Role of Dynamin, Src, and Ras in the Protein Kinase C-mediated Activation of ERK by Gonadotropin-releasing Hormone*

G-protein-coupled receptors are a large group of integral membranal receptors, which in response to ligand binding initiate diverse downstream signaling. Here we studied the gonadotropin-releasing hormone (GnRH) receptor, which uses Gq for its downstream signaling. We show that extracellular signal-regulated kinase (ERK) activation is fully dependent on protein kinase C (PKC), but only partially dependent on Src, dynamin, and Ras. Receptor tyrosine kinases, FAK, Gβγ, and β-arrestin, which were implicated in some G-protein-coupled receptor signaling to MAPK cascades, do not play a role in the GnRH to ERK pathway. Our results suggest that the activation of ERK by GnRH involves two distinct signaling pathways, which converge at the level of Raf-1. The main pathway involves a direct activation of Raf-1 by PKC, and this step is partially dependent on a second pathway consisting of Ras activation, which occurs in a dynamin-dependent manner, downstream of Src.

The family of receptors coupled to G-proteins (GPCR) is the largest group of integral membranal receptors involved in signal transduction (reviewed in Refs. 8 -12). Members of this group of receptors transmit their signals primarily via GTPbinding proteins (G-proteins), which are heterotrimeric proteins composed of ␣, ␤, and ␥ subunits. The G-proteins are classified according to the subtype of their ␣ subunit into four groups: G s , G i , G q , and G 12 . Upon activation, the heterotrimer dissociates into a GTP-bound, activated, ␣ subunit (G␣), and a dimer of G␤␥. Whereas a small number of common mechanisms is responsible for receptor Tyr kinase-mediated activation of ERK (13), GPCRs are thought to activate ERK by divers signaling pathways. All four G␣ subunits (12), the ␤␥ subunits (14), receptor-interacting proteins such as Src (15), ␤-arrestin (16), dynamin (17), and transactivation of receptor Tyr kinases by GPCRs (18) are capable of initiating several downstream signaling pathways culminating in MAPK activation. The various MAPK cascades appear to be important components in the downstream signaling events initiated by the GPCRs, although different receptors often use different mechanisms for this purpose (12).
Gonadotropin releasing hormone (GnRH) is a hypothalamic decapeptide, which serves as a key regulator of the reproductive system. In the pituitary, the signals of GnRH are transmitted via a specific GPCR (GnRH receptor; GnRHR), which upon activation interacts with the heterotrimeric G q protein (reviewed in Refs. 19,20). This interaction then initiates a variety of intracellular signaling events that culminate in the production and secretion of the leutinizing hormone and follicle-stimulating hormone. In recent studies it was shown that the JNK, ERK, and p38 MAPK cascades are activated in response to GnRH stimulation of ␣T3-1 cells (21)(22)(23)(24)(25)(26)(27)(28). The mechanism of JNK activation by GnRH has been shown to involve sequential activation of protein kinase C (PKC), Src, CDC42, and MEKK1 (28); and the mechanism of p38 MAPK activation by GnRH seems also to be PKC-dependent (27). The mechanism of ERK activation was studied by several laboratories over the past few years and was shown to involve PKC, and an unknown protein-tyrosine kinase (PTK) that is only partially sensitive to the PTK inhibitor genistein (reviewed in Ref. 12).
Here we show that the activation of ERK by GnRH in the ␣T3-1 cells involves two distinct signaling pathways, which converge at the level of Raf-1. One of these pathways involves a direct activation of Raf-1 by PKC, whereas the other pathway involves activation of Raf-1 by Src and Ras. Transactivation of EGF receptor, or activation through either ␤-arrestin or G␤␥, which play a role in the signaling of many other GPCRs to the ERK cascade, are not involved in the GnRHR to ERK pathway.
On the other hand, dynamin seems to be important for this pathway, because it is essential for the activation of Ras, in a PKC-independent manner. Thus, although the GnRHR to ERK signaling is mainly mediated by G q -PKC, another pathway involving dynamin is required for the Src-mediated activation of Ras, which supports the step of Raf-1 activation by PKC.
