Protein kinase C activates the MEK-ERK pathway in a manner independent of Ras and dependent on Raf.

Although the involvement of protein kinase C (PKC) in the activation of the mitogen-activated protein (MAP) kinase pathway has been implicated through experiments using 12-O-tetradecanoylphorbol-13-acetate (TPA), there has been no direct demonstration that PKC activates the MAP kinase pathway. A Raf-dependent intact cell assay system for monitoring the activation of MAPK/ERK kinase (MEK) and extracellular signal-related kinase (ERK) permitted us to evaluate the role of PKC isotypes in MAP kinase activation. Treatment of cells with TPA or epidermal growth factor resulted in the activation of MEK and ERK. The activation of the MAP kinase pathway triggered by epidermal growth factor was completely inhibited by dominant-negative Ras (RasN17), whereas the activation triggered by TPA was not, consistent with previous observations. The introduction of an activated point mutant of PKCdelta, but not PKCalpha or PKCepsilon, resulted in the activation of the MAP kinase pathway. The activation of MEK and ERK by an activated form of PKCdelta requires the presence of c-Raf and is independent of RasN17. These results demonstrate that activation of PKCdelta is sufficient for the activation of MEK and ERK and that the pathway operates in a manner dependent on c-Raf and independent of Ras.

Although the involvement of protein kinase C (PKC) in the activation of the mitogen-activated protein (MAP) kinase pathway has been implicated through experiments using 12-O-tetradecanoylphorbol-13-acetate (TPA), there has been no direct demonstration that PKC activates the MAP kinase pathway. A Raf-dependent intact cell assay system for monitoring the activation of MAPK/ERK kinase (MEK) and extracellular signal-related kinase (ERK) permitted us to evaluate the role of PKC isotypes in MAP kinase activation. Treatment of cells with TPA or epidermal growth factor resulted in the activation of MEK and ERK. The activation of the MAP kinase pathway triggered by epidermal growth factor was completely inhibited by dominant-negative Ras (RasN17), whereas the activation triggered by TPA was not, consistent with previous observations. The introduction of an activated point mutant of PKC␦, but not PKC␣ or PKC⑀, resulted in the activation of the MAP kinase pathway. The activation of MEK and ERK by an activated form of PKC␦ requires the presence of c-Raf and is independent of RasN17. These results demonstrate that activation of PKC␦ is sufficient for the activation of MEK and ERK and that the pathway operates in a manner dependent on c-Raf and independent of Ras.
Mitogen-activated protein kinases (MAP 1 kinases; ERK1 and ERK2) are common intermediates in intracellular signaling cascades involved in diverse cellular functions including growth and differentiation (1,2). The activation of MAP kinases requires the dual-phosphorylation of Thr and Tyr residues by activating kinases, MAP kinase kinases (MEK1 and MEK2). The activity of MAP kinase kinases is also regulated by phosphorylation, and the responsible kinases have been identified (3,4). c-Raf is a MAP kinase kinase kinase that directly phosphorylates and activates MEK1 and MEK2 (5,6). The extracellular stimuli that activate MAP kinases include insu-lin, EGF, platelet-derived growth factor, nerve growth factor, serum, phorbol esters, nicotine, okadaic acid, and activators of oocyte maturation. The signaling pathway involved in the activation of ERKs has been intensively studied for c-Raf and Ras, and the mechanism of their activation has been analyzed in detail. This has led to an understanding of the presence of a linear array of signaling pathway initiated by tyrosine kinases to activate Ras, c-Raf, MEKs, and ERKs (7). However, there remain many fundamental questions concerning the mode of activation of c-Raf and the signaling pathway from extracellular stimuli to the Raf-MEK-ERK pathway.
One of these questions concerns the role of protein kinase C (PKC), a family of enzymes activated through a pathway involving a diverse set of lipid metabolites activated by phospholipase C, phosphatidylinositol 3-kinases, and other molecules (8,9), in addition to phorbol esters such as TPA and phorbol 12,13-dibutyrate, in the activation of the Raf-MEK-ERK pathway. Treatment of cells with TPA results in the activation of c-Raf (10) and MAP kinase (11)(12)(13) within minutes, suggesting the involvement of PKC in the signaling pathway leading to MAP kinase activation. Although these observations suggest a link between PKC and MAP kinase activation, there has been no direct demonstration that some members of the PKC family actually activate the pathway. Furthermore, the target of PKC in the activation of the MAP kinase pathway, if PKC is involved, remains to be clarified. The involvement of Ras (14 -22) and c-Raf (19,21,(23)(24)(25)(26)(27)(28) in the TPA-induced activation of MAP kinases has been reported with rather paradoxical results.
