RBCK1 Negatively Regulates Tumor Necrosis Factor- and Interleukin-1-triggered NF-κB Activation by Targeting TAB2/3 for Degradation*

Inflammation is a homeostatic mechanism that limits the effects of infectious agents. Tumor necrosis factor (TNF) and interleukin (IL)-1 are two cytokines that induce inflammation through activation of the transcription factor NF-κB. Various studies have suggested that two homologous and structurally related adapter proteins TAB2 and TAB3 play redundant roles in TNF- and IL-1-mediated NF-κB activation pathways. Both TAB2 and TAB3 contain CUE, coiled-coil, and nuclear protein localization 4 zinc finger (NZF) domains. The NZF domains of TAB2/3 are critical for TAB2/3 to bind to Lys63-linked polyubiquitin chains of other adaptor proteins, such as receptor-interacting protein and TRAF6, which are two signaling proteins essential for TNF- and IL-1-induced NF-κB activation, respectively. In a search for proteins containing NZF domains conserved with those of TAB2/3, we identified RBCK1, which has been shown to act as an E3 ubiquitin ligase in iron metabolism. Overexpression of RBCK1 negatively regulates TAB2/3-mediated and TNF- and IL-1-induced NF-κB activation, whereas knockdown of RBCK1 by RNA interference potentiates TNF- and IL-1-induced NF-κB activation. RBCK1 physically interacts with TAB2/3 and facilitates degradation of TAB2/3 through a proteasome-dependent process. Taken together, our findings suggest that RBCK1 is involved in negative regulation of inflammatory signaling triggered by TNF and IL-1 through targeting TAB2/3 for degradation.

Proinflammatory cytokines, such as TNF 3 and IL-1, play critical roles in inflammatory processes and are involved in regulation of immune responses. Stimulation of cells with TNF or IL-1␤ initiates a cascade of signaling events leading to activation of transcription factors NF-B and AP1 (activator protein 1) and induction of proinflammatory cytokines (1,2). NF-B regulates expression of a large number of genes involved in immune responses, inflammation, cell survival, and cancer (3).
NF-B is sequestered in the cytoplasm and kept inactive in nonstimulated cells through binding to inhibitory IB (inhibitor of B) proteins. Following stimulation with cytokines, infectious agents, or radiation-induced DNA double-stranded breaks, the IB proteins are phosphorylated, ubiquitinated, and ultimately degraded in a proteasome-dependent manner, which frees NF-B from IB proteins to translocate to the nucleus, where it activates transcription of target genes (3).
Considerable progress has been made during the past couple of decades in identifying molecular components involved in TNF-and IL-1-triggered NF-B activation pathways. Binding of TNF to TNFR-I leads to its trimerization and the recruitment of adaptor proteins, including TRADD, TRAF2, and receptorinteracting protein (RIP) (1). RIP is subsequently polyubiquitinated in a Lys 48 -or Lys 63 -linked manner, among which Lys 63linked ubiquitination of RIP is required for the recruitment and activation of TAK1 (TGF-␤-activated kinase 1), a kinase associated with adapter proteins TAB2 (TAK1-binding protein 2) and TAB3 (TAK1-binding protein 3) (4,5). The activated TAK1 subsequently activates downstream kinases IKK␣ and IKK␤, which phosphorylate IB proteins, and leads to NF-B activation (6).
As described above, the adaptor proteins TAB2 and TAB3 are critical intermediates for both TNF-and IL-1-induced activation of NF-B. TAB2 deficiency was embryonic lethal due to liver degeneration and apoptosis. However, TAB2-deficient cells were still capable of activating IKK and JNK in response to TNF, IL-1, and lipopolysaccharide (22), whereas knockdown of both TAB2 and TAB3 inhibits IKK and JNK activation induced by TNF and IL-1, suggesting that TAB2 and TAB3 play redundant roles in TNF-and IL-1-triggered signaling (23). TAB2 and TAB3 are structurally related, and both contain an N-terminal CUE domain and C-terminal coiled-coil (CC) and nuclear protein localization 4 (Npl4) zinc finger (NZF) domains (24). TAB2 and TAB3 physically link TAK1 to RIP upon TNF stimulation and to TRAF6 upon IL-1␤ stimulation mainly through their respective NZF domains (4). Collectively, TNF stimulation leads to a RIP-TAB2/3-TAK1 complex, whereas IL-1 stimulation leads to the formation of a TRAF6-TAB2/3-TAK1 complex, both of which are critical for the signaling pathways leading to NF-B activation.
