Selective cleavage of the heregulin receptor ErbB-4 by protein kinase C activation.

The 180-kDa transmembrane tyrosine kinase ErbB-4 is a receptor for the growth factor heregulin. 125I-Heregulin binding to NIH 3T3 cells overexpressing the ErbB-4 receptor is rapidly decreased by 12-O-tetradecanoylphorbol-13-acetate (TPA) pretreatment. Immunologic analysis demonstrates that TPA treatment of cells induces the proteolytic cleavage of ErbB-4, producing an 80-kDa cytoplasmic domain fragment, which contains a low level of phosphotyrosine, and a 120-kDa ectodomain fragment, which is released into the extracellular medium. Cleavage of ErbB-4 was also enhanced by other protein kinase C activators, i.e. platelet-derived growth factor, ionomycin, and synthetic diacylglycerol, while protein kinase C inhibition or down-regulation suppressed the TPA stimulation of ErbB-4 degradation. TPA did not induce the degradation of related receptors (ErbB-1, ErbB-2, and ErbB-3) in the EGF receptor family. The phorbol ester-induced cleavage of ErbB-4 occurs within or close to the ectodomain, as the 80-kDa cytoplasmic domain fragment is recognized by antibody to the ErbB-4 carboxyl terminus and is membrane-associated. Coprecipitation experiments show that, while the 80-kDa ErbB-4 fragment is associated with the SH2-containing molecules PLC-γ1 and Shc, TPA did not induce the phosphorylation of these substrates in intact cells. In addition, kinase assays in vitro indicate that the 80-kDa fragment is not an active tyrosine kinase. These results show that protein kinase C negatively regulates heregulin signaling through the ErbB-4 receptor by the activation of a selective proteolytic mechanism.

The 180-kDa transmembrane tyrosine kinase ErbB-4 is a receptor for the growth factor heregulin. 125 I-Heregulin binding to NIH 3T3 cells overexpressing the ErbB-4 receptor is rapidly decreased by 12-O-tetradecanoylphorbol-13-acetate (TPA) pretreatment. Immunologic analysis demonstrates that TPA treatment of cells induces the proteolytic cleavage of ErbB-4, producing an 80-kDa cytoplasmic domain fragment, which contains a low level of phosphotyrosine, and a 120-kDa ectodomain fragment, which is released into the extracellular medium. Cleavage of ErbB-4 was also enhanced by other protein kinase C activators, i.e. platelet-derived growth factor, ionomycin, and synthetic diacylglycerol, while protein kinase C inhibition or down-regulation suppressed the TPA stimulation of ErbB-4 degradation. TPA did not induce the degradation of related receptors (ErbB-1, ErbB-2, and ErbB-3) in the EGF receptor family.
The phorbol ester-induced cleavage of ErbB-4 occurs within or close to the ectodomain, as the 80-kDa cytoplasmic domain fragment is recognized by antibody to the ErbB-4 carboxyl terminus and is membrane-associated. Coprecipitation experiments show that, while the 80-kDa ErbB-4 fragment is associated with the SH2-containing molecules PLC-␥1 and Shc, TPA did not induce the phosphorylation of these substrates in intact cells. In addition, kinase assays in vitro indicate that the 80-kDa fragment is not an active tyrosine kinase. These results show that protein kinase C negatively regulates heregulin signaling through the ErbB-4 receptor by the activation of a selective proteolytic mechanism.
Four transmembrane tyrosine kinases constitute the human epidermal growth factor (EGF) 1 receptor family: the EGF receptor or ErbB-1, ErbB-2, ErbB-3, and ErbB-4 (1). The extracellular domains of all ErbB molecules have two cysteine-rich domains; the cytoplasmic tyrosine kinase domains are highly conserved in primary sequence, while the cytoplasmic carboxyl termini are dissimilar in sequence but not size. Understanding the metabolic regulation of EGF receptor family members and their signal transduction pathways is of particular interest, as these receptors are implicated in a variety of aggressive carcinomas (2)(3)(4)(5)(6).