Transfection, Stimulation, and Harvesting of ␣T3-1 Cells-Subconfluent ␣T3-1 cells were transfected with 5 g of the examined plasmid. In case of cotransfection, 5 g of each plasmid were used. The transfection was carried out using the calcium phosphate technique. The total amount of plasmid was adjusted to 10 g with vector DNA in the control experiments. The transfection efficiency was 10 -30%, as determined by transfection with a plasmid that contained ␤-Gal and appropriate staining. Two days after transfection, the cells were serumstarved for 16 h and incubated for the desired time intervals with GnRH-a in the presence or absence of various inhibitors. After stimulation, cells were washed twice with ice-cold PBS, washed once with buffer A, and subsequently harvested in ice-cold buffer H. Cell lysates were centrifuged (20,000 ϫ g, 20 min), and the supernatant was assayed for protein content.
Western Blot Analysis-Cell supernatants, which contained the cytosolic proteins, were collected, and aliquots from each sample (20 g) were separated on 10% SDS-PAGE followed by Western blotting with the appropriate antibodies. Alternatively, immunoprecipitated antibodies were boiled in sample buffer and subjected to SDS-PAGE and Western blotting. The blots were developed with alkaline phosphatase or horseradish peroxidase-conjugated anti-mouse or anti-rabbit Fab antibodies (Jackson Laboratories).
Immunoprecipitation with Anti-HA Antibodies-For immunoprecipitation, protein A-Sepharose was mixed with HA antibodies at 37°C for 1 h, after which the beads were washed twice with PBS and twice in buffer H. Cell lysates (300 g) were added to the beads and swirled end to end at 4°C for 2 h. The immunocomplexes were washed once with radioimmune precipitation buffer, washed twice with 0.5 M LiCl in 100 mM Tris, pH 8.0, and finally washed with buffer A.
Ras Activation Assay-Cells were stimulated and washed as described above. Ras activation was assayed as described previously (30).
In brief, cells were lysed in Ral buffer, and lysates (300 g of protein) were incubated with 20 g of GST-Raf (RBD) and were washed three times in a buffer containing 20 mM Hepes, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, and 10% glycerol. The amount of Ras pulled-down was then assessed by Western blotting using anti-Ras antibody. To study Ras activity in transfected cells, a construct of GFP-Ras was used. After GnRH stimulation, the cells were lysed in Ral buffer and 300 g of protein was subjected for further treatment. The active, GTP-bound form of Ras was precipitated by the GST-Raf (RBD, 20 g) and washed as above, and the activated GFP-Ras was then detected with anti-GFP antibody.
Src Activity-Cell lysates (400 -500 g of protein in homogenization buffer containing 1% Triton X-100) were incubated with anti-c-Srcantibodies precoupled to protein A-Sepharose and swirled end to end at 4°C. The immunocomplexes were washed once with radioimmune precipitation buffer, twice with 0.5 M LiCl in 0.1 M Tris-HCl, pH 8.0, and once with buffer A. The washed immunoprecipitate were resuspended in a kinase assay buffer (28), and the c-Src activity was determined using acid-denatured enolase (3 M) as substrate in the presence of 20 M [␥-32 P]ATP (8000 cpm/pmol). The enzymatic reactions were terminated by the addition of sample buffer. The samples were then subjected to SDS-PAGE and autoradiography.
FAK Activity-Cell lysates (400 -500 g of protein in buffer H ϩ 1% triton X-100) were incubated with anti-FAK antibody coupled to protein A-Sepharose and swirled end to end at 4°C for 2 h. The immunoprecipitates were washed as above and mixed with sample buffer. Active FAK was determined by Western blot analysis using anti-phosphotyrosine antibody.
Raf-1 Activity-The activity of Raf-1 was determined by immunoprecipitation using an anti-C-terminal Raf-1 antibody and a subsequent in vitro kinase assay with recombinant MEK. This was performed in the same reaction mixture described for MEK except that 2 g of recombinant MEK1 were used instead of ERK in each reaction. The reactions were terminated by the addition of sample buffer, and the samples were subjected to SDS-PAGE analysis. The gels were blotted onto nitrocellulose membrane, and the phosphorylation of MEK was assessed by x-ray autoradiography.
