(Received for publication, March 4, 1997, and in revised form, May 5, 1997)
From the Department of Experimental Pathology and the
§ Department of Platelet Biology, The Holland Laboratory,
American Red Cross, Rockville, MD 20855, the ¶ Division of
Biology, Glaxo and Wellcome Research Institute, Research Triangle
Park, North Carolina 27709, and the ¶ Department of Anatomy and
Cell Biology, George Washington University,
Washington, D. C. 20037
Cortactin, a substrate of
pp60c-src and a potent filamentous actin
binding and cross-linking protein, is abundant in circulating platelets. After stimulation of platelet aggregation with collagen, cortactin undergoes a dramatic increase in tyrosine phosphorylation followed by a rapid degradation. The cleavage of platelet cortactin was
detected in lysates prepared using either Triton-containing buffer or
SDS-sample buffer. However, the degradation of cortactin was not
observed in platelets derived from a Glanzmann's patient, who lacked
functional integrin IIb
3 (GPIIb-IIIa). In
addition, the proteolysis of cortactin was abolished by treating
platelets before but not after collagen stimulation with EGTA or
calpeptin. Furthermore, recombinant cortactin was digested by
µ-calpain in vitro in a dose-dependent
manner, indicating that cortactin is a substrate for calpain. We also
observed that the calpain-mediated digestion in vitro is
dependent on the presence of a sequence containing a proline-rich
region and multiple tyrosine residues that are phosphorylated by
pp60c-src. Tyrosine phosphorylation by
pp60c-src up-regulates the activity of calpain
toward cortactin. Our data suggest that the calpain-mediated
proteolysis of tyrosine-phosphorylated cortactin may provide a
mechanism to remodel irreversibly the cytoskeleton in response to
platelet agonists.
Cortactin, an F-actin1 binding and cross-linking protein (1, 12), is a major target for tyrosine phosphorylation in response to signaling mediated by fibroblast growth factor (2), epidermal growth factor (3), integrin activation (4), bacteria-mediated phagocytosis (5), and v-src oncogene (6). Overexpression or amplification of the human cortactin gene (also called EMS1) is often associated with human malignancies (7, 8). In v-Src-transformed cells, cortactin has been found to co-localize with Src oncoproteins within podosomes, membrane-substratum contact structures (6). Analysis of cortactin phosphorylation in cells lacking the c-src gene (9) or following overexpression of c-Csk (10), a negative regulator for pp60c-src, has provided further compelling evidence that cortactin is an intrinsic substrate for pp60c-src.
The protein sequence of cortactin is unique because it contains six and
one-half 37-amino acid tandem repeats near the NH2 terminus, and a Src homology 3 (SH3) domain at the carboxyl-terminal end. Between the repeat and the SH3 domain is an -helix, a
proline-rich region, and multiple tyrosine residues. The amino acid
sequence of human cortactin within the repeat domain shares nearly
100% identity with the chicken and murine homologues and 70% with
HS1, a cortactin-related gene product (11), indicating that the repeat domain plays a fundamental role for cortactin (12). Indeed, the repeat
domain has been demonstrated as the binding site for F-actin (12). In
contrast, the sequence between the
-helix and the SH3 domain
exhibits less than 33% identity to HS1, but the function of this
region has not yet been identified.
We recently reported that there is abundant expression of cortactin in
megakaryocytes and platelets (13). While tyrosine phosphorylation of
cortactin has been described as a major phenomenon in
thrombin-stimulated platelets (14, 15), the significance of the
tyrosine phosphorylation is unknown. In the present study, we examined
the fate of cortactin in platelets stimulated by collagen. We found
that cortactin is degraded following tyrosine phosphorylation and that
the protease responsible for the cortactin degradation is a
calpain-related enzyme, which requires integrin
IIb
3. Furthermore, we provide in
vitro evidence that the sequence containing the proline-rich
region and multiple tyrosine residues targeted by pp60c-src is required for the
calpain-mediated cleavage. Finally, we demonstrated that tyrosine
phosphorylation of cortactin by pp60c-src
dramatically alters its susceptibility to calpain. These data suggest that tyrosine phosphorylation may play a role in the
calpain-mediated proteolysis of cortactin.
