Differences in Substrate Specificities of (X and /3 Platelet-derived Growth Factor (PDGF) Receptors CORRELATION WITH THEIR ABILITY TO MEDIATE PDGF TRANSFORMING FUNCTIONS*

Recombinant expression of either the a or B platelet- derived growth factor (PDGF) receptors in 32D he- matopoietic cells allows efficient coupling of PDGF with mitogenic and chemotactic signaling pathways inherently expressed by those cells. PDGF-BB stimulation of 32D-aR or BR cells results in anti-P-Tyr recovery of cellular proteins possessing similar as well as distinct phosphotyrosine signals. Comparison of the ability of each receptor to couple with known second messengers revealed that both receptors associated with and/or tyrosine phosphorylated phospholipase Cy (PLCy) and phosphatidylinositol 3-kinase (p85) with similar stoichiometry. However, the B platelet-derived growth factor receptor (PDGFR) was significantly more efficient at in vivo tyrosine phosphorylation of GTPase-activating protein (GAP). Similar differences in binding affinity for GAP were observed in NIH/3T3 cells which express both receptors. To quantitate the affinities of each receptor for GAP or PLC-y, we utilized baculovirus-expressed a and B PDGFRs purified by anti-P-Tyr affinity chromatography. Exposure of immunoblots containing bacterially expressed GAP or PLC-y to activated a or B PDGF receptors led to a comparable high affinity binding of each receptor to PLCy, transfection of 32D cells was performed by electroporation (31). Mass populations of stably transfected cells were selected by their ability to survive in growth medium containing mycophenolic acid (32). Transfection of NIH/3T3 fibroblasts was performed by the calcium-phosphate precipitation method (33) using 40 pg of calf thymus DNA and serial 10-fold Transformed foci were transfectants were for medium mycophenolic and Immunoprecipitation Analysis-32D cells were washed twice in Dulbecco's modified Eagle medium and incubated at 37 "C for 2 h in serum-free medium. The quiescent 32D cells were then stimulated with PDGF-BB (Upstate Biotechnology, Inc.) (100 ng/ml) for 5 min at 37 "C. For immunoprecipitation analysis, stimu- lated cells were treated with 5 mM diisopropyl fluorophosphate (DFP) at 4 "C for 5 min, and lysed in a P-Tyr buffer containing 50 mM HEPES, pH 7.5), 1% Triton X-100, 50 mM NaCl, 50 mM NaF, 10 mM sodium pyrophosphate, 5 mM EDTA, 1 mM Na3 VO,, 1 mM phenylmethylsulfonyl fluoride, 10 pg/ml aprotinin, 10 pg/ml leupep-tin, and 5 mM DFP. Soluble lysates (2 mg) were immunoprecipitated with anti-P-Tyr antibody (Upstate Biotechnology, Inc.). Immunoprecipitates were resolved on SDS-PAGE and immunoblotted with anti- P-Tyr, anti-PLCy (Upstate Biotechnology, Inc.), anti-p85 (Upstate Biotechnology, Inc.), or anti-GAP (34). Receptor-associated Phosphatidylinositol 3-Kinase Assays-For measurement of in uiuo PDGFR-associated phosphatidylinositol 3- kinase activity, quiescent 32D transfectants were exposed to PDGF-BB (100 ng/ml) for 5 min 37 "C, incubated with 5 mM DFP at 4 "C for 5 min, and lysed in a buffer containing 20 Tris (pH

prised of dimers of A and B chains encoded by distinct genes (3,4). All three PDGF isoforms (PDGF-AA, PDGF-BB, and PDGF-AB) have been identified (5)(6)(7). Moreover, they bind with different affinities to two related receptor molecules, designated a and 0 PDGFRs which are also encoded by distinct genes (8, 9). Accumulating evidence indicates that PDGF-induced receptor activation involves recruitment of receptor dimers (10)(11)(12)(13). In fibroblasts where both PDGFRs are expressed, PDGF-BB activates a& p@, and aa receptor dimers. In contrast, PDGF-AA induces formation of only aa receptor homodimers, while PDGF-AB is capable of recruiting aa as well as a(? dimers (14). Several PDGFR substrates have recently been identified. These include phospholipase C--y (PLCy) (15), GTPase-activating protein (GAP) (16,17), and the 85 kDa subunit of the phosphatidylinositol 3-kinase (18). Each has been shown to undergo rapid tyrosine phosphorylation and/or physical association with PDGFRs in response to PDGF triggering (15)(16)(17)(18)(19). PLCy hydrolyzes phosphatidylinositol 4,5-bisphosphate into two second messengers, 1,2-diacylglycerol and inositol 1,4,5-trisphosphate. The former activates protein kinase C, and the latter promotes the release of Ca2+ from intracellular stores (20). GAP enhances hydrolysis of ras-GTP to ras-GDP which normally inactivates ras function, but GAP may also be involved in RAS effector function (21,22). A phosphatidylinositol-3 kinase was initially identified in immunocomplexes with the v-src protein and later found to be physically associated with a variety of tyrosine kinases including PDGFRs (23)(24)(25)(26). Phosphatidylinositol 3-kinase phosphorylates the inositol ring of phosphatidylinositol at the D3 position, but there is yet no clue as to the biological functions of such metabolites (27, 28).