MEK Activity-The activity of transfected GFP-tagged MEK was assessed by immunoprecipitation with anti-GFP antibody followed by a subsequent in vitro kinase reaction. In brief, ␣T3-1 cells transfected with GFP-MEK were stimulated with GnRH-a and harvested in buffer H. The cell lysates were incubated with 5 g/assay of GFP antibody precoupled to protein G-Agarose and were washed as described for immunoprecipitation. Immunocomplex kinase reaction was carried out in a reaction mixture containing 1 g of recombinant ERK, 10 M MgCl 2 , 1.5 M DTT, 75 mM ␤-glycerophosphate, pH 7.3, 0.075 M sodium vanadate, 3 M PKI peptide, 1.25 mM EGTA, 10 M calmidazolium, and 20 M [␥-32 P]ATP (300 cpm/pmol) for 20 min at 30°C. The reactions were terminated by the addition of sample buffer, and the samples were subjected to SDS-PAGE analysis. The gels were blotted onto nitrocellulose membrane, and the phosphorylation of ERK was assessed by x-ray autoradiography.
PKC Activity-The activity of transfected ␦-epitope-tagged PKC⑀ was assessed using cellular fractionation. In brief, transfected ␣T3-1 cells were stimulated with GnRH-a, homogenized in buffer H, and centrifuged at 15,000 ϫ g. Pellets containing plasma membranes were washed twice in buffer H and suspended in buffer H containing Triton X-100. Translocated PKC in the membranes was determined by Western blot analysis using antibody to the PKC␦-epitope.

Role of PTKs in ERK Activation by GnRH-Various
MAPK cascades (ERK, JNK, and BMK) are activated in response to GnRH stimulation of ␣T3-1 cells. We have previously shown that the stimulation of JNK activity by GnRH is mediated by a unique pathway, which includes sequential activation of PKC, Src, CDC42, and probably also MEKK1 (28). PKC was implicated also in the activation of ERK by GnRH (25), but the other components involved in this pathway remained unclear. In this study we used anti-doubly phosphorylated ERK (DP-ERK) antibody to detect its phosphorylation and activation upon GnRH-a treatment of ␣T3-1 cells. ERK phosphorylation was detected 5 min after GnRH-a treatment (Fig. 1A), peaked at 15 min, and was slightly reduced 15 min later. No change was detected in the total amount of ERK as judged by the equal staining with anti-general ERK antibody (Fig. 1A). These results appear similar to the trend of ERK activation by GnRH (Ref. 25 and data not shown), indicating that the anti-DP-ERK antibody can serve as a tool to study ERK activation in ␣T3-1 cells.
To study the mechanism of ERK activation we first used inhibitors of various intracellular signaling cascades. As previously reported (25), the PTK inhibitor genistein partially (40%) inhibited, and the PKC inhibitor GF109203X abolished the GnRH-induced ERK activation. The MEK inhibitor PD98059 inhibited not only GnRH-stimulated ERK-activity but abolished also its basal activity, whereas the PI3K inhibitor wortmannin and the p38 MAPK inhibitor SB203580 had no significant effect on ERK activation. The results suggest that GnRH signaling toward ERK is mainly transmitted via PKC and to a lesser extent also by PTKs. Interestingly, the moderate effect of genistein on ERK activation was achieved under conditions where it completely abolished the GnRH activation of JNK (28). One possibility for this differential effect is that the PTK involved in the GnRH to ERK pathway is distinct from Src, which is the PTK operating in the GnRH-JNK pathway (28), and this distinct PTK is only mildly sensitive to genistein. To test this possibility, we examined the effects of additional PTK inhibitors on ERK activation by GnRH-a. We found (Fig. 1B) that the specific inhibitor of Src, PP1, which abolished endogenous src activation (data not shown), inhibited the GnRHinduced ERK activation to a similar extent as genistein. Similarly, the general PTK inhibitor AG18 had a small effect (ϳ25%) on the GnRH-induced ERK activation. However, the EGF receptor inhibitor AG1478, which abolished the EGFinduced ERK activation in ␣T3-1 cells (data not shown), had no effect on ERK activation by GnRH-a. The data suggest that Src is partially involved in the activation of ERK by GnRH but probably not via transactivation of EGF or other growth factor receptors as suggested for other GPCRs (31,32).