The EGF family of growth factors (7) specifically bind to the EGF receptor, while the heregulin family of growth factors (8,9) associate with the ErbB-3 and ErbB-4 receptors. Heregulin binds to ErbB-4 and ErbB-3, respectively, with high and moderate affinity (10), whereas a heterodimer of ErbB-2 and ErbB-3 is reported to constitute a second high affinity receptor (11,12). The association of heregulin with ErbB-3 or ErbB-4 is able to induce phosporylation of ErbB-2 by a mechanism that involves transphosphorylation and perhaps receptor heterodimerization (13,14). The EGF receptor also heterodimerizes with ErbB-2 (15,16) and perhaps ErbB-3 (10,17,18). These observations may underlie the clinical importance of coexpression of ErbB family members in many breast carcinomas (19,20).
Ligand-induced desensitization and down-regulation mechanisms are important aspects of the regulation of transmembrane receptors (21), and within the ErbB receptor family significant similarities and differences have emerged in attenuation mechanisms. Ligand binding to the EGF receptor rapidly induces receptor-mediated endocytosis through clathrincoated pits (22). The internalized complexes are subsequently degraded in lysosomes and, consequently, cell surface EGF receptors are down-regulated. In contrast to the EGF receptor, all of the other ErbB family members, including ErbB-4, are not rapidly internalized in the presence of ligand and do not exhibit ligand-enhanced metabolic turnover and downregulation (23,24).
Protein kinase C activity significantly influences desensitization of receptors in the ErbB family. EGF receptor activity is attenuated by protein kinase C through phosphorylation of cytoplasmic domain serine/threonine residues which decreases ligand binding affinity and tyrosine kinase activity (25)(26)(27). Evidence for a similar phosphorylation-dependent attenuation of ErbB-2 tyrosine kinase activity by protein kinase C has been presented (28). In addition, heterologous attenuation of the EGF receptor by the platelet-derived growth factor (PDGF) receptor is thought to be mediated through the activation of protein kinase C (29 -31). There are no reports of protein kinase C regulation of the ErbB-3 or ErbB-4 receptors. The data presented below show that protein kinase C activators downregulate the ErbB-4 receptor by activating a selective proteolytic mechanism.  (19). Polyclonal IgG to the carboxyl terminus (residues 1291-1308) of ErbB-4 were also purchased from Santa Cruz Biotechnology, Inc. Polyclonal antibodies to Shc were purchased from Transduction Laboratories. Antisera to phospholipase C-␥1 was described previously (32). Protein A-Sepharose and enhanced chemiluminescence (ECL) reagents were from Sigma and Immobilon-P membranes were from MCI. 125 I-Labeled protein A was a product of ICN, and horseradish peroxidase conjugated protein A was from Zymed. 125 I-Labeled goat anti-mouse IgG was purchased by ICN.
Immunoprecipitation and Western Blotting-T47-14 cells (70% confluent) in 60-mm culture dishes were starved overnight in 0.5% calf serum. Subsequent incubations were performed in 1.0 ml of basal medium (DMEM, 0.1% BSA, and 20 mM Hepes, pH 7.4). After the indicated treatments, cells were washed three times with ice-cold Ca 2ϩand Mg 2ϩ -free phosphate-buffered saline (CMF-PBS) and lysed with 300 l of ice-cold TGH buffer (1% Triton X-100, 10% glycerol, 20 mM Hepes, pH 7.2, 100 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, and 1 mM Na 3 VO 4 ). After scraping, the lysates were incubated for 30 min at 4°C with rocking. Insoluble material was then removed by centrifugation (14,000 ϫ g, 10 min) at 4°C, and the supernatant was collected. Protein concentration was assayed with the method of Bradford (34) using bovine serum albumin as the standard. To precipitate ErbB-4, approximately 1 g of ErbB-4 antibody was incubated (2 h at 4°C) with 300 g of cellular protein followed by a 1-h incubation with protein A-Sepharose CL-4B. Immune complexes were washed three times with TGH lysis buffer, resuspended in 1 ϫ Laemmli buffer (35), and boiled for 5 min. Subsequently, samples were electrophoresed in a 7.5% SDS-polyacrylamide gel and transferred to nitrocellulose membranes for Western blotting. Membranes were blocked by a 1-h incubation at room temperature with TBST buffer (0.05% Tween 20, 150 mM NaCl, 50 mM Tris, pH 7.4) containing 3% BSA. Subsequently, ErbB-4 antiserum (dilution, 1:3000) or phosphotyrosine antibody (dilution, 1:1000) was added for 2 h at room temperature with shaking. Membranes were then washed three times with TBST buffer, incubated either with 125 I-protein A for 1 h or horseradish peroxidase-conjugated protein A, and after washing with TBST buffer visualized by autoradiography (Kodak X-Omat AR film) or ECL, respectively.