The ability of PDGFRs to associate with and/or tyrosine phosphorylate known substrates has generally been investigated utilizing fibroblasts expressing both receptors. Thus, it has been difficult to quantitatively compare the ability of each receptor to interact with specific substrates. We have previously described 32D hematopoietic cell lines engineered to independently express equivalent levels of a or 0 PDGFRs.
We showed that ligand-induced stimulation of either receptor in this cell system leads to quantitatively similar levels of calcium mobilization, P I turnover, chemotaxis, and mitogenic responses (29). Our present studies were undertaken in an effort to compare the ability of each receptor independently to interact with known intracellular signaling molecules in vivo as well as in vitro.

MATERIALS AND METHODS
Transfection Analysis-The interleukin-3-dependent mouse hematopoietic cell line 32D has been described previously (30). DNA 9287 transfection of 32D cells was performed by electroporation (31). Mass populations of stably transfected cells were selected by their ability to survive in growth medium containing mycophenolic acid (32). Transfection of NIH/3T3 fibroblasts was performed by the calciumphosphate precipitation method (33) using 40 pg of calf thymus DNA as carrier and serial 10-fold dilutions of the indicated DNAs. Transformed foci were scored at 14-21 days. Depending upon the DNA construct, NIH/3T3 transfectants were selected for survival in growth medium containing either mycophenolic acid (32) or geneticin (750 rg/ml).
Purification of PDGF Receptors-Recombinant baculovirus containing human a and 0 PDGF receptor cDNA was used to infect Spodoptera frugrperda (SF9) cells at a multiplicity of infection of 1O:l. After a 60-h infection, SF9 cells were harvested, lysed in a buffer (50 mM HEPES, 150 mM NaCl, 1% Triton X-100) containing 1 mM Na3 VO. and 10 mg/ml of leupeptin, pepstatin, and aprotinin. PDGF receptors were purified from the lysate using an anti-phosphotyrosine column (36). The purity and concentration of PDGFRs preparation were determined by SDS-PAGE followed by silver stain analysis of purified materials. Generation of Recombinant Xgtll Phage and Expression of P-Gal Fusion Proteins-Xgtll-PLCy and Xgtll-GAP were generated by insertion of EcoRI-linkered BstEII fragment of PLCy (37) or Not1 fragment of GAP (38). For expression of p-gal fusion proteins, 10 ml of Escherichia coli (Y 1089) were grown to an ODW of 0.5 and infected with Xgtll-PLCy or Xgtll-GAP. Expression of P-gal fusion proteins containing either GAP or PLCy was induced by the addition of isopropyl-1-thio-0-D-galactopyranoside to a final concentration of 10 mM. After a 2-h treatment, 1-ml aliquots were pelleted and cells lysed in 100 pl of 2 X SDS sample buffer. I n Vitro Association of PLCy and GAP to Purified a or B PDGFRs-Whole cell lysates prepared from cells infected by Xgtll, Xgtll-PLCy, or Xgtll-GAP were separated on a 7.5% SDS-PAGE gel and transferred to Immobilon-P membranes (Millipore). Nonspecific binding was blocked by treating membranes with 3% nonfat dry milk in TTBS (25 mM Tris, pH 7.4, 150 mM NaC1,0.05% Tween 20) for 2 h. For receptor overlays, membranes were incubated with 3.5 ml of binding assay buffer (50 mM HEPES, 10 mM MgC12, 1 mM MnCI2, 0.1% Triton X-100, 50 PM ATP, 1 mM Na3 VO,) including purified PDGFR at indicated concentration for 2 h a t 25 "C. Membranes were washed twice then blotted with anti-a PDGFR or anti-phosphotyrosine monoclonal antibodies (UBI). After washing (2 X with TTBS), membranes were probed with '*'I-protein A, washed again (4 X with TTBS), then autoradiogramed by exposure to Kodak XAR-5 x-ray film. For estimating KD value of a or PDGFR for GAP and PLCy, the anti-P-Tyr signal was quantified by scanning on a Bio-Rad 620 densitometer. The amount of PDGFR bound to recombinant PLCy or GAP was estimated by comparing the anti-P-Tyr signal shown in Fig. 9 to the anti-P-Tyr signal derived from a known concentration of purified a or PDGFR. The data were then analyzed by the method of Scatchard (39).