Src, but Not FAK or EGF Receptor, Plays a Role in GnRH to ERK Signaling-Signaling by Src is often mediated via the focal adhesion kinase (FAK), which is usually instrumental in integrin signaling (33). We found that Src and FAK are activated in response to GnRH in ␣T3-1 cells, although the onset of Src activation appeared before that of FAK (Refs. 25 and 28 and Fig. 2). A useful tool in the study of GnRH signaling in ␣T3-1 cells is the co-overexpression of either constitutively active or dominant negative forms of upstream components together with a tagged form of the examined kinase (28). Therefore, we overexpressed hemagglutinin epitope-tagged ERK (HA-ERK2) in ␣T3-1 cells, stimulated the cells with GnRH, lysed the cells, immunoprecipitated the HA-containing proteins with anti-HA antibodies, and blotted with anti-DP-ERK antibodies. ERK activation measured by this method was essentially identical to that found for the endogenous ERK, indicating that overexpression can also serve as a useful tool in the study of ERK activation by GnRH. To study the role of Src/FAK in the GnRH to ERK signaling we co-overexpressed a dominant negative form of FAK (FRNK (34)) with the HA-ERK2 in ␣T3-1 cells. This FAK construct, which inhibits the activity of the endogenous FAK in ␣T3-1 cells (data not shown), had no effect on GnRH-induced ERK activation (Fig. 2B). This lack of effect, together with the late onset of FAK activation by GnRH-a stimulation, indicates that FAK is not involved in GnRH to ERK signaling. On the other hand, the C-terminal Src kinase (CSK), which acts as a dominant interfering mutant of Src (28,35) had a partial inhibitory effect on the GnRH to ERK pathway. This is demonstrated by a ϳ35% inhibition in the activation of ERK by GnRH-a caused by overexpression of CSK in ␣T3-1 cells (Fig. 2B). This result, together with the ϳ40% inhibition cause by PP1, strongly suggests that Src is partially involved in the activation of ERK by GnRH in ␣T3-1 cells.
It was recently reported that EGF receptor might be involved in GnRH signaling (36). However, our initial results indicated that there is no inhibition of GnRH-induced ERK activation by the inhibitor of EGF receptor AG1478 (Fig. 1B). To further confirm the lack of participation of EGF receptor in GnRH-ERK signaling, we cotransfected the dominant negative mutant of EGF receptor (K721A) with HA-ERK2 into ␣T3-1 cells. The cells were then serum-starved and activated with either EGF or GnRH-a followed by the determination of ERK activation. The dominant negative EGF receptor had no influence on ERK activation by GnRH-a (Fig. 3A) under conditions where it prevented ERK activation by EGF (Fig. 3B). These results, together with the lack of effect of AG1478 strongly suggest that EGF receptor is not involved in the activation of ERK by GnRH.
