Tumor Necrosis Factor Receptor–associated Factor 6 (TRAF6) Stimulates Extracellular Signal–regulated Kinase (ERK) Activity in CD40 Signaling Along a Ras-independent Pathway

Human embryonic kidney cell line 293 was maintained in DME supplemented with 10% FCS, 200 mM l-glutamine, and penicillin/streptomycin.

TRAF2, TRAF3, and human CD40 cDNA were obtained by the reverse transcription PCR (RT-PCR) by use of messenger RNA purified from B cell lymphomas, WEHI231 and Raji, with primers flanking the entire coding region, and then cloned into pMIKHygB, a gift from Dr. K. Maruyama (Tokyo Medical and Dental University, Tokyo, Japan). Deletion mutants of TRAF2 that lack a RING finger motif (amino acids [aa] 87–501) and the TRAF-C domain (aa 352–501) were constructed by PCR. Isolation of TRAF5 and TRAF6 cDNAs and construction of the expression vectors coding full-length TRAF5, full-length TRAF6, and the TRAF-C of TRAF6 were reported previously (9, 30). A deletion mutant of human CD40, designated CD40Δ246, that removes 32 aa at positions 226–257 from the cytoplasmic tail was previously described (9). Dominant-negative Raf-1 (aa 1–258) was constructed by reverse transcription PCR, according to Schaap et al. (37). Dominant-negative N17Ras (38) and glutathione-S-transferase (GST) fusion ERK2 (pGSTM1) were provided by Dr. G.M. Cooper (Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA) and Dr. K. Takishima (National Defense Medical College, Saitama, Japan), respectively. Construction of a dominant-negative mutant of MAPK/ERK-activating kinase (MEK)1, designated MEK1^DN, and a constitutively activated mutant of MEK1, designated MEK1^EE, was previously described (39). All cDNA fragments were subcloned into pMIKHygB.

ERK-dependent activator and reporter constructs were prepared according to Seth et al. (40, 41). In brief, the fusion of the GAL4 DNA-binding domain and the nuclear localization signal (aa 1–147) and NH₂-terminal transcriptional activation region of c-Myc (aa 2–103) were constructed in vector pGBT9 (Clontech, Palo Alto, CA), and its HindIII-SalI fragment was digested and subcloned into pMIKHygB, designated pGAL4MycN. A firefly luciferase reporter containing two copies of the GAL4 binding site adjacent to a thymidine kinase promoter was constructed into pGL2 (Promega, Madison, WI), designated pGAL4tkLuc. A previous study by Seth et al. (41) showed that, in cotransfected cells with the GAL4–c-Myc fusion construct and luciferase reporter, the phosphorylation state of the NH₂-terminal c-Myc transactivation domain was increased in an ERK-dependent manner, resulting in an increase in the activity of the luciferase reporter gene (41). Therefore, this ERK-dependent activator and reporter system allows one to monitor ERK activity by measuring the luciferase reporter activity. NFκB-dependent luciferase constructs containing three copies of an NFκB binding site from immunogloblin κ light chain enhancer and the thymidine kinase promoter was constructed into pGL2, designed pkBtkLuc. All cDNAs and constructs were confirmed by DNA sequencing.

An in vitro kinase assay for endogenous ERK activity was performed as previously described (17). For estimating exogenous ERK2 kinase activity, 293 cells were cotransfected with 1 μg of GST-ERK2 and various amounts of plasmid constructs by the calcium phosphate coprecipitation method. The total amount of DNA was adjusted with empty vector at 10 μg. When 293 cells were cotransfected with 1 μg of CD40 cDNA, the cells were stimulated with anti-CD40 mAb (2 μg/ml) 24 h after transfection. In some experiments, cells were analyzed for the expression of CD40 on the surface by FACScan^® (Becton Dickinson, Mountain View, CA) with FITC-conjugated anti–human CD40 mAb (5C3; PharMingen, San Diego, CA). The cell lysates were incubated with glutathione-coupled Sepharose 4B (Pharmacia, Uppsala, Sweden) at 4°C for 1 h. The Sepharose beads were pelleted down by centrifugation and washed twice with lysis buffer and with Tris-buffered saline (10 mM Tris-HCl, pH 7.2, 100 mM NaCl) supplemented with 5 mM benzamidine and 1 mM Na₃VO₄. The GST-ERK2 immobilized on Sepharose was directly subjected to a kinase reaction as previously described (17).

