Mechanism of down-regulation of c-kit receptor. Roles of receptor tyrosine kinase, phosphatidylinositol 3'-kinase, and protein kinase C.

The receptor tyrosine kinase Kit and Kit ligand (KL), encoded at the murine white spotting (W) and steel (Sl) loci, respectively, function in hematopoiesis, melanogenesis, and gametogenesis. To understand the mechanism of turnover of Kit in mast cells, mutant receptors generated in vitro were heterologously expressed in Wsb/Wsh mast cells lacking endogenous c-kit expression, and the effects of mutations on KL-induced internalization and ubiquitination/degradation of Kit were studied. Upon binding of KL, KL.Kit receptor complexes were rapidly internalized, and the turnover was accelerated by ubiquitin-mediated degradation. Inactivation of the Kit kinase resulted in a reduced rate of internalization of KL.Kit complexes, degradation of kinase-inactive receptor complexes was relatively slow, and receptor ubiquitination was absent. But abolishment of KL-induced receptor association and activation of phosphatidylinositol 3'-kinase and of tyrosine 821 autophosphorylation did not affect KL-induced internalization and ubiquitination/degradation of Kit. Furthermore, Kit receptors can be down-regulated by proteolytic cleavage induced by either activation of protein kinase C or by isopropyl alcohol. In summary, KL-induced internalization of KL.Kit complexes and ubiquitination/degradation require an active kinase. By contrast, proteolytic cleavage of Kit mediated by protein kinase C activation is independent of kinase activity.

rapidly internalized, and the turnover was accelerated by ubiquitin-mediated degradation. Inactivation of the Kit kinase resulted in a reduced rate of internalization of =.Kit complexes, degradation of kinase-inactive receptor complexes was relatively slow, and receptor ubiquitination was absent. But abolishment of KL-induced receptor association and activation of phosphatidylinositol 3"kinase and of tyrosine 821 autophosphorylation did not affect KL-induced internalization and ubiquitinatioddegradation of Kit. Furthermore, Kit receptors can be down-regulated by proteolytic cleavage induced by either activation of protein kinase C or by isopropyl alcohol. In summary, KL-induced internalization of =.Kit complexes and ubiquitinatiod degradation require an active kinase. By contrast, proteolytic cleavage of Kit mediated by protein kinase C activation is independent of kinase activity.
Ligand-induced down-regulation is an important aspect of the normal physiology of cell surface receptors. Upon ligand binding, receptor-ligand complexes are thought to be endocytosed via clathrin-coated pits. The receptor may then be recycled to the cell surface or targeted to lysosomes for degradation. Moreover, the receptor may be subjected to nonlysosomal degradation involving polyubiquitination of the protein. Consequently, cell surface levels of receptor are down-regulated. Understanding of the molecular steps involved in mediating receptor internalization and degradation has been facilitated using mutant receptors generated by in vitro mutagenesis. To elucidate the mechanism of ligand-induced receptor downregulation, the roles of kinase activity and receptor autophosphorylation have been studied for several receptor tyrosine kinases. These include the platelet-derived growth factor re-* This work was supported by Grant R37 CA 32926 from the National Institutes of Health and the American Cancer Society (to P. B.). The costs of publication of this article were defrayed in part by the payment tisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate of page charges. This article must therefore be hereby marked "aduerthis fact.   Mori et al., 1992;Sorkin et al., 1991), epidermal growth factor receptor (Prywes et al., 1986;Chen et al., 1987;Honneger et al., 19871, colony stimulating factor-1 receptor (CSF-1R) (Downing et al., 1989;Carlberg et al., 19911, and insulin receptor (Russell et al., 1987;McClain et al., 1987;Backer et al., 1989). Although kinase activity is essential for mediating most of the biological effects, a role for kinase activity in receptor down-regulation is not universal and differs depending on the receptor system and the cell type. In addition, most often these analyses were done in nonphysiological cell systems which normally lack endogenous expression of the receptors analyzed. Therefore, caution should be taken when