Activation of SAPK/JNK by TNF Receptor 1 Through a Noncytotoxic TRAF2-Dependent Pathway

See allHide authors and affiliations

Science  10 Jan 1997:
Vol. 275, Issue 5297, pp. 200-203
DOI: 10.1126/science.275.5297.200


Interaction of the p55 tumor necrosis factor receptor 1 (TNF-R1)-associated signal transducer TRADD with FADD signals apoptosis, whereas the TNF receptor-associated factor 2 protein (TRAF2) is required for activation of the nuclear transcription factor nuclear factor kappa B. TNF-induced activation of the stress-activated protein kinase (SAPK) was shown to occur through a noncytotoxic TRAF2-dependent pathway. TRAF2 was both sufficient and necessary for activation of SAPK by TNF-R1; conversely, expression of a dominant-negative FADD mutant, which blocks apoptosis, did not interfere with SAPK activation. Therefore, SAPK activation occurs through a pathway that is not required for TNF-R1-induced apoptosis.

Tumor necrosis factor is a pleiotropic cytokine that has growth modulatory, cytotoxic, and inflammatory activities (1). The effects of TNF are mediated by two distinct cell surface receptors of about 55 kD (TNF-R1) and 75 kD (TNF-R2), which are expressed on almost all nucleated cells (2). Proteins that are recruited to TNF-Rs after activation have been molecularly cloned, and their function has been partially characterized (3). The TNF-R1-associated death domain (TRADD) protein interacts with TNF-R1 upon TNF-induced trimerization, and its overexpression is sufficient to cause both activation of NFκB and apoptosis (4, 5). Other proteins associated with TNF-R1 by way of TRADD include the death domain protein FADD (5, 6) (which also participates in the transduction of Fas-induced apoptotic signals) and TRAF2 (5, 7). TRAF2 belongs to a family of signal transducers for both the TNF-R superfamily and interleukin-1 receptor I (8). The TRAF family is characterized by a conserved COOH-terminal TRAF domain and, with the exception of TRAF1, an NH2-terminal RING finger, which is relevant for signal transduction (8, 9). An additional molecule required for TNF-R1 signal transduction is RIP, a death domain-containing protein kinase that participates in both activation of NFκB and promotion of apoptosis (10). Whereas FADD is required for TNF-R1-induced apoptosis, the interaction of TRADD with TRAF2 is required for activation of NFκB and is dispensable for cytotoxicity (5, 9).

In addition to inducing apoptosis and activation of NFκB, cross-linking of TNF-R1 activates SAPK, also known as c-Jun NH2-terminal kinase (JNK). SAPK binds to and phosphorylates the transcription factor c-Jun within its NH2-terminal domain in cells exposed to environmental stresses [including TNF, ultraviolet (UV) light, protein synthesis inhibitors, and thermal stress] (11). The physiological consequences of SAPK activation have not been thoroughly defined. However, it has been shown that SAPK activity is required for apoptosis in nerve growth factor-deprived sympathetic neurons (12), as well as for stress-induced apoptosis in both leukemia cells and fibroblasts (13).

We therefore investigated the molecular mechanisms and the significance of TNF-induced activation of SAPK. Human embryonic kidney 293 cells contain very small amounts of endogenous TNF-R2, whereas TNF-R1 is constitutively expressed. Therefore, in these cells soluble TNF induces only TNF-R1 signaling (5, 9). Treatment of 293 cells with human recombinant TNF-α (hrTNF-α) induces a rapid and transient increase in SAPK activity that peaks at 15 min and reaches a maximal value of about three- to fivefold over the basal activity (Fig. 1) (14).

Fig. 1.

TNF activation of SAPK in 293 cells. Cells plated on gelatin-coated 60-mm culture plates were transfected by calcium phosphate coprecipitation, with 4 μg of HA-p46SAPKγ-pCDNA3 (14). Forty-eight hours after transfection, cells that had first been treated or not with Nac (50 mM, 4 hours) were stimulated with hrTNF-α (2500 IU/ml) for 15 min. HA-SAPKγ was precipitated by detergent lysates with monoclonal antibody to HA (12CA5), and its activity was determined (16) with GST-c-Jun(1–141) as substrate. The fold induction over the basal activity was quantitated with a PhosphorImager. The results (mean ± SD) of two experiments are shown. Control experiments were performed with mock-transfected cells.

Reactive oxygen species (ROS) such as superoxide, hydrogen peroxide, and hydroxyl radical function as second messengers in TNF-induced signal transduction. In cells treated with TNF-α, ROS are produced rapidly and mediate both cytotoxicity (15) and NFκB activation (16). ROS are also implicated in SAPK activation by some stimuli, such as UV-C rays (17). Treatment of 293 cells with N-acetylcysteine (Nac), a thiol anti-oxidant and glutathione precursor, had no apparent effect on the basal SAPK activity. However, initial treatment with Nac is sufficient to block TNF-α-induced activation of SAPK (Fig. 1), suggesting that this activity of TNF is also dependent on ROS.

