| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 32, 30083-30090, August 8, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶ ||
From the
Departments of Biochemistry,
Medicine, and
¶Oncology, Molecular Oncology Group, McGill
University Health Centre, Montreal, Quebec H3A 1A1, Canada
Received for publication, March 17, 2003 , and in revised form, May 19, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Gab1 (Grb2-associated binder 1) was the first mammalian gab gene cloned and was originally identified as a Grb2-binding protein in an interaction screen using a cDNA library prepared from glioblastoma (5). Gab1 is widely expressed and is phosphorylated downstream from numerous receptor tyrosine kinases, cytokine receptors, G protein-coupled receptors, and antigen receptors. In vivo, Gab1 is critical for embryonic development, because gab1-deficient mice die in utero displaying defects in the heart, placenta, and skin, as well as reduced liver size (6, 7). In vitro, Gab1 promotes cell survival, neurite outgrowth, and DNA synthesis in neuronal cells downstream from the TrkA receptor (8, 9) and regulates an invasive epithelial morphogenic program downstream from the HGF/Met receptor tyrosine kinase (1, 10, 11).
In epithelial cells, Gab1 is the major substrate for the Met receptor
tyrosine kinase (12), and,
upon tyrosine phosphorylation, Gab1 provides binding sites for proteins
involved in signal transduction, including the tyrosine phosphatase, SHP-2,
the p85 subunit of PI3'K, phospholipase C 
, as well as the Crk
adaptor protein (10,
11,
13–15).
The association of Gab1 with several of these proteins, as well as an intact
Gab1 pleckstrin homology domain, is required for the ability of Gab1 to
promote the morphogenic program of Madin-Darby canine kidney epithelial cells
downstream from the Met receptor
(1,
10,
11,
16). In contrast, the Gab2
protein fails to promote a morphogenic response, indicating that Gab1 and Gab2
are functionally distinct downstream from the Met receptor
(15,
17).
Gab1 and Gab2 contain highly conserved, but atypical binding sites for the
C-terminal SH3 domain of the adapter protein Grb2, and are recruited
indirectly via a Grb2-dependent interaction to the epidermal growth factor
receptor, the fibroblast growth factor receptor-1, and the interleukin-3
receptor beta common chain (
c)
(2,
18–21).
Whereas the recruitment of Gab2 to the Met receptor is strictly Grb2-dependent
(29,
54), the interaction between
Gab1 and the Met receptor is distinct, in that Gab1 is recruited by both a
Grb2-dependent and Grb2-independent mechanism
(12,
22). The Grb2-dependent
recruitment requires tyrosine 1356 in the Met receptor C terminus, which forms
a Grb2 SH2 domain binding site
(12), and intact Grb2 SH3
domain binding sites on Gab1
(18). The Grb2-independent
recruitment of Gab1 requires an 83-amino acid, proline-rich region of Gab1,
termed the Met binding domain (MBD). This interaction was initially identified
in a yeast-two hybrid interaction assay and requires the kinase activity of
the Met receptor and phosphorylation of tyrosine 1349 in the Met receptor C
terminus (22). Together,
tyrosines 1349 and 1356 are required for the full recruitment of Gab1 to the
Met receptor (12).
Based on the observation that docking proteins of the IRS, Dok, and FRS2 families all contain phosphotyrosine binding (PTB)-like domains, the Gab1 MBD has been proposed to be a PTB-like domain that would bind directly to a phosphotyrosine-containing motif involving Tyr-1349 in the Met receptor C terminus (22). However, the mechanism of interaction between the Gab1 MBD and Met has not been addressed.
