INTRODUCTION

Much attention has been devoted to the study of genes expressed aberrantly in transformed cells. In some instances, a role in transformation has been defined for the protein encoded by these genes (Dehbi and Bédard, 1992). In addition, transformation-regulated genes provide useful markers to investigate the signaling pathways activated by the oncoproteins. For instance, it has been shown that Ha-Ras is critical for the activation of the Egr-1 promoter and TIS10 gene (also known as Cox-2 and prostaglandin synthase) in v-src-transformed cells (Qureshi et al., 1992; Xie et al., 1994). Likewise the serine/threonine kinase c-Raf-1 is critical for the expression of Egr-1 but plays no role in the activation of TIS10 (Foster, 1993). Therefore, multiple pathways linked to Ha-Ras have been implicated in transcriptional activation by pp60^v-src. The nuclear targets of these oncoproteins are essential for gene expression and for the process of transformation itself. The expression of a dominant negative mutant of c-Jun blocks transformation by v-src and Ha-ras (Granger-Schnarr et al., 1992; Lloyd et al., 1991; Suzuki et al., 1994). Therefore, the oncogenic action of pp60^v-src and p21^ras is dependent on the activation of AP-1. Similar findings have been reported for the Ets family of transcription factors (Bruder et al., 1992; Wasylyk et al., 1994). In agreement with these observations, several components of the transduction pathways controlling the activity of AP-1 or Ets-like factors have oncogenic activity on their own. Moreover, the micro-injection of antibodies or the expression of a dominant negative mutant of Ha-Ras or c-Raf-1 abolish transformation by pp60^v-src (Qureshi et al., 1993; Smith et al., 1986). Hence, a more coherent view of cell transformation has emerged from the combined study of signaling pathways and gene expression.

Despite this recent progress, several aspects of cell transformation are poorly understood. The recent observation that a dominant negative mutant of Rac-1 or Rho-1 interferes with transformation by v-src or ras emphasizes the need to improve our knowledge of other members of this superfamily of small GTPases (Khosravi-Far et al., 1995; Qiu et al., 1995). The recent identification of JNK (SAPK) as a downstream effector of Rac-1 implies a role for the latter in the control of c-Jun activity (Foster, 1993; Minden et al., 1995; Olson et al., 1995). Since distinct functions such as intracellular localization, potential of trans-activation, DNA binding, dimerization, or expression can be regulated separately, it is likely that multiple transduction pathways cooperate in the activation of a single transcription factor.

We have previously characterized the promoter of the v-src responsive CEF-4/9E3 cytokine gene in chicken embryo fibroblasts (CEF)¹ (Dehbi et al., 1992). Three distinct promoter elements corresponding to an AP-1 binding site (or TPA responsive element, TRE) a PRDII/B domain, and a CAAT box are necessary for maximal expression in CEF transformed by the Rous sarcoma virus (RSV). Accordingly, this region of the promoter has been designated the v-src responsive unit (SRU) of the CEF-4 gene. In this report we characterize the transduction pathways involved in the activation of the TRE and PRDII domain of the promoter. Using minimal promoter constructs controlled by a single element, we show that the activated and dominant negative forms of Ras and the MEKK pathway regulate the activity of the PRDII and TRE binding factors. In contrast, a role for c-Raf-1 and protein kinase C (PKC) appears to be restricted to the activation of the TRE. Moreover, chronic TPA treatment had opposite effects on the activity of these elements. While the activation of a TRE-controlled promoter was reduced significantly by the prolonged treatment with phorbol esters, the activation of PRDII was not affected and was even superinduced in v-src transformed CEF. Further characterization confirmed that chronic TPA treatment results in the constitutive stimulation of PRDII and the CEF-4 promoter in CEF. Since CEF treated with TPA do not display a fully transformed phenotype, these results imply that constitutive activation through PRDII is not sufficient to cause transformation. The addition of calphostin C, a specific inhibitor of PKC, resulted in the down-modulation of the TRE activity and in the flattening or morphological reversion of RSV-transformed CEF. Hence, transformation was correlated with the activation of the TRE but not PRDII.

MATERIALS AND METHODS

Early passages of CEF were grown at 41.5 °C in Richter-improved minimal essential medium supplemented with 5% heat-inactivated newborn bovine serum (Cansera, Rexdale, Ontario, Canada), 5% tryptose phosphate broth, L-glutamine, penicillin, and streptomycin. Unless indicated, all experiments were done with actively growing cells. CEF were infected with the wild type Schmidt-Ruppin A strain of the Rous sarcoma virus (SR-A RSV) or with the temperature-sensitive mutant ts NY72-4 RSV. Cells were treated with various concentrations of tetradecanoyl phorbol acetate (TPA, Sigma) or calphostin C (Calbiochem) as indicated in the text or with the equivalent concentration of the diluent in control samples (0.1% ethanol and 0.1% dimethyl sulfoxide, respectively). A concentration of 25 mM Hepes, pH 7.6, was added routinely to the culture medium of transfected CEF to prevent excessive variations in pH.