interpreting the physiological significance of results. c-kit is allelic with the murine W locus (Chabot et al., 1988;Geissler et al., 1988) and encodes a transmembrane receptor tyrosine kinase belonging to the PDGF receptor subfamily (Besmer et al., 1986;Yarden et al., 1987;Qiu et al., 1988). The ligand of the c-kit encoded receptor (Kit), Kit ligand (KL), also known as steel factor, is encoded at the SI locus (Nocka et al., 1990a(Nocka et al., , 1990bWilliams et al., 1990;Zsebo et al., 1990;Copeland et al., 1990). Studies in bone marrow-derived mast cells (BMMC) show that Kit receptor activation initiates a cascade of molecular events. KL induces receptor dimerization and kinase activation. Subsequently, receptor autophosphorylation and phosphorylatiordactivation of downstream signaling molecules occur (Rottapel et al., 1991;Reith et al., 1991). Moreover, Kit can be down-regulated by KL, activation of protein kinase C, and calcium ionophores through independent mechanisms (Yee et al., 1993). KL accelerates the turnover of Kit by inducing receptor internalization and degradation. An active kinase has been shown to be essential for Kit to mediate its pleiotropic effects, but it is not known whether receptor kinase activity and receptor autophosphorylation are required for mediating ligand-induced down-regulation of Kit. Also, how activation of protein kinase C affects cleavage of Kit remains unclear.
To understand the mechanism of down-regulation of Kit, the roles of kinase activity and autophosphorylation in KL-induced receptor internalization and degradation were investigated. This was accomplished by using mutant Kit receptors in which: 1) the receptor kinase was inactivated by substitution of the conserved aspartic acid residue 790 with asparagine, known from the W4' allele (Tan et al., 1990); 2) tyrosine 719 and tyrosine 821 were substituted by the structurally similar phenylalanine. Tyrosine 719 within the kinase insert is essential for KL-induced association of the p85 regulatory subunit of phosphatidylinositol 3'-kinase (PI-S'-kinase) with Kit and activa- The abbreviations used are: PDGFR, platelet-derived growth factor receptor; CSF-lR, colony stimulating factor-1 receptor; KL  tion of PI-3"kinase (Serve et al., 1994); whereas tyrosine 821 is a potential autophosphorylation site. A myeloproliferative sarcoma virus-based retroviral expression vector (pGD) and Wh/ Wh mutant BMMC lacking endogenous c-kit expression were used for expression of normal and mutant Kit receptors in mast cells (Tono et al., 1992;Duttlinger et al., 1993). Inactivation of the Kit kinase reduced the rate of internalization and degradation of =.Kit complexes and inhibited receptor ubiquitination. Abolishment of KL-mediated receptor association and activation of PI-3"kinase and of autophosphorylation of tyrosine 821 had no effect on KL-induced internalization and ubiquitinatioddegradation of Kit. In addition, the mechanism underlying protein kinase C-induced cleavage of Kit was examined.
Introduction of Normal and Mutant Kit Receptors into W h / W h BMMC by Retroviral Infiction-The production of mutant c-kit constructs and introduction into Wh/Wh BMMC were described in detail elsewhere.2 Briefly, murine c-kit cDNA based on the published sequence (Qiu et al., 1988) but containing the insertion Gly-Asn-Asn-Lys at amino acid residues 512-513 in the extracellular domain (Hayashi et al., 1991) was used to generate the mutant Kit receptors by site-directed mu-Besmer, P., (1994) EMBO J., in press Serve, H., Yee, N. S., Stella, G., Sepp-Lorenzino, L., Tan, J. C., and tagenesis. The mutant cDNAs were subcloned into the retroviral vector pGD containing the myeloproliferative sarcoma virus (MPSV) long terminal repeat regulating c-kit expression. The constructs were transfected into E86 packaging cells by electroporation, and WhWh BMMC were subsequently infected by co-culture. Infected BMMCs were selected with G418 (400 pg/ml, Life Technologies, Inc.) and refed every 3-4 days for 3 weeks. Infected BMMC were sorted for Kit expression using ACK2-FITC by FACStaPLUS (Becton Dickinson). The mean fluorescence corresponding to the level of Kit in +/+ BMMC was used as cutoff for sorting. Sorted cells were maintained in RPMI complete medium supplemented with 10% Wehi-3 CM and 5% PWM-SCM and refed every 3 4 days. Two weeks later, sorted cells were reanalyzed for Kit expression using ACK2 by FACScan (Becton Dickinson).
