Identification of a domain within the carboxyl-terminal region of the beta platelet-derived growth factor (PDGF) receptor that mediates the high transforming activity of PDGF.

We have reported previously that a chimeric platelet-derived growth factor receptor (PDGFR) possessing the ligand binding domain of the alpha PDGFR and the intracellular domain of the beta PDGFR (alpha 340 beta 342 R) was markedly more efficient than the wild type alpha PDGFR (alpha RWT) in its ability to enhance PDGF-A transforming activity in NIH/3T3 fibroblasts. To determine the region within the cytoplasmic domain of beta PDGFR that confers this higher transforming activity, we generated several additional alpha/beta PDGFR chimerae. When a chimeric PDGFR possessing the first 933 amino-terminal amino acids from the alpha PDGFR and the final 165 amino acids from the carboxyl-terminal of the beta PDGFR (alpha 933 beta 942 R) was cotransfected with the PDGF-A gene into NIH/3T3 cells, it showed a similar high efficiency to enhance PDGF-A chain transforming activity as alpha 340 beta 342 R. However, when chimeric PDGFRs in which either the kinase insert domain (alpha beta RKI) or the last 79 amino acids from the carboxyl-terminal end of the beta PDGFR (alpha 1024 beta 1028 R) were substituted into alpha PDGFR sequences were cotransfected with PDGF-A, they showed similar low efficiencies in enhancing transforming activity as the alpha RWT. These results predicted that the 86 amino acids following the tyrosine kinase 2 domain of beta PDGFR (amino acid residues 942-1027) were responsible for the higher transforming activity of beta PDGFR. To confirm this finding, we next constructed a chimera in which amino acid residues 942-1028 of the beta PDGFR (alpha beta 942-1028R) were substituted for those in the alpha PDGFR. Cotransfection experiments indicated that alpha beta 942-1028R increased transforming activity of PDGF-A to similar extent as the alpha 933 beta 942R, or alpha 340 beta 342R. Therefore, our findings define a critical domain within the noncatalytic region of beta PDGFR intracellular domain that confers the higher focus forming activity mediated by the beta PDGFR.

We have reported previously that a chimeric plateletderived growth factor receptor (PDGFR) possessing the ligand binding domain of the ␣PDGFR and the intracellular domain of the ␤PDGFR (␣ 340 ␤ 342 R) was markedly more efficient than the wild type ␣PDGFR (␣RWT) in its ability to enhance PDGF-A transforming activity in NIH/ 3T3 fibroblasts. To determine the region within the cytoplasmic domain of ␤PDGFR that confers this higher transforming activity, we generated several additional ␣/␤PDGFR chimerae. When a chimeric PDGFR possessing the first 933 amino-terminal amino acids from the ␣PDGFR and the final 165 amino acids from the carboxyl-terminal of the ␤PDGFR (␣ 933 ␤ 942 R) was cotransfected with the PDGF-A gene into NIH/3T3 cells, it showed a similar high efficiency to enhance PDGF-A chain transforming activity as ␣ 340 ␤ 342 R. However, when chimeric PDGFRs in which either the kinase insert domain (␣␤RKI) or the last 79 amino acids from the carboxyl-terminal end of the ␤PDGFR (␣ 1024 ␤ 1028 R) were substituted into ␣PDGFR sequences were cotransfected with PDGF-A, they showed similar low efficiencies in enhancing transforming activity as the ␣RWT. These results predicted that the 86 amino acids following the tyrosine kinase 2 domain of ␤PDGFR (amino acid residues 942-1027) were responsible for the higher transforming activity of ␤PDGFR. To confirm this finding, we next constructed a chimera in which amino acid residues 942-1028 of the ␤PDGFR (␣␤ 942-1028 R) were substituted for those in the ␣PDGFR. Cotransfection experiments indicated that ␣␤ 942-1028 R increased transforming activity of PDGF-A to similar extent as the ␣ 933 ␤ 942 R or ␣ 340 ␤ 342 R. Therefore, our findings define a critical do-main within the noncatalytic region of ␤PDGFR intracellular domain that confers the higher focus forming activity mediated by the ␤PDGFR.