Previous studies had identified a 13-amino acid sequence, GMQVPPPAHMGFR, within the MBD, as critical for the MBD-Met interaction (15, 17). We show that this sequence is sufficient for interaction with the Met receptor, indicating that it is unlikely that the 83-amino acid Gab1 MBD forms a structured PTB-like domain. Moreover, we show that the structural integrity of the Met receptor, and residues upstream of tyrosine 1349 located in the C-terminal lobe of the kinase domain, are required for the Grb2-independent interaction with the Gab1 MBD. These results support the interpretation that the Gab1 MBD interacts with the Met receptor in a novel and previously unsuspected manner, where, instead of the expected interaction of a phosphotyrosine binding domain in Gab1 with a phosphotyrosine-containing motif in the Met receptor, we propose that the activated kinase domain of Met and the negative charge of phosphotyrosine 1349 engage the Gab1 MBD as an extended peptide ligand.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Grb2, and MBDGrb2-myc was previously described
(18). GST-MBD16 and GST-MBD16
R499A were generated by insertion of double-stranded oligonucleotides into
BamHI/EcoRI sites of pGEX 2TK; MBD16-F,
GATCCTTTGGAATGCAAGTACCTCCTCCTGCTCATATGGGCTTTAGAAGTTG; MBD16-R,
AATTCAACTTCTAAAGCCCATATGAGCAGGAGGAGGTACTTGCATTCCAAAG. Alanine-scanning
mutations in Tpr Met-PXM were generated via the QuikChange mutagenesis
protocol (Stratagene). Met C-terminal truncation mutants were generated by
insertion of stop codons at amino acids 1348 and 1353, respectively. Tpr Met A
and B mutants were p85 and Shc binding variants, respectively, and are
described in a previous study
(23). All mutants were
sequenced prior to use. Cell Culture, DNA Transfections, and Total Cell Extracts—293T cells were seeded at 1 x 106/100-mm Petri dish and transfected 24 h later by the calcium phosphate precipitation method (24) with 2 µg of DNA, whereas 293 cells were transfected with 10 µg of DNA by the SuperFect method (Qiagen). Cells were serum-starved in 0.1% fetal bovine serum for 24 h and harvested in either 0.5% Triton X-100 lysis buffer for co-immunoprecipitation assays (0.5% Triton X-100, 50 mM HEPES, pH 8.0, 150 mM NaCl, 10% glycerol, 2 mM EGTA, 1.5 mM MgCl2) or in RIPA buffer for GST-pull-down assays (0.05% SDS, 50 mM Tris, pH 8, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium desoxycholate). Inhibitors (10 µg of aprotinin/ml, 10 µg of leupeptin/ml, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium fluoride, and 1 mM sodium vanadate) were included in each lysis buffer. Following a 10-min incubation on ice, the lysates were centrifuged at 13,000 rpm for 10 min at 4 °C.
GST Fusion Proteins, in Vitro Association Assays, Immunoprecipitations,
and Western Blotting—Fusion proteins were produced in the
DH5
, or BL21 gold Escherichia coli strain, by induction with
isopropyl-1-thio--D-galactopyranoside. GST fusion proteins
(0.5–1 µg) were immobilized on glutathione-Sepharose beads for 30 min
at room temperature and washed three times prior to a 1-h incubation with the
indicated cell lysates (lysed in RIPA buffer). For some experiments, the
following peptides were added at the indicated concentrations: Gab1 MBD16
(FGMQVPPPAHMGFRSS) or Gab1 MBD16 R499A (FGMQVPPPAHMGFASS). The peptides were
synthesized at the Sheldon Biotechnology Center (Montreal, Quebec, Canada).
Immunoprecipitation and Western blotting were performed as previously
described (10).
Far Western Assay—Lysates prepared from 293T cells transiently expressing Tpr Met mutants (300 µg) were immunoprecipitated with anti-Met 144 antibody, loaded on an 8% polyacrylamide gel, transferred to membrane, and blocked overnight in 10% milk in TBST containing 1 mM sodium vanadate. GST-Grb2 and GST-MBD were eluted from glutathione-Sepharose beads with reduced glutathione, and the concentration was determined by Bradford assay. Purified protein (20 µg) was incubated with 2 µl of glutathione-conjugated horseradish peroxidase (G-6400, Sigma) for 30 min at room temperature. Washed membranes were then incubated with this reaction mixture in 10 ml of TBST for 30 min, prior to washing and ECL. Membranes were stripped and Western-blotted with anti-Met 144 antibody. The protocol is described in a previous study (25).