Flow Cytometry-Kit-ab binding sites were analyzed by FACS (Becton Dickinson) as described (Yee et al., 1993) except that ice-cold dissociation buffer (120 m M NaCl and 100 m~ acetic acid, pH 4.0) was used to dissociate surface-bound KL, and ACK2-FITC was used instead. The first-order rate constant (K) for the loss of Kit-ab binding sites was determined by linear regression analysis of log % Kit-ab binding sites versus time, and K = -2.303 x slope.
For Scatchard analysis, BMMC (1.5 x 105/ml) were incubated with lP5I-KL (3 x lo3 to 3 x lo6 cpm) in 80 pl of binding buffer (phosphatebuffered saline supplemented with 0.5% (w/v) bovine serum albumin and 0.02% (w/v) sodium azide) in a 96-well plate (Corning) for 1 h at room temperature. The cells were then layered onto calf serum and centrifuged at 400 x g for 5 min at room temperature. y-Emission from the cell pellets and supernatants (containing free 1261-KL) was measured by a Beckman Gamma 5500 counter. Specific binding of 12'I-KL was determined by using the following equation: specific binding = total binding (radioactivity bound in the absence of unlabeled KL)nonspecific binding (radioactivity bound in the presence of 500-fold excess of unlabeled KL). The average nonspecific binding accounts for less than 1.4% of total radioactivity added. Scatchard analysis was performed by plotting the "'I-KL boundfree ratio as a function of Y-KL bound. The correlation coefficients (r2) for the resulting linear regression lines are >0.98.
Metabolic Labeling and Immunoprecipitation Analysis-Metabolic labeling, immunoprecipitations, and SDS-PAGE were done as described (Yee et al., 1993) except that RIPA lysis buffer (50 m~ Tris-HC1, pH 7.5, 150 m M NaCl, 1% Triton X-100, 10 m M EDTA, 1% deoxycholate) was used instead. For immunoblotting, the proteins were transferred from the gel to the nitrocellulose blot (Transblot medium, Bio-Rad) at 60 V B, the amounts of Kit and the high molecular weight form of Kit were quantified using PhosphorImager (Molecular Dynamics) and are expressed as percent of those a t time 0. overnight at 4 "C. The blot was incubated in Tris-buffered saline, pH 7.6 (containing 5% bovine serum albumin) at 37 "C for 1 h; subsequently, anti-Kit serum (1:lOO) or anti-ubiquitin serum (1:lO) (Sigma) was added, and incubation was continued for 1 h a t room temperature. The blot was washed twice with Tris-buffered saline containing 0.1% Tween 20 and 3 times with Tris-buffered saline, and proteins were detected by the ECL method (Amersham).

KL-induced Internalization and Ubiquitination JDegradation of Kit-
To determine the effect of rmKL on the cell surface level of Kit in bone marrow-derived mast cells (BMMC), the level of Kit-ab binding sites was analyzed by FACS. The loss of Kit-ab binding sites followed first-order kinetics (Fig.  lA). Within the initial 10 min following binding of KL, the Kit-ab binding sites decreased rapidly at a rate of 3.3% min", and the rate constant was 1.4 min". Next, the rate of ligand internalization after binding to Kit was determined. "' I-KL was internalized at an initial rate of 3.8% rnin", and the internalization rate constant (Kin) was 1.5 x s-l (Fig. 1B). After reaching a peak at 15 min, the percentage of internalized 12'.I-KL decreased. Therefore, KL-induced loss of cell surface Kit-ab binding sites and the ligand itself was internalized.
