the journal of vol. no. 22, of pp. by for u. s. a ... · dimerization and activation of the kit...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 22, Issue of August 5, pp. 15970-15977,1992 Printed in U. S. A .

Dimerization and Activation of the Kit Receptor by Monovalent and Bivalent Binding of the Stem Cell Factor*

(Received for publication, March 17, 1992)

Sima Lev, Yosef Yarden$, and David Givolt From the Department of Chemical Immunology, the Weizmann Institute of Science, Rehovot 76100, Israel

The protooncogene c-kit encodes a tyrosine kinase receptor for the stem cell factor (SCF). Mutants of c- kit were shown to confer a pleiotropic defective phenotype and often display negative dominance in heterozygous mice. To explore the involvement of receptor dimerization in this genetic phenomenon, we employed both a human ligand, which does not recognize the murine receptor, and a rodent SCF, which binds to the human receptor with 100-fold reduced affinity as compared with human SCF. SCF binding to living cells was found to induce rapid and complete receptor dimerization that involved activation of the catalytic tyrosine kinase function. Although receptor dimerization can be attributed to the dimeric nature of the ligand, no dissociation of Kit dimers occurred at high excess of SCF, suggesting that receptor-receptor interactions are also involved in dimer stabilization. This was supported by in vitro formation of heterodimers between the human and murine Kit proteins through monovalent binding of species-specific human SCF. By coexpression of human and mouse Kit in murine fibro- blasts, we found that receptor heterodimerization in living cells involved an increase in the affinity of human Kit for rat SCF and also an accelerated rate of receptor down-regulation. When a human Kit mutant lacking the kinase insert domain was coexpressed with the murine wild-type receptor, we observed a significant decrease in both the activation of the intact tyrosine kinase and its coupling to an effector protein, namely phosphatidylinositol 3”kinase.

Our results favor a receptor activation model that assumes an initial step of monovalent ligand binding, followed by an intermediate receptor dimer bound by one arm of the ligand molecule. This model predicts the existence of an intrinsic receptor dimerization site and provides a structural basis for genetic dominance of mutant SCF receptors.

The binding of growth factors to their respective cellular receptors is the first event in signaling for cell growth and differentiation. A large group of growth factor receptors con- sists of transmembrane tyrosine kinases composed of an

*This work was supported in part by a grant from the Israel Cancer Research Fund, National Institutes of Health Grant CA 53120, a grant from the Wolfson Foundation administered by the Israel Academy of Sciences and Humanities, and the Forscheimer Center for Molecular Genetics. 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.

3 Recipient of a research career development award from the Israel Cancer Research Fund.

Incumbent of the Oscar H. and Anne Altshuler Chair of Immu- nology.

extracellular ligand-binding domain and a cytoplasmic kinase domain connected by a single transmembrane region (1). In general, truncated receptors devoid of the extracellular region show constitutive kinase activity, whereas intact receptors, in the absence of ligand, are essentially devoid of kinase activity (1,2). Hence, the extracellular region apparently exerts inhibitory constraints on the cytoplasmic kinase activity. Ligand binding to the extracellular region appears to relieve these constraints and renders the kinase domain enzymatically active. It has been demonstrated for several receptors that the major common mechanism for this change in kinase activity is due to a ligand-induced receptor dimerization (3- 6). Several important biological consequences result from the dimerization mechanism of kinase activation. First, dimerization facilitates transphosphorylation between the dimer partners (7-9). Second, it allows for cross-talk between receptors even when one of them is not occupied by the ligand (10, 11); and finally, it may lead to a dominant negative effect in heterozygous animals in which one receptor allele is mutated (12,13). The latter situation is well exemplified in the case of the Kit receptor (14-16). Kit is a tyrosine kinase receptor, related in its structure to the receptors for the platelet-derived growth factor (PDGF)’ and the macrophage growth factor (colony-stimulating factor-1) (17, 18). These receptors have in common five immunoglobulin-like domains at the extracellular region and a split kinase in the cytoplasmic domain. The protooncogene c-kit is allelic with the W (white spotting) locus of the mouse, and W mutants display germ line point mutations (rearrangements or deletions) in the c-kit gene (19- 21). Severe W mutants are lethal in the homozygous state and also display a strong mutant phenotype (white spotting, sterility, and severe anemia) in the heterozygous state, suggesting a dominant negative effect of the mutant allele (22- 24). Further evidence for the crucial involvement of Kit in the W phenotype was obtained when it was found that mutations in the Steel locus ( W ) , which result in a phenotype remarkably similar to W, affect the gene that encodes the ligand for Kit, also termed the stem cell factor (SCF) (25-32). SCF is a noncovalent homodimeric glycoprotein (31,33) that stimulates the proliferation of mast cells, melanocytes, and primitive hematopoietic progenitors.

Prior to the identification of SCF, we analyzed the activity of the Kit receptor by constructing a chimeric receptor between the epidermal growth factor (EGF) receptor and Kit in

The abbreviations used are: PDGF, platelet-derived growth factor; SCF, stem cell factor (prefixes h and r denote human and rat, respectively); EGF, epidermal growth factor; EDAC, l-ethyl-343- dimethylaminopropy1)carbodiimide hydrochloride; BS3, 3,3’-bis- (sulfosuccinimidy1)suberate; Ab, antibody; mAb, monoclonal antibody; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; EGTA, [ethylene- bis(oxyethylenenitrilo)]tetraacetic acid; Hepes, 4-(2-hydroxyethyl)- l-piperazineethanesulfonic acid; hKit, human Kit; mKit, mouse Kit.

15970

Dimerization of Kit 15971

which the tyrosine kinase activity of Kit was regulated by the heterologous growth factor, namely EGF (34, 35). A similar approach was used by others (36). These analyses revealed that stimulation of the Kit tyrosine kinase couples the receptor with a unique set of cytoplasmic signaling pathways and results in accelerated cell growth that, under certain condi- tions, confers a transformed phenotype (34,35).

To address the phenomenon of genetic dominance of some mutants of Kit/SCF receptor, we studied the initial events that follow SCF binding. The negative dominance is best explained by functional inactivation of heterodimers between a wild-type and a mutant receptor (14-16), and the ligand- induced dimer of Kit has been reported recently (37). Here we describe quantitative aspects of the phenomenon of Kit dimerization and correlate it with kinase activation. We further analyzed the formation of receptor heterodimers by ex- ploiting the human and mouse receptors and their very different affinities for rodent and human SCFs. Our results support a mechanism for dimerization that involves the receptor dimerization site in addition to the dimeric nature of the ligand. In addition, we employed mutagenesis of the Kit receptor to demonstrate negative dominance at the level of signal transduction in a model system.

