Dissociation of Platelet-derived Growth Factor (PDGF) Receptor Autophosphorylation from Other PDGF-mediated Second Messenger Events*

Activated p21“ alters the platelet-derived growth factor (PDGF) signal transduction pathway in fibro- blasts by inhibiting autophosphorylation of the receptor as well as by inhibiting the induction of the growth-related genes c-myc, c-fos, and JE. To elucidate the cause and effect relationships between receptor autophosphorylation and other second messenger events in the PDGF signaling pathway we created revertants of v-ras transformed cells by two methods: 1) the use of cAMP analogues, and 2) the introduction of a gene, Krev-1, which has been reported previously to revert rw transformed cells to normal morphology. Analysis of the revertants shows that the PDGF-mediated tyrosine phosphorylation of the 180-kDa PDGF receptor remains inhibited; however, the PDGF-mediated activation of phospholipase C and the induction of the growth-related genes c-myc, c-fos, and JE have been restored. These data suggest the presence of parallel pathways for PDGF signal transduction which are not dependent on autophosphorylation of the PDGF receptor. dimer

cellular activities or events occur rapidly after exposure of a cell to PDGF, including dimerization and autophosphorylation of the PDGF receptor, association and activation of several proteins (including phospholipase C (PLC), phosphatidylinositol-3 kinase (PI-3 kinase), and Raf-I), increased phosphatidylinositol turnover and calcium mobilization, activation of protein kinase C, and induction of a number of growth-related genes, including c-myc, c-fos, JE, and c-jun (Kumjian et al., 1989;Meisenhelder et al., 1989;Morrison et al., 1989;Wahl et al., 1989;. A causal or sequential relationship of these PDGF-induced phenomena to each other and to eventual DNA synthesis is suspected but not well established. The roles that cellular proto-oncogenes play in this cascade of events have also not been fully elucidated. One in particular, c-ras, may be implicated in PDGF signal transduction. Microinjection of anti-ras antibody arrests cells before they can enter S phase of DNA synthesis (Mulcahy et al., 1985). We and others have shown previously that mutated and activated ras genes (EJ-ras or Kirsten v-ras) block PDGFmediated DNA synthesis in murine fibroblasts and mitogenic signal transduction in a number of other cell types (Zullo and Faller, 1988;Lichtman et al., 1986;Lichtman et al., 1987). Fibroblasts containing an activated ras gene demonstrate an inhibition of PDGF-induced autophosphorylation of its receptor (Rake et al., 1991), diminished PI turnover (Parries et al., 1987), diminished PLC activity (Benjamin et al., 1987), diminished calcium mobilization (Benjamin et al., 1988), and inhibited induction of c-myc, c-fos, and J E (Zullo and Faller, 1988). We have therefore utilized cells containing activated ras genes as "mutants" in PDGF signaling to study second messenger systems in the PDGF mitogenic pathway. The ability to create revertants of the ras transformed phenotype then permits determination of which of the phenomena associated with PDGF binding revert as well and may thus be causally related.
Exposure to cAMP phenotypically reverts ras transformed cells to a more normal morphology (Anderson et al., 1974;Carchman et al., 1974) and leads to partial return of PI turnover and calcium mobilization after PDGF stimulation (Olinger et al., 1989). Kitayama et al. (1989) have isolated a gene, Krev-I, which encodes a 21-kDa protein possessing homology to the ras family and which reverts ras transformed cells to a flat morphology. We created new revertants of Kirsten v-ras transformed Balb/c 3T3 cells (KBalb) by two methods: 1) treatment of KBalb cells with CAMP; or 2) transfecting the Krev-1 DNA into KBalb cells. The revertant cells produced by either method possess morphologic and growth characteristics similar to the untransformed Balb/c 3T3 (Balb) fibroblasts.
In the revertant cells partial reconstitution of the PDGF signaling pathway is demonstrated. In response to PDGF, increases in PLC activity as measured by phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis, mobilization of calcium, and the induction of c-myc, c-fos, and JE were restored whereas the PDGF-mediated autophosphorylation of the PDGF receptor remained blocked. These observations demonstrate a dissociation between autophosphorylation of the PDGF receptor and the activation of a number of presumed second messengers of the PDGF signaling pathway.

MATERIALS AND METHODS
Cells, Cell Lines, and Cyclic AMP Treatment-Balblc 3T3 fibroblasts were obtained from the American Type Culture Collection.
Kirsten v-ras transformed Balb/c 3T3 fibroblasts (KBalb) were created as described previously (Zullo and Faller, 1988). Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% donor calf serum (GIBCO). The cell lines were treated with cyclic AMP by addition to the media of 2 mM Nfi-2'-O-dibutyryladenosine 3':5'-cyclic monophosphate (Sigma) for 48 h. As a control for the cyclic AMP studies, cells were also treated in 4 mM sodium butyrate to control for the release into the cells of 2 mol of butyrate for every mol of cyclic AMP.