Constructs of the CEF-4 promoter fused to the chloramphenicol acetyltransferase gene (CAT gene) are depicted in Fig. 1 and have been described previously (Dehbi et al., 1992). Briefly, plasmid Spe/CAT includes nucleotides 214 to +36 of the 5 flanking and transcribed region of the CEF-4 gene and contains an intact v-src-responsive unit (SRU). Plasmids µPRD/CAT and µTRE/CAT are equivalent constructs containing a mutation in the PRDII domain or the TPA-responsive element (TRE) of the CEF-4 SRU, respectively. The deletion of nucleotides 64 to 214 (deletion SpeI/NheI in plasmid S-N/CAT) in a larger fragment of the 5 flanking region of CEF-4 (spanning nucleotides 1312 to +36) results in the deletion of the SRU and in an inactive promoter (Dehbi et al., 1992). Constructs 4XPRD/CAT and 4XTRE/CAT consist of four copies of the PRDII domain or the TRE inserted in place of nucleotides 214 to 64 in plasmid S-N/CAT. Both constructs are v-src-responsive but are controlled by a single element of the SRU. Plasmids P3-CAT1/TATA and 3XTRE-CAT1/TATA were constructed by inserting three copies of the PRDII domain or the TRE in proximity of an heterologous minimal promoter. To this end, synthetic double-stranded oligonucleotides for PRDII (5-AGCTTCTG-3) or the TRE (5-AGCT-3) were inserted in the SalI site of plasmid pJFCAT1/TATA that includes a TATAAAA box and the initiation start site of human -globin gene. The minimal promoter of pJFCAT1/TATA has a weak basal activity that is not regulated by pp60^v-src in CEF (Dehbi et al., 1994). Equivalent constructs containing a mutation in the PRDII domain or the TRE were also generated with plasmid pJFCAT1/TATA (data not shown).

The plasmid expressing a dominant negative mutant of Ha-Ras pZIP M17 or M17RAS and the parental vector pZIPneo were a generous gift of Dr. G. Cooper (Feig and Cooper, 1988). The dominant negative mutant of Ha-Ras was also expressed from the RSV long terminal repeat by subcloning the XbaI-PstI fragment of pZIPM17 in the HindIII site of plasmid pRSV by blunt end ligation. Likewise, a v-Ha-ras sequence was subcloned in plasmid pRSV by isolating the BamHI fragment of plasmid pJCS1 (kindly provided by Dr. J. Stone). The same results were obtained when the dominant negative or activated form of Ha-Ras was expressed in plasmid pRSV or in the original vectors described above. Plasmid Raf375W encodes a kinase-deficient, dominant negative mutant of c-Raf-1 under the control of the RSV long terminal repeat (Bruder et al., 1992). An equivalent construct was also generated independently by inserting the dominant negative mutant of c-Raf-1 (kindly provided by Dr. S. Meloche) in plasmid pRSV. Plasmid BXB-Raf encodes a constitutively activated form of c-Raf-1 (Bruder et al., 1992). Constructs of wt MEKK and the dominant negative mutant of SEK have been described elsewhere (Sanchez et al., 1994; Yan et al., 1994) and were a generous gift of Dr. J. Woodgett. Plasmid pRSVgal consists of the lacZ gene under the control of the RSV long terminal repeat (Edlund et al., 1985).

Normal CEF and CEF infected with the temperature-sensitive mutant NY72-4 RSV were plated at a density of 2 × 10⁶ cells/100-mm dish 24 h prior to transfection. All transfections were done at the nonpermissive temperature of 41.5 °C by the DEAE-dextran method as described previously (Dehbi et al., 1992). A total amount of 30 µg of DNA consisting of 2 or 10 µg of the reporter plasmid for NY72-4 RSV-infected CEF and normal CEF, respectively, 2 µg of the pRSVgal internal control, 0.1-18 µg of effector plasmid (or parental vector, see the figure legends for the details of each transfection), and variable amounts of carrier salmon sperm DNA was transfected per 100-mm dish. Plasmid pRSVgal was used to correct for transfection efficiency since it is not regulated by pp60^v-src transformation (Dehbi et al., 1992; Edlund et al., 1985). Results obtained with equal amounts of total protein were also in close agreement with those based on equal levels of -galactosidase activity. All combinations of plasmids were investigated in duplicate or triplicate samples in at least two separate experiments. The errors indicated on bar graphs represent the deviation from the means of duplicate or triplicate samples. Increasing concentrations of effector plasmid or the corresponding parental vector (0.1-18 µg per 100-mm dish) were investigated to obtain maximal inhibition or stimulation by the dominant negative or activated forms of Ras, Raf, SEK, or MEKK and to control for possible squelching effects of the plasmid. Transformation was induced by transferring NY72-4 RSV-infected CEF to the permissive temperature of 37 °C for 24 h. Stimulation by phorbol esters was performed by a single addition of 800 nM TPA to the culture medium.

