Down-regulation of the filamentous actin cross-linking activity of cortactin by Src-mediated tyrosine phosphorylation.

Cortactin, a prominent substrate for pp60(c-src), is a filamentous actin (F-actin) binding protein. We show here that cortactin can promote sedimentation of F-actin at centrifugation forces under which F-actin is otherwise not able to be precipitated. Electron microscopic analysis after negative staining further revealed that actin filaments in the presence of cortactin are cross-linked into bundles of various degrees of thickness. Hence, cortactin is also an F-actin cross-linking protein. We also demonstrate that the optimal F-actin cross-linking activity of cortactin requires a physiological pH in a range of 7.3-7.5. Furthermore, pp60(c-src) phosphorylates cortactin in vitro, resulting in a dramatic reduction of its F-actin cross-linking activity in a manner depending on levels of tyrosine phosphorylation. In addition, pp60(c-src) moderately inhibits the F-actin binding activity of cortactin. This study presents the first evidence that pp60(c-src) can directly regulate the activity of its substrate toward the cytoskeleton and implies a role of cortactin as an F-actin modulator in tyrosine kinase-regulated cytoskeleton reorganization.

Cortactin (p80/p85) was initially discovered as a major phosphotyrosine-containing protein in v-Src-transformed chicken embryo fibroblasts (1). The murine homologue was independently isolated as a signaling molecule involved in the transition from G 0 to G 1 phase in response to fibroblast growth factor (2,3), and the human cortactin was found as an oncogene that is frequently amplified in subsets of tumors and tumor cell lines (4,5). A strong association of cortactin with F-actin 1 has been described (14). Consistent with its F-actin binding activity, cortactin primarily localizes within peripheral cell structures such as lamellipodia, pseudopodia, and membrane ruffles (9,14), which are enriched for cytoskeletal proteins. However, unlike many other F-actin-binding proteins, the protein sequence of cortactin features a unique structure characterized by six and a half 37-amino acid tandem repeats and a Src homology 3 (SH3) domain at the carboxyl terminus. Between the SH3 and the repeat domains are an ␣-helix domain and a sequence region rich in proline residues.
Recent evidence has indicated that cortactin is a prominent substrate for Src-related protein-tyrosine kinases (1, 6 -8). Furthermore, cortactin is implicated in signaling mediated by multiple extracellular stimuli including fibroblast growth factor (3), epidermal growth factor (9), thrombin (10), integrin (11), bacteria-mediated cell invasion (12), and mechanical strain (13). While tyrosine phosphorylation of cortactin is a profound phenomenon in response to many extracellular stimuli, the biological function of cortactin and the physiological role of tyrosine phosphorylation are not clear.
In an attempt to elucidate the function of cortactin, we have examined biochemical properties of cortactin. The study presented here demonstrated that cortactin is a potent F-actin cross-linking protein. Most importantly, the F-actin cross-linking activity is down-regulated upon phosphorylation mediated by pp60 c-src . Thus, cortactin may act as an important mediator for intracellular tyrosine kinases in regulating the cytoskeleton reorganization in vivo.

EXPERIMENTAL PROCEDURES
Preparation of Recombinant Cortactin-Murine cortactin was expressed in Escherichia coli as a glutathione S-transferase fusion protein in pGEX-2T plasmid and purified by affinity chromatography using glutathione-Sepharose (Pharmacia Biotech Inc.) as described previously (15). The glutathione S-transferase part of the fusion protein was removed by cleavage with bovine thrombin (ICN) in a digestion buffer (50 mM Tris-HCl, pH 8.2, containing 100 mM NaCl and 1 mM CaCl 2 ) for 3ϳ4 h at room temperature. The digested materials were loaded onto a DEAE-Sepharose FF (Pharmacia) column and eluted with 200 ml of elution buffer (20 mM Tris-HCl, pH 7.6, containing 1 mM MgCl 2 , 1 mM dithiothreitol, and 1 mM EGTA, and KCl with a gradient concentration from 20 to 600 mM). The fractions containing cortactin were pooled, and undigested fusion proteins were removed by additional chromatography using glutathione-Sepharose. The concentration of purified cortactin was determined by the D c protein assay (Bio-Rad) according to the manufacturer's instructions.
Preparation of Actin-Actin was purified from an acetone powder of rabbit skeletal muscle according to Pardee and Spudich (16). Pyrenelabeled actin was prepared as described by Kouyama and Mihashi (17). The labeled actin was further purified by chromatography using Sephadex G-150 (Pharmacia). Globular actin was polymerized into filaments by adding KCl, MgCl 2 , and ATP to the final concentrations of 134, 1, and 1 mM, respectively, and incubated for at least 4 h at room temperature.
Phosphorylation of Cortactin by pp60 c-src -Recombinant human pp60 c-src (18) was preactivated and maintained in a buffer containing 40 M ATP, 0.8 mM MgCl 2 , and 1 mg/ml bovine serum albumin at 0°C. To prepare tyrosine-phosphorylated cortactin, purified proteins were incubated with various amounts of preactivated pp60 c-src in 20 l of kinase buffer (50 mM Tris-HCl, pH 7.4, containing 5 mM MgCl 2 , 5 mM ATP, and 2 Ci of [␥-32 P]ATP, 6000 ci/mmol) at room temperature for 1 h. For F-actin cross-linking analysis, phosphorylated cortactin proteins were diluted to a final volume of 50 l of which the final concentrations of KCl and MgCl 2 were readjusted to 134 and 2 mM, respectively. To * This study was supported by National Institutes of Health Grant R29 HL52753 (to X. Z.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