Recently, G protein ␤␥ has been reported to be involved in the Ras-dependent activation of MAP kinases (29 -31), and G o protein ␣ has been reported to be involved in Ras-independent activation of MAP kinases (32). Furthermore, a protein tyrosine kinase, PYK2, has also been reported to be involved in Ras-dependent MAP kinase activation (33). The involvement of PKC in this novel pathway has also been suggested.
However, the above studies used TPA, raising the fundamental question of whether PKC is actually involved. The presence of a set of cellular proteins that bind to phorbol esters and the observation that their activities are modulated by phorbol esters also raises the question of whether the effect of TPA on MAP kinases actually involves PKC. Cellular phorbol ester receptors other than PKC include Ras-activating guanine nucleotide exchange factor and Rac-GTPase activating protein (34,35). In the present study, we addressed the question of whether PKC is actually involved in MAP kinase activation by TPA, and, if it is, which PKC isotype is involved. Using a series of PKC kinase-knockout mutants and mutants with constitutive kinase activity, we show that PKC␦ is actually involved in the signaling pathway from TPA to Raf-MEK-ERK activation that operates in a Ras-independent manner.
Cell Culture and Transfection-COS1 cells and NIH3T3 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum or 7% calf serum, respectively. For monitoring MEK and PKC activity by immunoprecipitation, COS1 cells were cotransfected with tag-MEK1 cDNA expression plasmid together with the cDNA expression plasmids of c-Raf, PKC, or RasN17, as indicated using a Gene Pulser apparatus (Bio-Rad) (49). The transfected cells were seeded in 10% fetal calf serum-DMEM at a density of 1.5 ϫ 10 6 cells/ 10-cm dish. After 24 or 48 h of culture, the cells were washed twice with ice-cold phosphate-buffered saline and scraped into 0.2 ml/dish of cell lysis buffer (20 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 10 mM 2-mercaptoethanol, 0.1 mg/ml leupeptin, 2 mM phenylmethylsulfonyl fluoride, and 0.5% Triton X-100). The cells were disrupted by brief sonication, and the supernatant was recovered by centrifugation. When indicated, cells were starved for 18 -20 h in serum-free DMEM and then stimulated with 100 ng/ml TPA or 10 ng/ml EGF for 10 min. The total amount of transfected cDNA was adjusted to the corresponding empty vector in all experiments.
Measurement of MEK Activity-For the immunoprecipitation of tag-MEK1, the supernatant from COS1 cells was preincubated with protein G-Sepharose 4 Fast Flow (Pharmacia Biotech Inc.) for 30 min at 4°C. After removing materials bound to protein G-Sepharose, the tag-MEK1 protein was immunoprecipitated with anti-T7-tag monoclonal antibody (Novagen), which was pre-adsorbed to the protein G-Sepharose by incubation for 1.5 h at 4°C. The Sepharose resin was washed four times with buffer A (10 mM Tris-HCl, pH 7.5, 500 mM NaCl, 0.5% Nonidet P-40, 2 mM phenylmethylsulfonyl fluoride, 0.1 mg/ml leupeptin, and 10 mM 2-mercaptoethanol) and then once with buffer B (20 mM Tris-HCl, pH 7.5, 0.5 mM EDTA, 0.5 mM EGTA, 10% glycerol, 2 mM phenylmethylsulfonyl fluoride, 0.1 mg/ml leupeptin, and 10 mM 2-mercaptoethanol), and the immunoprecipitates were mixed with 100 l of 18 mM HEPES, pH 7.5. The precipitate was recovered by centrifugation. The immunoprecipitated tag-MEK1 fixed on 10 l of protein G-Sepharose was suspended in 40 l of assay mix containing 18 mM HEPES, pH 7.5, 10 mM Mg(OAc) 2 , 50 M ATP, 2 Ci of [␥-32 P]ATP (Amersham Corp.), and 0.36 g of bacterially produced GST-ERK1 for 10 min at 30°C, and then 20 g of MBP (Sigma M1891) for 20 min at 30°C. One-half of the reaction mixture (20 l) was transferred onto a 2 cm ϫ 2-cm piece of p81 phosphocellulose paper (Whatman). The filters were washed four times with 75 mM phosphoric acid. Phosphorylation was quantitated by scintillation counting. The remaining materials were eluted with SDSpolyacrylamide gel electrophoresis sample buffer for Western blot analysis.
