(Received for publication, September 6, 1996, and in revised form, December 9, 1996)
From the Department of Physiology and Cell Biology, Albany Medical College, Albany, New York 12208
Integrin-mediated adhesion of cells to
extracellular matrix proteins triggers a variety of intracellular
signaling pathways including a cascade of tyrosine phosphorylations. In
many cell types, the cytoplasmic focal adhesion tyrosine kinase, FAK,
appears to be the initial protein that becomes tyrosine-phosphorylated in response to adhesion; however, the molecular mechanisms regulating integrin-triggered FAK phosphorylation are not understood. Previous studies have shown that the integrin 1,
3, and
5 subunit cytoplasmic domains all
contain sufficient information to trigger FAK phosphorylation when
expressed in single-subunit chimeric receptors connected to an
extracellular reporter. In the present study,
3
cytoplasmic domain deletion and substitution mutants were constructed
to identify amino acids within the integrin
3
cytoplasmic domain that regulate its ability to trigger FAK
phosphorylation. Cells transiently expressing chimeric receptors
containing these mutant cytoplasmic domains were magnetically sorted
and assayed for the tyrosine phosphorylation of FAK. Analysis of these
mutants indicated that structural information in both the
membrane-proximal and C-terminal segments of the
3
cytoplasmic domain is important for triggering FAK phosphorylation. In
the C-terminal segment of the
3 cytoplasmic domain, the
highly conserved NPXY motif was found to be required for
the
3 cytoplasmic domain to trigger FAK phosphorylation. However, the putative FAK binding domain within the N-terminal segment
of the
3 cytoplasmic domain was found to be neither
required nor sufficient for this signaling event. We also demonstrate
that the serine 752 to proline mutation, known to cause a variant of Glanzmann's thrombasthenia, inhibits the ability of the
3 cytoplasmic domain to signal FAK phosphorylation,
suggesting that a single mutation in the
3 cytoplasmic
domain can inhibit both "inside-out" and "outside-in" integrin
signaling.
Integrins are a family of /
heterodimeric adhesion receptors
used by cells to interact with their extracellular matrix (1). In
addition to mediating cell adhesion, integrins also function as
signaling receptors, integrating information from the extracellular environment (2-5). Integrin engagement triggers intracellular signals
that direct cell adhesion and regulate other aspects of cell behavior
including cell proliferation, differentiation, and survival. A major
integrin-triggered signal is a cascade of tyrosine phosphorylations. In
many cell types, an early integrin-triggered event is the tyrosine
phosphorylation and activation of the focal adhesion kinase,
FAK1 (2-5), which is followed by the
tyrosine phosphorylation of additional cytoskeletal and signaling
proteins, including paxillin, tensin, MAP kinase, cortactin, and
p130CAS (6-12). These phosphorylation events are likely to be
central to the assembly of adhesion-dependent signaling
complexes that regulate cell adhesion and other aspects of cell
behavior (2-5).
The integrin-triggered tyrosine phosphorylation of FAK induces the formation of specific adhesion-dependent signaling complexes by creating binding sites for SH2 domains of other signaling molecules, thus providing pathways linking integrin engagement with downstream signaling events. Integrin engagement triggers the autophosphorylation of FAK, which in turn generates a binding site for the SH2 domain of Src family kinases (13, 14). These interactions result in the further phosphorylation and activation of FAK (15) and have recently been shown to be required for the function of FAK in modulating the migratory behavior of cells (16), as well as the ability of constitutively activated FAK to inhibit apoptosis in some circumstances and function as a transforming agent in others (17). The phosphorylation of FAK at other tyrosine residues provides binding sites for additional SH2-containing proteins, such as phosphatidylinositol 3-kinase (18) and the adapter protein, GRB2 (19), both of which are important in growth factor signaling pathways. These phosphorylation events may therefore serve to link integrin engagement to downstream signaling pathways including the Ras and MAP kinase cascades known to become activated by integrin-mediated adhesion (8, 20, 22-24).
