(Received for publication, May 21, 1996, and in revised form, December 9, 1996)
From the Department of Biochemistry, Faculty of Medicine, Kagoshima University, 8-35-1 Sakuragaoka, Kagoshima 890, Japan
-Catenin is a 102-kDa protein exhibiting
homology to vincuin, and it forms complexes with cadherins or the
tumor-suppressor gene product adenomatous polyposis coli through
binding to
-catenin or plakoglobin (
-catenin). The incorporation
of
-catenin into the cadherin-catenin complexes is a prerequisite
for expression of the cell-adhesive activity of cadherins. Using an
in vitro assay system involving bacterially expressed
proteins, we localized a region in
-catenin required for molecular
interaction with
-catenin and plakoglobin. Analysis of various
truncated
-catenin molecules revealed that amino-terminal residues
48-163 are able to bind to
-catenin and plakoglobin. Consistent
with the observation that
-catenin and plakoglobin bind to the same
region of
-catenin,
-catenin competed with the binding of
plakoglobin to
-catenin and vice versa. Under the conditions used,
-catenin bound to
-catenin with higher affinity than did
plakoglobin. Scatchard analysis indicated that the affinity of the
interaction between
-catenin and
-catenin or that between
-catenin and plakoglobin was moderately strong
(Kd = 3.8 × 10
8 and 7.7 × 10
8, respectively). When transfected into L cells
expressing E-cadherin, the amino-terminal region of
-catenin (from
residue 1 to 226) formed complexes with
-catenin supporting the
in vitro binding experiment results.
Cadherins are a major group of calcium-dependent
cell-cell adhesion molecules that bind through a homophilic mechanism
and that are localized to specialized intercellular junctions called adherens junctions (1, 2). The cadherins are transmembrane proteins
possessing an extracellular calcium-binding segment and an
intracellular domain that is highly conserved (90% identity) among
most members of the family. Deletion of the conserved intracellular segment results in cadherin inactivation even if the extracellular binding domain seems to remain intact (3). The cytoplasmic domain of
cadherins interacts with three molecules termed catenins (
,
, and
), and the resultant complexes seem to associate with cortical actin
filaments (4). This interaction between cadherins and catenins is
essential for cadherin-mediated adhesion and the association of the
complexes with the cytoskeleton (5-7).
Two of the catenins have been cloned. -Catenin is homologous to
vinculin, a cytoskeleton-associated protein (8, 9).
-Catenin is
homologous to plakoglobin (a protein found in both adherens junctions
and desmosomes) and Armadillo (a segment polarity gene product in
Drosophila melanogaster (10-13)). Immunological data
suggest that
-catenin is identical to plakoglobin (14, 15).
Recent in vitro and in vivo experiments have
shown that -catenin and plakoglobin bind directly to the cytoplasmic
domain of E-cadherin, while
-catenin binds directly to
-catenin
or plakoglobin (16-20). The amino-terminal parts of
-catenin and plakoglobin have been shown to comprise their
-catenin-binding sites, and the central core region, which is composed of 13 copies of
the so-called Armadillo repeat, is involved in the association with
cadherins. The latter is also involved in the complex formation with
the adenomatous polyposis coli tumor-suppressor protein and in the case
of plakoglobin, with desmogleins (desmosomal cadherins). The region of
-catenin responsible for
-catenin and plakoglobin binding,
however, had not been identified. In this study, we report experiments
that revealed the region in human
-catenin responsible for the
binding of
-catenin and plakoglobin.
A full-length cDNA clone for human
-catenin has been described (20). To express
-catenin as a fusion
protein with the maltose-binding protein
(MBP)1 in Escherichia coli
cells, cDNA encoding the protein was cloned into an MBP fusion
vector (pMALc, New England Biolabs Inc.). cDNA fragments encoding
various regions of
-catenin were generated by using convenient
restriction enzyme sites within the cDNA clones or by means of the
polymerase chain reaction. The combinations of restriction enzymes used
were: BglII and SalI, BglII and
HindIII, MluI and SalI,
BglII and StuI, BglII and
SphI, BglII and XhoI, BglII
and ApaI, XbaI and SphI,
ApaI and SphI, and XhoI and
SphI. For the polymerase chain reaction, three sense primers
(5
GAAGATCTTCTAATAAGAAGAGAGG, 5
GAAGATCTAAAATTGCGAAGGAG, and
5
GAAGATCTGAGTTCGCAGATGAT) and two antisense primers
(5
GAAAGCTTCAAGATACCATCTTC and 5
GAAAGCTTGCCAACATCTTTCAA) containing a
BglII or HindIII recognition sequence at the
5
-end, respectively, were used. The reaction mixture was subjected to 30 cycles of denaturation (93 °C, 1 min), annealing (45 °C, 1 min), and extension (72 °C, 1 min). The cDNA fragments were
subcloned in frame into the vector, and the plasmid DNAs were
introduced into JM109 cells. MBP fusion proteins were purified by
affinity chromatography on columns of amylose resin (New England
Biolabs Inc.) as described previously (20).
