(Received for publication, August 2, 1996, and in revised form, November 7, 1996)
From SUGEN, Inc., Redwood City, California 94063
Cadherins are transmembrane receptors with an
extracellular domain that participates in homophilic cell to cell
adhesion and a cytoplasmic domain that associates with proteins called
catenins. Cadherin-mediated adhesion as well as adhesion-independent
functions for catenins play important roles in differentiation,
development, and malignant transformation. Mechanisms that regulate
steady-state catenin levels and cadherin-catenin complex stability are
poorly understood, but activities of both the Wnt-1 proto-oncogene and tyrosine kinases are implicated. Here I define, at the biochemical level, distinct mechanisms that modulate steady-state catenin levels.
Increased cadherin expression, providing more catenin binding sites,
leads to selective stabilization of the cadherin-associated population
of - and
-catenin, but not p120cas. In contrast,
expression of Wnt-1 leads primarily to increased stability of the
uncomplexed pool of
-catenin without effect on p120cas.
Significantly, the Wnt-1-induced stabilization of uncomplexed
-catenin is independent of cadherin expression. Transformation by
v-Src does not disrupt the catenin-cadherin complex despite the
phosphorylation of E-cadherin and
-catenin on tyrosine. In contrast
to the effects of Wnt-1, v-Src does not modulate the uncomplexed
population of
-catenin. p120cas is phosphorylated on
tyrosine by v-Src, and this is accompanied by a significant decrease in
the level of uncomplexed p120cas as well as a change in
behavior of p120cas upon biochemical fractionation. Taken
together these data suggest that p120cas and
-catenin are
regulated independently.
The cadherin cell to cell adhesion receptors play an important part in early embryonic development as well as in the morphogenesis of many tissue types and maintenance of the differentiated phenotype (1, 2). The cadherin adhesion system has also been shown to serve as a tumor invasion suppressor and is often defective in many tumor types that exhibit loss of cell to cell contacts and increased invasive and metastatic capability (3-5).
The cadherins are a family of transmembrane proteins that are localized
to the adherens type of intercellular junction (6, 7). The
extracellular domain of cadherin participates in
Ca2+-dependent, homophilic cell to cell
adhesion, and the cytoplasmic domain associates with a family of
proteins called catenins (6, 7). Three major catenin
species have been identified: -catenin, a vinculin family member (8,
9), and two related proteins,
-catenin and plakoglobin (
-catenin)
(10).
-Catenin and plakoglobin are mammalian homologs of a protein
originally described in Drosophila called Armadillo (11),
whose function is critical for normal segmental pattern formation
during development (12). Another mammalian protein, p120cas,
that is phosphorylated on tyrosine in response to both v-Src transformation (13) and activation of several receptor tyrosine kinases
(14, 15), has recently been included in the armadillo family
based on the presence of a repeating 42-amino acid sequence motif that
characterizes these proteins (16, 17), as well as the ability of
p120cas to associate with cadherin (18-21).
Several lines of evidence have shown that catenin binding to the
cytoplasmic domain of cadherin is required for effective cell to cell
adhesion (22, 23). -Catenin and plakoglobin form a mutually
exclusive association with cadherin (24, 25), and
-catenin is
incorporated into the complex through direct binding to either
-catenin or plakoglobin but not cadherin (26-28). The ability of
-catenin to anchor the cadherin complex to actin filaments of the
cell cytoskeleton is thought to be a key event in strengthening the
cadherin-mediated adhesive contact (8, 9, 23, 29).
One of the first insights to the regulation of catenin function came
from studies in the Drosophila system where it was shown that expression of the segment polarity gene, wingless,
leads to the post-transcriptional accumulation of Armadillo protein (30). In mammalian cells, expression of the wingless
homolog, the proto-oncogene wnt-1, results in increased
steady-state levels of -catenin and plakoglobin (24, 31),
stabilization of the cadherin-
-catenin complex, and increased cell
to cell adhesion (24). The specific role that cadherin serves in the
regulation of steady-state catenin levels by Wnt-1 is not known.