Expression and Purification of GST-ERK1-Rat ERK1 cDNA was cloned into the EcoRI site of the bacterial expression vector pGEX-3X to yield pGEX-3X:ERK1. The construct was transformed into Escherichia coli strain XL1-Blue, and recombinant GST-ERK1 protein was expressed and purified as reported (50).
ERK Mobility Shift Assay-For ERK mobility shift assay, the cells were transfected by calcium phosphate co-precipitation methods (44). After 6 h exposure to the calcium phosphate-DNA precipitate, the medium was changed to 7% calf serum-DMEM for NIH3T3 cells or 10% fetal calf serum-DMEM for COS1 cells. After an additional 18 h of culture, the cells were washed and lysed in SDS-polyacrylamide gel electrophoresis sample buffer. When indicated, cells were starved for 18 -20 h in serum-free DMEM for COS1 cells or 0.5% fetal calf serum-DMEM for NIH3T3 cells after exposure to the calcium phosphate-DNA precipitate for 6 h; then cells were stimulated with 100 ng/ml TPA, 10 ng/ml EGF, or 50 ng/ml platelet-derived growth factor for 5 min or 10 min. The cell lysate was briefly sonicated and then subjected to SDSpolyacrylamide gel (10%) electrophoresis, followed by Western blot analysis. The intensity of the band corresponding to the tag-ERK1 was quantified by a densitometer. The mobility shift of tag-ERK1 was expressed as the percentage of the shifted band in relation to the total amount of tagged ERK1.
Measurement of PKC Activity-For immunoprecipitation of PKC, the supernatant from COS1 cells was preincubated with protein A-Sepharose CL-4B (Pharmacia) for 30 min at 4°C. After centrifugation to remove materials bound to protein A-Sepharose, each PKC protein was immunoprecipitated with its respective anti-PKC antibody pre-adsorbed to the protein A-Sepharose by incubation for 1.5 h at 4°C. The Sepharose resin was washed in the same way as for the immunoprecipitation of tag-MEK1, and the immunoprecipitates were mixed with 100 l of 20 mM Tris-HCl, pH 7.5. The precipitate was recovered by centrifugation. Antibody against PKC␣ was obtained from guinea pigs immunized with bacterially synthesized protein with an amino acid sequence corresponding to the sequence 18 -672 of rabbit PKC␣ (42). Antibody against PKC␦ was obtained from rabbits immunized with synthetic oligopeptide with an amino acid sequence corresponding to the C-terminal sequence 656 -673 of rat PKC␦ (36). The immunoprecipitated PKC was fixed on 10 l of protein A-Sepharose and suspended in 40 l of assay mix containing 20 mM Tris-HCl, pH 7.5, 5 mM Mg(OAc) 2 , 0.2 mM CaCl 2 , 50 M ATP, 2 Ci [␥-32 P]ATP (Amersham), 0.01 mg/ml leupeptin, and 50 g/ml MBP4-14 as a substrate (36) with or without 25 g/ml phosphatidylserine (Avanti Polar Lipid, Inc.) and 50 ng/ml TPA (Sigma) for 20 min at 30°C. PKC activity was measured in the same way as MEK activity.
Western Blot Analysis-Following SDS-polyacrylamide gel electrophoresis, the separated proteins were electrophoretically transferred to a polyvinylidene difluoride membrane, and the membrane was soaked in phosphate-buffered saline containing 5% skimmed milk for 1 h at room temperature. ERK and MEK were detected using polyclonal antibodies raised against ERK1 (06 -182) and MEK1 (06 -235) (Upstate Biotechnology, Inc.), respectively. PKC isotypes were detected using monoclonal antibodies raised against PKC␣ (P16520), PKC␦ (P36520), and PKC⑀ (P14820) (Transduction Laboratories). Horseradish peroxidase-conjugated sheep anti-rabbit or mouse Ig (Amersham) was used as a secondary antibody for signal detection by the ECL detection system (Amersham).

