study.

In conclusion, to our knowledge, the evidence in this study identified an undiscovered role for TRIM25 in the regulating of TNF-α–induced NF-κB signaling. In the current study, TRIM25 mainly acted at the level of TRAF2 by K63-linked ubiquitination, then strengthened the interaction between TRAF2 and TAK1 or IKKβ. Thus, our findings gain an insight into the molecular mechanisms by which TRIM25 upregulates NF-κB signaling and provide a potential therapeutic target for inflammatory diseases and cancers in the future.

We thank Dr. Xiao-Feng Qin and Dr. Jun Cui for providing plasmids and reagents as a gift. We also thank Dr. Weicheng Liang for providing language support.

The authors have no financial conflicts of interest.