Activation of Raf-1 and Ras by GnRH-To understand the nature of the partial involvement of Src in the GnRH to ERK pathway, we studied the effect of GnRH on the upstream components of the ERK cascade, including the protein serine/threonine kinase Raf-1. After treatment of ␣T3-1 cells with GnRH-a and inhibitors, the cells were lysed, Raf-1 was immunoprecipitated, and its activity toward recombinant MEK was measured as described (37). Raf-1 activity was stimulated within 5 min after GnRH activation (data not shown); its activity peaked 10 min after stimulation and declined thereafter (Fig. 4). Similar to the inhibition of ERK activation by GnRH, we found that GF109203X completely inhibited Raf-1 activation by GnRH-a, whereas PP1 and genistein had a partial inhibitory effect (ϳ30% inhibition) and AG1478 had no significant influence on Raf-1 activation by GnRH-a. The small GTP-binding protein Ras, was also transiently activated by GnRH-a. Using a pulldown and a GTP loading assays, we found that activation of Ras was detected within 2 min from activation, peaked at 5-10 min, and declined thereafter (Fig. 5 and data not shown). However, the mechanism involved in this activation seems to be distinct from that of Raf-1 and ERK as judged from the differential sensitivity to the various inhibitors used. Thus, the Src inhibitors genistein and PP1 abolished the activation of Ras by GnRH-a, GF109203X had only a partial effect, and the EGF receptor inhibitor AG1478 had no effect upon Ras activation. These data indicate that, although both Ras and Raf-1 are activated in response to GnRH in ␣T3-1 cells, the upstream mechanism that leads to this activation is different. Indeed, when a dominant negative form of Ras (N-17 Ras) was transfected into the ␣T3-1 cells, it only partially (30%) inhibited ERK activation by GnRH-a (Fig. 6), whereas a constitutively active form of Ras (L61-Ras) caused a large elevation of ERK (18-to 22-fold above basal level). Therefore, although Ras is capable of activating the Raf-1/ERK pathway, GnRH activation of Raf-1/ERK is only partially Ras-dependent. The most plausible explanation for these data is that the main pathway operates via direct activation of Raf-1 by PKC (38), and that this activation requires only a minor contribution of activated Ras as previously suggested for the activation of Raf-1 by 12-O-tetradecanoylphorbol-13-acetate (39).
The Involvement of Dynamin, but Not ␤-Arrestin or G␤␥, in ERK Activation by GnRH-Recently, the ␤␥ subunits of Gproteins as well as ␤-arrestin and dynamin have been implicated in the G␣-independent GPCR-ERK signaling (reviewed in Ref. 12). For example, it was shown that G i -coupled receptor stimulation of mitogen-activated protein kinase is mediated by G␤␥-induced activation of Ras (40). To study the possible role of the ␤␥ subunit in GnRH to ERK signaling, we used a chimera of CD8 (which allows anchoring to the membrane) fused to the C terminus of ␤-adrenergic receptor kinase (ARK-C), which contains a G␤␥ binding domain (CD8-ARK-C). It has been previously shown that this chimera acts as a scavenger of the ␤␥ dimer (41,42). Although this construct was able to inhibit GPCR signaling toward ERK in COS7 cells, its overexpression had no significant effect on either the basal or the GnRHinduced activation of ERK in ␣T3-1 cells (Fig. 7), indicating that the signaling from GnRHR to ERK utilizes a G␤␥-independent pathway. Another protein that was implicated in the signal transmission of GPCRs is ␤-arrestin, which acts as a mediator of receptor internalization (43). Recently, it was also shown that ␤-arrestin can act as a scaffold protein and transmit the signals of Gq-coupled receptors toward ERK by forming a complex that contains internalized receptor, Raf-1, and activated ERK (44). We examined the possible involvement of ␤-arrestin using either an inactive form of this protein (V54D-␤-arrestin2), which inhibits the activity of all endogenous ␤-arrestins (43), or by overexpressing wild-type ␤-arrestin2, which should increase ERK activation by GPCRs (16, 45). Thus, we coexpressed these two constructs in ␣T3-1 cells together with HA-ERK2 and followed HA-ERK activation using anti-phospho ERK antibodies. As seen in Fig. 7, neither form of ␤-arrestin had any effect on ERK activation by GnRH. Both the G␤␥ subunits and ␤-arrestin do not seem to participate in the process of ERK activation by GnRH, although they can influence GPCR signaling in other systems.
As well as ␤-arrestin, dynamin seems to be a key regulator of the internalization processes of GPCR (46). As such, dynamin has also been implicated in GPCR signaling, including the activation of MAPKs by several GPCRs (16). The role of this protein in the ␣T3-1 cells was examined, as described for ␤-arrestin, by co-overexpression of either wild-type or a dominant negative form of dynamin (K44A-dynamin (16,47)) together with HA-ERK2. Although the wild-type form of dynamin had no significant influence on the activation of ERK by GnRH, the dominant negative form of dynamin partially inhibited (45%) both the basal and the GnRH-induced ERK activation (Fig. 8).