Antibodies and Reagents—Antibodies against the extreme C terminus (144 Met (26)) and the extracellular domain (DL-21, Upstate Biotechnology Inc., Lake Placid, NY) of human Met protein were used. RC20H was purchased from Transduction Laboratories, anti-GST and anti-Grb2 were from Santa Cruz Biotechnology (Santa Cruz, CA), anti-pY1349Met was from Cell Signaling (Mississauga, Ontario, Canada), and anti-Gab1 was from Upstate Biotechnology Inc.
Three-dimensional Modeling of Met—The catalytic domain of Met was modeled using Swiss Model Comparative Protein Modeling Server and was visualized with Swiss-Pdb viewer (27–29). The Met kinase domain was modeled on available structures of the insulin and IGF-1 receptors (Protein Data Bank ID codes, 1K3AA, 1IRK [PDB] , 1IR3A, and 1GAGA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Grb2)
(18). Previous studies had
identified 13 amino acids within the MBD as essential for Grb2-independent
recruitment of Gab1 (15,
17). To investigate the
requirements for Met receptor binding, we first established whether a 16-amino
acid peptide containing these 13 amino acids (MBD16,
FGMQVPPPAHMGFRSS (Fig.
1A)) would act as a competitor in association studies
in vitro. Addition of increasing levels of this peptide (10–250
µM) was able to compete the association of GST-MBDGrb2
from Met in an in vitro association assay
(Fig. 1B). This was
specific for the interaction of Met with the Gab1 MBDGrb2, because the
peptide failed to compete the association of a GST-Grb2 fusion protein with
Met (Fig. 1B) or the
association of a WT Gab1MBD fusion protein with Grb2
(Fig. 1C). Our
previous studies had shown that Arg-499 within the Gab1MBD was required for
the interaction of MBD with Met. Arg-499 is located within this 16-amino acid
peptide, and a peptide where Arg-499 was converted to Ala (Gab1MBD16-R499A)
failed to compete the association of the Gab1 MBDGrb2 with Met
(Fig. 1D).
|
We had previously demonstrated that insertion of these 16 amino acids into
Gab2 was sufficient to confer the capacity for Grb2-independent Met binding on
Gab2 (17). However, it was
unclear if this short sequence was sufficient to interact with Met
independently in a direct manner, or if surrounding amino acids in Gab2, which
share homology with the Gab1 MBD, were also required. To investigate this, we
created a fusion protein containing only these 16 amino acids fused to GST
(GST-MBD16). When used in an in vitro association assay, the
GST-MBD16 fusion protein was sufficient to associate with Met, albeit at a
lower efficiency than the full-length MBDGrb2. Moreover, consistent
with our previous studies, the conversion of Arg-499 to Ala abolished the
ability of the GST-MBD16 fusion protein to bind to Met
(Fig. 1E). No
association of Met was detected with GST alone, indicating that binding was
specific to the 16 amino acids in the Met binding domain, referred to as the
Met binding motif (MBM).
Structural Integrity of Met Is Required for Interaction with the Gab1
MBD—The observation that a 16-amino acid peptide derived from Gab1
can bind to Met is inconsistent with the Gab1 MBD forming a structured domain
in a manner similar to a PTB or SH2 domain. In general, an interaction between
two proteins requires at least one of the interacting regions to form a
structured domain. To assess whether the structural integrity of the Met
receptor, rather than the Gab1 MBD, is required for this interaction, we
performed in vitro association assays with denatured Met protein.