Since integrins have no intrinsic enzymatic activity, they must depend
on their association with other cellular proteins to initiate
intracellular signals. Integrin subunit cytoplasmic domains are
likely to serve as the major link between integrins and the
cytoskeleton and signal transduction apparatus of the cell (25).
subunit cytoplasmic domains have been demonstrated to be required for
many integrin-mediated processes, including cell adhesion, cell
spreading, cell migration, and fibronectin matrix assembly, as well as
to signal FAK phosphorylation (26-30). Additionally, cytoplasmic
domains of specific integrin
subunits expressed in the context of
single-subunit chimeric receptors are sufficient to both mimic and
inhibit various functions of ligand-occupied integrins (31-36),
further supporting the notion that
cytoplasmic domains interact
with specific cytoplasmic proteins required for endogenous integrin
function.
Chimeric receptors containing the integrin 1,
3, or
5 cytoplasmic domain connected to a
reporter consisting of either the IL-2 or CD4 receptor extracellular
domain were previously utilized to demonstrate that
subunit
cytoplasmic domains contain sufficient information to activate the FAK
signaling pathway (34, 35). Furthermore, this was found to be dependent
upon their primary structure, since the integrin
1B and
3B subunit cytoplasmic domains, which are modified by
alternative splicing, did not signal FAK phosphorylation (34, 37). In
the present study, we have used single-subunit chimeric receptors to
analyze the role of specific amino acids within the
3
cytoplasmic domain in signaling FAK phosphorylation. To do this, we
constructed chimeric receptors containing the extracellular and
transmembrane domains of the IL-2 receptor connected to mutant
3 cytoplasmic domains with specific amino acid deletions
or substitutions (Fig. 1). We found that amino acids within both the
N-terminal and the C-terminal segment of the
3
cytoplasmic domain were necessary to signal increases in the tyrosine
phosphorylation of FAK, and that the putative FAK binding domain within
the N-terminal segment of the
3 cytoplasmic domain was
neither sufficient nor necessary to signal FAK phosphorylation. We have
further defined an important role for the conserved NPXY
motif in this signal transduction event.
DNAs encoding specific amino acid substitutions and
deletions in the 3 cytoplasmic domain were generated by
the polymerase chain reaction (PCR), using plasmid DNA encoding the
3 chimera (32) as a template with oligonucleotide
primers described below. The only exception is the
3-d728-762 deletion mutant, which was constructed by
hybridizing the following complementary oligonucleotides: 5
-AGCTTCTCATCACCATCCACGACCGAAAAGAATTCTAAC-3
and
5
-TCGAGTTAGAATTCTTTTCGGTCGTGGATGGTGATGAGA-3
. For the
3-d718-741 mutant, the N-terminal PCR primer was
5
-AGGGACAAGCTTGCCAACAACCCACTG-3
and the C-terminal PCR primer was
5
-GAGTCTCTCGAGTTAAGTGCCCCGGTACGTG-3
. For the
3-723,726 (*/A) mutant, the N-terminal PCR primer was 5
-CTCATCTGGAAGCTTCTCATCACCATCCACGCCCGAAAAGCATTCGCTAAATTTGAG-3
and the
C-terminal PCR primer was 5
-GAGTCTCTCGAGTTAAGTGCCCCG-3
. The
N-terminal oligonucleotide primer for the remaining mutants was
5
-GGCTCACCTGGAAGCTTCTCATC-3
. The C-terminal primers were as follows:
for the
3-d742-762 mutant,
5
-GAGTCT-CTCGAGTCATGTGTCCCATTTTGCTCTGGC-3
; for the
3-756(N/A) mutant,
5
-GAGTCTCTCGAGTTAAGTGCCCCGGTACGTGATAGCGGTGAAGGTAGA-3
; for the
3-759(Y/A) mutant,
5
-GAG-TCTCTCGAGTTAAGTGCCCCGGGCCGTGATATTGGTGAA-3
; for the
3-752(S/P) mutant,
5
-GAGTCTCTCGAGTTAAGTGCCCCGGTACGTGATATTGGTGAAGGTAGGCGTGGCCTCTTTATACAG-3
; for the
3-751-753 (*/A) mutant,
5
-GAGTCTCTCGAGTTAAGTGCCCCGGTACGTGATATTGGTGAAGGCAGCCGCGGCCTCTTTATACAG-3
; for the
3-747 (Y/F) mutant,
5
-GAGTCTCTCGAGTTAAGTGCCCGCTACGTGATATTGGTGAAGGTAGACGTGGCCTCTTTAAACAGTGGGTTTGTTGC-3
; for the
3-747(Y/A) mutant,
5
-GAGTCTCTCGAGTTAAGGCCCCGGTACGTGATATTGGTGAAGGTAGACGTGGCCTCTTTAGCCAGTGGGTTGTTGGC-3
; and for the
3-744(N/A) mutant,
5
-GAGTCTCTCGAGTTAAGTGCCCCGGTACGTGATATTGGTGAAGGTAGACGTGGCCTCTTTATACAGGGCGTTGGCTGTGTCCCA-3
. The mutant cytoplasmic domains were inserted as
HindIII/XhoI restriction fragments into the
previously described plasmid vector (31). The construction of each
mutant was confirmed by DNA sequence analysis (38).