In this study, we found that the NZF-containing protein RBCK1 (RBCC protein interacting with protein kinase C1) interacts with TAB2 and TAB3 and targets them for degradation in a proteasome-dependent manner. Overexpression of RBCK1 inhibits TAB2/3-mediated and TNF-and IL-1-induced NF-B activation, whereas knockdown of RBCK1 by RNAi potentiates TNF-and IL-1-induced NF-B activation. Our findings suggest that RBCK1 plays an inhibitory role in inflammatory signaling pathways.
Transfection and Reporter Gene Assays-293 cells (ϳ5 ϫ 10 4 ) were seeded on 24-well dishes and transfected the following day by standard calcium phosphate precipitation. Where necessary, empty control plasmid was added to ensure that each transfection receives the same amount of total DNA. To normalize for transfection efficiency, 0.1 g of pRL-TK Renilla luciferase reporter plasmid was added to each transfection. Approximately 20 h after transfection, luciferase assays were performed using a dual specific luciferase assay kit (Promega, Madison, WI). Firefly luciferase activities were normalized based on Renilla luciferase activities. All reporter gene assays were repeated at least three times. Data shown were average values Ϯ S.D. from one representative experiment.
Co-immunoprecipitation and Western Blot Analysis-Transient transfection and co-immunoprecipitation, as well as endogenous coimmunoprecipitation experiments, were performed as described (27).

RBCK1
Interacts with TAB2 and TAB3-Previous studies have indicated that the CC domains of TAB2/3 are required for their ubiquitination and protein-protein interaction (23), whereas the NZF domains are critically involved in binding to polyubiquitin chains of RIP and TRAF6, which is key to TNFand IL-1-induced NF-B activation (4). To further understand the roles of TAB2/3 in TNF-and IL-1-triggered NF-B activation pathways, we sought to identify proteins containing NZF domains similar to those of TAB2/3. Data base searches led to identification of RBCK1, also called HOIL-1 (28), XAP3 (29), or UIP28 (30), which contains an NZF domain at its N terminus ( Fig. 1, A and B). In addition, it contains an intermediate CC domain and a C-terminal RING-IBR region, also known as RBR (31) or TRIAD domain (32), which contains two zinc finger domains and a cystine/histidine-rich motif in between (Fig.  1A). It has been shown that RING-IBR region-containing proteins are involved in E3 ubiquitin ligase activity (31).
Since RBCK1 contains an NZF domain that is related to the NZF domains of TAB2 and TAB3, we examined whether it associates with TAB2 and TAB3 through homophilic interactions. We transfected 293 cells with expression plasmids for HA-tagged TAB2 or TAB3 and FLAG-tagged RBCK1. Coimmunoprecipitation experiments indicated that RBCK1 interacted with both TAB2 and TAB3 (Fig. 1C).
To determine whether RBCK1 is associated with TAB2 and TAB3 under physiological conditions, we raised mouse polyclonal antiserum against recombinant human full-length RBCK1 and TAB2 (Fig. 1D). Anti-TAB2 also recognized TAB3 because of their structural similarity (data not shown). Endogenous coimmunoprecipitation experiments indicated that RBCK1 was associated with TAB2/3 in untransfected 293 cells (Fig. 1E).
To determine whether other proteins associated with TAB2 and TAB3 interact with RBCK1, we performed transient transfection and coimmunoprecipitation assays. As shown in Fig. 1F, RBCK1 interacted with TAK1, TRAF6, and RIP but not TRAF2 and TAB1 in these experiments.
Domain Mapping of the Interactions between RBCK1 and TAB2/3-It has been reported that the interaction between TAB2/3 and TAK1 is dependent on their respective CC domains (19,23). To determine which domain of RBCK1 facilitates its interaction with TAB2/3, we made a series of RBCK1 deletion mutants ( Fig. 2A), including aa 1-220 (carrying NZF), aa 221-500 (carrying CC and RING-IBR), aa 1-270 (carrying NZF and CC), and aa 271-500 (carrying RING-IBR). Coimmunoprecipitation assays indicated that RBCK1 (aa 1-270) interacted with TAB2 as strongly as its full-length did, and RBCK1 (aa 1-220) and RBCK1 (aa 221-500) mutants interacted with TAB2 to a lesser extent compared with the full-length RBCK1 (Fig. 2B). In the same experiment, RBCK1 (aa 271-500) did not interact with TAB2 (Fig. 2B). These observations suggest that the N terminus of RBCK1, which contains the NZF and CC domains, is essential for its interaction with TAB2. Similar results were obtained with TAB3 (data not shown).