MATERIALS AND METHODS

Cross-linking Reagents-l-Ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDAC) was from Sigma. 3,3'-Bis- (sulfosuccinimidyl) suberate (BS3) was from Pierce. Wheat germ agglutinin-agarose was obtained from Biomakor. 32Pi and Na'T were purchased from Amersham Corp. SCF was obtained from Amgen (Thousand Oaks, CA).

Cell Lines-The T18 cell line (38) was established by transfection of the human kit (hKit) cDNA cloned in a mammalian expression vector into Chinese hamster ovary cells using the calcium phosphate method (39). Transfectanta were selected by resistance to Geneticin (600 pg/ml). The MK37 cell line expressing mouse Kit was established by transfection of mouse kit cDNA cloned in the pZL retroviral vector (40) into $2 cells and selection in the presence of Geneticin (1 mg/ ml). Cell lines that express either the human KitACT mutant (Kit devoid of the COOH-terminal42 amino acids) or the human KitAKI mutant (Kit devoid of the kinase insert) were established as described (38). Three cell lines that coexpress mouse Kit (mKit) and hKit were established as follows: cell line MHAKI, in which hKitAKI was transfected into MK37 cells; cell line MHACT, in which mouse kit was transfected into NIH-3T3 cells expressing hKitACT; and MHKit, in which mouse kit was transfected into NIH-3T3 cells expressing hKit. All the transfections for these cell lines were performed together with pSVBhph, a plasmid that expresses hygromycin resistance and selected with hygromycin (50 pglml). The T18 cells were cultured in Ham's F-12 medium, whereas the other cell lines were cultured in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum (HyClone Laboratories, Logan, UT).

Antibodies and Immunoprecipitation-A rabbit antipeptide antibody directed against the COOH-terminal end of the human Kit receptor (Ab212), an antipeptide antibody directed against the kinase insert of the human Kit receptor (Ab213), and a mouse monoclonal antibody directed against the extracellular portion of the human c- kit receptor (mAb94) were prepared as described (38, 41). Anti- phosphotyrosine antibodies were prepared as previously described (34). A rabbit antibody to phosphatidylinositol 3"kinase was prepared against a synthetic peptide containing the COOH-terminal 15 amino acids of human p85a, the regulatory subunit of phosphatidyl-

tially as described (34). inositol 3"kinase (42). Immunoprecipitations were carried out essen-

Cell Labeling-For [35S]methionine labeling, confluent cells were washed three times with phosphate-buffered saline (PBS) and incubated for 12 h in methionine-free Dulbecco's modified Eagle's medium containing 10% dialyzed bovine calf serum and 50 pCi of [35S] methionine/ml. Labeling with 32Pi was performed for 4 h in phosphate-free Dulbecco's modified Eagle's medium containing 0.5% dialyzed bovine calf serum and 0.5 mCi of 3ZPi/ml.

Chemical Cross-linking of Kit in Intact Celk-Confluent cells in 65-mm plates were incubated with SCF in phosphate-buffered saline

for 90 min at 4 "C. The cells were then transferred to 22 "C and incubated for 1 h with 15 mM EDAC. The cross-linking reaction was terminated by extensive washing with PBS, followed by a 5-min incubation with PBS containing 150 mM glycine HC1 (pH 7.5).

Purification of Kit Receptor-The Kit receptor was purified from membrane preparations of Kit-expressing cells by wheat germ agglutinin chromatography. The membranes were prepared as described (43). Briefly, cells were washed with PBS and scraped into PBS containing 1 mM EGTA. Following a 10-min centrifugation at 1500 rpm, the cell pellet was treated with a hypotonic buffer (10 mM Hepes (pH 8), 1 mM EGTA, 1.5 mM MgC12, 1 mM phenylmethylsulfonyl fluoride, 10 pg/ml leupeptin, and 10 pg/ml aprotinin), incubated for 10 min on ice, and homogenized. The nuclei were precipitated by centrifugation at 1500 X g for 10 min. The membranes were precipitated from the supernatant by a 30-min centrifugation at 14,000 rpm at 4 "C. The membrane preparations were solubilized in lysis buffer (34) and incubated with wheat germ agglutinin-agarose for 90 min. Following column washing, the receptor was eluted with 0.2 M N- acetylglucosamine in 20 mM Hepes (pH 7.4).

Cross-linking of Purified Receptor-Wheat germ agglutinin-puri- tied receptor in 0.3 ml was incubated with SCF for 90 min at 4 "C. The samples were transferred to 22 "C, and the cross-linking reagent BS3 was added to a final concentration of 0.5 mM. After 30 min of incubation, the reaction was terminated by addition of glycine to 150 mM. The samples were diluted in lysis buffer and immunoprecipitated as described previously (34).

In Vitro Kinuse Assay-This was carried out on immunoprecipitates in 50 p1 of 20 mM Hepes (pH 7.5), 150 mM NaC1,20% glycerol, 0.1% Triton X-100 containing 10 mM MnClz and 5 pCi of [y-3'P] ATP for 20 min at 22 "C. The samples were washed with H', M', and L' washing solutions (34); boiled for 5 min in sample buffer; and separated by SDS-PAGE.

Radiolabeling and Ligand Binding Assay-SCF was labeled with lZ5I (Amersham Corp.) using chloramine T essentially as described (44). For direct binding analysis, confluent cells in 24-well dishes were washed with PBS and incubated for 90 min at 22 "C with increasing concentrations of '"I-SCF in 0.25 ml of binding buffer (RPMI 1640 medium containing 1 mg/ml bovine serum albumin and 25 mM Hepes (pH 7.5)). The dishes were rinsed four times with ice- cold PBS, the cells were solubilized in 1 ml/well solubilization buffer (0.1 N NaOH, 0.1% SDS), and radioactivity was counted in a y- counter. Nonspecific binding was determined by parallel binding experiments in the presence of a 100-fold excess of unlabeled SCF. For binding displacement analysis, confluent cells in 24-well dishes were incubated with a constant amount of '"I-SCF (-5% of saturation) and increasing concentrations of unlabeled SCF for 90 min at 22 "C (45). The cells were washed, solubilized, and counted as described above. The binding at each concentration was expressed as a percentage of the binding in the absence of competitor SCF.