Monoclonal antibody (mAb) 4F11 was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Monoclonal antibody against phosphotyrosine (RC20) was from Transduction Laboratories (Lexington, KY). The polyclonal antibody against the C-terminal part of murine cortactin was derived from mice immunized with a recombinant protein corresponding to amino acids 323-546. Prostaglandin E1, phenylmethylsulfonyl fluoride, Triton X-100, EGTA, benzamidine, leupeptin, and aprotinin were from Sigma. SDS-PAGE markers were from Bio-Rad. Protein A-Sepharose was from Pharmacia Biotech Inc. Purified µ-calpain derived from pig erythrocytes was obtained from ICN (Costa Mesa, CA). Calpeptin was from Biomol (Plymouth Meeting, PA). Sodium orthovanadate was from Fisher. Type I tendon collagen was from Chrono-Log Co. (Havertown, PA).
Isolation of Human PlateletsHuman blood (500 ml) from healthy volunteers was collected into 70 ml of CPD solution, containing 1.84 mg of sodium citrate, 1.78 mg of dextrose, 209 mg of citric acid, and 155 mg of monobasic sodium phosphate. Platelet-rich plasma was obtained by centrifugation at 200 × g for 16 min at ambient temperature. Blood from a Glanzmann's patient (female) was kindly provided by Robert Abel (Christina Hospital, Wilmington, DE). Citric acid and prostaglandin E1 were added to platelet-rich plasma to final concentrations of 4 mM and 1 µg/ml, respectively. The platelet-rich plasma was then centrifuged at 700 × g for 10 min. The platelet pellet was resuspended in washing buffer (4.26 mM NaH2PO4, 7.46 mM Na2HPO4, pH 6.5, containing 5.5 mM dextrose, 128 mM NaCl, 4.77 mM sodium citrate, 2.35 mM citric acid, and 3.5 mg/ml of bovine serum albumin) and centrifuged at 700 × g for 10 min. The pellet was then resuspended in a modified Tyrode-Hepes buffer (10 mM Hepes, pH 7.35, containing 136.7 mM NaCl, 5 mM glucose, 2.6 mM KCl, 13.8 mM NaHCO3, 1.0 mM MgCl2, 0.36 mM NaH2PO4, and 3.5 mg/ml bovine serum albumin) at 1 × 109 platelets/ml.
Analysis of Cortactin in Triton-solubilized PlateletsCollagen at a final concentration of 2.5 µg/ml was added to the washed platelets (0.6 ml of 1 × 109 cells/ml) in the presence of 1 mM CaCl2 in an aggregometer cuvette at 37 °C for the times indicated. Activated platelets were immediately lysed by adding 200 µl of 4 × Triton lysis buffer (200 mM Tris-HCl, pH 7.2, containing 4% Triton X-100, 20 mM EGTA, 40 µg/ml leupeptin, 40 µg/ml aprotinin, 4 mM phenylmethylsulfonyl fluoride, 4 mM benzamidine, and 4 mM Na3VO4). The lysates were centrifuged at 15,000 × g for 10 min. The pellet (insoluble fraction) was solubilized by adding an equal volume of 2 × SDS sample buffer (16). The soluble fractions were subjected to immunoprecipitation by mAb 4F11 (2.5 µg/ml) as described previously (2). The immunoprecipitates were washed once with 1 × Triton-lysis buffer, resuspended in 2 × SDS sample buffer, and analyzed by immunoblotting analysis with either mAb 4F11 or RC20 as described previously (2).
Analysis of Cortactin in SDS-lysed PlateletsStirred platelets (0.4 ml of 1 × 109 cells/ml) were stimulated with collagen at 37 °C and lysed by adding an equal volume of 2 × SDS sample buffer including 10 mM EGTA and 20 µM calpeptin. The platelet proteins were separated by SDS-PAGE, and cortactin was detected by immunoblot analysis.