Similarly, we made a series of TAB2 deletion mutants and determined which domains of TAB2 are required for its interaction with RBCK1 ( Fig. 2A). In transient transfection and coimmunoprecipitation experiments, full-length TAB2, TAB2 (aa 55-693), and TAB2 (aa 361-693) interacted with RBCK1 efficiently (Fig. 2C). In the same experiments, TAB2 (aa 1-665) interacted with RBCK1 weakly, whereas TAB2 (aa 1-360) did not interact with RBCK1 (Fig. 2C). These data suggest that TAB2 interacts with RBCK1 through its C-terminal CC-and NZF-containing fragment. Similar results were obtained with TAB3 and its deletion mutants (data not shown). Taken together, our results suggest that the N terminus of RBCK1 and the C terminus of TAB2/3 are required for their interactions.
Knockdown of RBCK1 Potentiates TNF-and IL-1-induced NF-B Activation-Since RBCK1 inhibits TNF-and IL-1-induced NF-B activation in overexpression experiments, we examined whether RBCK1 regulates TNF-and IL-1-induced FIGURE 1. RBCK1 is associated with TAB2, TAB3, TAK1, TRAF6, and RIP. A, a schematic presentation of human TAB2, TAB3, and RBCK1. C/H, cysteine/histidine-rich. B, an alignment of the amino acid sequences of the NZF domains from human RBCK1, TAB2, and TAB3. C, RBCK1 interacts with TAK1, TAB2, and TAB3. 293 cells (ϳ1 ϫ 10 6 ) were transfected with the indicated expression plasmids (5 g each). Cell lysates were immunoprecipitated (IP) with anti-FLAG (␣F) or mouse IgG antibody (Ab). The immunoprecipitates were analyzed by Western blot (WB) with anti-HA (top). Whole-cell lysates were analyzed by Western blots with anti-HA to determine the expression of TAK1, TAB2, and TAB3 (middle) and with anti-FLAG to determine the expression of RBCK1 (bottom). D, specificity of the antibodies against RBCK1, TAB2, and TAK1. 293 cells transfected with a control plasmid and an expression plasmid for FLAG-RBCK1, FLAG-TAB1, or HA-TAK1 were lysed. The lysates were analyzed by Western blots with mouse polyclonal anti-RBCK1, mouse polyclonal anti-TAB2, or rabbit polyclonal anti-TAK1 antibody. E, RBCK1 interacts with TAB2/3 and TAK1 under physiological conditions. 293 cell lysates were immunoprecipitated with mouse polyclonal anti-RBCK1 or mouse IgG. The immunoprecipitates were analyzed by Western blots with anti-TAB2 (top) or anti-TAK1 (bottom) antibody. F, RBCK1 interacts with TRAF6 and RIP but not with TRAF2 or TAB1. The experiments were performed as in C.

NF-B activation under physiological condition.
To test this, we made five RNAi plasmids targeting different sites of human RBCK1 mRNA. Transient transfection and Western blot analysis indicated that one of these RNAi plasmids (RBCK1-RNAi-4) could markedly inhibit the expression of transfected and endogenous RBCK1 in 293 cells, whereas the other RNAi plasmids had a little or no effect on RBCK1 expression (Fig. 4, A  and B). In reporter gene assays, knockdown of RBCK1 by RBCK1-RNAi-4 plasmid activated NF-B activation and potentiated TNF-and IL-1-induced NF-B activation (Fig. 4, C  and D). In similar experiments, knockdown of RBCK1 neither activated IRF-1 nor potentiated IFN-␥-induced IRF-1 activation (Fig. 4E). Collectively, these data suggest that RBCK1 is a physiological inhibitor of TNF-and IL-1-induced NF-B activation pathways.

