Ubiquitin-mediated degradation of active Src tyrosine kinase

Abstract

Src family tyrosine kinases are involved in modulating various signal transduction pathways leading to the induction of DNA synthesis and cytoskeletal reorganization in response to cell-cell or cell-matrix adhesion. The critical role of these kinases in regulating cellular signaling pathways requires that their activity be tightly controlled. Src family proteins are regulated through reversible phosphorylation and dephosphorylation events that alter the conformation of the kinase. We have found evidence that Src also is regulated by ubiquitination. Activated forms of Src are less stable than either wild-type or kinase-inactive Src mutants and can be stabilized by proteasome inhibitors. In addition, poly-ubiquitinated forms of active Src have been detected in vivo. Taken together, our results establish ubiquitin-mediated proteolysis as a previously unidentified mechanism for irreversibly attenuating the effects of active Src kinase.

Full text

Ubiquitin-mediated degradation of active Src
tyrosine kinase
Kimya F. Harris, Ikuo Shoji, Eric M. Cooper, Sushant Kumar, Hideaki Oda, and Peter M. Howley*
Department of Pathology, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115
Contributed by Peter M. Howley, August 5, 1999
Src family tyrosine kinases are involved in modulating various
signal transduction pathways leading to the induction of DNA
synthesis and cytoskeletal reorganization in response to cell-cell or
cell-matrix adhesion. The critical role of these kinases in regulating
cellular signaling pathways requires that their activity be tightly
controlled. Src family proteins are regulated through reversible
phosphorylation and dephosphorylation events that alter the
conformation of the kinase. We have found evidence that Src also
is regulated by ubiquitination. Activated forms of Src are less stable
than either wild-type or kinase-inactive Src mutants and can be
stabilized by proteasome inhibitors. In addition, poly-ubiquiti-
nated forms of active Src have been detected in vivo. Taken
together, our results establish ubiquitin-mediated proteolysis as a
previously unidentified mechanism for irreversibly attenuating the
effects of active Src kinase.
T
he Src family kinases are nonreceptor tyrosine kinases that
mediate a variety of cellular signaling pathways. Nine ver-
tebrate members of the Src kinase family are known: some are
expressed in a cell-type specific manner and others are expressed
more broadly. For example, Blk and Lck expression is limited to
lymphoid lineages. In contrast, Src, Fyn, and Yes are more
broadly expressed although their levels of expression do vary
among cell types. For instance, Src is abundant in brain and
platelets, whereas Fyn is abundant in brain, platelets, and T
lymphocytes (for review see ref. 1).
All Src family members share the same basic structural
features. Each possesses an amino-terminal Src homology (SH)
4 domain that contains a myristylation signal sequence required
for membrane localization. Adjacent to the SH4 domain is a
unique region followed by conserved SH3 and SH2 domains,
which mediate protein–protein interactions. The catalytic kinase
domain resides in the carboxyl-terminal half of the protein.
Finally, there is a short carboxyl-terminal tail domain that
functions to negatively regulate the kinase activity (for reviews
see refs. 1 and 2). An intramolecular interaction between the
SH2 domain and a phosphorylated tyrosine residue (amino acid
527 in chicken c-Src) within this tail domain alters the confor-
mation of Src such that the catalytic domain is confined to an
inactive state (closed conformation; ref. 3). The dephosphory-
lation of tyrosine 527 results in a release of the intramolecular
interaction with the SH2 domain relaxing conformational con-
straints. The molecule takes on an ‘‘open conformation’’ that
facilitates catalytic activation (3–5).
Src family kinases function in numerous signaling pathways
including those mediating DNA synthesis and proliferation.
Activated cell surface receptors interact with and signal through
the Src family kinases. For example, binding of ligand to the
platelet-derived growth factor receptor results in its association
with and activation of the Src family kinases Src, Fyn, and Yes
(6). In turn, the activated Src kinases trigger a cascade of events,
ultimately leading to entry into S phase and subsequent DNA
replication (7–9). Src kinases also have a role in progression
through the G
2
兾M transition of the cell cycle, suggesting that
they function in both the G
1
and mitotic checkpoints (10–12).
Several Src family members (Src, Fyn, and Yes) are also
important for transducing signals in response to cell-cell or
cell-matrix adhesion (13). Adherence of cells to the extracellular
matrix via integrin receptors results in the assembly of protein
complexes required to modify the cellular cytoskeleton (for
reviews see refs. 1, 14, and 15). One of the early responses to
integrin receptor engagement is the activation of Src family
kinases that phosphorylate substrates such as focal adhesion
kinase, paxillin, and p130
cas.
