Evidence for a bifurcation of the mitogenic signaling pathway activated by Ras and phosphatidylcholine-hydrolyzing phospholipase C.

NIH 3T3 cells stably transfected with the gene encoding phosphatidylcholine-hydrolyzing phospholipase C (PC-PLC) from Bacillus cereus display a chronic elevation of intracellular diacylglycerol levels and a transformed phenotype. We have used such PC-PLC-transformed cells to evaluate the roles of the cytoplasmic serine/threonine kinases Raf-1, protein kinase C (PKC) and protein kinase A (PKA) in oncogenesis and mitogenic signal transduction elicited by phosphatidylcholine hydrolysis. We demonstrate here that stable expression of dominant negative mutants of both PKC and Raf-1 lead to reversion of PC-PLC-transformed cells. Interestingly, expression of kinase defective PKC also reverted NIH 3T3 cells transformed by the v-Ha-ras oncogene. Activation of PKA in response to elevation of cAMP levels also lead to reversion of PC-PLC-induced transformation, implicating PKA as a negative regulator acting downstream of PC-PLC. On the other hand, inhibition or depletion of phorbol ester responsive PKCs attenuated but did not block the ability of PC-PLC-transformed cells to induce DNA synthesis in the absence of growth factors. These results clearly implicate both Raf-1 and PKC as necessary downstream components for transduction of the mitogenic/oncogenic signal generated by PLC-mediated hydrolysis of phosphatidylcholine and suggest, together with other recent evidence, a bifurcation in the signaling pathway downstream of PC-PLC.

NIH 3T3 cells stably transfected with the gene encoding phosphatidylcholine-hydrolyzing phospholipase C (PC-PLC) from Bacillus cereus display a chronic elevation of intracellular diacylglycerol levels and a transformed phenotype. We have used such PC-PLC-transformed cells to evaluate the roles of the cytoplasmic serine/threonine kinases Raf-1, protein kinase C (PKC) and protein kinase A (PKA) in oncogenesis and mitogenic signal transduction elicited by phosphatidylcholine hydrolysis. We demonstrate here that stable expression of dominant negative mutants of both PKC and Raf-1 lead to reversion of PC-PLC-transformed cells. Interestingly, expression of kinase defective PKC also reverted NIH 3T3 cells transformed by the v-Ha-ras oncogene. Activation of PKA in response to elevation of cAMP levels also lead to reversion of PC-PLC-induced transformation, implicating PKA as a negative regulator acting downstream of PC-PLC. On the other hand, inhibition or depletion of phorbol ester responsive PKCs attenuated but did not block the ability of PC-PLCtransformed cells to induce DNA synthesis in the absence of growth factors. These results clearly implicate both Raf-1 and PKC as necessary downstream components for transduction of the mitogenic/oncogenic signal generated by PLC-mediated hydrolysis of phosphatidylcholine and suggest, together with other recent evidence, a bifurcation in the signaling pathway downstream of PC-PLC.
Recently, evidence for a crucial role of phospholipase Cmediated hydrolysis of phosphatidylcholine (PC) 1 in mitogenic signaling in different mammalian cells and in the maturation of Xenopus oocytes has accumulated (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). In fact, exogenous addition of a phosphatidylcholine-hydrolyzing phospholipase C (PC-PLC) from Bacillus cereus is able to mimic both a significant portion of the mitogenic response to PDGF in Swiss 3T3 fibroblasts and the constitutive activation of protein kinase C (PKC) in v-ras-or v-src-transformed NIH 3T3 cells (5,12). Furthermore, constitutive expression of the gene (plc) encoding B. cereus PC-PLC leads to a chronic elevated level of intracellular DAG and oncogenic transformation of NIH 3T3 cells (13). Both in Xenopus oocytes and in murine fibroblasts, a variety of experimental approaches have revealed that PLC-mediated hydrolysis of PC, elicited either by the endogenous activity or by the exogenous addition of the bacterial enzyme, is located downstream of Ras (3,8,12,14,15). It has also been shown that PC-PLC may be involved in coupling Ras to activation of Raf (16). However, the mechanism whereby PC-PLC transduces mitogenic signals conveyed by p21 Ras remains to be elucidated. But, since PC-PLC generates the second messenger diacylglycerol (DAG) capable of activating PKC isozymes (17) possible downstream targets may include one or more specific PKC isozymes (12,18,19) or perhaps other hitherto undetected DAG-regulated kinases. A direct activation of Raf-1 by PCderived DAG may also be possible since the enzyme contains a cysteine finger in the regulatory domain similar to the DAG binding motifs of PKC and DAG kinase (20). The role of the atypical PKC subtype in mitogenic signal transduction is particularly interesting since it resembles Raf-1 both in terms of structural organization, insensitivity to Ca 2ϩ and phorbol esters, and a ubiquitous expression pattern (21,22). Furthermore, a requirement for the PKC isotype in the Ras-mediated maturation pathway in Xenopus oocytes and for serum-activated DNA synthesis in mouse fibroblasts has been documented (18,19), suggesting an important role for PKC in mitogenic signaling. Also, PKC, but not Raf-1, is required for stimulation of the stromelysin promoter via a PDGF/Ras/PC-PLC-responsive element (23).