Down-regulation Experiments-Confluent cells were washed with PBS and incubated with SCF in binding buffer at 37 "C for various periods of time. After extensive washing with ice-cold PBS, the cells were treated for 2 min with acidic PBS containing 60 mM acetic acid (pH 3.8) to dissociate receptor-bound SCF from the cell surface. The cells were then washed four times with PBS and incubated with '"I- SCF (20 ng/ml) for 90 min at 4 "C, followed either by cross-linking with EDAC, immunoprecipitation, and SDS-PAGE, or by direct counting of the cell-bound radioactivity.

RESULTS

Analysis of SCF Binding to Kit-Chinese hamster ovary cells overexpressing the human Kit receptor (T18) were used to measure the binding affinity of human SCF (hSCF) (Fig. 1). Scatchard analysis of the binding data revealed a single dissociation constant with a Kd of 2 X lo-' M and 1.75 X lo6 binding sites/cell. Displacement experiments with unlabeled SCF gave similar results, and the Kd values obtained from different experiments were between and 5 X 10"O M. To address the question of species specificity of receptor-ligand interactions, we used displacement analysis. The apparent affinities of hSCF for mouse Kit and of rat SCF (rSCF) for human and mouse Kit were determined, and the results are summarized in Table I. As shown, human SCF practically does not bind to mouse Kit, whereas rat SCF binds to mouse Kit with high affinity (Kd = 1-5 X lo-' M), but it has -100-

15972 Dimerization of K i t

A 0.

- 1 I

FIG. 1. Scatchard analysis of '"1-SCF binding to intact cells expressing Kit receptor. Confluent cells in 24-well dishes were incubated for 90 min at room temperature with different concentrations of '"I-SCF. Upper, direct binding data plotted as bound versw free ''"I-SCF concentration; lower, Scatchard plot of the data shown in A. Nonspecific binding was determined by parallel binding experiments in the presence of a 100-fold excess of unlabeled SCF.

TABLE I Receptor bindingparameters for SCF from different species

The dissociation constants for hSCF and rSCF to hKit were determined from displacement curves of '''1-hSCF with unlabeled hSCF and rSCF, respectively, whereas the dissociation constants for hSCF and rSCF to mKit were determined from displacement curves of '"I-rSCF with hSCF and rSCF, respectively. Each experiment was repeated four times, and the results represent the average values.

Receptor Ligand K d RM

hKit hSCF 0.5-1 hKit rSCF 10-50 mKit hSCF >700 mKit rSCF 1-5

fold reduced affinity for human Kit as compared with hSCF. These results are consistent with the observation of Martin et al. (29) on the dose-dependent proliferative effect of hSCF and rSCF on murine cells.

Ligand-induced Dimerization of Kit-We used chemical cross-linking to study the ligand-induced dimerization of human Kit in intact cells. T18 cells, either labeled with ["SI methionine or unlabeled, were incubated with increasing concentrations of SCF at 4 "C for 90 min. Then the cells were transferred to 22 "C and incubated with the cross-linking reagent EDAC for 1 h. Lysates from [3sS]methionine-labeled cells were immunoprecipitated with antipeptide antibodies directed against the COOH-terminal tail of Kit (Ab212) and analyzed by gel electrophoresis (Fig. 2A). Lysates from unlabeled cells were separated by SDS gel electrophoresis, transferred to nitrocellulose, and immunoblotted with Ab212 (Fig. 2B). The results obtained by these two independent methods showed that in the absence of the ligand, the receptor migrated as a monomer, whereas in the presence of the ligand, a new band of -320 kDa appeared. This protein band most likely represents a dimeric receptor-ligand complex. Dimerization appear? to include 50% of the receptors at 0.03 nM ligand and reaches cnrnpletion at 3 nM SCF, suggesting that receptor dimerization is complete a t saturation of the receptor's binding sites. It is noteworthy that no higher oligomers were observed, and even at very high concentrations of SCF (10

- 1 8 0

"116

- 1 8 0

- 116

FIG. 2. Dimerization of Kit receptor in response to SCF binding. [:"SS]Methionine-labeled ( A ) or unlabeled ( R ) T18 cells were incubated with the indicated concentrations of SCF for 90 min at 4 "C. The cells were transferred to 22 "C and incubated with 15 mM EDAC for 1 h. Following solubilization, the lysates were either subjected to immunoprecipitation with antibodies against Kit (Ab212) and analyzed by SDS-PAGE on a 5% acrylamide gel ( A ) or directly separated by 5.5% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with Ab212 ( B ) .

SCF : - - + +

- 84 "116

180

I I6

FIG. 3. SCF-induced tyrosine-phosphorylated dimers of Kit receptor. Upper, "'P-labeled T18 cells were incubated with (+) or without (-) SCF for 15 min a t 22 "C, followed by 15 min of incubation in the presence (+) or absence (-) of 15 mM EDAC. The cells were solubilized, and the tyrosine-phosphorylated proteins were immunoprecipitated with anti-phosphotyrosine antibodies. Following elution with phenyl phosphate, the eluates were immunoprecipitated with anti-Kit antibodies (Ab212), washed, and analyzed by 7.5% (left panel) or 5.5% (right panel) SDS-PAGE. Lower, cells were incubated with various concentrations of SCF for 90 min at 4 "C and then cross- linked with EDAC for 1 h. The cell lysates were immunoprecipitated with Ab212. The immunoprecipitates were washed and subjected to an in vitro kinase assay as described under "Materials and Methods." The samples were analyzed by 5.5% SDS-PAGE and autoradiography.

pg/ml), no monomers were observed (data not shown). Kinase Activity of Monomers and Dimers of Kit Receptor-

To examine the kinase activity of the monomeric and dimeric forms of Kit, we analyzed the ligand-dependent autophos- phorylation of the receptor in living cells and in vitro. Fig. 3 (upper) demonstrates that Kit undergoes tyrosine phosphorylation in response to binding of SCF. The oligomerization state of the phosphorylated receptor was determined by immunoprecipitation of the tyrosine-phosphorylated receptor from ["P]orthophosphate-labeled cells following stimulation with SCF and cross-linking with EDAC. Fig. 3 (upper right panel) demonstrates that when the cross-linking reagent was applied to the cells, only half of the tyrosine-phosphorylated

Dimerization of Kit 15973

receptor population appeared in dimers. This relatively low ratio (compare with Fig. 2) is presumably due to incomplete cross-linking during the short incubation time with EDAC (15 min). In fact, it is quite possible that all the phosphorylated receptor molecules are in the dimer form during stimulation in vivo. This possibility was supported by an in vitro kinase assay in which T18 cells were incubated with various SCF concentrations, followed by cross-linking with EDAC, immunoprecipitation with Ab212, and an in vitro kinase assay. As shown in Fig. 3 (lower), the dimer was the major ligand-dependent phosphorylated species, whereas the monomer showed only a low level of phosphorylation that was not affected by ligand binding. This result supports the conclusion that the tyrosine kinase activity is predominantly associated with receptor dimers in the Kit system.