These Src-mediated phosphoryla-
tion events have been associated with changes in cell adhesion,
motility, and shape (13, 16).
Because the Src family kinases affect both cell cycle progres-
sion and cytoskeletal organization, dysregulation may lead to
constitutive activation and cellular transformation (2). For ex-
ample, the viral derivative of Src, v-Src, and the point mutant
SrcY527F both lack the negative regulatory Tyr-527, leaving
them in the ‘‘open,’’ active conformation (17–21). Additional Src
variants bearing mutations in the SH2, SH3, and kinase domains
are also constitutively active presumably because of disruption of
the ‘‘closed’’ conformation of the protein (22, 23).
Although the reversible phosphorylation of the regulatory
tyrosine within the carboxyl terminus plays an important regu-
latory function, additional mechanisms may exist for controlling
the activity of Src family kinases. Indeed, our laboratory recently
has demonstrated that the Src family member Blk is regulated by
ubiquitin-mediated degradation (24). Specifically, the active
form of Blk is recognized by E6AP, an E3 ubiquitin-protein
ligase that promotes its ubiquitination and subsequent degrada-
tion (24–26). In this study, we show that Src itself is degraded in
a ubiquitin-dependent manner and that the active form is
specifically targeted for degradation. Taken together, these
results suggest that targeted degradation of active forms of the
Src family of tyrosine kinases may constitute an additional
mechanism for restricting the activity of these important signal-
ing proteins.
Materials and Methods
Plasmid Constructs. The pLNCX vectors encoding the chicken
c-Src, v-Src, c-Src(Y416F), and c-Src(K295R) were kind gifts of
Joan Brugge, Harvard Medical School (18, 20, 21, 27). pLNCX
vectors encoding c-Src(Y527F) and c-Src(E378G) were gener-
ated by subcloning ClaI fragments into pLNCX from plasmids
also given by Dr. Brugge (19, 21–23). pCMV4c-Src and pCMV4-
c-Src(E378G) were generated by subcloning KpnI–HindIII frag-
ments into pCMV4.
Cell Culture and Transfection. All cell lines were maintained in
DMEM (GIBCO兾BRL), supplemented with 100 units兾ml of
penicillin, 100
␮
g兾ml of streptomycin, and 10% FBS, at 37° in a
5% CO
2
incubator. csk⫹兾⫹, csk⫺兾⫺, src⫹兾⫹, and src⫺兾⫺
mouse embryo fibroblasts (MEFs; refs. 28 and 31) were kind
gifts of Sheila Thomas, Harvard Medical School. Transfections
Abbreviations: SH, Src homology; MEF, mouse embryo fibroblast; PVDF, poly(vinylidene
difluoride).
*To whom reprint requests should be addressed. E-mail: peter㛭howley@hms.harvard.edu.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
13738–13743
兩
PNAS
兩
November 23, 1999
兩
vol. 96
兩
no. 24

were performed by using standard electroporation procedures at
0.25 V for Cos-7 cells and 0.30 V for src⫺兾⫺ MEFs.
Antibodies. mAb 327 (anti-Src antibody; Oncogene Science and
generous gift of Dr. Thomas), polyclonal antibody PAb N16
(anti-Src antibody; Santa Cruz Biotechnology), polyclonal anti-
body PAb C11 (anti-actin antibody; Santa Cruz Biotechnology),
mAb 9E10 (anti-Myc antibody; Oncogene Science), and mAb
Ubi-1 (anti-ubiquitin antibody; Zymed) were used.
Cellular Lysate Preparation. Forty eight hours posttransfection,
cells were washed twice with PBS and incubated in 1 ml of
ice-cold RIPA buffer (150 mM NaCl兾1% NP40兾0.5% deoxy-
cholate兾0.1% SDS兾50 mM Tris䡠HCl, pH 7.5兾5
␮
g/ml PMSF兾5
␮
g/ml aprotinin兾5
␮
g/ml leupeptin) or NP-40 lysis buffer (100
mM Tris䡠HCl, pH 7.4兾120 mM NaCl兾1% NP-40兾5
␮
g/ml
PMSF兾5
␮
g/ml aprotinin兾5
␮
g/ml leupeptin) as indicated. Ly-
sates were sonicated and cleared of debris by centrifuging for 5
min at 4°C. For experiments requiring proteasome inhibitor
treatment, 48 hr posttransfection, cells were treated with 50
␮
M
Z-L
3
VS. (generous gift of Hidde Ploegh, Harvard Medical
School; ref. 30) for 8–14 hr. Control cells were treated with an
equal amount of DMSO and harvested as described above.
