c-Cbl binding and ubiquitin-dependent lysosomal degradation of membrane-associated Notch1.

Regulation of Notch1 activity is critical for cell fate decisions and differentiation of skeletal myoblasts. We have employed the skeletal myoblast cell line C2C12 to study posttranslational regulation of Notch1 protein levels during myogenesis. Although the major degradation pathway of the activated intracellular Notch1 fragment appears to involve ubiquitination and degradation by the 26 S proteasome, we provide evidence for an alternative catalytic pathway where the endogenous, transmembrane form of Notch1 is targeted to the lysosomal compartment. Immunoprecipitation analysis revealed ubiquitin-dependent accumulation of transmembrane Notch1 protein after treatment with the lysosomal inhibitor chloroquine but not after treatment with various proteasome inhibitors. This finding was supported by the observation that the transmembrane form of Notch1 was tyrosine-phosphorylated and specifically coprecipitated with the ubiquitin ligase c-Cbl. Our data suggest a regulatory mechanism down-regulating Notch1 protein levels already at the cellular surface, possibly with consequences for Notch-dependent signal transduction during terminal differentiation processes.

Regulation of Notch1 activity is critical for cell fate decisions and differentiation of skeletal myoblasts. We have employed the skeletal myoblast cell line C2C12 to study posttranslational regulation of Notch1 protein levels during myogenesis. Although the major degradation pathway of the activated intracellular Notch1 fragment appears to involve ubiquitination and degradation by the 26 S proteasome, we provide evidence for an alternative catalytic pathway where the endogenous, transmembrane form of Notch1 is targeted to the lysosomal compartment. Immunoprecipitation analysis revealed ubiquitin-dependent accumulation of transmembrane Notch1 protein after treatment with the lysosomal inhibitor chloroquine but not after treatment with various proteasome inhibitors. This finding was supported by the observation that the transmembrane form of Notch1 was tyrosine-phosphorylated and specifically coprecipitated with the ubiquitin ligase c-Cbl. Our data suggest a regulatory mechanism down-regulating Notch1 protein levels already at the cellular surface, possibly with consequences for Notch-dependent signal transduction during terminal differentiation processes.
A critical cellular signaling pathway important for decisions whether to differentiate or not consists of the evolutionary highly conserved members of the transmembrane Notch receptor family. Generation of the mature Notch receptor at the cell surface and eventually Notch1 activation itself requires at least three critical proteolytic cleavage steps. The 300-kDa Notch1 polypeptide precursor, respectively, is first processed by a furin-like protease activity in the Golgi apparatus yielding an N-terminal extracellular fragment and a C-terminal transmembrane fragment (N TM ). 1 These two subunits associate to form the heterodimeric, mature receptor at the cellular surface (1,2). Signaling mediated by Notch-dependent pathways involves the interaction of the receptors with specific cell surface ligands of the Delta or Jagged families. It could be shown that ligand-dependent activation results in two additional proteolytic cleavages of the 120-kDa N TM fragment, one cut occurring extracellularly and the other intracellularly, close to the transmembrane domain at valine 1744. This allows the nuclear translocation of a now activated, 110-kDa intracellular portion of Notch1 (N IC ) where it may bind and regulate the activity of specific transcription factors, such as CBF1/RBPJ (reviewed in Refs. 3 and 4). In all cellular systems tested so far, it has been described as difficult to detect these endogenous, activated Notch fragments by direct biochemical methods, leaving only the membrane-associated N TM fragments as detectable protein.
Studies of Notch1 function can be addressed particularly during formation of skeletal muscle structures, where myoblastic cells coordinately proceed through differentiation stages to give rise to fused myotubes, whereas other cells of the same origin remain undifferentiated to maintain a population of precursor cells. It has been reported that MyoD and MEF2c activity required for myogenesis can be inhibited by activated Notch1 via a yet unknown mechanism (5,6).
Tight down-regulation of the activity of proteins and transcription factors in particular is often achieved via ubiquitination and subsequent degradation, preferentially in the proteasome. Ubiquitination is the result of multienzyme processes, involving protein complexes of ubiquitin-activating and ubiquitin-conjugating enzymes and typically require ubiquitin-protein ligases, like c-Cbl, Itch, or Sel-10, that mediate the substrate specificity (reviewed in Refs. 7 and 8). Genetic and biochemical evidence in invertebrates has suggested that proteasomal degradation of activated forms of Notch may be required for the cessation of Notch signaling (9 -11). In human Jurkat cells, cotransfection experiments indicated an interaction of the Hect-type ubiquitin ligase Itch with ectopically expressed Notch1 proteins (12). Nevertheless, polyubiquitination and degradation of proteins by the 26 S proteasome is not the only known mechanism for protein destruction. For example, it has been described that certain membrane surface receptors such as epidermal growth factor receptor may become poly-or mono-ubiquitinated and be targeted to and degraded within lysosomes. In this context, ubiquitination seems to work as an endocytosis signal targeting proteins to the lysosomal degradation machinery (13)(14)(15).
