Rip2 Participates in Bcl10 Signaling and T-cell Receptor-mediated NF- (cid:1) B Activation*

Engagement of the T-cell receptor (TCR) initiates a signaling cascade that ultimately results in activation of the transcription factor NF- (cid:1) B, which regulates many T-cell functions including proliferation, differentiation and cytokine production. Herein we demonstrate that Rip2, a caspase recruitment domain (CARD)-containing serine/threonine kinase, plays an important role in this cascade and is required for optimal TCR signaling and NF- (cid:1) B activation. Following TCR engagement, Rip2 associated with Bcl10, a CARD-containing signaling com-ponent of the TCR-induced NF- (cid:1) B pathway, and induced its phosphorylation. Rip2-deficient mice were defective in TCR-induced NF- (cid:1) B activation, interleukin-2 production, and proliferation in vitro and exhibited defective T-cell-dependent responses in vivo . The defect in Rip2 (cid:2) / (cid:2) T-cells correlated with a lack of TCR-induced Bcl10phosphorylation.Furthermore,deficiencyinBcl10-dependentNF- (cid:1) B activation could be rescued in Rip2 (cid:2) / (cid:2) embryonic fibroblasts by exogenous wild-type Rip2 but not a kinase-dead mutant. Together these data define an important role for Rip2 in TCR-induced NF- (cid:1) B activation and T-cell function and highlight the significance of of in of

Many diverse stimuli activate NF-B by inducing the phosphorylation and destruction of inhibitory molecules known as the IBs that retain NF-B in the cytoplasm (1). The IB kinase (IKK) 1 complex, composed of two kinase subunits, IKK␣ and IKK␤, and a non-catalytic subunit, NEMO/IKK␥, is responsible for the phosphorylation of the IBs. The association of Bcl10 and CARMA1 (CARD11), two caspase-recruitment domain (CARD)-containing proteins, has been shown to be essential to the transduction of the signal from the T-cell receptor (TCR) to the IKK complex (2). Mice deficient for either Bcl10 or CARMA1 display profound defects in T-cell proliferation and cytokine production due to a lack of NF-B activation (3)(4)(5)(6)(7); however, the mechanism by which the CARMA1/Bcl10 complex activates IKK remains unclear.
In vitro experiments have indicated that Bcl10 undergoes phosphorylation when over expressed with its viral homologue, E10 or CARMA1 (8 -10). In these studies, Bcl10 phosphorylation correlated with its ability to activate NF-B, suggesting that this modification was required for NF-B activation. Indeed, the COOH-terminal domain of Bcl10 is rich in serine and threonine residues and has been proposed as the site of CARMA1-mediated phosphorylation (10). Since CARMA1 itself is not a kinase, the kinase responsible for Bcl10 phosphorylation has remained an open question.
Rip2 is a serine/threonine kinase that contains a CARD domain at its carboxyl terminus and has been shown to induce NF-B activation in over expression systems (11)(12)(13). Rip2 has also been shown to associate in vitro with members of the TRAF family, such as TRAF6, that plays an essential role in the innate immune response downstream of Toll-like receptors (TLRs) (14,15). In addition, Rip2 has been implicated in regulating both the innate and adaptive immune responses (16,17). Mice deficient in Rip2 mounted only an attenuated immune response against Toll-like receptor agonists such as lipopolysaccaride (LPS) (16,17). Interestingly, CD4 ϩ T-cells from Rip2-deficient mice were unable to proliferate efficiently in response to antigen-induced T-cell activation, but no mechanism was provided for this striking observation (16,17). We sought to define the role of Rip2 in antigen-induced NF-B activation and T-cell proliferation.

EXPERIMENTAL PROCEDURES
Generation of Rip2 Ϫ/Ϫ Mice-A targeting vector that removed exon I of Rip2 was electroporated into ES cells. Homologous recombinants were used to generate chimeric founder mice by microinjection into C57BL/6J blastocysts. Germ line transmission was confirmed by Southern blot analysis of genomic tail DNA. Two independent ES clone lines resulted in mice with identical phenotypes. All mice used in experiments were backcrossed onto C57BL/6 five to seven generations and were confirmed to be Ͼ95% C57BL/6 by PCR analysis of genomic tail DNA.