Transfection of Kreu-1 into KBalb-A plasmid containing both the Kreu-1 gene and the ne0 gene with each driven by an SV40 early promoter was the generous gift of Yoji Ikawa, Tsukuba Life Science Center, Japan. The plasmid was amplified by transformation into Escherichia coli HBlOl (Maniatis et al., 1982), purified, and used directly for electroporation into Balb and KBalb cells. After electroporation the cells were selected in Dulbecco's modified Eagle's medium containing 0.5 mg/ml G418 (Geneticin, GIBCO). After several weeks, isolated colonies containing clones of resistant cells were identified. Colonies with the flattest morphology were picked and expanded and then maintained in medium containing 0.5 mg/ml G418.
Cell Proliferation Assays-I X lo5 cells/well were plated in medium plus 10% donor calf serum into six-well tissue culture plates. To some wells dibutyryl cAMP was added to 2 mM. After 24 h the medium in some wells was changed to 0.5% serum. The number of cells in each well (in duplicate wells for each time point) was enumerated on days 0, 2, 4, and 6. I:'H]Thymidine Incorporation-5 X 10' cells/well were plated into 24-well tissue culture plates. At confluence cells were starved for 24 h in medium containing 0.5% donor calf serum, with some wells containing dibutyryl cAMP at 2 mM. After 24 h [3H]thymidine (Du Pont-New England Nuclear; specific activity, 20 Ci/mmol) was added to 5 pCi/ml. To some of the plates recombinant PDGF-BB (Amgen) or serum was added to final concentration of 10 ng/ml or IO%, respectively, and incubation at 37 "C was continued. After 24 h plating medium was removed, wells were washed twice with cold phosphatebuffered saline, and Triton X-100 was added to 1% to the wells to lyse the cells. Nucleic acids were harvested onto glass fiber filters, and incorporated thymidine was quantitated by scintillation counting. Each cell type and condition was tested in quadruplicate.
Phospholipase C Activity-Phospholipase C activity was measured in quiescent cells or cells stimulated with 40 ng/ml PDGF-BB by quantitating PIP, hydrolysis in the membrane fractions using the method of Jackowski et al. (1986). Each cell type was quantitated in quadruplicate.
Phospholipase C activity was also measured in anti-phosphotyrosine immunoprecipitates of quiescent cells or cells stimulated with 40 ng/ml PDGF-BB, using the technique of Wahl et al. (1988).
Calcium Mobilization-Cells were starved for 24 h in medium containing 0.5% donor calf serum. The spent medium was removed from the cells and saved. Cells were removed from the tissue culture plate with warm Versene (GIBCO), pelleted by centrifugation, resuspended in the spent medium, and incubated for 15-30 min at 37 "C. The cells were then pelleted and resuspended at 10' cells/ml in a buffered salt solution containing glucose and 1 mM INDO-1, incubated at room temperature for 20 min to allow uptake of the INDO-1, and washed once in 50 volumes of balanced salt solution to remove excess dye. Calcium mobilization in response to recombinant PDGF-BB was then determined using a stirred cuvette at 37 "C in an Aminico fluorometer with real-time recording at 405 and 480 nm.
Phospholipase C and Phorbol Ester Treatment of Balb and KBalb Cells-Phospholipase C (PLC) from Bacillus cereus was obtained as a suspension in 3.2 M ammonium sulfate, pH 6.0 (Boehringer Mannheim). The PLC was added to serum-starved cell cultures to a concentration of 2 units/ml. An equal volume of protein-free ammonium sulfate solution was added to control cultures. A final concentration of PLC greater than 2 units/ml or of ammonium sulfate greater than 45 mM was each toxic to the cells. Cells were treated for 60 min at 37 "C, and RNA was extracted and analyzed to assess the response of the growth-related genes to the PLC treatment.
Phorbol 12-myristate 13-acetate (PMA) was added to serumstarved cells to a final concentration of 200 ng/ml and incubated for 60 min at 37 "C. RNA was then isolated and transcripts quantitated by RNA blot analysis.
RNA Blot Analysis of Total Cellular RNA-Total cellular RNA was isolated by extraction with guanidine thiocyanate, separated on formaldehyde-agarose gels, and transferred to nitrocellulose or supported nitrocellulose membranes. RNA blot analysis was carried out as described previously (Offermann and Faller, 1989). Transcript levels were determined by scanning the autoradiographs with a Hewlett-Packard ScanJet, Macintosh IIcx, Image 1.2.
[3' P]DNA Probes-"P-Labeled DNA probes were made by the random oligonucleotide primer method (Feinberg and Vogelstein, 1983). The probes specific for c-myc exon 2 and c-fos have been described previously (Zullo and Faller, 1988). The JE probe was a 0.75-kilobase PstI fragment containing exons 3 and 4. The ATPasespecific probe was a PstI 3.7-kilobase fragment of a rat sodium potassium ATPase cy1 subunit cDNA (gift of Russell M. Medford, Emory University). Kreu-I mRNA was detected with a 2.0-kilobase BamHI-PstI fragment of the Kreu-1 plasmid used for transfections.