Although both PDGF-AA and PDGF-BB are mitogenic as well as chemotactic for cells possessing appropriate PDGFRs (28), we and others have shown previously that PDGF-BB is 10 -100-fold more efficient than PDGF-AA at inducing transformation of NIH/3T3 cells (29,30). It is known that PDGF-BB stimulates ␣as well as ␤PDGFRs, while PDGF-AA binds only to the ␣PDGFR. Thus, the greater transforming efficiency of PDGF-B could be due to a quantitative increase in the level of activated PDGFRs and/or differences in substrate specificity of ␣ versus ␤PDGFRs. We have observed previously that cotransfection of the PDGF-A chain gene with wild type ␣PDGFR (␣RWT) increased PDGF-A chain transforming activity by approximately 2-fold. In contrast, cotransfection of a chimeric receptor possessing ␣PDGFR PDGF-A ligand binding domain and the remaining sequences, including the catalytic domain from ␤PDGFR (␣ 340 ␤ 342 R), resulted in a 17-fold enhancement of PDGF-A chain transforming activity (25). Thus, our previous * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ findings indicated that the increased transforming activity of PDGF-B in comparison with PDGF-A could be due to distinct substrate specificities of the two PDGFRs. In the present manuscript, we have identified a specific region within the carboxyl-terminal of the ␤PDGFR that is responsible for mediating the enhanced focus formation of PDGF-B.

EXPERIMENTAL PROCEDURES
Materials-PDGF-AA, PDGF-BB, and monoclonal anti-Tyr(P) antibody were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Anti-␣PDGFR is a polyclonal antibody that was generated against bacterially expressed protein encoding amino acids 25-530 of the extracellular domain of ␣PDGFR (31). mAb-␣R1 is a monoclonal antibody raised by immunizing BALB/c mice with 32D cells expressing the human ␣RWT (32).
Generation of Cytoplasmic ␣/␤PDGFR Chimerae-The ␣/␤ PDGFR chimerae were generated by polymerase chain reaction. The entire coding region of each chimeric PDGFR was then subcloned into the LTR-2 mammalian expression vector (35). A detailed description will be provided upon request.
For in vitro kinase assays, total cell lysates (about 2 mg) were subjected to immunoprecipitation using mAb-␣R1. The immune complexes were extensively washed and incubated in 20 mM Tris (pH ϭ 7.5), 10 mM MgCl 2 , 5 mM MnCl 2 , 10 g/ml aprotinin, and 50 Ci of [␥-32 P]ATP (Amersham Corp.) for 15 min at room temperature. The immune complex reactions were then washed and electrophoretically separated on 8% SDS-polyacrylamide gel electrophoresis. The gels were dried and autoradiogrammed. The levels of PDGFRs and the kinase activities of each ␣/␤ chimerae in the various transfectants were quantitated using scanning densitometer (PDI, Inc.) and expressed relative to the signal derived from the ␣RWT. Immunoblot and immunoprecipitation analyses, NIH/3T3 transfection and cotransfection assays were performed as described previously (34).

RESULTS AND DISCUSSION
Generation of the Chimeric Receptors between the Cytoplasmic Domain of ␣and ␤PDGFR-We have reported that cotransfection of an expression vector containing the PDGF-A gene cDNA with another vector encoding a chimeric receptor possessing the first 340 amino acids of the ␣PDGFR fused to the remaining extracellular transmembrane and intracellular domain of the ␤PDGFR (␣ 340 ␤ 342 R) enhanced the transforming activity of PDGF-A by 17-fold. In contrast, a similar cotransfection experiment utilizing a ␣RWT expression vector increased PDGF-A transforming activity by approximately 2-fold only (25). Since the expression levels of ␣RWT and ␣ 340 ␤ 342 R were shown to be comparable, these findings strongly suggested that the greater transforming activity of PDGF-B compare with that of PDGF-A is due to distinct substrate specificities of the ␣and ␤PDGFR. Sequence comparison between ␣PDGFR and ␤PDGFR intracellular domains revealed that amino acids sequence within the juxtamembrane, tyrosine kinase 1 and tyrosine kinase 2 domains are 80 -90% homologous, while the amino acid sequences within the kinase insert (KI) and the carboxyl-terminal domains are less than 30% homologous (4). Therefore, we sought to more precisely define the region of ␤PDGFR that mediates the greater transforming activity by generating expression vectors containing cDNA that encoded various ␣/␤PDGFR chimerae (Fig. 1). The chimeric cDNAs were comprised mostly of the ␣PDGFR sequences. One chimeric receptor contained the kinase insert domain of the ␤PDGFR and another one contained most of tyrosine kinase 2 and all of the carboxyl-terminal domains of ␤PDGFR (designated as ␣␤RKI and ␣ 833 ␤ 842 R, respectively). From the remaining ␣/␤PDGFR chimerae, one chimera contained all of the carboxyl-terminal domain of ␤PDGFR (designated as ␣ 933 ␤ 942 R), while the other contained the final 79 amino acids of the ␤PDGFR carboxyl terminus (designated as ␣ 1024 ␤ 1028 R). Finally, we also constructed a chimera in which amino acid residues 942-1028 of the ␤PDGFR were substituted for those in the ␣PDGFR (designated as ␣␤ 942-1028 R).