Human fibroblasts,
transiently expressing the chimeric receptors, were harvested 16-18 h
after transfection. Positively transfected cells were sorted with
magnetic beads conjugated with goat anti-mouse IgG (Advanced Magnets,
Cambridge, MA) and mAb 7G7B6 to the IL-2 receptor (Upstate
Biotechnology Inc., Lake Placid, NY) as described previously (39).
Magnetically sorted cells were lysed with a buffer containing 50 mM Tris (pH 7.4), 150 mM NaC1, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EDTA, 1 mM sodium vanadate, 1 mM sodium fluoride, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 1 mM
phenylmethylsulfonyl fluoride. Protein concentrations were determined
using the Micro-BCA assay (Pierce). Cell lysates (15 µg) were
resolved under reducing conditions by 10% SDS-PAGE. Tyrosine
phosphorylation was analyzed by Western blotting using the mAb 4G10 to
phosphotyrosine (Upstate Biotechnology Inc.) and visualized by enhanced
chemiluminescence (Amersham). The filters were then stripped by
incubation at 75 °C for 40 min in stripping buffer containing 2%
SDS, 62.5 mM Tris-HCl, pH 6.8, and 100 mM -mercaptoethanol, and the level of FAK present in each sample was
assayed by reprobing the filter with mAb to FAK (Transduction Laboratories, Lexington, KY). Relative amounts of FAK and
phosphorylated FAK in each lysate were determined by densitometry using
a Millipore Bioimage 60S video densitometer utilizing Visage software.
For this purpose, exposures were chosen so that the signal from each lysate was in the linear range of the film. Expression levels of the
chimeric receptors were analyzed in each experiment by flow cytometry
using a Becton Dickinson FACScan flow cytometer as described previously
(32). In some experiments, magnetically sorted cells were lysed with a
1% Nonidet P-40 buffer (40), and equal amounts of lysates (300 µg)
were immunoprecipitated by incubation with polyclonal antibodies to
FAK. Immune complexes were collected with Protein A-Sepharose and then
analyzed under reducing conditions by 10% SDS-PAGE. Tyrosine
phosphorylation was then visualized by Western blotting as described
above.
To identify the amino acids within the 3
cytoplasmic domain required to signal increases in FAK phosphorylation,
we made a series of amino acid deletions and substitutions in the
3 cytoplasmic domain (Fig. 1) based on
three criteria: (a) the identity of amino acid residues
important for FAK binding in vitro (41), (b) the identity of the amino acids replaced by alternative splicing of
1B and
3B (25), and (c) the
regions of homology in the C-terminal regions of the
1,
3, and
5 cytoplasmic domains known to
contain sufficient information to trigger FAK phosphorylation (25, 34). These
3 cytoplasmic mutants were expressed in the
context of single-subunit chimeric receptors as described previously
(31, 32). Their ability to signal FAK phosphorylation was assayed using
a magnetic sorting procedure that not only enriched for positively
expressing cells but also triggered FAK phosphorylation in a manner
dependent upon the presence of the integrin
cytoplasmic domain
(34).