A Raf-dependent Pathway for MEK and ERK Activation in
Intact Cells-To address the question of whether PKC is actually involved in the activation of MAP kinase and, if it is, how PKC activates MAP kinase, we devised an assay system in COS cells where the effect of exogenous proteins in the activation of MEK1 and ERK1 could be evaluated. Introducing tagged MEK1 into COS cells, stimulating the cells with TPA or EGF, immunoprecipitating the tagged MEK1, and measuring the MEK1 activity in vitro using recombinant GST-ERK1 as a substrate permitted us to monitor the activation of MEK1 in response to TPA. Fig. 1A shows that treatment of cells with TPA or EGF results in 5-6-fold activation of MEK1 within 10 min. Furthermore, the co-expression of c-Raf potentiates the activation of MEK1 by TPA. The potentiation of MEK1 by c-Raf co-expression was also seen following EGF treatment, although the response was weaker. These results are consistent with previous observations that TPA-stimulated ERK activation is potentiated by the overexpression of c-Raf (19) and that the N-terminal fragment of c-Raf or the point mutant of c-Raf (K375W) suppresses the TPA-induced mobility shift of ERKs (21,23,28). Furthermore, the results demonstrate that this system can be used for the analysis of the Raf-dependent pathway for MEK1 and ERK1 activation.
To examine the involvement of Ras in the above system, we used a dominant-negative Ras mutant, Ki-RasN17. As shown in Fig. 1A, RasN17 only slightly suppresses the TPA-induced activation of MEK1, whereas it completely suppresses the EGF-induced activation of MEK1. Similar results were also obtained for ERK1 in COS cells and NIH3T3 cells, as shown in Fig. 1, B and C, where tagged ERK1 was used instead of tagged MEK1, and the upward shift in the electrophoretic mobility of tagged ERK1 was evaluated using an anti-ERK antibody. These results confirm the previous observations that the acti-vation of ERKs by TPA depends on c-Raf but not on Ras (19,21).
Dominant-negative Action of PKC Kinase-Knockout Mutants on TPA-induced MEK Activation-Since PKC is the major receptor for phorbol esters, PKC is the most obvious candidate for the substance that mediates the TPA-induced activation of MEK and ERK shown above. However, there have been few in vivo demonstrations that any PKC members are actually involved in the signaling pathway. Only PKC␤1 has been reported to potentiate the TPA-induced activation of c-Raf when overexpressed in insect cells (24). To address this issue, we next examined the effect of a series of kinase-knockout point mutants of PKC isozymes expressed ubiquitously in a wide variety of cells, including COS and NIH3T3 cells.
We introduced a series of PKC kinase-knockout mutants into COS cells and evaluated their effects on TPA-induced MEK1 activation. The kinase-knockout mutants were designed to have "activated conformation" by introducing mutation(s) into the pseudosubstrate region (see "Discussion"). Fig. 2 shows that the overexpression of the kinase-knockout mutants of PKC␣, PKC␦, and PKC⑀ all produce inhibition of the TPAinduced activation of MEK1 in a dose-dependent manner. However, this inhibition is incomplete at the highest DNA amount when the amount of the mutants was more than 100 times relative to their corresponding counterparts (Fig. 2D). The results suggest that PKC and/or its homologue is involved the signaling pathway from TPA to MEK1 activation. However, the results raise the question of the specificity of the kinase-knockout mutants on the action of each respective PKC isotype.
Constitutively Active PKC␦ Activates the MEK-ERK Pathway-To examine the specificity of the action of each PKC isotype directly, we next tried to determine the effect of the overexpression of PKC members on the TPA-induced activation of MEK1 and ERK1. However, we failed to detect any significant effect (data not shown), suggesting that the amount of endogenous PKC in COS and NIH3T3 cells is sufficient to mediate the TPA-induced activation of MEK1 and ERK1. Thus, we next constructed a series of constitutively active PKC mutants (Fig. 3A).