Comparison of the Ability of Each Chimera to Enhance PDGF-A Chain
Transforming Activity-To compare the ability of each chimera to enhance PDGF-A chain transforming activity, we cotransfected 1 g of expression vector containing the PDGF-A gene along with 1 g of expression vector containing each chimeric ␣/␤PDGFR into NIH/3T3 cells. As summarized in Table I, cotransfection of PDGF-A with the LTR-gpt vector alone resulted in formation of 10 -20 small foci. Cotransfection of PDGF-A chain with ␣RWT only weakly enhanced the number of large foci formed. Consistent with previous results, cotransfection of PDGF-A with ␣ 340 ␤ 342 R generated around 350 -400 large foci.
Cotransfection of PDGF-A chain with ␣␤RKI resulted in production of similar numbers of foci as those induced by cotransfection of PDGF-A with ␣RWT. In contrast, cotransfection of ␣ 833 ␤ 842 R with PDGF-A resulted in marked increase in transforming activity similar to that induced by cotransfection of PDGF-A with ␣ 340 ␤ 342 R. Moreover, ␣ 933 ␤ 942 R exhibited a similar ability to enhance the transforming activity of PDGF-A as that of ␣ 833 ␤ 842 R. These results suggest that substitution of KI domain of the ␣PDGFR with the KI domain of the ␤PDGFR does not enhance transforming activity. However, the replacement of the carboxyl-terminal domain of the ␣PDGFR with that of ␤PDGFR is sufficient to confer the higher transforming activity. Finally, we also showed that cotransfection of PDGF-A with ␣ 1024 ␤ 1028 R did not result in enhanced transformation, suggesting that the most distal 79 ␤PDGFR amino acids are not responsible for conferring the high transforming activity mediated by the ␤PDGFR. Since the expression vectors containing PDGF-A and ␣/␤PDGFR cDNAs each contain different drug-resistant markers (29,33), the number of cells expressing both cDNAs was also determined by analyzing the number of colonies that survived in the presence of both HAT/mycophenolic acid and geneticin. As shown in Table I, each cotransfection resulted in similar numbers of double marker-selected colonies, suggesting similar levels of plasmid DNA were transfected into the cells. Together, these results predicted that the 86 amino acids (amino acids 942-1028) following tyrosine kinase 2 domain of the ␤PDGFR may be responsible for its transforming activity.
Amino Acid Residues 942-1028 of the ␤PDGFR Are Critical for Mediating Its Higher Transforming Potential-To confirm that amino acid residues 942-1028 of the ␤PDGFR were sufficient for enhancing cellular transformation, we next generated a chimera designated ␣␤ 942-1028 R by substituting these ␤PDGFR residues for those normally expressed by ␣RWT. The ability of this chimera to increase PDGF-A transforming activity was then compared with that mediated by ␣RWT, ␣ 933 ␤ 942 R, and ␣ 1024 ␤ 1028 R. Results shown in Fig. 2 indicate that ␣␤ 942-1028 R increased the low transforming activity of PDGF-A to similar extent as that observed for the ␣ 933 ␤ 942 R. Consistent with our previous finding, ␣RWT and ␣ 1024 ␤ 1028 R each failed to appreciably enhance the low transforming activity of PDGF-A.
Since the level of receptor expression and their kinase activities have been documented to affect transforming activity of receptor tyrosine kinases expressed in NIH/3T3 cells (35,36), we next sought to examine the level of human chimeric ␣/␤PDGFR protein and its kinase activity in each transfectant. Total cell lysates prepared from NIH/3T3 cells transfected with ␣RWT, ␣ 933 ␤ 942 R, ␣ 1024 ␤ 1028 R, ␣␤ 942-1028 R, or vector alone were immunoprecipitated using a monoclonal antibody, mAb-␣R1, which recognizes human ␣PDGFR, but not endogenous murine ␣PDGFRs present in NIH/3T3 (31). The immune complexes were then subjected to either immunoblot analysis using anti-␣PDGFR serum or an in vitro kinase assay. Since the anti-␣PDGFR is directed against the extracellular domain of the ␣RWT, it was possible to compare the expression levels of each chimera and the wild type ␣PDGFR using this antibody. As shown in Fig. 3A, anti-␣PDGFR specifically detected proteins with molecular masses of approximately 190 kDa in lysates from cells transfected with the human ␣RWT or the ␣/␤ chimerae. Note that under these experimental conditions, the murine ␣PDGFR naturally expressed in NIH/3T3 cells, was not observed (Fig. 3A, lane 1). Quantitation of the amounts of the human receptor detected from each transfectant revealed that NIH/3T3 cells transfected with ␣ 933 ␤ 942 R, ␣ 1024 ␤ 1028 R, and ␣␤ 942-1028 R expressed about 1.7-, 0.7-, and 1.0-fold that of ␣RWT, respectively. Thus, the higher transforming activity of ␣␤ 942-1028 R was not due to higher protein expression, since the steady state level of this chimera was found to be nearly identical to ␣RWT level in the transfectants.