Proliferation Assays-Splenic B and T-cells and CD4 ϩ T were purified by negative selection using magnetic beads (Miltenyi Biotech) to Ͼ95% purity. Purified T-cells were activated with plate bound anti-CD3 (0 -10 g/ml) (BD Biosciences) with or without irradiated CD4-depleted * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. APCs or plate-bound anti-CD28 (0 -10 g/ml) (BD Biosciences), phorbol myristate acetate (PMA) (2 ng/ml) plus ionomycin (0.1 g/ml) (Sigma) in the presence or absence of IL-2 (50 ng/ml) (R & D Systems). B-cells were stimulated with anti-IgM (20 g/ml) (Jackson Laboratories), LPS (20 g/ml) (Sigma), or PMA (2 ng/ml) plus ionomycin (0.1 g/ml). Cells were harvested at 24, 48, 72, and 96 h after an 8-h pulse with [ 3 H]thymidine (1 Ci/well), and incorporation of [ 3 H]thymidine was measured using a Matrix 96 direct ␤ counter system (Hewlett-Packard). Data represent triplicate samples and are representative of at least three separate experiments.
Neonatal Heart Allograft-Neonatal hearts from BALB/c (H-2 d ) mice were surgically implanted behind the dorsum of the ear pinna of 12week-old male Rip2 Ϫ/Ϫ and wild-type mice (Both H-2 b ). Heart grafts were examined with a stereomicroscope at 10 -20-fold magnification every other day until rejection.

Isolation of Phosphorylated Proteins-Splenic T-cells from Rip2
Ϫ/Ϫ and wild-type mice (4 ϫ 10 7) cells/ml) were stimulated with 10 g/ml plate-bound anti-CD3 for 0, 15, and 30 min. Cells were lysed under denaturing conditions to disrupt protein-protein interactions and diluted to 0.1 mg/ml in phospho-lysis buffer (Qiagen). Phosphorylated proteins were separated using the phospho-protein purification kit (Qiagen) according to the manufacturer's instructions.

RESULTS AND DISCUSSION
Rip2 Associates with Bcl10 and Induces Its Phosphorylation-We investigated whether Rip2 could associate with molecules known to play essential roles in the TCR-induced signaling cascade. Initially, we tested whether Rip2 and Bcl10 could associate by overexpressing tagged versions of both proteins in 293 T-cells. V5-tagged Bcl10 could be co-immunoprecipitated with HA-tagged Rip2 (Fig. 1A). Interestingly, two bands representing Bcl10 were observed. The upper band was determined to be a hyperphosphorylated form of Bcl10, since it could be collapsed to the lower band by phosphatase treatment (Fig. 1A). Hyperphosphorylation of Bcl10 was also apparent by mobility shift in whole cell lysates from 293T-cells cotransfected with Rip2 and Bcl10, compared with the very low levels of phosphorylation seen with Bcl10 alone (Fig. 1B). To establish the domains of Rip2 responsible for hyperphosphorylation of Bcl10, mutants with deletions of either the kinase domain or the CARD domain were used in co-expression studies. Bcl10 hyperphosphorylation required both a functional kinase domain and CARD domain of Rip2 as neither mutant induced phosphorylation of Bcl10 (Fig. 1B). Moreover, phosphorylation of Bcl10 was specific for Rip2, as overexpression of RIP or Rip3 did not induce Bcl10 phosphorylation (Fig. 1C).