In the study presented here, we show that dominant negative mutants of both Raf-1 and PKC revert the transformed phenotype of cells stably transfected with the PC-PLC gene (13). Consistent with a downstream location of PC-PLC relative to Ras, we also found that v-Ha-ras-transformed cells acquired a normal, nontransformed phenotype following transfection with a dominant negative mutant of PKC. As recently demonstrated for both v-ras-and v-raf-transformed fibroblasts (24,25), PKA activation blocked both transformation, and the mitogenic signal elicited by chronic PLC-catalyzed hydrolysis of PC. Taken together, these results clearly show that PC-derived DAG acts upstream of Raf-1 and establish PKC as a novel downstream mediator of Ras/PC-PLC signaling.

MATERIALS AND METHODS
Plasmid Constructs-A PC-PLC expression vector allowing selection of hygromycin-resistant stable transfectants and expression of PC-PLC from the metallothionein IIa promoter was made by inserting a 1350base pair NheI-XhoI fragment from pOPLCmam (13) into NheI-XhoIcut pMEP4 (Invitrogen) generating pMT-PLC. For expression of a dominant negative mutant of Raf-1, an 817-base pair EcoRI-SalI fragment from p627 (ATCC 41050), encoding amino acids 1-257 of human c-Raf-1, was cloned into the KpnI and XhoI sites of pMT-hyg yielding pMTdnRaf. pMT-hyg was made by deleting a 4134-base pair EcoRV-StuI fragment containing the EBNA1-oriP region from pMEP4. In pMTdnRaf, the N-terminal regulatory domain of Raf-1 is expressed from the human metallothionein IIa promoter, and the vector also contains a hygromycin resistance gene for selection of stable transfectants. The expression vector for a kinase defective dominant negative mutant of PKC (pRcCMV mut ) has been described previously (18,26). The CAT reporter plasmids for assaying AP-1-and NF-B-mediated transactivation were obtained from A. S. Kekulé and P. H. Hofschneider and contain three copies of a consensus AP-1 binding site or two NF-B binding sites, respectively, inserted upstream of a minimal herpes simplex virus thymidine kinase promoter (27).
Cell Culture and Generation of Stably Transfected Cell Lines-NIH 3T3 fibroblasts (passage 123) were purchased from the American Type Culture Collection (ATCC CRL 1658) and cultured in Dulbecco's modified Eagle's medium supplemented with 10% calf serum (HyClone, Logan, Utah), penicillin (100 units/ml), and streptomycin (100 g/ml) (Life Technologies, Inc.) in a CO 2 incubator (5% CO 2 ) at 37°C. NIH 3T3 cells transformed by the v-Ha-ras oncogene (4,12) were grown in the same medium. The M1 cell line representing NIH 3T3 cells transformed by the gene (plc) encoding PC-PLC from B. cereus has been described previously (13). Stably transfected cell lines were established as described previously (13). To establish v-Ha-ras-transformed NIH 3T3 cell lines expressing the dominant negative mutant of PKC, transfection was performed with 3.56 g of pRcCMV mut and 0.44 g of pMT-hyg (9:1 molar ratio) using Lipofectamine according to the supplier's (Life Technologies, Inc.) instructions, and G418-and hygromycin-resistant clones were isolated as described previously (13).
Soft Agar Cloning-To assay anchorage-independent growth, 10 3 cells were mixed into 1 ml of top agarose containing 0.35% SeaPlaqueagarose in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and seeded onto 2 ml of solidified bottom agarose (0.7% SeaPlaque-agarose plus medium and serum) in triplicate 35-mm-diameter wells. The appropriate selective drugs were included in both top and bottom agar. The top agar was replenished every 4 days, and the incubations were performed for 21-30 days. Photographs were taken after staining the colonies overnight in 0.5% p-iodonitrotetrazolium violet (Sigma). When used, forskolin (Sigma) or GF 109203X (Calbiochem) was included in both top and bottom agar at final concentrations of 40 or 0.5 M, respectively.