Proteins from lysates of 293T cells transiently transfected with Met
expression plasmids were immunoprecipitated with Met antibody, boiled in 1%
SDS to promote protein unfolding, and boiled proteins were then added to the
in vitro association assay. As a control, the same assay was
performed using the same lysates that were not previously boiled. Importantly,
GST-Grb2 and GST-Grb2 SH2 domain fusion proteins bound equally to denatured or
non-denatured Met proteins (Fig.
2A). This was not surprising, because the interaction of
Grb2 with Met requires only a short phosphotyrosine motif corresponding to
Tyr-1356 (YVNV) in the Met C terminus and the structured SH2 domain of Grb2
(30,
31). In contrast, although the
GST-MBDGrb2 fusion protein associated with non-denatured Met protein,
it failed to bind Met from the boiled lysate containing denatured protein
(Fig. 2A). This
indicates that the interaction of the Gab1 MBD with the Met receptor requires
the structural integrity of Met. This is further supported by the inability of
the MBD to bind a denatured Met protein following SDS-PAGE and transfer to
nitrocellulose membrane in a Far Western blot assay. In contrast, a GST-Grb2
fusion protein efficiently binds Met in a far Western assay
(Fig. 2B).
|
Delineation of the Gab1 MBD Binding Site on Met—To define
the requirements in Met for interaction with the Gab1 MBM, we undertook a
structure-function approach using previously characterized mutants of Met. The
substitution of tyrosine 1349 with phenylalanine severely reduced the
association between GST-MBDGrb2 and Met in an in vitro
association assay (Fig.
3A) or by co-immunoprecipitation when both proteins are
transiently overexpressed in 293T cells
(Fig. 3B). In
contrast, the substitution of tyrosine 1356 with phenylalanine had little
effect on MBDGrb2 binding by either assay, whereas, as expected, a
mutant with both Tyr-1349 and Tyr-1356 substituted with phenylalanine residues
or a kinase inactive mutant (K1110A) were both unable to bind to a
MBDGrb2 fusion protein (Fig. 3,
A and B). This identifies an important role for
phosphotyrosine 1349 in the Grb2-independent interaction with the Gab MBD and
supports previous data obtained using a yeast two-hybrid interaction assay
(22).
|
To determine whether amino acids upstream or downstream from Tyr-1349 are
required for Grb2-independent recruitment of the Gab1 MBD, we used existing
mutants of the Met oncoprotein where amino acids 1353–1362 of Met were
replaced with amino acids distinct from those in the Met receptor but still
capable of recruiting signaling proteins. Mutant A contained amino acids
derived from the platelet-derived growth factor receptor, and mutant B
contained amino acids derived from the TrkA receptor, with binding sites for
the p85 subunit of PI3'K and the Shc adapter protein, respectively
(23). When subjected to an
in vitro association assay, both the A and B variants were capable of
associating with the MBDGrb2 fusion protein. As expected, substitution
of Tyr-1349 to Phe in either of these variants abolished binding
(Fig. 3C). Because the
sequence of the A and B inserts are distinct from Met, and B adds additional
amino acids, this indicates that residues 1353–1362 are not essential
for association with the Gab1 MBDGrb2. To determine whether residues in
Met C-terminal to 1362 are required, we constructed C-terminal truncation
mutants in the Met receptor (Fig.
3E). A Met protein lacking the last 38 amino acids
(1353–1390) associated with MBDGrb2 to similar levels as
WT, as revealed using an antibody that recognizes the extracellular domain of
Met. As expected, a Met protein that lacks the last 43 amino acids, including
Tyr-1349, (1348–1390) failed to bind
(Fig. 3D). This
indicates that residues in the C terminus of Met, downstream of Asn-1353 are
not required for interaction with the Gab1 MBD, and suggests that residues
upstream of Tyr-1349 may be required.
