Content uploaded by Benjamin Geiger
Author content
All content in this area was uploaded by Benjamin Geiger on Sep 28, 2015
Content may be subject to copyright.
c-Cbl/Sli-1 regulates endocytic sorting
and ubiquitination of the epidermal
growth factor receptor
Gil Levkowitz,
1
Hadassa Waterman,
1,5
Eli Zamir,
2,5
Zvi Kam,
2
Shlomo Oved,
1
Wallace Y. Langdon,
3
Laura Beguinot,
4
Benjamin Geiger,
2
and Yosef Yarden
1,6
Departments of
1
Biological Regulation and
2
Molecular Cell Biology, The Weizmann Institute of Science, Rehovot 76100,
Israel;
3
Department of Pathology, The University of Western Australia, Queen Elizabeth II Medical Center, Nedlands,
Western Australia 6907, Australia;
4
Molecular Oncology Unit, DIBIT, and Instituto di Neuroscienze e Biommagini del CNR,
H.S. Raffaele, Milan 20132, Italy
Ligand-induced down-regulation of two growth factor receptors, EGF receptor (ErbB-1) and ErbB-3, correlates
with differential ability to recruit c-Cbl, whose invertebrate orthologs are negative regulators of ErbB. We
report that ligand-induced degradation of internalized ErbB-1, but not ErbB-3, is mediated by transient
mobilization of a minor fraction of c-Cbl into ErbB-1-containing endosomes. This recruitment depends on the
receptor’s tyrosine kinase activity and an intact carboxy-terminal region. The alternative fate is recycling of
internalized ErbBs to the cell surface. Cbl-mediated receptor sorting involves covalent attachment of ubiquitin
molecules, and subsequent lysosomal and proteasomal degradation. The oncogenic viral form of Cbl inhibits
down-regulation by shunting endocytosed receptors to the recycling pathway. These results reveal an
endosomal sorting machinery capable of controlling the fate, and, hence, signaling potency, of growth factor
receptors.
[Key Words: Endocytosis; ErbB/HER; protein degradation; signal transduction; tyrosine kinase]
Received June 15, 1998; revised version accepted October 6, 1998.
The ErbB family of growth factor receptors carrying an
intrinsic tyrosine kinase activity constitutes a layered
signaling network that includes a large family of EGF-
like ligands that bind to four transmembrane receptors
capable of forming ten homo- and heterodimeric combi-
nations (for review, see Alroy and Yarden 1997). Whereas
the EGF receptor (ErbB-1) binds several growth factors
whose prototype is EGF, both ErbB-3 and ErbB-4 bind all
isoforms of the Neu differentiation factor (NDF). The
third layer of the network includes a large group of sig-
naling molecules sharing one of several types of phos-
photyrosine-binding (PTB) domains (e.g., an SH2 do-
main). Although nonidentical sets of signaling proteins
are recruited, all ErbB receptors, like their invertebrate
orthologs (Perrimon and Perkins 1997), funnel their sig-
nals into the mitogen-activated protein kinase (MAPK)
pathway. The kinetics of MAPK activation and its rela-
tive potency, however, display remarkable differences
that correlate with the rate of ligand-induced endocyto-
sis of receptors, termed down-regulation (Pinkas-Kra-
marski et al. 1996), and recruitment of c-Cbl by the ac-
tivated receptor (Levkowitz et al. 1996).
c-Cbl isthemammalian ortholog of the Sli-1 protein of
Caenorhabditis elegans (Yoon et al. 1995). Genetic evi-
dence indicated that Sli-1 negatively regulates signaling
downstream of the single nematode ErbB protein
(Jongeward et al. 1995). c-Cbl is a major cellular substrate
of tyrosine phosphorylation: It undergoes increased
phosphorylation in response to ligand-induced stimula-
tion of a variety of surface receptors, including the EGF
receptor, lymphokine receptors, immunoglobulin recep-
tors, antigen receptors, and integrin receptors (Thien and
Langdon 1997, and references therein). Overexpression
of c-Cbl attenuates signaling down-stream of the immu-
noglobulin E receptor (Ota and Samelson 1997) and the
T-cell receptor (Boussiotis et al. 1997), yet the mecha-
nism of Cbl action remains unknown. Here we report
that c-Cbl can increase the rate of degradation of ErbB-1,
but not ErbB-3. The underlying mechanism involves
transient physical associations between c-Cbl and
ErbB-1 in endosomes, and subsequent ubiquitination of
the degradation-destined receptors. An oncogenic viral
Cbl appears to interfere with the sorting function of c-
Cbl, thereby directing incoming receptors to the recy-
cling pathway.
5
These authors contributed equally to this work.
6
Corresponding author.
E-MAIL liyarden@weizmann.weizmann.ac.il; FAX 972-8-9344116.
GENES & DEVELOPMENT 12:3663–3674 © 1998 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/98 $5.00; www.genesdev.org 3663
Results
c-Cbl mediates selective degradation
of ligand-stimulated ErbB-1, but not ErbB-3
To investigate the possibility that EGF-driven ErbB-1 is
destined to lysosomal degradation because ErbB-1 can
interact with c-Cbl (Levkowitz et al. 1996), whereas
ErbB-3 is shunted tothe recycling pathway (Waterman et
al. 1998) because it cannot recruit c-Cbl, we transiently
overexpressed c-cbl and erbB-1 in Chinese hamster
ovary (CHO) cells. In addition to c-Cbl, we used two
deletion mutants that are schematically presented in
Figure 1A. These are a peptide-tagged amino-terminal
portion of c-Cbl analogous to the murine viral form, v-
Cbl, and the complementary deletion mutant, Cbl-C
(Fig. 1A). Cells were briefly stimulated with EGF and
tyrosine phosphorylation of the receptor analyzed by im-
munoblotting with anti-phosphotyrosine antibodies.
The results of this experiment indicated that ErbB-1 un-
derwent enhanced tyrosine phosphorylation in the pres-
ence of its ligand (∼10-fold), but c-Cbl overexpression
almost abolished this effect (Fig. 1B). Neither v-Cbl nor
Cbl-C were active, implying that the combination of
amino- and carboxy-terminal sequences is essential for
the effect of c-Cbl on receptor phosphorylation.
Reblotting with anti-ErbB-1 antibodies revealed that
c-Cbl overexpression led to degradation, rather than
catalytic inactivation, of ErbB-1 (Fig. 1B, bottom): prima-
rily the hyperphosphorylated form ofErbB-1, whose elec-
trophoretic mobility is retarded, was diminished (by
80%). This observation suggested that tyrosine phos-
phorylation of the surface-localized ErbB-1 molecules se-
lectively targets ErbB-1 to degradation through c-Cbl ac-
tion. Labeling of cell surface-exposed ErbB-1 molecules
by surface biotinylation confirmed this scenario (Fig.
1C): Overexpression of c-Cbl enhanced disappearance of
the hyperphosphorylated form of ErbB-1. Concomitant
with selective removal from the cell surface, signaling
downstream of ErbB-1 was down-regulated by c-Cbl.
This was exemplified by the ability of the ligand-stimu-
lated ErbB-1 to recruit one of its major substrates, the
Shc protein (Fig. 1C, bottom). In contrast with the EGF
receptor, the NDF-receptor (ErbB-3) was not affected by
c-Cbl (Fig. 1B). In conclusion, c-Cbl can down-regulate
ErbB-1 signaling, apparently by selective degradation of a
hyperphosphorylated form of the receptor.
Additional support for this model was provided by
analyses of ligand-induced redistribution of c-Cbl. Con-
sistent with the observation that EGF can induce physi-
cal association of its receptor with c-Cbl in living cells
(Bowtell and Langdon 1995; Galisteo et al. 1995; Meisner
et al. 1995; Tanaka et al. 1995), EGF treatment of cells
co-overexpressing ErbB-1 and ErbB-3 led to rapid changes
in the pattern of c-Cbl subcellular localization (Fig. 2A).
Whereas in untreated cells c-Cbl exhibited reticular/ve-
sicular distribution, it assumed a more punctate, vesicu-
lar-like, pattern in EGF-treated cells. In fact, some,
but not all, c-Cbl-bearing structures contained endocy-
tosed molecules of ErbB-1. No such redistribution of c-
Cbl, or ErbB-3, was noted on treatment with NDF (Fig.
