TRAF Family Member-associated NF-κB Activator (TANK) Inhibits Genotoxic Nuclear Factor κB Activation by Facilitating Deubiquitinase USP10-dependent Deubiquitination of TRAF6 Ligase*

Background: TANK has been shown to promote deubiquitination of TRAFs, but the mechanism involved is unclear. Results: TANK associates with MCPIP1 and USP10, thereby diminishing TRAF6 ubiquitination, which inhibits NF-κB activation by genotoxic stress and IL-1R/TLR activation. Conclusion: The deubiquitinating activity of TANK is achieved by its association with USP10. Significance: Modulating TANK may enhance cancer cell sensitivity to chemotherapy. DNA damage-induced NF-κB activation plays a critical role in regulating cellular response to genotoxic stress. However, the molecular mechanisms controlling the magnitude and duration of this genotoxic NF-κB signaling cascade are poorly understood. We recently demonstrated that genotoxic NF-κB activation is regulated by reversible ubiquitination of several essential mediators involved in this signaling pathway. Here we show that TRAF family member-associated NF-κB activator (TANK) negatively regulates NF-κB activation by DNA damage via inhibiting ubiquitination of TRAF6. Despite the lack of a deubiquitination enzyme domain, TANK has been shown to negatively regulate the ubiquitination of TRAF proteins. We found TANK formed a complex with MCPIP1 (also known as ZC3H12A) and a deubiquitinase, USP10, which was essential for the USP10-dependent deubiquitination of TRAF6 and the resolution of genotoxic NF-κB activation upon DNA damage. Clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9-mediated deletion of TANK in human cells significantly enhanced NF-κB activation by genotoxic treatment, resulting in enhanced cell survival and increased inflammatory cytokine production. Furthermore, we found that the TANK-MCPIP1-USP10 complex also decreased TRAF6 ubiquitination in cells treated with IL-1β or LPS. In accordance, depletion of USP10 enhanced NF-κB activation induced by IL-1β or LPS. Collectively, our data demonstrate that TANK serves as an important negative regulator of NF-κB signaling cascades induced by genotoxic stress and IL-1R/Toll-like receptor stimulation in a manner dependent on MCPIP1/USP10-mediated TRAF6 deubiquitination.

DNA damage-induced NF-B activation plays a critical role in regulating cellular response to genotoxic stress. However, the molecular mechanisms controlling the magnitude and duration of this genotoxic NF-B signaling cascade are poorly understood. We recently demonstrated that genotoxic NF-B activation is regulated by reversible ubiquitination of several essential mediators involved in this signaling pathway. Here we show that TRAF family member-associated NF-B activator (TANK) negatively regulates NF-B activation by DNA damage via inhibiting ubiquitination of TRAF6. Despite the lack of a deubiquitination enzyme domain, TANK has been shown to negatively regulate the ubiquitination of TRAF proteins. We found TANK formed a complex with MCPIP1 (also known as ZC3H12A) and a deubiquitinase, USP10, which was essential for the USP10-dependent deubiquitination of TRAF6 and the resolution of genotoxic NF-B activation upon DNA damage. Clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9-mediated deletion of TANK in human cells significantly enhanced NF-B activation by genotoxic treatment, resulting in enhanced cell survival and increased inflammatory cytokine production. Furthermore, we found that the TANK-MCPIP1-USP10 complex also decreased TRAF6 ubiquitination in cells treated with IL-1␤ or LPS. In accordance, depletion of USP10 enhanced NF-B activation induced by IL-1␤ or LPS. Collectively, our data demonstrate that TANK serves as an important negative regulator of NF-B signaling cascades induced by genotoxic stress and IL-1R/Toll-like receptor stimulation in a manner dependent on MCPIP1/USP10mediated TRAF6 deubiquitination.
DNA-damaging agents such as genotoxic chemotherapeutics and ionizing radiation are widely used as the mainstay of cancer therapy to treat various types of human malignancies such as breast cancer, prostate cancer, and colon cancer. However, acquired therapeutic resistance to the genotoxic agents significantly limits the efficacy of cancer treatments, resulting in cancer relapse and metastasis. DNA damage-induced activa-(PIASy), and the PIDD (p53-induced protein with a death domain)/RIP1 (receptor-interacting protein 1) complex (10 -12). Then SUMOylated NEMO could further interact with the DNA damage response apical kinase ATM (ataxia telangiectasia mutated), which directly phosphorylates and facilitates cIAP1/2-mediated monoubiquitination of NEMO (13,14). The nuclear monoubiquitination of NEMO promotes the export of the NEMO-ATM complex into the cytoplasm, where it further activates the cytoplasmic IKK and downstream NF-B signaling pathways. Moreover, Lys-63-linked polyubiquitination of ELKS, RIP1, and TRAF6 as well as the M1-linked linear ubiquitination of NEMO have been shown to modulate TAK1/IKK kinase complex activation in the cytoplasm upon genotoxic stimulation (15)(16)(17)(18).