Determination of Intracellular Levels of DAG-The DAG mass levels of subconfluent, serum-starved (0.5% calf serum for 32 h) cell cultures were determined by using the Amersham DAG assay reagents system kit (RPN 200) as described previously (13).
Antibodies, Immunostaining, and Immunoblotting-Expression of PC-PLC was monitored by immunoperoxidase staining of methanolfixed cells using an affinity-purified antibody raised against B. cereus PC-PLC as described previously (13). For the detection of PKC by immunoblotting, a rabbit anti-Xenopus PKC antiserum was used. The C-terminally truncated dnRaf protein was detected with a rabbit polyclonal antibody (K-153) raised against an N-terminal peptide of human Raf-1 (Santa Cruz Biotechnology). Cell extracts were made by adding boiling 1% SDS in 10 mM Tris-HCl (pH 7.5) directly onto the cells, followed by boiling of the extracts for 5 min. Immunoblots were developed using chemiluminescence employing a goat anti-rabbit IgG alkaline phosphatase-conjugated secondary antibody (Santa Cruz Biotechnology) and Lumiphos (Boerhinger Mannheim) as the substrate. Molecular weights were estimated using a prestained low moleculear weight protein standard (Bio-Rad).
Analysis of Ras-bound GDP and GTP-The amount of GTP bound to Ras was measured essentially as described previously (28). After in vivo labeling for 4 h in P i -free Dulbecco's modified Eagle's medium supplemented with 100 Ci/ml 32 P i (9,000 Ci/mmol, DuPont NEN), Ras was immunoprecipitated from the lysates for 40 min at 4°C by adding 5 l of agarose-conjugated anti-v-Ha-Ras (Y13-259) antibody (Santa Cruz Biotechnology). After extensive washing, GTP/GDP bound to Ras was eluted at 68°C for 20 min in 10 l of 2 mM EDTA (pH 8.0), 2 mM dithiothreitol, 0.2% SDS, 0.5 mM GTP, 0.5 mM GDP. Eluted nucleotides were separated by thin layer chromatography using PEI-cellulose plates containing fluorescent indicator (Merck) with 1 M Na 2 PO 4 (pH 3.4) as the running phase. The plates were briefly prewashed in methanol and air dried. Migration of the cold carrier was visualized by UV light (254 nm), and radioactive nucleotides were quantitated using a PhosphorImager (Molecular Dynamics). The results were expressed as (GTP/(GDP ϫ3/2 ϩ GTP)) ϫ 100%.
Transient Transfections and CAT Assays-Subconfluent cultures of the different stably transfected cell lines, in 100-mm-diameter Petri dishes, were transfected with 10 g of reporter plasmid following the calcium phosphate coprecipitation method. Plasmids were purified using a Quiaprep Spin Plasmid kit (Quiagen, Germany). The precipitates were left on the cells for 4 h in medium containing 10% serum, after which the cultures were washed once in phosphate-buffered saline and glycerol-shocked for 90 s in 15% (v/v) glycerol in HBS buffer (137 mM NaCl, 5 mM D-glucose, 0.9 mM NaH 2 PO 4 , 21 mM HEPES (pH 7.08)). The cell cultures were washed 3 times with phosphate-buffered saline and serum-starved for 24 h in 0.5% serum. Preparation of cell extracts and CAT assays were modified from Seed and Shen (29) and Pothier et al. (30). The cells were scraped in phosphate-buffered saline, collected by a brief centrifugation, and lysed by three cycles of freezing and thawing in 15 mM Tris-HCl (pH 8.0), 60 mM KCl, 15 mM NaCl, 2 mM EDTA, 0.15 mM spermidine, 1 mM dithiothreitol, 0.4 mM phenylmethylsulfonyl fluoride. Unsoluble material was removed by centrifugation, and total protein concentrations in the cellular extracts were determined with the Bio-Rad protein assay using bovine serum albumin as a standard. Aliquots of 10 g of total protein were mixed with 1 l of [ 14 C]chloramphenicol (56 mCi/mmol, 25 Ci/ml, Amersham Corp.), 5 l of n-butyryl-CoA (5 mg/ml) and 0.25 M Tris-HCl (pH 8.0) in a total reaction volume of 125 l. Following incubation at 37°C for 2 h, the reaction was stopped by extraction with 300 l of mixed xylene isomers (Merck), the phases were thoroughly mixed by vortexing 30 s and separated by centrifugation (3 min at 13,000 rpm in a microcentrifuge), and 280 l of the uppermost, organic phase was removed. The organic phase was washed 2 times by the addition of 150 l of Tris-HCl (pH 8.0). Acetylated n-butyryl-CoA in 200 l of the final organic phase was quantitated by liquid scintillation counting.