NY72-4 RSV-infected CEF were transfected with plasmid pcDNA3-HA-SAPK46 (5 µg/100-mm dish) in the presence or in the absence of a plasmid encoding SEK (5 µg) or MEKK (0.1 µg), or devoid of any insert (i.e. the parental vector pcDNA3). Two days later, a cell lysate was prepared from cells transferred to the permissive temperature for 24 h or maintained at the nonpermissive temperature for the duration of the experiment. When indicated cells were treated for 40 min with anisomycin (50 µg/ml) before lysis. Transfected CEF were washed in cold phosphate-buffered saline and incubated for 1 h in a lysis buffer consisting of 0.5% Nonidet P-40, 3 mM EDTA, 3 mM EGTA, pH 7.6, 0.3 M NaCl, 2 mM MgCl₂, 1 mM dithiothreitol, 5 mM benzamidine, 20 mM -glycerophosphate, 1 mM sodium vanadate, 1 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin and aprotinin on ice. The epitope-tagged SAPK was immunoprecipitated from the cleared lysate with the 12C5A monoclonal antibody (incubated for 1 h on ice; Boehringer Mannheim) and recovered with protein A-agarose. The immune complex was washed extensively in a buffer consisting of 50 mM NaCl, 0.1 mM EDTA, 1 mM sodium vanadate, 1 mM sodium fluoride, 0.05% Triton X-100, and 25 mM Hepes, pH 7.6, and then once in the kinase reaction buffer (20 mM MgCl₂, 20 mM -glycerophosphate, 1 mM dithiothreitol, 0.5 mM EGTA, 1 mM sodium vanadate, 0.5 mM sodium fluoride, and 25 mM Hepes, pH 7.6). Kinase assays were performed at 30 °C for 30 min in 30 µl of the kinase reaction buffer containing 20 µM ATP, 10 µCi of [-³²P]ATP, and 1.5 µg of purified GST-c-Jun (amino acids 5-89) as a substrate. The reactions were terminated by boiling in an SDS sample buffer, and the products were resolved on a 10% polyacrylamide gel followed by autoradiography. The phosphorylated GST-c-Jun proteins were quantitated in an Instant Imager (Packard).

Total cell RNA was isolated by the high salt-urea precipitation method as described previously (Auffrey and Rougeon, 1980; Bédard et al., 1987). An equal mass of total RNA (7.5 µg) was loaded per well and separated on a 1% agarose gel containing formaldehyde. The RNA was then transferred on a NYTRAN membrane (Schleicher & Schuell) and probed with radiolabeled cDNAs for the various ``CEF genes'' or glyceraldehyde-3-phosphate dehydrogenase (Simmons et al., 1989). Radiolabeling was done by random priming (Feinberg and Vogelstein, 1983).

Total proteins were isolated from CEF monolayers washed in cold phosphate-buffered saline and lysed by boiling in an SDS-containing buffer (2.3% SDS, 5% -mercaptoethanol, 10% glycerol, and 62.5 mM Tris, pH 6.8). Proteins were separated by electrophoresis on a 10% polyacrylamide gel, transferred electrophoretically onto a nitrocellulose membrane, and incubated with commercially available antibodies against protein kinase C or - (number 10267-011, Transduction Laboratories, Lexington, KY) as described previously (Gonneville et al., 1991). The immune complex was detected after incubation with a peroxidase-conjugated secondary antibody and revealed with a chemiluminescent substrate (ECL system, Amersham Canada, Oakville, Ontario).

RESULTS

The CEF-4 promoter is activated gradually in CEF transformed by a temperature-sensitive mutant of pp60^v-src (Dehbi et al., 1992). AP-1 and the PRDII binding protein accumulate with different kinetics in CEF infected with ts NY72-4 RSV. The level of AP-1 is nearly maximal within 6-12 h of activation of pp60^v-src, whereas the PRDII binding protein continues to accumulate well beyond this interval. Thus, induction through the TRE and PRDII may occur at different times. To address this question constructs of the CEF-4 promoter containing four copies of the TRE or PRDII were transfected in NY 72-4 RSV-infected CEF, and the activity of the reporter gene was measured at different times after transfer to the permissive temperature. As shown in Fig. 2A, the activation of the TRE was maximum within 12 h of temperature shift while the activity of the PRDII controlled promoter continued to increase thereafter. The same conclusion was reached with constructs consisting of a heterologous promoter fused to multiple copies of the TRE or PRDII (Fig. 2B). Therefore, the TRE and PRDII domain were activated with different kinetics in RSV-transformed CEF. This observation suggests that pp60^v-src acts on different signal transduction pathways to control the activity of AP-1 and the PRDII-binding factor.