Here we present data addressing a posttranslational regulation mechanism of endogenous Notch1 in the course of vertebrate myoblast differentiation. We provide evidence for the targeting of cell membrane-associated endogenous Notch1 (N TM ) to the lysosomal protein degradation pathway. Furthermore, we demonstrate that posttranslational modifications of Notch1 include specific tyrosine phosphorylation and interaction with the ubiquitin-ligating protein c-Cbl as well as monoubiquitination. Thus, for the first time our data indicate a proteasome-independent degradation mechanism of Notch1 possibly down-regulating the activity of its membrane-bound form in differentiating cells.

EXPERIMENTAL PROCEDURES
Cell Culture-The mouse skeletal myoblast line C2C12 was grown in Dulbecco's modified Eagle's medium/Glutamax (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum, sodium pyruvate, and penicillin/streptomycin at 7.5% CO 2 . For passaging, cells were kept at low densities, and medium was exchanged every other day. To induce terminal differentiation and fusion, cells were grown to confluency and then shifted to fusion medium containing 2% horse serum. Extracts for Western blot analysis and immunoprecipitations were grown in parallel and prepared at indicated time points.
Northern Blot Analysis-Total RNA was extracted from C2C12 cells stimulated to undergo fusion by serum withdrawal for 0 -7 days. RNA was isolated using the RNeasy protocol (Qiagen) according to the manufacturer's instructions. 10 g of total RNA was separated on 1% agarose gels and ethidium bromide-stained. RNAs were subsequently transferred overnight to a nylon membrane. Hybridization was carried out using the non-radioactive DIG System for Nucleic Acid Labeling and Detection (Roche Molecular Biochemicals) and an in vitro transcribed, digoxigenin-labeled murine Notch1-specific antisense probe (base pairs 5328 -6958) as described in the manufacturer's protocol.
To generate the GST-Notch1 expression plasmid, the intracellular domain of TAN1 (human Notch1) was PCR-amplified using the vector construct pGD-ICT (16) as template. The PCR fragment was cloned to the 3Ј end of the GST tag into the eukaryotic expression vector pEBG.
EDTA Treatment of C2C12 Cells-Confluent C2C12 cells, kept for 48 h in fusion medium, were washed with PBS and incubated for 15 min with 5 mM EDTA in PBS at 37°C. Control cells were incubated in PBS alone. The stimulation was followed by recovery for either 30 min or 4 h in fusion medium. Subsequently, cells were treated for Western blot analysis and immunoprecipitations as described below.
Proteasome and Lysosome Inhibition-Confluent C2C12 were kept for 48 h in fusion medium before proteasomal and lysosomal inhibitors were added for various stimulation periods. Inhibitors were resuspended either in Me 2 SO or PBS and, if not indicated otherwise, were used for stimulation as follows: proteasome inhibitor I at 3 M final; MG132 at 5 M final; clasto-lactacystin ␤Ϫlactone at 2 M final (all from Calbiochem). Chloroquine (Sigma) was used at a final concentration of 100 M. Controls were either performed with equivalent volumes of Me 2 SO for the proteasome inhibitors or with equivalent volumes of PBS for chloroquine stimulations.
For immunoprecipitation, 5 ϫ 10 6 to 1 ϫ 10 7 cells were harvested in 1 ml of lysis buffer. Lysates were cleared by centrifugation and mixed with 5 g of the precipitating antibodies as follows: rabbit anti-Notch1, generated by immunization of rabbits with a peptide corresponding to amino acids 1766 -1780 of the mouse Notch1 protein 2 ; polyclonal rabbit anti-ubiquitin (Dako); mouse anti-phosphotyrosine (PY99; Santa Cruz Biotechnology); and monoclonal mouse anti-GST (Santa Cruz Biotechnology). The monoclonal mouse anti-HA antibody was a kind gift of Victor Wixler, Erlangen, Germany. After incubation for 2 h on ice 50 l of protein G-Sepharose beads (Pierce) were added to each sample and incubated overnight at 4°C with gentle agitation. Beads were washed four times, resuspended in 1ϫ SDS loading buffer, and loaded immediately on denaturing 8% SDS-PAGE gels. Proteins were transferred and visualized as described above.
