Regulation of Ubiquitin Protein Ligase Activity in c-Cbl by Phosphorylation-induced Conformational Change and Constitutive Activation by Tyrosine to Glutamate Point Mutations*

c-Cbl down-regulates receptor tyrosine kinases by conjugating ubiquitin to them, leading to receptor internalization and degradation. The ubiquitin protein ligase activity of c-Cbl (abbreviated as E3 activity) is mediated by its RING finger domain. We show here that the E3 activity of c-Cbl is negatively regulated by other domains present in the amino-terminal half of the protein (the TKB and linker helix domains) and that this negative regulation is removed when the protein is phosphorylated on tyrosine residues. Protease digestion studies indicate that tyrosine phosphorylation alters the conformation of c-Cbl. We also show that mutation of certain conserved tyrosine residues to glutamate can constitutively activate the E3 activity of c-Cbl. In particular, a Y371E mutant shows constitutive E3 activity while retaining the ability to bind epidermal growth factor receptor (EGFR). The Y371E mutant also has altered protease sensitivity from wild type, instead resembling the proteolytic pattern seen with tyrosine-phosphorylated c-Cbl. Mutation of the homologous tyrosine residue in Cbl-b to glutamate also leads to E3 activation while retaining EGFR-binding ability. These studies argue that Tyr-371 plays a key role in activating the E3 activity of c-Cbl and that the Y371E mutant may partially mimic phosphorylation at that site. However, Tyr-371 point mutants of c-Cbl are still able to undergo phosphorylation-induced E3 activation, and we show that Tyr-368 can also be phosphorylated in addition to Tyr-371, and contributes to activation.

The c-Cbl proto-oncogene was first discovered as the cellular homologue of v-Cbl, a viral transforming gene from the Cas NS-1 murine retrovirus, which causes pre-B cell lymphomas and myelogenous leukemias in mice (1). The transforming gene v-Cbl is a truncation mutant of c-Cbl, which itself does not transform cells (2). Human c-Cbl encodes a widely expressed cytosolic protein of 906 amino acids, which is a prominent substrate of a variety of tyrosine kinases and undergoes binding interactions with a large number of intracellular signaling molecules (3,. Two other related genes exist in mammals, Cbl-b (9) and Cbl-3 (10), and Cbl homologues have been identified in Caenorhabditis elegans (11) and Drosophila (12).
The amino terminus of c-Cbl contains a conserved functional domain that binds phosphotyrosine, composed of a four-helix bundle, a Ca 2ϩ -binding EF hand domain, and a variant SH2 domain (13). These structural components together comprise a functional unit that has been called the TKB domain (for tyrosine-kinase binding). A short helical linker region connects the TKB domain to a RING finger domain that contains two bound zinc ions. The TKB, linker helix, and RING domains are well conserved among all Cbl family members.
Important clues regarding Cbl function have come from genetic studies in C. elegans and Drosophila, in which it was found that Cbl family members act to negatively regulate receptor tyrosine kinases (11,12). Subsequent studies in mammalian cells have confirmed this negative regulatory role of Cbl, and a mechanism for the down-regulation of receptor tyrosine kinases by Cbl has been provided by the discovery of ubiquitin protein ligase activity in c-Cbl, mediated by the RING finger domain (14 -19).
Ubiquitin protein ligases (abbreviated as E3s) 1 are part of a multienzyme system for conjugating ubiquitin to substrate proteins (reviewed in Refs. 20 -22). The first step in this process involves ATP-dependent formation of a thioester between the C terminus of ubiquitin and the active site cystine of ubiquitinactivating enzyme (E1). A second thioester intermediate is subsequently formed between ubiquitin and one of several ubiquitin-conjugating enzymes (abbreviated as E2s, or ubc). In the final step of the process, E2s act in concert with ubiquitin protein ligases (E3s) to form an iso-peptide bond between the carboxyl terminus of ubiquitin and a free amino group on the substrate protein. In most cases, ubiquitin conjugation results in multiubiquitin chains that target the substrate for degradation by proteasomes or, in some cases, lysosomes. In a few cases, conjugation of single ubiquitin moieties can produce a relatively stable post-translational modification of certain substrate proteins.
Two major families of E3s are known, HECT proteins (homology to E6-associated protein carboxyl terminus) and RING finger-containing proteins. Although not all proteins that contain a RING domain are E3s, a large and growing number of RING proteins have been found to have ubiquitin ligase activity. The RING domain is important for binding E2s, and a recent crystal structure of a complex of an amino-terminal fragment of c-Cbl and the E2, UbcH7 has provided precise details of that interaction (23).
The E3 activity of c-Cbl has been reported to be regulated by tyrosine phosphorylation, and a mutational study was done by Levkowitz et al. (19) to try to define the key tyrosine residues in c-Cbl responsible for that activation. In that study, a series of tyrosine to phenylalanine point mutants in c-Cbl were transfected into Chinese hamster ovary cells along with the EGF receptor, and the ability of EGF to induce ubiquitination of EGFR was measured. Wild type c-Cbl and Y to F point mutants at positions 92, 274, 291, 307, 337, and 368 were all able to enhance EGFR ubiquitination under these conditions, whereas Y371F was inactive. The authors concluded that Tyr-371 was the key site for phosphorylation-induced activation of the ubiquitin ligase activity of c-Cbl. However, this interpretation was subsequently called into question by the structural studies of Zheng et al. (23), who stated that this tyrosine residue was not solvent-accessible, and instead was internally hydrogenbonded to threonine 227. These authors implied that tyrosine to phenylalanine mutation at Tyr-371 would disrupt the structure of the enzyme, rather than merely removing a phosphorylation site. These two models for the function of Tyr-371 would appear to be mutually exclusive.
In the studies reported here, we further investigate the mechanism of phosphorylation-induced activation of the ubiquitin ligase activity of c-Cbl. As part of this work, we have attempted to resolve the controversy over which tyrosines are required for activation by generating gain-of-function mutations in c-Cbl to try to avoid the problems inherent in loss-offunction studies.