The extent and duration of NF-B activation is controlled by a dynamic balance between activation signaling and negative regulatory mechanisms, leading to the resolution of NF-B activation (19). Compared with signaling events mediating NF-B activation upon DNA damage, even less was known about how this signaling cascade is resolved. Lee et al. (20) found that a Sentrin/SUMO-specific protease, SENP2, was upregulated in response to genotoxic NF-B activation, which served as a negative feedback response to inhibit NF-B activation by attenuating NEMO SUMOylation in response to genotoxic stress. We showed recently that NF-B-dependent MCPIP1 (also known as ZC3H12A) induction negatively regulated the genotoxic NF-B signaling cascade by promoting USP10-mediated deubiquitination of NEMO, resulting in decreased NF-B activation upon DNA damage (21). Nevertheless, genetic deletion of either SENP2 or MCPIP1 in MEF cells was not sufficient to completely block the resolution of genotoxic NF-B activation, suggesting that additional negative regulatory mechanisms controlling genotoxic NF-B signaling remain to be elucidated.
TRAF family member-associated NF-B activator (TANK, also known as I-TRAF) could interact with the TRAF family members TRAF2 and TRAF3, thereby regulating TRAF-mediated signaling pathways (22)(23)(24). In the antiviral immune response following retinoic acid-inducible gene 1 activation, TANK may serve as an adaptor bridging TRAF3 with TBK1 and IKK⑀, which promotes phosphorylation and activation of IRF3/ IRF7 as well as induction of NF-B activation, leading to effective type I IFN production (25)(26)(27). Nevertheless, TANK has also been shown to negatively regulate NF-B activation (28,29). It has been found that NF-B activation upon TLR or BCR (B cell receptor) stimulation was augmented in macrophages and B cells isolated from Tank Ϫ/Ϫ mice compared with their wild-type counterparts. Moreover, TANK deficiency increased TRAF6 ubiquitination in response to TLR activation in macrophages, which may contribute to enhanced NF-B activation. However, the mechanism by which TANK inhibits TRAF6 ubiquitination remains enigmatic because TANK lacks the deubiquitination enzyme (DUB) domain and TANK does not interact with A20 and CYLD (cylindromatosis), two common DUBs that have been shown to inhibit NF-B signaling (28).
In this report, we show that TANK inhibits genotoxic stressinduced NF-B activation, which may be dependent on inhibition of TRAF6 ubiquitination . The TRAF2/3-interacting   domain of TANK and the TRAF-C domain of TRAF6 are essential for TANK association with TRAF6. Mechanistically, TANK  forms a complex with MCPIP1 and USP10, which promotes  USP10-dependent deubiquitination of TRAF6 in cells treated  with genotoxic stimuli. TANK is indispensable for TRAF6 deubiquitination by USP10 because deletion of TANK disrupts the  association between USP10 and TRAF6, thereby stabilizing TRAF6 ubiquitination upon DNA damage. Moreover, the TANK-MCPIP1-USP10 complex is also responsible for inhibition of TRAF6 ubiquitination in cells treated with LPS or IL-1␤. Collectively, our data suggest that TANK may serve as the adaptor for recruiting the USP10-MCPIP1 complex to polyubiquitinated TRAF6, which, in turn, disassembles the polyubiquitin chains anchored on TRAF6, resulting in the termination of NF-B activation in response to genotoxic stress and IL-1R/ TLR activation.