Figure 1. c-Cbl, but not v-Cbl, in-
creases degradation of ErbB-1 by its re-
moval from the cell surface. (A) Shown
are the domain structures of c-Cbl and
three derivative proteins. The following
structural motifs are represented: A 7
residue-long histidine stretch (7 His), a
positively charged basic domain (Basic),
a ring-finger domain (RF), a proline-rich
domain (Pro-Rich), and a leucine-zipper
(LZ). An influenza virus hemagglutinin
(HA) epitope tag was added to the
amino-terminal end of each protein.
cDNAs corresponding to the three natu-
ral forms of Cbl were transiently trans-
fected into CHO cells. Forty-eight hours
after transfection, cells were lysed and
whole cell lysates subjected to immuno-
blotting (IB) with an anti-HA antibody.
(B)ErbB-1(left)orErbB-3(right) were
transiently expressed in CHO cells by
cotransfection with plasmids encoding
the indicated Cbl proteins, or with a
control empty plasmid (Cont.). Cells were incubated for 45 min at 37°C with EGF or NDF (each at 100 ng/ml). Thereafter, whole cell
lysates were subjected to immunoprecipitation (IP) and immunoblotting (IB) with the indicated antibodies. (C) CHO cells were
co-transfected with an ErbB-1 expression vector together with a vector directing expression of c-Cbl or a control empty vector (Cont.)
Cell surface-exposed proteins were covalently labeled with biotin at 4°C. Soluble biotin was then removed and cells incubated at 37°C
for 45 min in the presence or absence of EGF. Cell lysates were subjected to immunoprecipitation (IP) with an anti ErbB-1 antibody
and the electrophoretically resolved proteins were probed either with horseradish peroxidase-conjugated Streptavidin, or with an
anti-Shc antibody. The locations of molecular mass markers are indicated in kilodaltons. Note that the three Shc isoforms (p66, p52,
and p46) associate with ErbB-1.
Levkowitz et al.
3664 GENES & DEVELOPMENT
2A). This observation led us to compare the effects of
c-Cbl with the abilities of the two ligands to down-regu-
late their binding sites. To this end, we stably expressed
c-Cbl, or v-Cbl, in cells already overexpressing ErbB-1.
Although the introduced Cbl proteins were detectably
expressed (Fig. 2B, bottom), c-Cbl- overexpressing cells
displayed a significantly lower amount of the ErbB-1 pro-
tein (Fig. 2B, left) and a six-fold decreased number of
EGF-binding sites relative to control- or v-cbl-trans-
fected cells (Fig. 2B, right). This basal down-regulation of
ErbB-1 was displayed by several independently selected
clones, and severely reduced the sensitivity of receptor
down-regulation assays. Therefore, we used transient co-
expression of a receptor and c-Cbl. Because of a high
level of receptor expression, EGF treatment of ErbB-1-
expressing cells led to only a moderate down-regulation
effect (Fig. 2C). The extent of ligand-induced down-regu-
lation displayed variation that apparently correlates with
receptor expression, but in all experiments receptor dis-
appearance was significantly enhanced on overexpres-
sion of c-Cbl (Fig. 2C). This observation indicates that
the amount of c-Cbl limits down-regulation when
ErbB-1 is overexpressed. In contrast, NDF treatment of
ErbB-3-expressing cells did not result in receptor down-
regulation, and c-Cbl exerted no significant effect on the
status of NDF-binding sites (Fig. 2C). In conclusion, c-
Cbl can enhance down-regulation of the EGF-receptor,
but not the NDF receptor, and this involves redistribu-
tion of c-Cbl into endocytic ErbB-1-containing vesicles.
Transient ligand-induced colocalization of c-Cbl
and ErbB-1 in endosomes
Because partial colocalization of ErbB-1 and c-Cbl was
observed on a 20-min-long incubation of cells with EGF
(Fig. 2A), we analyzed the kinetics of the phenomenon.
The results presented in Figure 3A demonstrate redistri-
bution of ErbB-1 and c-Cbl from their initial sites (mem-
branal and reticular, respectively), and partial colocaliza-
tion that peaks at 10–20 min. Concomitant with redis-
tribution, the double-stained vesicular structures
exhibited an apparent increase in fluorescence intensity.
Figure 2. c-Cbl colocalizes with ErbB-1, but not with ErbB-3, and accelerates its down-regulation. (A) CHO cells that stably co-
overexpress ErbB-1 and ErbB-3 (Tzahar et al. 1996) were plated on cover-slips and treated at 37°C with EGF (top)orNDF(bottom)for
the indicated periods of time. After fixation and permeabilization, cover slips were simultaneously stained with a polyclonal anti-Cbl
antibody and a mAb to ErbB-1 or to ErbB-3, as indicated. Antibody detection by immunofluorescence was performed as described
(Materials and Methods). (B) Stable CHO transfections expressing c-Cbl or v-Cbl were established by cotransfection of Cbl constructs
together with the pBABE/puro plasmid into CHO cells overexpressing ErbB-1 (Tzahar et al. 1996). Individual clones were screened by
immunoblotting with antibodies to ErbB-1 and to the HA tag of Cbl proteins (left, the locations of c-Cbl and v-Cbl protein bands are
indicated by arrows). To assess cell surface expression of ErbB-1, clones expressing c-Cbl or v-Cbl, as well as the parental ErbB-1
overexpressing cell line (−), were incubated for 90 min at 4°C with radiolabeled EGF (10 ng/ml). Cell-bound radioactivity is shown as
the average and range (bars) of duplicate determinations. (C) CHO cells were cotransfected with pairs of two plasmids: an ErbB-
expression vector and either a control empty pcDNA3 plasmid (䊊; ErbB-1; 䉭, ErbB-3), or a c-Cbl expression vector (䊉, ErbB-1; 䉱,
ErbB-3). Cell monolayers were subjected to a down-regulation assay 48 hr post-transfection. The results are expressed as the average
fraction (and range, bars) of original binding sites that remained on the cell surface after exposure to the nonlabeled ligand at 37°C.
Endosome sorting by Cbl
GENES & DEVELOPMENT 3665
This process was further analyzed by use of computer-
ized analysis of digital images. First, we defined the bor-
ders of each c-Cbl- containing particle, and then sepa-
rately determined the fluorescence intensity of c-Cbl and
ErbB-1 within the framed two-dimensional structure.
When this was repeatedly performed at different time
intervals, we obtained the results presented in Figure 3B,
in which each dot represents one vesicular structure. In
untreated cells, labeling for c-Cbl in individual particles
was very low, but the cells contained relatively large
numbers of small particles. Following 10 min of incuba-
tion, a coordinated increase in ErbB-1 and c-Cbl labeling
became evident and it displayed a linear correlation co-
efficient of 0.83. Colocalization was maximal after 20
min of incubation (r = 0.87) and then gradually declined.
Analysis of the total number of intracellular vesicles/
particles per cell (data not shown), as well as the extent
of colocalization (Fig. 3B, right) revealed a 25% decrease
in the average density of c-Cbl-positive particles at 10
and 20 min. The number of ErbB-1 vesicles gradually
increased and reached a maximal level at 10 min. Impor-
tantly, the maximal number of ErbB-1-positive vesicles
was much lower than the number of c-Cbl-positive
vesicles, suggesting that only a small subpopulation of
c-Cbl-rich vesicles contained endocytosed ErbB-1 mol-
ecules. This was confirmed by direct evaluation of the
extent of colocalization: Approximately 15% of c-Cbl-
positive vesicles also contained ErbB-1, implying that
nearly all of the ErbB-1-containing vesicles became as-
sociated with c-Cbl.
