Tumor necrosis factor-alpha-converting enzyme is required for cleavage of erbB4/HER4.

HER4 is a member of the epidermal growth factor receptor family and has an essential function in heart and neural development. Identification of two HER4 isoforms, HER4 JM-a and JM-b, which differ in their extracellular juxtamembrane region and in their susceptibility to cleavage after phorbol ester stimulation, showed that the juxtamembrane region of the receptor is critical for proteolysis. We now demonstrate that phorbol ester and pervanadate are effective stimuli for HER4 JM-a processing and that the HER4 JM-b isoform does not undergo cleavage in response to any of the stimuli studied. We also show that HER4 JM-a is not cleaved in cells lacking the metalloprotease tumor necrosis factor-alpha-converting enzyme (TACE) and that reexpression of TACE in these cells restores constitutive and regulated processing of HER4 JM-a. Moreover, we show that the sequence specific to the HER4 JM-a juxtamembrane region is sufficient to confer susceptibility to phorbol 12-myristate 13-acetate-induced cleavage of the HER2 receptor. In conclusion, we provide evidence that TACE is essential for the regulated shedding of the HER4 JM-a receptor.

The HER4/erbB4 receptor tyrosine kinase is a member of the EGF 1 receptor family, which also includes the EGF receptor (HER1 or erbB1), HER2 (erbB2 or Neu), and HER3 (erbB3) (1)(2)(3)(4). HER4 is a receptor for the neuregulins (NRGs) (5), a family of growth and differentiation factors that includes NRG (also known as acetylcholine receptor-inducing activity (6), glial growth factor (7), heregulin (8), and neu differentiation factor (9)), NRG2 (10 -12), NRG3 (13), and NRG4 (14). HER4 can also bind and be activated by heparin-binding EGF growth factor (15), betacellulin (16), and epiregulin (17), members of the EGF family which, unlike the NRGs, are also ligands for the EGF receptor. HER4 mRNA is expressed in several tissues such as heart, brain, kidney, and skeletal muscle (2), suggesting that this receptor is involved in the development and maintenance of a variety of organs and cell types. In fact, HER4 is essential for heart development: HER4-deficient mice die as embryos due to the lack of myocardial trabeculae (18). HER4 knockout mice also manifest neural defects, indicating that its function is critical for nervous system development as well. HER4 has also been implicated in the pathogenesis of different types of cancer (2,19,20). Thus, it is likely that the control of expression and function of HER4 is important in normal development as well as in disease.
In an earlier study, we reported the identification of a novel isoform of HER4 (HER4 JM-b) that differs from the original receptor (HER4 JM-a) in the juxtamembrane portion of the extracellular domain of the receptor. The juxtamembrane region of HER4 JM-b contains a stretch of 13 amino acids that replaces a stretch of 23 amino acids present in HER4 JM-a. The two receptor isoforms differ in their tissue distribution. The two isoforms also vary in their response to phorbol ester stimulation, which leads to a rapid down-regulation of surface expression of HER4 JM-a but not of HER4 JM-b. From this study, it was concluded that the JM-a isoform represents a cleavable form of HER4, whereas the JM-b form is uncleavable, and the differences in receptor processing as well as tissue distribution may reflect differences in their biological activities (21).
It has been shown by other investigators that the downregulation of surface HER4 in response to phorbol ester stimulation results from the shedding of the ectodomain of HER4 (22) and that a metalloprotease appears to be responsible for the cleavage (23). However, these studies did not identify the protease responsible for the cleavage of HER4 or the region of the receptor required for receptor shedding. Members of the metalloprotease-disintegrin family (also known as ADAMs) have been shown to participate in protein ectodomain shedding (reviewed in Refs. 24 -26). Among them, tumor necrosis factor-␣-converting enzyme (TACE or ADAM 17) is the metalloprotease whose catalytic function is best characterized. TACE was isolated based on its ability to cleave tumor necrosis factor-␣ precursor (27,28) and has been found to play a central role in protein ectodomain shedding of a variety of structurally and functionally unrelated transmembrane molecules (29 -31).
In the present study, we demonstrate that TACE is involved in the processing of the HER4 JM-a receptor in response to phorbol ester and pervanadate. Moreover, we show that the sequence of the juxtamembrane domain of HER4 JM-a is sufficient to make another HER family member, the HER2 receptor, susceptible to constitutive as well as regulated cleavage.
Plasmids-The cDNA encoding mouse TACE was obtained by KpnI/ NotI digestion from a pcDNA3-based construct and ligated into PCGN1 using the same restriction sites. The 1.3-kb cDNA encoding the dominant-negative mouse kuzbanian (MKUZDN) was obtained by a BamHI/ EcoRV restriction digestion of a pBluescript SK-based plasmid (32) and cloned into the PCGN1 vector using the same restriction sites.