RBCK1 Reduces TAB2/3 Protein Level in a Proteasome-dependent
Manner-Since RBCK1 interacts with TAB2/3 and inhibits TNFand IL-1-induced NF-B activation, we further investigated the molecular mechanisms responsible for this observation. Previously, it has been shown that RBCK1 is an E3 ubiquitin ligase and involved in proteasomedependent protein degradation in iron metabolism (28). Therefore, we determined whether RBCK1 could cause ubiquitination and degradation of TAB2 and TAB3. We transfected 293 cells with expression plasmids for TAB2 and TAK1 together with RBCK1 or its deletion mutants. Western blot analysis indicated that overexpression of RBCK1 caused down-regulation of TAB2 but not TAK1. In addition, RBCK1 (aa 221-500) and RBCK1 (aa 271-500) caused down-regulation of TAB2 but not TAK1, whereas RBCK1 (aa 1-220) and RBCK1 (aa 1-270) had no significant effect on TAB2 protein level (Fig. 5A). Similar results were obtained for TAB3 (data not shown). These results suggest that RBCK1 inhibits TAB2/3mediated NF-B activation through degradation of TAB2/3.
To determine whether the reduction of TAB2/3 protein levels by RBCK1 is dependent on ubiquitination, we transfected 293 cells with expression plasmids for HA-TAB2 and FLAG-ubiquitin, in combination with empty control plasmid or full-length RBCK1. Coimmunoprecipitation and Western blot analysis indicated that RBCK1 significantly increased ubiquitination of TAB2 (Fig. 5B). Furthermore, the addition of MG132, a specific proteasome inhibitor, rescued TAB2 from RBCK1-mediated degradation (Fig. 5C). Taken together, these observations suggest that RBCK1 negatively regulates TAB2/3-mediated NF-B activation by causing ubiquitination and proteasome-dependent degradation of TAB2/3.

DISCUSSION
Inflammation is a homeostatic mechanism that limits the effects of infectious agents. TNF and IL-1 are two proinflammatory cytokines, which induce inflammation through activation of the transcription factor NF-B. In this study, we identified RBCK1 as a negative regulator of TNF-and IL-1-induced NF-B activation. Identification of domains that mediate the interaction between RBCK1 and TAB2/3. A, a schematic presentation of RBCK1 and TAB2 deletion mutants. B, the CC and NZF domains of RBCK1 are required for its interaction with TAB2. 293 cells (ϳ1 ϫ 10 6 ) were transfected with expression plasmids for HA-TAB2 and FLAG-RBCK1 or its mutants (5 g each). Cell lysates were immunoprecipitated (IP) with anti-HA or mouse IgG. The immunoprecipitates were analyzed by Western blots (WB) with anti-FLAG antibody (top). Whole-cell lysates were analyzed by Western blots with anti-HA (middle) or anti-FLAG (bottom) to determine the expression levels of transfected plasmids. C, the CC and NZF domains of TAB2 are required for its interaction with RBCK1. 293 cells (ϳ1 ϫ 10 6 ) were transfected with the indicated expression plasmids (5 g each). Immunoprecipitation and Western blot analysis were performed with the indicated antibodies.
Various studies have indicated that two homologous and structurally related adapter proteins TAB2 and TAB3 play redundant roles in TNF-and IL-1-mediated NF-B activation pathways. TAB2 and TAB3 are associated with TAK1, and these protein complexes are recruited to RIP upon TNF stimulation and to TRAF6 upon IL-1 stimulation (4,19,23). Both TAB2 and TAB3 contain CUE, CC, and NZF domains. Among them, NZF domain is highly conserved, and the intact NZF domain is critical for TAB2/3 to bind to Lys 63 -linked polyubiquitin chains of other adaptor proteins, such as RIP and TRAF6, which are two signaling proteins essential for TNF-and IL-1induced IKK activation, respectively (4). In a search for proteins containing NZF domains conserved with those of TAB2/3, we identified RBCK1, which has been shown to act as an E3 ubiquitin ligase in iron metabolism (28). Besides an NZF domain, RBCK1 contains a CC motif and a RING-IBR region, indicating that RBCK1 might be multifunctional through association with other proteins.
Our coimmunoprecipitation results suggest that RBCK1 interacts with TAB2 and TAB3 both in the mammalian overexpression system and in untransfected cells (Fig. 1, C-E). In addition, RBCK1 interacts with TAK1, TRAF6, and RIP, signaling components that are associated with TAB2/3 (Fig. 1, C-F). These findings suggest that RBCK1 is physically associated with TAB2/3-containing protein complexes.