In addition to serving a pivotal role in modulation of cell adhesion,
recent evidence suggests that catenins can function in
cadherin-independent signaling pathways that regulate differentiation and cell proliferation (32, 33). In Drosophila, it has been proposed that accumulation of cytoplasmic Armadillo protein serves a
signaling function resulting in expression of genes required for
segmental pattern formation during development (34). In mammalian cells
responding to Wnt-1, the increased steady-state levels of catenins is
primarily due to a selective increase in the amount of uncomplexed,
monomeric -catenin and plakoglobin (35). While this free pool can
contribute to an increased stability of cadherin-catenin complexes
(24), a second adhesion-independent function is likely. The regulation
of p120cas by Wnt-1 and the consequences of tyrosine
phosphorylation on the regulation of free pools and complexed
populations of p120cas have not been examined.
Both -catenin and plakoglobin, but not p120cas, associate
directly with adenomatous polyposis coli
(APC)1 (19, 36, 37), a tumor suppressor
protein that is mutated in the majority of colorectal carcinomas (38).
Cells homozygous for mutant APC protein exhibit a significant increase
in a free, uncomplexed pool of
-catenin and plakoglobin (39), and
accumulating evidence indicates that APC functions in the Wnt-1
signaling pathway to directly regulate the stability of uncomplexed
-catenin and plakoglobin (35). In another study a mutant form of
-catenin, lacking the ability to bind
-catenin, leads to
transformation of NIH3T3 cells upon overexpression (40). Further
evidence for an adhesion-independent function for catenins comes from
studies in Xenopus. Overexpression of
-catenin or
plakoglobin in fertilized Xenopus oocytes leads to a
duplication of the dorsal embryonic axis during development (41, 42).
Portions of
-catenin that lack the
-catenin binding site are
still active in this assay, whereas overexpression of cadherin has no
effect (42, 43).
Several lines of evidence have implicated tyrosine kinases in the
modulation of cadherin-catenin complex formation and function. Transformation of cells by the v-Src tyrosine kinase oncogene leads to
a disruption of adherens junctions and an increase in invasive
properties and metastatic potential (44, 45). These phenotypes
correlate with phosphorylation on tyrosine of both cadherin and
-catenin (46-48). Activation of several receptor tyrosine kinases,
including receptors for hepatocyte growth factor (Met) and epidermal
growth factor, also leads to the phosphorylation on tyrosine of
cadherin,
-catenin, and plakoglobin (49, 50). Of interest is the
fact that the biological activities of these receptors include the
dissociation of cell monolayers and induction of cell motility for Met
(51) and cell rounding and membrane ruffling for the epidermal growth
factor receptor (52, 53). Transformation of human breast epithelial
cells by an activated form of Ras indirectly leads to phosphorylation
on tyrosine of
-catenin and p120cas and this correlates with
alterations in the adherens junctions and changes in cell morphology
(54). These studies implicate phosphorylation on tyrosine as a key
regulatory event in cadherin-catenin complex formation and function;
however, the biochemical mechanisms by which this is accomplished are
unknown, and the role of tyrosine phosphorylation in regulating free
catenin pools has not been examined.
In this study I have defined, at the biochemical level, distinct
mechanisms that modulate steady-state catenin levels. By increasing the
number of cadherin molecules, thus providing more catenin binding
sites, the complexed population of - and
-catenin, but not
p120cas, can be selectively increased without affecting the
uncomplexed, monomeric catenin pools. In contrast, the proto-oncogene
wnt-1 leads to an increased stability of the free pool of
-catenin but has no effects on
-catenin or p120cas.
Significantly, the effects of Wnt-1 on
-catenin are
cadherin-independent. In contrast to the effects of Wnt-1, v-Src
transformation does not regulate the uncomplexed population of
-catenin or the cadherin complex with
-catenin and
-catenin,
despite the phosphorylation of
-catenin and cadherin on tyrosine.
However, p120cas also becomes heavily phosphorylated on
tyrosine by v-Src, and this leads to a decrease in the uncomplexed
population as well as a change in distribution of p120cas
detected by biochemical fractionation.
L cells were obtained from ATCC. An expression vector pECE/ECad was constructed by subcloning the E-cadherin cDNA (55), generously provided by Rolf Kemler (Max Planck Institute, Freiburg, Germany), into an SV40 expression vector, pECE. A Wnt-1 expression vector, pRSVWnt-1, and pSV2Neo have been described (56, 57). A previously described v-Src expression vector, pMvSrc (58), was kindly provided by David Shalloway (Cornell University, Itnaca, NY). A vector encoding a hygromycin resistance marker, pREP4 was purchased from Invitrogen Corp.