To determine the kinetics of degradation of Kit, pulse-chase experiments were done by metabolic labeling of BMMC with [3sS]methionine/cysteine and incubating the labeled cells in the presence of excess unlabeled methionine to chase the 3sSlabeled receptors to the cell surface. The total amount of cellular Kit was analyzed by immunoprecipitation and SDSPAGE.
The half-life ( t h ) of Kit in the absence of ligand was 1-2 h (data not shown). Following incubation with KL, the level of Kit decreased with first-order kinetics and a tv2 of 26.8 min (Fig. 2,  A and B ) . The smear above Kit might represent a high molecular weight form of Kit, and it reached the highest level from 7 to 10 min after treatment with KL. Ubiquitin modification (polyubiquitination) has been implicated in intracellular degradation of a variety of proteins (Rechsteiner, 1987). Immunoblotting of Kit receptor using ubiquitin-specific antibody identified the presence of ubiquitin in the high molecular weight form of Kit (Fig. 3). Therefore, KL-accelerated degradation of Kit was associated with ubiquitination of the receptor. tant forms of Kit had been constructed using the c-kit cDNA. A tant constructs were cloned into the myeloproliferative sar-Kit receptor lacking kinase activity contained a substitution of coma virus-based retroviral expression vector (pGD) containing aspartic acid 790 with asparagine. Tyrosine 719 and adjacent a neomycin resistance gene, and E86 packaging cells were amino acids located in the kinase insert contain the binding transfected with the constructs to generate virus-producing cell site for the regulatory subunit p85 of PI-3"kinase (Serve et al., lines.2 WhWh BMMC lacking endogenous c-kit expression 1994). Tyrosine 821, located within the C-terminal kinase do-were infected with pGD-Kit viruses and infected cells were main, is a conserved residue for autophosphorylation in most selected by G418 and then sorted by FACS based on cell surface receptor tyrosine kinases (Hanks et al., 1987). Both tyrosine expression of Kit. To determine whether the exogenously ex-719 and tyrosine 821 had been changed to phenylalanine by in pressed receptors were appropriately expressed, the FACSvitro mutagenesis (Serve et al., 1994). Subsequently, the mu-sorted cells were metabolically labeled with [35Slmethionine/ cysteine and then analyzed for Kit by immunoprecipitation and SDS-PAGE. Wild-type BMMC were included for comparison. The total cellular amounts of Kit+, KitW42, K i t Y 7 I g F , and KitYR2IF expressed in the infected cells were comparable with each other but relatively low compared to that in wild-type BMMC (not shown). Furthermore, flow cytometric analysis indicated that the mean level of Kit-specific fluorescence on the infected cells was about 75% of that on wild-type BMMC (Fig. 4). The lack of kinase activity of KitW42 was confirmed by in vitro kinase assay and KL-stimulated tyrosine autophosphorylation in vivo (data not shown). Abolishment of KL-induced association of Kit with PI-3"kinase and PI-3"kinase activation for K i t Y 7 I g F was confirmed as well., BMMC expressing wild-type and mutant Kit receptors were then used to study receptor internalization, ubiquitination, and degradation in response to KL, as well as protein kinase C-induced proteolytic cleavage. First, the role of Kit tyrosine kinase in ligand-induced receptor internalization was studied. KL-induced loss of cell-surface Kit-ab binding sites on W h W h BMMC expressing Kit+ occurred at an initial rate of 3.0% min" (Fig. 5A). In BMMC expressing KitW42, Kit-ab binding sites were overall maintained at a relatively constant level throughout the incubation period. Consistent with this finding, the internalization of 1251-KL proceeded at a relatively low initial rate of 2.0% min" with a Kin of 4.3 x lo4 s-* in BMMC expressing KitW4*, as compared to that in cells expressing Kit+ (6.1% min" and 2.0 x s-', respectively) (Fig. 6, A and B). By contrast, KitY7IgF and KitYR2IF did not affect the rate of internalization of 1251-KL (not shown). As indicated by the Scatchard analysis (Fig. 6 0 , the ligand binding affinity was essentially the same for Kit+ and Kit'"42. Thus, it is unlikely that the impaired ligand-induced internalization of KitW42 is due to different binding affinity. Next, the rates of ligand-induced internalization of KitY7lgF and KitYR21F receptors were determined. The initial rates of decrease in Kit-ab binding sites on BMMC expressing K i t Y 7 I g F and KitYs21F were similar to that for Kit+ (2.7% min" and 2.9% rnin", respectively, versus 3.0% rnin") (Fig. 5B). Therefore, the tyrosine kinase activity of Kit is essential for ligand-induced internalization of the KL.Kit complex. However, abolishment of KL-induced PI-3"kinase activation and association with Kit and abolishment of autophosphorylation a t tyrosine 821 did not significantly affect this process.