Ligand Binding Assay-PDGF-AA was labeled by the chloramine T method (40), (specific activity 1 X lo5 counts/min/ng). NIH/3T3 transfectants were plated in 24-well microtiter plates at a density of 6 X lo5 cells/well. Adherent cells were washed in ice-cold HEPES binding buffer (25 mM HEPES, 150 mM NaC1, 5 mM KCl, 1.2 mM MgSO4, and 0.1% bovine serum albumin, pH 7.5) once and incubated for 2 h a t 4 "C with labeled PDGF-AA (1 ng/ml) preincubated in the presence of increasing concentration of unlabeled PDGF-AA. Free ligand was removed by washing cells three times with HEPES binding buffer. Cells were then lysed in 1.0% Triton X-100, and radioactivity in the extract of duplicate samples was measured in a gamma counter. Data obtained from cross-competition studies were then subjected to analysis by the method of Scatchard (39). Immunoprecipitates were electrophoretically separated, transferred to immobilon-P, and blotted with anti-P-Tyr antibody. a PDGFR or /3 PDFGR resulted in anti-P-Tyr recovery of a set of phosphotyrosine-containing proteins of similar respective sizes including p140, p85, p75, p70, and p65 ( Fig. 1). However, we also observed a distinct subset of proteins (p120, p95, and p57) which were more efficiently tyrosine phosphorylated by (3 PDGFR. In contrast, the CY PDGFR demonstrated more efficient tyrosine phosphorylation of proteins with estimated molecular masses of 48 and 50 kDa, respectively. These results were reproducible in several independent experiments, suggesting that the a and /3 PDGFR k' mases may have different specificities for certain signaling molecules in vivo.

Patterns of
Comparison of in Vivo Interactions of a and /3 PDGFR with Known Substrates-By use of 32D-aR and /3R cell lines, it was possible to quantitatively compare the ability of a and /3 PDGFRs to undergo ligand-induced association with and/or tyrosine phosphorylation of specific PDGFR substrates. For these studies, cells were either untreated or exposed to PDGF-BB, and total lysates were subjected to immunoprecipitation using anti-P-Tyr. The immune complexes were then immunoblotted with anti-P-Tyr or antisera directed against specific substrates. As shown in Fig (lanes 3 and 4 ) were either untreated (-) or treated (+) with PDGF-BB. Cell lysates were immunoprecipitated with a monoclonal antibody directed against phosphotyrosines. Immune complexes were then subjected to electrophoresis followed by immunoblot analysis using either anti-P-Tyr (panel A ) or anti-p85 (panel B ) . In panel C, the same immune complexes were subjected to the phosphatidylinositol 3-kinase assay.
GAP antibodies. The results in Fig. 3A demonstrate the presence of activated a PDGFRs, whose level was estimated to be 1.9-fold higher than that observed for activated /3 PDGFRs, based upon anti-P-Tyr signal intensities. At the same time, tyrosine phosphorylation of PLCy was estimated to be 1.3-fold higher in 32D-aR than that observed in 32D-DR. When normalized, the a PDGFR was around 70% as efficient as the /3 PDGFR at PLCy tyrosine phosphorylation.