The PCGN1 bicistronic vector is a pcDNA3 (Invitrogen)-based plasmid with a multiple cloning site down stream of the cytomegalovirus enhancer-promoter. The bovine growth hormone polyadenylation signal and transcription termination sequence are located 3Ј of a multiple cloning site. This plasmid leads to expression of myrEGFP and the neomycin resistance fusion gene under the control of the mini-elongation factor promoter. This vector was generated by Michael Lin (Children's Hospital Boston). All enzymes were from New England Biolabs. All clones were confirmed by sequencing.
Generation of the HER2 JM-a Expression Clone-A plasmid expressing full-length HER2 under the control of the long terminal repeat promoter was provided by S. W. Lee (Beth Israel Deaconess Medical Center, Boston, MA). To generate the chimeric receptor, we first generated a 108-bp HER4 JM-a fragment with HER2 flanking sequences by polymerase chain reaction using HER4 cDNA as a template and the following HER2/HER4 hybrid primers: F2 (5Ј-TCAACTGCACCCACT-CCTGTAACGGTCCCACTAGTCA-3Ј), and R1 (5Ј-ACGATGGACGTCA-GAGGGCTAGCATGTTGTGGTAAAGT-3Ј), which contains an AatII site. We also amplified a 119-bp fragment of the HER2 5Ј juxtamembrane region containing a SphI site, using full-length HER2 cDNA as a template. The following primers were used for this reaction: HER2 primer F1 (5Ј-CCCAGCGGTGTGAAACCTGA-3Ј), and HER2/HER4 hybrid primer R2 (5Ј-TGACTAGTGGGACCGTTACAGGAGTGGGTGC-AGTTGA-3Ј). Both 108-and 119-bp polymerase chain reaction products were isolated, pooled, and used as template for an additional polymerase chain reaction with the above-mentioned F1 and R1 primers. This reaction generated a 190-bp fragment that was then digested with SphI and AatII. This fragment was ligated on the 2.2-kb EcoRV-SacII fragment of HER2 that had been previously subcloned into the multiple cloning site of pBluescript SKϩ and cut with SphI and AatII. The 2.2-kb HER2 JM-a fragment was then cut with SacII and BstBI, and the 2041-bp fragment generated was ligated to the same sites of long terminal repeat-HER2 to generate long terminal repeat-HER2 JM-a.
ELISA-Cells were plated in 96-well plates and cultured to confluence. After the different treatments in serum-free medium, the cells were fixed with an equal volume of 8% paraformaldehyde solution (4% final concentration) for 15 min at room temperature. Cells were washed with PBS and then blocked with 3% bovine serum albumin-PBS for 1 h at room temperature. The mouse anti-human HER4 primary antibody (H4.77.16; NeoMarkers), which recognizes the extracellular region of HER4, was added to the wells at a 1:300 dilution in the same blocking solution and incubated overnight at 4°C with gentle rocking. The cells were then washed three times (10 min each) with PBS and incubated with a 1:1000 dilution of a biotinylated horse anti-mouse IgG (HϩL) (Vector Laboratories) for 1 h at room temperature. The cells were again washed with PBS, and the avidin-biotin-peroxidase complex (ABC Elite; Vector Laboratories) was used as directed for an additional hour at room temperature. After washing the plate with PBS, the coloring reaction was started by adding 100 l/well of the TMB peroxidase EIA substrate (Bio-Rad). The reaction was allowed to proceed for 5-10 min, and then 80 l of the supernatant of each well were transferred to 96 flat-bottomed flexible assay plates (Falcon), and the reaction was stopped with 40 l/well of 2N sulfuric acid. The resulting color intensity was measured in a Titertek Multiskan Plus reader at 450 nm. Each treatment point is the average of the computation of at least six wells. All experiments were repeated at least three times.
Immunostaining and Quantification of HER4-positive Cells-Transfected and untransfected cells in culture were fixed with 4% paraformaldehyde in PBS at room temperature for 20 min. Cells were washed with PBS and blocked with 3% bovine serum albumin (Sigma) in PBS (blocking solution) for 1 h at room temperature and incubated with the primary antibody (mouse anti-human HER4 (H4.77.16; NeoMarkers) or mouse anti-human HER2 (c-neu Ab-5; Oncogene Research Products)) diluted in blocking solution overnight at 4°C. The cells were washed with PBS, and the detection procedure was carried out using a Cy3 anti-mouse secondary antibody (Jackson ImmunoResearch Laboratories). Fields were counted first for GFP-positive cells (green fluorescence), and then the number of GFP-positive cells that were Cy3positive (red fluorescence) was calculated. Photographs from cells were obtained with a Nikon eclipse TE300 inverted microscope.