Construction of GST-Cbl Fusion Proteins and Cbl Point Mutants-
All GST-Cbl fusion proteins were made from corresponding pGEX-Cbl plasmids based on the vector pGEX-4T-1 (Amersham Biosciences). Inserts were constructed by PCR using Pfu polymerase (Stratagene, La Jolla, CA), a cDNA clone of human c-Cbl provided by W. Langdon (University of Western Australia, Perth, Australia) as template, and PCR primers containing compatible restriction sites. pGEX-Cbl-RING contains an insert encoding amino acids 358 -447 of c-Cbl preceded by an EcoR1 site and followed by a stop codon and an NotI site. pGEX-Cbl 1-480 contains a BamH1 site immediately preceding sequences encoding amino acids 1-480 of c-Cbl and followed by a stop codon and an NotI site. c-Cbl point mutants were constructed by PCR. All mutations were verified by sequencing. GST fusion proteins were produced in Escherichia coli strains DH5␣ or BL21 containing the appropriate pGEX plasmid as follows. 500-ml cultures were grown to log phase (A 600 of 0.5-0.7) and induced by addition of isopropyl-1-thio-␤-D-galactopyranoside to 0.1 mM final. After 3-4 h of induction, cells were harvested, rinsed with 50 mM NaCl, 20 mM Tris-HCl, pH 8.0, 1 mM EDTA, and frozen at Ϫ80°C. Cell pellets were later thawed and resuspended in 10 -15 ml of 150 mM NaCl, 20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM DTT, 1 mM PMSF, and 1 g/ml each of leupeptin, pepstatin, bestatin, and E64. Cells were lysed by the addition of 100 g/ml lysozyme and 1% Triton X-100 or 1% Thesit (Roche Applied Science) for 15-30 min followed by sonication. Lysates were cleared by centrifugation for 15 min at 25,000 ϫ g and then incubated with 0.5-1.0 ml of a 50% slurry of glutathione-Sepharose beads for 30 min to 1 h at 23°C or 2-15 h at 4°C with rotation. The beads were washed 3 ϫ 10 ml with the above buffer containing 0.1% Thesit and then placed in a column and eluted overnight at 4°C with 20 mM reduced glutathione, 100 mM Tris, pH 8.0, 50 mM NaCl, 0.01% Thesit at 1-2 ml/h. Protein-containing fractions were pooled, concentrated, and dialyzed first against 50 mM NaCl, 20 mM Tris-HCl, 1 mM DTT, pH 7.5, and then against the same buffer containing 50% glycerol and 0.01% Thesit. Dialyzed proteins were cleared of insoluble material by centrifugation at 13,000 ϫ g at 4°C for 15 min and stored at Ϫ20°C. Protein concentrations were measured by the Bio-Rad protein assay (Bio-Rad, Hercules, CA) using bovine serum albumin as the standard. Tyrosine-phosphorylated proteins were prepared in the E. coli strains TKX1 or TKB1 induced as described by the manufacturer (Stratagene) and lysed as described above except that the lysis buffer also contained 50 mM NaF and 1 mM sodium orthovanadate and lacked DTT. GST-Cbl-b 1-480 and the corresponding Y363E mu-tant were produced by similar methods from a human cDNA clone provided by Stanley Lipkowitz, NCI, National Institutes of Health, Bethesda, MD.
In Vitro Ubiquitination Assay-GST-ubiquitin was produced in E. coli using a pGEX-2TK-based expression plasmid obtained from Martin Scheffner, University of Cologne, Cologne, Germany. This plasmid was introduced into DH5␣ cells, and the cells were induced with isopropyl-1-thio-␤-D-galactopyranoside, and GST-ubiquitin was purified as described for the GST-Cbl fusion proteins above. Purified GSTubiquitin was 32 P-labeled in HMK buffer (100 mM NaCl, 20 mM Tris-HCl, pH 7.5, 12 mM MgCl 2 , 1 mM DTT) containing 0.8 Ci/l [␥-32 P]ATP ([␥-32 P]ATP was obtained from PerkinElmer Life Sciences) and 0.3 unit/l of bovine heart kinase (Sigma #P-2645) for 30 min at 12°C, then the reaction was quenched by addition of 10 volumes of stop solution (10 mM sodium phosphate, 10 mM sodium pyrophosphate, 10 mM EDTA, 1 mg/ml bovine serum albumin, pH 8.0). Glutathione-agarose was then added, and the mixture was rotated at 4°C for 30 min to 1 h. The beads were washed five times with PBS containing 0.1% Thesit and then treated with thrombin in the same buffer for 2 h at room temperature or overnight at 4°C with rotation. Following thrombin cleavage, PMSF was added to 2 mM final, and the beads were spun and washed three times with PBS to recover the [ 32 P]ubiquitin in the supernatant. Wheat E1 and human UbcH5b were expressed in E. coli from the plasmids pETUBA1 (Rick Vierstra, University of Wisconsin, Madison, WI, obtained from Mark Hochstrasser, Yale University, New Haven, CT) and pET15b-UbcH5b (Alan Weissman, NCI, National Institutes of Health, Bethesda, MD), and purified by covalent affinity chromatography on ubiquitin agarose as described (24). In vitro ubiquitination reactions were performed in 20 l of total volume containing 185 ng of E1, 1 g of UbcH5b, and 2 g of GST-Cbl fusion proteins in a buffer containing 50 mM KCl, 20 mM HEPES, pH 7.4, 5 mM MgCl 2 , 1 mM DTT, 1 mM ATP, and an ATP-regenerating system consisting of 10 mM creatine phosphate, 3.5 units/ml creatine phosphokinase, and 0.6 unit/ml of inorganic pyrophosphatase. [ 32 P]Ubiquitin was prepared to specific activities ranging from 3.5 ϫ 10 5 to 4.6 ϫ 10 6 cpm/g, and 100,000 -500,000 cpm was added per reaction. Reactions were set up on ice and then transferred to 30°C for various times; time points were quenched by adding an equal volume of 2ϫ SDS sample buffer followed by boiling.
Protease Digestions-GST fusion proteins were first cleaved with thrombin, treated with PMSF and glutathione-Sepharose to inactivate thrombin and remove GST, and then dialyzed and concentrated into 50 mM NaCl, 20 mM Tris-HCl, pH 7.5, 1 mM DTT, 50% glycerol. The resulting Cbl 1-480 proteins were stored at Ϫ20°C. For digestions, proteins were adjusted to 0.2 mg/ml in buffer lacking glycerol and digested with protease at 10 g/ml at 37°C. Proteinase K was obtained from Sigma, and digests were performed in the above buffer lacking glycerol.
Trypsin digestions were performed similarly using Sequencing Grade trypsin from Promega (Madison, WI) and the following buffer: 100 mM NaCl, 100 mM Tris-HCl, pH 7.9, 1 mM CaCl 2 , 1 mM DTT. Aliquots were removed at the indicated time points, treated with PMSF, and then boiled in SDS sample buffer. Samples were analyzed by SDS-PAGE on 12% gels followed by Coomassie staining.