Distribution of SCF between Kit Monomers and Dimers- To characterize the distribution of the ligand between monomeric and dimeric receptor forms in living cells, we incubated T18 cells with human '2sI-SCF, followed by cross-linking and immunoprecipitation with Ab212 as described above. Fig. 44 demonstrates that the ligand was almost exclusively present in the dimeric form of the receptor. Even at low SCF concentration, no ligand-bound receptor monomer was detectable, suggesting that ligand bound to receptor dimers is the major species at all SCF concentrations. Moreover, increasing concentrations of unlabeled ligand, although competing with "'1- SCF for binding to the receptor, did not lead to the appearance of labeled ligand-bound receptor monomers (Fig. 4 B ) . SCF is a noncovalently held dimeric molecule (31, 331, and as such it may be sufficient to induce receptor dimerization. Our results, however, imply that dimer formation may be driven also by an intrinsic receptor property. Otherwise, a t high ligand concentration, dimer dissociation is expected due to monovalent ligand binding to each receptor within the dimer. If indeed receptor dimers are held also by inter-receptor interactions, then receptor dimers may exist even if only one of the two sites is occupied by a ligand. To test this hypothesis, we analyzed dimer formation between the human and mouse Kit proteins.

Heterodimerizatwn between Human and Mouse Kit Recep- tors-The previous results suggest that in a mixture of human and mouse Kit, the binding of human SCF to human Kit would lead to dimerization with mouse Kit, even though the latter does not bind human SCF (Table I). To facilitate the analysis, we used the deletion mutant hKitACT, which lacks the COOH-terminal 42 amino acids (38). The wheat germ agglutinin-purified mouse Kit receptor and the human

A. [I?] !XF(nM) B. SCF (nM)

n r -n w m - 0 0 n c n

0

+' _"<" " . o-olUlo'VZN

m a - - Yr(kDa1 -i - 180 - - 116 . - 04 -

FIG. 4. Cross-linking of ""I-SCF to Kit receptor. Confluent T18 cells were incubated for 90 min at 4 "C either with different concentrations of lZ5I-SCF (A) or with 3 nM lZ5I-SCF in the presence of increasing concentrations of unlabeled SCF ( B ) . Cross-linking was performed with 15 mM EDAC for 1 h at 22 "C. The '"I-SCF-Kit complexes were immunoprecipitated with anti-Kit antibodies (Ab2121 and analyzed by 5.5% SDS-PAGE.

KitACT mutant were incubated either separately or together with '251-hSCF. Following binding, the cross-linking reagent BS3 was added, and the radiolabeled ligand-receptor complexes were immunoprecipitated either with mAb94, which recognizes the human extracellular region but not the mouse Kit receptor (Fig. 5B) , or with Ab212, which is directed against the COOH-terminal portion of the receptor and therefore recognizes only the mouse Kit receptor in this system. Fig. 5A demonstrates that Ab212 could not immunoprecipi- tate a Iz5I-hSCF cross-linked to Kit dimers unless the human receptor was present. This indicates that the human Kit receptor can form heterodimers with the mouse receptor even though only the human receptor site is occupied by the ligand. Conceivably, dimerization of the Kit receptor can be induced by monovalent ligand binding. Similar results were also obtained in living cells that coexpressed the mouse Kit receptor and either the human KitAKI (devoid of the kinase insert) or human KitACT mutant (data not shown). Since the fraction of heterodimers between human and mouse Kit in living cells was smaller in comparison with the homodimers of hKit receptors, it is possible that dimerization via monovalent ligand binding is relatively easily dissociable, whereas bivalent binding stabilizes the dimer form. To study the effect of monovalent ligand binding on the Kit receptor, we utilized cell lines that coexpressed the mouse and human Kit receptors.

Increased Binding Affinity and Receptor Down-regulation of hKit in Heterodimers-hSCF binds 100-fold more avidly than rSCF to human Kit, whereas mouse Kit binds only rSCF (Table I). To determine the apparent affinity of each ligand for the mouse and human Kit receptors, in cells that express either one type of receptor or both receptors, we carried out ligand displacement analyses. Fig. 6 shows that 50% displacement of '2sI-hSCF by unlabeled hSCF was obtained at 0.4 nM in cells that express only the human Kit receptor or in cells that coexpress the mouse and human Kit receptors. However, 50% displacement by rSCF was accomplished with 40 nM in cells that express the human receptor alone as compared with only 3 nM in cells that coexpress both mouse and human Kit. This observation was also reflected in a cross-linking experiment. In cells that express only hKit, no displacement of hSCF by rSCF was observed, whereas in cells that coexpress hKit and mKit, rSCF significantly reduced the binding of 9 - hSCF to hKit (Fig. 6, lower). The increase in affinity of rSCF for the human Kit receptor in the presence of the mouse Kit receptor is likely to result from heterodimerization of human

A. B. Kit I HKI MK H K M K I p A b : 'u

K l t : IpAb: 94 212 c ( K D a ) N N N O1 Mr(KDa) - w ".*..2q

- 180

-116

- 180

* -116

- 84 - 58 - 485

FIG. 5. Heterodimerization of human and mouse Kit receptors. A, the wheat germ agglutinin-purified hKitACT mutant ( H K t ) and mouse Kit ( M K ) were incubated either separately or together as shown. '*'I-hSCF was then added; and after 90 min a t 4 "C, the cross- linking reagent BS3 was added, and the incubation was continued for

precipitated either with Ab212 or with mAb94 as indicated and 30 min a t 22 "C. The receptor-ligand complexes were then immuno-

analyzed by 5.5% SDS-PAGE. B, SDS-PAGE analysis of immunoprecipitates ( I p ) of the [35S]methionine-labeled mouse Kit receptor either by Ab212 or by mAb94, demonstrating the specificity of Ab94.