Cell Membrane Extracts-Transfected COS-7 cells were cultured in 100-mm dishes for 3 days, and after treatments, the cells were washed with cold PBS and scratched off the plate with a "cell lifter" (Costar). The pelleted cells were resuspended in 1 ml of membrane buffer (10 mM Tris, pH 7.5, 10 mM NaCl, 1 mM EDTA, pH 8, 10% sucrose, 1 g/ml leupeptin, 1 g/ml pepstatin, 1 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride) and homogenized with pestle A of a Dounce homogenizer in ice. The homogenate was centrifuged for 5 min at 14,000 rpm in an Eppendorf centrifuge 5415C at 4°C. The pellets were resuspended in radioimmune precipitation buffer (50 mM Tris, pH 8, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% deoxycholic acid, 1 g/ml leupeptin, 1 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride) and spun down again under the same conditions, and the supernatant was retained. In some cases, to separate membrane from cytosol, the supernatants of the membrane buffer spin were ultracentrifuged at 200,000 ϫ g in a TLA120.1 rotor for 15 min at 4°C, and the pellets of this centrifugation were resuspended in lysis buffer (50 mM Tris, pH 8, 150 mM NaCl, 1% Nonidet P-40, 1 g/ml leupeptin, 1 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Protein content was measured using the Bio-Rad Protein Assay.
Immunoblotting-For COS-7 cell membrane extracts, aliquots containing the same amount of protein were diluted 1:1 in 2ϫ dithiothreitol sample buffer (125 mM Tris-HCl, pH 6.8, 20% glycerol, 6% SDS, 0.1 mg/ml bromphenol blue, and 100 mM dithiothreitol), resolved in SDSpolyacrylamide minigels, and transferred to Immobilon-P (polyvinylidene difluoride) membranes (Millipore) with a Mini Trans-Blot ® cell (Bio-Rad). Membranes were blocked, and the transferred proteins were probed with a mouse anti-HER2 (c-neu Ab-3; Oncogene Research Products) that recognizes an intracellular epitope of the receptor, a rabbit anti-c-insulin receptor ␤ (D. Vicent and C. R. Kahn; Joslin Diabetes Center, Boston, MA), or a mouse anti-APP (clone 22C11). For NIH 3T3 or immortalized embryonic fibroblasts expressing HER4, whole cells lysates in 2ϫ dithiothreitol sample buffer (starting from the same number of cells in each treatment) were prepared. Lysates were resolved by SDS-polyacrylamide gel electrophoresis and transferred to Immobilon-P. The membranes were blocked and incubated with a rabbit polyclonal anti-HER4 antibody that recognizes the intracellular region of the receptor (RB-284-P; NeoMarkers). All blots were incubated with the appropriate peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories), developed with Renaissance ® (NEN Life Science Products), and exposed to autoradiography film (X-OMAT TM Blue XB-1; Kodak).
Metabolic Labeling and Immunoprecipitation-Cells were grown to 80% confluence in 60-mm dishes and then labeled with 200 Ci/ml Tran 35 S-label in methionine-and cysteine-free Dulbecco's modified Eagle's medium ϩ 5% dialyzed FBS for 24 h. After a rinse with methionine and cysteine-free Dulbecco's modified Eagle's medium ϩ 5% dialyzed FBS, the cells were treated with 1 ml of serum-free Dulbecco's modified Eagle's medium with or without PMA. The supernatants were precleared by incubation with 25 l of protein A/G-Sepharose for 2 h at 4°C. The supernatants were recovered and incubated with 40 l of protein A/G-Sepharose containing 1 g of anti-HER4 antibody (H4.77.16; NeoMarkers) and incubated overnight at 4°C. Beads were then spun down, washed three times with cold PBS, resuspended in 2ϫ dithiothreitol sample buffer, and heated to 65°C for 10 min, and the supernatant was loaded onto a 5.5% polyacrylamide gel. The gel was dried and exposed to autoradiography film. Statistical Analysis-Unpaired t tests were performed using Stat-View software.

Phorbol Ester and Pervanadate Induce Cleavage of HER4
JM-a but not of HER4 JM-b-Previous studies showed that the HER4 receptor is endocytosis-impaired (33) but that the level of receptor surface expression can be regulated by receptor shedding, which can be induced by phorbol ester (22) and pervanadate (44). With the identification of different juxtamembrane isoforms of HER4, we demonstrated that PMA, a potent activator of protein kinase C (PKC), led to a reduction of radioactive NRG␤1 binding in cells expressing the JM-a but not the JM-b isoform (21). Thus, these results suggested that HER4 JM-b is not cleaved in response to PMA stimulation. We have now developed an ELISA to measure the relative amount of receptor present in the cell surface, using a monoclonal antibody against the extracellular domain of the HER4 receptor. We first used this assay to confirm our observation on the differences between the response of the two HER4 isoforms to PMA stimulation, using NIH 3T3 cell lines that express similar levels of the HER4 isoforms (Ref. 21; see Fig. 1B). When cells expressing HER4 JM-a were treated with PMA (100 ng/ml for 45 min), the ELISA showed a Ͼ50% reduction in the HER4 surface immunoreactivity. In contrast, when cells expressing HER4 JM-b were treated the same way, there was no change in HER4 surface immunoreactivity (Fig. 1A). These results replicate our previous finding that surface expression of the JM-a isoform but not the JM-b isoform is reduced in response to PMA stimulation.