Experimental Procedures
Cells, Plasmids, and Reagents-HEK293T cells, mouse embryonic fibroblast cells (wild-type and MCPIP1 Ϫ/Ϫ ), and human breast cancer cell line MDA-MB-231 cells were grown in DMEM supplemented with 10% fetal bovine serum. All cell lines were maintained in the presence of penicillin (100 IU/ml) and streptomycin (100 mg/ml) at 37°C, 5% CO 2 . The HA-TANK plasmid was generated by cloning the human TANK coding sequence into the pcDNA3 vector. The TANK siRNA sequence used was UCACUUCAACAGACUAUU-AUU, as reported previously (30). Two siRNAs targeting USP10 were used. One was synthesized with the sequence CCC UGA UGG UAU CAC UAA AGA UU, and the other was purchased from Cell Signaling Technology (catalog no. 7747). TANK or USP10 knockout cells were generated with the CRISPR-Cas9 system. The pX330-U6-Chimeric_BB-CBh-hSp-Cas9 vector was purchased from Addgene (catalog no. 42230). TANK-targeting constructs were generated as described previously (31). The small guide RNA sequences targeting TANK were AGCGTATGAAGCCTTCCGGCAGG (protospacer adjacent motif, PAM) and ATCCATGCATGCCTGCCGGA-AGG. The sgRNA (single-guide RNA) sequence targeting USP10 was GACTCCTCGATCTTCAGTTGAGG (PAM EMSA-The Ig-B oligonucleotide probe and conditions for EMSA have been described previously (32). The Oct-1 oligonucleotide used for control EMSA reactions was obtained from Promega. Gels were quantified with a Cyclone PhosphorImager (PerkinElmer Life Sciences).
Luciferase Assay-HEK293T cells were cotransfected with the pGL4.0 Luciferase reporter containing the NF-B target sequence and Tk-Rluc reporter (Promega). After 48 h, cells were treated and lysed, and the activity of firefly/Renilla luciferase in the lysates was measured with the Dual-Luciferase assay system (Promega).
To detect protein ubiquitination, cells were lysed with 1% SDS in IP lysis buffer at 95°C for 30 min. The cell lysates were then diluted with IP lysis buffer to reduce SDS to 0.1% and mixed with primary antibodies and protein G-Sepharose for incubation at 4°C overnight. The ubiquitin modification of precipitated proteins was examined by immunoblotting.
Cell Survival Assay-Cells were transfected with plasmids and treated as indicated. Then cells were stained with trypan blue, and live cells were counted with a TC20 automated cell counter (Bio-Rad). Data from three independent experiments were collected and plotted.
Statistical Analysis-The results are presented as mean Ϯ S.D. Student's t test was applied to analyze the significance. p Ͻ 0.05 was defined as statistically significant.

TANK Negatively Regulates Genotoxic Stress-induced NF-B
Activation-TANK has been shown to inhibit TLR-induced NF-B activation by attenuating TRAF6 polyubiquitination (28). Interestingly, TRAF6 ubiquitination has been reported to mediate genotoxic stress-induced NF-B activation (15). We reasoned that TANK may also negatively regulate genotoxic NF-B activation. To test this hypothesis, we overexpressed TANK in HEK293T cells and measured NF-B activation in response to TNF-␣ and genotoxic agents such as etoposide (VP16), camptothecin (CPT), and doxorubicin. As expected, TANK overexpression significantly decreased genotoxic agentinduced NF-B activation, as measured by EMSA (Fig. 1A) and luciferase reporter assay (Fig. 1B). Consistently, VP16 or CPTinduced IKK activation, IB␣ degradation, and RelA/p65 phosphorylation were abrogated in cells overexpressing TANK ( Fig.  1, C and D), suggesting that increased TANK expression diminishes the NF-B signaling cascade activated by genotoxic stimulation.
To further validate the role of TANK in modulating the genotoxic NF-B signaling pathway in human cells, we took advantage of the recently developed CRISPR-Cas9 gene-targeting tools (34) and generated two TANK-targeting CRISPR constructs harboring distinct sgRNAs as reported previously (31). We found both TANK-targeting CRISPR constructs effectively disrupted TANK gene expression in HEK293T cells (Fig. 1E, lanes 4 -6 and 10 -12), to which we hereafter refer as TANK KO cells. Accordingly, TANK deletion significantly enhanced NF-B activation by VP16 or CPT, which can be attenuated by reconstitution of TANK in TANK KO cells (Fig. 1E). In addition to chemotherapeutic drugs, ionizing radiation-induced NF-B activation was also attenuated or escalated in cells overexpressing TANK or with TANK deletion, respectively (Fig.  1F). To further confirm the results from TANK KO cells, we used siRNA to knock down TANK in HEK293T cells. Consistently, suppression of TANK expression by siRNA also substantially enhanced NF-B activation by genotoxic drugs (Fig. 1G). Moreover, NF-B-dependent transcription upon DNA damage was further enhanced in TANK KO cells, which could be attenuated by overexpressing TANK (Fig. 1H). Taken together, our data indicate that TANK is a negative regulator of the NF-B signaling cascade in cells exposed to genotoxic stress.