This inhibition of both basal and GnRH-stimulated activities is similar to the results obtained with dominant negative Ras (Fig. 6) and to some extent also to the results with CSK, suggesting that dynamin, Src, and Ras operate on the same signaling pathway.
Dynamin, but Not ␤-Arrestin, Is Involved in Ras Activation by GnRH-To confirm that ␤-arrestin is not involved in GnRHinduced stimulation of ERK and to further study the role of  dynamin in this process, we undertook to explore the influence on other components of the GnRH to ERK signaling pathway. First we examined whether ␤-arrestin has any effect on the GnRH activation of Ras. Thus, either wild-type or dominant negative forms of ␤-arrestin2 were cotransfected into ␣T3-1 cells together with GFP-Ras. After GnRH-a stimulation, the active, GTP-bound form of Ras was precipitated by the Ras binding domain of Raf-1 (Raf-RBD), and activated GFP-Ras was then detected with anti-GFP antibody. Similar to the effect on ERK2, no effect of the ␤-arrestin constructs on Ras activation could be detected under the conditions used (Fig. 9). Therefore, as was predicted by the lack of effect on ERK, ␤-arrestin does not seem to play a role also in the GnRH-induced signaling toward Ras. This observation is in agreement with the lack of ␤-arrestin involvement in GnRH-induced internalization of the GnRHR (48), suggesting that ␤-arrestin does not play a significant role in GnRH signaling.
Unlike ␤-arrestin, the other internalization mediator, dynamin appears to be involved in GnRH-induced GnRHR internalization (48) and, as demonstrated above (Fig. 8), in the activation of ERK by GnRH. We then undertook to elucidate the possible mechanism by which dynamin transmits the GnRH signals toward the downstream components of the ERK cascade. First, the possible involvement of dynamin in Ras activation by GnRH was examined by cotransfecting either wild-type or dominant negative (K44A) forms of dynamin together with GFP-Ras. Similar to the effect on ERK, the wildtype dynamin had no influence on Ras activity under the conditions examined. However, the dominant negative form of dynamin completely abrogated the GnRH activation of Ras under the conditions examined, indicating that dynamin lies upstream of Ras in the pathway that leads from the GnRHR. Similar to ERK, MEK activity was only partially inhibited by the dominant negative and to some extent also by wild-type dynamin (Fig. 9B). The fact that the inhibition of MEK activation by dominant negative dynamin was very similar to the inhibition of ERK activity makes it unlikely that dynamin  together with GFP-Ras, GFP-MEK, or epitope tagged-PKC⑀. Two days after transfection, the cells were serum-starved for 16 h and then either treated with GnRH-a (10 Ϫ7 M for 10 min for Ras and PKC and 4 min for MEK) or left untreated. Activated GFP-Ras was determined using a Western blot with anti-GFP antibody following Ras precipitation using GST-RBD as described under "Materials and Methods." Activity of transfected GFP-MEK was measured by immunoprecipitation with anti-GFP antibody and subsequent in vitro kinase reaction using recombinant ERK as a substrate (see "Materials and Methods"). To determine PKC translocation to the membrane, membranal fractions were collected and membranal (active) PKC was determined by Western blot analysis using antibody to the PKC epitope tag. The results in the bar graph represents the average of three experiments and represent the percentage of -fold activation of GnRH-stimulated enzyme in cells transfected with dominant negative or wild type dynamin as compared with activation (-fold) in cells transfected without dynamin constructs. influences the MEK-ERK level of the cascade as previously suggested in other systems (49). Moreover, unlike the inhibition of Ras, ERK, and MEK activities by dominant negative dynamin, there was no influence of this construct on the GnRH-induced membranal translocation of PKC⑀ (Fig. 9B), indicating that dynamin acts independently of this PKC. Because PKC⑀ is one of the main PKC isoforms that participate in GnRH signaling in ␣T3-1 cells (Ref. 50 and data not shown), the role of dynamin in the activation of Ras seems to be confined to its influence on the Src/Ras step without any significant influence on Raf-1 stimulation by PKC (Fig. 10). DISCUSSION GPCRs comprise the largest group of integral membranal proteins, which transmit their signals primarily via four groups of G-proteins, G s , G i , G q , and G 12 and also via receptorinteracting molecules and transactivation of growth factor receptors. The G-proteins and the other components are capable of transmitting signals from the receptor to MAPK cascades via distinct pathways, which often form an elaborate network of signaling cascades (12). However, the exact mechanisms, which are involved in each of these pathways, are still mostly obscure. Here we studied the receptor for GnRH, which was shown to operate via G q or G 11 in pituitary cells (51). This system is of particular interest because, in pituitary cells, GnRH does not promote growth and differentiation and does not constitute a stress signal and therefore might utilize a unique signaling pathway to activate the MAPK cascades (12,20).