Cell Fractionation-Cells were washed three times with ice-cold CMF-PBS, scraped into 800 l of hypotonic buffer A (20 mM Hepes, pH 7.4, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM Na 3 VO 4 ), and disrupted with a Dounce homogenizer. NaCl was then added to a final concentration of 100 mM. Nuclei and unbroken cells were removed by two low speed centrifugations (800 ϫ g, 5 min). The supernatants were ultracentrifuged (100,000 ϫ g, 50 min), and Triton X-100 was added to 1% to the high speed supernatants. The high speed pellets were gently rinsed twice with ice-cold CMF-PBS, resuspended in hypotonic buffer A supplemented with 1% Triton X-100 and 100 mM NaCl, and solubilized for 30 min at 4°C with rocking. Subsequently, insolubilized material was removed from the Tritonsolubilized pellets by centrifugation (100,000 ϫ g, 50 min). Equal volumes of cytosol and membrane fractions, representing equivalent numbers of cells, were precipitated with antibody to ErbB-4 and subjected to Western blot analysis.
In Vitro Kinase Assay-Cell monolayers were washed twice with ice-cold CMF-PBS and lysed in 300 l of TGH buffer without Na 3 VO 4 . Duplicate samples (300 g of protein) were immunoprecipitated (2 h, 4°C) with polyclonal antibody to ErbB-4 before protein A-Sepharose was added to each tube and incubated (1 h, 4°C). The immunoprecipitates were then washed twice with the TGH buffer without Na 3 VO 4 and twice with kinase buffer (20 mM Hepes, pH 7.4, 3 mM MnCl 2 , 15 mM MgCl 2 , 200 mM NaCl, 10 g/ml aprotinin, and 100 M Na 3 VO 4 ). Immune complexes were resuspended in 50 l of kinase buffer containing ATP (5 M), and to half the samples [␥-32 P]ATP (20 Ci per sample) was added. All tubes were then incubated at 37°C for 5 min. The reaction was stopped by adding 500 l of a buffer containing 5 mM sodium phosphate, pH 7.0, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 1 mM Na 3 VO 4 , 10 g/ml aprotinin, 1 mM ATP, and 5 mM EDTA. The immune complexes were washed twice with the same buffer, and 1 ϫ Laemmli buffer was added. Protein in each sample were then separated on a 7.5% polyacrylamide gel and the [␥-32 P]ATP-labeled samples were analyzed by autoradiography. Nonradioactive samples were analyzed for ErbB-4 protein by Western blotting with ErbB-4 antiserum, and the results were quantitated using a PhosphorImager.

TPA Down-modulation of 125 I-Heregulin
Binding to ErbB-4 -To determine whether protein kinase C activation could modulate heregulin binding to ErbB-4, 125 I-heregulin binding assays were performed on T47-14 cells, which express approximately 1 ϫ 10 6 ErbB-4 heregulin receptors per cell (24) following TPA pretreatment. Cells were incubated with or without TPA at 37°C for 30 min or 1 h and then incubated at 4°C with a near saturating concentration of 125 I-heregulin to measure cell surface receptors. The results, shown in Fig. 1, demonstrate that TPA preincubations of 30 and 60 min decreased 125 I-heregulin binding by 82% and 91%, respectively. These data show that the ligand binding capacity of ErbB-4 receptors can be dramatically down-regulated by the protein kinase C activator TPA.