Cells were cotransfected with either full-length CD40 or CD40Δ246 and ERK-dependent activator and reporter plasmids consisted of 500 ng of pGAL4tkLuc and 100 ng of pGAL4MycN, or NFκB-dependent plasmid of 20 ng of pkBtkLuc, 1 μg of β-actin β-galactosidase (β-gal), and various amounts of expression plasmids by the calcium phosphate coprecipitation method. Total amounts of DNA were adjusted with empty vector at 10 μg. The cells were stimulated with anti-CD40 mAb 24 h after transfection. The cells were lysed in PicaGene^TM Cell Culture Lysis Reagent LCβ (Toyo Ink, Tokyo, Japan) 24 h after incubation, and luciferase activity was measured by use of a luciferase assay kit, PicaGene^TM (Toyo Ink), followed by normalization of transfection efficiency by β-gal activity.

Overexpression of TRAF2, TRAF5, and TRAF6 results in its aggregation and causes activation of the NFκB signaling pathway, which could be similar to that induced by ligand-triggered receptor aggregation (6, 8, 9, 30, 33). Therefore, to investigate which TRAF proteins might participate in ERK activation, we cotransfected human embryonic kidney 293 cells with various TRAF expression vectors and an GST-ERK2 fusion construct. After transfection, the GST–ERK2 fusion protein was precipitated by use of glutatione Sepharose 4B from the cell lysate, and the ERK activity was analyzed by an in vitro kinase assay, with myelin basic protein (MBP) used as an exogenous substrate. As shown in Fig. 1 A, transient expression of full-length TRAF6, but not TRAF2, TRAF3, or TRAF5, stimulated exogenous ERK activity approximately fivefold above the vector control, suggesting that transient expression of TRAF6 stimulates ERK activity in the 293 cell line.

ERK is known to phosphorylate target protein within the cytosol (42), whereas, under certain conditions, ERK translocates to the nucleus to phosphorylate the nuclear substrate (41, 43–45). To investigate the effect of various TRAF proteins on ERK activity in the nucleus, we used an ERK- dependent activator and reporter system, established by Davis and coworkers (40, 41, and see Materials and Methods). Human 293 cells were transiently transfected with various TRAF expression vectors, together with an ERK-dependent activator, a fusion construct of the GAL4 DNA-binding domain and the nuclear localization signal/NH₂-terminal transactivation domain of c-Myc, and a luciferase reporter plasmid containing GAL4 binding sites adjacent to a thymidine kinase promoter, as an ERK-dependent reporter (40, 41). In addition, to confirm the activity of TRAF proteins, we cotransfected 293 cells with various TRAF expression vectors and the NFκB-dependent luciferase reporter.

As shown in Fig. 1,B a, transient expression of full-length TRAF6 triggered activation of the cotransfected ERK-dependent reporter in a dose-dependent manner to a maximum of approximately fourfold above the vector control, whereas a deletion of the RING finger and zinc fingers of TRAF6 (TRAF-C) abolished the activity of ERK activation. Consistent with previous reports (9, 33), transient expression of full-length TRAF6, but not TRAF-C of TRAF6, caused NFκB activation in the 293 cell line (Fig. 1,B b), suggesting that TRAF6 stimulates both ERK and NFκB activity through its NH₂-terminal region containing the RING finger and zinc fingers. Parallel experiments showed that transient expression of full-length TRAF2 and TRAF5 had no effect on ERK activity, whereas it stimulated NFκB activity efficiently (Fig. 1 C; 6, 8, 30–32). In this context, previous reports also ruled out the role of TRAF2 in ERK activation (34, 36). Transient expression of TRAF3 stimulated neither ERK nor NFκB activity, in agreement with previous reports (7, 9, 34). Taken together, these results suggest that TRAF6, but not TRAF2, TRAF3, or TRAF5, is a transducer of both ERK and NFκB activation.