In previous studies we and others have shown that the ERK (21,25), JNK (28), and p38MAPK (27) cascade are activated by GnRH in a PKC-dependent manner. However, although JNK activation seems to be fully dependent on Src and only partially (ϳ70%) on PKC, ERK activation was shown to be fully dependent on PKC but only partially (ϳ30%) on PTK. Because PTKs have been shown to play an important role in the activation of ERK by GPCR (12), the small effect of PTK inhibitors observed in our studies was surprising. We have ruled out the possibility that the PTK involved in the GnRH to ERK activation is only partially sensitive to genistein. This was demonstrated by showing that none of the general PTK inhibitors used here inhibit ERK activation by more than 40%. Moreover, the Src inhibitor PP1 and the inhibitory kinase CSK, which fully inhibit Src activity, also cause similar inhibitions. Because Src is activated by GnRH and has a major role in the transmission of signals to the JNK cascade, the most likely explanation for our results is that the activation of ERK is supported to some extent by Src, although the main pathway leading to ERK activation is PTK-independent.
PTK-induced activation of ERK is known to involve the small GTP-binding protein Ras. Interestingly, although GnRH causes a significant activation of Ras, a dominant negative form of Ras had only a minor inhibitory effect on ERK activation. Because the effect was similar to that exerted by Src, it is likely that Ras acts downstream of Src in a pathway that partially supports the activation of ERK (Fig. 10). Indeed, Ras activation by GnRH was abolished by the Src inhibitor, PP1. Interestingly, it has been previously demonstrated (39) that stimulation of PKC in COS-7 cells led to activation of Ras and formation of Ras⅐Raf-1 complexes, but the activation of Raf-1 by PKC was not completely blocked by dominant negative Ras. These data indicate that PKC activates Raf-1 by a mechanism distinct from that initiated by PTKs and that only a small amount of activated Ras is needed to allow the activation of Raf-1 by PKC. The fact that ERK activation is only partially dependent on Src but is fully dependent on PKC agrees well with this suggested model. Thus, we believe that, upon GnRH stimulation, Raf-1 is mainly activated by PKC but requires a small amount of active Ras. This possibility was demonstrated before in other experimental systems (38,39) and explains the partial inhibition of ERK by interfering mutants of Src and Ras (Fig. 10). It has previously been shown that calcium may be involved in the activation of ERK by GnRH in ␣T3-1 cells (25,52). According to our model this effect is probably on the Src/ dynamin/Ras branch of the pathway and not on the PKC-Raf-1 step, which was shown to be independent of calcium influx (52) and not on the Raf-1-MEK-ERK steps, which were shown to be calcium-independent in several cellular systems (1).
It should be noted that a recent study claimed that the activation of ERK by GnRH in ␣T3-1 cells is inhibited equally by PKC and PTK inhibitors and that Ras activation by GnRH is mediated by EGF receptor (36). The reason for these slight discrepancies between our results and the results reported by Grosse et al. (36) is not clear. However, one possible explanation could be the different lengths of serum starvation, which modify the content of many signaling components (53). We found that the minimal time needed for a complete removal of MAP kinase phosphatases and complete quiescence in ␣T3-1 cells is 14 h (data not shown). Alternatively, it is also possible that the ␣T3-1 cells are modified under different growing conditions to form distinct subpopulations that exhibit different repertoire of signaling molecules as was shown also in other cell lines such as PC12. These explanations can be used also for the different results obtained in regard to the activation of JNK by GnRH in different laboratories (54).