Solubilized Cell Kinase Assay-The various cell lines were subjected to tyrosine kinase assay by solubilization of membrane proteins as described previously (Rake et al., 1991). In brief, cells were grown to confluence and the media exchanged for methionine-free medium (GIBCO) containing ["S]methionine (Du Pont-New England Nuclear) at 200 pCi/ml for 4 h at 37 "C. The cells were then washed twice with phosphate-buffered saline, twice with the isotonic wash (120 mM KC1, 30 mM NaC1, 2 mM MnC12, 10 mM HEPES, pH 7.41, overlaid with solubilization buffer (isotonic wash, 0.1% Triton X-100, and 2 mM phenylmethylsulfonyl fluoride) at 750 pl/lOO-mrn plate, incubated at 0 "C for 5 min, and then scraped into 15-ml conical tubes. An additional 750 pl was used to rinse the plate, and the nuclei were pelleted. The supernatant was kept on ice. A Bradford protein assay kit (Bio-Rad) was used to determine the protein concentration. To 100 pg of solubilized protein were added 10 ng/ml recombinant PDGF-BB and 0.4 p~ ATP, and the reaction was incubated for 5 min on ice. The reaction was quenched with the addition of lysis buffer (1% Triton X-100, 10 mM Tris, pH 7.6, 5 mM EDTA, 50 mM NaC1, 30 mM Na,P2O7, 50 mM NaF, 0.1 mM Na3V04, 0.1% bovine serum albumin, 0.1% NaN3, 2 mM phenylmethylsulfonyl fluoride), and the resultant mixture was subjected directly to anti-phosphotyrosine immunoprecipitation.
Digitonin-permeabilized Tyrosine Kinase Assay-In a modification of the technique described by Erusalimsky et al. (1988) cells were grown to confluence in 35-mm wells and then starved in 0.5% calf serum for 24 h. After washing twice with phosphate-buffered saline and twice with isotonic wash (as above) the cells were overlaid with a permeabilization buffer consisting of isotonic wash, 40 pM digitonin (Sigma), 1 ~L M ATP, [Y-~'P]ATP (Du Pont-New England Nuclear) at 50 pCi/ml, with or without recombinant PDGF-BB at 10 ng/ml, and incubated at 37 "C for 5 min. The radioactive overlay was discarded, and the cells were collected in 600 pl of lysis buffer (as above). After centrifugation the pellet of debris was removed and the supernatant used for anti-phosphotyrosine immunoprecipitation.
Zmmunoblot Detection of PDGF Receptor-associated Tyrosine Phosphorylation-Cells were grown to confluence in 100-mm culture dishes. Monolayers were washed twice with phosphate-buffered saline, then l ml of HTG buffer (20 mM HEPES, pH 7.2, l% Triton X-100, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na,VO,) was added, the cells were dislodged with a cell scraper, the lysate was transferred to a microcentrifuge tube, cell debris was pelleted, and the specimens were normalized for protein content. The lysate was precleared by incubation with 100 p1 of a protein A-Sepharose (Sigma) suspension consisting of 250 mg of beads/4 ml of HTG buffer for 1 h at 4 "C, and 2 pl of antibody to PDGF type B

140.i7
receptor (gift o f Tom Daniel, Vanderhilt Universitv) was added to the mixture. which was rocked for 2 h at 4 " C . The receptor antihodv complexes were precipitated hy incubation with 100 pI o f protein A-Sepharose for I h. The heads were heated to 100 "C in the presence o f S I X , and equal protein amounts of the samples were run on a 5-1.5";' SDS-I'A(;l< and transferred electrophoretically to nitrocellulose. 1 yrosine phosphorvlation of the receptor was then detected bv Westr r n hlot with anti-phosphotvrosine antihody (gift of Tom lioherts. I h n a Farher Cancer Institute).

I n V~V O
Kinnsc Assny-The cells were grown to confluence in 35mm wells. The medium was exchanged for methionine-free medium with 200 pCi/ml [ '"S]methionine and incuhated for 4 h at 37 "C. T h e cells were then washed twice with phosphate-huffered saline, twice with isotonic wash, and then overlaid with stimulation huffer (isotonic wash, 1 0 ng/ml recornhinant PDGF-HR), and incuhated for 5 min at 37 "C while the cells were still adherent to the culture plates. The In~l'fer was removed, the cells lvsed with 800 pI of lvsis huffer (as d m v e ) , and the mixture transferred to l.5-ml Sarstedt screw-cap tubes for suhsequent anti-phosphotvrosine immunoprecipitation.