PKC␣ (␣R22A/A25E) has two point mutations in its pseudosubstrate region. The kinase activity of the mutant was evaluated after immunoprecipitation and showed clearly that PKC␣ (␣R22A/A25E) is fully active in the absence of cofactors (Fig. 3B). The introduction of the PKC␣ mutant with reporter plasmids containing TPA response elements (TRE-luciferase) resulted in gene expression without TPA stimulation (data not shown), consistent with the results of the in vitro kinase assay. PKC␦ (DR144/145A) also contains two point mutations and is fully active without cofactors (Fig. 3C). This is consistent with the previous observation that the introduction of PKC␦ (DR144/145A) into NIH3T3 cells results in the activation of reporter gene expression (TRE-tk-CAT) without any stimuli (44). PKC⑀ (⑀A159E) contains a point mutation in its pseudosubstrate region. The in vitro kinase assay using immunoprecipitated PKC⑀ failed because of the absence of an antibody able to immunoprecipitate PKC⑀ and monitor its kinase activity in a cofactor-dependent manner. However, the same mutant has been shown to be apparently constitutively active when monitored in terms of reporter gene expression (45). The overexpression of the respective wild-type or constitutively active mutant of PKC␣, ␦, or ⑀ was confirmed by Western blot analysis (Fig. 3A).
We next overexpressed the constitutively active PKC mutants and examined their effects on MEK1-ERK1 activation.
Overexpression of PKC␣, PKC␣ (␣R22A/A25E), PKC⑀, and PKC⑀ (⑀A159E) failed to activate MEK1 (Fig. 4A) or ERK1 (Fig.  4, B and C). Another constitutively active mutant of PKC␣, the kinase domain, also failed to activate MEK1 under our assay conditions (data not shown), although the same mutant induces TRE-CAT expression without any stimuli (51). However, the overexpression of PKC␦ resulted in a 3-fold enhancement of MEK1 activation, and the overexpression of its constitutively active mutant, PKC␦ (DR144/145A), resulted in an 8-fold enhancement of the MEK1 activation (Fig. 4A). Although the amount of each construct of PKC␣, ␦, or ⑀ cannot be directly compared (Fig. 3A), experiments using less (1 g) and more (5 g) amounts of the expression vector gave similar results (data not shown). The overexpression of PKC␦ (DR144/145A) also resulted in the activation of ERK1 (Fig. 4: B, COS cells, and C, NIH3T3 cells). The overexpression of wild-type PKC␦, but not other PKC isotypes, resulted in a slight activation (30% mobility shift) of ERK1 in NIH3T3 cells (Fig. 5, B and C). These results clearly show that PKC␦ is a likely mediate of the action of TPA in MEK1 and ERK1 activation in the assay system used in the present study.

PKC␦-mediated MEK/ERK Activation Requires c-Raf But Is Independent of Ras-
The ability of the PKC␦ mutant, PKC␦ (DR144/145A), to activate MEK1 and ERK1 prompted us to examine the involvement of Ras in the signaling pathway. As shown in Fig. 5A, PKC␦ (DR144/145A) cannot activate MEK1 in the absence of c-Raf, whereas c-Raf co-expression results in MEK1 activation. Furthermore, MEK1 activation by PKC␦ (DR144/145A) is not inhibited by RasN17. PKC␦ (DR144/145A) activates ERK1, as shown in Fig. 5, B and C. The activation of ERK1 requires the presence of c-Raf and is independent of the presence of RasN17 (Fig. 5, B and C). Thus, PKC␦ (DR144/ 145A) activates MEK1 and ERK1 in a manner dependent on c-Raf. Furthermore, activation occurs in the presence of RasN17.

DISCUSSION
The effect of TPA on the tyrosine phosphorylation of the protein p42 was observed before the discovery of MAP kinases (52)(53)(54), and the effect of TPA on the activation of MAP kinase was also reported in the original identification of MAP kinases  (11)(12)(13). Since TPA directly activates PKC, it has been implicated in the signaling pathway from TPA to MAP kinase activation. However, it is now evident that there are proteins in addition to PKC isotypes that are potential cellular phorbol ester receptors. These include Vav, a guanine-nucleotide exchange protein known to be involved in the activation of Ras (34), and n-chimaerin, a GTPase activating protein, which inactivates another small G-protein, Rac (35). In the present study, we used an intact cell assay system in which the TPAinduced activation of MEK1 and ERK1 occurs in a Raf-dependent manner and demonstrated that the kinase-knockout mutants of PKC␣, PKC␦, and PKC⑀ inhibit the TPA-induced activation of MEK1, indicating that PKC or its homologues are involved in the signaling pathway from TPA to MEK-ERK. More importantly, we showed that a constitutively active mutant of PKC␦ causes the activation of MEK1 and ERK1. This is one of the most important points of the present study, since it excludes the possibility that TPA-induced MAP kinase activation requires phorbol ester receptors other than PKC isotypes. Quite interestingly, the constitutively active mutants of PKC␣ and PKC⑀ failed to activate MEK1 and ERK1 in a similar assay system. These results show that PKC␦ is actually involved in and is sufficient for the activation of MEK and ERK. Furthermore, PKC␦ activates MEK1 and ERK1 in the presence of RasN17, indicating that the signaling pathway operates in a Ras-independent manner.