Differences in the abilities of the
3 and
3B cytoplasmic domains to trigger
FAK phosphorylation suggested either that the C-terminal amino acids of
the
3 cytoplasmic domain are required for signaling FAK
phosphorylation or that the C-terminal amino acids of the
3B cytoplasmic domain inhibit this signaling event. To
distinguish between these possibilities, three deletion mutants were
constructed:
3-d742-762,
3-d728-762,
and
3-d718-741 (Fig. 1). Chimeric receptors containing
these mutant cytoplasmic domains or the wild-type
3
cytoplasmic domain were transiently expressed in normal human
fibroblasts and then assayed for their ability to trigger FAK
phosphorylation (Fig. 2). Increases in tyrosine phosphorylation of FAK were observed in lysates from cells expressing chimeric receptors containing the wild-type
3
cytoplasmic domain, whereas cells expressing a chimeric receptor
lacking the entire
3 cytoplasmic domain did not show
this increase in phosphorylation in agreement with previously published
results (34). The
3 deletion mutant,
3-d742-762, which lacked the C-terminal amino acids
that differ between the
3 and
3B
cytoplasmic domains, consistently did not signal FAK phosphorylation
above background levels (Figs. 2 and 6). This result suggests that the
C-terminal region of
3B does not play a negative role,
but rather that amino acids in the C-terminal region of
3 are required to signal FAK phosphorylation. The
deletion mutant,
3-d728-762, containing the first 12 membrane-proximal amino acids of the
3 cytoplasmic domain reported to bind to FAK in vitro (41), also was not
sufficient to signal increases in tyrosine phosphorylation of FAK above
background levels (Figs. 2 and 6), further suggesting that additional
C-terminal amino acids are required for this signaling event.
Similarly, chimeric receptors containing the deletion mutant,
3-d718-741, which contained only the 21 C-terminal
amino acids of
3 that are replaced in
3B,
also did not signal FAK phosphorylation, indicating that these amino
acids are not sufficient to signal FAK phosphorylation (Figs. 2 and 6).
These results suggest that the sequence information in both the
N-terminal and C-terminal regions is required for the
3
cytoplasmic domain to trigger FAK phosphorylation. Although the level
of expression of each chimeric receptor varied from each other and from
one experiment to the next, differences in the ability of various
chimeric receptors to trigger FAK phosphorylation were not found to be
due to these differences in expression levels.
The Level of Expression of Chimeric Receptors Versus Their Ability to Signal FAK Phosphorylation
We also directly tested whether
differences in the level of expression of individual chimeric receptors
affected their ability to signal FAK phosphorylation. Since the amount
of DNA used for transfection directly influenced the average level of
transfected gene expression in individual cells, as well as the total
number of positively expressing cells observed after transfection, we transfected cells with either 15 or 40 µg of plasmid DNA encoding the
chimeric receptors containing either the wild-type 3
cytoplasmic domain or the deletion mutant,
3-d728-762.
We analyzed chimeric receptor expression levels in these different
transfected cell populations by flow cytometry. Transfection with
either 15 or 40 µg of DNA encoding the chimeric receptor containing
the wild-type
3 cytoplasmic domain generated populations
of cells with mean fluorescence intensities of 42.9 and 135.99, respectively. Similarly, transfection with 15 or 40 µg of DNA
encoding the
3-d728-762 deletion mutant generated
populations of cells with mean fluorescence intensities of 124.59 and
352.99, respectively. When we assayed these different populations of
cells for FAK phosphorylation, we consistently found that the wild-type
3 cytoplasmic domain signaled FAK phosphorylation when
expressed at either higher or lower levels, whereas the deletion mutant
3-d728-762 was consistently unable to signal tyrosine
phosphorylation of FAK, even at higher levels of expression (Fig.
3).
The Conserved Acidic Residues Asp and Glu in the N-terminal Segment of the
The chimeric receptor
3-d728-762, which contains the putative FAK binding
domain, was unable to signal FAK phosphorylation, indicating that in
some circumstances the putative FAK binding is not sufficient to
trigger FAK phosphorylation (Fig. 2). To determine whether the putative
FAK binding domain was required for the
3 cytoplasmic
domain to signal FAK phosphorylation, we also generated a chimeric
receptor,
3-723,726(*/A), that contained a mutant
3 cytoplasmic domain with alanine substitutions at the acidic residues, aspartic acid (Asp-723) and glutamic acid (Glu-726). Previous studies had shown that these specific mutations inhibit FAK
binding in vitro to peptides modeled after the
membrane-proximal region of
cytoplasmic domains (41).
Interestingly, chimeric receptors containing these alanine
substitutions consistently triggered FAK phosphorylation to levels
similar to those induced by chimeras containing the wild-type
3 cytoplasmic domain (Fig. 4,
A and B, and Fig. 6). These results indicate that
the putative FAK binding domain is neither sufficient nor required for
the
3 cytoplasmic domain to trigger FAK
phosphorylation.