FIG. 1. Sedimentation of cortactin-cross-linked F-actin. A, pyrene-labeled filaments at a concentration of 4 M in the presence (shaded bar)
or absence (open bar) of cortactin (400 nM) were incubated for 30 min at room temperature and subsequently subjected to centrifugation at different forces as indicated. Fluorescence in the supernatant was measured, and its decrease was used as an indication of the precipitation of F-actin. B, cortactin at different concentrations was incubated with pyrene-labeled F-actin for 30 min followed by centrifugation at 26,300 ϫ g for 10 min. The F-actin cross-linking activity was determined by fluorescence measurement as described above. The value of each point represents the mean of four independent experiments. C, the pellets from experiment B were resolved in SDS sample buffer, and aliquots of each sample were applied to a SDS-PAGE gel. Lane 1, without cortactin; lanes 2-5, with cortactin at concentrations of 26, 53, 106, and 212 nM, respectively. The bands corresponding to cortactin and actin are indicated. D, pyrene-labeled F-actin at a concentration of 4 M was incubated with cortactin (200 nM) for the times indicated and subsequently centrifuged at 84,000 ϫ g for 2 min. F-actin cross-linking was determined as described above. confirm and quantitate phosphorylated cortactin, aliquots of the reactions were combined with an equal volume of 2 ϫ SDS sample buffer (19) and fractionated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). The gel was stained with Coomassie Blue, and the phosphorylated proteins were visualized by autoradiography.
F-actin Cross-linking Assay-Purified cortactin at the indicated concentrations in 50 l of TKM buffer (50 mM Tris-HCl, pH 7.4, containing 134 mM KCl and 1 mM MgCl 2 ) was mixed with an equal volume of 8 M pyrene-labeled F-actin and incubated for 30 min at room temperature. The mixture was then immediately centrifuged at 26,300 ϫ g for 10 min at room temperature in a Beckman TL-100 centrifuge. The supernatant was carefully transferred to a new tube and mixed with 300 l of TKM buffer containing 2 M phalloidin (Sigma). The fluorescence in the supernatant was recorded on an LS50B luminescence spectrometer (Perkin-Elmer) at an excitation wavelength of 370 nm with a slit of 2.5 nm and an emission wavelength of 410 nm with a slit 6 nm, respectively. The decrease in the supernatant fluorescence reflects the precipitation of F-actin due to cross-linking. In some experiments, the precipitated pellets were directly analyzed by SDS-PAGE.
Electron Microscopic Analysis-F-actin or the mixture of F-actin and cortactin was mixed with an equal volume of 1% phosphotungstic acid in 0.1 M phosphate buffer, pH 7.4. The stained samples (10 l) were absorbed to a 0.25% Formvar 15/95E resin-(Sigma) coated film on a gold grid. The grid was then air-dried overnight. Transmission electron micrographs were taken with a Philips M12 microscope.