Biochemical analyses that involved labeling of surface
ErbB-1 molecules with biotin, or selective immunopre-
cipitation of the membranal EGF receptors, indicated
that only the endocytosed receptor fraction physically
recruited c-Cbl (data not shown), consistent with the en-
dosomal site of interaction implied by the immunofluo-
rescence results. Indeed, separation of the endosomal
fraction from CHO cells by use of the method described
previously (Wada et al. 1992) and Rab-5 as an endosome
marker revealed that c-Cbl is a resident protein of the
endosome (Fig. 3C). In addition, physical interaction be-
tween c-Cbl and ErbB-1 was detectable in the endosomal
fraction only if cells were treated with EGF prior to lysis
and subcellular fractionation. c-Cbl and ErbB-1 redistri-
Figure 3. Time dependence of EGF-induced ErbB-1 and c-Cbl co-localization in endosomes. (A) CHO cells that coexpress ErbB-1 and
ErbB-3 were either fixed (0 min) or first incubated with EGF for 5 min at 37°C. Thereafter, EGF was removed and incubation continued
for the indicated time intervals. Double staining of ErbB-1 and c-Cbl was performed and visualized as described in Materials and
Methods. (B, left) For each time point, digital images of three representative cells were segmented according to the labeling of Cbl. The
fluorescence of c-Cbl and ErbB-1 labeling was calculated for each segmented vesicle. The scatter plots (arbitrary units) present ErbB-1
fluorescence vs. the c-Cbl fluorescence in each segmented vesicle. The correlation coefficient (r) indicates the strength of a linear
correlation between ErbB-1 and c-Cbl fluorescence. (B, right) Cbl-segmented vesicles, which showed ErbB-1 positivity above a thresh-
old, were considered Cbl-vesicles with ErbB1. (Top) Percentage of those vesicles from the total number of Cbl-segmented vesicles at
each time point. (Bottom) Percentage of ErbB-1 vesicles with Cbl from the total number of ErbB1-segmented vesicles. (C) CHO cells
were cotransfected with ErbB-1 and c-Cbl expression vectors and cells were incubated for 15 min at 37°C with EGF. Control
monolayers were mock stimulated (−). Endosomes were prepared as described in Materials and Methods and solubilized (1% Triton
X-100) for 30 min at 4°C. Cell lysates were cleared and subjected to immunoprecipitation (IP) and immunoblotting (IB) with the
indicated antibodies. Both the endosomal marker protein Rab-5 and c-Cbl were significantly enriched in the isolated fraction relative
to other fractions that werecollected (bottom;data notshown). (D) CHOcells werecotransfected withErbB-1 andan expressionvector
encoding c-Cbl fused inframe to a green fluorescence protein(GFP-Cbl). Forty-eighthours after transfection, cells were incubatedwith
Texas-red-labeled EGF (0.5 µg/ml) for 30 min at 4°C and then either transferred to 37°C (right), or left at 4°C (left), for an additional
incubation of 15 min.
Levkowitz et al.
3666 GENES & DEVELOPMENT
bution was independently supported by use of a fluores-
cently labeled EGF, andc-Cbl fused to a GFP. The results
presented in Figure 3D confirmed that the ligand-occu-
pied receptor molecules colocalize with Cbl-containing
vesicles.
c-Cbl and v-Cbl differently affect sorting
of endocytosed ErbB-1 molecules
Determination of the rate of intracellular accumulation
of EGF revealed that neither c-Cbl nor v-Cbl affected the
rate of ligand uptake (Fig. 4A), implying that Cbl acts at
a step distal to receptor binding and coated pit-mediated
internalization. This conclusion was also supported by
short time (1–5 min) ligand uptake experiments, and
analyses of the rate of EGF degradation, which was not
affected by c-Cbl or v-Cbl (data not shown). On the other
hand, examination of the receptor’s fate, by use of a
down-regulation assay indicated that c-Cbl, unlike
v-Cbl, significantly accelerated the rate of receptor
down-regulation (Fig. 4B). A surprising effect of v-Cbl,
however, was revealed on the background of the exten-
sive c-Cbl-induced down-regulation (Fig. 4B). Following
a rapid phase of partial ligand-induced down-regulation,
v-Cbl caused an almost complete recovery of ErbB-1 to
its original surface level. This effect of v-Cbl was sensi-
tive to monensin (Fig. 4B, right), a carboxylic ionophore
that exerts diverse intracellular effects, including disrup-
tion of recycling of EGF-occupied ErbB-1 molecules after
their endocytosis (Basu et al. 1981). Conceivably,
whereas c-Cbl can direct ErbB-1-loaded endosomes to
degradation, v-Cbl antagonizes this sorting function, and
directs the receptors to the alternative recycling path-
way.
Kinase activity and an intact carboxy-terminal region
are obligatory for Cbl-induced down-regulation
of ErbB-1
If c-Cbl plays a causative role in receptor down-regula-
tion, as implied by our results, then its interaction with
ErbB-1 may depend on kinase activity and autophos-
phorylation sites. This question was addressed by coex-
pressing a series of ErbB-1 mutant proteins, together
with c-Cbl, in CHO cells. The mutant proteins, along
with a wild-type receptor, are schematically depicted in
Figure 5A. They include a kinase-defective form (Kin
−
), a
deletion mutant lacking 214 amino acids of the carboxy-
terminal region (denoted Dc-214), and a protein whose
five major tyrosine autophosphorylation sites were re-
placed by phenylalanines (denoted F5). In addition, we
analyzed a receptor mutated at tyrosine 974 (denoted
A974) because this site has been implicated in the inter-
actions of ErbB-1 with the AP-2 adaptor of clathrin-
coated pits (Sorkin et al. 1996). Likewise, a receptor mu-
tated at both tyrosine residues 974 and 992 (denoted
Y5,6A) was analyzed because these sites are homologous
to the reported Cbl-docking site of the ZAP-70 tyrosine
kinase (Lupher et al. 1997). Transient expression of the
mutated proteins in CHO cells resulted in receptor lev-
els similar to that of the wild-type ErbB-1 (Fig. 5A). The
following aspects of Cbl interaction with ErbB-1 were
analyzed by use of this cellular system: ligand-induced
physical association, as well as tyrosine phosphoryla-
tion, of c-Cbl, recruitment of Shc, and receptor down-
regulation. The resultsof these assays are summarized in
a table (Fig. 5A), and presented in Figure 5, B and C. The
results obtained with Y5,6A are presented only in the
table.
Evidently, two features of ErbB-1 are essential for pro-
Figure 4. v-Cbl promotes receptor recycling,
whereas c-Cbl induces receptor down-regula-
tion. (A) Ligand internalization analyses. CHO
cells were cotransfected with an ErbB-1 vector
along with one of the following plasmids:
pcDNA3 (control, 䊊), c-Cbl expression vector
(䊉), or a plasmid directing v-Cbl expression
(䊏). Cell monolayers were treated for 2 hr at
4°C with
125
I-labeled EGF (at 10 ng/ml) and
then transferred to 37°C for the indicated pe-
riods of time. The fraction of internalized li-
gand was determined by use of a low-pH wash.
Each data point represents the average ±
S.E.
(bars) of triplicate measurements. (B) CHO
cells were cotransfected withan ErbB-1-encod-
ing plasmid along with an expression vector
encoding v-Cbl (䊏), c-Cbl (䊉), both v- and c-
Cbl (⽧), or with an empty vector (control, 䊊).
Cells were rinsed and incubated at37°Cfor the
indicated periods of time with EGF (at 250 ng/
ml). Sister cultures were similarly treated, ex-
cept that monensin (100 µ
M) was added to the
medium. Down-regulation assays were per-
formed as described in Materials and Methods.
Endosome sorting by Cbl
GENES & DEVELOPMENT 3667
ductive interaction with c-Cbl. These are the intrinsic
tyrosine kinase activity, and the presence of an intact
carboxy-terminal region. The five autophosphorylation
sites of ErbB-1 arenot crucial for the interactionbetween
c-Cbl and ErbB-1 (Fig. 5), although they were essential for
physical association with another ErbB-1 substrate,
namely Shc (Fig. 5B). In accordance with their inability
to interact with c-Cbl, the Kin
−
and the Dc214 mutants
displayed markedly reduced ligand internalization and
down-regulation (Sorkin et al. 1992; Chang et al. 1993).