Preparation of Subcellular Fractions Enriched for Endosomal and Lysosomal Compartments-C2C12 cells were grown to confluency as described above and shifted for 48 h to Dulbecco's modified Eagle's medium containing 2% horse serum. Cells were washed 3 times with PBS and lysed in 2 ml of cold homogenization buffer (0.25 M sucrose; 25 mM KCl; 5 mM MgCl 2 ; 10 mM Hepes-NaOH, pH 7.4, supplemented with protease inhibitors) per 20-cm plate. After disruption of the cells in a Teflon-glass homogenizer (Potter-Elvehjem), nuclei, membranes and other heavy organelles were pelleted by a 10-min centrifugation step at 1000 ϫ g. Supernatants were then centrifuged at 3000 ϫ g for 10 min, and the supernatants were removed again and centrifuged at 17,000 ϫ g for 10 min to harvest the lysosome-enriched fraction. Pellets from the 1000 and 17,000 ϫ g centrifugation steps were resuspended in RIPA buffer and analyzed on denaturing SDS-PAGE gels as described.

Notch1
Protein Levels Are Regulated during Differentiation of Myogenic Cells-Myogenic cells in culture are capable of proliferating in the presence of rich nutrients and undergo terminal differentiation and fusion after a shift to low serum levels. To analyze changes in Notch1 expression and protein levels during myoblast differentiation, C2C12 myoblasts were cultured in 10% serum, and terminal differentiation was induced by a shift to medium containing 2% horse serum. Although Notch1 mRNA (ϳ8.0 kb) was detectable up to 7 days after induction of terminal differentiation with a maximum at day 2 ( Fig. 1A), the amount of N TM , representing a 120-kDa protein, appeared to be regulated more significantly (Fig. 1B). Western blot analysis with a Notch1-specific antibody revealed that Notch1 protein increased from low levels at day 0 and 1 to highest levels ϳ48 h after induction of terminal myoblast differentiation via low serum cultivation. On days 3 and 4 Notch1 levels decreased slightly and reached low to undetectable levels within 7 days of terminal myoblast differentiation (Fig. 1B).
To demonstrate that the Notch1-specific signal shown in Fig.  1B represented the transmembrane anchored N TM fragment, C2C12 cells were treated with 5 M EDTA. EDTA treatment has been shown to mimic Notch-ligand binding, causing the release of the activated intracellular N IC fragment (17). Using Notch1 specific antibodies in a Western blot demonstrated the emergence of a smaller 110-kDa protein fragment, when cells were treated with EDTA followed by a 30 min of recovery in medium (Fig. 2, lanes 5 and 7). This 110-kDa fragment, representing the activated N IC protein, is not detectable when cells were incubated in PBS only (Fig. 2, lanes 1-4). The 110-kDa band disappeared when cells were allowed to recover for 4 h after EDTA treatment, very likely reflecting the rapid degradation of N IC (Fig. 2, lanes 6 and 8). This analysis indicates that the Notch1 protein detectable in our experiments represented membrane-associated 120-kDa N TM fragments.
N TM Fragments Are Associated with a Ubiquitinated Protein Complex in C2C12 Cells-During terminal differentiation of the myogenic cell line C2C12, Notch1 protein levels were significantly reduced. As shown in Fig. 1B, Notch1 protein increased at early stages of myoblast fusion with the highest levels at day 2 and declined at later time points of differentiation to low to undetectable levels. Therefore, the question was addressed whether N TM is regulated posttranslationally via ubiquitination and subsequent proteolytic degradation. For this purpose immunoprecipitation assays were performed using antibodies specific for ubiquitin and Notch1 to determine whether endogenous N TM is an integral part of ubiquitinated protein complexes or even ubiquitinated itself. As shown in Fig.  3A, lane 1, N TM fragments can be precipitated and detected specifically by using a polyclonal Notch1 antibody. We were also able to pull down N TM protein, employing a ubiquitinspecific antibody (USAB) in immunoprecipitation experiments (Fig. 3A, lane 2). The appearance of a single Notch1 band from lysates precipitated with USAB may either indicate mono-ubiquitinated N TM or N TM that is specifically associated with a ubiquitinated protein complex.
To demonstrate the specificity of the Notch1-specific antibody, a series of control precipitations was performed (Fig. 3B). Precipitations from lysis buffer alone (Fig. 3B, lane 1) or from the indicated cell lysate with an isotype control antibody (Fig.  3B, lane 2) did not reveal any specific immunoprecipitate in Western blot analysis using the Notch1-specific antibody. Percolating of the cellular lysate with an isotype control antibody prior to an immunoprecipitation with Notch1 antibody did not abolish the specific N TM precipitation (Fig. 3B, lane 3), whereas this was the case when the precipitating antibody was coincubated with its specific blocking peptide (Fig. 3B, lane 5).