For identification of phosphorylation sites, 50 g of P-Cbl 1-480 was adjusted to 1% SDS and 5 mM DTT and boiled for 5 min. After cooling, 10 mM iodoacetamide was added, and the sample was incubated 30 min at room temperature in the dark. The sample was then diluted 10-fold into 1% Thesit, 100 mM NaCl, 100 mM Tris-HCl, pH 7.9, 1 mM CaCl 2 and digested with 2.5 g of trypsin for 7 h at 37°C followed by 15 h at room temperature. PMSF was then added to a final concentration of 2 mM followed by 10 l of a 50% slurry of agarose beads containing immobilized monoclonal anti-phosphotyrosine antibody (4G10, Upstate Cell Signaling Solutions, catalog number 16-101). The sample was rotated at room temperature for 3 h, and the beads were then spun out and washed three times with PBS and three times with H 2 O. All supernatant was removed, and the beads were eluted with 0.1% trifluoroacetic acid in acetonitrile, for analysis by MALDI-TOF mass spectrometry. Expected masses were calculated using the program ProteinProspector. 2 Cell Lysate Preparation and EGFR Binding Assay-The human mammary epithelial cell line MCF12A was obtained from the University of Colorado Cancer Center Tissue Culture Core Facility and cultured in a 50:50 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 supplemented with 5% horse serum, 1 mM L-glutamine, 10 g/ml insulin, 0.5 g/ml hydrocortisone, 0.1 g/ml cholera toxin, 20 ng/ml EGF, 100 units/ml penicillin, and 100 g/ml streptomycin. Horse serum was from Sigma, and charcoal-stripped fetal calf serum was obtained from HyClone (Logan, UT). Murine epidermal growth factor was obtained from Collaborative Biomedical Products (Bedford, MA). All other media components were from Invitrogen. Quiescent cells were obtained by overnight culture (16 h) in media containing 2% charcoalstripped serum instead of 5% horse serum and lacking other growth factors. EGF stimulation was done by adding 100 ng/ml EGF to quiescent cells, followed by incubation at 37°C for 5 min prior to lysis. Cells were harvested by rinsing in ice-cold PBS, and then lysed by the addition of 0.8 ml per 10-cm dish of ice-cold EB (50 mM NaCl, 50 mM NaF, 10 mM Tris, pH 7.4, 5 mM EDTA, 1% Triton X-100, 1 mM sodium orthovanadate) containing the following protease inhibitors: 100 units/ml aprotinin, 1 mM PMSF, and 1 g/ml each of leupeptin, pepstatin, bestatin, and E64. Cells were scraped from the dish, and the lysates were clarified by spinning in a microcentrifuge at 13,000 ϫ g for 30 min at 4°C. Clarified lysates were stored at Ϫ80°C.
For EGFR binding, 10 g of purified GST fusion proteins were added to 0.5 ml of MCF12A lysate containing 0.5 mg of protein in EB plus protease inhibitors. Reactions were incubated at 4°C for 1 h, and then 40 l of a 50% slurry of glutathione-Sepharose beads (Amersham Biosciences) was added, and the tubes were rotated at 4°C for an additional hour. The tubes were then spun briefly to pellet the beads, the beads were washed four times with EB buffer, and the bound proteins were resolved by SDS-polyacrylamide gel electrophoresis. The resolved proteins were electrotransferred to polyvinylidene difluoride membranes (Immobilon-P, Millipore, Bedford, MA) for Western blotting using EGFR antibodies from Santa Cruz Biotechnology (Santa Cruz, CA, catalogue no. SC-03). Detection of proteins by immunoblotting was conducted using the Enhanced Chemiluminescence system (ECL) according to the manufacturer's recommendations (Amersham Biosciences).

RESULTS
The TKB Domain and/or Linker Helix of c-Cbl Modulate the E3 Activity of the RING Domain in a Phosphorylation-dependent Manner-The carboxyl-terminal half of c-Cbl and Cbl-b is dispensable for EGF-induced ubiquitination of EGFR and PDGFR; fragments containing approximately the first 480 amino acids are sufficient, as is Cbl-3, which is only 474 amino acids in length (10,18,19). To investigate the regulation of the E3 activity of c-Cbl, we have therefore focused our investigations on this region of the molecule. We first set out to establish an in vitro ubiquitination system, using only purified components expressed in E. coli. The system contains 32 P-labeled ubiquitin, E1, UbcH5b as E2, GST-c-Cbl fusion proteins as E3s, and an ATP-regenerating system. We first examined the behavior of a GST fusion construct containing amino acids 358 -447 of c-Cbl, essentially the RING finger domain and a few flanking amino acids (GST-Cbl-RING). A similar construct has been reported to demonstrate ubiquitin ligase activity in vitro (18). Similar to previous reports, we find that GST-Cbl-RING is readily auto-ubiquitinated ( Fig. 1, lanes 7-12). Auto-ubiquitination of E3s is commonly observed and provides a convenient assay for ubiquitin ligase activity in vitro. No E3 activity was observed with GST alone (Fig. 1, lanes [1][2][3][4][5][6] or in the absence of ATP, E1, or E2 (data not shown). In contrast to GST-Cbl-RING, a GST fusion protein containing the entire amino-terminal half of c-Cbl (GST-Cbl 1-480) showed much less activity in this assay, suggesting that the intrinsic E3 activity of the RING domain may be inhibited by interactions from the TKB domain and/or linker helix region present in this larger construct (Fig.  1, compare lanes 13-18 with 7-12; note also that ubiquitinated GST-Cbl-RING species exhibit increased mobility due to the smaller size of this construct).
To test whether this putative inhibition might be modulated by tyrosine phosphorylation, we next prepared a tyrosine-phosphorylated form of GST-Cbl 1-480, (P-GST-Cbl 1-480) using the bacterial strain TKX1, which contains an inducible elk1 tyrosine kinase. The protein purified from this strain was confirmed to contain phosphotyrosine using the phosphotyrosine-specific monoclonal antibody 4G10 (data not shown). It should be noted that we have no data at present indicating whether or not the pattern of tyrosine phosphorylation in our bacterially produced protein accurately reflects the tyrosine phosphorylation occurring in mammalian cells. However, we do find that bacterially produced P-GST-Cbl 1-480 has dramatically enhanced ubiquitin ligase activity over the unphosphorylated protein ( Fig. 1, compare lanes 19 -24 with 13-18; note also that E3 activity is evident even in the zero time point for P-GST-Cbl 1-480, in which the sample remained on ice rather than incubating at 30°C). To rule out any possible effects of the GST portion of these fusion proteins, we prepared Cbl 1-480 and P-Cbl 1-480 free from GST by treating the fusion proteins with thrombin. Cbl 1-480 and P-Cbl 1-480 prepared in this way retained the ubiquitin ligase activities of the parent GST fusion proteins (data not shown). Taken together, these data argue that the TKB domain and/or linker helix region of c-Cbl regulate the intrinsic E3 activity of the RING domain in a phosphorylation-dependent manner.