The Ras-dependent protein kinase cascade leading from various receptors to ERK is dependent on the protein kinase Raf-1, which activates a dual-specificity kinase, MEK (42). To investigate the role of Ras, Raf-1, and MEK1 in TRAF6-mediated ERK activation, we cotransfected 293 cells with a full-length TRAF6 and ERK-dependent activator and reporter, or GST-ERK, together with a dominant-negative mutant of H-Ras (N17Ras; 38), c-Raf-1, which contains only the NH₂-terminal regulatory domain (DNRaf; 37), or MEK1, in which Asp²⁰⁸ was replaced with Ala (MEK1^DN; 39). Some experiments were performed in parallel with the experiments depicted in Fig. 4. As shown in Fig. 2, coexpression of N17Ras had no effect on the activity of the cotransfected ERK-dependent reporter (Fig. 2,A) or on the exogenous ERK kinase activity (Fig. 2 B), which were increased five- to sixfold above the vector control by transient expression of TRAF6. The results suggest that TRAF6 may stimulate ERK activity along a Ras-independent pathway.

The dominant-negative DNRaf and MEK1^DN suppressed TRAF6-mediated ERK activity in the 293 cell line, but not entirely (Fig. 2, A and B). Although ERK activation is essentially induced by growth factors via a Ras-mediated pathway (42), the result could be interpreted as evidence that the diversified signal cascades regulate ERK activity, sometimes in a cell-type–specific manner, via c-Raf-1 and other possible MAPK kinase kinase in a Ras-independent fashion (46–51).

The truncated mutant of TRAF proteins that lack the RING finger or the RING finger and zinc fingers acts as a transdominant-negative mutant to suppress NFκB activation in signaling through TNFR1, TNFR2, IL-1R, CD30, and CD40 receptor (6, 8, 9, 28, 30, 31, 33), in which the dominant-negative mutant may act specifically by replacing the corresponding endogenous wild-type TRAF protein.

To investigate whether TRAF6 could be involved in ERK activation by CD40, we cotransfected 293 cells with the TRAF-C of TRAF6 and ERK-dependent activator and reporter, or GST-ERK, together with full-length CD40 cDNA. Because a previous report suggested that the NH₂-terminal CD40 cytoplasmic tail at positions 210–225 could be essential for TRAF6 binding and could function for NFκB activation (9), the involvement of TRAF6 in ERK activation by CD40 was also investigated by transient transfection of the TRAF-C of TRAF6 and ERK-dependent activator and reporter, or GST-ERK, into 293 cells coexpressing a deletion mutant of CD40, CD40Δ246 (9), that removes the COOH-terminal cytoplasmic tail at positions 226–257. Previous results showed that TRAF6 interacts with the CD40Δ246 construct in vivo (9).

As shown in Fig. 3, A and B, cross-linking of full-length CD40 and CD40Δ246 caused ERK activation approximately four- to sixfold above the level of unstimulated cells, comparable to the magnitude of the endogenous ERK activity that was transiently increased in 293 transfectants which constitutively expressed CD40 after stimulation with anti-CD40 mAb (data not shown). As shown in Fig. 3, A (a) and B both ERK-dependent reporter activity (Fig. 3,A) and the exogenous ERK kinase activity (Fig. 3 B) mediated by full-length CD40 were suppressed by coexpression of TRAF-C of TRAF6 in a dose-dependent manner to a maximum of 30–35% of inhibition, compatible with the notion that TRAF6-dependent and -independent pathways could be involved in ERK activation by CD40 (see below). Coexpression of the TRAF-C of TRAF6 caused prominent suppression of ERK activity mediated by cross-linking of CD40Δ246 in a dose-dependent manner, suggesting a major role of TRAF6 in ERK activation by CD40Δ246. It is not likely that the transdominant-negative mutant of TRAF6 nonspecifically diminished the activity of the ERK-dependent activator and reporter, because activation of the ERK-dependent reporter mediated by coexpression of catalytically active MEK1 (39) in the 293 cell line was not prevented by coexpression of TRAF-C of TRAF6 (data not shown).