The L cells were transfected with either pSV2Neo alone (control) or
were co-transfected with pECE/ECad plus pSV2Neo using LipofectAMINE
reagent (Life Technologies, Inc.) as directed by the manufacturer.
After 48 h the cells were either harvested for transient
expression analysis as indicated or were subjected to drug selection
using Geneticin (G418) (Life Technologies, Inc.). Stable clones of
transfected cells were selected, grown, and analyzed for E-cadherin
protein expression by Western immunoblot. Multiple transfected clones
were analyzed with similar results (for example see Fig. 1). Most of he
experiments shown in this report were performed with one clonal line,
ECad2-2. Two of the E-cadherin-transfected L cell lines, including
clone ECad2-2, were co-transfected with pMvSrc plus pREP4, and stable
cell lines were selected using hygromycin B (Sigma).
v-Src expression was verified by Western immunoblot analysis. Multiple
v-Src transfected clones were analyzed with similar results. The
experiments shown in this report were performed with either of two cell
lines, ECad/Src 6 or ECad/Src 20. For Fig. 4A, cells were
observed with an Axiovert 100 microscope using a 10× projective lens,
and photographs were taken using Polaroid 57 film.
Antisera
Polyclonal antisera made in rabbits against
-catenin- and
-catenin-specific peptides were produced and
characterized as described (24). The rabbit polyclonal antiserum
against cadherin was generated as described (59) and was generously
provided by James Nelson (Stanford University, Stanford, CA). A mouse
monoclonal antibody against p120cas was purchased from
Transduction Laboratories. Monoclonal antibodies 327 and 2-17 against
Src protein were kindly provided by Joan Brugge (Ariad
Pharmaceuticals). A monoclonal antibody directed against
phosphotyrosine, PY-20, was purchased from ICN Biochemicals.
For immunoprecipitations and affinity precipitations, cells were extracted with a buffer consisting of 1% Nonidet P-40, 10 mM sodium phosphate, pH 7.0, 0.15 M sodium chloride, and a mixture of protease and phosphatase inhibitors. Monolayers of tissue culture cells were extracted for 30 min on ice and scraped from the culture dish, and the mixture was centrifuged at 12,000 × g for 15 min at 4 °C. The protein concentration of clarified extracts was determined using a BCA reagent kit (Pierce). Cell extracts used for total protein analysis only were made with immune precipitation buffer or Laemmli sample buffer as needed.
Antibodies for immunoprecipitation were added, as indicated for each figure, to clarified extracts adjusted for protein concentration. Extracts were incubated with antisera at 4 °C for approximately 2 h, and immune complexes were collected by binding to protein A-Sepharose beads (Pharmacia Biotech Inc.), followed by three washes with extraction buffer. The immunoprecipitates were subjected to SDS-PAGE and immunoblot analysis as indicated.
For the subcellular fractionation experiment (Fig. 8), one dish each of
ECad2-2 and ECad/Src cells were fractionated into Triton X-100-soluble
and -insoluble fractions as described previously (24). Protein
equivalent aliquots of each fraction were immunoprecipitated with an
antiserum against p120cas, and the immunoprecipitates as well
as aliquots of total protein were analyzed by SDS-PAGE and Western
immunoblotting as indicated.
Affinity Precipitation
For affinity precipitation, clarified cell extracts were incubated with Sepharose beads (Pharmacia) to which either control GST protein or a GST-E-cadherin fusion protein was bound. After incubation at 4 °C for 1 h the Sepharose beads were collected by centrifugation in a microcentrifuge, washed three times with immunoprecipitation buffer, and solubilized in Laemmli sample buffer for SDS-PAGE and Western immunoblot analysis. The GST fusion protein for E-cadherin consists of bacterial glutathione S-transferase fused in frame to the entire cytoplasmic domain of E-cadherin (35). The GST-E-cadherin fusion protein and the control GST protein were purified from bacterial lysates.