Immunoblotting. For each experiment, equal amounts of total
protein were resolved by 8% SDS兾PAGE. After SDS兾PAGE,
proteins were transferred to poly(vinylidene difluoride) (PVDF)
membrane in 12.5 mM Tris䡠HCl-100 mM glycine for 2 hr at 60
V. The membrane then was incubated in a blocking solution of
5% nonfat dried milk in TNET (10 mM Tris䡠HCl兾2.5 mM
EDTA兾50 mM NaCl兾0.1% Tween20) with rocking. After wash-
ing three times in TNET, the membrane was incubated in
primary antibody for at least 2 hr, washed three times, and
incubated in secondary antibody (horseradish peroxidase linked;
Amersham Pharmacia) for 1 hr. Membranes then were washed
three times in TNET and developed with enhanced chemilumi-
nescence reagents (NEN).
Pulse–Chase Experiments. src⫺兾⫺ cells were transiently trans-
fected with the pLNCX-Src constructs described above. After
plasmid transfection, the cells were pooled and divided equally
among five 6-cm
2
dishes. Forty eight hours posttransfection, cells
were washed twice in PBS and then incubated in DMEM lacking
methionine and cysteine for 1 hr at 37°C. Each 6-cm
2
dish was
labeled with 100
␮
Ci of
35
S methionine兾cysteine (Express Pro-
tein Labeling Mix; NEN) for a 45-min pulse, washed three times
with PBS, and chased with DMEM containing 10% FCS and
100-fold excess
L-methionine. At each time point, cells were
harvested in 0.5 ml of NP-40 lysis buffer as described above, and
equivalent amounts of cell lysate were immunoprecipitated by
using mAb 327. Immune complexes bound to the Sepharose
beads were washed three times in NP-40 lysis buffer and released
into gel loading buffer (2% SDS兾60 mM Tris䡠HCl, pH 6.8兾10%
glycerol兾0.1% bromophenol blue兾292 mM
␤
-mercaptoethanol)
by boiling for 3 min. Samples were resolved by 8% SDS兾PAGE,
dried, and exposed to film. Results were quantitated by using a
Molecular Dynamics PhosphorImager (Storm860).
In Vivo
Ubiquitination Assay. Cos-7 cells were cotransfected with
pCMV4 expressing either c-Src or c-SrcE378G and pCMV4Myc-
Ub. Where indicated, cells were treated with proteasome inhib-
itors as described above. Lysates were harvested in 1 ml of RIPA
buffer and immunoprecipitated with N16 anti-Src antibodies.
Immunoprecipitates were separated by SDS兾PAGE, transferred
to PVDF membrane, and immunoblotted as described above by
using mAb 9E10. For csk⫺兾⫺ cells, lysates were harvested and
immunoprecipitated as described above. Immunoprecipitates
were separated and transferred as above and immunoblotted
with mAb Ubi-1.
Results
Previous work from our laboratory has demonstrated that the
Src family tyrosine kinase Blk is degraded via the ubiquitin
pathway (24). Specifically, we found that activated Blk is ubi-
quitinated and is a substrate for E6AP, an E3 ubiquitin-protein
ligase. Our studies with E6AP and Blk suggested to us a model
in which the ubiquitin-mediated degradation of activated forms
of the Src family kinase might be a more general mechanism by
which this family of signaling proteins is regulated. To test this
model, we chose to investigate whether Src itself was regulated
by ubiquitination and proteolysis.
Activated Forms of Src Have Reduced Steady-State Protein Levels. To
test whether activated Src is less stable, steady-state Src levels
were analyzed in csk⫹兾⫹ and csk⫺兾⫺ MEF cells (29, 31). Csk
(C-terminal Src kinase) is responsible for the negative regulatory
tyrosine phosphorylation of Src at Tyr-527 (32). In cells from
Csk-deficient mice, Src activity is increased 11-fold as compared
with cells from a wild-type littermate (31). However, the steady-
state levels of Src were reduced 5.4-fold from the levels observed
in wild-type cells (Fig. 1), suggesting that steady-state levels of
Src are inversely proportional to their activity (29, 31). Half-life
experiments using cyclohexamide treatment to block protein
synthesis demonstrated that whereas Src protein from csk⫹兾⫹
cells remained stable for the duration of the experiment (2 hr),
Src protein levels from csk⫺兾⫺ cells were undetectable after 90
min (data not shown). These results demonstrated that the
difference in steady-state levels was caused by a decrease in the
protein stability of endogenous Src in csk⫺兾⫺ cells, presumably
because it is activated.