Preparation of Recombinant Cortactin ProteinsTo prepare
GST-cortactin, a DNA fragment of 182 bp was generated by polymerase
chain reaction. The oligonucleotide ACTCGTGGATCCTGGAAAGCCTCTGCA was
used as the 5 primer and contained a BamHI site; the
oligonucleotide CTTGAGCGTCTGGTGTT was used as the 3
primer and
contained an Xmn1 site. The amplified fragment was ligated
with a DNA fragment derived from the digestion of a cDNA clone
encoding the murine cortactin (2) with Xmn1 and
EcoRI, and the ligated product was then cloned into
BamHI and EcoRI sites of PGEX-2T (Pharmacia). The
cortactin variant Cort
496-546 was prepared as follows.
A DNA fragment of 370 base pairs was prepared by polymerase chain
reaction. The oligonucleotide CGAGAGAGCTCAGCGGATGGCC was used as the 5
primer and contained a SacI site; the oligonucleotide
ACTGCAGAATTCTAGATGGCTGTGATGCC was used as the 3
primer and contained
an EcoRI site and a stop codon. The amplified fragment was
cloned into a DNA fragment derived from the digestion of GST-cortactin
with SacI and EcoRI. Cort
375-546 was prepared in the same way as Cort
496-546 except that
the oligonucleotide CGAGAGAGCTCAGTAACGGATCGCCAAAGAA was used as the 5
primer and contained a stop codon and a SacI site. All polymerase chain reaction-generated fragments were confirmed by DNA
sequencing. Cortactin and its variants were expressed in
Escherichia coli as glutathione S-transferase
fusion proteins and purified by affinity chromatography using
glutathione-Sepharose as described previously (17). The purified
glutathione S-transferase fusion proteins were further
digested with thrombin, and the glutathione S-transferase-free proteins were purified using
glutathione-Sepharose and DEAE-Sepharose (1).
Purified
recombinant cortactin (3.6 µg) was incubated with µ-calpain for 90 min at different concentrations in 40 µl of reaction buffer (50 mM Tris-HCl, pH 7.36, containing 134 mM KCl, 1 mM MgCl2, 75 µM EGTA, and 75 µM CaCl2). The reaction was terminated by
adding an equal volume of 2 × SDS sample buffer, and the proteins
were separated by a gradient SDS-PAGE gel (4-20%, w/v). The digested proteins were visualized by either Coomassie Blue staining or immunoblotting with mAb 4F11 or a polyclonal antibody directed against
a peptide encoding the amino acid sequence from the -helix to the
SH3 domain.
To
evaluate the role of cortactin in platelet aggregation, we examined
tyrosine phosphorylation of cortactin in collagen-stimulated platelets.
Activated platelets were lysed using a Triton X-100-containing buffer.
The soluble fractions were subjected to immunoprecipitation with 4F11,
a mAb recognizing the repeat domain of cortactin (6). The pellets were
solubilized in SDS sample buffer. Proteins in both fractions were
immunoblotted using either a polyclonal antibody against cortactin or a
mAb against phosphotyrosine. As shown in Fig.
1A, stimulation of stirred platelets with
collagen caused a dramatic increase in the level of tyrosine
phosphorylation of cortactin after 15 s and a maximum
phosphorylation at 45 s, which was concomitant with platelet
aggregation (data not shown). However, the level of phosphorylated
cortactin declined slightly after 1 min of stimulation, and this
coincided with a decrease in the level of cortactin in the soluble
fraction (Fig. 1B).
We examined the possibility that a reduced amount of cortactin in the soluble fraction in response to collagen could be a consequence of the cytoskeletal translocation that has been described previously in thrombin-stimulated platelets (15, 18). Immunoblot analysis of cortactin in the insoluble fraction demonstrated that the stimulation of platelets with collagen enhanced tyrosine phosphorylation of multiple proteins including those that migrated at the positions for cortactin (Fig. 1C). However, the amount of cortactin associated with the insoluble fraction was only transiently increased during the period from 30 to 45 s and diminished afterward (Fig. 1D), suggesting that platelet cortactin, in either the soluble or insoluble fractions, was degraded after collagen stimulation. The degradation of cortactin appears not caused by a nonspecific proteolysis because pp60c-src associated with the pellets was not degraded under the same conditions even after a prolonged stimulation (Fig. 1D).