MCPIP1 Is Required for TANK-dependent Repression of NF-B Activation by DNA Damage-Previous studies have
indicated that TANK-associated TBK1 could inhibit NF-B signaling by phosphorylating the canonical IKK complex (35,36). We found that pharmacological inhibition of TBK1 with amlexanox did not affect TANK-dependent inhibition of genotoxic NF-B activation (data not shown), suggesting that additional TANK-associated partners may be involved in TANKmediated NF-B inhibition. The association between TANK and NEMO has been reported previously (37). Moreover, TANK has also been identified as a MCPIP1-associated protein in a yeast two-hybrid screen (38). We confirmed the association between TANK and NEMO or MCPIP1 in cells treated with VP16 (Fig. 2, A and B). Interestingly, TANK overexpression was able to inhibit NF-B activation by genotoxic agents in wildtype but not in MCPIP1-deficient MEFs (Fig. 2C), suggesting that MCPIP1 may be required for TANK to inhibit genotoxic NF-B activation. Reciprocally, although overexpressing MCPIP1 was able to partially decrease NF-B activation by genotoxic drugs in TANK KO cells (Fig. 2D, compare lanes 11 and 12 with lanes 8 and 9), the extent of MCPIP1-dependent NF-B suppression was substantially dampened. These data indicate that MCPIP1 and TANK are mutually required for effective suppression of genotoxic NF-B activation by overexpressing either MCPIP1 or TANK.
To further characterize the interaction between MCPIP1 and TANK, we generated a series of MCPIP1 truncation/deletion mutants (    further suggested that the N-terminal fragment (amino acids 1-31) of TANK was essential for the association between TANK and MCPIP1.
TANK-associated USP10 Is Essential for TANK to Inhibit Genotoxic NF-B Activation-We showed recently that MCPIP1 association with USP10, a bona fide DUB, is required for MCPIP1-dependent inhibition of genotoxic NF-B signaling (21). USP10 removes the linear ubiquitin chain attached to NEMO en bloc, thereby attenuating NF-B activation signaling upon DNA damage. Because we found that TANK formed a complex with MCPIP1, and both were mutually required for NF-B inhibition, we asked whether USP10 is involved in TANK-dependent negative regulation of NF-B. To this end, we depleted USP10 with siRNA in HEK293T cells. We found that USP10 knockdown almost abrogated TANK-dependent inhibition of NF-B activation in VP16-treated cells (Fig. 3A), indicating that USP10 is critical for TANK to negatively regulate NF-B signaling upon DNA damage. We further confirmed the inducible association between TANK and USP10 in cells treated with VP16 ( Fig. 3B), which suggested that TANK-MCPIP1-USP10 may form a functional complex in response to genotoxic stress. Further coimmunoprecipitation analyses showed that TANK was dispensable for the interaction between MCPIP1 and USP10, whereas MCPIP1 was essential for TANK to interact with USP10 (Fig. 3, C and D). Therefore, MCPIP1 likely served as a bridge to anchor both TANK and USP10 into the same complex.
USP10 Is Required for TANK-promoted Deubiquitination of TRAF6 -In response to TLR activation, TRAF6 ubiquitination was enhanced in mouse cells deficient of TANK (28). We found that VP16-induced TRAF6 polyubiquitination was decreased significantly in cells overexpressing TANK (Fig. 4A, lane 3), suggesting that TANK may also promote TRAF6 deubiquitination in response to genotoxic stress. This notion was supported by our observation that TRAF6 ubiquitination was further enhanced in TANK KO cells upon VP16 treatment (Fig. 4B). In parallel, we found that overexpression of wild-type USP10, but not its catalytic-inactive mutant (C424A), also abolished VP16induced TRAF6 polyubiquitination (Fig. 4A, lanes 4 and 5). Considering the lack of a DUB domain in TANK and the inducible association between TANK and USP10, we speculated that USP10 may facilitate TANK-dependent deubiquitination. As expected, we found that overexpression of TANK failed to decrease genotoxic stress-induced TRAF6 polyubiquitination in USP10-depleted cells (Fig. 4C). In accordance, we found that VP16-induced TRAF6 ubiquitination was comparably enhanced in cells deficient of TANK or USP10 (Fig. 4D). Reciprocally, overexpression of USP10 was able to diminish TRAF6 polyubiquitination in wild-type cells, but not in TANK KO cells, upon genotoxic treatment (Fig. 4E, compare lanes 3  and 4). Moreover, the defect of USP10 in deubiquitinating TRAF6 in TANK KO cells can be rescued by ectopic expression of TANK (Fig. 4E, compare lanes 4 and 5). These results indicate that USP10 is the essential DUB for TANK to inhibit DNA damage-induced ubiquitination of TRAF6, whereas TANK is indispensable for TRAF6 deubiquitination by USP10.