To confirm that the decreased cell surface expression of HER4 JM-a after phorbol ester treatment is due to receptor cleavage, we analyzed the effects of PMA on HER4 by Western blot and metabolic labeling. NIH 3T3 cells expressing either HER4 JM-a or JM-b were treated with vehicle or PMA, and the levels of HER4 were monitored by Western blot using an antibody directed against the intracellular domain of the receptor. Whereas PMA treatment resulted in a small increase in the intensity of the band corresponding to the full-length receptor A, NIH 3T3 cells expressing HER4 JM-a (clone AC2/G3) or HER4 JM-b (clone BC15/H12) were treated with vehicle (Ⅺ) or PMA (100 ng/ml for 45 min; f). The cells were then fixed, and the levels of HER4 expression were measured by an ELISA. PMA induced a dramatic reduction in HER4 JM-a cell surface immunoreactivity (p Ͻ 0.0001) but not in HER4 JM-b reactivity. B, whole cells lysates of cells treated as described above were subjected to Western blot with an antibody directed to the intracellular domain of HER4. PMA treatment of cells expressing HER4 JM-a led to a reduction in the amount of fulllength receptors (180-kDa band) and the appearance of an 80-kDa band corresponding to the intracellular domain of the receptor. In contrast, PMA did not induce a reduction in the 180-kDa band and the appearance of the 80-kDa band in HER4 JM-b-expressing cells. C, cells were metabolically labeled before exposure to PMA, and after treatment, supernatants were subjected to immunoprecipitation with an antibody directed to the extracellular domain of HER4. A 120-kDa protein could be detected only in PMA-treated HER4 JM-a-expressing cells. D, the effects of other treatments on HER4 shedding were tested by ELISA. Cells expressing HER4 JM-a (Ⅺ) and HER4 JM-b (f) were treated with vehicle (Control), thapsigargin (TG), forskolin (F), okadaic acid (OA), calyculin A (CA), or pervanadate (PV), and the levels of HER4 surface immunoreactivity were measured. Only pervanadate induced HER4 JM-a shedding to a level similar to that obtained with PMA. Calyculin A treatment induced a small but significant reduction in HER4 JM-a surface immunoreactivity. None of the treatments induced significant cleavage of HER4 JM-b. The data represent the mean Ϯ S.E. of three independent experiments.
in HER4 JM-b-expressing cells, this band was dramatically reduced by PMA treatment in the HER4 JM-a-expressing cells (180 kDa; Fig. 1B; see also Ref. 21). Concomitant with the reduction in full-length receptor, lysates of PMA-treated HER4 JM-a-expressing cells displayed an 80-kDa band, which most probably corresponds to the intracellular domain of the receptor (see Ref. 22). This band was absent from the PMA-treated HER4 JM-b-expressing cells.
Cleavage of cell surface receptors is expected to result in the release of their extracellular domain into the medium. This possibility was tested by metabolic labeling and immunoprecipitation. Cells expressing either HER4 JM-a or JM-b were metabolically labeled with [ 35 S]methionine and [ 35 S]cysteine and treated with vehicle or PMA, and the culture medium was subjected to immunoprecipitation with an antibody directed against the extracellular domain of HER4. When immunoprecipitates from PMA-treated HER4 JM-a-expressing cells were analyzed by gel electrophoresis, only a 120-kDa band was observed (Fig. 1C). This band was absent from untreated HER4 JM-a-expressing cells and from the untreated or treated HER4 JM-b-expressing cells. Thus, these results show that in response to PMA treatment, HER4 JM-a is cleaved, resulting in an 80-kDa intracellular fragment and the release to the medium of a 120-kDa extracellular domain fragment.
It has been reported that HER4 shedding (most probably the JM-a form) is induced not only by PMA but also by pervanadate, an inhibitor of tyrosine phosphatases (44). Thus, it is apparent that more than one signaling pathway regulates surface HER4 expression. We tested whether other signaling pathways can induce HER4 shedding, and whether the difference between the JM-a and JM-b isoforms extends to pervanadate and other stimuli. Cells were incubated in the absence or presence of thapsigargin (1 M), a blocker of the endoplasmic reticulum ATPase that leads to increased cytoplasmic calcium, forskolin (10 M), an activator of adenylate cyclase that leads to activation of protein kinase A, okadaic acid (10 nM) and calyculin A (10 nM), inhibitors of protein phosphatases 1 and 2A, or pervanadate (1 mM). The effects of these drugs on surface HER4 expression were tested by ELISA. Only pervanadate induced HER4 JM-a shedding similar to that induced by PMA (Fig. 1D). Calyculin A had a small effect on surface HER4 JM-a expression. The results indicate that shedding of the HER4 JM-a is predominantly regulated by PKC and tyrosine phosphorylation and that the JM-b isoform is consistently resistant to cleavage.