Pulse-Chase AnalysisFor the pulse-chase analysis shown in
Fig. 2, replicate cultures of the indicated cell lines were incubated
in the absence of methionine for 20 min, pulse-labeled for 30 min with
250 µCi/ml [35S]methionine, washed twice, and incubated
in medium containing excess unlabeled methionine for 0, 0.5, 1.0, 2.0, and 4.0 h. At each indicated time point, the cells were extracted
with immunoprecipitation buffer and equivalent aliquots of each extract
were immunoprecipitated with an antibody against -catenin.
Gel Electrophoresis and Western Immunoblot Analysis
For all experiments the washed immune complex pellets were boiled for 5 min in Laemmli sample buffer, divided as appropriate to generate duplicate gels for Western immunoblot analysis, and separated in a 10% SDS-polyacrylamide gel. Gels consisting of 35S-labeled samples were analyzed by fluorography using Amplify as directed by the manufacturer (Amersham Corp.), and gels with non-radioactive samples were electrophoretically transferred to Immobilon-P filter paper (Millipore Corp.) for Western immunoblot analysis.
For Western immunoblots, filters were first incubated in a blocking solution consisting of Tris-buffered saline solution containing 0.1% Tween 20 and either 5% bovine serum albumin fraction V (for phosphotyrosine immunoblots) or 5% nonfat dry milk (for all other immunoblots). Filters were next incubated with antibodies, added in blocking solution at a 1:1000 dilution, for 2 h at room temperature. The filters were washed for 1 h with multiple changes of Tris-buffered saline plus 0.1% Tween-20 and subsequently incubated for 30 min with blocking solution containing either horseradish peroxidase-conjugated protein A (Amersham) for polyclonal antisera or horseradish peroxidase-conjugated anti-mouse imunoglobulin (Amersham) for monoclonal antibodies. The filters were washed as above and developed using an enhanced chemiluminescence (ECL) detection system (Amersham).
L cells have previously been used for studies of
cadherin function since they normally lack detectable expression of
classic cadherins and fail to participate in
calcium-dependent cell to cell adhesive contacts (60).
However, expression of exogenous cadherins in these cells leads to
calcium-dependent cell to cell adhesion (60). For this
study, L cells were transfected with either an expression vector for
E-cadherin or with a control vector conferring only neomycin
resistance. Individual transfected cell clones were selected and
analyzed for E-cadherin expression by Western immunoblot with a
cadherin-specific antibody. As in previous studies, no expression of
cadherin was detected in the parental L cell line, whereas the
individual E-cadherin transfectants expressed substantial levels of
E-cadherin protein (60) (Fig. 1). Steady-state expression levels of catenins in the transfected cells were next examined by Western immunoblot using antibodies specific for each of
the catenins. Very low levels of both -catenin and
-catenin were
observed in the parental L cells and the control transfectants. In
contrast, all of the cell lines expressing E-cadherin showed markedly
increased steady-state levels of both
-catenin and
-catenin (Fig.
1). Both the
-catenin and
-catenin proteins were detected in a
complex with E-cadherin (Fig. 1). Unlike
- and
-catenin, little
or no plakoglobin (
-catenin) protein expression was detected in
either the parental L cells or in the E-cadherin-expressing cell lines
(data not shown). In addition to the prototype catenins (
,
, and
), the steady-state levels of p120cas, a Src substrate that
has recently been included in the armadillo family along
with
- and
-catenin (13, 16), were also examined. Fig. 1 shows
that p120cas protein is readily detected in the parental L
cells and the control transfectants, but in contrast to
- and
-catenin, the steady-state level of p120cas remains
essentially unchanged upon introduction of E-cadherin. Recent studies
have shown that p120cas can associate with cadherin (18-21).
Similarly, p120cas was detected in complex with E-cadherin in L
cells (Fig. 1); however, in contrast to results with
- and
-catenin, very little p120cas appears to associate with
E-cadherin (see below). These results indicate that the steady-state
levels of
- and
-catenin, which are normally very low in L cells,
can be significantly increased upon introduction of E-cadherin
expression. One clonal line of E-cadherin-transfected L cells, ECad2-2,
was used for subsequent analysis.