To verify whether Notch1 is a target for ubiquitin conjugation, Notch1 was expressed in a heterologous cell system using a transient transfection protocol (Fig. 4). COS1 cells were transiently transfected with a hemagglutinin (HA)-tagged ubiquitin-expression construct and with an expression vector, encoding the intracellular Notch1 (GST-N IC ) domain fused to a GST tag. These constructs were transfected either alone or together into COS1 cells, and lysates were prepared 48 h posttransfection and subjected to immunoprecipitation analysis with antibodies specific for ubiquitin, Notch1, or the GST and HA tags. Subsequently, Western blot analyses were performed using tag-specific antibodies. It is noteworthy that the fusion protein of GST-tagged Notch1 has a predicted molecular mass FIG. 1. mRNA and protein levels of Notch1 in C2C12 cells. A, total RNA of C2C12 cells induced to differentiate by serum withdrawal for up to 7 days was isolated, and 10 g of each sample were separated on agarose gels. Day 0 indicates RNA from proliferating cells growing in 10% fetal calf serum. Northern hybridization was performed with an in vitro transcribed, digoxigenin-labeled Notch1-specific fragment. Notch1-specific signals at ϳ8.0 kb were visualized by using the nonradioactive DIG System for Nucleic Acid Labeling and Detection. Equal loading was assessed by ethidium-bromide staining of the agarose gel (lower panel). B, equal amounts of whole cell lysates from C2C12 cells induced to differentiate by growth factor withdrawal for the indicated time points were separated on an 8% denaturing SDS-PAGE gel and blotted onto a polyvinylidene difluoride membrane. N TM -specific signals (120 kDa) were detected with a goat anti-Notch1-specific polyclonal antibody (gt ␣-Notch1) and a donkey anti-goat IgG horseradish peroxidase-coupled secondary antibody. Signals were finally visualized using the ECL system and autoradiography. Day 0 indicates protein lysates from proliferating cells growing in 10% fetal calf serum. The signal for the 120-kDa specific N TM protein is indicated. WB, Western blot.  1-4). Notch1-specific signals were visualized by immunoblotting as described in Fig. 1B. Lines indicate the 120-kDa membrane-bound N TM fragment as well as the activated 110-kDa NIC fragment (lanes 5 and 7 only).

FIG. 3. Immunoprecipitation of N TM .
A, cellular lysates of C2C12 cells induced to differentiate for 2 days in low serum were prepared and subjected to immunoprecipitation (IP) analysis with either a rabbit anti-Notch1 polyclonal serum (rb ␣-N, lane 1) or a rabbit anti-ubiquitinspecific polyclonal antibody (rb ␣-Ubi, lane 2). Immunoprecipitates were separated on 8% SDS-PAGE gels and blotted. Notch1-specific signals were detected as described in Fig. 1B. Lines indicate the 120-kDa N TM fragment. The asterisk indicates a band corresponding to the 300 -330-kDa precursor form of Notch1. B, to verify the specificity of Notch1-specific immunoprecipitations, whole cellular lysates were prepared from C2C12 cells that had been induced to differentiate for 2 days and subjected to immunoprecipitation experiments. Either lysis buffer alone (lane 1) or the indicated cellular lysate (lane 4) was precipitated with the Notch1-specific antibody. In addition, the extract was either incubated with an isotype control antibody (rabbit anti-IgG) alone (lane 2) or with the isotype control antibody prior to the Notch1 antibody (lane 3). Lane 5 shows an immunoprecipitation analysis of the indicated lysate with the Notch1 antibody competing with its specific blocking peptide. Immunoprecipitates were separated on 8% SDS-PAGE gels, and Notch1-specific signals were detected as described. WB, Western blot. of 150 kDa, whereas a mono-ubiquitinated Notch1-GST is only slightly larger, ϳ158 kDa. As shown in Fig. 4, the specific conjugation of HA-tagged ubiquitin to the GST-N IC could be verified in this experimental system. Specific signals for the GST-tagged Notch1 protein were detected only in COS1 cells transfected with GST-N IC alone or together with the HA-ubiquitin expression vector. Most importantly, as shown in Fig. 4A, lane 8, the GST-N IC corresponding band also appears when lysates from cotransfected cells were precipitated with the HA tag antibody and detected with an anti-GST antibody, indicating the conjugation of HA-ubiquitin to GST-N IC . Using a Notch1 antibody for precipitation of cotransfected COS1 cells, the positive reaction of the HA-tag specific antibody in the subsequent Western blot analysis confirms this finding (Fig.  4B, lane 7). The single band at ϳ150 kDa corresponding to GST-N IC in Fig. 4A, lane 8, as well as in Fig. 4B, lane 7, may indicate covalent coupling of a mono-ubiquitin subunit to a significant portion of the intracellular Notch1 domain.