Analyzing the Phosphorylation-induced E3 Activity of c-Cbl Using Tyrosine to Phenylalanine Point Mutations: Tyr-268 Is Not Required for Activation-Previous investigators have examined the question of which tyrosine residues in c-Cbl are critical for activation of its ubiquitin ligase activity, however the question remains controversial. Analysis is complicated by In vitro ubiquitination reactions employing 32 P-labeled ubiquitin were performed as described under "Experimental Procedures" using purified proteins produced in E. coli. All reactions contained E1, UbcH5b as E2, an ATP-regenerating system, and various GST-Cbl fusion proteins as E3 as indicated at the top of the figure. Lanes 1-6 labeled as "GST," GST alone; lanes 7-12 labeled as "GST-Cbl-RING," a GST fusion with amino acids 358 -447 of c-Cbl (the RING domain and a few flanking amino acids); lanes 13-18 labeled as "GST-Cbl 1-480," a GST fusion of the entire amino-terminal half of c-Cbl (including TKB domain, linker helix, and RING domains); and lanes 19 -24 indicated "P-GST-Cbl 1-480," the same construct as lanes 13-18 but produced under conditions that phosphorylated tyrosine residues. Reactions were assembled on ice then transferred to 30°C and incubated for various times (0 -40 min) as shown at the top of the figure. Reactions were terminated by the addition of SDS sample buffer followed by boiling. Samples were analyzed by SDS-PAGE on 8% gels, followed by autoradiography. Lane numbers are indicated at the bottom of the figure, and the positions of molecular weight markers are indicated on the side. the large number of residues to be considered; c-Cbl contains 22 tyrosine residues. However, only six of these are completely conserved between the five known Cbl family members, corresponding to residues 268, 274, 291, 307, 337, and 371 of c-Cbl. Two additional tyrosine residues are conserved in four of the five Cbl family members (corresponding to residues 92 and 368 of c-Cbl). Levkowitz et al. (19) approached the question by generating a series of tyrosine to phenylalanine point mutants in c-Cbl; they found that Tyr to Phe point mutations at sites 92, 274, 291, 307, 337, and 368 retained activity in their assay, whereas a Y371F mutant was inactive. However, as discussed earlier, the interpretation of this result has been controversial, with Zheng et al. (23) arguing from their structural studies that a Y371F mutation would disrupt the structure of c-Cbl, rather than merely removing a regulatory phosphorylation site (23). If one accepts this argument for the moment, what other conclusions can be drawn from the data of Levkowitz et al.? Because none of the other Tyr to Phe point mutations tested abolished E3 activity, two possibilities present themselves. The first is that the key tyrosine residue required for phosphorylationinduced E3 activation was not among those tested. The second possibility is that more than one tyrosine residue in c-Cbl can activate E3 activity through phosphorylation, i.e. no single tyrosine residue is absolutely required (except for the structural role played by Tyr-371). Finally, a third possibility is that the conclusions of Levkowitz et al. about phosphorylation of Tyr-371 are correct, and the arguments of Zheng et al. are incorrect. We will examine each of these possibilities in turn.
The mutational study by Levkowitz et al. was extensive, but it did not include Tyr-268; this tyrosine residue is the only completely conserved tyrosine not examined in their study. To investigate whether this conserved tyrosine residue might be critical for E3 activation, we constructed a Y268F point mutant of GST-Cbl 1-480 and analyzed its behavior in vitro in both unphosphorylated and tyrosine-phosphorylated forms. As shown in Fig. 2, unphosphorylated Y268F behaves similarly to wild type GST-Cbl 1-480, with a very low level of auto-ubiquitination seen in this assay ( Fig. 2A, compare lanes 1 and 4). The tyrosine-phosphorylated version of Y268F also behaves similarly to phosphorylated wild type, i.e. it still appears to be fully activated by phosphorylation ( Fig. 2A, compare lanes 2 and 5). Phosphorimaging quantitation of this data is shown in Fig. 2B and confirms that the Y268F mutation does not significantly abrogate the E3 activity of c-Cbl as measured by this in vitro assay. Thus we do not believe that Tyr-268 is essential for phosphorylation-induced E3 activation of c-Cbl.
Tyrosine to Glutamate Point Mutations Constitutively Activate the E3 Activity of c-Cbl-The data presented above, in combination with the data of Levkowitz et al., indicate that none of the conserved tyrosine residues of c-Cbl, other than Tyr-371, is absolutely essential for E3 activity. If we accept the argument of Zheng et al. that the Y371F mutant lacks E3 activity due to structural disruption rather than lack of a phosphorylation site, we must consider the possibility that more than one tyrosine residue in c-Cbl can activate E3 activity through phosphorylation; i.e. no single tyrosine residue is absolutely required (except for the proposed structural role played by Tyr-371). Of course it is also possible that one of the non-conserved tyrosines is the key phosphorylation site, but we consider this possibility less likely. If more than one tyrosine is involved, the number of possible permutations grows large and makes analysis by multiple Tyr to Phe mutation unattractive. Attempting to construct inactive mutants through Tyr to Phe mutation also suffers from the earlier problem that more than one interpretation is possible. We therefore chose to try a different approach; we constructed a series of tyrosine to glutamate point mutants in c-Cbl, in an attempt to partially mimic the effect of tyrosine phosphorylation by introducing fixed negative charge. Although glutamate is not expected to closely mimic phosphotyrosine, the effect we are looking for is a gain of function, rather than a loss of function, and thus likely to be informative, if observed. In Fig. 3, we analyze the E3 activity of a series of Tyr to Glu point mutations in each of the six conserved tyrosines of c-Cbl 1-480. Equal amounts of purified GST fusion proteins were used in the in vitro ubiquitination assay and compared with wild type and tyrosine-phosphorylated wild type protein. It can be seen that several of the Tyr to Glu mutants did indeed show increased constitutive E3 activity in this assay. In particular, Tyr to Glu mutations at positions 307, 337, and 371 show strongly increased activity over unphosphorylated wild type (Fig. 3, compare lanes 8 -10 with lane  3), and Y274E also appears to have slightly increased activity Y371E Mutation in c-Cbl Enhances Binding to Activated EGFR-Two possibilities could explain the activation of ubiquitin ligase activity observed in the Y274E, Y307E, Y337E, and Y371E mutants. The first is that, as hoped, tyrosine to glutamate mutations are partially mimicking the effects of tyrosine phosphorylation at these sites, and therefore some or all of these sites might represent physiological phosphorylation sites involved in the activation of the E3 activity of c-Cbl. However, an alternative explanation might be that these mutations simply disrupt the structure of the TKB and/or linker helix domains, and this disruption removes the negative regulation exerted by these domains on the RING E3 activity. Indeed, the crystal structures of c-Cbl fragments obtained by Meng (13,23). In addition, as described earlier, Zheng et al. (23) have argued that Tyr-371 makes important internal structural contacts. These are precisely the tyrosine residues at which glutamate substitution increases E3 activity. Therefore, to assess the integrity of the TKB domain, we assayed the ability of these constructs to bind to the EGF receptor.