These results are consistent with previous findings indicating that PDGF-BB stimulation induces similar levels of PI hydrolysis in 32D-aR or /3R cells (29). In striking contrast, GAP appeared to be a much better in vivo substrate of the p PDGFR as compared to CY PDGFR under the same assay conditions (Fig. 3C). Similar results were obtained with at least three independently derived mass cultures of 32D cells expressing either a or /3 PDGFR. The respective levels of immunoreactive GAP, PLCy, and phosphatidylinositol 3-kinase p85 in total cell lysates from 32D-cuR were similar to those in 32D-PR (data not shown).
Comparison of in Vivo Interaction of CY and /3 PDGFRs with GAP-To further explore the mechanism underlying the lower stoichiometry of ligand-induced GAP association and/ or tyrosine phosphorylation by the CY PDGFR, we compared the abilities of activated CY and p PDGFRs to interact with GAP in NIH/3T3 fibroblasts where both PDGFRs are normally expressed. Accordingly, confluent NIH/3T3 cells were rendered quiescent by incubation overnight in serum-free medium and then exposed to a saturating concentration of PDGF-AA or PDGF-BB for 5 min, after which cell lysates were subjected to immunoblot analysis with anti-P-Tyr antibody. As shown in Fig. 4 A , addition of either homodimer led to tyrosine phosphorylation of a similar level of PDGFRs. The same cell lysates (2 mg) were also immunoprecipitated with anti-GAP and immune complexes subjected to immunoblot analysis using anti-P-Tyr or anti-GAP.
As shown in Fig. 4B, PDGF-AA and PDGF-BB both stimulated equivalent levels of anti-GAP recoverable p64/62 (42). However, PDGF-AA induced a 9-fold lower level of tyrosinephosphorylated GAP and anti-GAP recoverable PDGFRs as compared to PDGF-BB.  lanes 1 and 2 ) or (s' PDGFR  (lanes 3 and 4 ) were either untreated (-)  GAP were immunoprecipitated by anti-GAP antiserum in cells stimulated by PDGF-AA or PDGF-BB. Since differences observed in the level of anti-GAP recoverable PDGFRs could not be attributed to differences in levels of activated PDGFRs (see Fig. 4A), our results imply that the decreased level of GAP tyrosine phosphorylation by activated a PDGFRs reflects its lower affinity for GAP.
T o further establish that differences observed in GAP tyrosine phosphorylation did not reflect differences in PDGF binding specificities of the two PDGFRs, we examined the ability of PDGF-AA to stimulate GAP tyrosine phosphorylation in cells engineered to express a similar level of either the a PDGFR or a chimeric PDGFR. The latter, designated the domains were substituted by the homologous domain of the a PDGFR (9), conferring a PDGFR binding properties (41). Accordingly, NIH/3T3 cells were transfected with vector alone (NIH), a PDGFR (NIH-a340R), or a340P342R (NIH-a340@342R), marker selected, and subjected to binding studies using '251-PDGF-AA. As shown in Fig. 5, B and C, both NIH-a340R and NIH-a340p342R cell lines exhibited similar levels of PDGF-AA-binding sites/cell. Each showed 1.5-fold more sites than observed in the parental NIH cells, which expressed 6.5 X lo4 sites/cell (see Fig. 5A). a340 / 3 342 R, comprises a / 3 PDGFR, in which the first 3-IgG-like T o examine the ability of PDGF-AA to stimulate tyrosine phosphorylation of GAP in each cell line, cultures were exposed to saturating concentrations of PDGF-AA or -BB. Total cell lysates (2 mg) were then immunoprecipitated with anti-P-Tyr and the immune complexes subjected to immunoblot analysis using anti-P-Tyr (panel D), anti-PLCy (panel E ) , or anti-GAP (panel F). The results in Fig. 5 0 indicate that PDGF-AA induced equivalent levels of activated PDGFRs in NIH-cu~~OR or NIH-a340@342R, which were -50% higher than that observed in NIH cells. Under these conditions, PDGF-AA stimulated anti-P-Tyr recovery of PLCy with similar relative stoichiometry (see Fig. 5 E ) . In contrast, the level of PDGF-AA-stimulated GAP tyrosine phosphorylation in NIH-@ R was -6-fold higher than that observed in NIH-cY~~OR and 9-fold greater than observed with the parental NIH cells (Fig. 5F). These results strongly argue that the catalytic domain of @ PDGFR exhibits significantly greater ability than the a PDGFR to induce GAP tyrosine phosphorylation and receptor association.