The Ras-Raf pathway is important in the transformation by v-src since the micro-injection of neutralizing antibodies or the expression of a dominant negative mutant of these proteins decrease the tumorigenicity and interfere with morphological transformation of v-src-transformed cells (Qureshi et al, 1993; Smith et al., 1986). To examine the role of this pathway, minimal promoter constructs of the TRE and PRDII were co-transfected with a plasmid expressing a dominant negative mutant of Ha-Ras or c-Raf-1. The effect of the wild type or activated form of the proteins was investigated in parallel. Levels of the CAT enzyme were determined in cells maintained at the nonpermissive temperature of 41.5 °C or transferred to 37 °C for 24 h. The activity of the PRDII-controlled promoter was markedly reduced by the Ras dominant negative mutant in transformed CEF and to a lesser extent in CEF maintained at the nonpermissive temperature (Fig. 3B). Likewise the activity of the TRE-controlled promoter was inhibited by the dominant negative mutant at both temperatures (Fig. 3C). The expression of v-Ha-ras strongly stimulated the TRE but had a more modest effect on the PRDII-controlled promoter (Fig. 3A). This poor inducibility of PRDII was observed repeatedly in transient expression assays over a wide range of plasmid concentrations suggesting that at least in these experimental conditions Ha-Ras is not a potent inducer of PRDII. Since the dominant negative mutant interfered with activation by v-src, Ha-Ras may nevertheless interact with a component of the pathway controlling PRDII.

A similar investigation was carried out for c-Raf-1. A reduction in the activity of the TRE was observed in cells expressing the dominant negative mutant (plasmid RAF375W in Fig. 3C). In all experiments, however, the inhibition never exceeded 35% of the total activity of the promoter. In contrast, the activation of PRDII was not affected by the co-transfection of plasmid RAF375W (Fig. 3B). The expression of the activated form of c-Raf-1 enhanced the activity of the TRE but did not stimulate the PRDII-controlled promoter, in agreement with the results obtained with the dominant negative mutant (Fig. 3A). In fact the activity of PRDII was consistently repressed by the activated form of the kinase. Therefore, we conclude that c-Raf-1 controls in part the activation of AP-1 but plays no inductive role in the pathway controlling PRDII in v-src-transformed CEF.

The activity of c-Jun (and AP-1) is dependent on dephosphorylation of residues at the C terminus of the protein (controlling DNA binding) as well as phosphorylation and potentiation of the activation domain of the protein (located at the N terminus) (Binetruy et al., 1991; Boyle et al., 1991). Since v-src is a potent inducer of JNK, the kinase(s) responsible for phosphorylation of the activation domain of c-Jun (also known as SAPK) (Kyriakis et al., 1994; Minden et al., 1995; Xie and Herschman, 1995), we examined the role of this pathway in the function of the TRE and PRDII domain. The expression of a dominant negative mutant of SEK1 (SAPK/ERK kinase-1) reduced the activity of the TRE-controlled promoter in both normal and transformed CEF (Fig. 4D). This reduction ranged from 15 to 30% of total promoter activity in different experiments. In addition, the dominant negative mutant of SEK was generally just as efficient on the activity of the TRE at the nonpermissive and permissive temperature suggesting that the SAPK pathway was also involved in the control of the TRE activity in normal CEF (data not shown). A more pronounced effect of the SEK dominant negative mutant was observed on the activity of PRDII with inhibitions ranging from 35 to 55% (Fig. 4C). In addition, the inhibition was consistently more pronounced at the permissive temperature. We then examined the effect of MEKK, a kinase that is responsible for the activation of SEK and occupies a position analogous to that of c-Raf-1 in the MEK-ERK pathway. Since overexpression of MEKK may also lead to stimulation of MEK (Lange-Carter et al., 1993; Yan et al., 1994), various amounts of the MEKK expression vector were co-transfected in NY72-4 RSV-infected CEF. These results are presented in Fig. 4, A and B. Higher concentrations of MEKK enhanced modestly the activity of the TRE. In contrast, the activity of PRDII was stimulated markedly at all concentrations of plasmid DNA (Fig. 4A). Therefore, PRDII was more responsive to the action of MEKK, and a significant stimulation of the promoter was observed with DNA concentrations that had no significant effect on the TRE (such as 0.1 µg of the MEKK plasmid DNA; note the difference in scale for the y axis in A and B). The action of MEKK on PRDII was also more potent at the permissive temperature, an effect that was less pronounced on the TRE. Therefore, we conclude that MEKK and the SAPK pathway are important for the activation of the PRDII-binding factor by pp60^v-src. In addition, this pathway is important for the activity of AP-1 at both the permissive and nonpermissive temperature suggesting that it is at least partially activated in normal CEF. To resolve this issue, we measured the activity of JNK/SAPK in normal and transformed CEF. To this end, an expression construct encoding an HA-tagged SAPK was transfected in NY72-4 RSV-infected CEF. SAPK was isolated by immunoprecipitation from cell lysates prepared from CEF transferred to the permissive temperature for 24 h or maintained at the nonpermissive temperature for the duration of the experiment. In addition, cells at the nonpermissive temperature were treated with anisomycin for a period of 40 min prior to cell lysis. GST-c-Jun was phosphorylated in vitro by the immunoprecipitated SAPK and analyzed by polyacrylamide gel electrophoresis and autoradiography. The results shown in Fig. 5A indicate that the activation of the thermolabile pp60^v-src did not result in a significant induction of SAPK activity in CEF (lanes 3 and 4). The same experiment performed with total cell lysate (thus looking at total GST-c-Jun kinase activity) led to the same conclusion although a modest activation (of less than 3-fold) was often observed in CEF stably transformed by SR-A RSV (data not shown). Treating nontransformed CEF with anisomycin, a potent inducer of SAPK (Kyriakis et al., 1994), enhanced the activity of the ectopically expressed kinase, but again this induction was modest (lane 5). It is possible that a limiting component of this pathway accounts for the poor activation of SAPK in our experimental conditions. Therefore, we also co-transfected vectors encoding SEK and MEKK and quantitated the level of SAPK activity at the permissive and nonpermissive temperature. The results of this experiment are shown in Fig. 5B. The addition of SEK resulted in a small increase of SAPK activity (lanes 8 and 9), whereas MEKK led to a significant enhancement of GST-c-Jun phosphorylation activity in the presence or in the absence of SEK (lanes 10-13). In all cases, however, the transfer to the permissive temperature did not result in a marked increase in SAPK activity. Therefore, the activity of SAPK was not strongly stimulated by the activation of pp60^v-src in CEF, in apparent contradiction with the results obtained by other investigators in NIH 3T3 cells (Minden et al., 1995; Xie and Herschman, 1995). These results will be discussed below.