N TM Fragments Are Targeted to a Chloroquine-sensitive Degradation Pathway-It was reported previously (18) that polyubiquitin chains, consisting of at least four ubiquitin polypeptides, target proteins for specific degradation by the proteasome. However, mono-ubiquitination was described for different transmembrane receptor proteins and discussed as a regulatory mechanism for the localization and activity of proteins (14,19,20). In particular, it is believed that mono-ubiquitination serves as a signal to trigger the internalization of membrane-bound proteins and to regulate the activity of components of the endocytotic/lysosomal machinery (reviewed in Ref. 21). To test our hypothesis that N TM is targeted either to the proteosomal or to the endosomal/lysosomal pathway of protein degradation in the course of terminal differentiation of myoblasts, confluent C2C12 cells were incubated in fusion medium for 2 days to enhance endogenous Notch1 synthesis. The cells were subsequently treated for various hours with different proteasomal inhibitors. Neither treatment with the highly specific proteasome inhibitor clasto-lactacystin ␤-lactone (Fig. 5A) nor with the proteasome inhibitor-1 nor with MG132 (data not shown) revealed significant accumulation of overall or USABprecipitable N TM species. Remarkably, the inhibitor concentrations used were sufficient to induce enrichment of ubiquitinated protein species in general, as shown in Western blot analysis (data not shown). In addition, no accumulation of N TM fragments was detected during longer incubation times with these inhibitors (data not shown).
To investigate whether Notch1 might be targeted to the lysosomal/endosomal machinery of protein degradation, analogous experiments were performed, incubating C2C12 cells with the lysosomal inhibitor chloroquine for various time points (Fig. 5B). Noticeably, whereas the overall N TM levels appear not to change during incubation with chloroquine (Fig. 5B,  lanes 4 -6), a USAB-precipitable N TM subfraction appears to be strongly enriched within 6 h of chloroquine treatment (Fig. 5B,  lanes 10 -12), indicating that N TM is posttranslationally regulated by a chloroquine-sensitive pathway. Control incubation of C2C12 cells with PBS resulted in unchanged levels of N TM protein (Fig. 5B, lanes 1-3) and no enrichment of USAB-pre-

FIG. 5. Treatment of C2C12 cells with proteasomal and lysosomal inhibitors.
C2C12 cells were induced to differentiate by serum withdrawal for 2 days and were subsequently treated for 1, 3, and 6 h with the respective inhibitors added to the medium. Cellular lysates were prepared and subjected to immunoprecipitation (IP) analysis with a Notch1 (rb ␣-Notch1) or a ubiquitin (rb ␣-Ubi)-specific antibody. Immunoprecipitated proteins were resolved on denaturing 8% SDS-PAGE gels, and signals were detected with a goat anti-Notch1 antibody (gt ␣-Notch1). A, cells were incubated with the proteasomal inhibitor clasto-lactacystin ␤-lactone (c-Lacta ␤-L., lanes 4 -6) or with an equivalent volume of the solvent Me 2 SO (DMSO) for control purposes (lanes  1-3). B, C2C12 cells were treated for 1, 3, and 6 h with either the inhibitor chloroquine (lanes 4 -6 and 7-9) or with an equivalent volume of PBS as control (lanes 1-3). Note the difference in exposure time of ECL-treated filters: 20 min (A) and 1 min (B). cipitable protein. These data indicate that N TM may be regulated posttranslationally by a chloroquine-sensitive pathway leading to lysosomal degradation rather than being targeted to the proteasomal degradation machinery.
c-Cbl Coprecipitates with N TM in C2C12 Myoblasts-It was reported previously (22) that various transmembrane receptor proteins are internalized by a c-Cbl-mediated receptor sorting which involves covalent attachment of ubiquitin, tyrosine phosphorylation events, and subsequent lysosomal degradation, thereby controlling the fate and signaling capacity of growth factor receptors. Detailed analysis of the protein motifs within the intracellular portion of the Notch1 protein revealed the presence of two potential c-Cbl-docking sites at the C terminus of the Notch1 protein, close to the PEST domain, mediating protein stability and turnover rates of proteins. The consensus sequences for c-Cbl-docking sites were reported to span the amino acid sequence motif Tyr-Xaa-Xaa-Xaa-Pro (23). As shown in Fig. 6, the last 200 amino acids of the intracellular portion of Notch1 contain two potential c-Cbl-docking sites with the amino acid sequences YQGLP and YSSSP. Additionally, using the NetPhos 2.0 prediction program, the second potential c-Cbl-docking site (YSSSP) is predicted to become specifically tyrosine-phosphorylated, a prerequisite for binding of c-Cbl to its substrate (24). Based on these findings, we asked whether c-Cbl might be an integral component of a protein complex containing Notch1 in myogenic cells. This could provide an explanation not only for ubiquitination events but also for the targeting of membrane-associated Notch1 to the lysosomal degradation machinery.
We started to determine the relative amounts of c-Cbl protein during terminal differentiation of C2C12 cells after a shift to low serum levels for a period of 7 days. As shown in Fig. 7A, the c-Cbl protein is constitutively synthesized during this period. Immunoprecipitations with a Notch1-specific antibody and subsequent Western blot analysis revealed a specific interaction between endogenous Notch1 and c-Cbl (Fig. 7B). In this context, it is important to note that c-Cbl and N TM have a nearly identical molecular mass of 120 kDa.