To test this hypothesis further, we analyzed the steady-state
levels of a series of active and inactive mutants of Src kinase
transfected into src⫺兾⫺ MEFs (28). The Rous sarcoma virus
derivative of Src, v-Src, encodes a protein with 10–20 times the
activity of c-Src (18, 20). c-SrcY527F lacks the negative regula-
tory Tyr-527 and exhibits 5–10 times the kinase activity of c-Src
(19, 21). c-SrcE378G contains a mutation in the kinase domain
and is 20 times more active than c-Src (22, 23). We also
characterized the steady-state levels of two inactive mutants of
c-Src: c-SrcY416F, which lacks a positive regulatory tyrosine
required for maximal kinase activity, and c-SrcK295R, which
lacks the critical lysine in the ATP-binding pocket, rendering it
completely inactive (20, 21, 27).
Western blot analysis performed on src⫺兾⫺ cells transfected
with the above constructs demonstrated that the levels of the
activated forms of Src, namely v-Src, c-SrcY527F, and
Fig. 1. Steady-state levels of endogenous Src protein. Equivalent amounts of
whole-cell lysate from src⫺兾⫺ and csk⫺兾⫺ MEFs and matched wild-type MEFs
were immunoprecipitated with mAb 327, separated by SDS兾PAGE, transferred
to PVDF membrane, and probed with mAb 327.
*
indicates the antibody heavy
chain.
Harris et al. PNAS
兩
November 23, 1999
兩
vol. 96
兩
no. 24
兩
13739
BIOCHEMISTRY

c-SrcE378G, were much lower than the levels of wild-type c-Src
protein (Fig. 2A, compare lanes 1–4). Quantitation of the Src
levels indicated a reduction in the steady-state protein by 3-fold,
4-fold, and 10-fold for c-SrcY527F, c-SrcE378G, and v-Src,
respectively, compared with c-Src levels. This inverse correlation
of protein levels to kinase activity has been noted previously for
v-Src and c-SrcY527F (20). The less active c-SrcY416F and the
completely inactive c-SrcK295R both showed steady-state levels
equivalent to or greater than c-Src (Fig. 2A, lanes 5 and 6). There
was a 1.4-fold increase in the steady-state levels of c-SrcY416F
and a 1.3-fold increase in the levels of c-SrcK295R compared
with c-Src (Fig. 2B). These results were confirmed in multiple
separate experiments using either RIPA or NP-40 lysis buffers
(data not shown). Each blot was probed for actin to control for
protein loading. The comparable levels of c-Src, Y416F, and
K295R forms of Src are consistent with our model because c-Src
is found predominantly in the inactive form (20, 21).
Activated Forms of Src Kinase Are Rapidly Degraded. We next tested
whether the decreased steady-state levels of the activated forms
of Src were caused by degradation and a shortened protein
half-life. Pulse–chase analyses therefore were performed on
individual Src proteins. The half-life of wild-type c-Src was
approximately 5–6 hr. Previous reports also have indicated that
c-Src is a relatively stable protein (33, 34). In contrast to the
stability of c-Src, we found that the half-lives of the active forms
of the kinase, c-SrcE378G and c-SrcY527F, were 1 hr and 3 hr,
respectively. The decreased stability observed for the active
forms of Src is consistent with the reduction in steady-state levels
being caused by increased protein turnover. Furthermore, the
less active and inactive forms of Src were more stable than the
wild-type c-Src. Both c-SrcY416F and c-SrcK295R proteins had
half-lives greater than 7.5 hr (Fig. 3).
Active Src Kinase Is a Target for the Ubiquitin Degradation Machinery.
We next examined whether the differential stabilities of the various
forms of the Src kinase were the result of the susceptibility of the
activated forms to ubiquitination. Our previous results with Blk
indicated that the preferential proteolysis of the activated forms of
the kinase involved ubiquitination and subsequent degradation by
the proteasome (24). To determine whether activated Src was being
degraded by the proteasome, we examined the effect of the
proteasome inhibitor, Z-L
3
VS, on the levels of Src protein (30). We
compared the effects of proteasome inhibitor on the steady-state
levels of c-Src (predominantly inactive) and the active form
c-SrcE378G. As shown in Fig. 4, there was a minimal 1.2-fold
increase in c-Src protein level in the presence of proteasome
inhibitor; however, there was a dramatic 3.3-fold increase in the
c-SrcE378G protein levels. This result, demonstrating the stabili-
zation of active Src by a proteasome inhibitor, confirmed the
involvement of the proteasome in the degradation of activated Src.
The slight stabilization of c-Src might be caused by a low level of
active wild-type c-Src or the normal turnover of inactive c-Src. The
proteasome inhibitor had no effect on actin levels.