To examine a possible role for TRAFs in activation of SAPK, we cotransfected 293 cells with TRAF expression vectors together with a HA-p46SAPKγ-pCDNA3 plasmid (14). Forty-eight hours after transfection, HA-SAPKγ was recovered by immunoprecipitation with a monoclonal antibody to the HA epitope (clone 12CA5), and its activity was measured in an immune complex kinase assay with glutathione-S-transferase (GST)-c-Jun as a substrate (14). Expression of TRAF1 had no effect on SAPK activity; in contrast, expression of TRAF2 increased SAPK activity in a dose-dependent manner to a maximum of about 15-fold relative to the control (Fig. 2, A and B). The amount of SAPK activation obtained with TRAF2 is greater than that obtained with TNF-α (Fig. 2A), consistent with the presence of limiting amounts of endogenous TNFR1-TRADD-TRAF 2 complexes.

Fig. 2.

Effects of TRAFs on SAPK activity in 293 cells. (A) TRAF2 activates SAPK. 293 cells were transfected with 4 μg of HA-p46SAPKγ-pCDNA3 (14) together with 4 μg of pRK-TRAF1 or pRK-TRAF2 (9). DNA concentration was always held constant by the addition of empty vector. Forty-eight hours after transfection detergent lysates were prepared, and the activity of exogenous transfected p46SAPKγ was determined (14). Expression of TRAF1 and TRAF2 in lysates from transfected cells is shown. Values shown are averages (mean ± SD) of four independent experiments. A Ser 63,73→Ala GST-Jun mutant was not phosphorylated by immunoprecipitated HA-SAPKγ (19). (B) Dose dependency of TRAF2-induced SAPK activation. The extent to which TRAF2 activated p46SAPKγ was dose dependent but did not further increase when amounts of pRK-TRAF2 greater than 4 μg per plate were used. (C) SAPK activation by TRAF2 is reduced by the anti-oxidant N-acetylcysteine. 293 cells were transiently cotransfected with 4 μg of HA-p46SAPKγ-pCDNA3 together with 4 μg of either pRK-TRAF2 (9) or pCMV-ΔMEKK (18). Forty-eight hours after transfection cells were treated with Nac (50 mM, 4 hours), and SAPK activity was determined (14). Nac had no effects on SAPK activation induced by either heat shock or osmotic stress (19), indicating that its inhibitory effect is selective for a subset of SAPK activators.

We next evaluated the effects of Nac on TRAF2-mediated activation of SAPK. Incubation with Nac (4 hours) reduced the extent to which SAPK was activated by TRAF2 by about 40% (Fig. 2C). The induction of SAPK by MEK kinase (MEKK)—a protein kinase that phosphorylates and activates SEK, which in turn phosphorylates and activates SAPK (18)—is unaffected by Nac. This suggests that ROS are downstream of both TNF-R1 and TRAF2, but are upstream of MEKK. A longer incubation with Nac caused a greater inhibition of TRAF2-induced SAPK activity, but also caused some toxicity (19).

Activation of NFκB by TRAF2 requires the presence of the NH2-terminal RING finger (9). In a similar manner, a TRAF2 mutant that lacks the first 86 amino acids, TRAF2(87–501) (9), had no effect on SAPK (Fig. 3A) even when expressed in large amounts. This TRAF2 mutant is still capable of both homotypic aggregation and interaction with TRADD; therefore, it acts as a dominant-negative inhibitor of both TNF-R1-, TNF-R2-, and TRADD-induced activation of NFκB (5, 9). Overexpression of TRAF2(87–501) resulted in a complete suppression of TNF-induced SAPK activation (Fig. 3B), which indicates that TRAF2 is required for SAPK activation by TNF-R1. In contrast, osmotic stress-, UV-C-, and anisomycin-mediated activation of SAPK was not affected by TRAF2(87–501) (Fig. 3C).

Fig. 3.

A TRAF2 deletion mutant lacking the NH2-terminal RING finger, TRAF2(87–501), acts as a dominant-negative inhibitor of TNF-induced SAPK activity. (A) The RING finger is required for SAPK activation by TRAF2. 293 cells transfected with various amounts of pRK-TRAF2(87–501) (9), and 4 μg of HAp46SAPKγ-pCDNA3 were analyzed for SAPK activity 48 hours after transfection as described (14). As a positive control, the effect of wild-type TRAF2 on SAPK activity is shown. 293 cell lysates from duplicate plates were subjected to protein immunoblot analysis with a rabbit polyclonal antibody raised against the COOH-terminus of TRAF2 (Santa Cruz). (B) TRAF2(87–501) blocks SAPK activation by TNF. HA-p46SAPKγ-pCDNA3 was cotransfected into 293 cells with either 4 μg of pRK-TRAF2(87–501) or vector. Forty-eight hours after transfection cells were stimulated with hrTNF-α (2500 UI/ml) for 15 min; detergent lysates were prepared and the activity of HA-SAPKγ was measured (14). Expression of TRAF2(87–501) in lysates from transfected cells is shown. (C) Effects of dominant-negative TRAF2 on SAPK induction by osmotic stress, UV-C, or anisomycin. 293 cells transfected with pRK-TRAF2(87–501) or vector plasmid were stimulated as indicated.