Comparison of Transforming Actiuity of PDGF-AA Coexpressed with Exogenous a or ap PDGFRs-We have previously demonstrated that PDGF-BB is 10-100-fold more efficient than PDGF-AA at inducing transformation of NIH/3T3 cells (43). It is known that PDGF-BB stimulates a as well as @ PDGFRs (41). Thus, the greater transforming efficiency of PDGF-BB could be due to a quantitative increase in the level of activated receptors, differences in a and / 3 PDGFR substrate specificity, or both. T o investigate these possibilities, we utilized the ap chimeric PDGFR (cY''O@~~R) and compared transforming activity of PDGF-A cotransfected with either wild type a PDGFR or the chimeric PDGFR. As shown in Fig. 6 and Table I, transfection with PDGF-A resulted in -2.3 X 10' transformed foci/picomole of DNA (panel 6 B ) , while cotransfection of PDGF-A with the wild type a PDGFR resulted in a 2-fold higher level of focus formation as compared to that observed with PDGF-A alone (panel 6E). COtransfection with the chimeric ab PDGFR (a340p342R) led to a greater than 17-fold increase in PDGF-A transforming efficiency (panel 6F). These results indicate the ( 3 PDGFR catalytic domain is more efficient in mediating PDGF transforming function due to differences in substrate specificity of a and @ PDGFRs.

I n Vitro Affinity for GAP Is Significantly Lower for a
Compared to @ PDGFRs-In an effort to directly compare their affinities for GAP, we examined the ability of either tyrosine phosphorylated a or PDGFR expressed in the baculovirus system to associate with bacterially expressed recombinant GAP or PLCy immobilized on a solid matrix. Accordingly, the Xgtll was engineered to express the fusion protein composed of coding region of either GAP (Xgtll-GAP) or PLCr (Xgtll-PLCy) fused to the @-galactosidase (pgal) gene. Bacteria infected with recombinant bacteriophage were then induced to express p-gal, &gal fused to PLCy (@gal-PLCy) or GAP (P-gal-GAP), respectively. Proteins in lysates were then subjected to SDS-PAGE analysis and transferred to nitrocellulose. Filters were then stained with Amido Black or immunoblotted with antibodies specific to P-gal, GAP, or PLCy.
As shown in Fig. 7A, Amido Black staining confirmed the expression of similar levels of recombinant protein in each of the bacterial lysates. Fig. 7B shows that the majority of proteins whose expression was specifically induced in bacteria infected by Xgtll-GAP or Xgtll-PLCy showed specific reactivity with anti-@-gal antibody. This antibody also recognized a major species of 116 kDa, consistent with the size of the pgal protein itself, in total lysate prepared from bacteria in-

Comparison of transforming activity of LTR-2 ( A ) , PDGF-A ( D ) , PDGF-A/aPDGFR ( E ) , and PDGF-A/o~~'B'~~R ( F ) . NIH/3T3 cells ( B ) , PDGF-B (C), PDGF-AILTR-2
struct by means of calcium-phosphate precipitation technique (33). Plates were stained with Giemsa a t 14 days after transfection. \ " fected by Xgtll alone. When identical filters were immunoprobed with anti-GAP peptide serum, a heterologous series of proteins with a molecular mass ranging from 120-190 kDa was specifically detected in lysate of Xgtll-GAP-infected bacteria (Fig. 7C). Moreover, immunodetection of these proteins was specifically blocked by preincubation of the antiserum with the homologous GAP peptides (Fig. 70). In contrast, the anti-PLCy serum detected a predominant single band with molecular mass of -210 kDa specific to Xgtll-PLCy bacterial lysates (Fig. 7 E ) . These results demonstrated specific expression of P-gal-GAP and p-gal-PLCy fusion proteins in each bacteria.