We note that in some experiments basal association be-
tween c-Cbl and akinase-defective mutant of ErbB-1 was
detectable (cf. Figs. 5B and 7B), but in no case was it
ligand dependent. Interestingly, the carboxy-terminally
deleted mutant of ErbB-1 mediated both prolonged acti-
vation of the MAPK pathway (data not shown) and en-
hanced tyrosine phosphorylation of c-Cbl (Fig. 5B, bot-
tom). Nevertheless, it was refractory to the c-Cbl-in-
duced down-regulation effect (Fig. 5C), implying that
c-Cbl phosphorylation is insufficient for degradation of
Figure 5. Structural determinants of ErbB-1 that are essential for functional interactions with c-Cbl. (A) Schematic representation of
ErbB-1 mutants and their interactions with c-Cbl. The domain structure of ErbB-1 is shown by boxes that correspond to the double
cysteine-rich domain of the extracellular (EC) region, the transmembrane domain (TM), the juxtamembrane domain (JM), the tyrosine
kinase (TK) domain, and the carboxy-terminal tail (CT). The five major tyrosine autophosphorylation sites, along with the ␣-adaptin
tyrosine-based internalization signal [Y
974
; (Sorkin et al. 1996)] are indicated. The ATP-binding lysine residue (K
721
) was mutated to
an alanine residue in the kinase-defective mutant (Kin
−
). A carboxy-terminal deletion mutant (Dc214) lacking 214 carboxy-terminal
amino acids, and an ErbB-1 mutant in which the five major tyrosine phosphorylation sites were mutated to phenylalanine (F5) have
been described previously (Sorkin et al. 1996). A double tyrosine to alanine mutant (Y5,6) is also shown. A summary of the results
shown in B and C is presented in the table. The histogram presents the results of an assay that determined the binding of radioactive
EGF to the surface of cells transiently expressing the indicated mutants. (B) Monolayers of CHO cells were separately cotransfected
with plasmids encoding the indicated ErbB-1 mutants together with a c-Cbl-encoding vector. Sister plates were incubated for 15 min
at 37°C with or without EGF (at 100 ng/ml). Thereafter, whole cell lysates were subjected to immunoprecipitation (IP) and immu-
noblotting (IB) withthe indicated antibodies. (C) Theindicated ErbB-1 mutants were introduced into CHO cells by cotransfection with
a control vector (䊊) or a c-Cbl plasmid (䊉). EGF-induced down-regulation assay was then performed.
Levkowitz et al.
3668 GENES & DEVELOPMENT
ErbB-1. On the other hand, complex formation between
c-Cbl and ErbB-1 correlated with down-regulation (Fig.
5A), suggesting that stable physical association of c-Cbl
with the receptor is critical for directing ErbB-1 to deg-
radation.
c-Cbl increases ligand-induced ubiquitination
of ErbB-1
Next, we addressed the effect of c-Cbl on receptor deg-
radation by using chloroquine, an inhibitor of prelyso-
somal/lysosomal proteolysis. Chloroquine exerted no ef-
fect on the limited ability of EGF to induce degradation
of an overexpressed ErbB-1 in CHO cells (∼20% of ErbB-1
molecules underwent degradation following EGF treat-
ment, regardless of chloroquine presence, Fig. 6A). Co-
overexpression of c-Cbl significantly enhanced ligand-
induced degradation of ErbB-1 (80% of ErbB-1 molecules
underwent degradation), but chloroquine was able to
partly attenuate this Cbl-mediated enhanced degrada-
tion (only 44% of ErbB-1 molecules underwent degrada-
tion). Conceivably, c-Cbl affects receptor processing up-
stream to the chloroquine-sensitive late endosomal step.
Because ligand-induced ubiquitination of ErbB-1 pre-
cedes its intracellular degradation, and it depends on en-
docytosis of the ligand-receptor complexes (Galcheva-
Gargova et al. 1995), we analyzed the ability of c-Cbl to
affect receptor ubiquitination. In line with previous re-
ports, we detected a ligand-induced increase in ErbB-1
ubiquitination (Fig. 6B). By overlapping anti-ErbB-1 im-
munoblots with the ubiquitin signals, we learned that
the ubiquitinated fraction of ErbB-1 was minor (<5%,
open arrow in Fig. 6B). Overexpression of c-Cbl dramati-
cally increased the amount of ubiquitin that underwent
covalent attachment to ErbB-1 (Fig. 6B), implying that
the cellular level of Cbl critically controls the extent of
receptor ubiquitination.
In experiments that are not presented, we found that
treatment of cells with an inhibitor of proteasomal ac-
tivity, MG132, increased the extent of receptor ubiqui-
tination in cellsoverexpressing c-Cbl. Therefore, we ana-
lyzed the effect of MG132 on the fraction of surface-
exposed receptors by using a down-regulation assay. The
presence of MG132 decreased not only the basal EGF-
induced disappearance of ligand binding sites, but also
the accelerated down-regulation that was promoted by
an overexpressed c-Cbl (Fig. 6C). Because inhibition of
either c-Cbl (Fig. 4B) or proteasomes (Fig. 6C) can en-
hance recycling of ErbB-1, butblocking lysosomal hydro-
lases allows no recycling (data not shown), it is likely
that c-Cbl acts upstream to the proteasomal and lyso-
somal degradation processes.
Ligand-dependent ubiquitination corresponds
to the down-regulation activity of mutant Cbl
and ErbB-1 proteins
The contention that c-Cbl canaccelerate the degradation
rate of receptor molecules by increasing their ubiquiti-
nation predicts that mutant Cbl proteins, which are un-
able to down-regulate ErbB-1, will be defective in induc-
ing receptor ubiquitination. The two oncogenic forms of
c-Cbl, v-Cbl, and 70Z-Cbl, a deletion mutant lacking 17
internal amino acids (Langdon et al. 1989), were either
unable to enhance receptor down-regulation and ubiqui-
tination or elevated receptor expression (Fig. 7A). Like-
Figure 6. c-Cbl-induced down-regulation in-
volves an increase in ErbB-1 ubiquitination.
(A) Chloroquine sensitivity. The wild-type
form of ErbB-1was expressed in CHO cells by
cotransfection of an erbB-1-encoding plasmid
together with either a c-Cbl-expression vec-
tor or an empty vector (Cont.). Cells were in-
cubated for 45 min at 37°C in the absence or
presence of EGF (at 100 ng/ml) and chloro-
quine (CQ, 0.1 m
M). Cell lysates were pre-
pared and subjected to immunoprecipitation
(IP) and immunoblotting (IB)with anti-ErbB-1
antibodies. (B) CHO cells were transfected
and treated as in A. Cell lysates were sub-
jected to immunoprecipitation (IP) with anti-
bodies to ErbB-1 and immunoblotting (IB)
with antibodies to either ubiquitin (Ub) or
ErbB-1. (Closed arrowheads) The major band
of ErbB-1; (open arrowheads) the minor frac-
tion that underwent ubiquitination. (C)
ErbB-1 was transiently expressed in CHO
cells by cotransfection with either an empty
expression vector (control, 䊊) or a c-Cbl ex-
pression vector (䊉). EGF-induced down-regu-
lation of ErbB-1 was determined in the pres-
ence or absence of the proteasomal inhibitor
MG132 (10 µ
M).
Endosome sorting by Cbl
GENES & DEVELOPMENT 3669
wise, mutant ErbB-1 proteins that cannot stably recruit
c-Cbl displayed no ubiquitination. These are a kinase-
defective receptor, and a carboxy-terminally deleted
ErbB-1 (Fig. 7B). However, a receptor whose five major
autophosphorylation sites were mutated (F5) retained
enhanced ubiquitination (data not shown), in line with
its ability to recruit c-Cbl (Fig. 5). Because the defective
mutants were unable to associate with c-Cbl and accel-
erate receptor down-regulation and degradation (Figs. 5
and 7B), we concluded that functional interaction with
c-Cbl is necessary for receptor degradation, down-regu-
lation, and ubiquitination. The relative order of these
processes and their presumed compartmentalized orga-
nization are discussed below and summarized in a model
(Fig. 8).