The Oncogenic Potential of PC-PLC Is Similar to That of
Activated v-Ha-Ras-We have recently demonstrated that stable expression of the gene (plc) encoding PC-PLC from B. cereus is oncogenic to NIH 3T3 cells (13). In that study, G418-resistant cell lines expressing the plc gene either from the murine mammary tumor virus long terminal repeat (M clones) or the Rous sarcoma virus long terminal repeat (R clones) promoters were used. To increase our ability to conveniently select doubly transfected cell lines, we have now established NIH 3T3 cell lines transfected with pMT-PLC (see "Materials and Methods") containing the plc gene under the control of the human metallothionein IIa promoter and carrying a hygromycin resistance gene. This allows the introduction of other genes into the plc transformed cells using expression vectors containing either G418 or hygromycin resistance markers. When PC-PLC was expressed from the metallothionein IIa promoter in pMT-PLC, the cell lines isolated (P clones) acquired an even more transformed phenotype than our previously described plc-transformed cell lines (13). These new cell lines displayed a highly transformed phenotype, similar to v-Ha-ras-transformed cells, as evidenced by increased colony sizes in soft agar, a more potent induction of DNA synthesis in the absence of serum growth factors, and shorter doubling times than our previously reported plc expressing cell lines. This correlated with higher expression levels of PC-PLC as visualized by immunostaining of cells with an affinity-purified antibody against B. cereus PC-PLC and a higher level of the intracellular DAG mass than the M clones described before (see Figs. 2 and 3, and Table I for specific data). The fact that the P clones closely mimicked the behavior of v-ras-transformed NIH 3T3 cells may indicate that constitutive expression of PC-PLC leads to activation of critical downstream targets of Ras. Thus, the hygromycin-resistant P18 clone and the previously characterized G418-resistant M1 clone were selected for further studies aimed at identifying downstream effectors of the mitogenic signal elicited by chronic hydrolysis of PC.
The Transformed Phenotype of NIH 3T3 Cells Constitutively Expressing PC-PLC Is Reverted by Dominant Negative Mutants of Either Raf-1 or PKC-PC-PLC has been shown to be located downstream of Ras in the insulin-stimulated maturation pathway of Xenopus oocytes (3) and to release NIH 3T3 cells from the block to proliferation imposed by the dominant negative Ras N-17 mutant (14). Since Ras is known to function upstream of Raf-1 in the pathway leading to activation of mitogen-activated protein kinases and it has recently been suggested that activation of an endogenous PC-PLC activity couples Ras to activation of Raf (16), we asked whether Raf was required for plc-mediated transformation of NIH 3T3 fibroblasts. To answer this question, the M1 clone was transfected with a dominant negative Raf mutant (dnRaf), where only the N-terminal regulatory domain of c-Raf-1 (20, 31) was expressed from a vector containing the hygromycin resistance gene, and stable doubly transfected G418-and hygromycin-resistant cell lines were established. Expression of dominant negative mutants of Raf-1 have been shown to cause reversion of both ras-and src-transformed NIH 3T3 cells (32,33). PKC has been shown to be required both for maturation of Xenopus oocytes (19) and for the induction of DNA synthesis by serum in murine fibroblasts (18). To determine if this kinase also could be a necessary downstream component in the signaling pathway(s) triggered by PC-PLC action, we transfected P18 cells with a plasmid harboring a G418 resistance gene and encoding a kinase-defective mutant of PKC (dnPKC), previously shown to act in a dominant negative manner (18,26), and doubly transfected cell lines were isolated. The expression of dominant negative Raf-1 and PKC proteins was verified by immunoblotting (Fig. 1). The expression of dnRaf and dnPKC in M1 and P18, respectively, lead the cells to revert to a flat, contact-inhibited, nontransformed phenotype, even though they still expressed PC-PLC as demonstrated by immunostaining (Fig. 2). Strikingly, of 12 out of 12 M1-dnRaf clones and seven out of eight P18-dnPKC clones analyzed, all