In order to understand the mechanism leading to
increased steady-state -catenin levels in response to E-cadherin
expression, the half-life of
-catenin was examined in a pulse-chase
experiment. Replicate dishes of parental L cells and ECad2-2 cells were
briefly pulse-labeled with [35S]methionine followed by a
chase in the absence of label. At each indicated time point, cells were
harvested and extracts were immunoprecipitated with a
-catenin-specific antibody. Fig. 2 shows that the
half-life of
-catenin in the parental L cells is very short, with
most of the newly synthesized protein gone within 30 min. In contrast, the half-life of
-catenin is dramatically increased in the presence of E-cadherin, with most of the protein still remaining after 4 h
of chase.
We
previously showed that Wnt-1 expression significantly prolongs the
half-life of -catenin in AtT20 and C57MG cells (24), primarily by
increasing the stability of uncomplexed, monomeric pools of this
protein (35). In order to compare the effects of Wnt-1 and E-cadherin
on steady-state levels as well as free pools of both
-catenin and
p120cas, parental L cells were transfected either with an
expression vector for Wnt-1 or with an expression vector for
E-cadherin. Two days following transfection, the steady-state levels of
both
-catenin and p120cas were examined by Western
immunoblot analysis of equivalent amounts of cell extract. Fig.
3 shows that both Wnt-1 and E-cadherin are independently
capable of inducing an increase in the steady-state level of
-catenin in L cells, illustrated by Western blot analysis of total
cell extracts. The steady-state levels of p120cas were not
affected by either Wnt-1 or E-cadherin expression. To examine whether
the increased levels of
-catenin localized to an uncomplexed
population, the free pool of
-catenin was selectively isolated from
equivalent amounts of detergent cell extract by affinity precipitation
using a purified GST-fusion protein containing the cytoplasmic domain
of E-cadherin. We have previously shown that this method allows the
identification of free pools of catenins which are available for
association with the exogenous binding site, while the endogenous
catenin molecules that are already in a complex with APC or cadherin
are not selected by this affinity precipitation protocol (35). As shown
in Fig. 3, the GST-cadherin fusion protein identifies a free pool of
-catenin in the Wnt-1-transfected L cells, whereas no free pool is
detected in the parental L cells or the E-cadherin transfectants. In
contrast, an equivalent amount of uncomplexed p120cas can be
isolated with the GST-cadherin fusion protein from extracts of both the
Wnt-1 and E-cadherin transfectants as well as the parental L cells.
Expression of v-Src in ECad2-2 Cells
Previous studies have shown that v-Src transformation leads to tyrosine phosphorylation of cadherin and catenins, and it has been suggested that this contributes to the concomitant loss of cell to cell adhesion observed in these v-Src-transformed cells (46-48). In order to examine in biochemical detail the consequences of tyrosine phosphorylation on cadherin-catenin complex formation and free pools of catenins, the ECad2-2 cells were transfected with an expression vector for v-Src and stable cell lines were selected. The v-Src-expressing cells showed morphological alterations characteristic of transformation by this oncogene (Fig. 4A). The biochemical analysis in Fig. 4B shows that v-Src is readily detected in the transfected cells, and these cells have elevated tyrosine phosphorylation of cell proteins as expected. Either of two clones of v-Src-expressing ECad2-2 cells, ECad/Src 6 or ECad/Src 20, were used for further analysis.
Phosphorylation of E-cadherin andIn agreement with
previous studies (46-48), immunoprecipitation of either -catenin or
E-cadherin followed by immunoblot analysis with a
phosphotyrosine-specific antibody shows that both of these proteins
become phosphorylated on tyrosine in the ECad/Src cells but not in the
ECad2-2 cells (Fig. 5A). Tyrosine
phosphorylation of
-catenin was observed both directly by
immunoprecipitation with a
-catenin-specific antibody and indirectly
by immunoprecipitating
-catenin in complex with E-cadherin using a
cadherin-specific antibody. No phosphorylation of
-catenin was
observed in either the ECad2-2 or ECad/Src cells (data not shown). Fig.