15974 Dimerization of Kit

CELL LINE : !+Kit .,MKKij

rSCF : - + - + Mr(kDo)

- I80

"116

FIG. 6. Displacement analysis of hSCF binding. Upper, cells that either express the human Kit receptor or coexpress human and mouse Kit were incubated with a constant concentration of '"I-hSCF and increasing concentrations of either unlabeled hSCF or unlabeled rSCF for 90 min a t 22 "C. Cell-associated radioactivity was determined after extensive washing and expressed as a percentage of ""I- hSCF binding when present alone. The displacement curves of '*'I- hSCF by hSCF were identical in both cell lines (0). The displacement curves of ""I-hSCF by rSCF were determined by four independent experiments on cells that express human Kit alone (0) or that coexpress human and mouse Kit (0). Lower, cells that express the human Kit receptor (H-Ki t ) or that coexpress human and mouse Kit (MH-Ki t ) were incubated with '*'I-hSCF in the presence (+) or absence (-) of unlabeled rSCF (150 ng/ml). After 90 min a t 4 "C, the cross-linking reagent EDAC was added for 1 h a t 22 "C. The receptor- ligand complexes were immunoprecipitated with mAb94 and analyzed by 5.5% SDS-PAGE. The autoradiograph is shown together with the apparent molecular masses of marker proteins.

CELL LINE : H-Kit MH-Kt1

Time(hr) : C(kDa1

. I80

I I 6

FIG. 7. Down-regulation of human Kit receptor by rSCF. Cells that express human Kit (H-Ki t ) or that coexpress human and mouse Kit (MH-Ki t ) were incubated with rSCF (100 ng/ml) for the indicated periods of time a t 37 "C. Following washing with acidic PBS, the cells were incubated with "'I-hSCF for 90 min at 4 "C. The cross-linking reagent EDAC was added and incubated with the cells for 1 h a t 22 "C. Receptor-ligand complexes were then immunoprecipitated with mAb94. The samples were washed and analyzed by 5.5% SDS-PAGE and autoradiography.

and mouse Kit. I t is possible that human Kit in the heterodimers undergoes a conformational change that increases its affinity for rSCF. An analogous observation was made for the PDGF receptors, where the binding of PDGF-AB to the a- receptor induces a conformational change that facilitates ligand binding to the b-receptor (46-48). This phenomenon in Kit was supported by independent experiments in which we compared the effect of rSCF on the down-regulation of human Kit when expressed either alone or together with mouse Kit. As shown in Fig. 7, preincubation of cells expressing human Kit alone with rSCF at 37 "C induced only limited down- regulation of the human Kit receptor, whereas cells expressing

both the mouse and human receptors displayed remarkable down-regulation of the human Kit protein in a time-dependent manner. Quantitative analysis of receptor down-regulation in both cell lines was performed by measuring the binding of '251-hSCF to the receptors remaining after incubation with rat SCF (Table 11). It is shown that in the cell line that coexpressed both receptors, rSCF down-regulated human Kit almost as well as hSCF did. In contrast, in cells expressing only human Kit, there was only a small down-regulation effect by rSCF.

Inhibition of Kinase Activation and Phosphatidylinositol 3'- Kinase Association by hKit Mutant in Heterodimers-Some W mutants display a severe phenotype in the heterozygous state. This phenomenon is thought to be due to heterodimers between wild-type and mutant receptors, where the mutant receptor inhibits signal transduction by the wild-type receptor (14-16). To characterize the effect of a mutant receptor on signal transduction by the wild-type receptor, we utilized cell lines that coexpress wild-type mKit and a hKit mutant (hKitAKI) (38) that is devoid of the kinase insert. hKitAKI exhibits a partially impaired tyrosine kinase activity and does not associate with phosphatidylinositol 3'-kinase (38). I t therefore experimentally represents a partially kinase-defective mutant of the Kit receptor. We compared the tyrosine phosphorylation of mKit and its association with phosphatidylinositol 3"kinase in cells that either express mKit alone or coexpress mKit and hKitAKI. To allow differential receptor analysis, we used a monospecific antibody (Ab213) directed against the kinase insert (38) as the mKit-specific antibody and mAb94 as the hKitAKI-specific antibody. Fig. 8 (A and B ) shows that upon stimulation by rat SCF, the level of tyrosine phosphorylation of mKit was markedly reduced by the presence of hKitAKI. This effect was observed using either immunoprecipitation with Ab213 and immuno- blotting with anti-phosphotyrosine antibodies (Fig. 8A) or immunoprecipitation with anti-phosphotyrosine and blotting with Ab213 (Fig. 8B). The reduction in tyrosine phosphorylation of mKit in the presence of hKitAKI was not due to a difference in the level of the expressed protein, as is evident from the results shown in Fig. 8C. Furthermore, no reduction in tyrosine phosphorylation was observed in a cell line that coexpressed the two wild-type receptors, namely hKit and mKit (data not shown).

We have shown previously that the kinase insert of Kit is the binding site for phosphatidylinositol 3'-kinase (38). Therefore, the hKitAKI mutant was used to analyze its effect on phosphatidylinositol 3"kinase association with wild-type

TABLE I1 Down-regulation of hKit by rSCF or hSCF

Cells that express human Kit (HKit) or coexpress human and mouse Kit (MHKit) were incubated for the indicated periods of time with either rSCF (150 ng/ml) or hSCF (75 ng/ml) or binding buffer alone at 37 "C. After washing, the cells were incubated with '*'I-hSCF for 90 min a t 22 "C to determine residual receptors. Cell-associated radioactivity was determined after extensive washing and expressed as percentage of ligand binding prior to down-regulation with SCF. The concentration of rSCF was 100 ng/ml, and that of hSCF was 50 ne/ml.

Binding of '2sII-hSCF after down-regulation

Time HKit MHKit

rSCF hSCF rSCF hSCF h %

0.5 96 48 65 58 1.5 87 34 48 35 3 76 28 38 27 5 71 23 30 25

Dimerization of Kit 15975 A

CELL LINE MK MHAKI SCF: -7 &lkDal

- IS0

- I16

- 84 - 58

I P Ab: 213 I 6 Ab: P T Y R

C CELL LINE:,MK ,YAK!