The constitutively active form of PKC was reported in the original identification of PKC (55). Deletion of the regulatory domain of PKC␣ or PKC␤ allows them to be constitutively active; their expression results in the activation of c-fos promoter and TRE elements and the induction of Xenopus oocyte maturation (51, 56, 57). However, the regulatory domain of PKC has also been suggested to be involved in substrate recognition, and its deletion might result in the generation of a mutant with altered substrate specificity (58). Another approach to the construction of the constitutive active PKC is to introduce a point mutation into the pseudosubstrate region, which is suggested to mask the substrate recognition site of the kinase domain (59). PKC␣ (A25E) shows increased cofactorindependent activity in vitro and increased activity of collagenase-CAT expression as compared to the wild type (60). However, whether this mutation results in full activation without cofactors remains to be clarified. PKC␣ (R22A/K23A) and (R27A) show cofactor-independent activity, as judged by in vitro kinase assay (61). These mutants show high phosphorylation levels in vivo and are located in the cytoplasmic and particulate fractions but are not fully active in inducing Xenopus oocyte maturation (61). Thus, we introduced two point mutations into the pseudosubstrate region of PKC␣ to construct PKC␣ (␣R22A/A25E), which shows completely cofactorindependent activity (Fig. 3B) and induces reporter expression (data not shown). The constitutively active mutant of PKC␦ (DR144/145A) has been shown previously to be active, as judged by its induction of TRE-CAT expression (44). In the present study, we showed that this mutant does, in fact, have cofactor-independent activity (Fig. 3C). Although we failed to confirm that PKC⑀ (⑀A159E) shows cofactor-independent activity in the in vitro kinase assay, the mutant was shown to be active, as judged by its activation of reporter genes such as TRE-CAT, NFAT-CAT, and NFB-CAT (45), strongly supporting the idea that this mutant is active in vivo. This excludes the possibility that the failure of these mutants to activate MEK and ERK is due to the fact that they are not fully active. One possible explanation for the failure of PKC␣ and PKC⑀ to activate MEK and ERK is that they overactivate MEK and ERK and cause their own down-regulation. However, this possibility is unlikely since smaller amounts of DNA also failed to activate ERK1 (Fig. 4, B and C). Thus, the failure of the PKC␣ and PKC⑀ mutants to activate MEK and ERK may reflect the intrinsic properties of the PKC isotypes.
The dominant-negative effects of the kinase-knockout mutants of PKC␣, PKC␦, and PKC⑀ on the TPA-induced activation of MEK1 might suggest the involvement of these PKC members and/or other PKC homologues. However, the dominantnegative effect was incomplete at the highest level of the kinase-knockout mutant expression (Fig. 2D). Furthermore, there was no isozyme specificity making a clear contrast to the results obtained by constitutively active mutants. This might suggest that the dominant-negative effect was caused by the competition of TPA rather than their target protein(s). In addition, TPA might activate MEK and ERK through a pathway that does not involve PKC.
There are several reports that focus on PKC isotypes and their potential for MAP kinase activation. Marquardt et al. (24) have reported that the co-expression of PKC␤1 and c-Raf in insect cells results in a potentiation of the TPA-induced activation of c-Raf when the activity of the immunoprecipitated c-Raf was evaluated by an in vitro GST-MEK activation assay. It has been reported that purified PKC␣ phosphorylates c-Raf and activates autokinase activity (62,63). Note that the authors did not examine the activity of c-Raf in terms of MEK activation, and that there is a report that purified brain PKC (cPKC mixture) phosphorylates c-Raf in vitro and activates its autokinase activity but does not activate its MEK phosphorylation activity (64). The results of Marquardt et al. (24) showing that PKC␤1 potentiates the TPA-induced activation of c-Raf to phosphorylate MEK suggests the involvement of PKC␤1 in TPA-induced c-Raf activation. However, the use of TPA does not directly show that PKC␤1 is sufficient to activate c-Raf. The present demonstration using constitutively active mutants of PKC isotypes that PKC␦ but not PKC␣ or PKC⑀ activates MEK1 and ERK1 clearly indicates that the activation of PKC␦ is sufficient to activate the MEK-ERK pathway and strongly suggests that PKC␦ is a likely activator of the MEK-ERK pathway.