A Zinc-dependent Metalloprotease Is Responsible for the Shedding of HER4 JM-a-Metalloproteases have been shown to participate in shedding of the ectodomain of transmembrane proteins (34). To test whether the effects of PMA on the surface level of HER4 JM-a are due to the activation of a metalloprotease, we tested the effects of two structurally distinct metalloprotease inhibitors: 1) the metal chelator PNT, and 2) the hydroxamic acid-derived inhibitor IC-3 (also called TAPI-2). Addition of PNT at a low concentration (500 M) significantly reduced the effects of PMA on HER4 JM-a cleavage by 40% (p ϭ 0.0015). The effect of PNT on cleavage was reversed by adding ZnCl 2 to the medium, but not by adding CaCl 2 (data not shown), suggesting that the inhibitory effect of PNT is due to its metal-chelating activity and that the ion required for the proteolytic activity is zinc.
IC-3 was more effective than PNT in preventing the PMAinduced ectodomain shedding of HER4 JM-a (Fig. 2). IC-3 by itself induced a small but not statistically significant increase in HER4 levels, possibly through the reduction in constitutive cleavage of HER4 (see below). Simultaneous application of PMA and IC-3 not only completely prevented HER4 JM-a proc-essing but resulted in an increased number of cell membrane receptor (37% above control; p ϭ 0.0326). This could result from the blockade of cleavage by IC-3 and the induction of receptor insertion into the membrane by PMA. This is supported by our previous finding that PMA leads to a rapid increase of surface expression of the uncleavable isoform HER4 JM-b (Ref. 21; see also Fig. 1B). The effects of IC-3 and PNT implicate a zinc-dependent metalloprotease as the enzyme that processes HER4 JM-a in response to phorbol ester.
TACE Is Necessary for the Shedding of HER4 JM-a-Because metalloprotease inhibitors have been shown to revert PMA-induced HER4 shedding (present data; Ref. 23), we tested whether TACE was necessary for the proteolysis of this receptor. TACE is a disintegrin metalloprotease with a central role in the shedding of the ectodomain of several proteins (35). For these experiments, we generated HER4 JM-a-expressing cell lines using immortalized embryonic fibroblasts from mice homozygous for a mutation that deletes a domain essential for TACE metalloproteinase activity and from their wild-type littermates (29). When HER4-TACEϩ/ϩ cells were treated with PMA or pervanadate, surface HER4 JM-a expression was reduced (Fig. 3A). This reduction was similar to that seen in NIH 3T3 cells treated with PMA (Fig. 1A). In contrast, HER4 surface immunoreactivity was not reduced by PMA or pervanadate in the HER4-TACEϪ/Ϫ cells (Fig. 3A). Because HER4 regulated shedding is completely abolished in the absence of TACE, we concluded that TACE is necessary for the induction of HER4 JM-a cleavage. To confirm that the reduction in HER4 surface expression observed in the TACEϩ/ϩ cells is due to receptor cleavage, we used Western blot analysis and metabolic labeling. As in NIH 3T3 cells, when TACEϩ/ϩ cells were treated with PMA, there was a reduction in full-length receptor, the appearance of an 80-kDa HER4 intracellular fragment (Fig.  3B), and the appearance of a 120-kDa HER4 soluble extracellular domain (Fig. 3C). In contrast, there were no apparent changes in the HER4 protein in TACEϪ/Ϫ cells. Thus, these results show that HER4 JM-a is cleaved in TACEϩ/ϩ cells but not in TACEϪ/Ϫ cells. To further analyze the involvement of the metalloprotease TACE in HER4 processing, we transiently expressed TACE in the TACEϪ/Ϫ cells expressing HER4 JM-a (clone B7/G2). This was done by transfecting the cells with a bicistronic vector expressing GFP and mTACE. We then monitored the expression of HER4 JM-a in the GFP-expressing cells. As shown in Fig. 4A, when the vector alone (GFP) was transfected into the TACE knockout cells, PMA treatment did not have any effect on HER4 immunoreactivity, as in the experiment described above. However, when the plasmid expressing mTACE (GFP-mTACE) was transfected into the knockout cells, a reduction in HER4 surface immunoreactivity was evident even without PMA stimulation. The reduction in HER4 immunoreactivity was even more dramatic when the transfected cells were treated with PMA (Fig. 4B). The fact that overexpression of TACE stimulates HER4 cleavage in the absence of PMA suggests that this metalloprotease might be also involved in the basal or constitutive cleavage known to occur in this receptor (23). Together, these results demonstrate that TACE is necessary for the shedding of the extracellular domain of HER4 JM-a.