A. Kit+
We then examined the effects of kinase inactivation and mutations a t tyrosine 719 and tyrosine 821 on ligand-induced receptor degradation. Following treatment of BMMC with KL, Kit+ rapidly decreased with a half-life ( t~) of 21.7 min (Fig. 7, A  and B ) . In contrast, degradation of Kitw42 was relatively slow (tM = 52.1 min). Only slightly reduced rates of degradation were observed for K i t Y 7 I g F (tH = 30.3 min) and KitYRZ1F (tv2 = 28.3 mid.
Therefore, receptor kinase activity is required for KL-induced efficient degradation of Kit. But, KL-induced activation and association of PI-3I-kinase with Kit and autophosphorylation at tyrosine 821 are not essential for these processes.
Phosphorylation of the ubiquitin-activating enzyme E , and the ubiquitin carrier enzyme E, by protein kinase C and protein tyrosine kinase, respectively, has been shown to enhance the activity of E, and E, (Kong and Chock, 1992). Immunoblotting of Kit with ubiquitin-specific antibody showed that KT, induced polyubiquitination of Kit+, KitY7'", and KitYR2IF, but not Kitwd2 (Fig. 8 A ) . On the other hand, down-regulation of protein kinase C by pretreatment of BMMC with TPA did not affect polyubiquitination of Kit (Fig. 8B). Therefore, an active receptor tyrosine kinase but not protein kinase C is necessary for KLinduced polyubiquitination of Kit.
Mechanism of TPA-induced Proteolytic Cleavage of Kit-We had previously shown that activation of protein kinase C by A.