had lost the ability to reinitiate DNA synthesis following serum starvation (data not shown). One P18-dnPKC clone, which was both G418-and hygromycin-resistant but did not display any reversion of phenotype, did not express dnPKC clone (Fig. 1). As seen from Fig. 3, the ability of the plc clones (M1 and P18) to display anchorageindependent growth in soft agar was completely blocked by expression of either the dnRaf-or the dnPKC mutants. As mentioned above, the soft agar colony size of the P18 clone was significantly larger than the M1 clone. To ensure that the expressed PC-PLC in M1-dnRaf and P18-dnPKC, shown by immunostaining against PC-PLC (Fig. 2), had retained enzymatic activity, the DAG mass levels in these clones were determined and compared with their respective parental cell lines (Table I). Of note, the DAG levels were still elevated, and even increased from the parental plc-expressing cell lines. (Table I). The P18 clone displayed a significantly higher DAG mass level than M1, as compared with the parental NIH 3T3 cells. The P18 clone also showed a shorter doubling time and grew to a higher saturation density in 10% serum than M1. However, by coexpression of either dnPKC or dnRaf, the doubling times were increased, and the saturation densities were reduced to about the same values as for NIH 3T3 cells and vector control cell lines (Table I). As shown in Fig. 4A, the ability to induce DNA synthesis in the absence of added mitogens was reduced to background levels in the plc clones stably transfected with either of the dominant negative kinase mutants. As a control, stable expression of wild-type PKC in plc-transformed cells had no effect on induction of DNA synthesis or growth in soft agar (data not shown). By transient transfection assays using CAT reporter plasmids containing either binding sites for the transcription factors NF-B or AP-1 inserted upstream of a minimal herpes simplex virus thymidine kinase promoter (27), we found that in plc-transformed cells, the transactivation potential of both NF-B and AP-1 was induced in the absence of growth factors. However, following stable expression of dnPKC, these parameters were reduced to background levels (Fig. 4B). Taken together, all of these data strongly suggest that both Raf-1 and PKC are required for PC-PLC-induced transformation and that both kinases are located downstream of PLC-mediated hydrolysis of PC in the mitogenic signaling cascade.
A Dominant Negative Mutant of PKC Is Able to Revert v-ras-transformed Cells-Having found that a dominant negative mutant of PKC reverted plc-transformed cells, we next asked whether expression of this mutant also could cause reversion of cells transformed by v-ras. v-Ha-ras-transformed NIH 3T3 cells were cotransfected with the dnPKC expression vector and a plasmid carrying a hygromycin resistance gene (see "Materials and Methods"). After selection of stably transfected clones, the expression of dnPKC was demonstrated by immunoblotting (Fig. 5A). We found that expression of dnPKC in v-ras expressing cells led to reversion to a flat, contactinhibited, nontransformed phenotype, although these cells still contained activated Ras as verified by immunoprecipitation of Ras followed by analyses of GTP/GDP ratios in the precipitates (Fig. 5B). The cell doubling time was significantly increased (from 15 to 26 h), the saturation density decreased (2.4-fold), and the ability to induce DNA synthesis in the absence of growth factors or to form colonies in soft agar was completely abolished (see Fig. 5, C and D). In fact, of 15 independent clones analyzed, all had lost the potent ability of the parental v-rastransformed cells to induce DNA synthesis in the absence of serum (data not shown). As a separate control, we measured activation of the transcription factors AP-1 and NF-B by transient transfection assays with CAT reporter constructs in serum-starved v-ras cells expressing dnPKC compared with parental v-ras cells. As evident from Fig. 6, transactivation of the CAT gene via AP-1 or NF-B binding sites was completely abolished. These results are completely consistent with the notion that PKC is located downstream of Ras and PC-PLC and serves as a necessary component in Ras-mediated mitogenic signaling and transformation.