The cDNA clones for human
plakoglobin and -catenin have been described (20, 21). The
full-length
-catenin cDNA with a BamHI 5
-3
SalI orientation in Bluescript II KS(+) vector was excised
by digestion with BamHI and SalI and then cloned
into the BamHI/SalI sites of the glutathione
S-transferase (GST) fusion protein (pGEX-4T3, Pharmacia
Biotech Inc.) vectors. The GST fusion protein vectors containing
cDNA for the entire coding region of plakoglobin have been
described (20).
The binding of GST fusion
proteins to MBP fusion proteins was visualized as follows. Purified MBP
fusion proteins (50 ng) were subjected to sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then
electroblotted onto nitrocellulose filters as described previously
(20). The filters were incubated in phosphate-buffered saline
containing 5% nonfat dried milk for 30 min and with GST fusion
proteins (100 µg/ml) for 2 h. The filters were then washed with
phosphate-buffered saline containing 0.05% Tween 20. The MBP fusion
proteins bound to proteins on the filters were detected by incubation
with affinity-purified anti-GST antibodies followed by horseradish
peroxidase-labeled F(ab)2 fragments of goat anti-rabbit
IgG (Jackson ImmunoResearch Laboratories, Inc.) and 4-chloro-1-naphthol
as a substrate. For competitive binding assay of
-catenin and
plakoglobin, affinity-purified antibodies against
-catenin or
plakoglobin were used to detect the respective fusion protein bound to
-catenin. To quantify the binding, the peroxidase-labeled antibody
was replaced by the 125I-labeled F(ab)2
fragment of goat anti-rabbit IgG (7.5 µCi/µg, DuPont NEN), and
bound radioactivity was determined using a scintillation counter.
Rabbit antibodies against GST were prepared and
purified as described previously (20). A monoclonal antibody against
GST was purchased from Santa Cruz Biotechnology. Monospecific
antibodies against the carboxyl-terminal part of plakoglobin or the
full-length -catenin have been described previously (20, 21). A
monoclonal antibody (12CA5) directed against hemagglutinin (HA) was
kindly provided by Dr. A. Yoshimura (Kurume University, Fukuoka,
Japan).
The -catenin fusion proteins (10 µg)
containing residues 1-906 (the entire protein, MBP/
-CN-(1-906) or
residues 48-163 (MBP/
-CN-(48-163) were electroblotted onto
nitrocellulose filters after SDS-PAGE. Each filter was incubated with
varying concentrations (0-20 µg/ml) of GST/
-catenin or
GST/plakoglobin, and the bound GST/
-catenin or GST/plakoglobin was
detected by incubation with monoclonal anti-GST antibodies followed by
the horseradish peroxidase-labeled F(ab
)2 fragment of
rabbit anti-mouse IgG. After washing, each filter was incubated for 2 min with orthophenylenediamine (0.4 mg/ml) in 0.1 M citric
phosphate buffer, pH 5.0, containing 0.012% H2O2. The peroxidase reaction was stopped with
2 M sulfuric acid. The reaction was quantified by measuring
the optical density at 492 nm.
-Catenin cDNA encoding the amino-terminal region
(from residue 1 to 226) was subcloned in frame into a vector (BX-C-Flu, a gift from Dr. Yoshimura) containing a 9-amino acid HA sequence (YPYDVPDYA) using an oligonucleotide (5
AATTCGATATCGGC) as a linker. Thus the cDNA codes for truncated
-catenin (residues 1-226) and a sequence (EFDIA) followed by the HA tag sequence. The cDNA was cloned into the mammalian expression vector pCAGGS neo (a gift from Dr.
K. Yamamura, Kumamoto University, Kumamoto, Japan), which contains an
enhancer derived from cytomegalovirus and the
-actin promoter (22)
as described previously (21). A mouse L cell line expressing E-cadherin
(EL-8) was established as described previously (4) and used for
transient expression experiments. EL-8 cells (1 × 106) were transfected with an expression vector
(pC-
CN1-226) by electroporation using a Bio-Rad Gene Pulser set at
700 V and 25 microfarads. After 48 h, the transfected cells were
analyzed as described below. To isolate stable transfectants,
Madin-Darby canine kidney cells (1 × 106) were
transfected with pC-
CN1-226 as described above except that the
voltage was set at 600 V. Stable transfectants were selected and cloned
as described before (4). The cells were lysed in 10 mM
Tris-HCl buffer, pH 7.5, containing 0.5% Nonidet P-40, 1 mM EDTA, and 1 mM phenylmethylsulfonyl
fluoride. The truncated
-catenin with the HA tag was collected with
the anti-HA monoclonal antibody 12CA5, which had been preabsorbed to
protein A-Sepharose CL-4B. The immune complex was washed with a washing
buffer (10 mM Tris-HCl, pH 7.5, 0.5% Nonidet P-40, 0.5 M NaCl, 1 mM phenylmethylsulfonyl fluoride).