Residues Upstream of Tyr-1349 Are Required for Association with the
Gab1 MBD—To test this hypothesis, we performed alanine-scanning
mutagenesis on residues surrounding Tyr-1349 and upstream of 1353. Amino acids
1340–1352 were individually substituted with alanine residues and tested
for their ability to associate with a GST-MBDGrb2 fusion protein.
Substitution of several residues upstream of Tyr-1349 in Met was found to
significantly decrease the association with the GST-MBDGrb2 fusion
protein. Notably, substitution of Phe-1341, Phe-1344, Ile-1345, Gly-1346, and
His-1348 severely diminished the association, and as shown, the Y1349F Met
mutant protein failed to associate with the GST-MBDGrb2 fusion protein
(Fig. 4B), even though
all mutant proteins were expressed at similar levels. Importantly, mutant Met
proteins were tyrosine-phosphorylated and capable of associating with GST-Grb2
to similar levels as WT Met (Fig.
4C). Hence, the reduction in their association with
GST-MBDGrb2 was not due to a change in the overall tyrosine
phosphorylation of the mutants. We conclude that these mutations specifically
alter the ability of Met proteins to associate with Gab1 MBDGrb2,
without affecting the ability of Met to associate with other interacting
partners such as Grb2. This also implies that the structure of the Met
receptor kinase domain is not compromised, such that it could no longer
phosphorylate tyrosine residues in the C terminus. Indeed, as shown using a
phospho-specific Tyr-1349 antibody, all mutant proteins are
tyrosine-phosphorylated on Tyr-1349 (Fig.
4C), indicating that the diminished association of
GST-MBDGrb2 is not due to the inability of Met mutant proteins to
phosphorylate Tyr-1349. These data identify that residues upstream of Tyr-1349
are required for binding to the Gab1 MBD.
|
A Negative Charge at Residue 1349 Is Sufficient for MBD
Binding—Phosphorylation of Tyr-1349 is required for binding to the
MBDGrb2 (Fig. 3, A and
B). However, the role of this phosphorylation is unclear,
because the association of the Gab1 MBDGrb2 requires structural
integrity in Met and hence does not bind Met in a similar manner to the
association of PTB or SH2 domains with linear phosphopeptide motifs
(32–35).
Because Tyr-1349 is localized two amino acids downstream from the last helix
of the Met kinase domain, and Met structural integrity was required for
association with the MBDGrb2, we reasoned that phosphorylation of
Tyr-1349 may play a role in modulating the structure of Met, allowing
association with the Gab1MBDGrb2 protein. To investigate this, we
created mutant proteins with tyrosine to glutamic acid substitutions of
Tyr-1349 and Tyr-1356, to determine whether the MBDGrb2 requires the
actual phosphotyrosine entity for binding to Met or just requires the negative
charge of the phosphotyrosine. Similar substitutions have been used previously
to mimic the effects of a negatively charged phosphotyrosine on conformational
changes, without allowing the binding of SH2 domain-containing proteins
(36). Whereas a Y1349F mutant
protein was unable to associate with the MBDGrb2, the substitution of
Tyr-1349 with glutamic acid (Y1349E) rescued the ability of the Y1349F mutant
Met protein to bind to MBDGrb2, although at a reduced level to the WT
protein. Moreover, association of MBDGrb2 was observed, even when both
tyrosines 1349 and 1356 were substituted for glutamic acid residues,
Y1349E/Y1356E (Fig.
5A). This was specific for the MBDGrb2, because
mutants Y1349E/Y1356E or Y1356E were unable to bind GST-Grb2
(Fig. 5B). Consistent
with this, following transient transfection assays, a Y1349E/Y1356E mutant
induced tyrosine phosphorylation of the endogenous Gab1 protein in 293 cells,
whereas in cells expressing the Y1349F/Y1356F mutant Gab1 was only basally
phosphorylated (Fig.