To investigate further the role of the ubiquitin pathway in Src
proteolysis, we examined whether ubiquitinated forms of Src
Fig. 2. Steady-state levels of wild-type and mutant Src proteins. (A)(Upper)
Src⫺兾⫺ MEFs were transfected with pLNCX vector expressing c-Src, v-Src,
Y527F, E378G, Y416F, or K295R forms of Src, or pLNCX alone. Cells were
harvested after 48 hr, and 50
␮
g of whole-cell lysate from each transfection
was separated by SDS兾PAGE, transferred to PVDF membrane, and probed with
mAb 327. (Lower) The same membrane reprobed with anti-actin antibody
C11. (B) Quantitation of the results shown in A by using densitometry.
Fig. 3. Pulse–chase analysis of wild-type and mutant Src proteins. Src⫺兾⫺
MEFs were transfected with pLNCX vector expressing c-Src, or the following
Src mutants: Y527F, E378G, Y416F, or K295R. Forty eight hours posttransfec-
tion, cells were pulse-labeled with 100
␮
Ci
35
S Met兾Cys for 45 min. Plates were
incubated in ‘‘chase’’ media with 100-fold excess methionine and then har-
vested at the indicated times. Lysates were immunoprecipitated by using mAb
327 and separated by SDS兾PAGE. Gels were quantitated with a PhosphorIm-
ager, and each time point was normalized to the 0-hr time point.
13740
兩
www.pnas.org Harris et al.

could be detected in vivo. For these experiments c-Src or
c-SrcE378G were transfected into Cos-7 cells together with a
plasmid expressing Myc-tagged ubiquitin. Ubiquitinated forms
of c-SrcE378G were readily detected as a smear of higher
molecular weight bands in the cells coexpressing c-SrcE378G
and Myc-tagged ubiquitin (Fig. 5A, lanes 7 and 8). Although only
a faint ladder of multiubiquitinated Src was detected, the high
molecular weight smear is characteristic of multiubiquitinated
proteins and is similar to those found in previous reports using
Myc-tagged ubiquitin (35, 36). The addition of proteasome
inhibitor did not greatly stabilize the ubiquitinated forms of
c-SrcE378G, which is consistent with previous reports demon-
strating the overall stability of Myc-tagged ubiquitin protein
conjugates (37). We next examined the ubiquitination state of
endogenous Src in csk⫺兾⫺ cells where Src is predominantly
active. After treatment with proteasome inhibitor for 14 hr, cell
lysates were immunoprecipitated with anti-Src antibodies and
then analyzed by immunoblotting with anti-ubiquitin antibodies.
As shown in Fig. 5B, we could detect the characteristic high
molecular weight smear indicative of ubiquitination. In longer
exposures of the transfection experiment (Fig. 5A) and in
experiments using csk⫹兾⫹ cells we also were able to detect
evidence of high molecular bands indicative of ubiquitinated
c-Src (data not shown). Because the half-life of c-Src is 5–6 hr
(Fig. 3), the 14-hr proteasome inhibitor treatment period would
account for at least two half-lives of c-Src. This result suggests
that normal turnover of c-Src also may be regulated by ubiq-
uitination. Nonetheless, these results demonstrate the in vivo
ubiquitination of endogenous Src. Taken together with the
stabilization of Src observed with proteasome inhibitors, these
data confirm a role for ubiquitination in the regulation of Src
through targeted proteolysis of its activated forms (Fig. 6).
Discussion
Our laboratory recently has shown that activated forms of the Src
family member Blk are specifically degraded by ubiquitin-
mediated proteolysis (24). In the present study we have extended
this observation to Src itself and have examined whether this
mechanism of regulation is a property of other members of this
kinase family. Using csk⫺兾⫺ cells, which lack the negative
regulator of Src, and mutant forms of Src as sources of active
kinase, we demonstrated a decrease in the steady-state levels of
the active forms of Src. Pulse–chase experiments revealed
shorter half-lives for activated forms of c-Src, suggesting that the
decreased protein levels were the result of increased protein
turnover. In contrast, Src mutants fully or partially impaired in
their catalytic activity were quite stable and were present at
Fig. 4. Steady-state levels of Src in the presence of the proteasome inhibitor
Z-L
3
VS. (Upper) Src
⫺兾⫺
MEFs were transfected with pLNCX vector expressing
c-Src, the active E378G mutant, or vector alone. Forty eight hours after
transfection, cells were treated with either 50
␮
M Z-L
3
VS in DMSO or DMSO
alone as a negative control. Eight hours after treatment, cells were harvested,
separated by SDS兾PAGE, transferred to PVDF membrane, and probed with
mAb 327. The faint band in the LNCX lane is a background band. (Lower) The
same membrane reprobed with anti-actin antibody C11.