The Interaction between TANK and TRAF6 Was Required for Inhibition of Genotoxic Stress-induced NF-B Activation-The indispensable role of TANK in mediating USP10-dependent  TRAF6 deubiquitination suggested that TANK may directly interact with TRAF6, which recruits USP10 to the ubiquitinated TRAF6. TANK has been shown to interact with TRAF1/ 2/3 via a TRAF-binding domain located at the center region (24). Our in vitro pulldown assay indicated that the TRAF-binding domain is also required for TANK association with TRAF6 (Fig. 5A). A C-terminal region of TANK, which was identified as part of NEMO-binding domain (37), was also important for the TANK-TRAF6 interaction. In accordance, deletion of the TRAF-binding domain (amino acids 170 -191) completely abrogated the interaction between TANK and TRAF6 (Fig. 5B). The essential region required for TRAF6 to interact with TANK was mapped to the TRAF-C domain (Fig. 5C), which is consistent with that identified in other TRAF family members (23,24). To examine the role of TANK in mediating TRAF6 and USP10 interaction, we performed co-IP analyses in both wild-type and TANK KO cells. As we expected, the genotoxic treatment-induced association between TRAF6 and USP10 was abolished in TANK KO cells (Fig. 5D). Similarly, genotoxic treatment also promoted the interaction between TRAF6 and MCPIP1, which was diminished by TANK deletion. In contrast, MCPIP1 deficiency did not affect the interaction between TANK and TRAF6 in cells exposed to genotoxic drugs (Fig. 5E). Nevertheless, TANK KO cell reconstitution with the MCPIP1 binding-deficient TANK mutant (31-425, Fig. 2G) failed, whereas TANK WT reconstitution was sufficient to suppress TRAF6 ubiquitination by VP16 (Fig. 5F). These data suggested that TANK served as an essential adaptor to assemble MCPIP1-USP10 and TRAF6 into the same complex, thereby diminishing TRAF6 polyubiquitination. In support of this notion, we found that the TRAF-binding deficient mutant of TANK failed to inhibit TRAF6 ubiquitination in cells treated with VP16 (Fig. 5G). Moreover, only wild-type TANK, but not its TRAF binding-deficient mutant, was able to attenuate NF-B activation by genotoxic drugs (Fig. 5H), indicating that TANK association with TRAF6 plays a critical role in mediating TANK-MCPIP1-USP10-dependent inhibition of TRAF6 ubiquitination and subsequent NF-B activation upon DNA damage. TANK also Facilitates the Deubiquitination of NEMO in Response to Genotoxic Stress-We found that NEMO was attached to the linear ubiquitin chain upon DNA damage, which is required for genotoxic NF-B activation (18). In a negative feedback response, NF-B-up-regulated MCPIP1 could facilitate the USP10-dependent removal of the linear ubiquitin chain from NEMO, thereby inhibiting NF-B activation upon DNA damage (21). Because we found that genotoxic treatment also induced the interaction between NEMO and TANK ( Fig.  2A), we speculate that TANK may also regulate NEMO ubiquitination by genotoxic stress. Indeed, we found that VP16induced NEMO ubiquitination was substantially enhanced in TANK KO cells (Fig. 6A), suggesting that TANK may inhibit NEMO ubiquitination. This TANK-mediated suppression of NEMO ubiquitination is specific because we did not detect a significant change in ELKS ubiquitination, another signaling event required for genotoxic NF-B activation (16), in TANK KO cells upon genotoxic stimulation (Fig. 6B). However, TANK deletion only partially reduced USP10-dependent NEMO deubiquitination (Fig. 6C, compare lanes 2 and 4). Consistently, we were able to detect an interaction between MCPIP1-USP10 and NEMO in TANK KO cells, although at a reduced level (Fig. 6D). All of these results suggest that MCPIP1-USP10 may mediate the deubiquitination of NEMO upon genotoxic stress in both a TANK-dependent and TANK-independent manner.