To determine whether Rip2 was involved in Bcl10-dependent signaling pathways, we studied the interaction of endogenous proteins in Jurkat cells stimulated with cross-linking antibodies to CD3-TCR. Rip2 and Bcl10 consistently associated in a transient and time-dependent manner after TCR engagement (Fig. 1D, bottom panel). Induction of phosphorylated Zap-70 confirmed TCR activation (Fig. 1E). We next examined the phosphorylation status of Bcl10 using phosphoserine-specific antibodies. Lysates from anti-CD3 treated and untreated Jurkats were immunoprecipitated using antibodies to Bcl10 and Western blots were performed using antibodies for phosphoserine and Bcl10. Serine phosphorylated Bcl10 was detected after 15-min treatment with anti-CD3 (Fig. 1F, top panel) and treatment of immunoprecipitates with phosphatase (PPase) significantly diminished levels of serine-phosphorylated Bcl10. Phosphorylation of endogenous Bcl10 was also apparent after treatment with anti-CD3, as evidenced by a slower migrating band that could be collapsed by treatment with phosphatase (Fig. 1F, bottom panel). Taken together, these results were consistent with Rip2 binding Bcl10 upon TCR engagement and inducing its phosphorylation.
Defective T-cell Proliferation and Function in Rip2 Ϫ/Ϫ Mice-To examine the effects of Rip2 on T cell activation in an in vivo setting, we generated Rip2-deficient mice by homologous recombination. Rip2 Ϫ/Ϫ T-cells were deficient in anti-CD3 FIG. 1. Rip2 associates with Bcl10 and induces Bcl10 phosphorylation. A, 293T-cells were co-transfected with HA-tagged Rip2 and V5-tagged Bcl10. Cell lysates were immunoprecipitated with antibodies to HA, V5, or as a negative control, Myc, and precipitates were immunoblotted for V5. In some cases, immunoprecipitates were treated with alkaline phosphatase (ϩCIP). Western blots for anti-HA and anti-V5 were also performed on lysates prior to immunoprecipitation. B, 293Tcells were transfected with V5-tagged Bcl10 together with wild-type Rip2 or mutants lacking the kinase domain (⌬KD) or the CARD domain (⌬CARD), and lysates were analyzed by Western blot with anti-V5, anti-Rip2, or anti-actin antibodies. C, 293T-cells were transfected with V5-tagged Bcl10 together with Myc-tagged Rip2, RIP, or Rip3, and lysates were analyzed by Western blot with anti-V5, anti-Myc, or antiactin antibodies. D, Jurkat cells were stimulated with anti-CD3 for the indicated times, and lysates were immunoprecipitated with anti-Bcl10 or anti-Rip2. Immunoprecipitates were subjected to Western blot analysis using anti-Bcl10. E, whole cell lysates from Jurkat cells treated as above were immunoblotted using phospho-specific antibodies for Zap-70. F, Jurkat cells treated as above were immunoprecipitated with anti-Bcl10 and immunoblotted using antibodies specific for phosphoserine or Bcl10. In some cases, immunoprecipitates were treated with phosphatase (PPase). These data are representative of at least three separate experiments. (P-Bcl10, phosphorylated Bcl10). WB, Western blots. induced proliferation (Fig. 2, A and B). This defect could not be rescued by co-stimulation with anti-CD28 or activation using PMA in combination with calcium ionophore (ion) (Fig. 2C). The levels of IL-2 produced after treatment with anti-CD3 alone, anti-CD3 with anti-CD28, or with PMA and ionomycin were drastically reduced compared with wild-type T-cells (data not shown). Moreover, addition of exogenous IL-2 was not able to rescue the defect in proliferation in Rip2 Ϫ/Ϫ T-cells after stimulation (Fig. 2D). Consistent with previous reports, B-cell proliferation in response to PMA/ionomycin, IgM, and LPS was comparable between Rip2-deficient and wild-type B-cells (data not shown) (16). Taken together, these results suggested that the defect in proliferation in Rip2 Ϫ/Ϫ mice was confined to T-cells and likely due to impairment upstream of IL-2 gene transcription and NF-B activation.