5B illustrates that tyrosine phosphorylation of E-cadherin
and
-catenin in the ECad/Src cells does not change the level or
stoichiometry of the E-cadherin complex with
-catenin or
-catenin. To demonstrate this point, protein equivalent aliquots of
both ECad2-2 and ECad/Src cells were compared by immunoprecipitation of
E-cadherin, followed by Western immunoblot analysis for E-cadherin,
-catenin, or
-catenin. These data show that the overall level as
well as the ratio of E-cadherin to
-catenin are unchanged in the
presence of v-Src (Fig. 5B). Although
-catenin did not
become phosphorylated on tyrosine in the v-Src-transformed cells, I
determined whether tyrosine phosphorylation of
-catenin or
E-cadherin in these cells led indirectly to alterations in the ability
of
-catenin to associate with
-catenin and thus integrate into
the cadherin-catenin complex. The Western immunoblot shown in Fig.
5B demonstrates that the amount of
-catenin in
association with E-cadherin, presumably through binding to
-catenin
is unchanged in response to v-Src transformation. Analysis of
steady-state levels of E-cadherin,
-catenin, and
-catenin showed
no change in the presence of v-Src (data not shown). Pulse-chase
analysis further showed that tyrosine phosphorylation does not affect
the half-life of
-catenin (data not shown).
Phosphorylation of p120cas by v-Src: No Effect on Steady-state Levels
p120cas, an armadillo
family member, was originally identified as an abundant substrate
for tyrosine phosphorylation in v-Src-transformed cells (13). Here I
have examined whether this phosphorylation alters the steady-state
level of p120cas or its ability to complex with cadherin.
Immunoprecipitation of p120cas followed by Western immunoblot
with a phosphotyrosine-specific antibody illustrates that
p120cas is heavily phosphorylated on tyrosine in the ECad/Src
cells but not in the ECad2-2 cells (Fig. 6).
Phosphorylation of p120cas on tyrosine has no effect on its
steady-state level, illustrated both by examination of p120cas
levels in total cell extracts and by immunoprecipitation of
p120cas followed by Western immunoblot analysis for
p120cas protein (Fig. 6). The amount of p120cas
identified in association with E-cadherin is very small by comparison to the overall pool, but there is no apparent change in this population (Fig. 6).
Tyrosine Phosphorylation Has No Effect on the Uncomplexed Pool of
To examine potential effects of tyrosine
phosphorylation on free pools of -catenin and p120cas, an
affinity precipitation experiment was performed using the GST-cadherin
protein followed by Western immunoblotting for either
-catenin or
p120cas. Very little uncomplexed
-catenin can be detected in
ECad2-2 cells, requiring extended exposure times of the Western
immunoblot, and this is unchanged by tyrosine phosphorylation upon
v-Src transformation (Fig. 7). On the other hand, a
substantial free population of p120cas can be identified in the
ECad2-2 cells, and this is decreased significantly in the ECad/Src
cells (Fig. 7). Also of note is the finding that
tyrosine-phosphorylated p120cas can be efficiently isolated
from the ECad/Src cell extract by affinity precipitation with the
GST-cadherin protein (Fig. 7).
Altered Subcellular Fractionation of p120cas upon v-Src Transformation
Since the uncomplexed pool of p120cas is diminished upon v-Src transformation, it was of interest to determine if the subcellular distribution of p120cas was also altered. For this purpose ECad2-2 and ECad/Src cells were fractionated into Triton X-100-soluble and -insoluble fractions, and protein equivalent aliquots of each fraction were immunoprecipitated with an antiserum against p120cas. Western immunoblot analysis of the immunoprecipitates as well as of total protein aliquots from each fraction shows that a small amount of p120cas is detected in a Triton X-100-insoluble fraction from ECad2-2 cells; the majority is in a Triton X-100-soluble fraction (Fig. 8). V-Src transformation leads to a decrease in the Triton X-100-insoluble population of p120cas with a corresponding increase in the Triton X-100-soluble fraction (Fig. 8). Most of the tyrosine-phosphorylated p120cas is localized to a Triton X-100-soluble fraction from ECad/Src cells (Fig. 8).
Cadherin-mediated cell to cell adhesion as well as adhesion-independent functions for catenins have been implicated in the modulation of many aspects of multicellular differentiation, cell proliferation, and malignant transformation of epithelial cells (1-5). In this study I have defined, at a biochemical level, distinct mechanisms leading to changes in steady-state catenin levels. Furthermore, I have characterized the impact of tyrosine phosphorylation on regulation of both free pools of catenins as well as cadherin-catenin complex stability.