Influence of TPA on ErbB-4 Protein-There are several possible mechanisms by which TPA might modulate 125 I-heregulin binding to the ErbB-4 receptor. To determine whether TPA alters the cellular level or structure of ErbB-4 protein, cells were incubated with TPA in the presence or absence of heregulin for 30 min or 6 h, and detergent lysates were assayed for ErbB-4 protein by immunoprecipitation and Western blotting. As shown in Fig. 2A (lane 4), TPA rapidly stimulated proteolytic cleavage of the ErbB-4 receptor as indicated by the loss of native receptor (180 kDa) and the appearance of an 80-kDa fragment. Activation of this proteolytic event while evident at T47-14 cells in six-well dishes were washed twice with DMEM, and 1.0 ml of binding medium (DMEM plus 0.1% BSA) was added. TPA (100 ng/ml) was added to the indicated cells for 30 min or 1 h, and the cultures were incubated at 37°C. Cell monolayers were then rapidly washed with ice-cold DMEM and incubated for 90 min at 4°C in binding medium containing 125 I-heregulin (100 ng/ml). After incubation, the cells were washed twice with ice-cold DMEM and lysed in 1 N NaOH. Quantitation of the amount of 125 I-heregulin specifically bound was performed with a ␥ spectrometer (Beckman). Nonspecific 125 Iheregulin binding was determined in parallel cultures containing an excess of unlabeled heregulin. 30 min was not sustained over a longer (6 h) period of time. No decrease in ErbB-4 receptor was induced by heregulin ( Fig. 2A,  lanes 2 and 4) at either time point.
The 80-kDa fragment reacts with antibody to the ErbB-4 carboxyl terminus and its molecular mass corresponds to the entire ErbB-4 cytoplasmic domain. Anti-phosphotyrosine blotting of ErbB-4 precipitates (Fig. 2B, lane 11) shows that the 80-kDa fragment is tyrosine-phosphorylated. However, the extent of 80-kDa tyrosine phosphorylation was not further increased by simultaneous treatment with TPA and heregulin (Fig. 2B, lane 13). The tyrosine phosphorylation level of the 80-kDa fragment was equivalent to that of the full-length receptor in untreated cells (Fig. 2B, lane 8) and not enhanced by the presence of heregulin. Pretreatment of the cells with heregulin (100 ng/ml) for 30 min did not prevent the TPAmediated generation of the 80-kDa fragment (data not shown). The data in Fig. 3 demonstrate that the TPA-induced cleavage of ErbB-4 is both time-and concentration-dependent. At TPA concentrations of 50 ng/ml or more, ErbB-4 cleavage could be detected by 10 min and was maximal at approximately 30 -60 min. Coordinately, TPA induced a transient decrease in the level of the 180-kDa native ErbB-4 at 30 -60 min. Quantitatively, the TPA-mediated decrease of full-length ErbB-4 correlated well with the TPA-mediated increase of the 80-kDa fragment. The transient nature of TPA-induced cleavage of ErbB-4 is consistent with the reported time courses of other cellular responses to TPA (36,37). In addition, the kinetics and agonist concentration of the TPA-stimulated degradation of ErbB-4 are consistent with that shown in Fig. 1 for the TPA-induced decrease in 125 I-heregulin binding.
Specificity of ErbB-4 Cleavage and Localization of the 80-kDa Fragment-To determine whether protein kinase C activation was selective for ErbB-4 or could induce the degradation of other ErbB family members, NIH 3T3 cells overexpressing each of the ErbB family members were assayed. In addition, cells expressing a chimeric ErbB-4 receptor (composed of the extracellular and transmembrane domains of the EGF receptor and the cytoplasmic domain of ErbB-4) were employed (24). Each of these transfected cells, which express receptor numbers or protein level comparable to that of the ErbB-4 expressing T47-14 cells, was treated with vehicle alone or TPA for 30 min or 1 h. After solubilization, each ErbB receptor was analyzed by immunoprecipitation and Western blotting using antibody specific for each receptor. As shown in Fig. 4, TPA produced degradation only of the wild-type ErbB-4 receptor. No detectable degradation was noted for the EGF receptor or ErbB-3, although TPA did induce the appearance of minor degradation products in ErbB-2-overexpressing cells. Other bands in the ErbB-1 and ErbB-3 blots are nonspecific.
It is worthwhile to note (Fig. 4, lanes 10 -12) that TPA did not induce proteolytic cleavage of the chimeric ErbB-4 receptor, which contains EGF receptor transmembrane and extracellular domain sequences. This result and the 80-kDa molecular mass of the ErbB-4 fragment suggest that TPA-induced cleavage of ErbB-4 occurs within the transmembrane or extracellular domains. If so, the 80-kDa ErbB-4 fragment should be membrane-localized. Therefore, cell fractionation experiments were performed. Cells expressing wild-type ErbB-4 receptor were incubated in the absence or presence of TPA, and following homogenization, cytosolic and membrane fractions were prepared by differential ultracentrifugation. As shown in Fig.  5, the 80-kDa fragment as well as the full-length ErbB-4 receptor were detected only in the membrane fraction. Washing the membrane fraction with 1 M NaCl did not remove the 80-kDa fragment (data not shown), consistent with the fragment being anchored to the membrane by a transmembrane domain.