The ability of the substitution mutant, 3-723,726(*/A),
to signal FAK phosphorylation and the inability of the deletion mutant,
3-d728-762, containing only the putative FAK binding
site to signal FAK phosphorylation was also examined by
immunoprecipitation. Again, chimeric receptors containing either the
wild-type
3 cytoplasmic domain or the substitution
mutant,
3-723,726(*/A), were able to signal FAK
phosphorylation, whereas the deletion mutant containing only the
putative FAK binding domain was not (Fig. 4C).
To ensure that our lack of signal with the chimeric receptor containing the putative FAK binding domain was not due to the loss of phosphorylated FAK with the magnetic beads in our original lysates, we also compared the level of FAK phosphorylation in lysates obtained by boiling the cell-magnetic bead complexes under reducing conditions in a buffer containing 2% SDS (Fig. 4D). These results further confirm our findings that the putative FAK binding domain is not sufficient to trigger FAK phosphorylation above background levels.
Conserved Amino Acids in the C-terminal Segment of theTo identify the amino acids within the C-terminal
portion of the 3 cytoplasmic domain that are required to
signal FAK phosphorylation, amino acid substitution mutants were
generated (Fig. 1). Initial substitutions were targeted to the
NPXY motif which is conserved in the
1,
3, and
5 cytoplasmic domains (25). Three
amino acid substitutions were made in this conserved motif (Fig. 1) and
assayed as above. Chimeric receptors containing alanine substitutions for either asparagine 744, the
3-744(N/A) mutant, or
tyrosine 747, the
3-747(Y/A) mutant, consistently did
not signal FAK phosphorylation (Figs. 5 and 6), even when expressed at
much higher levels (data not shown). These results indicate that the
NPXY motif is important for integrin intracellular function
in this signal transduction event. However, chimeric receptors
containing the conservative substitution of phenylalanine for tyrosine
747 were able to signal FAK phosphorylation (Figs. 5 and
6), in agreement with published reports that a
phenylalanine substitution for tyrosine in the NPXY motif of
the
1 cytoplasmic domain does not inhibit FAK
phosphorylation (30). These results indicate that the NPXY
motif is important for integrin-triggered FAK phosphorylation and that
although an aromatic amino acid is required at residue 747, it need not
be tyrosine, indicating that phosphorylation at this residue is not required for integrin-triggered FAK phosphorylation.
Similar to the NPXY motif, the NXXY motif is also
conserved in the 1,
3, and
5 cytoplasmic domains (25). To determine the
contribution of this motif to the ability of the
3
cytoplasmic domain to signal FAK phosphorylation, we constructed two
additional chimeric receptors,
3-756(N/A) and
3-759(Y/A), containing alanine substitutions for either
asparagine 756 or tyrosine 759, respectively (Fig. 1). Chimeric
receptors containing these mutant
3 cytoplasmic domains
were able to signal FAK phosphorylation, although to a consistently
lower level compared with the wild-type
3 cytoplasmic domain (Figs. 5 and 6). These results indicate that the
NXXY motif is not strictly required for the
3
cytoplasmic domain to signal the phosphorylation of FAK, but may
regulate the level of FAK phosphorylation induced by integrin
engagement.
In addition to targeting the conserved NPXY and
NXXY motifs, we also tested the consequences of the
naturally occurring serine 752 to proline mutation (Fig. 1). This
mutation is responsible for a variant of Glanzmann's thrombasthenia,
in which the absence of platelet aggregation is due to the inability of
the IIb
3 integrin to become activated in
response to platelet agonists (42). Thus, the serine 752 to proline
mutation makes
IIb
3 defective in
"inside-out" signal transduction (42-44). Chimeric receptors
containing this mutant
3 cytoplasmic domain,
3-752(S/P), consistently could not trigger FAK
phosphorylation (Figs. 5 and 6), indicating that it was also defective
in "outside-in" signal transduction. However, chimeric receptors
containing the mutant
3 cytoplasmic domain,
3-751-753(*/A), containing alanine substitutions at
residues threonine 751, serine 752, and threonine 753 (Fig. 1) signaled
an increase in tyrosine phosphorylation of FAK (Fig. 5), although at a
consistently lower level than the wild-type
3
cytoplasmic domain (Fig. 6). These results demonstrate that, although
the substitution of proline at residue 752 has a negative effect on
this signal transduction event, there is not a strict requirement for a
serine residue at this position.