5B). This ability of the Y1349E/Y1356E mutant to bind and
phosphorylate Gab1 in the absence of phosphotyrosines 1349/1356 also
correlated with the ability of this mutant to transform fibroblasts
(Fig. 5C). Whereas the
Y1349E/Y1356E mutant of Tpr-Met fails to phosphorylate Gab1 and is unable to
transform fibroblasts (Fig.
5C)
(37–40),
the Y1349E/Y1356E mutant transformed fibroblasts to a low level (6 foci/µg
of DNA). These results indicate that negative charge at Tyr-1349 is essential
for binding of MBDGrb2.
|
Residues Critical for MBD Binding Lie within the Met Receptor Kinase
Domain—From sequence alignments, residues identified through
alanine-scanning mutagenesis as being critical for MBD binding (Phe-1341,
Phe-1344, and Ile-1345, Fig.
4A) are located within the kinase domain of the Met
receptor (41). Kinase domains
are extremely well conserved among serine/threonine and tyrosine kinases and
consist of two subdomains. The N-terminal lobe is composed of a five-stranded
-sheet (1–5) and one -helix (helix C), whereas
the larger C-terminal lobe contains at least two strands
(7/8) and 7–8 helices (D, E,
EF, F, and I/J)
(42). ATP is coordinated
primarily by the N-terminal lobe, whereas substrate peptide binding and
catalysis are performed by residues in the C-terminal lobe (reviewed in Ref.
43). Although the structure of
the Met receptor kinase domain has not been solved, the three-dimensional
structure of the Met kinase domain modeled on the insulin and IGF-1 receptors
predicts that these residues lie within the C-terminal lobe
(Fig. 6A). Notably,
Phe-1341 is located in the loop between -helix I and -helix J,
whereas Phe-1344 and Ile-1345 compose -helix J
(Fig. 6B), indicating
that the binding of the GST-MBDGrb2 to Met requires residues within the
kinase domain. The remainder of the residues critical for Gab1 MBD binding
(Gly-1346, His-1348, and pY1349) did not appear in the model because they are
located C-terminal to the kinase domain. Phe-1341, Phe-1344, and Ile-1345 are
predicted to lie in close contact (less than 5 Å) to residues in the
-helix E (Lys-1179, Ile-1182, Leu-1186, and Gln-1187), as well as
Lys-1215 located in the loop between -strand 7 and -strand 8
(Fig. 6B). Because we
have shown that the MBD cannot bind to a denatured Met protein
(Fig. 2A), the
location of these residues in the kinase domain supports our observation that
the structural integrity of the kinase domain itself may be required.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
A Peptide Motif within the Gab1 MBD Is Sufficient to Interact with the Met Receptor—A GST fusion protein containing only 16 amino acids derived from the Gab1 MBD (GST-MBD16), referred to as the Met binding motif (MBM), was sufficient to interact with Met. Consistent with this, only residues found within the MBD16 peptide (amino acids 487–499) were identified by random PCR mutagenesis as critical for Met binding (15). However, interaction of the Gab1 MBD16 was reduced when compared with the full-length GST-Gab1MBD fusion protein indicating that the surrounding residues in the Gab1MBD likely contribute to a higher affinity binding (Fig. 1E). Residues outside of a minimal domain-binding motif can contribute to affinity (46). For example, a full-length Nef protein can bind to SH3 domains 300 times more efficiently than a peptide corresponding to known SH3 domain interacting PXXP region in Nef (47). In support of this, in a yeast two-hybrid binding assay, amino acids 450–532 corresponding to the Gab1 MBD were found to be necessary and sufficient for binding of Gab1 to Met, whereas shorter constructs retained only minimal binding to Met (22). Together this supports a role for residues in Gab1 outside of the Met binding motif for efficient binding to Met.