Sprangler et al. (1989) reported that the accumulation of the CEF-4 mRNA is dependent on PKC in RSV-transformed CEF. However, the component of CEF-4 induction controlled by PKC is unknown. Therefore, we examined the effect of calphostin C, a specific inhibitor of PKC, on the accumulation of the CEF-4 mRNA and on the activity of a TRE- and PRDII-controlled promoter. Since CEF-4 is regulated by different mechanisms after activation of pp60^v-src (Dehbi et al., 1992; Gonneville et al., 1991), NY 72-4 RSV-infected CEF were treated with 1 µM calphostin C at different times after transfer to the permissive temperature. Cells were incubated for a total of 5 h in all cases. Northern blotting analysis indicated that the basal expression at the nonpermissive temperature and the early phase of CEF-4 expression following activation of pp60^v-src were not affected by calphostin C (Fig. 6A; lanes 1-4). In contrast, the level of CEF-4 mRNA was reduced significantly at 8 and 12 h after transfer to the permissive temperature (a 50-60% decrease; lanes 5-8). This period of CEF-4 induction corresponds to the transition from an exclusively post-transcriptional to a predominantly transcriptional level of regulation (Dehbi et al., 1992; Gonneville et al., 1991). Little effects were observed after this period (lanes 9-10). Interestingly, the accumulation of the mRNA for other v-src inducible genes (CEF-5, CEF-10, and CEF-147; Simmons et al., 1989) was not affected by calphostin C.²

The effects of PKC inhibition were also studied in transient expression assays. 72-4 RSV-infected CEF were treated with calphostin C the day after transfection and transferred to the permissive temperature for a period of 24 h. No change was observed in the activation of the PRDII controlled promoter (Fig. 6B). In contrast, the activity of the TRE was greatly reduced by calphostin C suggesting that PKC is involved in the activation of AP-1. The same result was obtained when cells were treated with staurosporine, a different inhibitor of PKC.³ CEF were also treated chronically with a high concentration of phorbol esters to down-regulate PKC. Three different constructs of the CEF-4 promoter were transfected in NY72-4 RSV-infected CEF treated with 800 nM TPA for a period of 48 h. In agreement with the results obtained with calphostin C, chronic treatment with TPA decreased markedly the activation of the TRE-controlled promoter (Fig. 6C). In contrast the activity of the PRDII-controlled promoter was enhanced in the same experiment suggesting again that PKC is not required for the activation of PRDII by pp60^v-src. The superinduction of PRDII was unexpected since it was not observed in cells treated with the PKC inhibitors. Moreover, we found that the intact CEF-4 promoter was also stimulated by the chronic TPA treatment of RSV-infected CEF (Fig. 6C). Hence, the action of phorbol esters may be more complex than suggested initially by the effects of calphostin C. This question will be addressed below.

The addition of calphostin C had a striking effect on the morphology of RSV-transformed CEF (Fig. 7). We consistently observed that cells treated for 5 h with 1 µM calphostin C were flatter and appeared morphologically normal. Criss-crossing was greatly reduced, and cells began to form a monolayer of contact-inhibited cells in regions of higher cell density. The most likely interpretation of this result is that calphostin C interfered with the transformation of CEF by pp60^v-src. We found no qualitative or quantitative difference in the pattern of phosphotyrosine-containing proteins in cells treated with calphostin C or with the diluent alone. Likewise, calphostin C at the concentration of 1 µM did not inhibit the activity of pp60^v-src in the IgG kinase assay (data not shown). Therefore, the phenotype induced by calphostin C was not the result of inhibition of pp60^v-src. Other specific inhibitors of PKC such as CGP 41 251 (Meyer et al., 1989) and GF109203X (Toullec et al., 1991) had a less pronounced effect but also resulted in the flattening of RSV-transformed CEF.⁴ This result implies a critical role for PKC in the transformation of CEF by RSV.