The importance of FAK activity is evident by the observation that
FAK deficient embryos fail to develop normally, with an overall
phenotype similar to fibronectin deficient embryos, supporting the
notion that FAK activity is an important downstream component of
signals generated by integrins (57). FAK activity has been implicated
in regulating cell survival, proliferation, and migration (16, 17, 58).
Although integrin subunit cytoplasmic domains are known to be an
important link in the pathway by which cell adhesion activates FAK (30,
34, 35), the molecular mechanisms regulating this signal transduction
event are not fully understood. Several different experimental
approaches are likely to be required to delineate the molecular
mechanisms leading from integrin engagement to FAK phosphorylation. The
approach we have taken, expressing wild-type and mutant
3 cytoplasmic domains as separate domains connected to
an extracellular reporter, allows examination of the role of specific
amino acid motifs within the
3 cytoplasmic domain in
triggering FAK phosphorylation, independent of their role in upstream
events such as ligand binding and cell adhesion. This is important,
since mutations have been identified in the
3
cytoplasmic domain that inhibit both these processes (27, 52, 53).
Using this approach, we have demonstrated 1) that elements in both the
membrane-proximal and C-terminal segments of the
3
cytoplasmic domain are necessary to trigger FAK phosphorylation; 2)
that the conserved acidic residues in the membrane proximal region,
proposed to be important for FAK binding to the integrin, are not
required for this signaling event; 3) that the conserved NPXY motif is required for integrin
cytoplasmic domains
to trigger FAK phosphorylation, although a conservative tyrosine to
phenylalanine mutation is tolerated; and 4) that the naturally
occurring serine 752 to proline mutation that inhibits
IIb
3 activation also inhibits the ability
of the
3 cytoplasmic domain to trigger FAK
phosphorylation.
Although FAK can bind to peptides modeled after the membrane-proximal
region of cytoplasmic domains, FAK has not been co-precipitated with integrins. Therefore, it is not known whether FAK directly interacts with the
cytoplasmic domains of endogenous heterodimeric integrins and whether the linkage of FAK and integrins differs for
different integrin heterodimers and in different cell types. Nonetheless, clustering unoccupied integrins on the cell surface was
shown to result in the co-clustering of only FAK and tensin, as well as
the tyrosine phosphorylation of FAK (49), suggesting that additional
protein interactions are not required for FAK phosphorylation at least
when signaled by the clustering of some unoccupied integrins. However,
clustering ligand-occupied integrins or clustering integrins with
antibodies that mimic the clustering of ligand-occupied integrins was
shown to result in the co-clustering of additional cytoskeletal and
signaling proteins including actin, talin,
-actinin, paxillin, Src,
and FAK (49-51), as well as the phosphorylation of FAK (49). This
suggests that ligand occupancy may either direct the association of
additional proteins with FAK and tensin, or may dictate alternative
interactions between integrins and these and other cytoplasmic
components. Interestingly, clustering chimeric receptors containing a
cytoplasmic domain resulted in the colocalization of the same
cytoplasmic proteins as was observed for the clustering of
ligand-occupied integrins,2 as well as
triggering FAK phosphorylation (34). These results suggest that the
intracellular protein interactions mediated by the transfected chimeric
receptors are similar to those mediated by ligand-occupied
integrins.
Interestingly, the 1B and
3B cytoplasmic
domains contain the putative FAK binding domain (41); however, they do
not signal FAK phosphorylation (34, 37), suggesting that either the
amino acids inserted by alternative splicing interfered with the
ability of FAK to bind to
cytoplasmic domains within the cell, or
that these alternative amino acids inhibited additional or alternative protein interactions required for this signaling event. Our results with mutant cytoplasmic domains lacking either the membrane-proximal segment or the C-terminal segment indicated that structural information in both regions of the
3 cytoplasmic domain is important
for the regulation of integrin-triggered FAK phosphorylation and that the putative FAK binding domain alone is not sufficient to trigger FAK
phosphorylation. Consistent with this notion,
1
cytoplasmic domain deletion mutants lacking the C-terminal segment also
did not signal FAK phosphorylation (30). The ability of the chimeric receptors containing substitutions in the putative FAK binding domain
to signal FAK phosphorylation further suggested that FAK binding to
this region is not required for the integrin
3
cytoplasmic domain to trigger this signaling event. These results
suggest that perhaps integrin-cytoplasmic protein interactions other
than FAK binding to its putative binding site on the
3
cytoplasmic domain may couple integrins to the FAK signaling pathway,
and that additional structural information for this interaction must be
encoded in the C-terminal region of
cytoplasmic domains.