Consistent with the Met binding motif in Gab1 being sufficient to bind Met,
using multiple structure prediction methods, including PHD
(48–50)
and PSIPRED (51,
52), the Gab1 MBD is predicted
not to form any secondary structure and, from this data, is unlikely to
function as a domain (data not shown). This is likely due in part to the
proline-rich nature of the MBD; 25% of the residues are prolines in comparison
to the expected 6%. The amino acid proline is established as a potent breaker
of both -helical and -sheet structures in globular proteins
(53–56).
Moreover, in addition to the Met binding motif, the MBD contains binding sites
for other proteins, including the p85 SH2 domain (two binding sites), the Grb2
SH3 domain, and Erk1/2 (15,
18,
57,
58). These sites are
non-overlapping, suggesting that the MBD is relatively extended and
accessible. We propose then that the MBD is actually a loosely structured
region with a central Met binding motif and surrounding residues that provide
additional contacts with the Met receptor.
The Integrity of the Met Kinase Domain Is Required for Gab1 Interaction—In general, an interaction between two proteins requires at least one of the interacting regions to form a structured domain. The conclusion that the MBD functions as a peptide ligand led us to speculate that the Met receptor may contain a structured domain that interacts with the MBM peptide. In support of this, we have shown through denaturation studies that the structural integrity of Met is required for its interaction with the MBD (Fig. 2, A and B). This is in contrast to the interaction of Met with Grb2, which requires only a short YVNV phosphotyrosine-containing motif involving Tyr-1356 in the C terminus of Met, and the structured SH2 domain of Grb2 (Fig. 2, A and B). In addition, we have shown that the interaction between the MBD and Met requires residues upstream of Tyr-1349, as well as the presence of a phosphotyrosine or a negatively charged residue at 1349 (Figs. 3, 4, 5). Several of these residues (Phe-1341, Phe-1344, and Ile-1345) lie within the end of the C-terminal lobe of the kinase domain. Together these data indicate that the interaction of the Gab1MBD with Met requires the structural integrity of the Met receptor kinase domain.
Although the Met catalytic domain has not been crystallized, the modeling
of the Met receptor kinase domain on structures of the insulin and IGF-1
kinase domains indicated that residues required for MBD binding are located in
the loop between -helix I and -helix J (Phe-1341) and comprise
-helix J (Phe-1344 and Ile-1345)
(Fig. 6). All kinase domains
whose structures have been solved thus far have been done so in the absence of
their C-terminal regions, with the exception of the Tie2 receptor
(59). Therefore, it is unclear
how the 47-amino acid C terminus of Met, including Gly-1346, His-1348, and
pY1349, critical for MBD binding, fit into the model. The crystal structure of
Tie2 suggests that the C-terminal tail blocks access to the substrate binding
site of the kinase domain and must undergo a conformational change upon
activation of the receptor to expose both the substrate binding site and
tyrosines in the C terminus required for binding of signaling proteins
(59). The Met C terminus may
adopt a similar conformation, because a peptide of the Met receptor
multifunctional docking site (amino acids 1345–1363) can bind to the Met
kinase domain in an undefined manner, and inhibit the kinase activity of the
receptor (60).
We propose that residues identified through alanine-scanning mutagenesis in the C-terminal lobe of the kinase domain function to create a binding surface for the MBM peptide and that phosphorylation of Tyr-1349 located at the junction between the kinase domain and the C terminus is required for a conformational change in Met that exposes a binding surface for Gab1 (Fig. 7, model). Consistent with this, the substitution of Tyr-1349 for a charged glutamic acid is sufficient for Gab1 binding and phosphorylation, whereas a similar substitution at 1356 abrogates the association of the Grb2 SH2 domain (Fig. 5B). However, we cannot rule out the possibility that mutation of these residues may affect the binding of the proline-rich MBM to another location in the Met kinase domain.
|
The peptide binding surfaces of domains that bind to proline-rich ligands,
including SH3, WW, GYF, and EVH1, tend to contain elongated patches of
aromatic and hydrophilic residues that create a contiguous binding surface.