Chronic TPA treatment did not inhibit the activation of the CEF-4 promoter by pp60^v-src and in fact resulted in a superinduction. To characterize this response, we examined the expression of the CEF-4 mRNA in normal CEF treated with a high concentration of TPA for increasing periods. The addition of a single dose of TPA at 800 nM was sufficient to reduce the steady state level of PKC and PKC for a period of at least 48 h (Fig. 8A and data not shown). Rapid accumulation of the CEF-4 mRNA was observed in cells treated with the phorbol ester. This expression was biphasic but remained elevated throughout the duration of the treatment (Fig. 8B). The same conclusion was reached when TPA was added at the concentration of 400 or 300 nM or when mRNAs were isolated 60 h after addition of the drug. Prolonged accumulation of the CEF-4 mRNA was observed in TPA-treated CEF cultured in the presence or in the absence of serum (i.e. in actively growing and in serum-strarved cells; data not shown). Hence, chronic TPA treatment resulted in the constitutive activation of CEF-4. We also analyzed the expression of three other v-src inducible genes (the CEF genes) (Simmons et al., 1989). The mRNA for CEF-5 (the avian homologue of Egr-1) and CEF-147 (prostaglandin synthase or Cox-2) accumulated rapidly in response to TPA but did not show the same level of high constitutive expression observed with CEF-4 (Fig. 8B). Two transcripts were detected by the CEF-147 probe. The upper transcript, which includes an intron, does not encode a functional prostaglandin synthase gene product while the lower transcript encodes the functional protein (Xie et al., 1991). CEF-10 was originally described by Simmons et al. (1989) as a phorbol ester repressible gene. Hence, the level of the CEF-10 mRNA decreased soon after the addition of TPA but was expressed at a markedly elevated level at later time points. Thus, CEF-4 and CEF-10 belong to a class of genes whose expression is v-src-inducible and enhanced constitutively in cells chronically treated with TPA. The expression of c-fos was only detected at the 1-h time point in this experiment (data not shown).

Run-on transcription analysis indicated that the expression of CEF-4 was regulated predominantly at the level of transcription in conditions of chronic TPA treatment.⁴ Therefore, we examined the activity of the TRE- and PRDII-controlled promoter in response to a prolonged TPA treatment. In this experiment, CEF were transfected with the P3-CAT1/TATA or 3XTRE-CAT1/TATA vector and treated for increasing periods with a single dose of TPA at 800 nM. As shown in Fig. 8C, both the TRE and PRDII domain were activated early on by TPA. However, at later time points, the activity of the TRE-controlled promoter decreased significantly while that of PRDII was superinduced. Thus, the expression of PRDII but not the TRE was enhanced markedly in cells chronically treated with TPA. Since the CEF-4 promoter is also induced in conditions of chronic TPA treatment, the results of Fig. 8C suggest that PRDII but not the TRE is critical at later time points. This question was addressed by the transfection of constructs containing a mutation of the TRE, PRDII domain, or CAAT box of the CEF-4 promoter. As shown in Fig. 8D, a significant reduction in the activity of the promoter was observed only with the construct harboring a mutation in the PRDII element. Therefore, constitutive activation of the CEF-4 promoter was dependent predominantly on the PRDII domain in conditions of chronic TPA treatment. This contrasts with the expression in RSV-transformed CEF that requires all three elements of the promoter (Dehbi et al., 1992). These results are intriguing because CEF treated with TPA are elongated but do not display the morphology of fully transformed cells. They are also incapable of growing in soft agar.⁴ Therefore, the constitutive activation of CEF-4 did not correlate with transformation in conditions of chronic TPA treatment. Since the treatment with calphostin C resulted in morphological reversion, these observations suggest that the constitutive activation of CEF-4 is a marker of transformation when AP-1 is involved but not when dependent solely on the PRDII domain of the promoter.

DISCUSSION

Several results indicated that multiple signaling pathways control the activation of the CEF-4 promoter. Individual elements of the SRU of CEF-4, namely the TRE and PRDII, were activated with different kinetics when fused to a minimal promoter (Fig. 2). The activity dependent on the TRE was maximum within 8 h of pp60^v-src activation while that of the PRDII-controlled promoter continued to increase after this interval. Likewise, the TRE binding activity (AP-1) investigated by electrophoretic mobility shift assay was nearly maximum within 6 h of activation of a thermolabile pp60^v-src, while again the PRDII binding factor required more than 12 h to reach a maximum level (Dehbi et al., 1992). Despite these differences, we observed that p21^ras is essential for the activation of both factors (Fig. 3). This conclusion is in agreement with the results of other investigators who reported a similar finding for the activation of the SRE and CRE (Qureshi et al., 1992; Xie and Herschman, 1995). It is also consistent with earlier results demonstrating that transformation by v-src is blocked by the injection of p21^ras antibodies (Smith et al., 1986). Therefore, p21^ras occupies a central position in the signaling pathways activated by pp60^v-src.