Fig. 5. In vivo ubiquitination of active c-Src. (A) Cos-7 cells were transiently transfected with pCMV4 vector alone (lane 1) or pCMV4 expressing c-Src (lane 2),
E378G (lane 3), Myc-tagged ubiquitin (lane 4) alone, or in the combinations indicated (lanes 5–8). For each transfection, 1 mg of lysate was immunoprecipitated
with polyclonal antibody (PAb) N16 (anti-Src), separated by SDS兾PAGE, transferred to PVDF membrane, and probed with mAb 9E10 (anti-myc). Cells used for lanes
6 and 8 were treated with 50
␮
M Z-L
3
VS for 14 hr before harvesting.
*
indicates a background band. (B) csk⫺兾⫺ cells were treated with either 50
␮
M Z-L
3
VS in
DMSO or DMSO alone as a negative control. One milligram of lysate was immunoprecipitated with PAb N16 (anti-Src), separated by SDS兾PAGE, transferred to
PVDF membrane, and probed with mAb Ubi-1 (anti-ubiquitin).
Harris et al. PNAS
兩
November 23, 1999
兩
vol. 96
兩
no. 24
兩
13741
BIOCHEMISTRY

higher steady-state levels. Proteasome inhibitors stabilized only
the active forms of Src, implicating the proteasome machinery in
the proteolysis of the active forms of Src. We also demonstrated
the in vivo ubiquitination of kinase active Src, thus establishing
the involvement of ubiquitination in the regulation of Src.
Regulation of Src by ubiquitin-mediated degradation provides
a previously unrecognized mechanism for limiting Src activity. A
major switch for the activation and inactivation of Src involves
the phosphorylation and dephosphorylation of tyrosine residues
within the C terminus and kinase domains of the protein.
Regulation by reversible phosphorylation allows the kinase to
alternate between active and inactive states. The specific pro-
teolysis of the kinase-active forms of Src by ubiquitination as
demonstrated in this manuscript provides an alternative mech-
anism for restricting Src kinase activity. The irreversible nature
of ubiquitin-mediated proteolysis allows the cell to control and
attenuate the activity of a mitogenic protein.
Although ubiquitin-mediated degradation has not previously
been shown to be a mechanism for regulation of Src activity, the
correlation between increased activity and decreased protein
stability has been noted previously. Using NIH 3T3 cells stably
expressing various Src constructs, Kmiecik and Shalloway (20)
demonstrated that v-Src and c-SrcY527F have higher specific
kinase activities and significantly lower steady-state protein
levels than c-Src. In addition, other studies have shown that v-Src
has a short-half life that varies among different viral strains,
suggesting that the stability of v-Src depends on mutations
unique to each strain. Protein synthesis inhibitors greatly in-
creased the half-life of v-Src in these studies, indicating that
additional proteins are required for v-Src degradation (38, 39).
Unregulated, constitutive activation of Src induces cellular
transformation presumably through activation of mitogenic and
cellular adhesion pathways. In most cases, transformation by Src
occurs only when activated forms of the kinase are present;
however, there are data suggesting that elevated levels of c-Src
are transforming. Johnson et al. (40) found that c-Src transfected
into NIH 3T3 cells could induce low levels of foci. When these
transformed foci were examined, Src levels were consistently
higher than those found in transfected, nontransformed NIH
3T3 cells (average ratio of 5:1). Confirmation of the wild-type
status of the c-Src in these foci led these researchers to propose
that a level of c-Src above a certain threshold was oncogenic.
Defects in Src ubiquitination could cause Src protein levels to
exceed a ‘‘threshold’’ level and result in cellular transformation.
In fact, various human cancers, including breast cancer, exhibit
elevated Src levels despite the absence of any detectable Src
mutations (41). Elevated levels of Src in these cancers possibly
could result from mutations in genes that regulate Src levels,
possibly in those that participate in the proper targeting of Src
to the ubiquitin proteasome pathway.
Although we have established that the ubiquitin兾proteasome
pathway is involved in degrading activated Src, the specific
enzymes involved in the process have yet to be determined. The
E3 ubiquitin-protein ligases are the critical enzymes responsible
for substrate recognition by the ubiquitin machinery (for review
see ref. 42). E6AP was the first mammalian E3 enzyme to be
identified (26). In cooperation with the human papillomavirus
E6 protein, E6AP mediates the ubiquitination of p53 (25, 26).