There are apparently paradoxical observations about the role of Ras in the phorbol ester-induced activation of MAP kinases. For example, in PC12 cells and NIH3T3 cells, the TPA-induced activation of MAP kinases involves Ras (14,16,17,33). On the other hand, RasN17 fails to inhibit TPA-induced MAP kinase activation in Rat 1 cells, COS cells, and 293 cells (18,19,21). It should be noted that these are all observations using TPA or phorbol 12,13-dibutyrate and not a direct examination of the role of PKC. Considering that TPA directly modulates effectors of Ras, the action of TPA might involve these proteins in addition to PKC. There is a report in which cell-free extracts of Xenopus oocytes were used to show that an active fragment of brain PKC (cPKC mixture) activates MAP kinases and is inhibited by RasN17 (26). This suggests the presence of a Ras-dependent pathway involving PKC isotypes such as PKC␣ (Fig. 5D).
c-Raf has been shown to phosphorylate and activate MEK directly (5). The overexpression of c-Raf potentiates the TPAinduced activation of MEK1 and ERK1 (Fig. 1A) (19), suggesting the involvement of c-Raf in the signaling pathway from TPA to MEK-ERK. Consistent with this is the report that the N-terminal fragment of c-Raf, or a point mutation of c-Raf (K375W), inhibits the TPA-induced activation of MAP kinase in COS cells and in 293 cells (21,23,27,28). Furthermore, the activation of c-Raf in TPA-treated COS cells has been shown in vitro using immunoprecipitated c-Raf and MEK as substrates (19). However, there is also a report that a kinase-knockout mutant of c-Raf fails to inhibit the phorbol 12,13-dibutyrateinduced activation of MAP kinase in LA-90 cells (25). Several lines of evidence indicate that PKC phosphorylates c-Raf in TPA-treated cells. In vitro, PKC␣ phosphorylates c-Raf, presumably at site S499 (63), and activates c-Raf in terms of a kinase activity against a peptide spanning the c-Raf autophosphorylation site. The phosphorylation of c-Raf at S499 occurs in vivo when A293 cells are treated with TPA. TPA treatment of insect cells co-expressing PKC␣ and c-Raf results in the activation of a c-Raf kinase activity against a peptide spanning the c-Raf autophosphorylation site, and the c-Raf S499A mutant is not activated under the same condition (63). These in vivo results confirm the notion obtained from in vitro results that PKC␣ phosphorylates c-Raf at S499 and activates its autokinase activity. Carroll and May reported that rat brain PKC (cPKC mixture) phosphorylates c-Raf at S497 and S619 and activates c-Raf kinase activity against histones III (65). However, as mentioned before, there has been no direct demonstration that c-Raf phosphorylated by PKC has an increased capacity to activate MEK. It is not clear whether PKC␦ directly activates c-Raf. Whether PKC␦ interacts directly with c-Raf and activates its activity toward MEK remains the subject for future experiments.
FIG. 5. Constitutively active PKC␦ activates MEK1 and ERK1 in a Ras-independent manner. A, activation of MEK1. COS1 cells (5.0 ϫ 10 6 cells) were co-transfected with tag-MEK1 (5 g) cDNA expression plasmid together with the cDNA expression plasmids of c-Raf (5 g), DR144/145A (1 g), or RasN17 (5 g) as indicated. After 24 h of culture on 10-cm dishes, the cells were harvested. tag-MEK1 was immunoprecipitated, and its activity was evaluated. Values represent the means of three independent experiments; bars, S.D. B and C, activation of ERK1. COS1 (B) or NIH3T3 (C) cells (1.8 ϫ 10 5 cells) cultured on 15-mm dishes were co-transfected with tag-ERK1 (2 g) and MEK1 (1 g) cDNA expression plasmids together with the cDNA expression plasmids of c-Raf (1 g), PKC␦ (1 g), DR144/145A (1 g), or RasN17 (2 g) as indicated. tag-ERK1 was visualized using anti-ERK1 antibodies. The figure shows the results of typical experiments, and similar results were obtained in three separate experiments. D, schematic drawing of the proposed mechanism for MAP kinase activation by PKC.