Discussion
Endocytosis of the ErbB-1 is one of the best-characterized
routes of induced internalization of ligand-receptor com-
plexes (for review, see Sorkin and Waters 1993; Trow-
bridge et al. 1993). The most rapid endocytic pathway,
which utilizes clathrin-coated pits and vesicles, is char-
acterized by saturability: The rate of internalization de-
creases with increasing receptor occupancy (Lund et al.
Figure 7. Effect of Cbl proteins on ubiq-
uitination of ErbB-1 and its mutants. (A)
CHO cells were cotransfected with a plas-
mid encoding ErbB-1 together with vectors
directing the expression of the indicated
Cbl proteins. An empty vector was used
for control (Cont.). Cell monolayers were
treated for 15 min at 37°C with EGF (100
ng/ml). Thereafter, we eitheranalyzed cell
lysates by immunoprecipitation (IP) and
immunoblotting (IB) with the indicated
antibodies (left, arrowheads are as in Fig. 6)
or performed a ligand-binding assay as de-
scribed in the legend to Fig. 2B. (B) Mono-
layers of CHO cells were transfected with
a plasmid expressing c-cbl or a control
empty vector (−) together with vectors en-
coding the wild-type form (WT) of ErbB-1,
or the indicated mutants. Cell monolayers
were treated with EGF as in A and their
whole lysates subjected to immunopre-
cipitation (IP) with an antibody directed to
the extracellular portion of ErbB-1. Immunoblotting (IB) was performed with an antiserum to ubiquitin, or with an antibody directed
to the most carboxy-terminal 14 amino acids of ErbB-1. To confirm expression of the Dc214 mutant of ErbB-1, which was not
recognized by the immunoblotting antibody, we performed a ligand-binding assay on living cells (data not shown).
Figure 8. Proposed model of ligand-in-
duced endocytosis of ErbB-1. The model
summarizes the major steps of receptor
endocytosis and indicates their presumed
time scale. Ligand binding to ErbB-1 mol-
ecules, probably by elevating autophos-
phorylation (encircled P), induces their in-
teractions with clathrin-coated areas of
the plasma membrane, which rapidly in-
vaginate to form coated pits. c-Cbl may
not affect excision of the pit to form a
coated vesicle and the subsequent rapid
uncoating process. c-Cbl recruitment to
endosome-located ErbB-1 molecules tags
them for ubiquitination (Ub) and subse-
quent degradation through the combined
action of prelysosomal/lysosomal acid hy-
drolases, as well as by proteasomal pro-
teinases. v-Cbl shunts receptors to the de-
fault pathway, which involves recycling of
vesicles back to the cell surface. This step
is inhibitable by monensin.
Levkowitz et al.
3670 GENES & DEVELOPMENT
1990). Another limiting step is the sorting of internalized
receptors to lysosomal degradation (French et al. 1994).
On transient overexpression of ErbB-1 in CHO cells (∼2–
6×10
5
receptors/cell) the rapid endocytic pathway is
practically saturated and very low receptor down-regula-
tion occurs (Fig. 2C). Overexpression of c-Cbl on this
cellular background revealed a Cbl-dependent limiting
step that can regulate receptor degradation (Figs. 2C and
4B). Similarly, the negative regulation by c-Cbl of an-
other tyrosine kinase, Syk, was detectable only on Cbl
overexpression with recombinant vaccinia constructs
(Ota and Samelson 1997).
Down-regulation is the net result of receptor degrada-
tion and recycling. Because the rate of EGF internaliza-
tion was not detectably affected by c-Cbl overexpression
(Fig. 4A), but receptor down-regulation was accelerated
by c-Cbl (Fig. 4B), we propose that Cbl acts at a post-
coated pit step. In agreement with this difference, a ki-
nase-defective ErbB-1 mutant, which cannot interact
with Cbl, retained the ability to internalize EGF mol-
ecules and escaped degradation (Felder et al. 1990). The
observation that a co-overexpressed v-Cbl can reverse
c-Cbl’s action by shunting internalized receptors to a
monensin-sensitive pathway (Fig. 4B) led us to the propo-
sition that the recycling endosome (Trowbridge et al.
1993) is the site of Cbl’s action. In support of this possi-
bility, c-Cbl was localized to a vesicular compartment
(Figs. 2A and 3A) and it was fractionated with an endo-
somal fraction (Fig. 3C). Because receptor ubiquitination
was enhanced by c-Cbl (Fig. 6), and its structural require-
ments appear to reflect c-Cbl recruitment (cf. Figs. 7 and
5A), it is conceivable that the sorting function of c-Cbl
involves receptor ubiquitination. The scheme presented
in Figure 8 incorporates these conclusions into a model
that integrates data from the present and previous stud-
ies. According to this model, c-Cbl is not involved in the
entrapment of EGF receptors by coated regions of the
plasma membrane, and in the subsequent rapid invagi-
nation and scission that form coated pits and coated
vesicles, respectively. However, ∼1–2 min after binding
of EGF to thecell surface, c-Cbl becomes associatedwith
ErbB-1-containing vesicles (Fig. 3). On the basis of quan-
tification of fluorescent images, we assume that all ma-
ture ErbB-1-containing vesicles associate with c-Cbl, but
this recruitment engages only a small fraction of the cel-
lular pool of Cbl (Fig. 3B).
How exactly c-Cbl is recruited by ErbB-1 remains un-
clear. One possiblity that we have not addressed is the
involvement of several lysosome targeting motifs lo-
cated at the carboxy-terminal region of ErbB-1 (Opresko
et al. 1995). Although none of five tyrosine autophos-
phorylation sites located at this region appears essential
for Cbl-ErbB-1 interactions, we cannot exclude the pos-
sibility that tyrosine autophosphorylation of ErbB-1 also
plays a role in c-Cbl recruitment. For example, phos-
phorylation of tyrosine residues we have not mutated
may allow ErbB-1 to bind to c-Cbl either directly or in-
directly. Because c-Cbl isfound in living cells complexed
to several proteins, like Crk, phosphatidylinositol 3⬘-ki-
nase, and the Grb2 adaptor protein (Meisner et al. 1995),
it is possible that c-Cbl/Grb2, or other complexes, asso-
ciate with the carboxy-terminal region of ErbB-1 through
compensatory tyrosine phosphorylation sites.
Biochemical and morphological studies that utilized a
kinase-defective mutant of ErbB-1 revealed causal rela-
tionships between receptor occupancy, kinase activity,
and the rate of internalization (Honegger et al. 1987;
Chen et al. 1989; Hopkins et al. 1990). However, the role
played by the catalytic activity of ErbB-1 in lysosomal
targeting is currently unclear. According to one possibil-
ity, targeting is independent of kinase activity (Wiley et
al. 1991), but an alternative model implies that the in-
trinsic kinase can interrupt receptor recycling (Felder et
al. 1990). According to this scenario, the early endocytic
pathways of wild-type and kinase-impaired receptors are
identical, but after 10–20 min the pathways diverge at
the multivesicular body (MVB): Wild-type ErbB-1, des-
tined for degradation, localizes to internal vesicles,
whereas kinase-defective ErbB-1, destined for recycling,
localizes to surface membranes of the MVBs. The coin-
cidental peak of c-Cbl recruitment by ErbB-1-containing
endosomes (Fig. 3), together with the ability of c-Cbl to
promote down-regulation of wild type, but not a kinase-
defective mutant of ErbB-1 (Fig. 5), strongly support the
identification of c-Cbl as the protein substrate whose
recruitment by internalized receptors allows transloca-
tion into internal vesicles of the MVB. Conceivably, ad-
ditional endosomal molecules are involvedin the sorting
mechanism. Examples ofcandidate sorting moleculesin-
clude annexin I (Futter et al. 1993), SNX1, a homolog of
a yeast vacuolar sorting protein that accelerates down-
regulation of ErbB-1 (Kurten et al. 1996), phosphatidyli-
nositol 3⬘-kinase, (Joly et al. 1995), and Grb-2 (Wang and
Moran 1996).