The specific binding of a or / 3 PDGFRs to GAP, phospha-tidylinositol3-kinase or PLCy has been shown to be depend-ent on specific tyrosine phosphorylation of these receptor molecules. Thus, to compare their in uitro association with bacterially expressed GAP or PLCy, we utilized recombinant PDGFR molecules produced in SF9 insect cells. At high levels of expression in this system, PDGFR molecules became tyrosine phosphorylated, allowing their enrichment by means of anti-P-Tyr immunoaffinity chromatography (36). Accordingly, lysates of bacterially expressed p-gal, p-gal-GAP, or pgal-PLCy proteins were subjected to SDS-PAGE, and immunoblots were incubated in the presence of saturating concentrations of baculovirus a PDGFRs. Filters were then washed extensively and immunoblotted with either anti-a PDGFR monoclonal antibody (Fig. 8A) or anti-P-Tyr (Fig.  8B). Fig. 8 shows that the CY PDGFR bound to GAP and a and / 3 PDGFR Substrate Specificity P L C r specifically, since both anti-P-Tyr and anti-a PDGFR antibodies detected proteins with a similar pattern to those observed with anti-GAP or anti-PLCr antibodies (see Fig. 7, C and E ) . In the absence of PDGFR preincubation, neither anti-a PDGFR or anti-P-Tyr antibodies showed evidence of nonspecific binding to &gal or P-gal-GAP or /3-gal-PLCy proteins (data not shown). We next sought to quantitatively compare the abilities of a and / 3 PDGFRs to associate with GAP and PLCr. Accordingly, lysates of bacterially expressed p-gal-PLCr and &gal-GAP were subjected to immunoblot analysis using increasing concentrations of baculovirus expressed a or / 3 PDGFRs. After extensive washing, specifically absorbed PDGFRs were immunodetected with anti-P-Tyr since this antibody detected both PDGFR molecules with comparable efficiency. As shown in Fig. 9, panels A and B, incubation of an increasing concentration of a or p PDGFRs with bacterially expressed P L C r led to increasing anti-P-Tyr signal. In each case, the major band corresponded in size to the 210-kDa @-gal-PLCy fusion product observed with antisera directed against @gal or PLCr (Fig. 7, B and E ) . Smaller bands likely represented partial breakdown products containing P L C r immunologic determi- Transfection was performed with calf thymus DNA (40 pg) as carrier and serial dilution of indicated plasmids by the calciumphosphate precipitation method (33). Focus formation was scored at 14-21 days and transforming efficiency was calculated in focus forming activity/picomole of DNA (ffu/pmol). Marker selection was performed in medium containing mycophenolic acid (32) or geneticin (750 pg/ml). Data represent mean values of at least three independent experiments performed in duplicate. Standard error was less than 10% of the mean. Relative transforming efficiency was calculated by dividing the ffu foci/pmol of DNA relative to that induced by PDGF-A cotransfected with LTR-2 expression vector (41). nants. Quantitation of anti-P-Tyr signal intensities revealed no significant differences with a or p PDGFRs (Fig. 9, A and  B ) . Affinity constants determined by Scatchard analysis revealed KD values of 0.9 and 0.6 nM for the a and p PDGFR, respectively (see Fig. 10).  C and D). Each receptor showed a similar pattern of binding to similar size products representing bacterial recombinant proteins retaining GAP determinants, the largest of which corresponded to the intact @-gal-GAP fusion product. I t can be further observed that -5-fold higher concentration of a PDGFR as compared to p PDGFR was required to achieve a comparable signal intensity of P-Tyr immunoreactive bands (Fig. 9, C and D). When subjected to Scatchard analysis, the estimated KD value of the p PDGFR for GAP was 0.5 nM. In contrast, the affinity constant of the a PDGFR toward GAP was -3 nM (Fig. 10). These results correlated well with the results of in vivo analysis and implied that the relative inefficiency of GAP tyrosine phosphorylation by the a PDGFR i n vivo reflects a lower binding affinity.

DISCUSSION
Our present studies demonstrate differences in substrate specificities of a and / 3 PDGFRs as indicated by their distinct patterns of PDGF-activated anti-P-Tyr recoverable phosphoproteins in a common target cell. These results are consistent with findings of Eriksson et al. (44) using a different cell system. By comparing activities of the two receptors with respect to several known PDGFR substrates, we also demonstrated similarities as well as striking differences in their abilities to interact with specific target molecules in vivo and in vitro.