Kuzbanian (KUZ) Is Not Involved in HER4 JM-a Cleavage-KUZ (or its mammalian homologue ADAM 10) is another metalloprotease that has been implicated in the processing of several membrane proteins (32,36). KUZ is the ADAM family member most structurally similar to TACE (25,37), and it has been shown to cleave several proteins that are also substrates for TACE, including tumor necrosis factor-␣ (27,28,38) and APP (39). Therefore, we decided to determine whether KUZ was also involved in the cleavage of HER4 JM-a by using MKUZDN. KUZDN is a truncated form of the protease that blocks kuzbanian function in Drosophila, Xenopus, and mammalian cells (32). 2 A plasmid expressing MKUZDN was transfected into the NIH 3T3 HER4 JM-a-expressing cells, and the effects of PMA treatment on HER4 surface expression were monitored by immunofluorescence (Fig. 5A). We found that PMA treatment reduced the expression of surface HER4 to the same extent in cells transfected with the control plasmid or with the plasmid expressing MKUZDN (Fig. 5B). Although expression of MKUZDN did not affect HER4 processing, it reduced the constitutive processing of APP under these conditions in our system (data not shown), as it did in a previous report (39), confirming that it is a functional reagent. Thus, these results suggested that KUZ activity is not necessary for the PMA-induced shedding of HER4 and reinforce the evidence that TACE is the metalloprotease that cleaves HER4 JM-a.
Inserting the JM-a Domain in the HER2 Receptor Makes It Cleavable-In contrast to HER4, only one juxtamembrane sequence has been identified for HER2, and this receptor has been shown not to be cleaved in response to PMA (22). To test whether the juxtamembrane sequences specific to the JM-a isoform of HER4 are sufficient to confer PMA-induced receptor cleavability, we replaced the juxtamembrane region of HER2 with that of HER4 JM-a, hence generating a chimeric receptor we named HER2 JM-a (Fig. 6B). We set out to test whether the chimeric receptor could be cleaved in response to PMA in COS-7 cells. These cells were selected because large amounts of protein can be expressed by transient transfection, because HER4 JM-a in COS-7 cells is cleaved in response to PMA (Fig.  6A), and because these cells express TACE (data not shown). Plasmids expressing the wild-type HER2 or HER2 JM-a were transfected into COS-7 cells, the cells were treated with vehicle or PMA, and surface HER2 immunofluorescence was visualized with an antibody specific for the extracellular region of HER2. PMA treatment resulted in an almost complete disappearance of HER2 immunoreactivity in cells expressing HER2-JM-a but did not show any changes on cells expressing wildtype HER2 (Fig. 6C), suggesting that the JM-a sequence is sufficient to bestow PKC-activated cleavability to the HER2 receptor.
We further analyzed the processing of the HER2 receptor by Western blotting, using an antibody that recognizes the intracellular portion of HER2. In cells transfected with the wild-type HER2 receptor, a 185-kDa band was identified in both unstimulated and PMA-treated cells (Fig. 6D). This 185-kDa form 2 D. J. Pan, personal communication.

FIG. 3. PMA and pervanadate do not induce shedding of HER4 JM-a in TACE knockout (؊/؊) cells.
The effects of PMA and pervanadate were tested on HER4 JM-a-expressing cell lines generated from immortalized embryonic fibroblasts from TACEϪ/Ϫ and TACEϩ/ϩ mice. A, the effects of PMA and pervanadate were tested using the ELISA. Whereas in TACEϩ/ϩ cells, PMA and pervanadate induced HER4 shedding (PMA, clone G10/B5/F11, p ϭ 0.0002; clone C8/G2, p ϭ 0.0045; pervanadate, clone G10/B5/F11, p ϭ 0.0001), HER4 surface expression in the TACEϪ/Ϫ cells was not reduced after exposure to PMA or pervanadate. The data represent the mean Ϯ S.E. of three independent experiments. B, whole cells lysates of cells treated with PMA were subjected to Western blot with an antibody directed to the intracellular domain of HER4. PMA treatment of TACEϩ/ϩ cells (clone G10/B5/F11) led to a reduction in the amount of full-length receptors (180-kDa band) and the appearance of an 80-kDa band corresponding to the intracellular domain of the receptor. In contrast, PMA did not induce a reduction in the 180-kDa band and the appearance of the 80-kDa band in TACEϪ/Ϫ cells (clone B7/G2). C, cells (clone G10/ B5/F11 and clone B7/G2) were metabolically labeled before exposure to PMA, and after treatment, supernatants were subjected to immunoprecipitation with an antibody directed to the extracellular domain of HER4. A 120-kDa protein could be detected in PMA-treated TACEϩ/ϩ cells, but not in TACEϪ/Ϫ cells. of HER2 corresponds to the cell-associated full-length receptor (40), and its levels were not altered by the addition of PMA, whereas the levels of mature APP were altered by the addition of PMA (mature APP is a substrate for TACE, and its levels are reduced after PMA addition; Ref. 30). In contrast, when cells were transfected with HER2 JM-a under control conditions, a novel