Rip2 Participates in Bcl10 Signaling
Previous in vivo experiments on Rip2 Ϫ/Ϫ mice tested T-cell responsiveness using models such as Listeria challenge and T-celldependent antibody responses (16,17), which also involve participation of TLRs and other innate signaling cascades through adjuvant and bacterial components. Therefore, to test T-cell responsiveness in vivo, we designed a high bar functional test, the ability to participate in a graft rejection response, which does not require major driver co-signals from pathways of the innate immune system. Hearts from allogeneic neonate BALB/c (H-2 d ) mice were transplanted into the ear pinna of wild-type and Rip2deficient mice (H-2 b ) and allograft survival monitored. While all hearts were rejected by wild-type mice by day 15, over 50% of the neonate hearts were still beating in Rip2-deficient mice and continued to function for an additional 5 days (Fig. 2E). Therefore, Rip2-deficient mice rejected heart allografts much less readily than wild-type mice, consistent with our in vitro data and a defect in normal T-cell activation and function.
Defective NF-B Activation in Rip2-deficient Cells-To determine the molecular basis of the impairment in T-cell receptor signaling in the absence of Rip2, we analyzed pathways activated by TCR engagement in wild-type and Rip2 Ϫ/Ϫ T-cells. T-cells from wild-type and Rip2 Ϫ/Ϫ mice were treated with plate-bound anti-CD3 or TNF␣, and lysates were assessed by Western blot using phospho-specific antibodies to IB␣. IB␣ was rapidly phosphorylated and degraded in wild-type T-cells but not in Rip2 Ϫ/Ϫ cells (Fig. 3A). In contrast, treatment of both wild-type and Rip2 Ϫ/Ϫ T-cells with TNF␣ promoted equivalent phosphorylation and degradation of IB␣ (Fig. 3B). Hence, NF-B signaling downstream of other surface receptors remained intact in Rip2 Ϫ/Ϫ mice.
TCR engagement also elicits activation of the RAS/MAPK (mitogen-activated protein kinase) pathway. Western blotting using phospho-specific anti-ERK1/2 antibodies demonstrated that ERK-1 and ERK-2 were phosphorylated with similar kinetics in wild-type and Rip2 Ϫ/Ϫ T-cells after TCR engagement (Fig. 3C, upper panel). Similarly, activation of the JNK signaling pathway post-TCR engagement was equivalent in both wild-type and Rip2-deficient T-cells (Fig. 3C, lower panel). These results confirmed that the defect was specific for NF-B signaling downstream of the TCR and that parallel pathways activated by TCR engagement remained intact.
To address the role of Rip2 kinase activity in Bcl10-dependent NF-B activation, we transfected MEFs from wild-type and Rip2 Ϫ/Ϫ mice with Bcl10 and a luciferase reporter for NF-B. While Bcl10 could induce NF-B activation in wild-type MEFs, NF-B activation by Bcl10 was significantly decreased in Rip2 Ϫ/Ϫ MEFS (Fig. 3D). Transfection of exogenous wild-type Rip2, but not a kinase-dead mutant, K47A, could rescue Bcl10induced NF-B reporter activity in Rip2 Ϫ/Ϫ MEFs (Fig. 3D). Therefore, the kinase activity of Rip2 is required for optimal Bcl10-induced NF-B activation.
Bcl10 Is Phosphorylated after TCR Engagement in Wild-type but Not Rip2 Ϫ/Ϫ Mice-Taken together, our data suggested

FIG. 3. Defective NF-B activation and Bcl10 phosphorylation in Rip2deficient cells. Purified T-cells from Rip2
Ϫ/Ϫ and wild-type (WT) mice were stimulated with 10 g/ml plate-bound anti-CD3 (A) or 10 ng/ml TNF␣ for 0 -30 min (B). Lysates were subjected to Western blotting using antibodies to IB␣ and phospho-IB␣. C, purified T-cells from Rip2 Ϫ/Ϫ and wild-type mice were stimulated with 10 g/ml CD3 for 0 -30 min, and lysates were subjected Western blotting using antibodies for phospho-ERK1/2, p44, and phosho-JNK. D, Rip2 Ϫ/Ϫ and wild-type MEFs were transfected with an NF-B luciferase reporter and Bcl10 with or without either wildtype Rip2 or a kinase-dead (KD) mutant Rip2. These data are representative of four separate experiments. Purified T-cells from Rip2 Ϫ/Ϫ and wild-type T-cells were stimulated with 10 g/ml CD3 for 0 -30 min. Cell lysates were separated into non-phosphorylated (NPh) fractions and phosphorylated (Ph) fractions using phospho-specific columns. Purified fractions were Western blotted for Bcl10 and phospho-ERK (E) and HSP60 (F).