Interactions of receptor kinases with some known substrates is dependent upon phosphorylation of specific tyrosine residues within the receptor kinase (45). In addition, these interactions are dependent upon the presence of particular regions within the substrates, designated src homology 2 (SH2) domains (46). The most well characterized interactions of receptor kinases with their substrates involve those with the phosphatidylinositol 3-kinase p85 regulatory subunit. Accumulating evidence indicates that phosphorylation of 2 different tyrosine residues within the respective kinase insert  E). (35,45,(47)(48)(49)  A appears to be required (15). Our present findings demonstrate similar efficiency of PLCy tyrosine phosphorylation by the two receptors in uiuo, as well as comparable affinities for their in uitro interaction with PLCy. Thus, the abilities of these related receptors to interact with certain substrates appear to be very similar.
In striking contrast to our findings with phosphatidylinositol 3-kinase and PLCy, we observed marked differences in the ability of the CY and p PDGFRs to associate with and tyrosine phosphorylate GAP in uiuo. In addition, we demonstrated around 5-fold lower affinity of a PDGFR interaction with GAP both in vivo and in uitro. These findings suggest that differences observed in a and p PDGFR tyrosine phosphorylation of GAP may not be limited to differences in the extent of GAP tyrosine phosphorylation (i.e. lower affinity of activated a PDGFR for GAP), but also may relate to differences in preferred sites of GAP tyrosine phosphorylation.
Recently, two independent laboratories have reported that the ability of the p PDGFR to associate with GAP was determined by phosphorylation of tyrosine 771 located within the carboxyl half of its kinase insert domain (45,49). Sequence comparison between the two PDGFRs indicates that this tyrosine is conserved within the a PDGFR. However, further inspection of this alignment reveals that amino acid residues surrounding tyrosine 771 within the / 3 PDGFR (homologous to tyrosine 762 in the a PDGFR) has significantly diverged between the two receptors. Thus, a more detailed molecular genetic analysis of kinase insert domains of the a PDGFR may help in determining whether or not sequence differences between the receptors in this region could contribute to a lower affinity between GAP and the a PDGFR.
In 32D hematopoietic cells, activation of either independently expressed PDGFR results in similar mitogenic and chemotactic responses (29). Thus, the differences observed in GAP phosphorylation do not directly correlate with either of these biological effects. The p PDGFR is very efficient at phosphorylating GAP (16). Thus, it is not possible to exclude that a threshold requirement is met by the a PDGFR despite its lower affinity GAP interaction. A number of other receptor kinases exhibit low affinity interactions with GAP, PLCy, or both and yet are capable of stimulating good mitogenic re-

aPDGFR(nM)
, $ ,$ $ ,+' , , , + Densitometric analysis was performed on anti-P-Tyr signals shown in Fig. 9. Following standardization with anti-P-Tyr signals derived from known concentrations of purified a or p PDGFRs, the data were then analyzed by the method of Scatchard (39).
sponses (51-54). Thus, there may also be significant redundancy in signaling molecules triggered by the same receptor, such that absence of any one substrate does not substantially affect major biologic responses. Our present studies demonstrate that a chimeric PDGFR with a PDGFR ligand binding properties and the @ PDGFR catalytic domain was markedly more efficient than the a PDGFR at enhancing PDGF-A transforming function in NIH/3T3 fibroblasts. We conclude that differences in a and p PDGFR substrate specificities likely account for the greater ability of the , f 3 PDGFR to mediate PDGF transformation in these cells. Recently, Eriksson et al. (44) reported differences in the ability of LY and @ PDGFRs to induce actin reorganization in porcine aortic endothelial cell lines. All of these findings support the concept that PDGF signaling is determined not only by the levels of a and @ PDGFR expression, but also by their respective interactions with distinct substrates in different cell types. The PDGF ligand/receptor system is further complicated by the existence of three PDGF isoforms which can induce formation of different combinations of activated receptor dimers (14). Thus, it is likely that the PDGF isoforms present in a particular microenvironment as well as relative levels of expression of the two receptors influence the biological responses mediated by these related receptors.