RESULTS
Morphologic a n d Growth Characterktics of Normal, Transformed, and Revertant Cells-The characteristic morphology of Ralh/c 3T3 fibroblasts at confluence is t h a t of flat appearing cells that demonstrate contact inhibition. Kirsten v-rascontaining fibroblasts (KRalh), in contrast, appear small and refractile with extended processes and overgrow the monolayer at confluence (Fig. 1). Revertants in the transformed phenotype were created by transfection with the Kreu-I cDNA and selection of clones resistant to G418. T h e m R N A analysis of these clones (called KR) demonstrated the presence of Kreu-1 transcripts (Fig. 2). Alternatively, the cells were treated with 2 mM dihutyryl CAMP for 48 h. In  the morphology of hoth of these types of revertants was compared with the parental KRalb cell line. Neither of these revertants approached the appearance of the untransformed Ralb cells; however, they were clearly flatter than their parental cell line, KRalh. They also demonstrated contact inhibition. The changes noted in the CAMP-treated cells became apparent after exposure for 24-48 h to 2 mM dihutvnl cAMJ'. Neither the CAMP treatment nor the Krrc-1 transfection affected the levels of v-ras as detectable bv mRNA analysis (not shown). Therefore, the changes noted in the two types of revertants were not caused bv the trivial explanation of having reduced the levels of rus in the cells.

Serum reyuircmcnts o f res-cnntnininz Hnlh rclls and
The serum requirements of these revertant cell lines was assessed. After growth in 0.5T calf serum or IOr; calf serum for a predetermined number of davs, the increase in cell numher was determined (Table I) like the (untransformed) Balb cells in that they showed no growth in 0.5% serum, demonstrating a restoration of their serum requirements paralleling their morphological reversion. DNA Synthetic Response of ras-containing Balb Cells and Their Revertants to PDGF or Serum-The ability of the normal Balb cells, the v-ras transformed cells, and the CAMPtreated or Kreu-1 -transfected revertants to respond to PDGF-BB mitogenically was assessed by thymidine incorporation. Tritiated thymidine incorporation over 24 h in Balb cells increased approximately 20-fold in response to 10 ng/ml PDGF-BB (Table 11). No such increase in response to PDGF-BB was observed in the KBalb cells (which already possessed a high basal incorporation) although treatment with 10% serum did lead to a small increase. Treatment of KBalb cells with CAMP or the introduction of the Kreu-1 gene produced dramatic results. The basal level of thymidine incorporation during serum deprivation fell, approaching the low level found in the Balb cells (Table 11). However, no significant increase in thymidine incorporation was observed in response to PDGF-BB. As a positive control, treatment with 10% serum resulted in a 2-10-fold increase in incorporation for the Krev-1 -transfected cells and the CAMP-treated cells, respectively (compared with a 52-fold increase seen in the Balb cells). This minimal response to serum in the revertants demonstrates that the cells are still capable of responding to growth factors in serum other than PDGF.
Phospholipase C Activity in Response to PDGF-BB-PLC activity in membrane preparations of our cell lines was examined. PIPz hydrolysis increased 25% in the normal Balb cell membranes upon treatment with PDGF-BB (Fig. 3A). Membranes from v-ras-infected KBalb cells demonstrated a 16% decrease in PLC activity in response to PDGF-BB (Fig.  3A). The differences between the Balb cells and the KBalb cells were highly significant ( p < 0.05). Membranes from the revertants, as represented by the KBalb/cAMP in Fig. 3A, had an increase of 24% in their PLC activity upon PDGF-BB stimulation, a statistically significant difference ( p < 0.05) from the decrease seen in the KBalb cells.
We also examined the PLC activity that was recoverable in phosphotyrosine immunoprecipitates of quiescent and stimulated cells. In these preparations, PIP2 hydrolysis increased 28% in the normal Balb cells upon treatment with PDGF-BB ( Fig. 3B; p < 0.05). Similar to the results obtained using the membrane preparations, immunoprecipitates from the v-ras- FIG. 3. Activation of PLC after PDGF stimulation is restored in the revertants. Panel A, PLC activity is represented as cpm of tritiated inositol 1,4,5-trisphosphate ( P a ) released after incubation of tritiated PIP, with quiescent or stimulated membranes isolated from Balb, KBalb, and KBalb/cAMP cells. The values shown represent the PLC activities demonstrable in the basal state and after stimulation with 40 ng/ml PDGF-BB. In response to PDGF, PLC activity in the Balb cells increased by 25%; in the KBalb cells, PLC activity decreased by 16%; and in the revertants, PLC activity increased by 24%. These differences were significant (p < 0.05). Panel B, PLC activity was measured as above but in anti-phosphotyrosine immunoprecipitates from Balb, KBalb, KBalb/cAMP, and KR6-5 cells. In response to PDGF, PLC activity in the Balb cells increased by 28%, and in the KBalb cells, PLC activity decreased by 44%. The KBalblcAMP revertants showed a 17% decrease after PDGF-BB stimulation whereas the KR6-5 revertants showed a 24% increase. This 24% increase in the KR6-5 cells was significant compared with the 44% decrease seen in the KBalb cells ( p < 0.05).
infected KBalb cells demonstrated a 44% decrease in PLC activity in response to PDGF-BB (Fig. 3B). Phosphotyrosine immunoprecipitates from the KBalb/cAMP revertants demonstrated a 17% decrease in phosphotyrosine-associated PLC activity after PDGF-BB stimulation in contrast to the 24% increase seen in the membranes. This decrease seen in the phosphotyrosine immunoprecipitates from the CAMP-treated cells was not significantly different from the decrease seen in the untreated KBalb cells (Fig. 3B; p not significant). Immunoprecipitates from the Kreu-1 -containing revertants KR6-5 demonstrated a significant increase of 47% in PLC activity after PDGF-BB stimulation ( Fig. 3B; p < 0.05).