To localize the sites for -catenin
and plakoglobin binding in
-catenin, different regions of the
-catenin molecule (Fig. 1) were expressed as
maltose-binding protein (MBP) fusion proteins in E. coli and
then purified to homogeneity by affinity chromatography on amylose
resin. Each fusion protein migrated on an SDS-PAGE gel as a band
corresponding to the expected size calculated from the sequence (Fig.
2, A and C). The ability of the
-catenin fusion proteins to bind to
-catenin was assessed using a
blot overlay assay as described under "Materials and Methods."
As shown in Fig. 2B, GST/-catenin bound to the
-catenin fusion proteins containing residues 1-906 (the entire
protein, MBP/
-CN-(1-906) and residues 1-449 (the amino-terminal
half of the protein, MBP/
-CN-(1-449), but it did not bind to a
protein spanning residues 326-906 (the carboxyl-terminal two-thirds of
the protein, MBP/
-CN-(326-906). The binding appeared to be specific
for these portions of the MBP fusion proteins as no binding was
observed between GST/
-catenin and MBP alone. Furthermore, GST itself
had no ability to bind to MBP/
-CN-(1-906) or MBP/
-CN-(1-449)
(data not shown), indicating that the
-catenin portion of
GST/
-catenin is responsible for the binding between GST/
-catenin
and the MBP fusion proteins. The binding of GST/
-catenin to the MBP
fusion proteins was abolished if GST/
-catenin was boiled for 5 min
immediately before the blot overlaying was performed (data not shown).
The results indicate that
-catenin binding to
-catenin is limited
to the amino-terminal 449 residues of the
-catenin molecule.
To further define the -catenin-binding site within the
amino-terminal 449 residues of the
-catenin molecule, we expressed a
series of
-catenin fusion proteins with deletions from the amino
terminus or from the carboxyl terminus of MBP/
-CN-(1-449). These
proteins were then examined as to their ability to bind to
-catenin.
Although the
-catenin fusion proteins containing residues 48-163
exhibited such an ability, those with deletions in residues 48-163 did
not (Fig. 2D). These results suggest that the
-catenin-binding site in
-catenin is localized within residues 48-163 (as summarized and shown in Fig. 1).
The high degree of sequence identity (82%) between
the -catenin-binding site of
-catenin and that of plakoglobin
(20) suggested that the binding site of
-catenin for
-catenin and plakoglobin is the same. To determine whether this is the case or not,
we analyzed the binding of the plakoglobin fusion protein with GST
(GST/plakoglobin) to the
-catenin fusion proteins as above. As in
the case of GST/
-catenin binding, GST/plakoglobin bound to the
-catenin fusion proteins containing residues 48-163 but not to the
-catenin fusion proteins without these residues (Fig.
2E). These results suggest that the plakoglobin-binding site
in
-catenin is also localized within residues 48-163.
To confirm that -catenin and plakoglobin bind to the same region of
-catenin, a competition experiment was carried out. When
GST/plakoglobin was included on the incubation of GST/
-catenin with
an
-catenin fusion protein, the binding of GST/
-catenin to the
-catenin fusion protein decreased with increasing amounts of
GST/plakoglobin and vice versa (Fig. 3). The binding of
GST/plakoglobin, however, seems to be more sensitive to the presence of
the competitor. The presence of an approximately 12-fold molar excess
of GST/plakoglobin reduced the binding of GST/
-catenin, but the
presence of an approximately 4-fold molar excess of GST/plakoglobin did
not (Fig. 3A). The binding of GST/plakoglobin to an
-catenin fusion protein was significantly reduced by the presence of
a 4-fold molar excess of GST/
-catenin and almost completely
inhibited by the presence of a 12-fold molar excess of GST/
-catenin
(Fig. 3B). These results may suggest that
-catenin has a
higher affinity to
-catenin than plakoglobin.
To better characterize the interaction of -catenin or plakoglobin
with
-catenin, Scatchard analysis was carried out (Fig. 4). Increasing concentrations of GST/
-catenin or
GST/plakoglobin were incubated with MBP/
-catenin (residues 1-906)
fixed on nitrocellulose membranes followed by quantification of the
amounts of the bound GST/
-catenin or GST/plakoglobin. The results
demonstrated that GST/
-catenin bound to
-catenin with a
dissociation constant (Kd) of 3.8 × 10
8 M and that GST/plakoglobin bound to
-catenin with a Kd of 7.7 × 10
8 M. When the truncated
-catenin with
residues 48-163 was subjected to the same analysis,
Kd values of 1.5 × 10
8
M and 5.6 × 10
8 M were
obtained for GST/
-catenin and GST/plakoglobin, respectively.