There is increasing evidence that Ha-Ras controls several pathways that cooperate in transformation (White et al., 1995). Thus far the pathway best understood involves the interaction of Ras and the c-Raf-1 serine/threonine kinase. Several observations had linked these two proteins functionally. In addition, recent results indicated that Ha-Ras interacts physically with Raf resulting in the translocation of the serine/threonine kinase to the membrane where it is activated (Koide et al., 1993; Van Aelst et al., 1993; Vojtek et al., 1993). This pathway and perhaps also a direct interaction of c-Raf-1 with pp60^v-src (Fabian et al., 1993) appear to be involved in the activation of a TRE-controlled promoter in RSV-transformed CEF (Fig. 3). Downward and his colleagues (1994) described an alternative pathway depending on the direct interaction of Ha-Ras with the phosphatidylinositol-3-OH kinase. However, we did not see any effect of wortmannin, a potent inhibitor of phosphatidylinositol-3-OH kinase, on the activation of the TRE or PRDII element, suggesting that this pathway is not involved in the activation of AP-1 and PRDII binding factor in CEF.³

Finally, Ha-Ras interacts at least functionally with other members of the superfamily of monomeric G-proteins. The expression of a dominant negative mutant of Rac1 impairs the activation of the JNK/SAPK pathway by ras and v-src implying a role for this protein in the activation of c-Jun (and AP-1; Coso et al., 1995; Minden et al., 1995; Olson et al., 1995). We have also observed an effect of the SEK dominant negative mutant on the activity of the TRE (Fig. 4). However, this effect was rather modest and not restricted to transformed cells. Moreover, the measure of SAPK/JNK activity indicated that little activation of this kinase(s) occurs in response to the action of pp60^v-src (Fig. 5). Our results are in contradiction with those reported by other groups (Minden et al., 1995; Xie and Herschman, 1995). The discrepancy may stem from the investigation of different cell systems, namely CEF and NIH 3T3 cells. Catling et al. (1993) reported that the activation domain of c-Jun is hyperphosphorylated even in resting CEF and that the dephosphorylation of the DNA binding domain of c-Jun accounts primarily for the increase in AP-1 activity in v-src-transformed CEF. Our results support this conclusion since (a) we did not observe any significant induction of SAPK activity upon activation of pp60^v-src; (b) treating cells with anisomycin, a potent inducer of SAPK (Kyriakis et al., 1994), resulted only in a modest increase in activity; and (c) activating pp60^v-src led to a marked increase in TRE-binding activity in CEF (Fig. 5) (Dehbi et al., 1992). Therefore, these results are consistent with the idea that nontransformed CEF contain a substantial level of SAPK activity and that pp60^v-src adds little to the activity of this kinase(s).

We did observe, however, that the PRDII-controlled promoter was highly responsive to the action of MEKK (Fig. 4A). In agreement with this observation we also found that a significant reduction in the activity of PRDII resulted from the expression of a dominant negative mutant of SEK. Interestingly, the action of this pathway on PRDII appeared to be more potent in transformed cells. It is possible that a MEKK-dependent pathway unrelated to the SEK/SAPK cascade is involved in this activation. However, to our knowledge this putative pathway has not been described. Alternatively, the transformation-specific action of MEKK may reflect the availability of a new target whose expression or localization is itself controlled by pp60^v-src. While we have shown that p50NF-B1 is part of the PRDII-binding factor of RSV-transformed CEF,⁵ we do not know presently the identity of the partner of p50 in these cells. The overexpression of IB diminished the activity of PRDII suggesting that the partner of p50 is regulated by IB. The modest effect of the activated Ha-Ras protein on PRDII raises other questions on the role of this protein in the pathway controlling the activity of the p50 complex.