band with a molecular mass of approximately 85 kDa appeared together with the 185-kDa band. Moreover, when these HER2 JM-a-expressing cells were treated with PMA, the intensity of the 185-kDa band was reduced, and the intensity of the 85-kDa form increased substantially (Fig. 6D). The fact that the 85-kDa fragment is present even without PMA stimulation suggests that the chimeric HER2 receptor containing the JM-a sequence in the juxtamembrane region may be processed in a constitutive manner and that PMA stimulation results in more extensive cleavage. These results show the ability of the JM-a region to give rise to a cleavable HER2 receptor and suggest that the 85-kDa band corresponds to the intracellular portion of HER2 JM-a produced after processing of the extracellular domain by a protease, most probably TACE. DISCUSSION Changes in the level of expression of members of the EGF receptor family of receptor tyrosine kinases can have dramatic effects on cell function, whereas overexpression can lead to transformation (41,42). Thus, regulation of the number of functional receptors on the cell surface is a key event in the biology of the cell. The members of the EGF receptor family differ widely in the mechanism that regulates their cell surface expression. Whereas the EGF receptor (HER1) undergoes rapid internalization and down-regulation after ligand binding, none of the other family members do (33). HER2 has been recently reported to undergo cleavage, probably by a matrix metalloprotease, in a PKC-independent manner (43). Mechanisms that may regulate surface expression of HER3 have not yet been identified. The HER4 receptor has been shown to be regulated by cleavage through a PKC-dependent proteolytic pathway (Refs. 21 and 22; present study) and through a PKC-independent pathway (Ref. 44; present study). We have also previously shown that alternative spliced forms of HER4 (JM-a and JM-b) differ in their susceptibility to this proteolytic cleavage in that the HER4 JM-a isoform but not the JM-b isoform can be cleaved after phorbol ester stimulation (21). Now we demonstrate that: 1) PMA and pervanadate are the most effective stimuli for processing of HER4 JM-a, 2) the HER4 JM-b isoform does not undergo cleavage in response to any of the signaling pathways tested (PKC, protein kinase A, phosphatases, or calcium), 3) the JM-a juxtamembrane region is sufficient to permit cleavage of the HER2 receptor; and 4) most importantly, TACE is the Zn metalloprotease directly involved in the cleavage of HER4 JM-a.
TACE is a zinc-dependent metalloprotease-disintegrin mem-FIG. 6. Introduction of the specific HER4 JM-a sequences into HER2 results in constitutive as well as regulated HER2 cleavage. A, PMA induces HER4 JM-a shedding in COS-7 cells. COS-7 cells were transfected with the HER4 JM-a plasmid and treated with vehicle or PMA, and the levels of surface HER4 were measured by ELISA. B, schematic drawing of wild-type HER2 and chimeric HER2 JM-a receptors. The 14 amino acids of the juxtamembrane region of wild-type HER2 (gray oval) were replaced by the JM-a sequence of the HER4 isoform (gray rectangle). C, immunofluorescence of surface expression of HER2. COS-7 cells were transfected with a plasmid expressing wild-type HER2 (middle) or HER2 JM-a (bottom) or were not transfected (top) and stimulated with 500 ng/ml PMA for 25 min (right) or left unstimulated (left). The level of HER2 surface expression was detected by immunofluorescence using an antibody against the extracellular domain of HER2. Whereas HER2 immunoreactivity was not reduced in cells expressing HER2, PMA led to a dramatic reduction in HER2 JM-a expression. Scale bar, 10 m. D, Western blot analysis of receptor expression in cell membranes. Cell membrane extracts from COS-7 cells transfected with the HER2 and HER2 JM-a plasmid and treated with or without PMA were subjected to SDS-polyacrylamide gel electrophoresis, and the blot was probed with an antibody directed against the intracellular domain of HER2. Whereas untransfected COS-7 cells lack significant HER2 expression, HER2 immunoreactivity was detected in the transfected cells. The levels of full-length HER2 (p185) expression were similar in unstimulated and PMA-stimulated cells. Cells expressing HER2 JM-a, even before PMA stimulation, contained two HER2-immunoreactive bands, p185 (the full-length receptor) and p85 (the intracellular portion of the receptor), indicating receptor cleavage. PMA stimulation led to a substantial reduction in the amount of p185 and an increase in p85. The same membrane was probed with anti-insulin receptor ␤ subunit (c-IR␤) used to control for loading and to demonstrate that not all transmembrane receptors get cleaved. The bottom panel shows that PMA induces cleavage of the mature ␤-amyloid precursor protein (␤APP) in all cells, leading to a decrease in cell-associated mature ␤-amyloid precursor protein. ␤-Amyloid precursor protein was used as a positive control for increased TACE activity induced by PMA in the cells.