Another point of interest is the mechanism of Src activation by GnRH. We have previously shown that Src is activated by GnRH via a mechanism that is partially (ϳ70%) dependent on PKC (28). Similarly, we found that the activation of Ras is only partially dependent on PKC, but fully dependent on Src, supporting the notion that Ras is activated by Src. Such pathway of sequential activation of Src-Ras-ERK was reported also for some other GPCRs (12,(55)(56)(57). However, the fact that Src and Ras are only partially dependent on PKC raises the question as to what might be the other pathway involved in the activation of Src/Ras. An answer to this question may come from the recent observations that, upon GPCR stimulation, Src can be activated in a G␣-independent manner and therefore we un- dertook to study the role of these additional signaling molecules as outlined below. Many GPCRs were shown to transmit their signal through transactivation of either EGF receptor-or cytoskeleton-associated PTKs (FAK and PYK). Our results indicate that those components are not involved in GnRH signaling to ERK. Thus, EGF receptor does not seem to be activated in response to GnRH (25), and the specific inhibitor of EGF receptor (AG1478) or dominant negative form of the EGF receptor had no effect on GnRH-induced ERK stimulation. Moreover, FAK does not seem to participate in GnRH to ERK signaling, because a dominant negative FAK had no effect on ERK activation by GnRH. Finally, PYK does not seem to be expressed to any detectable level in ␣T3-1 cells as judged by immunoblotting and immunoprecipitation experiments (data not shown), and therefore is unlikely to participate in the GnRH to ERK pathway.
Although the G␣ subunits are important transducers of GPCR signaling, dissociated ␤␥ subunits have been implicated in the transmission of GPCRs signaling as well. Thus, ␤␥ dimers can act via PTKs (such as Src), via a direct activation of Ras or via a direct activation of either the protein serine/ threonine kinase KSR-1 or activation of phosphatidylinositol 3-kinase (58,59). In addition, GPCRs can operate via ␤-arrestin-and dynamin-mediated internalization (16), and ␤-arrestin may serve as a scaffold for additional signaling molecules and initiate a second wave of G-protein-independent, heptahelical receptor-mediated signals that activate the MAPK cascades (12). Interestingly, in the ␣T3-1 system, we found that neither a scavenger of ␤␥ dimer nor the dominant negative form of ␤-arrestin affect the GnRH-induced ERK activation. However, overexpression of dominant negative dynamin reduced the activation of both basal and GnRH-induced ERK activation indicating that dynamin, unlike the other upstream components examined, may participate in the PKC-independent activation of Src/Ras.
Recently it was shown that, in addition to its role in the internalization of GPCRs, dynamin is also necessary for the direct activation of ERK by MEK (49). This may indicate that the inhibition by the dominant negative dynamin is downstream of Raf-1 and not upstream of the Src/Ras pathway. To determine the site of dynamin action, we examined its role on several components of the cascade. Interestingly, we found that GnRH-induced Ras activity was significantly inhibited (ϳ80%) by the dominant negative dynamin, whereas PKC activity was not affected under the same conditions. Because Src activation is mostly dependent on PKC (ϳ70%), inhibition by the dominant negative dynamin upstream of Src at the PKC-independent pathway should have resulted in only ϳ30% inhibition of Ras activation, which is much smaller than the complete inhibition obtained. Moreover, the inhibition of the GnRH-induced MEK activation was similar to that of ERK activation. Therefore, we suggest that the dynamin is required for the process of Ras activation by Src, and the step of Raf-1 activation by PKC is probably not affected by the dominant negative dynamin under the conditions used.
In summary, we studied here the mechanism of ERK activation by GnRH in the pituitary derived ␣T3-1 cell line. We show that ERK activation is fully dependent on PKC but only partially dependent on Src, Ras, and dynamin. Because it has previously been shown that Raf-1 activation by PKC is only partially dependent on active Ras, our results are best explained by the involvement of two distinct pathways in the GnRH-mediated stimulation of Raf-1/ERK. One of these pathways involves a direct activation of Raf-1 by PKC; the other involves Ras activation by Src and dynamin.