In Fig. 4, equal amounts of the indicated GST fusion proteins were added to cell extracts prepared from MCF12A cells, a human mammary epithelial cell line containing abundant EGF receptors. Cell extracts were prepared from both quiescent (Fig. 4B) and EGF-stimulated (Fig. 4A) cells. Control experiments indicated that EGFR in the EGF-stimulated cell extracts was phosphorylated on tyrosine residues as expected and that EGFR in extracts of quiescent cells lacked detectable phosphotyrosine (not shown). GST fusion proteins, and any cellular proteins bound to them, were then recovered by binding to glutathione-agarose. After washing, bound proteins were analyzed by SDS-PAGE followed by Western blotting with anti-EGFR antibody. In Fig. 4, it can be seen that GST alone does not bind EGFR, nor does GST-Cbl-RING, which lacks the TKB domain (Fig. 4, A and B, lanes 2 and 3). However, as expected, wild type GST-Cbl 1-480 does bind EGFR, but only the activated, tyrosine-phosphorylated form from EGF-stimulated cells (compare lanes 4 in Fig. 4, A and B). Furthermore, our bacterially prepared tyrosine-phosphorylated P-GST-Cbl 1-480 also binds EGFR in an activation-specific manner (Fig.  4, A and B, lanes 5). Of the six tyrosine to glutamate point mutants in c-Cbl, only Y371E binds to activated EGFR (Fig.  4A, compare lane 11 with lanes 6 -10). Consistent with the structural studies mentioned above, activated EGFR no longer binds to the Tyr to Glu mutants at positions 274, 307, and 337, which lie in the TKB binding pocket (Fig. 4A, lanes 7, 9, and  10). It is notable that another Cbl mutant known to disrupt EGFR binding, G306E, is adjacent to one of these sites. We also do not see significant EGFR binding to Y268E or Y291E (Fig.  4A, lanes 6 and 8).
The behavior of the Y371E mutant in these studies is notable. Not only does Y371E retain the ability to bind to the activated EGF receptor, it actually appears enhanced over wild type (Fig. 4A, compare lane 11 with lane 4). In this regard it is worth noting that we also find reproducibly greater binding of EGFR to phosphorylated wild type GST-Cbl 1-480 than unphosphorylated GST-Cbl 1-480 (Fig. 4A, compare lanes 4 and  5), although this enhanced binding is not as pronounced as that seen in the Y371E mutant. These data indicate that the TKB  (23). They suggested that phosphorylation of Tyr-371 would have to involve a significant structural rearrangement of the linker-TKB and linker-E2 interfaces. We therefore set out to look for evidence of such structural rearrangement.
If tyrosine phosphorylation of c-Cbl significantly alters its conformation, this may be reflected in altered susceptibility to proteases. To examine this possibility directly we performed protease digestion studies on purified Cbl 1-480 and P-Cbl 1-480 free from GST. Our purified Cbl 1-480 and P-Cbl 1-480 preparations show variable amounts of slightly smaller species, which likely represents slight proteolytic cleavage of the carboxyl terminus occurring during purification from the bacterial lysates (Fig. 5, A and B, lanes 1 and 2). We prepared equal concentrations of phosphorylated and unphosphorylated c-Cbl 1-480 and subjected these proteins to protease digestions in parallel, at a 20:1 ratio of substrate protein to protease. At various time points, aliquots of the reactions were removed and analyzed by SDS-PAGE and Coomassie staining. Fig. 5A shows the results of digestion with proteinase K, a protease with broad cleavage specificity. Panel B shows the results of a digest with trypsin, which has a more restricted cleavage specificity. In both cases, the tyrosine-phosphorylated form of Cbl 1-480 is attacked much more readily than the unphosphorylated protein. Furthermore, in the trypsin digests, a different pattern of proteolytic intermediates is observed (Fig. 5B, lanes 3-10, compare even and odd pairs of lanes). These data indicate that tyrosine phosphorylation of c-Cbl causes the protein to adopt a different conformation, possibly a more open conformation, more accessible to added proteases.
The Y371E Point Mutant of c-Cbl Has an Altered Conformation-The protease susceptibility studies above provide direct evidence that tyrosine phosphorylation of c-Cbl alters its conformation. We have also presented evidence above that mutation of tyrosine 371 to glutamate may partially mimic the effects of tyrosine phosphorylation. We therefore asked whether the Y371E mutant of c-Cbl has an altered conformation from wild type, as evidenced by protease susceptibility. Fig. 6 shows that indeed the Y371E mutant of c-Cbl is digested differently by proteases than unphosphorylated wild type. In this experiment equal concentrations of wild type and Y371E  Fig. 5. Both proteins were unphosphorylated. Samples in A were digested with proteinase K; in B, trypsin was employed. In C, unphosphorylated Y371E mutant c-Cbl 1-480 (evennumbered lanes) is compared with tyrosine-phosphorylated wild type c-Cbl 1-480 (odd-numbered lanes) in a trypsin digest. mutant c-Cbl 1-480 protein, cleaved from GST, were incubated with either proteinase K (Fig. 6A) or trypsin (Fig. 6B), as in the previous experiment, and aliquots were removed over time. As noted in the previous experiment, some heterogeneity is evident in the starting material; nonetheless, it is evident that the two proteins differ in both the rate of digestion and the pattern of proteolytic intermediates formed. Similar to the case of phosphorylated c-Cbl, Y371E is initially degraded more rapidly than unphosphorylated wild type c-Cbl. Interestingly, at later time points, a fragment of the Y371E mutant appears to have increased stability over wild type (Fig. 6, for example compare lanes 15 and 16 in panel A, or lanes 13 and 14 in panel B). In the tryptic digestion of Y371E, a discrete band is prominent with a mobility just above the 37-kDa marker. A similar discrete tryptic fragment was prominent in the digestion of tyrosine phosphorylated wild type c-Cbl (see Fig. 5, lanes 4, 6, 8,  and 10). The molecular weight markers on that gel might suggest a fragment of different size, however the markers on the gels in Fig. 5 were commercial pre-stained markers that exhibited quite broad bands on the gel, and thus they provide only a rough guide to the true electrophoretic mobilities. The markers used in Fig. 6 were much sharper and more accurate. For example, c-Cbl 1-480 has a predicted molecular mass of ϳ54 kDa, more in agreement with the electrophoretic markers used in Fig. 6. To more accurately determine if the tryptic digestion of phosphorylated c-Cbl 1-480 and Y371E mutant c-Cbl shared any similarities in rate or pattern of digestion, the two proteins were digested at equal concentrations under identical conditions and run side by side, as is shown in Fig. 6C. Phosphorylated wild type protein is present in the odd-numbered lanes, and mutant unphosphorylated protein is shown in the even-numbered lanes. It can be seen that a discrete tryptic fragment of ϳ38 kDa, relatively resistant to attack, is prominent in both digests. The digestion pattern of the Y371E mutant more closely resembles that of tyrosine-phosphorylated c-Cbl rather than the unphosphorylated form. This is evidence that the Y371E mutant has an altered conformation from wild type and provides further evidence that the Y371E mutant may be a model for tyrosine-phosphorylated c-Cbl.