Treatment of C2C12 cells with the lysosomal inhibitor chloroquine for 1, 3, and 6 h after the induction of terminal differentiation for 48 h showed increasing amounts of c-Cbl protein physically associated with N TM after stimulation with chloroquine (Fig. 8). This interaction strengthens the evidence that the cell membrane-associated Notch1 (N TM ) can be targeted to the endosomal pathway of protein degradation, possibly mediated by the ubiquitin-ligating enzyme c-Cbl and the attachment of ubiquitin residues.
Notch1 Is Tyrosine-phosphorylated and Accumulates Upon Chloroquine Treatment of C2C12 Cells-As described previously (25), receptor proteins, which are targeted to the endosomal protein degradation machinery, are often tyrosine-phos-phorylated after activation by ligand binding. Furthermore, tyrosine phosphorylation is also a prerequisite for c-Cbl binding to its substrate (24,25). To investigate whether Notch1 is  7. c-Cbl protein synthesis and association with Notch1 in C2C12 cells. A, whole cell lysates of confluent C2C12 cells induced to differentiate by growth factor withdrawal were prepared and separated on 8% denaturing SDS-PAGE gels. c-Cbl-specific signals were detected with a rabbit anti-c-Cbl-specific polyclonal antibody (rb ␣-Cbl) and a goat anti-rabbit IgG horseradish peroxidase-coupled secondary antibody. Day 0 indicates protein lysates from proliferating cells growing in 10% fetal calf serum. The line points to the 120-kDa c-Cbl-specific signal. Based on the homology features shared by all members of the Cbl protein family, it appears likely that the visible 90-kDa fragment results from a cross-reactivity of the c-Cbl antibody either with a Cbl isoform or with another member of the Cbl protein family, possibly CARP90 (38). B, cellular lysates of C2C12 cells induced to differentiate for 48 h in low serum as described were subjected to immunoprecipitation analysis with a rabbit anti-Notch1 polyclonal serum. Immunoprecipitates were separated on 8% SDS-PAGE gels, and c-Cbl-specific signals were detected using a rabbit anti-c-Cbl-specific antibody and a goat anti-rabbit IgG-specific antibody coupled to horseradish peroxidase. A c-Cbl-specific signal was detected in the anti-Notch1-specific immunoprecipitate, indicating that c-Cbl interacts specifically with the N TM fragment of Notch1 (lane 1). As a control, whole cellular lysate was loaded in parallel (lane 2). WB, Western blot.  1-3 and 7-9). Cellular lysates were subjected to immunoprecipitation either with a Notch1-specific rabbit antibody (rb ␣-N) or with a rabbit anti-c-Cbl antibody (rb ␣-Cbl). Immunoprecipitated (IP) proteins were resolved on denaturing 8% SDS-PAGE gels, and specific signals were visualized on membranes with either rabbit ␣-Cbl or with a goat anti-Notch1 polyclonal antibody (gt ␣-N), as indicated. WB, Western blot. also tyrosine-phosphorylated under these conditions, we applied a phosphotyrosine-specific polyclonal antibody (Fig. 9A). Only after treatment of terminally differentiating C2C12 cells with the endosomal inhibitor chloroquine an accumulation of tyrosine-phosphorylated N TM -protein can be observed (Fig. 9A,  lanes 4 -6), whereas in lysates of PBS-stimulated C2C12 cells no tyrosine-phosphorylated N TM protein is evident (Fig. 9A,  lanes 1-3). Confirming results were obtained in a reciprocal analysis when lysates of chloroquine-treated C2C12 cells were immunoprecipitated with Notch1-specific antibodies and analyzed in Western blots using the monoclonal phosphotyrosinespecific antibody (data not shown).
In addition, we analyzed whether c-Cbl is tyrosine-phosphorylated also and whether the levels of phosphorylated c-Cbl protein are regulated after treatment with the lysosomal inhibitor chloroquine. As shown in Fig. 9B, c-Cbl is tyrosinephosphorylated, but the levels remain unchanged after treatment with either PBS (Fig. 9B, lanes 1-3) or chloroquine (Fig.  9B, lanes 4 -6). From these data, it is conceivable that the N TM fragment becomes specifically tyrosine-phosphorylated. This feature might be recognized by the ubiquitin-protein ligase c-Cbl leading to ubiquitination and subsequent targeting of Notch1 to the endosomal/lysosomal protein degradation machinery. This may contribute to the termination of Notch1mediated signal transduction in the course of terminal differentiation of myoblast cells.