The relationships between proteasomal and lysosomal
degradation processes is unclear at present. However, it
appears likely that ubiquitination occurs already in a
prelysosomal compartment. In accordance with this sce-
nario, we were unable to detect ubiquitinated ErbB-1
molecules on the surface of CHO cells by using biotin
labeling (data not shown). This observation is consistent
with reports that blocking endocytosis of either the
growth hormone receptor (Govers et al. 1997) or the EGF
receptor (Galcheva-Gargova et al. 1995), can prevent
ubiquitination. The situation is clearly different in yeast
cells: Ubiquitination of the Ste2p receptor, a G protein-
coupled receptor for the ␣ factor, is necessary for receptor
endocytosis (Hicke and Riezman 1996). In fact, transfer
of ErbB-1 to the lysosome may involve proteasomal ac-
tivity because inhibition at this step still allows some
receptor recycling (Fig. 6C). We note that degradation of
several other growth factor receptors such as the insulin-
like growth factor-1 receptor (Sepp Lorenzino et al.
1995), the PDGF-receptor (Mori et al. 1995), and the Met
receptor for the scatter factor (Jeffers et al. 1997), also
depend on proteasomal activity. Thus, complete degra-
dation of internalized ErbB-1 molecules probably in-
volves simultaneous proteolysis by hydrolases and pro-
teasomal proteinases at the late endosome and/or at the
lysosome.
Endosome sorting by Cbl
GENES & DEVELOPMENT 3671
Genetic analyses of mutant worms and flies indicated
that c-Cbl functions as a major negative regulator of in-
tercellular inductive processes controlling vulva
(Jongeward et al. 1995; Yoon et al. 1995) and eye (Meisner
et al. 1997) development, respectively. In both examples
growth factor signaling is funneled into the Ras-MAPK
pathway (for review, see Perrimon and Perkins 1997).
The analogous mammalian signaling machinery, which
is mediated by the ErbB network, also feeds into the
Ras-MAPK pathway (for review, see Alroy and Yarden
1997). However, unlike ErbB-1, which is strongly
coupled to c-Cbl, the two NDF receptors do not interact
with c-Cbl (Levkowitz et al.1996). Consequently, ErbB-3
and ErbB-4 undergono down-regulation following stimu-
lation with NDF (Figs. 1B and 2C; data not shown), and
no translocation of c-Cbl into receptor-containing endo-
somal vesicles occurs in response to NDF binding to
ErbB-3 (Fig 2A). These differences indicate that ErbB-1 is
subject to negative regulation by c-Cbl, whereas ErbB-3
and ErbB-4 escape Cbl-mediated attenuation. Conse-
quently, the mitogenic signal elicited by EGF is signifi-
cantly less potent than that induced by NDF binding to
ErbB-3 (Pinkas-Kramarski et al. 1996).
Our finding with ErbB-1 and ErbB-3 may be relevant to
the many other growth factor receptors that interact
with c-Cbl. Thus, it will be interesting to test the pre-
diction that most c-Cbl-coupled receptors will undergo
ligand-induced ubiquitination. In addition to ErbB-1, the
receptors for CSF-1 (Wang et al. 1996), PDGF (Mori et al.
1995), and the scatter factor (Jeffers et al. 1997) are also
candidates for c-Cbl-mediated down-regulation through
a ubiquitin-dependent process. Moreover, c-Cbl may
control also nonreceptor tyrosine kinases, such as Syk
(Ota and Samelson 1997), and Src (Tanaka et al. 1996). If
verified, this function may account for the ubiquitous
expression of c-Cbl, and for its general involvement in
apparently unrelated signaling pathways.
Materials and methods
Materials and antibodies
A recombinant form of NDF-
1177–246
was provided by Amgen
(Thousand Oaks, CA). Texas red-labeled EGF was from Molecu-
lar probes (Eugene, OR). Radioactive materials were purchased
from Amersham (Buckinghamshire, UK). Iodogen was from
Pierce. Biotin-X-NHS, MG-123, and lactacystin were from Cal-
biochem. Rabbit anti-c-Cbl (C-15) antibodies, as well as anti-
ErbB-3 antibodies, and a monoclonal antibody (mAb) to phos-
photyrosine antibody were from Santa-Cruz Biotechnology.
Murine mAbs to human ErbB-1 and ErbB-3 were from NeoMar-
kers (Fremont, CA). For immunoblot analysis of ErbB-1, Rab-5,
and Shc, we used antibodies from Transduction Laboratories
(Lexington, KY). Anti-ubiquitin antibody was kindly provided
by Drs. S. Yokota (Yamanashi Medical University, Japan) and
A. Amsterdam (Weizmann Institute, Rehovot, Israel). The anti-
hemagglutinin (HA) mAb was purchased from Boehringer
Mannheim.
Construction and transfection of expression vectors
To subclone c-cbl and 70Z-cbl into the pcDNA3 expression
vector (Invitrogen) containing the HA sequence tag (a gift from
Dr. Y. Haupt, Hebrew University, Jerusalem,Israel), we inserted
the cDNAs into the BamHI and XbaIorBamHI and XhoI sites
of pcDNA3, respectively. To generate thev-Cbl and Cbl-C trun-
cation mutants, we similarly subcloned cDNA fragments cor-
responding to amino acids 1–357 and 358–906, respectively. A
GFP–Cbl expression vector was generated by replacement of
Cbl’s stop codon with a SmaI site and insertion into the KpnI
and SmaI sites of pEGFP-N1 (Clontech). Expression vectors
(pcDNA3) containing the full-length cDNA of human ErbB-1
and ErbB-3 were described previously (Tzahar et al. 1996). The
ErbB-1 deletion mutant lacking 214 carboxy-terminal amino ac-
ids [Dc-214 (Sorkin et al. 1992)], as well as the F5 mutant in
which the five major tyrosine autophosphorylation sites were
mutated to phenylalanine residues (Soler et al. 1993), and the
A974 in which tyrosine 974 was replaced by an alanine (Sorkin
et al. 1996), have been described. All three mutants were sub-
cloned into pcDNA3. To generate a kinase-defective ErbB-1 (ly-
sine 721 mutated to alanine), and the double tyrosine to alanine
mutant Y5,6A (tyrosine residues 974 and 992 mutated to al-
anines), we used oligonucleotide-directed mutagenesis with T7-
DNA polymerase. Expression vectors were introduced to CHO
cells by the Lipofectamine transfection method (GIBCO-BRL).
The total amount of DNA in each transfection was normalized
with pcDNA3 plasmid. Twenty-four hours following transfec-
tion, cells were split and assayed 24 hr later.
Ligand binding, internalization analyses, and endosome
purification
Recombinant human ligands were labeled with Iodogen, and
their binding to cell monolayers and internalization were per-
formed as described (Waterman et al. 1998). An endosomal frac-
tion was prepared as described (Wada et al. 1992) with the fol-
lowing modifications: monolayers of CHO cells were scraped,
washed with PBS and resuspended in homogenization buffer
(250 m
M sucrose, 3 mM imidazole, 1 mM EDTA) containing
protease and phosphatase inhibitors. Cells were homogenized
(20 strokes, pestle B in a Dounce homogenizer) and centrifuged
for 10 min at 1500g. The sucrose concentration in the post-
nuclear supernatant was adjusted to 1.15
M, overlaid with 1.00
and 0.25
M sucrose cushions and centrifuged at 200,000g for 1.5
hr (Beckman SW41-Ti rotor). Endosomes were collected at the
0.25–1.00
M sucrose interface. The endosomal marker protein
Rab-5 was significantly enriched in this fraction relative to
other fractions of the sucrose gradient.
Receptor down-regulation and surface biotinylation assays
Ligand-induced receptor down-regulation was measured as fol-
lows: cells grown in 24-well plates were incubated at 37°C for
up to 2.5 hr with or without 250 ng/ml EGF or NDF in binding
buffer, and then rinsed with ice-cold binding buffer. Surface-
bound ligand molecules were removed by use of a low pH wash
(Pinkas-Kramarski et al. 1996). The number of ligand-binding
sites on the cell surface was then determined by incubating cells
at 4°C with the corresponding radiolabeled ligand for at least 1
hr. Biotinylation of cell surface proteins was performed as we
described previously (Waterman et al. 1998).