Examination of primary sequence and tertiary crystal structures of c-Cbl for tryptic cleavage sites suggests a possible identity for the tryptic fragment of ϳ38 kDa resistant to proteolytic cleavage in the Y371E mutant and tyrosine-phosphorylated wild type. The structures forming the TKB domain (four helix bundle, EF-hand domain, and divergent SH2 domain) pack tightly together and are likely to form a relatively resistant core. Because glutamic acid substitution at position 371 stabilizes the tryptic fragment, this suggests that the linker helix is also part of this resistant fragment. A tryptic peptide containing residues 54 -382 of c-Cbl appears most likely. Such a fragment has a predicted size of 38,329 and would include the entire tightly folded TKB domain as well as the linker helix, while removing less tightly structured residues at the amino terminus, and essentially the entire RING finger domain.
It may be noted that the tryptic digest pattern of tyrosinephosphorylated c-Cbl 1-480 in Fig. 6C differs somewhat from that in Fig. 5B. This may be due to differences in glycerol concentrations between the two digests (10% in Fig. 5B versus 20% in Fig. 6C), with the higher glycerol concentration lending greater stability to the protein.
A Y363E Mutation in Cbl-b Behaves Similarly to c-Cbl Y371E-Based on all the evidence discussed above, we believe that the TKB and/or linker helix domains of c-Cbl inhibit the ability of the RING finger domain to function as a ubiquitin ligase and that phosphorylation at tyrosine 371 causes a conformational change in the protein that removes this negative regulation, resulting in enhanced ubiquitin ligase activity. We would also argue that our Y371E mutant partially mimics the effect of tyrosine phosphorylation at this site. If this model is correct, we would predict that tyrosine to glutamate point mutation would have similar effects in other Cbl family members. To test this prediction we prepared GST fusion proteins containing the amino-terminal half of Cbl-b (GST-Cbl-b 1-480), in both wild type and mutant forms (Cbl-b Y363E; tyrosine 363 in Cbl-b is the homologue of Tyr-371 in c-Cbl). In Fig. 7A, we examine the behavior of wild type and Y363E GST-Cbl-b fusion proteins in our in vitro ubiquitination assay. We find that wild type GST-Cbl-b shows a low but evident level of auto-ubiquitination in this assay, and that the Y363E mutant has much greater activity (Fig. 7A, compares lanes 5 and  6). This behavior is very similar to that of c-Cbl (Fig. 7A,   FIG. 7. A Y363E 2 and 3). We also tested the ability of these GST-Cbl-b proteins to bind to activated EGFR. In Fig. 7B, equal amounts of purified GST fusion proteins were incubated with cell lysates prepared from EGF-stimulated MCF12A cells. Following incubation, GST proteins, and any cellular proteins bound to them, were recovered by binding to glutathione-agarose. The presence of bound EGFR in these complexes was then assayed by Western blotting. It can be seen that both wild type and the Y363E mutant of Cbl-b bind activated EGFR in this assay, whereas GST alone does not (Fig. 7B, lanes 2-4). Similar to the findings in c-Cbl, Tyr to Glu mutation at tyrosine 363 in Cbl-b does not disrupt the function of the TKB domain and in fact may even enhance EGFR binding (Fig. 7B, compare lanes  3 and 4).
Tyrosines 368 and 371 of c-Cbl Can Both Be Phosphorylated-We propose that the constitutively active Y371E c-Cbl mutant partially mimics the effect of tyrosine phosphorylation at this site. To confirm this model, we next sought to identify the sites of tyrosine phosphorylation in our bacterially produced c-Cbl 1-480 protein. We electrophoresed equal amount of Cbl 1-480 and P-Cbl 1-480 on SDS-polyacrylamide gels and performed in-gel tryptic digestions followed by MALDI-TOF of the resulting peptides. Peptides containing 9 of the 14 tyrosine residues present in c-Cbl 1-480 were identified in this analysis. Peptides containing tyrosines 83, 92, 102, 114, 141, 268, 274, 291, and 337 were recovered from both samples in approximately equal amounts, however, phosphorylated forms of these peptides were not observed. We therefore believe that these tyrosine residues are not significantly phosphorylated in our bacterially produced protein. However we were unable to recover the tryptic peptide containing tyrosine 371 (which also contains tyrosine 368) under these conditions from either sample. In addition, we did not recover peptides containing tyrosines 235, 307, or 455. Peptides containing these residues may be too hydrophobic to elute from the gel under the conditions employed. We therefore pursued a different strategy to attempt to recover tyrosine-phosphorylated peptides. P-Cbl 1-480 was reduced and denatured by treatment with DTT and SDS at 95°C. The protein was then alkylated and diluted into nonionic detergent for trypsin digestion. Phosphotyrosine-containing peptides were recovered by binding to the anti-phosphotyrosine monoclonal antibody 4G10 immobilized on agarose beads. MALDI-TOF analysis of eluate from the beads revealed two peptides with masses expected for phosphorylated tryptic fragments of c-Cbl. Both are predicted to derive from the tryptic peptide VTQEQYELYCEMGSTFQLCK from residues 361-382 of c-Cbl and correspond to peptides containing either one or two phosphate groups (Fig. 8, peaks A and C). Although the tyrosine residue in the singly modified peptide cannot be exactly identified from this analysis, recovery of the doubly modified peptide identifies both tyrosines 368 and 371 as sites of phosphorylation in our bacterially produced protein. No other phosphopeptides were recovered in this analysis. The third peptide peak shown in Fig. 8 (peak B) has the mass expected for the non-phosphorylated peptide DAFQPHHHHHHHLSPHP-PGTVDK, residues 31-53 of c-Cbl, apparently adsorbed nonspecifically to the agarose beads. This combination of approaches has provided direct data on the phosphorylation status of 11 of the 14 tyrosine residues present in P-Cbl 1-480; phosphorylation is detected only at Tyr-368 and Tyr-371, although peptides containing 3 tyrosines (235, 307, and 455) were not recovered.