The TANK-MCPIP1-USP10 Complex Inhibits IL-1R/TLRmediated NF-B Activation by Deubiquitinating TRAF6 -Both TANK and MCPIP1 have been shown to suppress NF-B activation by IL-1␤ treatment or TLR activation (28,39). Deubiquitination of TRAF proteins was identified as the underlying mechanism by which TANK and MCPIP1 inhibit NF-B signaling. Interestingly, neither TANK nor MCPIP1 harbor a DUB domain, suggesting that the TANK/MCPIP1-associated DUB is required for the deubiquitination of TRAF proteins. Because we found that TANK-MCPIP1-USP10 formed a complex and promoted TRAF6 deubiquitination in response to genotoxic stress, we wondered whether the TANK-MCPIP1-USP10 complex could play a role in suppressing IL-1R/TLR-mediated NF-B activation. To test this hypothesis, we depleted USP10 with siRNA in HEK293T cells and treated the cells with IL-1␤. We found that knockdown of USP10 significantly enhanced IL-1␤induced NF-B activation (Fig. 7A). Consistent results were observed in cells where USP10 was depleted with another siRNA targeting a different sequence (Fig. 7B). Moreover, ectopic expression of USP10 suppressed IL-1␤-induced NF-B activation and diminished the enhanced NF-B activation by IL-1␤ in USP10-depleted cells. These results were further confirmed by a NF-B reporter assay showing that depletion of USP10 enhanced IL-1␤-induced NF-B transactivity (Fig. 7C). In accordance, TRAF6 ubiquitination was also remarkably increased in cells depleted of USP10, suggesting that USP10 may negatively regulate IL-1␤-induced NF-B activation by deubiquitinating TRAF6 (Fig. 7D).
Consistent with previous reports (28,29), TRAF6 ubiquitination was increased substantially in TANK KO HEK293T cells treated with IL-1␤, whereas overexpression of TANK suppressed TRAF6 ubiquitination (Fig. 7E). Moreover, deletion of TANK significantly enhanced IL-1␤-induced NF-B-dependent transcription (Fig. 7F). In parallel, USP10 was indispensable for the TANK-promoted deubiquitination of TRAF6 because knockdown of USP10 abrogated TRAF6 deubiquitination by TANK overexpression (Fig. 7G). Because we found that TANK was an essential partner for USP10 to deubiquitinate TRAF6, we examined whether ectopic USP10 could inhibit IL-1␤-induced TRAF6 ubiquitination in the absence of TANK. As shown in Fig. 7H, overexpressing USP10 was able to inhibit TRAF6 ubiquitination in wild-type cells but not in TANK KO cells (compare lanes 3 and 7). In parallel, we found that LPSinduced TRAF6 ubiquitination was resistant to TANK overexpression in MCPIP1-deficient MEFs (Fig. 7I). Together, our data indicate that USP10/MCPIP1/TANK likely also form a deubiquitinating complex in response to IL-1R/TLR activation, which, in turn, suppresses NF-B activation by diminishing TRAF6 ubiquitination.
TANK Overexpression Decreases Cell Survival upon DNA Damage-We found that increased expression of the anti-apoptosis genes BIRC2, BIRC3, and BCL2L1 by VP16 treatment was further enhanced in TANK KO cells (Fig. 8A). Moreover, TANK overexpression abrogated the induction of cIAP1, cIAP2, and BCL-XL in both wild-type and TANK KO cells. In parallel, TANK deletion also further increased the VP16-induced transcription of the cytokine genes TNF, IL6, and CXCL8, which was suppressed by reconstitution of TANK (Fig.  8B). Our previous studies have demonstrated that the transcription of these genes in response to genotoxic stress is regulated by NF-B, which may contribute to reduced apoptosis and prolonged cell survival upon DNA damage (5,16,18). Accordingly, we found that genotoxic drug-induced apoptosis was attenuated in TANK KO cells compared with wild-type cells (Fig. 8C, compare lanes 3 and 7). In contrast, overexpressing TANK increased cell apoptosis in both HEK293T and MDA-MB-231 cells exposed to genotoxic drugs (Fig. 8, C and   FIGURE 5. The TRAF-binding domain of TANK is required for USP10 to deubiquitinate TRAF6. A, recombinant full-length GST-TANK or varying mutants harboring different binding domains were incubated with lysates from HEK293 cells expressing FLAG-TRAF6. TANK-associated TRAF6 was determined by IP with glutathione beads, followed by immunoblotting with anti-FLAG antibody. A diagram of the TRAF6 function domain is shown at the bottom. RING, really interesting new gene domain; CC, coiled coil domain; FL, full length. B, HEK293T cells were transfected with wild-type HA-TANK or a TANK mutant with TRAF-binding domain depletion (⌬170 -191). Whole cell lysates were subjected to IP with anti-HA and followed by immunoblotting as indicated. C, Similar analyses as in B were carried out in HEK293T cells transfected with HA-TRAF6 wild-type or a TRAF-C depletion mutant. D, wild-type and TANK KO HEK293T cells were treated with VP16 (10 M, 2 h). Whole cell lysates were immunoprecipitated with antibody against TRAF6. The precipitates were immunoblotted as indicated. E, wild-type and MCPIP1 Ϫ/Ϫ MEFs were treated with doxorubicin (DOX, 2 g/ml, 2 h), and whole cell lysates were immunoprecipitated with anti-TANK antibody, followed by immunoblotting with the indicated antibodies. F, TANK KO cells were transfected with a vector control, TANK WT, or TANK (31-425) mutant. Cells were treated with VP16 (10 M, 2 h), and TRAF6 ubiquitination was analyzed by TRAF6 IP, followed by immunoblotting with antiubiquitin (Ub). Input whole cell lysates were analyzed by immunoblotting using the indicated antibodies. G and H, HEK293T cells were transfected with HA-TANK or TRAF-binding domain-depleted TANK mutants as shown. Cells were then treated with VP16 (10 M, 2 h) or CPT (10 M, 2 h). Whole cell lysates were subjected to analyses of TRAF6 ubiquitination (G) or analyzed by EMSA (NF-B and Oct1) and immunoblotting (H). Normalized NF-B activation (NF-B/Oct1) is shown as -fold induction. D). In agreement, TANK KO cells were more resistant to VP16 treatment compared with wild-type cells, whereas TANK overexpression significantly reduced cell survival upon genotoxic treatment (Fig. 8E). All of these data further support the hypothesis that TANK may promote genotoxic stress-induced cell apoptosis through inhibiting NF-B activation.

Discussion
Studies in the last decade have provided significant insights into how genotoxic stress induces NF-B activation through an atypical nuclear-to-cytoplasmic signaling cascade that may play critical roles in mediating acquired therapeutic resistance to chemotherapy and radiotherapy in cancer cells (2,8). In this genotoxic NF-B signaling pathway, a variety of protein posttranslational modifications, such as SUMOylation, poly(ADPribosyl)ation, phosphorylation, and ubiquitination, were found to be involved in mediating signal transduction. Multiple forms of ubiquitination, including monoubiquitination and polyubiquitination with Lys-63, Met-1, and Lys-48 linkages, have been shown to play distinct roles in transducing the nuclear DNA damage signal to the cytoplasmic IKK complex, leading to NF-B activation (2,40). Because these posttranslational modifications are reversible, they may serve as signaling switches to regulate the extent and duration of genotoxic NF-B activation. Indeed, a recent study showed that a SUMO protease, SENP2, was induced by NF-B activation in response to DNA damage, which, in turn, inhibited NEMO SUMOylation, leading to decreased genotoxic NF-B activation (20). We found that MCPIP1 transcription was enhanced by NF-B upon genotoxic stress, which served as a scaffold protein to promote USP10-dependent deubiquitination of NEMO, thereby attenuating NF-B activation (21). In this study, we further showed that TANK was a member of the MCPIP1-USP10-associated complex induced by genotoxic stress. NF-B-mediated up-regulation of MCPIP1 enhanced complex formation among MCPIP1, USP10, and TANK, which played an important role as a negative feedback response to attenuate NF-B activation by diminishing polyubiquitination of both NEMO and TRAF6 (Fig. 8F).   (41,42). We found that MCPIP1-associated USP10 was able to remove the linear polyubiquitin chain from NEMO in cells exposed to genotoxic stimuli, likely by cleaving isopeptide bond between NEMO and the first ubiquitin and having the Met-1-linked linear chain removed en bloc. In contrast, USP10 was able to disassemble the Lys-63-linked chain by cleaving the individual ubiquitin moiety (21). Another well characterized DUB, CYLD, was also found to cleave both the Lys-63 chain and linear chain in vitro (43) and form a complex with OTULIN to suppress linear ubiquitination in cells (44). We found that overexpression of CYLD or OTULIN diminished NEMO linear ubiquiti-  . TANK overexpression promotes cell apoptosis upon DNA damage. A and B, wild type and TANK KO HEK293T cells were transfected with HA-TANK or a control vector. After 48 h, cells were treated with VP16 (10 M, 4 h), and expression of the antiapoptosis genes cIAP1, cIAP2, and