Since TPA treatment of T47-14 cells generates an 80-kDa membrane-associated, cytoplasmic domain fragment of ErbB-4, the receptor extracellular domain may be released into the medium. Therefore, following TPA treatment, the medium was analyzed with an antibody to the extracellular domain of ErbB-4. Since this antibody recognized only the native form of this receptor, immunoprecipitation, but not Western blot analysis, was employed. T47-14 cells were incubated for 24 h with 35 S-labeled methionine, the medium was replaced with nonradioactive medium, and the cells were incubated for 30 min at 37°C in the absence or presence of TPA. The medium was then immunoprecipitated with antibody to the ErbB-4 ectodomain. As shown in Fig. 5B (lane 8), a radiolabeled band of approximately 120 kDa was detected in the medium of cells treated with TPA. These data indicate that TPA induces cleavage of the ErbB-4 receptor, producing a membrane-associated cytoplasmic domain of 80 kDa and a soluble extracellular domain of 120 kDa. While the extracellular domain of ErbB-4 is likely to contain disulfide bonds, a reducing gel was employed to detect the extracellular 120-kDa fragment. Therefore this fragment most likely is a single polypeptide chain.
Characterization of the TPA-stimulated Cleavage Process-While the capacity of TPA to specifically activate most protein

FIG. 2. Analysis of ErbB-4 protein.
T47-14 cells were treated with vehicle alone, heregulin (100 ng/ml), TPA (100 ng/ml), or heregulin and TPA simultaneously for the indicated periods of time. Cell lysates (300 g) were precipitated with polyclonal antibodies to ErbB-4 receptor. Panel A, the immunoprecipitates were electrophoresed and blotted with ErbB-4 antiserum, and the bound antibody was visualized by the addition of 125 I-labeled protein A and autoradiography. Panel B, the Western blot was stripped (62.5 mM Tris-HCl, pH 6.8, 2% SDS, and 100 mM 2-mercaptoethanol) at 60°C for 30 min and blotted with phosphotyrosine (PY) antibodies, and bound antibody was visualized by ECL. The position of the 180-kDa full-length ErbB-4 receptor and its 80-kDa degradation product are indicated by arrows.
kinase C isozymes is established, it is possible that the observed TPA-induced cleavage of ErbB-4 might be mediated by an unknown action of the phorbol ester. Therefore, other mechanisms of modulating protein kinase C activity were tested. Protein kinase C down-regulation was achieved by prolonged TPA stimulation (20 h) which is known to down-regulate many, but not all, protein kinase C isozymes (38). As shown in Table  I (experiment 1), protein kinase C depletion blocked ErbB-4 cleavage by a subsequent addition of fresh TPA. A different means of affecting protein kinase C inhibition was also tested by briefly pretreating cells with GF 109203X, a selective inhibitor of Ca 2ϩ -sensitive protein kinase C isozymes, prior to TPA addition (39). The data in Table I (experiment 1) demonstrate that pretreatment with this inhibitor suppressed TPA-induced cleavage of ErbB-4. These results indicate that proteolysis of ErbB-4 in the presence of TPA requires protein kinase C activity.
We also employed non-TPA protein kinase C agonists and assayed their capacity to stimulate ErbB-4 proteolysis. As shown in Table I (experiment 2), the Ca 2ϩ ionophore ionomycin, the synthetic diacylglycerol diC 8 , and PDGF each induced significant ErbB-4 degradation. However, the individual effects of these agonists were not as great as that observed in response to TPA, which is a particularly potent and nonmetabolizable protein kinase C activator. The data in Table I also show that FIG. 4. Selectivity of the TPA-induced proteolytic cleavage within the ErbB receptor family. Cells expressing the EGF receptor, ErbB-2, ErbB-3, ErbB-4, and the chimeric ErbB-4 receptor were incubated at 37°C for 30 min or 1 h in the absence or presence of TPA (100 ng/ml). Cell lysates (500 g) were subjected to precipitation using antibodies specific for each ErbB receptor. The immunoprecipitates were electrophoresed, transferred to nitrocellulose membranes, and blotted with the appropriate antibodies. Bound antibody was detected with 125 I-protein A. The 80-kDa fragment amounts were quantitated with a PhosphorImager (Panel B). The maximum level of the 80-kDa fragment was obtained after a 1-h treatment with 100 ng/ml TPA, and this was set at 100%. the TPA-induced cleavage of ErbB-4 is not inhibited by cycloheximide pretreatment and, therefore, does not require the synthesis of new proteins.