IPAb: 2139421394 &lkDal

' "-116 - IS0

0 CELL LINE MK M H M

rSCF r*T Mr(k0o) -

IPAb: P-TYR IEAb: 213

- IS0

-I I 6

- 84 - 5s - 4 8 5

D CELL LINE: MK MHAU

I P Ab: AcK2 I6 Ab: PI3K

I so I16 84 58

FIG. 8. Inhibition of mouse Kit tyrosine phosphorylation and association with phosphatidylinositol 3"kinase by hKitAKI mutant. Quiescent cells that either express mouse Kit ( M K ) or coexpress the mouse Kit and hKitAKI mutant ( M H A K I ) were stimulated either with hSCF or with rSCF for 10 min at 37 "C and analyzed as follows. A, cell lysates were immunoprecipitated ( I P ) with antipeptide antibodies directed against the kinase insert (Ab213) and immunoblotted ( I B ) with anti-phosphotyrosine antibodies. B, cell lysates were immunoprecipitated with anti-phosphotyrosine antibodies and immunoblotted with Ab213. C, the expression level of each receptor in both cell lines was determined by immunoprecipitation with mAb94 or with Ab213. Immunoprecipitates were separated by 7.5% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with Ab212. D, association of phosphatidylinositol 3"kinase with the mouse Kit receptor. Following stimulation with rat SCF, the cells were solubilized, and the mouse Kit receptor was immunoprecipitated with the mouse-specific monoclonal antibody AcK2 (31). The immunoprecipitates were washed, separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with anti-pS5a antibodies. The p85a associated with the mKit receptor in each cell line is indicated by the arrow. The antibodies on the immunoblots were detected by the ECL method (Amersham Corp.) using protein A coupled to horseradish peroxidase.

mKit in cells coexpressing both receptors. Cells that express mKit alone or coexpress mKit and hKitAKI were stimulated with rSCF, and mouse Kit was immunoprecipitated with a mouse-specific monoclonal antibody (49). After separation of the immunocomplexes by gel electrophoresis, the gel was blotted with anti-phosphatidylinositol 3"kinase antibodies, which detect the regulatory p85a subunit of phosphatidylinositol 3"kinase. The results demonstrate that the association of p85 with mKit was markedly reduced in the cell line that coexpressed mKit and hKitAKI. These results suggest an inhibitory effect of the defective hKitAKI on tyrosine phosphorylation and downstream events in the signal transduction by mKit.

DISCUSSION

The binding of ligands to cell receptors is a prerequisite for delivering signals into the cell. I t is generally accepted that receptor oligomerization is the major consequence of ligand binding and is in fact essential for signal transduction (3-6, 37,50,51). Generally, the receptors for growth factors contain one ligand-binding site. The ligands, on the other hand, vary from being monomers (EGF, transforming growth factor-cy, fibroblast growth factor), to noncovalent dimers (SCF), to covalent dimers (PDGF, colony-stimulating factor-1). The mechanism of receptor oligomerization therefore must be separated form the oligomeric nature of the ligand. In this report, we have analyzed ligand-induced dimerization of Kit

using two different ligands, hSCF and rSCF, and two different receptors, human and mouse Kit. Since human Kit binds rSCF with only low affinity and mouse Kit binds rSCF but not hSCF (Table I), this approach allowed us to analyze several parameters of the relationships between the valence of ligand binding and receptor dimerization and activation. The binding of hSCF to hKit demonstrates several important features of ligand-induced receptor dimerization. Dimeriza- tion is complete, and even at low SCF concentrations, the ligand is associated mainly with receptor dimers. At ligand saturation (-1 nM), only receptor dimers are observed. The fact that no higher oligomers are present may suggest that the bivalence of the ligand is responsible for dimer formation. This, however, is inconsistent with the finding that ligand- labeled monomers are not observed either. Particularly strik- ing is the fact that even a t high ligand excess (Fig. 4), no monomeric ligand-labeled receptors exist in spite of the fact that the unlabeled ligand displaced the labeled ligand propor- tionally to its concentration. It is conceivable that at high ligand excess, all the binding sites are occupied, and therefore most of the ligand will be bound monovalently; nevertheless, only dimeric receptor complexes are observed. This indicates that the dimeric form of the receptor-ligand complex is independent of the bivalence of the ligand, as may be the situation with monovalent ligands and their receptors (e.g. EGF receptor) (4-6, 50, 52). We therefore conclude that ligand binding exposes, probably through a conformational change, a dimerization site in the receptor itself that keeps receptors as dimers even though each binding site may be occupied by a different ligand molecule through monovalent binding. To further examine this interpretation and to test whether monovalent binding to one receptor can induce dimers, we analyzed the formation of heterodimers between the human and mouse receptors through the binding of hSCF, which does not bind to mKit. The results demonstrate that in this situation, human-mouse Kit heterodimers are formed. We therefore conclude that monovalent binding of hSCF to hKit can induce the formation of human-mouse Kit heterodimers through a dimerization site in the receptor itself. This interpretation is in variance with the mechanism of dimerization suggested by others for the PDGF receptor (53) and Kit (37). In these models, the only factor that determines dimer formation is the bivalence of the ligand. This conclusion was partly based on the observation of a bell-shaped curve for dimerization when ligand dose dependence of dimers was analyzed. Our results a t high ligand excess and with human-mouse Kit heterodimers are inconsistent with a model of simple interaction between a bivalent ligand and a monovalent receptor. Rather, we prefer a model of ligand-induced conformational change in the receptor's extracellular region (Fig. 9) that exposes a receptor dimerization site and appears to increase receptor affinity for the ligand.

The mechanisms of receptor dimerization shown in Fig. 9 are supported by several observations on the response to SCF of cells that coexpress human and mouse Kit. First, the affinity of hKit for rSCF is increased by -10-fold in these cells (Fig. 6). Second, the down-regulation of hKit in cells coexpressing mKit and hKit is markedly different from that observed in cells expressing only hKit (Fig. 7). Whereas in the latter rSCF does not induce down-regulation of hKit (Table 11), in the cells that coexpress both receptors, the down-regulation of hKit by rSCF is comparable to that observed with hSCF in cells expressing hKit alone. Taken together, the results imply that upon binding of rSCF, hKit must be associated with mKit as heterodimers in cells that coexpress both receptors. These results show similarity to

15976 Dimerization of Kit

1. LOW-AFFINITY 2. CONFORMATIONAL 3. UNSTABLE 4. HIGH-AFFINIlY 5. DIMERSAT LIGAND BINDING CHANGE DIMER LIGAN0 BINDING EXCESS LIGAND (MONOVALENT) (STABLE DIMER)

FIG. 9. Sequential model of binding and activation of Kit by SCF. The ligand dimer that first binds to a monomeric receptor (step 1 ) induces a conformational change that exposes the receptor’s dimerization site (step 2). This facilitates rapid dimerization of receptors (step 3) , which may be stabilized by bivalent ligand binding (step 4 ) and remain in the dimer form even at excess ligand (step 5). The model allows for the formation of dimers by monovalent ligand binding, for heterodimer formation between receptors from different species, and for heterodimer formation between receptors with different ligand specificity.

those obtained with the PDGF receptors as it was shown that the affinity of PDGF-AB for the 8-receptor increases in the presence of the a-receptor. Furthermore, down-regulation of the 8-receptor by PDGF-AB is enhanced in the presence of the a-receptor (46-48). Finally, the formation of heterodimers can explain the reduced tyrosine phosphorylation of mKit coexpressed with the hKitAKI mutant when stimulated with rSCF (Fig. 8). hKitAKI lacks the putative major autophos- phorylation sites of the receptor, i.e. the tyrosine(s) in the kinase insert, and does not bind phosphatidylinositol 3’- kinase (38). Therefore, it represents a functionally defective mutant of Kit. The results indicate that this mutant causes a reduction in the phosphorylation of mKit when stimulated by rSCF. We also observed a significant reduction in the amount of the phosphatidylinositol 3”kinase associated with the activated mKit. Hence, the formation of heterodimers affects the binding affinity, internalization, and down-regulation of the receptor as well as downstream functions such as phosphorylation and association with cellular substrates. Presum- ably, transphosphorylation within the dimer is involved in the functional dominance of the mutant receptor.