Residues Phe-1341, Phe-1344, and Ile-1345 on Met are aromatic and/or
hydrophilic and are predicted to lie in close contact with side chains of
residues in -helix E (Lys-1179, Ile-1182, Gly-1183, Leu-1186, and
Gln-1187), as well as Lys-1215 located in the loop between -7 and
-8 (Fig. 6B). A
potential binding surface for the MBD ligand is present in the region
comprising helices E, I, and J (Fig.
6B). However, the presence of two basic residues
(Lys-1179 in -helix E and Lys-1215 in the loop between -7 and
-8) could be a deterrent to MBM binding. Activation of the kinase and
phosphorylation of tyrosines with the activation loop, as well as
phosphorylation of Tyr-1349, could lead to a conformational change that would
position the MBM binding region away from these basic residues, or may be
required to expose the binding site for the MBM
(Fig. 7, model).
The peptide binding surface of SH3 domains generally also contains at least one negatively charged acidic residue that interacts with a positively charged basic residue located within the proline-rich ligand. The MBM contains a basic residue, Arg-499, that is critical for interaction with Met (15, 17) (Fig. 1, D and E). However, the MBD binding region on Met contains only one acidic residue, Glu-1347, that was not essential (Fig. 4B). Alternatively, because phosphorylation of Tyr-1349 or the creation of a negative charge is critical (Fig. 2, A and B), the positively charged Arg-499 in the MBM may interact with the negatively charged phosphate group on Tyr-1349, in a manner reminiscent of, but opposite to, that of SH2 domains and phosphotyrosine ligands. Hence, in a similar manner to an SH3 domain, the kinase domain of Met may contain a contiguous binding surface of aromatic and hydrophilic residues, in addition to a negatively charged amino acid (phosphotyrosine 1349), that engage the prolines and the critical arginine, respectively, of the MBM.
Although the Grb2-independent interaction is specific for Gab1 and Met, specificity in binding to particular receptors has also been shown for other docking/adapter proteins. For example, the Grb7/Grb10/Grb14 family of adapter proteins, in addition to their SH2 domains, contain novel receptor-specific interaction domains (BPS) that allow them to interact differentially with receptors (61, 62). Moreover, a domain in IRS-2, the kinase regulatory loop-binding (KRLB) domain, interacts specifically with the insulin receptor but not the highly related IGF-1 receptor (63). Intriguingly, MBD, BPS, and KRLB all bind to kinase domains, and all three interactions require receptor kinase activity (61, 64). It is not clear yet whether the BPS or the KRLB are indeed domains or, like the MBD, would also function as extended peptide motifs. The presence of domains or motifs in docking proteins that interact differentially with a subset of receptor tyrosine kinases may thus be a common mechanism through which docking proteins can modulate distinct biological responses downstream from receptor tyrosine kinases. Solving the three-dimensional structure of Met, in conjunction with Gab1, will permit a clearer understanding of this novel interaction, and could provide an approach to interfere specifically with the invasive response triggered by Gab1 downstream from the Met receptor.
|
FOOTNOTES |
|---|
|| A Scientist of the Canadian Institutes of Health Research. To whom correspondence should be addressed: Molecular Oncology Group H5–10, McGill University Health Centre, 687 Pine Ave. West, Montreal, Quebec H3A 1A1, Canada. Tel.: 514-842-1231 (ext. 35845); Fax: 514-843-1478; E-mail: morag{at}molonc.mcgill.ca.
1 The abbreviations used are: Gab1, -2, -3, Grb2-associated binders
1–3; HGF, hepatocyte growth factor; PI3K, phosphatidylinositol 3-kinase;
MBD, Met binding domain; PTB, phosphotyrosine binding; RIPA, radioimmune
precipitation assay buffer; GST, glutathione S-transferase; IGF-1,
insulin-like growth factor 1; MBM, Met binding motif; WT, wild type; KRLB,
kinase regulatory loop-binding domain; BPS, between PH and SH2.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