Phosphatidylinositol-3 Kinase Activity in the Revertants-Assays for phosphotyrosine-associated PI-3 kinase activity utilizing the technique of Kaplan et al. (1986) demonstrated a significant increase in PI-3 kinase activity in the Balb cells upon stimulation with PDGF-BB. KBalb cells contained a higher basal PI-3 kinase activity than did Balb cells, and this activity did not change upon PDGF-BB stimulation. In both types of revertants (the CAMP-treated KBalb cells and the Krev-1 -containing revertants), the PI-3 kinase activity was returned to the lower basal level seen in the Balb cells, but there was no significant change seen after treatment with PDGF-RR (data not shown).
Calcium Mobilization in Response to PDGF-RR-A technique for examining stimulus-dependent calcium mobilization in fibroblast cells in suspension was developed to permit analysis of the response of an entire population of cells rather than relying on the pooling of single cell analyses. Basal fluorescence of the INDO-1-loaded cells was identical for each cell type, indicating that intracellular pools of calcium did not differ in the presence of the v-ras or Kreu-1 genes. The presence of v-ras markedly inhibited the mobilization of calcium which is normally observed in Balb cells in response to PDGF-BR, decreasing it by 80-9296 (Fig. 4). Transfection of KBalb cells with Kreu-1 resulted in substantial restoration of the intracellular calcium response to PDGF-BB, to 80-8.796 of the RaIh levels of calcium mobilization (Fig. 4). Both the characteristic 5-s delay between the addition of PDGF-RR and the beginning of the calcium flux and the distinctive shape of the INDO-1 profile were restored in the Kreu-Icontaining cells (tracings not shown). Cyclic AMP-treated KRalb cells were not evaluable by the method used, as the pretreatment with CAMP inhibited loading with the indicator dye.
PDGF-RR-mediated Induction of Growth-related Genes-We have shown previously that in v-ras-containing cells, PDGF-RR stimulated induction of the growth-related genes c-myc, c-fos, and J E was inhibited. This inhibition was at the level of transcription and was apparent within 30 min of vras introduction into the cell (Zullo and Faller, 1988). Although uninducible by PDGF-RR, c-myc mRNA expression appeared to become somewhat deregulated when Ralb cells transformed by v-rm were carried over many generations, such that basal steady-state levels of c -m y mRNA are higher in serum-deprived v-ras-containing cells than in the untransformed parental cell line.
The revertants were subjected to mRNA analysis to examine the transcript levels of these genes, before and after PDGF-BR stimulation, in comparison with the parental lines. Ralb cells showed a strong induction of c-fos transcripts by :?0 min aft.er treatment with PDGF-RB, whereas c-fos tran-scripts in the PDGF-treated KRalb cells were only detectable on a longer exposure and were at least 100-fold less than the untransformed cells by quantitation using densitometric scanning (Fig. 5 , top). The revertant cell lines (KR-1 or KHalb/ CAMP) demonstrated strong induction of c-fos mRNA in response to PDGF-RR, with the level of mRNA induction approaching that of normal Ralh cells (Fig. 5 , top). In the Kreu-1-containingcells (KR-l), the level of induction of c-fos mRNA was dependent on the cell density; that is, c-fos transcripts in more confluent cultures were somewhat less inducible (not shown). The CAMP-treated cells demonstrated full induction of c-/os regardless of cell density.
The KRalb cells had an undetectable base-line J E mRNA level and a very low level of induction after PDGF-RH stimulation when compared with the Ralb cells (Fig. 5 , hottom).
In both the Kreu-]-containing revertants and in the CAMPtreated KRalb cells, there was a strong increase in the level of J E mRNA seen after PDGF-HR stimulation, but there was also a rise in the basal (unstimulated) level. Some constitutive expression of c-myc mRNA was ohserved in the serum-deprived KRalh cells, but this expression was unchanged in response to PDGF-RH (Fig.  5 , hottom). The Kreu-I-containing revertants (KR-1 or KR-6) and the CAMPtreated revertants demonstrated two significant changes in comparison with the KRalb cells: 1) reduction of the basal levels of transcripts for c-mvc in serum-deprived cells; and 2 ) restoration of PDGF-RR inducibility. The KR-and CAMPtreated cell lines had c -m y mRNA levels that were undetect- K H a h rells had a reduction in the calcium mohilization in response to I'IXF-HH to 8 and 20% of the levels seen in the Ralh cells at doses of 5 and 40 ng/ml, respectively. In  revertants levels of calcium mobilization in response to PDGF-HH were restored to 83 and 80rE of those seen in the Halt) cells at 5 and 4 0 ng/ml, respectively. Cyclic AMP-treated KRalh cells were not evalunhle by the method used. as the pretreatment with CAMP int1it)ited londing with the indicator dye. able in the quiescent unstimulated state.