Recently, we generated a secreted recombinant extracellular portion of Kit that fully retained ligand binding activity. We have demonstrated that this protein undergoes ligand- induced dimerization in a way similar to that described here for the full-length receptor (41), suggesting that the putative receptor dimerization site resides in the extracellular region of the receptor. This also implies that the kinase activity of the receptor is not required for its dimerization. On the other hand, the results presented in Fig. 3 clearly show a correlation between dimer formation and kinase activity, implying that dimerization is a prerequisite for kinase activation.

In conclusion, our results favor the following model of ligand binding and activation of the Kit receptor (Fig. 9). Accordingly, the initial receptor-ligand complex is monovalent. This induces a conformational change in the receptor, and concomitantly a dimerization site is exposed. We predict that the putative site is monovalent (namely binds only one receptor) and resides on the extracellular domain. Although an activated dimerization site can interact with an unoccupied receptor, the result is a relatively unstable dimer as reflected by human-mouse Kit heterodimers (Fig. 5). Dimer stabilization is achieved by bivalent ligand binding, which “opens” the other dimerization site and also contributes the noncovalent intra-SCF bond. Conceivably, the formation of this stable dimer is facilitated by the relatively high affinity of binding of the second ligand arm, as indicated by the data shown in

Fig. 6. This effect may be attributed to the retarded diffusion of the single-arm bound SCF. An interesting question that is related to this possibility is why receptor dimers do not undergo dissociation at large excess of the ligand? One possibility is shown in Fig. 9, namely a second ligand binds to the dimer, but trimers and tetramers are not formed because of the monovalence of the dimerization site.

The dimerization model and our findings that a mutant receptor reduces kinase activity and coupling to an effector protein (Fig. 8) may provide a structural basis for understand- ing the phenomenon of negative dominance in heterozygous W mice. The negative effect either may result from lack of transphosphorylation within the heterodimer or may be due to the absence of a major phosphorylation site, as is the case in our model system (hKitAK1). Such transdominant inhibitory actions of mutant receptors are not limited to Kit and have been recently demonstrated for the PDGF receptor (13) and the EGF receptor (54). Therefore, the results we presented may be relevant also to the physiology of other receptors for growth factors.

Acknowledgments-We thank K. Zsebo (Amgen) for recombinant rSCF and hSCF, Etta Livneh for the pZL retroviral vector, M. A. Dieckmann for pSV2hph, S. Nishikawa for monoclonal antibody AcK2 against mouse Kit, and E. Peles for antibodies to phosphotyrosine.

REFERENCES

2. Ullrich, A., and Schlessinger, J. (1990) Cell 61,203-212 1. Yarden, Y., and Ullrich, A. (1988) Annu. Reu. Biochem. 67,443-478

3. Li, W., and Stanley, E. R. (1991) EMBO J. 10,277-288 4. Schlessinger, J. (1988) Biochemistry 27,3119-3123 5. Yarden, Y., and Schlessinger, J. (1987) Biochemistry 26,1434-1442 6. Yarden, Y., and Schlessinger, J. (1987) Biochemistry 26,1443-1451 7. Honegger, A. M., Kris, R. M., Ullrich, A., and Schlessinger, J. (1989) Proc.

Natl. Acud. Sci. U. S. A. 86,924-929 8. Kelly, J. D., Haldeman, B. A., Grant, F. J., Murray, M. J., Seifert, R. A.,

Bowen-Popet, D. F., Cooper, J. A., and Kazlauskas, A. (1991) J. Biol. Chem. 266,8987-8992

9. Ohtsuka, M., Roussel, M. F., Sherr, C. J., and Downing, J. R. (1990) Mol.

10. Goldman, R., Ben-Levy, R., Peles, E., and Yarden, Y. (1990) Biochemistry Cell. Biol. 10,1664-1671

11. Wada, T., Quian, X., and Green, M. I. (1990) Cell 6 1 , 1339-1347 29,11024-11028

12. Treadway, J. L., Morrison, B. D., Soos, M. A., Siddle, K., Olefsky, J., Ullrich, A., McClain, D. A., and Pessin, J. E. (1991) Proc. Natl. Acud.

13. Ueno, H., Colbert, H., Escobedo, J. A., and Williams, L. T. (1991) Science Sci. U. S. A. 88,214-218

14. Nocka, K., Tan, J. C., Chiu, E., Chu, T. Y., Ray, P., Traktman, P., and

15. Pawson, T., and Berenstein, A. (1990) Trends Genet. 6,350-356 16. Tan, J. C., Nocka, K., Ray, P., Traktman, P., and Besmer, P. (1990) Science

17. Qiu, F., Prabir, R., Karen, B., Barker, P. E., Jhanwar, S., Ruddle, F. H.,

18. Yarden, Y., Kuan , W.-J., Yang-Feng, T., Coussens, L., Schlessinger, J.,

19. Chabot, B., Stephanson, D. A., Chapman, P., Besmer, P., and Bernstein,

20. Geisller, E. N., Ryan, M. A., and Housman, D. E. (1988) Cell 66,185-192 21. Nocka, K., Majumder, S., Chabot, B., Ray, P., Cervone, M., Bernstein, A.,

22. Geissler, E. N., McFarland, E. C., and Russel, E. S. (1981) Genetics 97,

262,844-848

Besmer, P. (1990) EMBO J. 9 , 1805-1813

247 , 209-212

and Besmer, P. (1988) EMBO J. 7,1003-1011

Francke, U., andiullrich, A. (1987) EMBO J. 6,3341-3351

A. (1988) Nature 336,8&89

and Besmer, P. (1989) Genes & Deu. 3,816-826

RR7-Rfi l 23. Russel, E. S. (1979) Adu. Genet. 20,357-459 24. Silvers, W. K. (ed) (1979) Coat Colors of Mice: A Model for Gene Action and