A novel pathway linking p21^ras and PKC is implied by the recent findings described by Feig and co-workers (1995). Indeed these investigators reported that Ras induces the activity of Ral GTPases by interacting and activating directly Ral-GDS, the exchange factor for the Ral GTPases (Jiang et al., 1995). The activation of Ral results in the induction of a phospholipase D activity that in conjunction with a phosphatidic acid phosphatase generates diacylglycerol, a physiological inducer of PKC. Several results suggest that PKC is important in the activation of AP-1. Indeed, the addition of calphostin C reduced the activation of a TRE-controlled promoter by pp60^v-src. The same conclusion was reached when CEF were treated chronically with high concentrations of phorbol esters to deplete the cell of PKC. In these conditions the activity of the TRE was also impaired (Fig. 6). Phorbol esters stimulate AP-1 by inducing the dephosphorylation of the DNA binding domain of c-Jun, a process that is consistent with the increase in TRE binding activity observed in RSV-transformed CEF (Boyle et al., 1991; Dehbi et al., 1992). The exact mechanism of action of PKC is not known. The activity of a phosphatase may be regulated or that of a kinase responsible for the phosphorylation of the C terminus of c-Jun (Lin et al., 1992). GSKIII is a candidate for this kinase since it is phosphorylated and inactivated by PKC, and it is capable of phosphorylating the DNA binding domain of c-Jun at least in vitro (Boyle et al., 1991; Goode et al., 1992). In agreement with the critical role played by AP-1 in transformation by v-src, the inhibition of TRE activity caused by calphostin C resulted in the appearance of a more normal cell morphology (Fig. 7). Since we did not observe any effects of the drug on the pattern of phosphotyrosine-containing proteins nor on the activity of pp60^v-src as determined by IgG phosphorylation in vitro and neither on the expression of most CEF genes in RSV-transformed cells, we conclude that the action of calphostin C did not affect the activity of pp60^v-src. The simplest explanation for these observations is that calphostin C inhibited the activity of PKC thus interfering with the function of AP-1, an essential factor of v-src transformation (Granger-Schnarr et al., 1992; Lloyd et al., 1991; Suzuki et al., 1994). Several investigators have reported the activation and translocation of PKC to the membrane of v-src-transformed cells (Diaz-Laviada et al., 1990; Zang et al., 1995). Our results are consistent with a model of constitutive activation of PKC without complete down-regulation in the presence of pp60^v-src. Despite the existing evidence on the translocation of PKC, little is known about the mechanism employed by pp60^v-src to control this process. Moscat and co-workers (1990) (see also Martins et al, 1989 have suggested that the activation of a phosphatidylcholine-specific phospholipase C may be involved in the constitutive activation of PKC by pp60^v-src. However, we did not observe any change in the activity of the TRE or PRDII domain in RSV-transformed CEF treated with D609, a specific inhibitor of this enzyme (data not shown; Schutze et al., 1992). Likewise, the expression of a dominant negative mutant of PKC that appears to be stimulated by the activity of the phosphatidylcholine-specific phospholipase C (Dominguez et al., 1993) had no effect on the TRE- and PRDII-controlled promoter. Hence, it is unlikely that pp60^v-src controls the activity of AP-1 and the PRDII-binding factor by activating a phosphatidylcholine-specific phospholipase C in CEF. Separate investigators concluded that PKC can also activate c-Raf-1 directly (Kolch et al., 1993; Sozeri et al., 1992); however, others have questioned the validity of this conclusion (MacDonald et al., 1993). It is not clear at this point if PKC regulates the activity of AP-1 through a c-Raf-1-dependent pathway in RSV-transformed CEF.

While the activation of AP-1 correlated with transformation, a different picture emerged from the study of PRDII. Indeed the chronic stimulation of CEF with TPA enhanced the activity of PRDII and led to the constitutive induction of the CEF-4 promoter and accumulation of the CEF-4 mRNA (Fig. 8). In this situation the PRDII domain appeared to be solely required for the activation of the CEF-4 promoter. The study of a TRE- and PRDII-controlled promoter in transient expression assays indicated that the induction of the former is transient, whereas the latter is activated in a prolonged manner in response to a high concentration of TPA (Fig. 8C). Since CEF stimulated chronically with TPA do not display a standard transformed cell morphology and are unable to grow in soft agar, they do not appear to be fully transformed.⁴ Hence, the constitutive activation of the PRDII binding factor is not sufficient to cause transformation. We have attempted to express a dominant negative mutant of p50 NF-B1 in RSV-transformed CEF to determine whether this factor is required for transformation by v-src but have failed to obtain a conclusive answer. Indeed, the inhibition of NF-B appears to be deleterious to the cell since no expression of the mutant p50 was ever observed. In fact, the deletion of the p50 cDNA inserted in a retroviral vector was generally observed in these cells.⁶ The reason for the prolonged induction of CEF-4 and PRDII activity is unknown. Like CEF-10, a phorbol ester-repressible gene (Simmons et al., 1989), CEF-4 is expressed at a high level in conditions leading to the depletion of PKC and - (Fig. 8A). Hence it is possible that CEF-4 is regulated both positively and negatively by PKC. Since multiple isoforms of PKC co-exist in the cell, it is possible that each function is carried out by a different enzyme. It is also possible that chronic TPA stimulation leads to a prolonged translocation and activation of one of the PKC isoforms. In this case, down-regulation would not account for the induction of CEF-4.

Several questions remain on the signaling pathways and transcription factors induced by v-src transformation. Iba and his co-workers (1994) reported that c-Jun and Fra-2 are the predominant components of AP-1 in RSV-transformed CEF. Using specific antibodies we confirmed this conclusion and demonstrated that all AP-1 complexes bound on the CEF-4 TRE contain Fra-2. A large fraction of these complexes also included c-Jun although another member of the Jun family may also exist in the AP-1 complex.²). The presence of Fra-2 is intriguing because it is generally regarded as an inhibitor of AP-1 since it has a relatively weak trans-activation potential (Suzuki et al., 1994; Yoshioka et al., 1995). However, the comparison of TRE activity in TPA and v-src-transformed CEF indicated that AP-1 is equally potent in these two conditions.³ Since c-fos is induced transiently in response to TPA stimulation (thus resulting in the formation of the strong c-Jun/c-Fos dimer), it appears that a potent Fra-2-containing dimer of AP-1 functions in RSV-transformed CEF. It is perhaps not surprising that Fra-2 is expressed in these cells since the expression of this gene is controlled by AP-1 (Kovary and Bravo, 1992; Yoshida et al., 1995). Hence one might expect that Fra-2 will be expressed in cells with elevated AP-1 activity. Whether or not the activity of Fra-2 is altered (potentiated) in v-src-transformed CEF remains to be investigated.