To confirm that the apparent degradation of cortactin was not the
result of a protease released during the Triton-mediated lysis, we
analyzed cortactin in platelet lysates that were prepared by direct
lysis in SDS-sample buffer. The results from these experiments were
compared with the pattern of cortactin degradation prepared in Triton
X-100 buffer. As shown in Fig. 2A,
significant amount of degraded cortactin was detected in the SDS-lysed
whole platelets after collagen stimulation, although the extent of the
degradation, especially at early phases of stimulation (30 and 45 s), appeared to be less than that of Triton-lysed platelets. However,
the degradation patterns in both lysates are similar (Fig.
2B).
Calpain-related Protein Is the Major Protease Responsible for the Proteolysis of Cortactin in Platelets
Calpain is a family of
calcium-dependent cysteine proteases that are abundantly
present in platelets and are activated during platelet aggregation (19,
20). As shown in Fig. 3A, EGTA treatment of
platelets significantly inhibited the degradation of cortactin as
compared with untreated platelets. Furthermore, treatment with calpeptin, a specific membrane-permeable peptide-derivative inhibitor for calpain (21), resulted in the same reduction of cortactin degradation (Fig. 3B, part a). However, when a
lysis buffer containing either EGTA or calpeptin was used to lyse
activated platelets, no significant inhibition of cortactin degradation
was observed (Fig. 3B, parts b and c).
This result further confirms that the degradation of cortactin
primarily occurs prior to platelet lysis.
Degradation of Cortactin in Platelets Is Dependent on
In platelets, the influx of
calcium can be regulated by the activation of
IIb
3 (22, 23), a major integrin on the
surface of platelets. To evaluate the role of
IIb
3 in the proteolysis of cortactin, we
examined tyrosine phosphorylation of cortactin in platelets from a
Glanzmann's patient. As shown in Fig. 4, normal platelets exhibited a 60% reduction in the amount of intact cortactin after 3 min of collagen stimulation. In contrast, no significant reduction was found with the Glanzmann's platelets, suggesting that
the degradation of cortactin requires
IIb
3 under identical conditions.
Interestingly, the induction of tyrosine phosphorylation of cortactin
in response to collagen in the Glanzmann's platelets was not impaired
(Fig. 4), implying that tyrosine phosphorylation of cortactin is a
process independent of
IIb
3.
Cleavage of Cortactin by µ-Calpain in Vitro Is Dependent on the Presence of a Sequence Containing the Proline-rich Region and Multiple Tyrosine Residues
Purified µ-calpain (calpain-I) digests
recombinant murine cortactin in vitro in a
dose-dependent manner (Fig. 5A).
At a concentration of 6.2 µg/ml of calpain, approximately 90% of the
cortactin proteins were digested to multiple fragments. Interestingly,
many of the digested fragments were reactive to mAb 4F11, which
specifically recognizes the repeat domain of cortactin (12), but not to
an antibody directed against the region between the repeat and the carboxyl terminus (Fig. 6A). This implies
that the sequence in this region may be more susceptible to calpain. To
verify this, we analyzed two cortactin variants,
Cort496-546, which lacks the SH3 domain, and
Cort
375-546, which lacks the sequence from the
proline-rich region to the carboxyl terminus (Fig. 6B,
upper part). As with the wild-type cortactin, the mutant Cort
496-546 was efficiently digested by calpain (Fig. 6B, lower part). In contrast, little digestion of
the mutant Cort
375-546 was detected under the same
conditions, indicating that the sequence of amino acids 375-496, which
contains the proline-rich region and multiple tyrosine residues, may be
involved in the calpain-mediated proteolysis.
Src-mediated Tyrosine Phosphorylation Increases the Susceptibility of Cortactin to Calpain
Amino acids 375-496 contain multiple
tyrosine residues that can be targeted for phosphorylation by
pp60c-src.2 Thus, we
performed a calpain digestion of cortactin phosphorylated by
pp60c-src. Fig. 7 shows that most
phosphorylated cortactin proteins were digested nearly completely
within 2 min. In contrast, significant amounts of nonphosphorylated
cortactin remained even after 20 min of digestion under the same
conditions. However, when partially digested phosphorylated cortactin
was analyzed by SDS-PAGE and compared with nonphosphorylated cortactin,
we did not observe any significant difference in the two patterns (Fig.