Influence of TPA on ErbB-4 Kinase Activity-While TPA induces the cleavage of the ErbB-4 receptor, it is plausible that the resulting 80-kDa fragment might be an activated tyrosine kinase. In vivo and in vitro assays were performed to assess the kinase activity of the 80-kDa ErbB-4 fragment.
To determine whether the TPA cleaved fragment was an active tyrosine kinase in intact cells, cells were incubated with TPA or heregulin and the tyrosine phosphorylation of Shc and PLC-␥1 were assayed. As shown in Fig. 6A, the tyrosine phosphorylation of Shc and PLC-␥1 by heregulin activation of the native ErbB-4 receptor is evident (Fig. 6A, lanes 2 and 5), while TPA treatment did not stimulate the tyrosine phosphorylation of either substrate (Fig. 6A, lanes 3 and 6). When unfractionated lysates from cells treated with TPA or heregulin were probed by anti-phosphotyrosine blotting, there was no evidence of TPA-induced increases in total cell tyrosine phosphorylation (data not shown). However, the Shc and PLC-␥1 immunoprecipitates from TPA-treated cells do indicate the presence of coprecipitating proteins whose migration corresponds to that of the 80-kDa ErbB-4 fragment or the 180-kDa native ErbB-4 receptor. Therefore, the blots were stripped and reprobed with antibody to ErbB-4 (Fig. 6B). The results show that both the 80-kDa and native ErbB-4 molecules are coprecipitated with these two substrates. The amount of Shc or PLC-␥1 protein present in each precipitate was similar regardless of cell treatment (data not shown). These results suggest that the 80-kDa ErbB-4 fragment, although able to associate with Shc and PLC-␥1, does not phosphorylate these potential substrates in intact cells.
The tyrosine kinase activity of the 80-kDa ErbB-4 fragment was also assayed in an immune complex assay in vitro. Control and TPA-treated cells were solubilized and ErbB-4 molecules immunoprecipitated from each lysate. The immunoprecipitates were then resuspended in kinase buffer containing unlabeled Immunoprecipitates were separated on a polyacrylamide gel, transferred to nitrocellulose, and analyzed by Western blotting with anti-ErBb-4. Bound antibody was detected with 125 I-labeled protein A and autoradiography. Solid arrows indicate positions of the full-length receptor and the 80-kDa fragment. Panel B, subconfluent T47-14 cells grown in 60-mm dishes were incubated for 24 h in methionine-free Dulbecco's modified Eagle's medium containing 1% dialyzed fetal calf serum and Trans 35 S-label (100 Ci/ml). After one wash, 0.5 ml of methionine-free Dulbecco's modified Eagle's medium containing 1% dialyzed fetal calf serum, 20 mM Hepes, pH 7.4, and 0.1% BSA was added. Cells were then treated with vehicle alone or TPA (100 ng/ml) for 30 min at 37°C. The medium was then recovered and subjected to precipitation using an antibody to the extracellular domain of ErbB-4. The immunoprecipitates were extensively washed and loaded onto a polyacrylamide gel. The gel was dried and exposed to x-ray film. The open arrow indicates the position of the 120-kDa extracellular ErbB-4 domain.

TABLE I
Influence of protein kinase C modulators on formation of the 80-kDa ErbB-4 fragment Cells were incubated at 37°C with the indicated modulators of protein kinase C for the indicated periods of time and subsequently assayed by Western blotting for the ErbB-4 80-kDa fragment, as discussed under "Experimental Procedures." 80-kDa fragment amounts were quantitated using a densitometer (Bio-Rad) and the amounts of the 80-kDa fragment obtained with TPA treatment considered 100%. For each treatment 10-, 30-, and 60-min time points were analyzed, and the time at which the amount of the 80-kDa fragment was maximal is reported. Normalizing for protein levels, the 80-kDa fragment is estimated to be 20% as active as the native ErbB-4 molecule.