The Induction of DNA Synthesis in PLC-transformed Cells Is Largely Independent of Phorbol Ester-responsive PKC-PC-PLC action generates DAG, a potent activator of both classical and novel PKC subtypes (17). In NIH 3T3 cells, only the ␣ subtype of classical PKCs, the ␦ and ⑀ subtypes of novel PKCs, and the atypical PKC subtype are expressed. Other PKC isoforms are not expressed at all or at very low levels (34,35). PKC is not activated by either phorbol esters or short chain DAGs (22,36). The ␣PKC has been shown to phosphorylate Raf-1 both in vitro and in vivo (37,38), but this does not stimulate the activity of Raf-1 toward its natural substrate mitogen-activated protein kinase/extracellular signal-regu-lated kinase kinase (39). To investigate the possible contribution of DAG-responsive PKC to the mitogenic signal elicited by PC-PLC, we measured the induction of DNA synthesis in the presence and absence of the novel PKC inhibitor GF 109203X and following depletion of PKC by long term treatment with the phorbol ester TPA (5). GF 109203X is reported to be completely nontoxic and acts as a highly selective PKC inhibitor able to block both classical PKC subtypes and the novel ␦ and ⑀ subtypes but not PKC (19, 40 -42). As seen from Fig. 7A, blockade of phorbol ester-responsive PKC (␣PKC), by either down-regulation or direct inhibition, attenuated somewhat the magnitude of the mitogenic response (25-35% inhibition) of v-ras and plc-transformed cells, indicating that PKC may contribute but that it is by no means essential for mitogenesis. For PDGF-stimulated cells, TPA down-regulation and direct inhibition with GF 109203X produced opposite effects, with downregulation being somewhat stimulatory and GF 109203X giving 25% inhibition of DNA synthesis. This may reflect the fact that GF 109203X is a less potent inhibitor of the growth inhibitory ␦-isoform (41) than ␣and ⑀PKC (40), whereas long term treatment with TPA will down-regulate all of these three PKC isoforms (5). Together, our results indicate that both PKC-dependent and -independent mechanisms are involved but that phorbol ester-responsive PKCs are not necessary for the induction of DNA synthesis in NIH 3T3 cells elicited by PDGF, activated Ras, or overexpression of PC-PLC. This conclusion is also supported by the fact that long term treatment of v-Haras-or plc-transformed cells with GF 109203X at concentrations up to 3 M did not change the transformed phenotype of these cells or abolish their ability to form colonies in soft agar. In fact, the number and sizes of colonies in soft agar showed a similar slight reduction of about 25% as described above for the effect on DNA synthesis (data not shown).

Protein Kinase A Is a Negative Regulator of Mitogenesis Functioning Downstream of PC-PLC-A number of research
groups have recently demonstrated that activation of PKA by elevation of cAMP levels blocks the Ras-dependent activation of the mitogen-activated protein kinase pathway (28,43,44). Treatments with the adenylate cyclase-stimulating agent for-skolin or cAMP analogues such as 8-bromo-cAMP cause reversion of both Ras-and Raf-transformed cells (24,25). The inhibition by PKA has been shown to be due to direct phosphorylation of Raf (24,44). On this background and as a further means to investigate the location of PC-PLC action in this pathway, we assayed the effect of forskolin and 8-bromo-cAMP on the induction of DNA synthesis following serum starvation of cells transformed by v-Ha-ras or plc (P18). As shown in Fig. 7B, both forskolin and 8-bromo-cAMP strongly inhibited the mitogenic response in all of these cell lines as well as the PDGF response of parent NIH 3T3 cells. As a control, treatment with 8-bromo-cGMP had no effect. Furthermore, forskolin  The data are expressed as -fold increase compared with control NIH 3T3 cells in more than three independent experiments performed in duplicate.
c Doubling time in 10% serum was determined by counting viable cells every day for 4 days. d Saturation density was the number of cells in culture 4 days after the cultures reached confluency. P18 was counted 2 days after confluency due to detachment and loss of viable cells by prolonged incubation. The M1 clone started to detach after 4 days, while the parental and reverted cell lines displayed contact inhibition with growth arrest at confluency. The data are expressed as number of cells ϫ 10 6 and represent three independent experiments performed in duplicate.
e Colony formation in soft agar was scored after 21-28 days.