When we further analyzed the requirements in the C-terminal segment of
the 3 cytoplasmic domain for this signaling event, we
found that the NPXY motif present in the C-terminal segment of many integrin
cytoplasmic domains is important in the regulation of FAK phosphorylation. The NPXY motif may function by
regulating the overall conformation of the
cytoplasmic domain,
thereby influencing many protein interactions involving the
cytoplasmic domain involved in several integrin-mediated processes.
Alternatively, the NPXY motif may itself interact with
specific cytoplasmic components necessary to trigger FAK
phosphorylation. One of these proteins may be the cytoskeletal protein,
talin, since peptides spanning this motif can inhibit the association
of talin and integrins (56). Consistent with this notion, the
C-terminal segment of the
1 cytoplasmic domain has been
shown to be required for the association of both FAK and talin with
integrins (40, 51), as well as FAK phosphorylation (30). This suggests
a correlation between the association of talin and integrins with FAK
phosphorylation.
In contrast to these findings, others have recently demonstrated that
recombinant heterodimeric IIb
3 integrin
receptors containing only the putative FAK binding site of the
3 cytoplasmic domains were able to signal FAK
phosphorylation (45). These results differ from ours and those of Guan
et al. (30), which suggest that amino acids C-terminal to
the putative FAK binding site are also required to signal FAK
phosphorylation. Although the reasons for these differences are not
known, it seems likely that signaling pathways leading to FAK
phosphorylation may differ for different integrin heterodimers and
different cell types (46). Recently, it has been reported that
integrins can associate with transmembrane 4 proteins and that these
associations occur with only specific integrin heterodimers (47),
including
IIb
3 (48). Since these
associations may promote integrin clustering by extracellular interactions (47), they may serve in some cases to bypass a requirement
for cytoskeletal interactions mediated by the C-terminal segment of
integrin
cytoplasmic domains.
Several laboratories have mutagenized the conserved motifs within the
3 cytoplasmic domain and have assayed several different integrin functions. The inhibitory phenotypes reported are strikingly similar to those reported here for the function of the
3
cytoplasmic domain in triggering FAK phosphorylation. Mutations in the
NPXY motif have recently been shown to inhibit the
regulation of the affinity state of the platelet integrin,
IIb
3 (52), as well as other
integrin-mediated processes such as cell adhesion, cell spreading, cell
migration, and integrin localization to adhesion sites (53-55).
However, our results are the first demonstration to our knowledge of a
role for this motif in a specific post-ligand binding integrin
signaling event. Mutation of serine 752 to proline had similar
inhibitory effects (52, 54). In addition, mutations in the
NXXY motif, as well as alanine substitutions for serine 752, had intermediate phenotypes, not only on FAK phosphorylation described
in this report, but also on cell spreading and the regulation of
integrin affinity (52, 54), suggesting that the NXXY motif may modulate the binding of a protein required for integrin function in
these processes.
The similar inhibitory phenotypes described in this study and those
discussed above suggest that conserved motifs targeted in these studies
may function to regulate the overall conformation of the cytoplasmic domain. Therefore, mutations in these motifs may have
global effects on many different protein interactions. However, it is
also possible that since
intracellular domains are small in size,
yet function in many different integrin-mediated processes, they may
encode overlapping binding sites for several cytoplasmic proteins.
Additional mutagenesis, functional assays, and protein binding assays
will be necessary to dissect the specific interactions required to
mediate specific integrin functions.
We thank Drs. Jane Sottile, Randy Morse, and Anthony Mastrangelo for helpful comments during the preparation of this manuscript; Drs. Jun-Lin Guan and J. Thomas Parsons for generously providing polyclonal antibodies to FAK; and Todd Christian, David Devernoe, and Dr. Anthony Mastrangelo for help with flow cytometry.