E6AP is the prototype of a family of E3 proteins known as HECT
domain proteins (homology to E6AP carboxyl terminus), all of
which share similarities in their catalytic domains (43, 44). Based
on our previous results demonstrating a role for E6AP in
targeting Blk for degradation, E6AP, or perhaps a related HECT
family member, is a strong candidate for the E3 responsible for
recognizing activated forms of Src. In preliminary coimmuno-
precipitation experiments, we have detected an interaction
between E6AP and Src (K.F.H., E.M.C., and P.M.H., unpub-
lished results). Additional experiments, however, will need to be
performed to establish the specific ubiquitin protein ligase that
is involved in the degradation of Src.
In addition to identifying the E3 enzyme involved, it will be
important to determine the specific degradation signal(s) in the
active form of Src that is recognized by the ubiquitin machinery.
Activated Src adopts an open conformation in which tyrosine
416 becomes phosphorylated (2). It is possible that either the
change in conformation status or the phosphorylation of specific
residues such as tyrosine 416 serves as the degradation signal.
The stability of c-Src K295R argues for the latter possibility. This
mutant can adopt the open conformation, yet remains stable,
suggesting that conformation alone is not sufficient to target Src
for degradation. There also may be a requirement for interac-
tions with additional cellular proteins before recognition of Src
as a substrate for the ubiquitination machinery. These mecha-
nisms are not mutually exclusive and must be carefully explored.
The work presented in this report demonstrates the targeted
degradation of active forms of Src by the ubiquitin兾proteasome
pathway. These results reveal a previously unidentified cellular
mechanism for irreversibly limiting Src kinase activity. Further
investigation and elucidation of this pathway is necessary to gain
a general understanding of the cellular regulation of Src activity
and how this regulation may be disrupted in cancer cells.
We thank J. Brugge for plasmids expressing wild-type and mutant Src
kinases. We thank S. Thomas for anti-Src antibodies and src⫹兾⫹,
src⫺兾⫺, csk⫹兾⫹, and csk⫺兾⫺ MEFs and H. Ploegh for Z-L
3
VS
proteasome inhibitor. We are grateful to J. Brugge, A. Hudson, L.
Decker, and W. Kao for a critical review of this manuscript. This work
was supported by a grant from the National Institutes of Health
(R01-CA64888). K.F.H. was supported by a postdoctoral fellowship from
the National Institutes of Health兾National Cancer Institute (award
number 1 F32 CA81727–01). S.K. was supported by a postdoctoral
fellowship from the American Cancer Society (award number PF-4309).
1. Thomas, S. M. & Brugge, J. S. (1997) Annu. Rev. Cell. Dev. Biol. 13, 513–609.
2. Brown, M. T. & Cooper, J. A. (1996) Biochim. Biophys. Acta 1287, 121–149.
3. Xu, W., Doshi, A., Lei, M., Eck, M. J. & Harrison, S. C. (1999) Mol. Cell 3,
629–638.
4. Taylor, S. J. & Shalloway, D. (1993) Curr. Opin. Genet. Dev. 3, 26–34.
5. Xu, W., Harrison, S. C. & Eck, M. J. (1997) Nature (London) 385, 595–602.
6. Ralston, R. & Bishop, J. M. (1985) Proc. Natl. Acad. Sci. USA 82, 7845–7849.
7. Twamley-Stein, G. M., Pepperkok, R., Ansorge, W. & Courtneidge, S. A.
Fig. 6. Proposed model for the targeted degradation of active Src. Phos-
phorylation of Tyr-527 maintains Src in a ‘‘closed’’ inactive conformation.
Dephosphorylation of pTyr-527 and autophosphorylation of Tyr-416 releases
the kinase into an open, fully active state. The active form of Src is specifically
targeted for ubiquitin-mediated proteolysis. The E3 ubiquitin-protein ligase
that recognizes active Src has not yet been identified.
13742
兩
www.pnas.org Harris et al.

(1993) Proc. Natl. Acad. Sci. USA 90, 7696–7700.
8. Roche, S., Koegl, M., Barone, M. V., Roussel, M. F. & Courtneidge, S. A.
(1995) Mol. Cell. Biol. 15, 1102–1109.
9. Barone, M. V. & Courtneidge, S. A. (1995) Nature (London) 378, 509–512.
10. Chackalaparampil, I. & Shalloway, D. (1988) Cell 52, 801–810.
11. Fumagalli, S., Totty, N. F., Hsuan, J. J. & Courtneidge, S. A. (1994) Nature
(London) 368, 871–874.