(40 M) and 8-bromo-cAMP (0.5 mM) also caused reversion of the transformed morphology of ras-and plc-transformed cells and abolished the ability of plc-transformed cells to form colonies in soft agar (data not shown). Thus, our results confirm previous observations on the negative role played by PKA on the Ras-dependent activation of Raf and further demonstrate that PC-PLC is not able to bypass the block to mitogenesis and transformation imposed by PKA activation. This also corroborates the finding that PC-PLC is acting upstream of Raf, which is thought to be the target of the inhibitory action of PKA. DISCUSSION A number of recent reports have firmly established that PC hydrolysis is critically involved in growth factor-mediated mitogenic signal transduction (1-11, 14 -16). Hydrolysis of PC can be catalyzed either by PC-PLC or PC-PLD (2, 6). However, PC-PLC action seems to be responsible for the sustained increase in cellular DAG levels observed in fibroblasts upon growth factor stimulation or following transformation by v-ras or v-src oncogenes (5,10,14,15). Importantly, PC-PLC has been shown to act downstream of Ras and upstream of the serine/threonine kinase Raf-1 (3, 14 -16). Consistent with a critical role of PC-PLC in mitogenic signal transduction, we have recently shown that constitutive expression of the gene (plc) encoding B. cereus PC-PLC leads to transformation of NIH 3T3 cells and that the transformed phenotype is completely dependent on plc expression, resulting in a chronic increase of the cellular DAG mass (13). In the present report, we extend upon these findings and show that there seems to be a direct relationship between the expression level of the bacterial PC-PLC, the resulting DAG levels, and the extent of oncogenic transformation achieved as evaluated by parameters such as cell morphology, growth pattern, cell doubling times, size of soft agar colonies, and induction of DNA synthesis in the absence of growth factor addition. We also show that like v-ras-transformed cells, plc-transformed cells possess activated AP-1 and NF-B transcription factors following serum starvation. Thus, transformation by plc closely mimicks the behavior of v-Ha-rastransformed NIH 3T3 cells, perhaps indicating that constitutive expression of PC-PLC leads to activation of most, if not all, of the downstream targets of Ras. In keeping with this notion, we found that we could revert the transformed phenotype of plc expressing cells by coexpression of a dominant negative mutant of Raf-1. This result corroborates the findings of Cai et al. (16) who, by employing colony formation assays with cotransfection of PC-PLC and dominant negative Raf-1 mutants, recently demonstrated that PC-PLC is unable to bypass the block to proliferation caused by dominant negative Raf mutants. They also found that treatment with D609, a specific inhibitor of PC-PLC enzyme activity (45), completely inhibited the activation of Raf-1. In addition, we demonstrate for the first time that PKC is required for transformation and growth factor-independent DNA synthesis in both v-ras-and plc-transformed fibroblasts. Although our present results permit the conclusion that both Raf-1 and PKC are required for transduction of the signal elicited by PLC-catalyzed hydrolysis of PC, this study does not address the mechanism(s) of activation of these kinases. The mechanism(s) of activation of Raf is not completely understood. However, it seems clear that the N-terminal regulatory domain of Raf binds to the effector loop of GTP-bound Ras and that Ras may serve to ferry the Raf kinase to the plasma membrane where it is subsequently activated (46 -52). At the membrane, Raf is part of a complex including GTPbound Ras, mitogen-activated protein kinase/extracellular signal-regulated kinase kinase-1, and mitogen-activated protein kinase. Following phosphorylation, mitogen-activated protein kinase is released from this complex (47,53). GTP-bound Ras is not by itself able to activate the kinase activity of Raf (34,52), suggesting that an additional signal is needed. Recent results pinpoint membrane-associated second messengers as likely candidates for the second signal needed for full activation of Raf (16,46,48). Thus, this signal could be provided by DAG generated by PC hydrolysis interacting with the cysteine finger in the regulatory CR1 domain of raf-1 p74 (16,20). Alternatively, PC-derived DAG may activate a kinase(s) distinct from phorbol ester-sensitive PKCs, which then can activate Raf by direct phosphorylation (55,56). Whatever the mechanism, our present results and those of Cai et al. (14,16) together strongly implicate a crucial role for PC-derived DAG in this mitogenic signaling pathway and place this step downstream of Ras but upstream of Raf. Additional support for this notion is provided by our demonstration that activation of PKA by treatment with forskolin and/or 8-bromo-cAMP reverted the transformed phenotype of plc-transformed cells as evidenced by morphological reversion, loss of ability to form colonies in soft agar and lack of induction of DNA synthesis in the absence of growth factors. Since Raf-1 has been shown to be the target of the inhibitory action of PKA (24,44), our results with forskolin and 8-bromo-cAMP confirm the downstream location of Raf-1 relative to PC-PLC. Furthermore, we show that plc-transformed cells do not contain elevated levels of GTP-bound Ras. Together with the previous finding that expression of PC-PLC is able to relieve the block to proliferation imposed by expression of the N-17 dominant negative mutant of Ras (14), this strongly supports the conclusion that PC-PLC acts downstream of Ras.