N TM Is Present in Subcellular Fractions Enriched for Lysosomes in C2C12 Cells-To address the question whether USAB-precipitable N TM is targeted to the endosomal/lysosomal compartment of C2C12 cells, we employed differential centrifugation in order to obtain subcellular fractions, enriched for lysosomes and containing endosomes. The N TM fragment is present in the 1000 ϫ g fraction containing nuclei as well as in the 17,000 ϫ g fraction containing the majority of lysosome organelles (Fig. 10). For control reasons we also analyzed these lysates for the presence of the ubiquitin-ligating protein c-Cbl. c-Cbl was detectable in the analyzed extracts, although the level of c-Cbl in the 17,000 ϫ g pellet was markedly reduced as compared with the 1000 ϫ g pellet (Fig. 10). To verify the relative purity of the analyzed fractions, Western blot analysis was also performed with monoclonal antibodies specific for the endosomal marker protein EEA-1 and the nuclear protein nucleoporin (Fig. 10). This analysis provides further evidence for the targeting of Notch1 to the lysosomal compartment during the process of terminal differentiation of the myoblast cell line C2C12. DISCUSSION Ubiquitination and subsequent degradation of cellular proteins, especially of receptors and transcription factors, serves as a major mechanism to regulate their activity. Several lines of evidence from the literature have suggested a connection between Notch and the ubiquitin/proteasome-dependent degradation pathway. In Drosophila melanogaster, the E3 class ubiquitin-protein ligase Suppressor of deltex has been shown to FIG. 9. Accumulation of tyrosine-phosphorylated N TM after treatment of C2C12 cells with the lysosomal inhibitor chloroquine. Confluent C2C12 cells were induced to differentiate by serum withdrawal for 2 days and were subsequently treated for 1, 3, or 6 h with either the inhibitor chloroquine (100 M) or an equivalent volume of PBS. A, cellular lysates were subjected to immunoprecipitation (IP) analysis with a phosphotyrosine-specific mouse monoclonal antibody (m ␣-P-Tyr). Immunoprecipitated proteins were resolved on denaturing 8% SDS-PAGE gels, and blots were developed with a goat anti-Notch1specific polyclonal antibody (gt ␣-N). The left panel represents controltreated cells (lanes 1-3), and the right panel (lanes 4 -6) shows chloroquine-treated cells. Lines indicate the 120-kDa N TM fragment. The asterisk indicates a band corresponding to the 300 -330-kDa precursor form of Notch1. B, cellular lysates were subjected to immunoprecipitation analysis with a phosphotyrosine-specific mouse monoclonal antibody (m ␣-P-Tyr). Immunoprecipitated proteins were resolved on denaturing 8% SDS-PAGE gels, and blots were developed with a rabbit anti c-Cbl antibody (rb ␣-c-Cbl). days in low serum media were prepared by differential centrifugation as described under "Experimental Procedures," separated on denaturing 8% SDS-PAGE gels, and blotted. The levels of the N TM protein (120 kDa) and the c-Cbl protein (120 kDa) were determined by using Notch1-(gt ␣-Notch1) and c-Cbl (rb ␣-Cbl)-specific antibodies in Western blot analysis. To verify the specificity of the 1000 and 17,000 ϫ g fractions, mouse monoclonal antibodies directed against the Early Endosome Associated protein (EEA-1, 180 kDa) and against nucleoporin (62 kDa) were employed. interact specifically with Notch and is capable of suppressing its activity (10). Furthermore, dominant-negative mutations in different proteasome subunits do affect cell fate decisions in Drosophila with an enhancement of Notch signaling activity (11). Finally, Itch, a mouse ortholog of Suppressor of deltex interacts specifically with Notch1 in mice (12). In addition, Gupta-Rossi et al. (26) demonstrated very recently that the mammalian ubiquitin-ligase SEL-10 could bind to an activated form of Notch1 protein, overexpressed in 293T and HeLa cells. Nevertheless, these studies rather suggest a role for ubiquitination and proteasomal degradation of activated Notch, possibly taking place in the nucleus.
To address the biological relevance of ubiquitination for the stability of the transmembrane Notch1 protein during terminal differentiation processes in myogenesis, we have employed the myoblastic cell line C2C12 that allows the analysis of Notchdependent differentiation leading from mononucleated precursor cells to multinucleated, terminally differentiated myotubes. We were interested whether the endogenous Notch1 receptor protein is regulated at the cellular periphery before being activated and translocated into the nucleus.
Recently, various experimental settings have demonstrated that ubiquitination of certain transmembrane receptors, particularly of receptor tyrosine kinases, serves as a signal for their targeting to the endosomal/lysosomal compartment where they may undergo proteolytic degradation (22,(27)(28)(29). Here, ubiquitination is not only discussed in the context of protein degradation but also in connection with the internalization and trafficking of integral receptors where mono-ubiquitination itself can serve as an internalization signal (15,20). Our results of the transient cotransfection experiments in Fig.  4 indicate such mono-ubiquitination of the ectopically expressed and GST-tagged version of the Notch1 protein. In this context, it is noteworthy that internalization of transmembrane receptors may not necessarily require ubiquitination of the receptor itself but may also depend on associated proteins of the ubiquitination machinery, which mediate the critical signal for receptor internalization and degradation (30).