Residues 1-220 of
To prove that the results obtained in in
vitro binding experiments reflect the activity of -catenin
in vivo, we expressed the amino-terminal region of
-catenin in living cells. We constructed an expression vector
containing a cDNA encoding residues 1-226 of
-catenin and a
sequence of an epitope of HA. The presence of the HA tag at the
carboxyl terminus of the truncated
-catenin enabled us to collect
the expressed protein by immunoprecipitation. An L cell line expressing
E-cadherin (and thus also expressing
-catenin) was transiently
transfected with an expression vector, and the expressed
-catenin
was collected with the anti-HA monoclonal antibodies to determine
whether or not it could bind to
-catenin. After SDS-PAGE and
transfer to a nitrocellulose membrane, the collected materials were
probed with antibodies to
-catenin. As shown in Fig.
5,
-catenin was precipitated by anti-HA antibodies from the cells transfected with the expression vector but not from the
cells transfected with a control vector. Therefore, the truncated
-catenin containing amino-terminal residues 1-226 was able to bind
to
-catenin. To estimate the amount of
-catenin bound to the
mutant
-catenin in cells, Madin-Darby canine kidney cells were
transfected with the same vector, and stable transfectants were
isolated. The amounts of
-catenin co-precipitated with the truncated
-catenin were compared with those of
-catenin in the total cell
lysates. Quantitative analysis revealed that about 5% of
-catenin
in the lysates was associated with the mutant
-catenin (data not
shown).
Using an in vitro binding system, we analyzed the
molecular interaction of -catenin with
-catenin and plakoglobin
and localized the binding site for the latter two molecules in
-catenin. Furthermore, by expressing the amino-terminal region of
-catenin in living cells, we obtained evidence that the region can
bind to
-catenin in vivo. During the preparation of this
manuscript, we noticed a report that the amino-terminal 606 amino acids
of
-catenin bind to
-catenin in a yeast two-hybrid system (23).
Our results indicating that amino-terminal residues 48-163 of
-catenin bind to
-catenin are consistent with this observation
and also further extend it. Furthermore, we showed that plakoglobin and
-catenin bind to the same region in
-catenin. Scatchard analysis
revealed that
-catenin binds to
-catenin with higher affinity
(Kd = 3.8 × 10
8 M)
than plakoglobin (Kd = 7.7 × 10
8
M).
cDNA cloning revealed that -catenin is a vinculin-related
protein. Vinculin is a cytoskeletal protein associated with both cell-cell and cell-extracellular matrix adherens-type junctions (24,
25). The homology between vinculin and
-catenin is restricted to
three major regions in the amino-terminal, central, and
carboxyl-terminal parts of the two proteins. Vinculin has been shown to
bind to F-actin (26) and talin (27). The latter protein can bind
directly to the cytoplasmic domain of
1 integrins (28,
29), which are members of the integrin superfamily of transmembrane
heterodimeric glycoproteins. The actin-binding site of vinculin has
been localized to the carboxyl-terminal region (26), whereas the
talin-binding site is in the amino-terminal region of residues 1-258
(30). Therefore, we found similarities between
-catenin and vinculin not only in the primary sequence but also in the position of the region
used for hierarchical molecular interactions in cell adhesion molecule
complexes. The amino-terminal region of both proteins is involved in
the interaction with a molecule (
-catenin in the case of
-catenin
and talin in the case of vinculin) that binds directly to the cell
adhesion molecules (cadherins and integrins, respectively). The 116 amino acid residues identified in human
-catenin as the
-catenin-
and plakoglobin-binding site in this study, however, show only a 19.8%
identity to the sequence of human vinculin. Therefore, it is not
surprising that these two proteins interact with distinct proteins
despite their overall structural similarity.
The alternative splicing of an mRNA primary transcript is a
widespread means of generating structurally and functionally distinct protein isoforms and contributes to tissue-specific and developmentally regulated patterns of gene expression. The presence of splice variants
of -catenin and
N-catenin (a neural form of
-catenin) has been
reported (31, 32), although their biological activities have not been
determined. The variant forms included a 24- or 48-amino acid insertion
in their carboxyl-terminal regions, respectively. The present study
suggests that these variant forms have similar, if not identical,
abilities to bind to
-catenin and plakoglobin.
We thank Drs. Akihiko Yoshimura and Ken-ichi Yamamura for the reagents, Drs. Jun-ichiro Tsutsui and Tomo Hashiguchi for suggestions, and Kumiko Sato for secretarial assistance.