Phosphorylation of Tyr-371 Mutants Enhances E3 Activity-Identification of phosphorylation at tyrosine 368 in addition to tyrosine 371 led us to ask if phosphorylation at this site has functional significance. Levkowitz et al. found that a Y368F mutant in c-Cbl was still able to ubiquitinate EGFR, indicating that phosphorylation at this site is not absolutely required for the E3 activity of Cbl. Might it contribute to activation, however? If phosphorylation of Tyr-371 alone mediates Cbl activation, we would predict that tyrosine phosphorylation of our Y371F and Y371E mutants would have no effect upon their E3 activity. However this is not what we observed; as shown in Fig. 9, tyrosine phosphorylation of both the inactive Y371F mutant and the constitutively active Y371E mutant increased their E3 activities over the unphosphorylated forms (Fig. 9,  compare lanes 4 and 5, and lanes 6 and 7). When these data were quantitated by phosphorimaging and normalized to an activity of 100% for phosphorylated wild type Cbl, the Y371F mutant is found to increase from 9.5% to 39.5% upon phosphorylation, and the Y371E mutant increased from 32.9% to 79.8%. These data indicate that tyrosine 371 is not the sole site of phosphorylation responsible for activation of the ubiquitin ligase activity of Cbl.
Phosphorylation of Tyr-368, Tyr-371 Double Mutants Does Not Alter Their E3 Activity-The data of Fig. 9 indicate that tyrosine phosphorylation at sites other than Tyr-371 (possibly FIG. 8. Both Tyr-368 and Tyr-371 of c-Cbl are phosphorylation sites. P-Cbl 1-480 was digested with trypsin and then immunoprecipitated with anti-phosphotyrosine antibodies immobilized on agarose beads as described under "Experimental Procedures." Bound peptides were eluted and analyzed by MALDI-TOF mass spectrometry. Shown is a portion of the mass spectrum with three peaks labeled A, B, and C, and their measured masses are shown above. The mass measured for peak A is in good agreement with that expected for the tryptic peptide VTQEQYELYCEMGSTFQLCK, residues 363-382 of c-Cbl, containing one phosphate group (predicted mass 2594.07). The mass measured for peak C is in good agreement with that expected for the same tryptic peptide containing two phosphate groups (predicted mass 2674.03). Peak B corresponds to the non-phosphorylated peptide DAFQPHHHH-HHHLSPHPPGTVDK (residues 31-53 of c-Cbl with expected mass of 2665.27), apparently adsorbed non-specifically to the agarose beads.
Tyr-368, for example) can contribute to E3 activation in c-Cbl. Our mass spectroscopy data of Fig. 8 positively identified both Tyr-368 and Tyr-371 as phosphorylation sites. If these two residues are the only significant sites, we would predict that Tyr-368, Tyr-371 double mutants would no longer show an increase in E3 activity upon phosphorylation. We tested this prediction by constructing FF and EE double mutants at Tyr-368 and Tyr-371 and preparing GST-Cbl fusion proteins in both unphosphorylated and tyrosine-phosphorylated forms. We then measured their activities in the in vitro ubiquitination assay. We found that the Y368F/Y371F double mutant had a low basal level of E3 activity, similar to unphosphorylated wild type, and this basal level was not increased upon tyrosine phosphorylation (data not shown). The Y368E/Y371E double mutant in contrast, was constitutively activated, and in fact appeared to have somewhat greater activity than the Y371E single mutant (Fig. 10A, compare lanes 1 and 2). However, the ubiquitin ligase activity of the EE double mutant did not further increase upon phosphorylation (Fig. 10A, compare lanes 2  and 3). Control experiments confirmed that the phosphorylated forms of the FF and EE double mutants did in fact contain phosphotyrosine, as evidenced by reactivity with the 4G10 monoclonal anti-phosphotyrosine antibody, although the extent of phosphorylation was significantly reduced (appearing less than 5% of the level of the phosphorylated Y371E single mutant, data not shown). In Fig. 10C, we have assessed whether glutamate mutation at Tyr-368 affects ability to bind to activated EGFR. It can be seen that Y368E mutation, singly or in combination with Y371E, does not disrupt the ability of c-Cbl to bind activated EGFR (Fig. 10C, compare lanes 5 and 7  with lanes 3 and 6), indicating that alteration of this residue does not disrupt the TKB domain. The Y274E point mutant, which does disrupt the TKB domain, is included here as a negative control (Fig. 10C, lane 4). Because Tyr-368, Tyr-371 double mutants no longer showed phosphorylation-induced activation, we conclude that these two residues are the critical phosphorylation targets for ubiquitin ligase activation. DISCUSSION In this study we have used an in vitro ubiquitination system to explore the phosphorylation-induced activation of the ubiquitin ligase (E3) activity of c-Cbl. Activation of the E3 activity of c-Cbl by tyrosine phosphorylation has been reported by others in several systems (19,25). We show here that the intrinsic E3 activity of the isolated RING finger domain of c-Cbl is negatively regulated by the amino-terminal TKB and/or linker helix domains when these are included in a larger construct. We also show that the activation of the ubiquitin ligase activity of c-Cbl by tyrosine phosphorylation appears to result from removal of the inhibitory effects of the amino-terminal domains. We provide direct evidence that tyrosine-phosphoryl-  7) were added to lysates of EGF-stimulated MCF12A cells, as described previously. After 1 h at 4°C, glutathione-Sepharose was added, and bound proteins were collected, washed to remove nonspecifically bound proteins, and analyzed by SDS-PAGE followed by immunoblotting with antibodies to EGFR. Shown is an EGFR immunoblot visualized with chemiluminescence. Lane numbers are shown at the bottom of the panel, and the positions of molecular weight markers are indicated on the side. Lane 1 labeled as "Lysate," 10 g of unfractionated cell lysate; all other lanes (2-7) contain washed protein complexes bound to glutathione-Sepharose. GST fusion proteins used to form the complexes are indicated at the top of the panel. ated c-Cbl adopts a different conformation than the unphosphorylated form, based on protease sensitivity studies. This phosphorylation-induced conformational change provides a mechanism to explain the regulation of E3 activity.