To distinguish between lysosomal and proteasomal pathways of degradation, we were using either the lysosomal inhibitor chloroquine, which blocks endosomal protein degradation by inhibiting the acidification of the lysosomes or various proteasome inhibitors. The latter included the highly specific proteasome inhibitor clasto-lactacystin ␤-lactone. Incubation with proteasome inhibitors did not result in significant changes of N TM protein levels (Fig. 5A). Only treatment of C2C12 cells with the lysosomal inhibitor chloroquine leads to a significant accumulation of ubiquitin-associated N TM protein (Fig. 5B,  lanes 10 -12). These findings provide strong evidence that N TM fragments are targeted preferentially to the lysosomal/endosomal machinery of protein degradation rather than to the 26 S proteasome-mediated proteolytic pathway. In our hands and as described previously (31)(32)(33), proteasomal inhibitors as well as chloroquine are capable of blocking differentiation and fusion of myogenic cells or increasing cell death rates after prolonged incubation. Nevertheless, short term exposure to these substances has been proven to be sufficient in many cellular systems, leading to the accumulation of otherwise degraded proteins (31)(32)(33). 3 The proto-oncogenic ubiquitin-ligase c-Cbl (34) plays a pivotal role in the context of ubiquitination and targeting of receptors to the endosomal/lysosomal pathway (22,29). The 120-kDa c-Cbl polypeptide chain contains a phosphotyrosine recognition domain, an SH2 motif, and a RING finger ubiquitin ligase domain that recruits E2 ubiquitin-conjugating enzymes (35). c-Cbl acts as a negative regulator for several growth factor receptors, including the receptors for the epidermal growth factor, the platelet-derived growth factor, and the colony-stimulating factor-1 (22,29,36). It was reported that c-Cbl plays a key role for their ubiquitination and degradation in the course of endosomal/lysosomal sorting, thereby representing a key mechanism for the removal of these receptors from the cell surface. Consequently, this may reduce the potential of receptor-mediated signaling within the cell. The notion of c-Cbl-dependent modulation of Notch1 function is strongly supported by the presence of two potential docking sites for c-Cbl (23,24) within the C-terminal part of the intracellular portion of the Notch1 protein. Furthermore, we have shown that c-Cbl specifically interacts with Notch1, and we propose that this interaction is mediated via binding of c-Cbl to a tyrosine-phosphorylated form of the N TM protein. Interaction of c-Cbl with target receptors apparently occurs in response to their tyrosine phosphorylation and requires phosphorylation of c-Cbl itself (25). The recruitment of c-Cbl by phosphorylated N TM may therefore be a key factor for the down-regulation of the receptor and its targeting to the endosomal/lysosomal protein degradation machinery.
It is apparent from Fig. 9A that the full-length prepro-form of Notch1 appears to be constitutively tyrosine-phosphorylated. This may indicate that tyrosine-specific phosphorylation events take place very early after the synthesis of the fulllength precursor Notch1 protein, even before the receptor is present on the cell surface. It is tempting to speculate that phosphorylation patterns of the full-length Notch1 precursor protein might be required for specific proteolytic processing events resulting in the formation of an active Notch1 receptor complex consisting of N-terminal extracellular fragment and N TM and the formation of the activated intracellular form of Notch1 (N IC ). Alternatively, phosphorylation may be a requirement for specific intracellular trafficking of the Notch precursor protein. Further analysis of this particular observation will be addressed in our ongoing research.
Based on these data we postulate a novel mechanism for the regulation of Notch1 activity during terminal differentiation of myoblast cells. Lysosomal degradation appears to be an important regulatory mechanism to ensure a timely limited activity of the developmental essential protein Notch1. Dysregulation of Notch1 activity may result in formation of tumors as it was described in the hematopoietic cell lineage where a translocation of the intracellular portion of Notch1 leads to constitutive activity of the protein and to the formation of acute lymphoblastic T cell leukemias (16).
Removal of Notch1 from the cell surface by endocytosis may allow cells to regulate their predisposition to respond to Notch ligands present in their vicinity. Endocytosis of Notch1 may regulate both the extent of a receptor-mediated response and the specificity of the response. For example, it was described that a short duration of signaling of receptor tyrosine kinases in the cell line PC12 promotes proliferation, whereas a longer duration of the signal results in differentiation (37). Thus, changing the signaling kinetics and the signaling magnitude of receptor proteins might be an important parameter in the regulation of cellular responses. In this regard, it is of special interest whether the regulation of Notch1 observed in C2C12 myoblasts represents a cell type-specific feature or whether cessation of Notch1 function via ubiquitination and endosomal degradation is also detectable in other systems. Future research should be addressed to characterize this system in a broader range of cells and tissues.