An increase in availability of -catenin binding sites, by increasing
the expression level of E-cadherin, leads to an increased stability of
- and
-catenin in complex with cadherin, but notably no increase
in the free pool of
- or
-catenin is observed.
-Catenin associated with cadherin exhibits a greatly extended half-life and
appears to be selectively protected from a rapid degradation mechanism
that operates in normal cells to restrict the availability of an
uncomplexed
-catenin population (35). Similar results were obtained
in another study, where an increase in the steady-state level of
plakoglobin was observed upon co-transfection, into L cells, of
expression vectors for plakoglobin and the desmosomal cadherins,
desmoglein or desmocollin, when compared to expression of the
plakoglobin vector alone (61). The increased steady-state level of
plakoglobin resulted from a 15-20-fold decrease in plakoglobin protein
degradation and correlated with complex formation between plakoglobin
and desmoglein or desmocollin. In our studies no increase in endogenous
plakoglobin levels was detected in response to the expression of
E-cadherin in L cells, despite the fact that plakoglobin can
effectively bind to E-cadherin in other cell types (24, 25). This is
perhaps due to an inherent inability of our L cell line to express
plakoglobin. Taken together these findings suggest that catenin protein
stability can be increased upon co-ordinate interaction with other
proteins that participate in junctional complex formation.
In contrast to the stabilization of - and
-catenin by E-cadherin
expression, p120cas is readily detectable in the parental L
cells lacking cadherin and is not stabilized further upon introduction
of E-cadherin. This is consistent with the fact that very little
p120cas binds to E-cadherin (20, 21) and thus is not likely to
be significantly stabilized by complex formation with an increased availability of cadherin binding sites. Furthermore, in normal cells
cadherin may exhibit a greater affinity for
-catenin than p120cas. Examination of the uncomplexed population of
p120cas in L cells revealed that this pool was not affected by
E-cadherin expression.
A second distinct mechanism leading to increased steady-state
-catenin levels is through an increase in uncomplexed, monomeric pools as a consequence of Wnt-1 signal transduction. As a result of
increasing the free pools, the steady-state levels of both
-catenin
and plakoglobin are increased by Wnt-1 signaling in several cell types
(24, 31, 35), whereas there is no effect of Wnt-1 on
-catenin levels
(24). I have extended this analysis to show that the regulation of
p120cas is not affected by Wnt-1 expression. While
p120cas is considered an armadillo family member,
based on sequence analysis and cadherin binding, the data presented
here indicate that the regulation of p120cas is is independent
of the regulation of
-catenin and plakoglobin since the free pools
are not affected by Wnt-1 expression. Here I also show that the
mechanism utilized by Wnt-1 to increase the free pool of
-catenin
does not require the presence of E-cadherin since this can occur in
parental L cells in the absence of E-cadherin expression. This suggests
that the previously observed stabilization of the
-catenin-cadherin
complex by Wnt-1 must be a secondary consequence of the increased free
pool of
-catenin. The mechanism by which Wnt-1 increases free
catenin pools is not completely defined, but it appears to modulate a
signal transduction pathway that includes APC, in complex with
-catenin and plakoglobin, and the serine-threonine kinase, GSK3,
both of which function in concert to restrict
-catenin levels in
normal cells (32, 35). Since p120cas does not associate with
APC (19), this may explain the lack of effect of Wnt-1 on
p120cas levels. This would also predict that changes in GSK3
activity would not affect p120cas levels.
Several studies have shown that transformation by the v-src
oncogene leads to phosphorylation on tyrosine of E-cadherin and -catenin, and it has been proposed that these phosphorylations could lead to a destabilization of the cadherin-catenin complex, resulting in loss of cell to cell adhesion characteristic of the transformed cell (46-48). Here I have examined directly the stability of the cadherin-catenin complex, the half-life of
-catenin, and the
availability of uncomplexed catenin pools in response to v-Src-induced transformation. Transformation of E-cadherin-expressing L cells with
v-Src results in tyrosine phosphorylation of E-cadherin,
-catenin,
and p120cas but not
-catenin. However, tyrosine
phosphorylation appears to have no effect on the steady-state levels of
these proteins. Tyrosine phosphorylation also has no effect on the
stability or stoichiometry of complexes of E-cadherin with
-catenin,
-catenin, and p120cas. In agreement with our findings, no
effect of v-Src was seen on cadherin complexes with either
-catenin
or p120cas in Madin-Darby canine kidney cells (62).