Interaction, pp. 206-241, Springer-Verlag, New York 25. Anderson, D. M., Lyman, S. D., Baird, A., Wignall, J. M., Eisenman, J.,

Rauch, C., March, C. J., Boawell, H. S., Gimpel, S. D., Cosman, D., and

26. Flanagan, J. G., and Leder, P. (1990) Cell 63,185-194 Williams, D. E. (1990) Cell 63,235-243

27. Co eland, N. G., Gilbert, D. J., Cho, B. C., Donovan, P. J., Jenkins, N. A., Eosman, D., Anderson, D., Lyman, S. D., and Williams, D. E. (1990) Cell

28. Huan E , Nocka, K., Beier, D. R., Chu, T.-Y., Buck, J., Lahm, H.-W., 6 3 , 175-183

29. Martin, F. H., Suggs, S. V., Langley, K. E., Lu, H. S., Ting, J., Okino, K. Wefine;, D., Leder, P., and Besmer, P. (1990) Cell 63,225-233

R., Morris, C. F., McNiece, I. K., Jacobsen, F. W., Mendiaz, E. A,, Birkett, N. C., Smith, K. A,, Johnson, M. J., Parker, V. P., Flores, J. C., Patel, A. C., Fisher, E. F., Erjavec, H. O., Herrera, C. J., Wypych, J., Sachdev, R. K., Pope, J. A., Leslie, I., Wen, D., Lin, C.-H., Cupples, R.

30. Williams, D. E., Eisenman, J., Baird, A., Rauch, C., Van Ness, K., March, L., and Zsebo, K. M. (1990) Cell 63,203-211

C. J., Park, L. S., Martin, U., Mochizuki, D. Y., Boswell, H. S., Burgess, G. S., Cosman, D., and Lyman, S. D. (1990) Cell 63,167-174

31. Zsebo, K. M., Wypych, J., McNiece, I. K., Lu, H. S., Smith, K. A., Karkare,

”. I”

Dimerization of Kit 15977 S. B., Sachdev, R. K., Yuschenkoff, V. N., Birkett, N. C., Williams, L. 43. Yarden, Y., Harari, H., and Schlessinger, J. (1985) J. Biol. Chem. 260, R., Satya al, V. N., Tung, W., Bosselman, R. A., Mendiaz, E. A., and Langley, 8. E. (1990) CeU 63,195-201

32. Zsebo, K. M., Williams, D. A., Geissler, E. N., Broudy, V. C., Martii, F. 45. Lax, I., Johnson, A., Howk, R., Sap, J., Bellot, F., Winkler, M., Ullrich, A., H., Atkins, H. L., Hsu, R.-Y., Birkett, N. C., Okino, K. H., Murdock, D. Vennstrom, B., Schlessinger, J., and Givol, D. (1988) MOL Cell. Bwl. 8, C., Jacohsen, F. W., Lan ley, K. E., Smith, K. A., Takeishi, T., Cattanach, B. M., Galli, S. J., and {uuggs, S. V. (1990) CeU 63.213-224

1970-1978

33. Arakawa, T., Yphantis, D. A., Lary, J. W., Nahri, L. O., Lu, H. S., 46. Hammacher, A. Mellstrom, K., Heldin, C.-H., and Westermark, B. (1989)

EMBO J. 8, i489-2495 Prestrelski, S. J., Clogston, C. L., Zsebo, K. M., Mendiaz, E. A,, Wypych, 47. Heidarm, A. M., Pierce, M. H., Yu, J.-C., Lombari, D., Artrip, J. E., J., and Langley, K. E. (1991) J. BioL Chem. 266,18942-18948

34. Lev, S., Givol, D., and Yarden, Y. (1991) EMBO J. 10,647-654 Fleming, P. T., Thomason, A., and Aaronson, S. A. (1991) J. Biol. Chem.

36. Herhst, R., Lammers, R., Schlessinger, J., and Ullrich, A. (1991) J. Biol. 35. Lev, S., Yarden, Y., and Givol, D. (1990) MOL cell. io^ 10,6064-6068 48. Karnakara, p.9 hi s., Khan, s. A., and Bishayee, S. (1991) Bhhemktry

Chem. 266,19908-19916 49. Nishikawa, S. Kusakahe M., Yoshinage, K., 0 awa, M., Hayashi, S.-I. 37. Blume-Jenaen, P., Claesson-Welsh, L., Siegbahn, A., Szeho, K. M., Wes- Kunisada, ‘f., Era, T., hkakura, T., and Nishifawa, S.-I. (1991) EMBd

termark, B., and Heldin, C. H. (1991) EMBO J. 10,4124-4128 38. Lev, S., Givol, D., and Yarden, Y. (1992) Pm. Natl. Acad. Sci. U. S. A. 89, i:: :$::; ~ ~ ~ ; 1 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ! ~ , ~ ~ ~ o ~ , ~ . ~ ~ ~ ~ 8 ~ , ~ ~ ~ ~ ~ ~ ,

678-682 39. W W r , M., Sweet, R., Sim, K. G., Wold, B., Pellicer, A,, Maniatis, T., 52. L ~ , i., M ~ M , A. K., Avera, c,, Hurwitz, D. R Rubinstein, M Ullrich

40. Eldar, H., Zisman, Y., Ullrich, A., and Livneh, E. (1990) J. Biol. Chem. Silverstein, S., and Axel, R. (1979) Cell 16, 777-785 A., Stroud, R. M., and Schlessinger, J. (1991) J.’BioL Chem. 266,13828-1

265,13290-13296 13833 41. Lev, S., Yarden, Y., and Givol, D. (1992) J. Biol. Chem. 267,10866-10813 53. Heldin, C.-H., Emlund, A., Rorsman, C., and Ronnstrand, L. (1989) J. 42. Escowo, J. A., Navankasattusas, s., Kavanaugh, W. M., Milfay, D., Fried, 54. Kashles O., Yfvden, Y., Fischer, R., Ullrich, A., and Schlessinger, J. (1991)

Bwl. Chem. 264,8905-8912

315-319 44. Hunter, W. M., and Greenwood, F. C. (1962) Nature 194,495-496

266,20232-20237

30, 1761-1767

J. 10,2111-2118

M. and Bowen-Pope D. F. 6689) J. Biol. Chem. 264,8771-8778

V. A., and Williams, L. T. (1991) Cell 66,75-82 Mol. keU. Bwl. 11,1454-1463

the journal of vol. no. 22, of pp. by for u. s. a ... · dimerization and activation of the kit...

Documents