7). Therefore, it is likely that tyrosine phosphorylation enhances the
efficiency of calpain-mediated digestion without altering its cleavage
sites.
It is unclear whether platelet calpain-mediated proteolysis occurs during lysis of cells or within aggregated platelets (24, 25). Our data indicate that the proteolysis of cortactin occurs within activated platelets. We detected the degradation of cortactin in whole platelets prepared by direct lysis in a SDS-sample buffer (Figs. 2 and 4). In addition, the calpain inhibitors EGTA and calpeptin block the degradation when they are applied before but not after platelet activation. Finally, it appears that the calpain-mediated proteolysis of cortactin is not the result of a nonspecific proteolysis, because pp60c-src, another substrate for calpain (26), was not degraded under the same conditions that allow cortactin proteolysis (Fig. 1D). However, it should be pointed out that the degree of the proteolysis of cortactin in activated platelets appears to vary depending on the method of lysing platelets. There is more extensive degradation found in Triton-lysed platelets than in SDS sample buffer (Fig. 2). This may be due to the fact that the soluble cortactin becomes more vulnerable to calpain released after lysis.
In agreement with other reports (15, 18), we observed that cortactin undergoes a transient translocation into the Triton-insoluble fraction between 30 and 45 s after collagen stimulation (Fig. 1D); however, the role of the translocation in this proteolysis of cortactin is not clear. While cortactin is a potent F-actin binding and cross-linking protein, the presence of F-actin does not apparently change the efficiency of the proteolysis of cortactin in vitro (data not shown). Furthermore, a cortactin mutant able to bind to F-actin but lacking the sequence from the proline-rich region to the carboxyl terminus is not efficiently digested by calpain (Fig. 6B). Hence, it is unlikely that the F-actin binding is a rate-limiting step for the calpain digestion.
Many cytoskeleton-associated proteins have been reported to be
substrates for calpain. These include actin-binding proteins (27),
vitronectin (28), protein-phosphotyrosine phosphatase 1B (29), integrin
3 subunit (30), talin (27), spectrin (31), and protein
kinase C (32). As with many of those substrates, proteolysis of
cortactin appears to be dependent on
IIb
3
because it does not occur in Glanzmann's platelets lacking functional
IIb
3 (Fig. 4). We have found, however,
that the absence of
IIb
3 does not affect
collagen-induced tyrosine phosphorylation of cortactin. Our finding is
in agreement with a previous report, which also showed increased
tyrosine phosphorylation in thrombin-treated platelets derived from
Glanzmann's patients (15). Furthermore, tyrosine phosphorylation of
cortactin can be detected after 15 s of stimulation (Fig.
1A). This is prior to platelet aggregation, which occurs
30-45 s after stimulation. Thus, tyrosine phosphorylation of cortactin
is kinetically correlated with the activation of pp60c-src, which occurs in the early phase prior to
the activation of
IIb
3 during platelet
stimulation (33). These data indicate that the Src-mediated tyrosine
phosphorylation of cortactin could be involved in the calpain-mediated
digestion. The importance of Src in the digestion of cortactin is
further highlighted by our findings that the digestion of recombinant
cortactin by µ-calpain is dependent on the presence of a sequence
containing multiple tyrosine residues targeted by
pp60c-src (Fig. 6), and the efficiency of the
digestion of cortactin in vitro is dramatically increased by
pp60c-src (Fig. 7).
Calpain-digested cortactin in vitro has significantly less F-actin cross-linking activity (data not shown). Interestingly, the F-actin cross-linking activity can be also down-regulated by tyrosine phosphorylation without degradation (1). These dual mechanisms regulating cortactin may be required to ensure the irreversible shape change associated with activated platelets. It is also noteworthy that both mechanisms involve the same structural region between the proline-rich motif and the SH3 domain. This may suggest the importance of this region in the regulation of cortactin function. Since calpain and cortactin are widely expressed in many mammalian cells, future studies using a structure-function approach should reveal the significance of calpain-mediated cleavage of cortactin in cellular cytoskeletal reorganization.
We thank Graham Jamieson and Allan Mufson for critical reading of the manuscript and Diana Norman for expert secretarial support.