These observations imply that the morphological transformation and
apparent loss of cell to cell contact induced by v-Src are not a
consequence of disruption of the cadherin-catenin complex itself upon
tyrosine phosphorylation. Although -catenin is not phosphorylated on
tyrosine nor is it dissociated from the cadherin complex, it is
possible that the tyrosine phosphorylation of
-catenin and cadherin
compromises the ability of the associated
-catenin to anchor the
complex to the cell cytoskeleton and stabilize cell to cell adhesion.
Alternatively, since many of the known substrates for v-Src are
cytoskeletal-associated proteins (63), tyrosine phosphorylation of one
of these other proteins may destabilize the interaction of the
cadherin-catenin complex with the cytoskeleton. v-Src transformation
also leads to phosphorylation of components of the cell-matrix adhesion
system such as integrins and the focal adhesion kinase (63). By
phosphorylation of multiple substrates within the cell-cell and
cell-matrix adhesion systems, the striking morphological alterations
characteristic of v-Src-induced transformation may result from a
disruption of the balance between these distinct adhesion systems.
Recent evidence in fact shows that changes in cadherin expression are
reflected in altered expression of genes encoding fibronectin and
integrins and altered cell-matrix adhesion properties (64, 65).
Alternatively, tyrosine phosphorylation of the E-cadherin--catenin
complex may be important for other functions of these proteins aside
from adhesion. For example, tyrosine phosphorylation of the
cadherin-catenin complex may serve to recruit various signal transduction effectors, similar to the consequences of receptor tyrosine kinase phosphorylation upon ligand-induced activation (66).
These effectors could include proteins such as receptor tyrosine
kinases or the tyrosine phosphatase mu, proteins that have been
identified in association with the cadherin-catenin complex (49, 67,
68).
Recent evidence suggests that the uncomplexed populations of catenins
can serve a signal transduction function in response to Wnt-1 and
Wingless proteins (32-34). In this study I have also examined
uncomplexed catenin populations in the v-Src-transformed L cells and
found that in contrast to the effects of Wnt-1, there is no change in
the free pool of -catenin in response to v-Src phosphorylation and
transformation. Conversely, Wnt-1 does not appear to regulate free
pools of p120cas, whereas the free pool of p120cas is
significantly decreased upon v-Src transformation and tyrosine phosphorylation. The tyrosine phosphorylation and corresponding decrease in uncomplexed p120cas is also accompanied by an
increase in the Triton X-100-soluble pool of p120cas at the
expense of the Triton X-100-insoluble population. This implies that
p120cas, upon phosphorylation on tyrosine, may show increased
association with other cytoplasmic effector molecules, perhaps
including the tyrosine kinase FER (69). Furthermore, the altered
detergent solubility of p120cas may suggest an altered
subcellular distribution. These findings support the conclusion that,
at the biochemical level, p120cas is regulated independently
from
-catenin and plakoglobin. Thus, while both Wnt-1 and v-Src
induce transformation-dependent changes in cell
morphologies and modulate uncomplexed catenin pools, the mechanisms are
different.
Cadherins and catenins participate in the modulation of the epithelial phenotype, changes that have profound consequences for tissue differentiation and malignant transformation. Here I have characterized, at the biochemical level, several mechanisms that regulate the free pools of catenins and the complexes that form with cadherin. Our data support a model whereby uncomplexed catenin populations are tightly regulated and serve a cadherin independent signal transduction function in response to both Wnt-1 and v-Src. These studies contribute to a biochemical understanding of the phenotypes produced by oncogenes, growth factors, and their receptors.
I thank James Nelson for the antiserum against E-cadherin, Rolf Kemler for the E-cadherin cDNA clone, and David Shalloway for the v-Src expression vector. I thank John Forsman and James Rice for help with the preparation of the GST fusion proteins and Rhea Pugliese for assistance with the v-Src transfections. I am grateful to Beverly Smolich for critical comments on the manuscript.