DISCUSSION
Activators of protein kinase C modulate the activity and/or structure of many transmembrane molecules encoding growth factors and their receptors within the EGF/ErbB family. Regulation of the EGF receptor and its ligands by protein kinase C occurs at several levels: stimulation of transcription of mRNA encoding the ligands heparin-binding EGF (40,41) and amphiregulin (42); stimulation of the cleavage of the transmembrane forms of transforming growth factor-␣ (43-45) and heparinbinding EGF (46) to produce diffusable, extracellular, low molecular weight ligands; and attenuation of EGF receptor functions by phosphorylation of serine/threonine residues within its cytoplasmic domain (47). ErbB-2 does not bind a known ligand, but rather forms heterodimeric complexes with the EGF receptor (15,16), as well as ErbB-3 (11,12) and ErbB-4 (14). ErbB-2 is phosphorylated by protein kinase C and its tyrosine kinase activity decreased (28).
Heregulin isoforms, which serve as ligands for the ErbB-3 and ErbB-4 members of the EGF receptor family (10), are also synthesized as transmembrane molecules and protein kinase C activation stimulates their cleavage leading to the release of soluble extracellular growth factors (48). The data in this report demonstrate that protein kinase C activation regulates the ErbB-4 receptor by inducing a proteolytic cleavage that yields a membrane-localized cytoplasmic domain of 80 kDa and a soluble ectodomain of 120 kDa. This is the first demonstration of protein kinase C control of transmembrane receptors that directly bind heregulin. Whether ErbB-3 is directly modulated by protein kinase C activity is unknown, although our studies demonstrate that any such regulation would not likely include stimulation of proteolysis. Hence, the protein kinase C-stimulated cleavage of ErbB-4 is unique within the EGF receptor family of related transmembrane molecules.
Protein kinase C activation does induce the cleavage of sev- eral other receptor tyrosine kinases, including the colony-stimulating factor-1 receptor (150 kDa), the c-kit receptor (150 kDa), and the receptor tyrosine kinase Axl (140 kDa). In each case, as with ErbB-4, proteolytic cleavage occurs at or near the extracellular face of the plasma membrane generating soluble ectodomains of 100 kDa for the colony-stimulating factor-1 receptor (49), 95 kDa for the c-kit receptor (50,51), and 80 kDa for the Axl receptor (52). However, the protein kinase C activated protease(s) for any growth factor precursor or transmembrane receptor has not been identified. Cleavage of these growth factor receptor tyrosine kinases produces cytoplasmic fragments that each contains a complete tyrosine kinase domain. There is no evidence, however, that these membranelocalized or cytoplasmic tyrosine kinases are activated. Our experiments demonstrate that the tyrosine kinase fragment of ErbB-4 does not phosphorylate substrates in intact cells following TPA treatment and is also inactive in kinase assays in vitro. The lack of activity in these tyrosine kinase molecules may be due to the loss of dimerization capacity which is generally facilitated by the ligand-binding extracellular domains. The ErbB-4-derived 80-kDa cytoplasmic domain fragment does contain phosphotyrosine and is, therefore, able to associate with SH2 domain-containing molecules (Shc and PLC-␥1) that are ErbB-4 tyrosine phosphorylation substrates. The question may arise as to why this fragment is tyrosine-phosphorylated in cells, if its kinase domain is not active. There are two plausible rationales for this circumstance. The basal level of autophosphorylation of ErbB-4 is high, and the tyrosine phosphate present on the 80-kDa fragment may have been present on the native receptor prior to proteolysis. Alternatively, the native ErbB-4 receptor or another tyrosine kinase may utilize the 80-kDa fragment as a substrate.
We detect the ErbB-4 ectodomain as a 120-kDa fragment present in the medium of cells exposed to TPA. In many polypeptide hormone systems, soluble receptor ectodomains are able to bind ligand and may, in some cases, serve as negative regulators. We do not have evidence as to whether the ErbB-4 120-kDa ectodomain fragment detected in these studies is able to bind heregulin. However, Dong et al. (53) have expressed the ectodomain of ErbB-4 as a soluble fusion protein and shown that it is able to block the activity of neural differentiation factor, a heregulin isoform.