A debate has existed concerning the role of Tyr-371 in the activation of the E3 activity of c-Cbl, with mutational data and structural data seemingly at odds. Because loss of function mutants are always subject to multiple interpretations, we attempted here to generate activating mutations in c-Cbl, in which the negative regulation of the amino-terminal TKB and linker helix domains upon the RING was removed or reduced. Substitution of conserved tyrosine residues with glutamate was employed in the hope that introduction of fixed negative charge might, in some modest way, partially mimic the effects of tyrosine phosphorylation. At four of the six completely conserved tyrosine residues in Cbl, we found that mutation to glutamate resulted in point mutants showing increased autoubiquitination in our assay, indicating that altering the TKB and/or linker helix domain structure can indeed remove or reduce the negative regulatory effects exerted on the RING E3 domain. However, in vivo, Cbl proteins bind and ubiquitinate receptor tyrosine kinases, in addition to undergoing auto-ubiquitination. Thus the TKB domain is critical for biological function, and activating the E3 activity of Cbl in vivo must occur by structural changes that do not abolish TKB function. In three of the four mutants where constitutive activation was observed, the function of the TKB domain was compromised, resulting in loss of the ability to bind activated EGFR. However, mutation of tyrosine 371 to glutamate proved a notable exception. The Y371E mutant not only retains ability to bind activated EGFR, this binding appears enhanced over wild type and thus echoes the enhancement of EGFR binding observed with phosphorylation of wild type c-Cbl. We therefore suggest that the Y371E mutant may be a model for tyrosine-phosphorylated c-Cbl. We generalized this model by studies with Cbl-b and show that an homologous Cbl-b-Y363E mutant behaves similarly. To confirm our model we identified Tyr-371 as a site of phosphorylation in our bacterially produced protein and unexpectedly found that Tyr-368 was also phosphorylated. Tyr-368 was not included in our original set of mutants, because this site is not present in Cbl 3. Tyrosine 368 is conserved in all other Cbl family members, however, and we show that phosphorylation at this site also contributes to E3 activation. Tyrosine 368 can also be mutated to glutamate without loss of TKB domain function, and Y368E/Y371E double mutants appear to have increased constitutive E3 activity over the single point mutants.
Our data support the earlier conclusions by Levkowitz et al. that phosphorylation of Tyr-371 plays a key role in activating the E3 activity of c-Cbl. How can we reconcile these data with Zheng et al.'s structural studies? It must be noted that x-ray crystallographic studies of necessity provide a static picture of a stable conformation. The extremely powerful and detailed information provided by crystallography is unable to reveal dynamic movements of a protein in solution. We would argue then, that there must be sufficient flexibility in the conformation of c-Cbl for the internally buried Tyr-371 residue to somehow become available for phosphorylation. One possibility is that, in solution, the structure proposed by Zheng et al. might be in equilibrium with some other conformation of c-Cbl in which Tyr-371 is more accessible. Alternatively, binding interactions between c-Cbl and another protein, perhaps Src, might alter the conformation of c-Cbl sufficiently to provide access for phosphorylation. Zheng et al. state that, if Tyr-371 were to be phosphorylated, this would "result in significant structural change in the linker-TKB and linker-E2 interfaces." Our stud-ies do not provide the level of resolution necessary to confirm these predictions. Nonetheless, our data do provide direct evidence that tyrosine phosphorylation of c-Cbl alters its conformation. Furthermore, our Y371E mutant also has an altered conformation, possibly mimicking the effect of tyrosine phosphorylation.
The protease digestion studies performed here indicate that tyrosine phosphorylation of c-Cbl, or mutation of tyrosine 371 to glutamate, both increase the stability of a tryptic fragment of ϳ38 kDa. How might this be achieved? As discussed earlier, this tryptic fragment is likely to consist of residues 54 -382 of c-Cbl, containing the entire TKB and linker helix domains. Introduction of fixed negative charge at residue 371 through phosphorylation or mutation is expected to disrupt the interaction of residue 371 with threonine 227, and the negatively charged side chain would be unlikely to remain in the hydrophobic interior occupied by Tyr-371 in the crystal structure. For tyrosine phosphorylation or glutamate mutation at residue 371 to stabilize such a peptide, one mechanism would be the formation of a salt bridge between the negative charge at residue 371 and a nearby positively charged residue. Examination of the published crystal structures does not reveal any nearby positive charges in the hydrophobic interior region occupied by the Tyr-371 side chain. However, rotation of the linker helix could allow the side chain of residue 371 to reach the surface where it might form a salt bridge with arginine 343 or perhaps with another of the several positively charged residues located in this surface vicinity. More detailed structural studies of our Y371E mutant or the Y368E/Y371E double mutant might be of some interest as a model for structural changes occurring upon phosphorylation.
Our data demonstrating conformational changes induced in c-Cbl by tyrosine phosphorylation are in agreement with suggestive earlier findings by Yokouchi et al. (25). These authors performed in vitro binding studies with purified GST-UbcH7, purified c-Cbl, ATP, and SF9 cell extracts containing or lacking Src. In the absence of Src, c-Cbl was observed to co-precipitate with GST-UbcH7. When cell extracts containing Src were employed in this system, c-Cbl no longer co-precipitated with GST-UbcH7. The authors had previously shown that under similar conditions, c-Cbl was tyrosine-phosphorylated by Src. (Furthermore, our laboratory has also published data indicating that Src family kinases phosphorylate c-Cbl (26).) Omission of ATP from the reactions, or pretreatment of the SF9 cells with the Src inhibitor PP1, restored c-Cbl binding to GST-UbcH7 and led to the presence of variable amounts of Src co-precipitating in the complex. These data lend support to the idea that tyrosine phosphorylation of c-Cbl alters its conformation, as evidenced by its ability to stably bind UbcH7.
Tyrosines 368 and 371 of c-Cbl have also received much attention in another context. Single amino acid deletion of either tyrosine confers transforming ability to c-Cbl (27). Four other transforming mutants of c-Cbl are known: point mutation of tyrosine 371 to alanine, Y371A (28), and the deletion mutants v-Cbl (a truncation after residue 355) (2), 70Z-Cbl (an internal deletion of residues 366 -382) (2), and p95Cbl (an internal deletion of residues 366 -477) (29). All of the currently known transforming mutants of c-Cbl lack the ability to ubiquitinate activated EGFR while retaining the ability to bind to it. It has therefore been proposed that they act as dominant negative mutants. However, several other mutants that bind to activated EGFR and are unable to ubiquitinate it do not transform cells, including Y371F. Thus it has not been clear what specific features are required for transforming ability in c-Cbl; inability to conjugate ubiquitin is insufficient. Thien et al. (28) have studied this issue in some detail and note that all of the known transforming mutants in c-Cbl disrupt the ␣ helical linker domain between the TKB domain and the RING finger. This linker helix also makes contacts with bound UbcH7. Thien et al. (28) suggest that disruption of specific interactions between TKB, RING, and bound UbcH7 may be critical for transforming ability rather than simply the loss of E3 activity. We have shown here that phosphorylation of Tyr-368 and Tyr-371, both located in this linker helix region, is critical for altering the conformation of c-Cbl and activating ubiquitin ligase activity. Clearly, further work will be required to fully understand both the normal cellular functions of c-Cbl and how disruptions of c-Cbl function can transform cells.