Immunoprecipitation and immunoblotting analyses
Subconfluent CHO cells were grown in 10-cm culture dishes,
washed briefly with PBS, and incubated for 10 min at 37°C with
the indicated recombinant growth factors (each at 100 ng/ml).
Levkowitz et al.
3672 GENES & DEVELOPMENT
To stop activation, cells were washed with ice-cold PBS and
kept on ice. Cell lysates were prepared at 4°C. Immunoprecipi-
tation, gel electrophoresis, and immunoblotting were per-
formed as we described previously (Levkowitz et al. 1996).
Immunofluorescence
Cells grown on cover-slips were rinsedwith serum-freemedium
and then treated in the absence or presence of ligands for 5 min
at 37°C. Thereafter, the medium was replaced and incubation
was continued for the indicated time intervals. Cells were fixed
for 15 min with 3% paraformaldehyde in PBS. For immunoflu-
orescent labeling, cells were permeabilized for 10 min at 22°C
with PBS containing 1% albumin and 0.2% Triton X-100. For
double labeling, cover-slips were incubated for 1 hr at room
temperature with mAb 111.1 or mAb 90 (anti-ErbB-1 or -3, re-
spectively) in combinationwith anti-Cbl antibodies. After wash
with PBS, the cover-slips were incubated with Cy3-conjugated
goat-anti-rabbit F(ab⬘)
2
and FITC-conjugated goat-anti-mouse
F(ab⬘)
2
-specific antibodies (Jackson ImmunoResearch Laborato-
ries) for an additional 1 hr. Finally, the coverslips were mounted
in Elvanol (Hoechst, Frankfurt).
Quantitative immunofluorescence microscopy
The system used for quantitative fluorescence microscopy and
image analysis will be described in detail elsewhere (E. Zamir,
B.-Z. Katz, K. Yamada, B. Geiger, and Z. Kam, in prep.). Speci-
mens were examined by use of a Zeiss Axioskope microscope
(Oberkochen, Germany) with a 100/1.3 plan-Neofluar objec-
tive. Images were acquired by a scientific, 12-bit, charged-
couple devise (CCD) camera (Model C220, Photometrics Co.).
Particles were segmented by the Water algorithm following
high-pass filter (subtracting from each pixel the average over
4 × 4 µm area around the particle). The parameters in Water
were adjusted to the dimension of the particles, and were kept
constant for all the analyses. To analyze the relationships be-
tween Cbl-labeled structures and ErbB-1-labeled vesicles,
vesicles were segmented separately and fluorescence intensities
measured for both Cbl- and ErbB1-segmented vesicles. Thetotal
number of Cbl- or ErbB1-containing vesicles was divided by the
total cell area to yield the average number of vesicles per 10
3
µm
2
. Cbl-segmented vesicles that had FITC fluorescent inten-
sity higher than a constant threshold were defined as Cbl-
vesicles with ErbB1, and their percentagefrom the total number
of Cbl-segmented vesicles was calculated. Similarly, the per-
centage of ErbB1-containing vesicles with Cbl from the total
number of ErbB1-segmented vesicles was obtained.
Acknowledgments
In cherished memory of Yarden Weinberg (1994–1998) who died
of rhabdomyosarcoma. We thank Drs. S. Yokota and A. Amster-
dam for anti-ubiquitin antibodies, Amar Sahay for constructing
the pcDNA3/F5 mutant, and Sara Lavi for technical help. This
work was supported by a grant from the National Institutes of
Health (CA72981).
The publication costs of this article were defrayed in part by
payment of page charges. This article must therefore be hereby
marked ‘advertisement’ in accordance with 18 USC section
1734 solely to indicate this fact.
Note added in proof
It has been reported very recently that Cbl enhances ubiquiti-
nation and degradation of the PDGF receptor [S. Miyake et al.
(1998) Proc. Natl. Acad. Sci. 95: 7927–7932].
References
Alroy, I. and Y. Yarden. 1997. The ErbB signaling network in
embryogenesis and oncogenesis: Signal diversification
through combinatorial ligand-receptor interactions. FEBS
Lett. 410: 83–86.
Basu, S.K., J.L. Goldstein, R.G.W. Anderson, and M.S. Brown.
1981. Monensin interrupts the recycling of low density lipo-
protein receptors in human fibroblasts. Cell 24: 493–502.
Boussiotis, V.A., G.J. Freeman, A. Berezovskaya, D.L. Barber,
and L.M. Nadler. 1997. Maintenance of humanT cell anergy:
Blocking of IL-2 gene transcription by activated Rap1. Sci-
ence 278: 124–128.
Bowtell, D.D. and W.Y. Langdon. 1995. The protein product of
the c-cbl oncogene rapidly complexes with the EGF receptor
and is tyrosine phosphorylated following EGF stimulation.
Oncogene 11: 1561–1567.
Chang, C.-P., C.S. Lazar, B.J. Walsh, M. Komuro, J.F. Collawn,
L.A. Kuhn, J.A. Tainer, I.S. Trowbridge, M.G. Farquhar,
M.G. Rosenfeld, H.S. Wiley, and G.N. Gill. 1993. Ligand-
induced internalization of the epidermal growth factor re-
ceptor is mediated by multiple endocytic codes analogous to
the tyrosine motif found in constitutively internalizing re-
ceptors. J. Biol. Chem. 268: 19312–19320.
Chen, W.S., C.S. Lazar, K.A. Lund, J.B. Welsh, C.-P. Chang,
G.M. Walton, C.J. der, H.S. Wiley, G.N. Gill, and M.G.
Rosenfeld. 1989. Functional independence of the epidermal
growth factor receptor from a domain required for ligand-
induced internalization and calcium regulation. Cell 59: 33–
43.
Felder, S., K.Miller, G. Moehren, A. Ullrich, J. Schlessinger, and
C.R. Hopkins. 1990. Kinase activity controls the sorting of
the epidermal growth factor receptor within the multive-
sicular body. Cell 61: 623–634.
French, A.R., G.P. Sudlow, H.S. Wiley, and D.A. Lauffenburger.
1994. Postendocytic trafficking of epidermal growth factor-
receptor complexes is mediated through saturable and spe-
cific endosomal interactions. J. Biol. Chem. 269: 15749–
15755.
Futter, C.E., S. Felder, J. Schlessinger, A. Ullrich, and C.R. Hop-
kins. 1993. Anexin I is phosphorylated in the multivesicular
body during the processing of the epidermal growth factor
receptor. J. Cell Biol. 120: 77–83.
Galcheva-Gargova, Z., S.J. Theroux, and R.J. Davis. 1995. The
epidermal growth factor receptor is covalently linked to
ubiquitin. Oncogene 11: 2649–2655.
Galisteo, M.L., I. Dikic, A.G. Batzer, W.Y. Langdon, and J.
Schlessinger. 1995. Tyrosine phosphorylation of the c-cbl
proto-oncogene protein product and association with epider-
mal growth factor (EGF) receptor upon EGF stimulation. J.
Biol. Chem. 270: 20242–20245.
Govers, R., P. van Kerkhof, A.L. Schwartz, andG.J. Strous. 1997.
Linkage of the ubiquitin-conjugating system and the endo-
cytic pathway in ligand-induced internalization of the
growth hormone receptor. EMBO J. 16: 4851–4858.
Hicke, L. and H. Riezman. 1996. Ubiquitination of a yeast
plasma membrane receptor signals its ligand-stimulated en-
docytosis. Cell 84: 277–287.
Honegger, A.M., T.J. Dull,S. Felder,E. vanObberghen, F. Bellot,
D. Szapary, A. Schmidt, A. Ullrich,and J. Schlessinger. 1987.
Point mutation at theATP bindingsite of EGF receptor abol-
ishes protein-tyrosine kinase activity and alters cellular
Endosome sorting by Cbl
GENES & DEVELOPMENT 3673
routing. Cell 51: 199–209.
Hopkins, C.R., A. Gibson, M. Shipman, and K. Miller. 1990.
Movement of internalized ligand-receptor complexes along a
continuous endosomal reticulum. Nature 346: 335–339.
Jeffers, M., G.A. Taylor, K.M. Weidner, S. Omura, and G.F.
Vande Woude. 1997. Degradation of the Met tyrosine kinase
receptor by the ubiquitin-proteasome pathway. Mol. Cell.