In response to PDGF-RH, c-mvc transcripts were induced to levels comparable to t,hose found in the PDGF-RR-stimulated Balb cells (Fig. 5 , bottom).

h'xogenous PLC or I' MA Induces c-myc Transcripts in ras-
containing Halh Cells-Calcium immobilization and induction of the growth-related gene c -m y by PDGF are generally considered to be events dependent upon activation of PLC. Since these two events are also inhibited in v-ras-containing cells the possibility existed that. the inhibition of PDGF-RH signal transduction in the KHalh cells was at the level of PLC activation. To test this possibility serum-starved KRalb cells were treated directly with PLC or with PMA, a known activator of protein kinase C.
The results of RNA blot analysis showed that both PLC and PMA were able to increase the mRNA levels of the cmyc transcripts in the KRalb cells (Fig. 6). These results suggest that the block to PDGF-RR signal transduction in the v-ras-containing cells is a t or before the level of PLC activation and the protein kinase C-mediated components of the signal cascade. We have shown previously that t,here is a hlockade at the level of receptor phosphorylation in v-rascontaining cells (Rake et al., 1991). This new result demonstrates that the signaling pathway distal to the activation of PLC is still intact in the KRalh cells.

PDGF Receptor Autophosphoplation Remains Blocked-
Ralb cells demonstrate autophosphorylation of the PDGF receptor when stimulated by PDGF-RR. When Ralb fihroblasts were labeled metabolically with [%]methionine and the cytoplasmic membranes solubilized in 0.1% Triton X-100, the autophosphorylated receptor could be immunoprecipitated with an anti-phosphotyrosine antibody and identified as the hand migrating at 180 kDa (Fig. 7, top). Under these same conditions autophosphorylation of the PDGF receptor in KRalh cells was not demonstrable (Rake et al., 1991). Furthermore, the revertants obtained after exposure of KHalb fibroblasts to 2 mM dihutyryl CAMP for 48 h did not tyrosine phosphorylate the PDGF receptor after stimulation with PDGF-RR (Fig. 7, top). The Kreo-I-containing revertants (KR) demonstrated the same inhibition to tyrosine phosphorylation of the PDGF receptor in [:"S]methionine-labeled cells (not shown), and this was confirmed further in the Kreo-I -containing revertants utilizing different assay conditions: digitonin permeabilization, labeling with [y-:'I'P]ATP, and immunoprecipitation with the same anti-phosphotyrosine antibody (Fig. 7, bottom).
A second independently isolated and distinct anti-phosphorinw c s p,c OMA I NOkIF cz c PQlA FIG. 6. E x o g e n o u s I'1.C or P M A i n d u c e s c-myc t r a n s c r i p t s i n r a s -c o n t a i n i n g Ihlh cells. Chlluent Ihlh and v-ros-containing Hall) (KI3nlt)) C C I I~ w c r r serum starved (0.5"; serum) for 24 h and then tretlted with eitl1c.r 2 units/ml purified H. crrrus PIX, 20 ng/ml I'MA. 10''; calf serum (CY), or no addition (nonr) for 90 min, at which time the cells were lysed with kwanidine thiocyanate and their RNA purified. The total cellular RNA (20 cg/lane) was size fractionnted on a formaldehyde-agarose gel, transl'erred t o a nitrocellulose filter, and the filter hyl)ridized simultaneously with radiolaheled prohes for c -m y a n d A'l'l'ase (as a control). The addition of 10' ; calf serum was used as a positive control for induction of c-mvc levels in KHaIl). An autoradiogram is shown here. There were significant increases in the Ievels of c-myr transcripts in hoth the Ihlh and KI3aIt) cells in response to both P I X and I'MA. tyrosine antibody was used to immunoblot lysates that had been immunoprecipitated with an anti-PDGF receptor t.ype d antibody. As can he seen in Fig. 8, the stimulation of h l b and CAMP-treated Ralb cells with PDGF-HR resulted in strong tyrosine phosphorylation of the 180-kDa receptor. A greater than 10-fold inhibition of the receptor phosphorylation is seen in the KRalh cells, and there is no demonstrahle reversion of this inhihition in the KRalb/cAMP-treated or Kreo-1 -containing (KR-6) cells (Fig. 8). These results confirm the identity of the 180-kDa hand and provide an alternative system for demonstrating receptor autophosphorylation or its inhibition. Tyrosine phosphorylation of other cellular proteins in response to PDGF-RR could he detected in the anti-phosphotyrosine blots of PDGF receptor type [f immunoprecipitates of Ralb cells (Fig. 8) (-), lysed, and immunoprecipitated with anti-I'DGF receptor type @ antibody. Equal amounts of protein were loaded into each well, electrophoresed on polyacrylamide gels, and transferred to nitrocellulose. Probing with an anti-phosphotyrosine antibody showed significant increases in receptor tyrosine phosphorylation in the Ralb (lanes I and 2) and Balb/cAMP cells (Innes 5 and 6 ) whereas autophosphorylation of the PDGF receptor in the KBalb cells was inhibited greatly (lanes 3 and 4 ) . In neither of the revertants, KBalb/cAMP-treated (lanes 7 and 8 ) or Kreu-l-containing (KR-6, lanes 9 and IO), was the inhihition reversed. Also noted were proteins