Kit+ KitW42
Time ( KL for up to 10 min a t 37 "C. B , wild type BMMC were pretreated with or without TPA (100 ng/ml) for 24 h to down-regulate protein kinase C. The cells were incubated in the presence or absence of KL for 10 min a t 37 "C. The presence of ubiquitin associated with Kit was determined by immunoprecipitation of cell lysates usingACK2 followed by immunoblotting using anti-ubiquitin serum. Same results were obtained from an independent experiment. TPA decreased the cell surface level a s well as the total cellular amount of Kit, and this was associated with the release of the extracellular domain of Kit (ED100 kDa protein) into the medium. In an attempt to understand the mechanism underlying TPA-induced cleavage of Kit, a panel of protease inhibitors with different substrate specificities was used to characterize the cleavage process in BMMC. Only EDTA inhibited the release of the EDlOO kDa protein to background level as in non-TPAtreated cells (Fig. 9A ), suggesting that TPA-induced cleavage of Kit involves a protease whose activity depends on divalent cation. By contrast, isopropyl alcohol alone is suficient to elicit release of the EDlOO kDa protein (Fig. 9B) possibly by exposing Kit to a surface protease. Moreover, a synergistic effect was achieved by using a combination of isopropyl alcohol and TPA. However, protein kinase C-induced release of the EDlOO kDa protein occurred in TPA-treated BMMC expressing Kit'""', at least as efficiently a s in the cells expressing Kit+ (Fig. 9C). Therefore, proteolytic cleavage of Kit can be accomplished by TPA or isopropyl alcohol, each of which may act by independent mechanisms. ~itY'/l!?F or KitYR21F DISCUSSION Ligand-induced down-regulation of receptors serves as an attenuation mechanism for growth factor-mediated signaling (Wells et al., 1990;Mori et al., 1993). KL binding to its receptor Kit accelerates the turnover of Kit by inducing internalization of receptor-ligand complexes, followed by polyubiquitination and degradation of Kit. We studied the mechanism of downregulation of Kit by using mutant receptors expressed in Wh/ Wh BMMC lacking endogenous c-Kit expression. Since these cells contain the normal machinery for endocytosis and degradation of Kit, our results provide information about Kit receptor desensitization in a unique relevant cellular context. The Kit receptor in the W"' allele contains a single amino acid substitution (D790N) and lacks kinase activity (Tan et al., 1990). Aspartic acid 790 of Kit is analogous with aspartic acid 184 of cyclic adenosine monophosphate-dependent protein kinase and is thought to play a role as the catalytic base in the phosphate transfer reaction (Knighton et al., 1991). KL-induced internalization of Kitw"' was shown to be defective, and this was reflected in a slower rate of degradation of Kitw"' and lack of polyubiquitin modification of Kitw"' receptor. Similarly, kinase-inactive CSF-1 receptor does not undergo accelerated ligand-induced receptor turnover (Downing et al., 1989;Carlberg et al., 1991). Nevertheless, KL still facilitated degradation of Kit'""', although to a lesser extent than the normal receptor, suggesting the existence of alternate pathways for receptor degradation such a s lysosomal degradation known from studies with PDGF receptor down-regulation (Sorkin et al., 1991). The slower turnover of Kit'""' may a t least in part contribute to the dominant phenotype of the W"' allele (Ray et al., 1991). Under such conditions, a net accumulation of Kit' ""' homodimers and Kit""'2/Kit+ heterodimers at the cell surface in response to KL would be expected. In contrast, an EGF receptor mutation, in which receptor internalization is defective, was previously shown to produce an amplified mitogenic signal (Chen et al., 1989;Wells et al., 1990).
The requirement of an active receptor kinase for efficient down-regulation of Kit implies a role for tyrosine phosphorylation of the receptor itself and/or cellular substrates. In previous studies with CSF-1 receptor, deletion of the kinase insert was shown to affect ligand-induced receptor degradation but not receptor internalization (Carlberg et al., 1991). But, phenylalanine substitutions of tyrosines 699 and 708 in the kinase insert region and tyrosine 809 of the CSF-1 receptor had no A and B, 3sS-labeled BMMC were treated with or without TPA(100 ng/ml) in the presence of various protease inhibitors (A) or isopropyl alcohol ( B ) for 60 min a t 37 "C. C, BMMC expressing Kit+, Kit'v42, KitY519F, and KitYR2IF were incubated in the presence or absence of TPA for 60 min a t 37 "C. The cell-free supernatants were analyzed for the ED100 kDa protein by immunoprecipitation and 7.5% SDS-PAGE. Time of exposure for autoradiogram: 12 days. Same results were obtained in two