Biol. 17: 799–808.
Joly, M., A. Kazlauskas, and S. Corvera. 1995. Phosphatidylino-
sitol 3⬘-kinase activity is required at a postendocytic step in
platelet-derived growth factor receptor trafficking. J. Biol.
Chem. 270: 13225–13230.
Jongeward, G.D., T.R. Clandinin, and P.W. Sternberg. 1995. sli-
1, a negative regulator of let-23-mediated signaling in C. el-
egans. Genetics 139: 1553–1566.
Kurten, R.C., D.L. Cadena, and G.N. Gill. 1996. Enhanced deg-
radation of EGF receptors by a sorting nexin. Science
272: 1008–1010.
Langdon, W.Y., J.W. Hartley, S.P. Klinken, S.K. Ruscetti, and
H.C. Morse III. 1989. v-cbl, an oncogene from a dual-recom-
binant murine retrovirus that induces early B-lineage lym-
phomas. Proc. Natl. Acad. Sci. 86: 1168–1172.
Levkowitz, G., L.N. Klapper, E. Tzahar, A. Freywald, M. Sela,
and Y. Yarden. 1996. Coupling of the c-Cbl protooncogene
product to ErbB-1/EGF-receptor but not to other ErbB pro-
teins. Oncogene 12: 1117–1125.
Lund, K.A., L.K. Opresko, C. Starbuck, B.J. Walsh, and H.S.
Wiley. 1990. Quantitative analysis of the endocytic system
involved in hormone-induced receptor internalization. J.
Biol. Chem. 265: 15713–15723.
Lupher, M.L., Jr., Z. Songyang, S.E. Shoelson, L.C. Cantley, and
H. Band. 1997. The Cbl phosphotyrosine-binding domain se-
lects a D(N/D)XpY motif and binds to the Tyr292 negative
regulatory phosphorylation site of ZAP-70. J. Biol. Chem.
272: 33140–33144.
Meisner, H., B.R. Conway, D. Hartley, and M.P. Czech. 1995.
Interactions of Cbl with Grb2 and phosphatidylinositol 3⬘-
kinase in activated Jurkat cells. Mol. Cell. Biol. 15: 3571–
3578.
Meisner, H., A. Daga, J. Buxton, B. Fernandez, A. Chawla, U.
Banerjee, and M.P. Czech. 1997. Interactions of Drosophila
Cbl with epidermal growth factor receptors and roleof Cblin
R7 photoreceptor cell development. Mol. Cell. Biol.
17: 2217–2225.
Mori, S., K. Tanaka, S. Omura, and Y. Saito. 1995. Degradation
process of ligand-stimulated platelet-derived growth factor
beta-receptor involves ubiquitin-proteasome proteolytic
pathway. J. Biol. Chem. 270: 29447–29452.
Opresko, L.K., C.P. Chang, B.H. Will, P.M. Burke,G.N.Gill, and
H.S. Wiley. 1995. Endocytosis and lysosomal targeting of
epidermal growth factor receptors are mediated by distinct
sequences independent of the tyrosine kinase domain. J.
Biol. Chem. 270: 4325–4333.
Ota, Y. and L.E. Samelson. 1997. The product of the proto-on-
cogene c-cbl: A negativeregulator of the Syk tyrosine kinase.
Science 276: 418–420.
Perrimon, N. and L.A. Perkins. 1997. There must be 50 ways to
rule the signal: The case of the Drosophila EGF receptor.
Cell 89: 13–16.
Pinkas-Kramarski, R., L. Soussan, H. Waterman, G. Levkowitz,
I. Alroy, L. Klapper, S.Lavi, R.Seger, B.Ratzkin, M.Sela, and
Y. Yarden. 1996. Diversification of Neudifferentiation factor
and epidermal growth factor signaling by combinatorial re-
ceptor interactions. EMBO J. 15: 2452–2467.
Sepp Lorenzino, L., Z. Ma, D.E. Lebwohl, A. Vinitsky, and N.
Rosen. 1995. Herbimycin A induces the 20 S proteasome-
and ubiquitin–dependent degradation of receptortyrosine ki-
nases. J. Biol. Chem. 270: 16580–16587.
Soler, C., L. Beguinot, A. Sorkin, and G. Carpenter. 1993. Tyro-
sine phosphorylation of ras GTPase-activating protein does
not require association with the epidermal growth factor re-
ceptor. J. Biol. Chem. 268: 22010–22019.
Sorkin, A. and C.M. Waters. 1993. Endocytosis of growth factor
receptors. BioEssays 15: 375–382.
Sorkin, A., K. Helin, C.M. Waters, G. Carpenter, and L. Begui-
not. 1992. Multiple autophosphorylation sites of the epider-
mal growth factor are essential for receptor kinase activity
and internalization. Contrasting significance of tyrosine 992
in the native and truncated receptors. J. Biol. Chem.
267: 8672–8678.
Sorkin, A., M. Mazzotti, T. Sorkina, L. Scotto, and L. Beguinot.
1996. Epidermal growth factor receptor interaction with cla-
thrin adaptors is mediated by the Tyr974-containing inter-
nalization motif. J. Biol. Chem. 271: 13377–13384.
Tanaka, S., L. Neff, R. Baron, and J.B. Levy. 1995. Tyrosine
phosphorylation and translocation of the c-Cbl protein after
activation of tyrosine kinase signaling pathways. J. Biol.
Chem. 270: 14347–14351.
Tanaka, S., M. Amling, L. Neff, A. Peyman, E. Uhlmann, J.B.
Levy, and R. Baron. 1996. c-Cbl is downstream of c-Src in a
signaling pathway necessary for bone resorption. Nature
383: 528–531.
Thien, C.B. and W.Y. Langdon. 1997. EGF receptor binding and
transformation by v-cbl is ablated by the introduction of a
loss-of-function mutation from the Caenorhabditis elegans
sli-1 gene. Oncogene 14: 2239–2249.
Trowbridge, I.S., J.F. Collawn, and C.R. Hopkins. 1993. Signal-
dependent protein trafficking in the endocytic pathway.
Annu. Rev. Cell Biol. 9: 129–161.
Tzahar, E., H. Waterman, X. Chen, G. Levkowitz, D. Karunaga-
ran, S. Lavi, B.J. Ratzkin,and Y. Yarden. 1996. A hierarchical
network of inter-receptor interactions determines signal
transduction by NDF/neuregulin and EGF. Mol. Cell Biol.
16: 5276–5287.
Wada, I., W.H. Lai, B.I. Posner, and J.M. Bergeron. 1992. Asso-
ciation of the tyrosine phosphorylated epidermal growth fac-
tor receptor with a 55-kD tyrosine phosphorylated protein at
the cell surface and in endosomes. J. Cell Biol. 116: 321–330.
Wang, Y., Y.G. Yeung, W.Y. Langdon, and E.R. Stanley. 1996.
c-Cbl is transiently tyrosine-phosphorylated, ubiquitinated,
and membrane-targeted following CSF-1 stimulation of
macrophages. J. Biol. Chem. 271: 17–20.
Wang, Z. and M.F. Moran. 1996. Requirement for the adapter
protein GRB2 in EGF receptor endocytosis. Science
272: 1935–1938.
Waterman, H., I. Sabanai, B. Geiger, and Y. Yarden. 1998. Al-
ternative intracellular routing of ErbB receptors may deter-
mine signaling potency. J. Biol. Chem. 273: 13819–13827.
Wiley, H.S., J.J. Herbst, B.J. Walsh, D.A. Lauffenburger, M.G.
Rosenfeld, and G.N. Gill. 1991. The role of tyrosine kinase
activity in endocytosis, compartmentation, and down-regu-
lation of the epidermal growth factorreceptor. J. Biol. Chem.
266: 11083–11094.
Yoon, C.H., J. Lee, G.D. Jongeward, and P.W. Sternberg. 1995.
Similarity of sli-1, a regulator of vulval development in C.
elegans, to the mammalian proto-oncogene c-Cbl. Science
269: 1102–1105.
Levkowitz et al.
3674 GENES & DEVELOPMENT