at 85, 74, 60, and 42 kDa, which after tyrosine phosphorylation appeared to have co-immunoprecipitated with the PDGF receptor in the Balb (lane 2 ) and the Ralb/cAMP (lane 6 ) cells but not in the other cells. tyrosine phosphorylation of protein bands a t 145, 100,85, 74, and 42 kDa (Fig. 9). No unequivocal identification of these other proteins seen in Figs. 8 and 9 has been established, but the molecular masses of the 145-, 85-, 74-, 60-, and 42-kDa bands suggest that they may be the previously identified substrates of membrane-associated receptor tyrosine kinase activity: PLC, PI-3 kinase, Raf-1, src, and pp42, respectively (Meisenhelder et al., 1989;Kaplan et al., 1987;Morrison et al., 1989;Kypta et al., 1990;Rossomando et al., 1989;. Phosphorylation of these presumptive receptor kinase substrates in response to PDGF-BB was never observed in the KBalb cells and was not restored in the cAMP or in the Kreu-1 revertant cells. Of note, the inhibition of tyrosine phosphorylation of the 100-kDa protein in response to PDGF was an inconstant finding in that occasional experiments did demonstrate PDGF-BB-stimulated phosphorylation of the 100-kDa protein in the KBalb cells using the in vivo method.

3-C A T P a s e rcl)w c -r n y c
None of the other PDGF receptor substrate proteins found in the KBalb cells or in the revertants demonstrated any significant tyrosine phosphorylation under either of the conditions shown in Figs. 8 and 9.

DISCUSSION
The presence of activated ray proteins in the cell blocks the PDGF-BB-mediated activation of the suspected second messenger components of the signal transduction pathway, including PDGF receptor autophosphorylation, phosphorylation of the PDGF receptor-associated substrates, activation of phospholipase C, calcium mobilization, and induction of the growth-related genes. These potential second messengers were initially thought to be ''linked,'' consistent with a linear model of signal transduction in which one event activates the next. If this model were correct then inhibition of any point in this signaling cascade by v-ras would block activation of all of the subsequent components of the pathway; accordingly, reversion of this blockade would result in reconstitution of all the second messenger elements. Alternatively, if more than one distinct and independent signal are generated by the binding of PDGF-BB to its receptor (a branching model for the signal transduction cascade) then revertants of ray transformed cells might demonstrate reconstitution of certain signaling components while others would remain blocked. Increasing indirect evidence has been accumulating which suggests that the presumed second messengers of PDGF signaling may be acting independently of one another, consistent with a branching model for signal transduction . A comparison of the v-ras-containing cells we describe here with the morphological revertants produced by cAMP or Kreu-1 provides further evidence for a branching model by demonstrating uncoupling among some of the second messengers and accordingly aids in the assignment of specific second messengers to one or another arm of such a model.
The mechanisms by which Kreu-1 or cAMP induce morphologic reversion have not yet been elucidated. The Kreu-1 gene encodes a 21-kDa ray-like protein that is identical to the previously reported smg21 and rap1 and which reverts v-ras transformed cells to a more normal morphology (Kitayama et al., 1989;Pizon et al., 1988). Cyclic AMP causes a similar reversion (Carchman et al., 1974;Anderson et al., 1974) and has been documented to reverse a t least partially the inhibition by v-ray of PDGF-BB-induced calcium mobilization and PI turnover (Olinger et al., 1989). The Kreu-1 transfectants and the CAMP-treated KBalb cells we describe have growth characteristics and serum requirements that approach those of the normal Balb fibroblasts; yet their mitogenic response to PDGF-BB, as measured by thymidine incorporation, is clearly not restored, and it is evident that only certain of the PDGF second messenger signals which were blocked by v-ras have been reconstituted in the revertants.
The PDGF-BB-mediated induction of c-myc, c-fos, and JE transcripts was restored substantially in the revertant cells. One point is noteworthy regarding the c-myc mRNA levels. In the revertants there was not only a restoration of the induction of c-myc mRNA in response to PDGF-BB, but there was also a reversal of the deregulation of c-myc mRNA levels which is characteristic of chronically v-ras-infected cells. Deregulation of c-myc expression resulting in high constitutive c-myc mRNA levels is a phenotype common to many transformed cell lines and naturally occurring tumors and is thought to contribute to tumorigenesis (Bading and Moelling, 1990;Forgue-Lafitte et al., 1989;Zajac-Kaye and Levens, 1990;Kakkis et al., 1989;Suchy et al., 1989). In our hands, Balb cells containing activated ras genes also have this phenotype. It may be significant that the morphological revertants also demonstrate reversal of this deregulated phenotype.