brane-anchored glycoprotein, which belongs to the ADAM/metalloprotease/disintegrin/cysteine-rich protein family. These family of proteases is involved in the processing of many cleavable integral membrane proteins (45,46), but the physiological substrates for only a few of them have been identified. TACE (ADAM 17) was identified as the enzyme responsible for cleaving tumor necrosis factor-␣ precursor (27,28). It has also been shown to be the regulated ␣-secretase for the Alzheimer amyloid protein precursor (30) and to be involved in the processing of other cell surface proteins such as L-selectin, transforming growth factor-␣, and TRANCE (29,31). ADAM 10 or its homologue, KUZ, has been implicated in the shedding of the cell fate determinants Delta and Notch (32,36), and it has been proposed to be the constitutive and regulated ␣-secretase for the Alzheimer amyloid protein (39). Finally, metalloprotease/disintegrin/cysteine-rich protein 9/meltrin/ADAM 9 has been shown to be involved in the processing of the membrane-anchored heparin-binding EGF-like growth factor (47). Other proteases that are not susceptible to inhibition by metal-chelating agents and whose activity is not up-regulated by phorbol esters may also be involved in the control of the number of integral cell membrane proteins (43,48). The present study shows that TACE is also necessary for the cleavage of the JM-a isoform of HER4, and our data suggest that TACE is responsible for the regulated and possibly also the constitutive cleavage of this receptor tyrosine kinase.
We have previously demonstrated that HER4 exists in vivo in two alternative forms that differ in their juxtamembrane regions (21), and, more recently, an isoform lacking a PI3kinase binding domain has also been described (49). Because only one HER4 gene has been identified (50), these isoforms are most likely generated by alternative splicing of a single HER4 RNA precursor molecule. Expression of HER4 isoforms with different functional properties, i.e. difference in the regulation of cell surface expression and in the activation of signaling pathways, is likely to result in different cellular responses to the same ligands. Alternative splicing has been shown to create functional diversity of other receptor tyrosine kinases. For example, two alternatively spliced forms of fibroblast growth factor receptor 2 exist that differ in their affinity for either fibroblast growth factor 2 or fibroblast growth factor 7 ligands (51). Our previous studies showed that HER4 JM-a and HER JM-b have similar affinities for their various ligands and appear to differ only in their susceptibility to cleavage by metalloproteases (Ref. 21; the present study).
With the exception of HER4, none of the other HER receptors have been observed to be cleaved after PKC stimulation. To test whether a specific juxtamembrane region of the HER4 JM-a can bestow the capability of being cleaved to other members of the EGF family of receptors after PKC activation, we replaced the normal 14-amino acid juxtamembrane region of the human HER2 with the 23 amino acids forming JM-a (see Fig. 6B). Cells expressing this chimeric HER2 receptor experienced PMA-regulated shedding, showing that the JM-a domain enables cleavage of at least one of the other members of the HER family. Interestingly, the HER2 JM-a receptor was also processed in a constitutive manner, and the cleavage was dramatically increased by phorbol ester stimulation. Basal hydrolysis of normal HER4 has been reported (23), and our data show that reintroducing TACE in HER4 JM-a-expressing TACE-null cells leads to a reduction in extracellular receptor immunoreactivity even in the absence of PMA and that PMA stimulation increased the processing substantially. Altogether, these results indicate that TACE is implicated in the constitutive and regulated processing of JM-a-containing receptors. This is different from other cases in which different proteases appear to regu-late the basal and regulated shedding of receptors, e.g., in the cases of APP ␣-secretases (26,30) or Notch cleavage (in which KUZ is only responsible for the regulated processing; Ref. 36).
Comparison of sequences of JM-a with those of other juxtamembrane regions susceptible to cleavage by metalloproteases (for a review, see Ref. 34) reveals little or no common pattern. It has also been shown that the cleavage of some membrane proteins occurs independently of amino acid sequence (52). There are some similarities in the flexibility profile, and the presence of proline residues in most of these domains could create "turns" and thereby present the intervening sequence to a membrane protease. These considerations raise questions about the roles of the cleaved amino acid sequences and the distance as determinants of cleavage specificity. In our study, replacement of the noncleavable HER2 juxtamembrane region by that of the HER4 JM-a receptor resulted in a cleavable chimeric receptor. Therefore, we have clearly shown that there are some structural determinants in the JM-a juxtamembrane sequence that make it accessible for cleavage.
The physiological significance of the ectodomain shedding is still unclear. It remains to be elucidated whether the HER4 JM-a shedding constitutes merely a process of efficient and rapid down-regulation or whether the released extracellular domain of HER4 can bind neuregulin or any other of its ligands in the extracellular milieu in vivo blocking its effects, or may lead to some kind of autocrine and/or paracrine signaling events through the transmembrane forms of the ligands in the cells. Further understanding of the regulation of this EGF family of receptors and their signal transduction pathways will be of particular interest because these receptors are implicated in regulating normal development (18,53,54) and cancer (20,55,56).