We evaluated the role of FADD, a death domain-containing protein that acts as a transducer of both Fas- and TNF-R1-initiated apoptotic signals (5, 6). FADD contains a COOH-terminal death domain, which mediates the association with the related death domains of Fas and TRADD (5, 6), and an NH2-terminal domain that is required for death induction and mediates the interaction with a downstream death effector known as MACH, FLICE, or Mch4 (20). A FADD mutant that lacks this NH2-terminal effector domain, FADD(80–205), acts as a dominant-negative inhibitor of both TNF-R1-, TRADD-, and Fas-induced apoptosis (5, 6). Overexpression of FADD(80–205) had no effect on basal SAPK activity. More importantly, it did not impair the ability of TNF to activate SAPK (Fig. 4), indicating that FADD is not a component of the TNF-R1-triggered signaling cascade resulting in SAPK activation.

Fig. 4.

Dominant-negative FADD does not interfere with SAPK activation by TNF. A FADD deletion mutant, FADD(80–205), that lacks the first 79 NH2-terminal amino acids (5, 6), which comprise the death effector domain (5, 6), was expressed in 293 cells [4 μg of pRK-FADD(80–205)] together with HA-p46SAPKγ. Forty-eight hours after transfection cells were stimulated with TNF-α (2500 IU/ml, 15 min), and HA-SAPKγ activity was measured as above (14). Expression of Flag-tagged FADD(80–205) was analyzed by protein immunoblotting in lysates from transfected 293 cells with a rabbit polyclonal antibody to Flag (Santa Cruz). The upper band is nonspecific.

We next studied the correlation between SAPK activation and TNF-induced apoptosis in HeLa cells. HeLa cells express predominantly TNF-R1 and are readily killed by TNF in the presence of protein- or RNA-synthesis inhibitors. In these cells, TNF activated SAPK about fourfold over basal activity; treatment with TNF plus actinomycin D (ActD) caused a similar activation of SAPK and induced cell death (Table 1). Activation of SAPK by either TNF or TNF plus ActD was completely blocked by TRAF2(87–501), the expression of which did not block apoptosis. In contrast, an increase in the number of apoptotic cells after treatment with TNF alone was observed in cells that express TRAF2(87–501); this effect may be due to the blockade of TNF-R1-induced activation of NFκB, which recently has been reported to protect against TNF-induced death (21). Thus, TNF-α-induced SAPK activation can be blocked at a membrane-proximal level without detrimental effects on cell death signaling. In a reciprocal manner, FADD(80–205) blocked apoptosis induced by TNF plus ActD but did not affect activation of SAPK (Table 1).

Table 1.

Cell death and activation of SAPK by TNF and TNF-R1-associated proteins in HeLa cells. For apoptosis assays, HeLa cells were transiently transfected with a β-galactosidase expression vector (pCDNA HislacZ, 1 μg) in the presence or absence of 5 μg of the indicated expression constructs encoding TRAF1, TRAF2, TRAF2(87–501), or FADD(80–205), respectively (5, 9). Forty-eight hours after transfection, the cells were treated with TNF in the presence or absence of ActD as indicated. After 12 hours the cells were fixed and stained with X-Gal. The data are expressed as the mean percentage (±SEM) of blue cells exhibiting signs of apoptosis (that is, intensely staining, shrunken blue cells showing loss of adherence) as a fraction of the total number of cells counted (shown in parentheses) (25). SAPK activation assays were performed as described (14). The data are from three independent experiments.

View this table:

Our results confirm that induction of apoptosis by TNF-R1 requires FADD; however, FADD is not required for activation of SAPK by TNF-α, demonstrating that the TNF-α-induced apoptotic pathway is separate from the pathway that leads to SAPK activation. Conversely, TRAF2 is both necessary and sufficient for SAPK stimulation, and yet TRAF2 expression does not result in (and is not required for) cytotoxicity (Table 1).

The correlation of SAPK activation by TNF-α and TRAF2 with activation of NFκB suggests that SAPK may cause NFκB activation. Indeed, activation of SAPK by MEKK results in NFκB activation, and kinase-inactive MEKK blocks TNF-induced NFκB activation in some cellular models (22). However, in our system dominant-negative SEK did not block TRAF2-mediated SAPK activation (19), suggesting a possible divergence of the pathways leading to the activation of SAPK and NFκB downstream of TRAF2. TRAF2 is also a component of the signaling apparatus of CD40 (9), a TNF-R-related transmembrane receptor whose cross-linking triggers activation of both NFκB and SAPK (9, 23). Because TRAF2 is required for CD40-induced NFκB activation (9), it may also mediate SAPK activation by this receptor.

Note added in proof: Liu et al. reported that recruitment of TRAF2 to the TNF-R1 complex mediates activation of SAPK/JNK (24).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
View Abstract

Navigate This Article