independent trials. significant effects on ligand-induced receptor degradation. In agreement with this, mutation of the analogous tyrosine 821 of Kit did not affect KL-induced internalization and degradation of the receptor. By contrast, CSF-lRYsogF abolishes cell proliferation (Roussel et al.,199OJ and KitYR21F cell proliferation and survival.2 A potential role for phosphatidylinositol 3"kinase (PI-3"kinase) in intracellular protein transport had been suggested based on sequence homology between the pll0 catalytic subunit of mammalian PI-3"kinase and the yeast VPS34 protein (Panayoutou and Waterfield, 1992). Our results demonstrate that lack of KL-induced association of Kit with PI-3"kinase and consequently lack of activation of PI-3"kinase did not significantly affect internalization and ubiquitinatiod degradation of Kit. Recently, studies concerning the role of PI-3"kinase in the internalization of PDGFR showed that in Hep G2 cells, exogenously expressed normal and mutant PDG-FRb require functional PI-3"kinase binding sites for ligandinduced endocytosis (Joly et al., 1994). However, PI-3"kinase binding site mutations in hPDGFRb exogenously expressed in porcine aortic endothelial cells did not affect receptor internalization (Mori et al., 1994). In contrast, the PDGFR src binding site Tyr-579 in the juxtamembrane domain appeared to control ligand-induced internalization of PDGFRb in porcine aortic endothelial cells (Mori et al., 1994). The contradicting results concerning the role of PI-3"kinase in internalization of PDGFR may suggest that the cellular context in which receptor function is studied determines the signal transduction pathways for receptor internalization. It is possible that the endocytic machinery is cell type-specific such that the requirement of PI-3'kinase for endocytosis of PDGFR differs depending on the cellular context. This underscores the importance of the cellular context in the study of biological functions of receptor signaling.
KL was recently shown to induce adhesion of BMMC to a fibronectin matrix (Dastych and Metcalf, 1994). Interestingly, BMMC expressing KitY719F are impaired in their ability to mediate cell adhesion: implying a role for PI-3"kinase in this process. Therefore, KL-induced activation of PI-3"kinase as well as Tyr-821 autophosphorylation are important for the biological functions of Kit, but not for regulating cell surface levels of the receptor.
Ubiquitin modification is thought to play a fundamental role in mediating protein degradation. Ubiquitin modification of proteins involves several enzymatic steps that are well documented. A requirement for kinase activity in mediating ubiquitination may suggest that receptor autophosphorylation and association with a docking protein containing an SH2 domain or the phosphorylation of a cellular substrate mediate the ratelimiting step in the ubiquitination reaction. In agreement with this notion, PDGFRb with phospholipase C-yl binding site mutations expressed in porcine aortic endothelial cells blocked growth factor-induced ubiquitin modification (Mori et al., 1993). It remains to be determined whether a similar mechanism mediates ubiquitin modification of Kit.
Activation of protein kinase C induces down-regulation of Kit receptors, and this down-regulation is associated with proteolytic cleavage of Kit, releasing the extracellular domain of the receptor (ED100 kDa protein) into the medium (Yee et al., 1993). While protein kinase C activity is essential for this effect to occur, tyrosine kinase activity of Kit is dispensable, and this finding agrees with that for CSF-1R (Downing et al., 1989). While the protease facilitating down-regulation of Kit requires divalent cation, isopropyl alcohol induces proteolytic cleavage of Kit, and isopropyl alcohol-induced cleavage is enhanced with TPA. This may suggest that exposure of Kit to the protease may be facilitated by isopropyl alcohol, possibly by modifying membrane fluidity, although the involvement of protein kinase C in isopropyl alcohol-induced cleavage cannot be excluded. Recently, Koike et al. (1993) showed that KL activated phospholipase D that led to the formation of 1,2-diacylglycerol, a physiological activator of protein kinase C. Moreover, protein kinase C activation increases serinelthreonine phosphorylation of Kit resulting in inhibition of autophosphorylation of Kit (Blume-Jensen et al., 1993). Taken together, KL may activate protein kinase C and, thus, down-regulate Kit as a negative feedback mechanism by 1) cleavage of Kit to reduce ligand-binding capacity and 2) by impairment of autophosphorylation of Kit. Henceforth, a diversity of means for receptor down-regulation exists, and this may be important as alternative mechanisms to ensure receptor desensitization.