that Rip2 functions to regulate T-cell activation by phosphorylating Bcl10. Therefore, we wished to establish whether the lack of NF-B activation observed in Rip2 Ϫ/Ϫ T-cells correlated with a lack of Bcl10 phosphorylation after TCR engagement. Wildtype and Rip2-deficient T-cells were treated with ␣-CD3, and cell lysates were fractionated using a phosphoserine/threonine column. Under these lysis conditions, all protein-protein interactions are disrupted, and only phosphorylated proteins bind the column, while unphosphorylated proteins flow through. Western blotting of the phosphorylated protein fractions using antibodies to phospho-ERK and Bcl10 revealed that while phosphorylated ERK1/2 could easily be detected in the purified phosphorylated fractions of both wild-type and Rip2 Ϫ/Ϫ T-cells (Fig. 3E, middle panel), Bcl10 was only present in the purified phosphorylated fractions of ␣-CD3 treated wild-type T-cells (Fig. 3E, top panel). By contrast, similar levels of Bcl10 were detected in the non-phosphorylated fractions from wild-type and knock-out T-cells (Fig. 3E, bottom panel). To confirm that no unphosphorylated proteins contaminated the phosphorylated protein fraction, lysates from both fractions were Western blotted using antibodies for Hsp60. Hsp60 was abundant in the non-phosphorylated fraction but undetectable in the phosphorylated protein fraction (Fig. 3F). These data demonstrate that Bcl10 is phosphorylated in mouse primary T cells after TCR stimulation, and deficiency of Rip2 precludes phosphorylation of Bcl10.
Herein we provide evidence for the importance of Rip2 in TCR-mediated NF-B activation and Bcl10-dependent signaling. Phosphorylation of Bcl10 occurs after TCR engagement, and lack of phosphorylation correlates with a defect in NF-B activation and T-cell proliferation. Earlier studies have shown that Bcl10 is phosphorylated upon over expression of CARMA1; however, the importance of phosphorylation in T-cell signaling was unclear. Our data suggest that phosphorylation of Bcl10 by Rip2 plays a key role in signaling between the TCR and the IKK complex.
Recent reports (4 -7, 19 -21) have demonstrated that CARMA1 is critically involved in TCR-induced NF-B activation. It remains unclear whether Bcl10 phosphorylation is required for its association with CARMA1. The kinetics of the association between Rip2 and Bcl10 and subsequent Bcl10 phosphorylation in Jurkat cells correlates with the kinetics of the published interaction between CARMA1 and Bcl10 in Jurkat T-cells (19). Phosphorylation of Bcl10 may either facilitate its recruitment to lipid rafts or serve to activate other key molecules in downstream signaling events that ultimately activate the IKK complex. For example, MALT1/paracaspase, a death domain-containing caspase-like molecule, has also been shown to associate with Bcl10 and enhance NF-B activation (22,23) and is also required for TCR-induced proliferation, cytokine production, and NF-B activation (24).
Our data are consistent with previous reports that Rip2deficient mice suffer from defects in the adaptive immune response due to lack of antigen-induced T-cell proliferation and NF-B activation (16,17). Similar to published results, we also observed a defect in cytokine production in macrophages stimulated with LPS and other Toll-like receptors, demonstrating an additional defect in innate immunity (data not shown) (16,17). Since Rip2 associates with key signaling molecules in both the adaptive and innate immune responses, such as Bcl10 and TRAF6 respectively, it is reasonable that the absence of this promiscuous kinase would impinge on multiple signaling pathways and result in broad ranging deficits in immune system function.