The mechanism for PDGF-mediated induction of c-myc, cfos, and JE transcripts may be independent of the PDGF receptor kinase activity, or at least of receptor autophosphorylation, since in neither type of revertant was there significant ligand-induced tyrosine phosphorylation of the PDGF receptor; nor was there a restoration of a PDGF-mediated increase in phosphatidylinositol-3 kinase activity in either revertant. This kinase-independent restoration of growthrelated gene induction may be mediated by PLC, as the ability of PDGF to induce both growth-related gene transcription and PLC activation reverted coincidentally in the Krev-land CAMP-treated cells. The revertant cells demonstrate a clear restoration of PDGF-inducible PLC activity by direct measurement of PIP, hydrolysis, and by calcium mobilization, which is presumed to be a PLC-dependent phenomenon. PDGF-inducible PLC activity was restored in these cells without any detectable ligand-stimulated receptor tyrosine kinase activity.
In view of current speculation regarding the role of receptor tyrosine kinase activity in growth factor signaling and our finding of a restoration of PDGF-mediated induction of both PLC activity and growth-related gene transcription, it was striking that the kinase activity of the PDGF receptor was not restored in the Kreu-1 or dibutyryl CAMP-treated revertants. The mechanism of the delivery of a PDGF-mediated signal resulting in the induction of c-myc, c-fos, and JE in the absence of phosphorylation of the receptor remains undefined. There may be a conformational change in the PDGF receptor which is sufficient to activate some other second messenger pathway distinct from the PDGF receptor tyrosine kinasedependent pathway and which results in the activation of the growth-related genes. This tyrosine kinase-independent pathway would then potentially represent a branch point in signal transduction for PDGF. Both the kinase-independent and kinase-dependent arms of the pathway appear to be necessary for the full delivery of the proliferative signal to the cell, yet each arm could be dissociated from the other one under conditions like those described here.
We propose that the PDGF-receptor kinase-independent arm of the signaling pathway might be mediated by activation of phospholipase C. We have demonstrated substantial restoration of PIP, hydrolysis and calcium mobilization in response to PDGF-BB in the Kreu-1-or CAMP-induced revertants, demonstrating reactivation of PLC activity. Furthermore, we have shown that treatment of Balb or KBalb cells with phorbol esters or exogenous phospholipase C results in a pattern of growth-related gene expression identical to that seen in PDGF-treated Balb cells, which would indicate that events distal to the activation of PLC are not directly inhibited in the v-ras-containing cells. It is possible that the PDGF receptor transmits a signal to phospholipase C via a conformational change that is sufficient to lead to the induction of the genes c-myc, c-fos, and JE (and c-jun'). Our in vivo kinase assays did not demonstrate increased tyrosine phosphorylation of the 145-kDa protein. When the PLC activities in the phosphotyrosine fractions of soluble cell lysate were examined no consistent pattern for PLC activity could be demonstrated among the revertants despite a consistent restoration of PLC activity in both types of revertants. These findings suggest that changes in tyrosine phosphorylation of PLC may not be the mechanism for its activation in these particular revertants. However, until the phosphorylated state of receptorassociated and unassociated PLC in the revertants is examined more directly phosphorylation of PLC by the receptor or by other activatable kinases in the revertants cannot be ruled out. Nonetheless, it is likely that the tyrosine kinase activity of the PDGF receptor itself is necessary for the initiation of certain other signaling pathways in the cell which together with the non-kinase-dependent signals are required for DNA synthesis and cell growth.
It is noteworthy that induction of growth-related gene expression, including c-myc, by PDGF-BB in the revertants is not sufficient for stimulation of DNA synthesis. This is unexpected in light of previous reports that microinjection of c-myc protein is sufficient to move cells into S phase of the cell cycle (Kaczmarek et al., 1985). There are a number of trivial explanations for these divergent findings, including dose effects, e.g. the amount of c-myc protein delivered by microinjection is supraphysiologic whereas the induction of c-myc mRNA we noted in the revertant cells is likely to be of the same magnitude as that seen in the normal Balb cells. Yet, our findings indicate the likely presence of at least two signaling pathways induced by ligand binding to the PDGF receptor which remain distinct beyond the induction of the immediate early genes and which together must be activated for effective growth factor-mediated DNA synthesis.
In summary, we have utilized revertants of v-ras-containing (PDGF-unresponsive) fibroblasts induced by either treatment with CAMP or transfection with Kreu-1 to study the PDGF signaling pathway. These cells did not revert the block in autophosphorylation and presumed kinase activity of the PDGF receptor but have demonstrated restoration of the PDGF-BB-mediated increase in PLC activity and induction of c-myc, c-fos, and JE. This provides evidence that autophosphorylation of the intact PDGF receptor can be dissociated from PLC activation and the induction of the growthrelated genes. These findings may suggest the presence of a PDGF receptor tyrosine kinase-independent portion of the PDGF signal transduction pathway, a portion that needs to be coupled with the kinase-dependent portion for the full delivery of the signal into the cell.