BCL-XL (A) as well as the cytokine genes TNF-␣, IL-6, and IL-8 (B) was measured by quantitative RT-PCR. GAPDH expression was used as the internal control. Normalized -fold change data were from three independent experiments and are shown as mean Ϯ S.D. C, wild-type and TANK KO HEK293T cells were transfected with HA-TANK as shown. Cells were treated with VP16 (1 M) for 24 h and whole cell lysates were subjected to immunoblotting using the indicated antibodies. Arrow, endogenous TANK; arrowhead, HA-TANK. D, MDA-MB-231 cells were transfected with HA-TANK or a control vector. Cells were treated with doxorubicin (DOX, 0.5 M) for 24 h and analyzed as in C. E, wild-type and TANK KO HEK293T cells were transfected with HA-TANK or a control for 24 h. Cells were then treated with 1 M VP16 for the indicated times, and the survival fraction of the cells was determined by trypan blue staining. The cell survival fraction data are from three independent experiments and are shown as mean Ϯ S.D. F, a model illustrating the TANK/MCPIP1/USP10-mediated negative feedback response to suppress NF-B activation via deubiquitinating TRAF6 and NEMO. *, p Ͻ 0.05; **, p Ͻ 0.01. nation by genotoxic stress, whereas deletion of TANK had little impact on CYLD/OTULIN-dependent deubiquitination of NEMO (data not shown). These observations suggest that TANK selectively participates in USP10-dependent deubiquitination of NEMO, likely through interaction with MCPIP1.
USP10 has been shown to regulate the cellular response to DNA damage by deubiquitinating and stabilizing p53, which requires its translocation into the nucleus and phosphorylation by ATM (45). Moreover, USP10 could stabilize Beclin1 by inhibiting its ubiquitination, which reciprocally stabilizes USP10 and promotes p53-dependent apoptosis (46). A recent report found that USP10 could interact with and deubiquitinates SIRT6, which enhanced SIRT6 stability and resulted in SIRT6/p53-dependent suppression of c-Myc oncogenic activity (47). All of these studies indicate that USP10 may play a tumor-suppressive role by deubiquitinating and stabilizing critical tumor suppressors. Here we show that USP10 may also suppress tumor progression by inhibiting NF-B signaling. In cancer cells exposed to genotoxic treatment, USP10 may diminish NF-B activation by removing polyubiquitin chains attached on NEMO and TRAF6, which disrupts the signaling scaffold required for effective IKK activation, thereby sensitizing cancer cells to cytotoxic treatment.
Our data indicate that TANK-associated MCPIP1-USP10 may also suppress NF-B activation in the inflammatory response upon IL1R/TLR activation. It has been found that LPS-induced ubiquitination of TRAF2, TRAF3, and TRAF6 was increased substantially in MCPIP1 Ϫ/Ϫ macrophages (39). Similarly, TRAF6 ubiquitination was enhanced significantly in Tank Ϫ/Ϫ B cells and macrophages upon TLR stimulation (28). Consistently, escalated NF-B activation and hyperinflammation were observed in both Zc3h12a Ϫ/Ϫ and Tank Ϫ/Ϫ mice. Although both TANK and MCPIP1 were implicated in promoting the deubiquitination of TRAF proteins, the underlying mechanism was elusive because neither of the proteins harbor a well defined DUB domain. Our study demonstrated that USP10 forms a complex with TANK and MCPIP1 and is essential for TANK-dependent deubiquitination of TRAF6 in response to IL-1␤ treatment and genotoxic stress, suggesting that USP10 may be the missing link responsible for the DUB activity of the TANK-MCPIP1-associated complex.
In summary, the results presented in this study imply that TANK negatively regulates genotoxic NF-B activation by diminishing the ubiquitination of TRAF6 and NEMO, which relies on the formation of a deubiquitinating complex comprised of TANK, MCPIP1, and USP10. TANK plays an essential role as a scaffold protein, bridging the interaction between TRAF6 and MCPIP1-USP10, which facilitates USP10-dependent deubiquitination. Our data further suggest that this TANK-MCPIP1-USP10 complex may also be responsible for inhibiting the ubiquitination of TRAF family proteins in response to IL-1R/TLR activation, which may prevent hyperinflammation and other pathological processes because of uncontrolled activation of NF-B signaling. Therefore, modulating the TANK-mediated inhibition of NF-B activation may serve as a valuable therapeutic approach to mitigate cancer therapeutic resistance and prevent autoimmune diseases.