As for Raf-1, the mechanism(s) of activation of the atypical subtype of PKC is not completely understood. However, it is clear that the enzyme is not activated by phorbol esters or short-chain DAGs (22,36). Recently, it was reported that in vitro PKC can be activated by phosphatidylinositol 3,4,5trisphosphate, a product of PI 3-kinase (57). If this mechanism is active in vivo, a direct link between PI 3-kinase activation following binding to activated tyrosine kinase receptors and stimulation of PKC can be envisioned. Interestingly, Ras may contribute to the activation of PI 3-kinase by directly interacting with the catalytic p110 subunit (58,59). This would place both activation of PI 3-kinase and PKC downstream of Ras. There is also evidence that PKC is activated in vivo by treatment of NIH 3T3 cells with sphingomyelinase C capable of generating the lipid second messenger ceramide and that ceramide can activate PKC in vitro (60). Thus, PKC may be regulated by different lipid mediators. Different lines of evi-dence suggest that PC-PLC is acting upstream of PKC (19,26,60). Considering the fact that PKC is activated by ceramide, the downstream location relative to PC-PLC is completely consistent with the prevalent model for tumor necrosis factor ␣ signaling where an acidic sphingomyelinase C is activated by DAG generated by a membrane-bound PC-PLC (45). As previously shown for Raf-1, it has recently been demonstrated that PKC interacts with Ras both in vitro and in vivo and that the in vivo interaction is dependent on GTP-bound active Ras and takes place between the N-terminal regulatory domain of PKC and the effector domain of Ras (61). Thus, analogous to Raf, the PKC-Ras interaction may serve to bring PKC to the membrane where its kinase activity is induced by a lipid mediator(s). This activation is probably direct since, contrary to Raf-1 (56), there is no evidence for a role of phosphorylation/dephosphorylation in the activation of the kinase activity of PKC. Thus, the same role for PC-derived DAG proposed for Raf-1 activation above may be applicable to PKC, which is similar to Raf-1 in its structural organization. Hitherto, experiments directly addressing binding and activation by physiological relevant PC-derived DAG species have not been performed for these kinases. Alternatively, as outlined above, PC-derived DAG may activate a sphingomyelinase C, which in turn produces ceramide that will directly activate PKC.
In view of the proven direct interaction between Ras and PKC, our results with expression of the kinase-defective mutant of PKC in v-ras cells could be explained as simply due to a blockade of the ability of Ras to interact productively with both PKC, Raf-1, and PI 3-kinase as well as other presently unknown downstream targets that may be dependent on binding to the effector domain of activated Ras. However, this model fails to explain why dominant negative mutants of both PKC and Raf is able to revert plc-transformed cells since PC-PLC has been convincingly shown, by different experimental approaches, to act downstream of Ras (3, 14 -16, this work). Colony formation assays have revealed that cotransfection of activated PKC is not able to relieve the block to proliferation of NIH 3T3 cells imposed by a dominant negative Raf-1 mutant. Likewise, an activated form of Raf-1 did not abolish the FIG. 7. The mitogenic signal elicited by PC-PLC is not dependent on classical, TPA-responsive PKC subtypes but is blocked following treatment with PKA activating agents. A, neither depletion of TPA-responsive PKC subtypes or inhibition of PKC by GF 109203X blocked the mitogenic signal elicited by v-ras or PC-PLC (P18 and P12) expression. To down-regulate PKC TPA was added to the culture medium during serum starvation (24 h) to a final concentration of 500 ng/ml (Long term TPA). The PKC inhibitor GF 109203X (500 nM) was added to quiescent cultures 18 h prior to cell harvesting and determination of [ 3 H]thymidine incorporation. Short term stimulation with 100 ng/ml TPA (Short term TPA) was performed for 1 h following serum starvation. Addition of GF 109203X totally abolished the mitogenic response resulting from short term stimulation with TPA (data not shown). Open bars indicate the induction of DNA synthesis observed in the absence of growth factor addition. B, increasing the intracellular cAMP level either indirectly by forskolin treatment or directly by addition of 8-bromo-cAMP (8-Br-cAMP) inhibited DNA synthesis induced by both PC-PLC and v-ras, while 8-bromo-cGMP (8-Br-cGMP) was without effect. Additions of 10 ng/ml PDGF (BB homodimer), 0.5 mM 8-bromo-cAMP, 0.5 mM 8-bromo-cGMP, or 0.1 mM forskolin are indicated (ϩ). The results in A and B are expressed as means with standard errors of three independent experiments performed in triplicate. The mitogenic response to 10 ng/ml PDGF (BB) was used to set the 100% level. growth inhibitory action of a dominant negative PKC mutant (61). Furthermore, expression of dominant negative PKC in NIH 3T3 cells did not inhibit PDGF-stimulated Raf-1 phosphorylation of mitogen-activated protein kinase/extracellular signal-regulated kinase kinase-1. 2 Also, a kinase defective mutant of PKC, but not a similar mutant of Raf-1, is able to block the activation of the stromelysin promoter mediated through a novel palindromic PDGF, Ras and PC-PLC responsive element (23). Together with our present results, these findings suggest a bifurcation of the mitogenic signaling pathway downstream of PC-PLC with Raf-1 and PKC located on separate branches. Thus, more than one signaling pathway need to be activated in order to bring about mitogenesis or oncogenic transformation.