The level of ligand-induced receptor autophosphorylation is a good measurement for receptor tyrosine kinase activity (37). Therefore, we next subjected immune complexes prepared using mAb-␣R1 to an in vitro kinase assay. Quantitation of data presented in Fig. 3B demonstrated that the level of receptor autophosphorylation of ␣ 933 ␤ 942 R, ␣ 1024 ␤ 1028 R, and ␣␤ 942-1028 R was 0.5-, 1.1-, and 0.5-fold that of ␣RWT, respectively. Since the ␣␤ 942-1028 R exhibited lower autokinase activity as compared with that shown by ␣RWT, our findings sug-gest that the higher transforming activity of this chimera is not due to the higher level of receptor tyrosine kinase activity.
In previous studies we showed that the higher focus forming activity of PDGF-B was due to distinct biochemical properties of the ␤PDGFR intracellular domain (25). In the present report, we have utilized a PDGF-A focus formation enhancement assay to localize a region within the ␤PDGFR carboxyl terminus that is responsible for this effect. This region designated as the minimal transforming domain of the ␤PDGFR is 86 amino acids in length and most likely represents an independent functional domain, since attempts to further dissect this domain led to a reduction in transforming activity (data not shown). The minimal transforming domain of ␤PDGFR has minimal effect on expression level or kinase activity of the receptors, suggesting that the difference in transforming activity mediated by ␣PDGFR and ␤PDGFR may be due to their different substrate specificity. Activation of the ␤PDGFR has been reported to induce significantly higher levels of tyrosine phosphorylation of RasGAP and Nck than ␣PDGFR (25,27). However, the binding sites within ␤PDGFR for both Nck and RasGAP are within its kinase insert domain (13, 18 -20). Since substitution of ␣PDGFR KI domain with that of ␤PDGFR did not confer the high transforming activity to the ␣PDGFR, our findings suggest that Nck or RasGAP are not likely to be important for PDGF-induced transformation. Furthermore, we have found that Shc, which is involved in mediating Ras activation (24,38,39), is tyrosine-phosphorylated to a similar extent by activated ␣ and ␤PDGFR and each ␣/␤PDGFR chimerae expressed in 32D transfectants. 2   that Shc is the major downstream signaling molecule that regulates the higher transforming activity of ␤PDGFR versus ␣PDGFR is unlikely (40).
The minimal transforming domain of the ␤PDGFR contains four tyrosine residues. Among these tyrosines 966 and 970 are not phosphorylated in vivo. 3 In contrast, tyrosine residues 1009 and 1021 have been shown to undergo ligand-dependent tyrosine phosphorylation and association with Syp and PLC␥, respectively (11)(12)(13)(14)(15). The ␣PDGFR has been shown previously to interact with PLC␥ and Syp with a lower efficiency (25,27). In addition we have shown that the low transforming activity of PDGF-A mediated by the ␣PDGFR can be abolished by mutation of the PLC␥ binding site (41). Therefore, our data indicate that PLC␥ and/or Syp may be involved in the higher transforming activity of PDGF-B mediated by the ␤PDGFR. Interestingly, it has been reported recently that Tyr 1009 and Tyr 1021 are also required for efficient ligand-induced ubiquitination of the activated ␤PDGFR leading to proper degradation of the ligandactivated receptor (42). In addition, the minimal transforming domain of the ␤PDGFR contains a hydrophobic region that has been shown previously to be essential for proper ligand-induced internalization and down-regulation of activated ␤PDGFRs (43). Thus, it would be remiss not to consider the possibility that the higher transforming activity of the ␤PDGFR may also be due to the difference in the ligand-induced post-translational regulation of the ␤PDGFR. Further studies to dissect the role of this region in the ligand-induced receptor processing and substrate phosphorylation will potentially help us to elucidate the exact molecular basis for PDGF-induced transformation of NIH/3T3 cells by the ␤PDGFR.