In control experiments, transient expression of the TRAF-C of TRAF6 caused prominent suppression of the activity of cotransfected NFκB-dependent reporter mediated by CD40Δ246 in the 293 cell line (Fig. 3,C b), consistent with previous reports (9). In contrast to ERK activation, the activation of NFκB activity by full-length CD40 was greater than that by CD40Δ246, and ∼50% of the activity was suppressed by coexpression of the TRAF-C of TRAF6 (Fig. 3 C a). Because of a similar frequency and the level of expression of full-length CD40 and CD40Δ246 in the 293 cell line in transient transfection assays (data not shown), the results seem compatible with the notion that distinct regions of the CD40 cytoplasmic domain could be required for full activation of NFκB in the 293 cell line.

To investigate the role of Ras, Raf-1, and MEK in ERK activation by CD40, we cotransfected 293 cells with the ERK-dependent activator and reporter, or NFκB-dependent reporter as a control, and with N17Ras, DNRaf, or MEK1^DN, together with either full-length CD40 or CD40Δ246. The experiments were performed in parallel with some of the experiments depicted in Fig. 2,A. Cotransfection of N17Ras, DNRaf, and MEK1^DN caused ∼70% inhibition of ERK activity mediated by cross-linking of full-length CD40 (Fig. 4,A a). In contrast, coexpression of N17Ras and MEK1^DN did not affect NFκB activation by both full-length CD40 and CD40Δ246, whereas DNRaf marginally suppressed NFκB activation (Fig. 4 B). In conjunction with the previous observation indicating activation of the Ras-mediated pathway by CD40 stimulation in B cells (52), the results support the possibility that the Ras-Raf-MEK–mediated pathway could be involved in ERK activation, but not NFκB, by cross-linking of CD40. Transient expression of N17Ras resulted in ∼30% inhibition of ERK activity mediated by CD40Δ246, compatible with the notion that TRAF6 is a major transducer of ERK activation by CD40Δ246, whose activity is mediated by a Ras-independent pathway.

Taken together, these results support the possibility that CD40 may activate ERK via Ras-dependent and -independent pathways in which TRAF6 could be involved. The inhibition by a dominant-negative mutant of MEK1 in ERK activity by full-length CD40 was more prominent than in ERK activity mediated by TRAF6 and by CD40Δ246, in which TRAF6 may play a major role. Because the MEK isoforms MEK1 and MEK2 are an ERK-specific kinase (53), these results might imply that preferential usage of MEK isoforms would differ in TRAF6-dependent and -independent pathways in CD40 signaling. In this context, it has been suggested that the signaling pathways leading to activation of MEK isoforms appear to differ, and that each MEK isoform could be preferentially used in a particular external stimulus in a certain type of cell (54–56). Further study is needed to clarify the signaling pathway coupling TRAF6 to ERK activation.

Given that full ERK activity is stimulated through TRAF6-dependent and -independent pathways in CD40 signaling, a similar extent of ERK activation by full-length CD40 and CD40Δ246 in the 293 cell line led to the idea that ERK activity could be regulated in full-length CD40 signaling in 293 cells, probably via activation of MAPK phosphatase or downregulation of MEK activity (57–60). In contrast to ERK activation in murine splenic resting B cells (17, 18), it was reported that CD40 stimulation did not cause ERK activation in human tonsil B cells (15, 16). Whether ERK activation is regulated in CD40 signaling, depending on the maturation stage of B cells, would be worth analyzing.

We are grateful to Drs. G.M. Cooper (Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA), K. Maruyama (Tokyo Medical and Dental University, Tokyo, Japan), and K. Takishima (National Defense Medical College, Saitama, Japan) for reagents. We are also grateful to Drs. Tsunetsugu-Yokota and Z. Matsuda (NIID, Tokyo, Japan) for helpful discussions and reading of the manuscript, and to Mr. H. Shimizu, Ms. M. Takizawa, and K. Ishii for technical help.

This work was supported by a grant from the Agency of Technology and Science of the Japanese government and partly from the Ministry of Education and Science and Japanese Human Sciences Foundation to T. Takemori.