12. Roche, S., Fumagalli, S. & Courtneidge, S. A. (1995) Science 269, 1567–1569.
13. Klinghoffer, R. A., Sachsenmaier, C., Cooper, J. A. & Soriano, P. (1999) EMBO
J. 18, 2459–2471.
14. Juliano, R. L. & Haskill, S. (1993) J. Cell Biol. 120, 577–585.
15. Schwartz, M. A. (1993) Cancer Res. 53, 1503–1506.
16. Hanks, S. K. & Polte, T. R. (1997) BioEssays 19, 137–145.
17. Takeya, T. & Hanafusa, H. (1983) Cell 32, 881–890.
18. Hunter, T. (1987) Cell 49, 1–4.
19. Cartwright, C. A., Eckhart, W., Simon, S. & Kaplan, P. L. (1987) Cell 49, 83–91.
20. Kmiecik, T. E. & Shalloway, D. (1987) Cell 49, 65–73.
21. Piwnica-Worms, H., Saunders, K. B., Roberts, T. M., Smith, A. E. & Cheng,
S. H. (1987) Cell 49, 75–82.
22. Levy, J. B., Iba, H. & Hanafusa, H. (1986) Proc. Natl. Acad. Sci. USA 83,
4228–4232.
23. Bjorge, J. D., Bellagamba, C., Cheng, H. C., Tanaka, A., Wang, J. H. & Fujita,
D. J. (1995) J. Biol. Chem. 270, 24222–24228.
24. Oda, H., Kumar, S. & Howley, P. M. (1999) Proc. Nat. Acad. Sci. USA 96,
9557–9562.
25. Huibregtse, J. M., Scheffner, M. & Howley, P. M. (1993) Mol. Cell. Biol. 13,
775–784.
26. Scheffner, M., Huibregtse, J. M., Vierstra, R. D. & Howley, P. M. (1993) Cell
75, 495–505.
27. Snyder, M. A., Bishop, J. M., McGrath, J. P. & Levinson, A. D. (1985) Mol. Cell.
Biol. 5, 1772–1779.
28. Soriano, P., Montgomery, C., Geske, R. & Bradley, A. (1991) Cell 64, 693–702.
29. Nada, S., Yagi, T., Takeda, H., Tokunaga, T., Nakagawa, H., Ikawa, Y., Okada,
M. & Aizawa, S. (1993) Cell 73, 1125–1135.
30. Bogyo, M., McMaster, J. S., Gaczynska, M., Tortorella, D., Goldberg, A. L. &
Ploegh, H. (1997) Proc. Natl. Acad. Sci. USA 94, 6629–6634.
31. Imamoto, A. & Soriano, P. (1993) Cell 73, 1117–1124.
32. Nada, S., Okada, M., MacAuley, A., Cooper, J. A. & Nakagawa, H. (1991)
Nature (London) 351, 69–72.
33. Iba, H., Cross, F. R., Garber, E. A. & Hanafusa, H. (1985) Mol. Cell. Biol. 5,
1058–1066.
34. Hirai, H. & Varmus, H. E. (1990) Mol. Cell. Biol. 10, 1307–1318.
35. Ward, C. L., Omura, S. & Kopito, R. R. (1995) Cell 83, 121–127.
36. Hofmann, F., Martelli, F., Livingston, D. M. & Wang, Z. (1996) Genes Dev. 10,
2949–2959.
37. Ellison, M. J. & Hochstrasser, M. (1991) J. Biol. Chem. 266, 21150–21157.
38. Ziemiecki, A., Friis, R. R. & Bauer, H. (1982) Mol. Cell. Biol. 2, 355–360.
39. Sefton, B. M., Patschinsky, T., Berdot, C., Hunter, T. & Elliott, T. (1982)
J. Virol. 41, 813–820.
40. Johnson, P. J., Coussens, P. M., Danko, A. V. & Shalloway, D. (1985) Mol. Cell.
Biol. 5, 1073–1083.
41. Kolibaba, K. S. & Druker, B. J. (1997) Biochim. Biophys. Acta 1333, F217–F248.
42. Ciechanover, A. & Schwartz, A. L. (1994) FASEB J. 8, 182–191.
43. Huibregtse, J. M., Scheffner, M., Beaudenon, S. & Howley, P. (1995) Proc. Natl.
Acad. Sci. USA 92, 2563–2567.
44. Huibregtse, J. M., Scheffner, M. & Howley, P. M. (1994) in Cold Spring Harbor
Symposia on Quantitative Biology (Cold Spring Harbor Lab. Press, Plainview,
NY), Vol. 59, pp. 237–245.
Harris et al. PNAS
兩
November 23, 1999
兩
vol. 96
兩
no. 24
兩
13743
BIOCHEMISTRY