In reporter gene assays, overexpression of RBCK1 inhibits TAB2/3-mediated and TNF-and IL-1-induced NF-B activation (Fig. 3, A and B), whereas knockdown of RBCK1 by RNAi potentiates TNF-and IL-1-induced NF-B activation (Fig. 4D). These results suggest that RBCK1 is a physiological suppressor of TNF-and IL-1-induced NF-B activation pathways.
Previously, it has been demonstrated that the RING-IBR domain exerts E3 ubiquitin ligase activity and is involved in proteasome-dependent protein degradation (31). Our results indicate that the RING-IBR domain is essential for RBCK1 to confer its ubiquitination and proteasome-dependent degradation of TAB2/3 (Fig. 5). Consistently, the RBCK1 mutants that contain the CC and RING-IBR domain also inhibit TAB2/3mediated NF-B activation. Furthermore, it has been reported that TAB2/3 are modified by phosphorylation and ubiquitination upon stimulation, which seems to be required for their adaptor functions (23). Taken together, these observations support the following model of the role of RBCK1 in TNF-and IL-1-induced signaling. In this model, RBCK1, which is an E3 ubiquitin ligase, is associated with TAB2/3 through their respective CC and NZF domains. This association causes TAB2/3 ubiquitination and degradation and therefore negatively regulates TNF-and IL-1-induced NF-B activation. Fur-  Reporter assays were performed at 48 h after transfection. D, knockdown of RBCK1 potentiates TNF-and IL-1-induced NF-B activation. Transfection and reporter assays were performed similarly as in C, except that transfected cells were treated with TNF (10 ng/ml), IL-1␤ (10 ng/ml), or left untreated for 8 h before reporter assay. E, knockdown of RBCK1 has no significant effect on IFN-␥-induced IRF-1 activation. Transfection and reporter assays were carried out as in D, except that IRF-1 luciferase plasmid was used as reporter plasmid, and cells were treated with IFN-␥ (100 ng/ml). GAPDH, glyceraldehyde-3phosphate dehydrogenase.
ther research could be done to find out the exact mechanism by which RBCK1 exerts the degradation of TAB2/3 in quiescent cells or stimulated cells, if the efficient experimental system is available.
Domain mapping experiments suggest that the N-terminal CC-and NZF-containing fragment of RBCK1 (aa 1-270) and the C-terminal CC-and NZF-containing fragment of TAB2 (aa 361-693) are responsible for their interactions (Fig. 2), whereas the C-terminal RING-IBR domain of RBCK1 (aa 271-500) is responsible for catalyzing the ubiquitination of TAB2 (Fig. 5). In our experiments, RBCK1 (aa 221-500) interacts with TAB2/3 weakly but can still efficiently inhibit TAB2/3-mediated NF-B activation. The simplest explanation for this observation is that the concentration of overexpressed RBCK1 (aa 221-500) is high, and a tight interaction between RBCK1 (aa 221-500) and TAB2/3 is not required for RBCK1 (aa 221-500)mediated ubiquitination of TAB2/3. This is further supported by the observation that RBCK1 (aa 271-500) could also inhibit TAB2/3-mediated NF-B activation, although to a lesser degree compared with RBCK1 (aa 221-500) (Fig. 3). In reporter assays, RBCK1 (aa 1-270) also inhibited TAB2/3-mediated NF-B activation (Fig. 3). Although we do not know the exact mechanism behind this observation, one possible explanation is that RBCK1 (aa 1-270) blocks recruitment of other signaling components in the NF-B activation pathways to TAB2/3. Host inflammatory responses must be strictly controlled by a variety of negative regulators, because hyperactivation of inflammatory pathways is usually harmful. So far, a number of suppressors of inflammatory signaling have been identified. For example, IRAK-M, a homolog of IRAK1 that is induced by lipopolysaccharide stimulation after 24 h, interacts with MyD88 to negatively regulate its signaling to TRAF6 (36). A20 inhibits TNF-, IL-1-, TLR-, and RIG-Iinduced NF-B activation pathways by its ubiquitin-editing function (5,37). ␤-Arrestin inhibits NF-B activation at both upstream and transcriptional levels (37,38). The inhibitory factors alleviate the inflammatory responses, thereby preventing the host from mounting extraordinary inflammatory responses. The identification of RBCK1 as a novel inhibitor of TNF-and IL-1-induced NF-B activation pathways provides an additional target and approach for controlling inflammation.