RESULTS AND DISCUSSION
In an effort to investigate the function of cortactin, the murine cortactin was expressed in E. coli as a glutathione Stransferase fusion protein. The glutathione S-transferase-free cortactin was used for the evaluation of its F-actin cross-linking activity by virtue of a co-sedimentation assay (20). To determine optimal conditions for the assay, cortactin was incubated with F-actin for 30 min and subjected to centrifugation at different forces from 13,000 to 30,000 ϫ g for 10 min. Under these conditions, neither filaments in the absence of cortactin (Fig. 1A) nor cortactin alone could be precipitated (data not shown). However, the presence of both cortactin and F-actin resulted in a dramatic increase in the amount of F-actin associated with pellets in a dose-dependent manner (Fig. 1B). The half-maximum effect requires an approximately 1:100 molecular ratio of cortactin to actin, and the maximum sedimentation (80% of total F-actin) can be reached in the presence of a 1:36 ratio of cortactin to actin. The sedimentation of F-actin induced by cortactin is also correlated with the association of cortactin with F-actin, as shown by their co-sedimentation at a ratio of 1 cortactin molecule to approximately 15 actin subunits (Fig.   1C), which is in agreement with the stoichiometry for the binding of cortactin to actin (14). The two bands of 85 and 90 kDa shown on the gel most likely reflect different conformations of cortactin, since only one band of 85 kDa was visualized when SDS-PAGE was performed in the presence of 5 M urea (data not shown). A kinetic study further demonstrated that cortactin-mediated F-actin precipitation is a rapid process (Fig.  1D). Incubation with cortactin for 2 min resulted in a sedimentation of more than 60% of F-actin, although the maximum sedimentation (80% of total F-actin) only occurred at 30 min after interaction. The cross-linked F-actin in the presence of cortactin was further examined by electron microscopy after negative staining. Filaments prepared in the absence of cortactin displayed individual single strands (Fig. 2, A and C). In the presence of cortactin, however, most F-actin strands became thicker and formed a bundle-like structure (Fig. 2, B and D).
Since activities of many F-actin cross-linking proteins are dependent on Ca 2ϩ and pH, we examined the effects of Ca 2ϩ and pH on the activity of cortactin. In the presence of either EGTA or various concentrations of Ca 2ϩ , the F-actin crosslinking activity of cortactin was not affected (data not shown). However, when the cross-linking assay was performed at different pH values, the optimal sedimentation of F-actin was observed in a range from pH 7.3 to 7.5. The F-actin sedimentation induced by cortactin at pH 6.9 and 8.2 is only approximately 30% of that at pH 7.4 (Fig. 3). The apparent effect of pH on the F-actin cross-linking is not due to its potential effect on actin polymerization, since pH has little influence on the stable polymerization of actin (Fig. 3).
The purified cortactin can be efficiently phosphorylated by pp60 c-src exclusively at tyrosine residues in a manner depending on time and the amount of Src. By incubating cortactin with 500 nM pp60 c-src at room temperature for 1 h, a maximum phosphorylation of cortactin was reached (Fig. 4A). When the phosphorylated cortactin was used in the co-sedimentation assay, a dramatic inhibition for the F-actin cross-linking was observed (Fig. 4B). The treatment of cortactin with 62.5 nM pp60 c-src reduced the efficiency of F-actin sedimentation from near 60% to approximately 38%, and that with 500 nM pp60 c-src reduced further the sedimentation of F-actin to 10%. In a control experiment where cortactin was treated with the buffer only but in the absence of pp60 c-src , no reduction of the F-actin cross-linking was detected (Fig. 4B).
The apparent decrease in the F-actin cross-linking was not due to a possible inhibitory activity of pp60 c-src , since a solution in the presence of pp60 c-src itself did not have any detectable effect on F-actin cross-linking (data not shown). Furthermore, we examined whether tyrosine phosphorylation is essential for the inhibition of cortactin's F-actin cross-linking activity. We carried out the phosphorylation of cortactin in a kinase buffer in the absence of Mg 2ϩ , on which the kinase activity of pp60 c-src is dependent (21). As shown in Fig. 5A, the Src-mediated tyrosine phosphorylation of cortactin was abolished in the absence of Mg 2ϩ . When the Src-treated cortactin in the absence of Mg 2ϩ was mixed with F-actin and subsequently subjected to the co-sedimentation analysis, a significant amount of F-actin was detected in the pellet (Fig. 5B, column 4). Although the level of sedimentation of F-actin induced by cortactin and pp60 c-src in the absence of Mg 2ϩ was about 33% lower than that by non-Src-treated cortactin (Fig. 5B, compare columns 2 and 4), the lower efficiency could be the result of a trace amount of Mg 2ϩ present in the F-actin buffer, which may have partially restored the kinase activity of pp60 c-src . Indeed, when Src-treated cortactin in the absence of Mg 2ϩ was further incubated with F-actin, cortactin was able to be phosphorylated to the extent of approximately 25% of that in the regular kinase buffer (Fig. 5A,   FIG. 3. Effects of pH on the F-actin cross-linking  column 4). Taken together, these data demonstrate that the inhibition of the F-actin cross-linking activity of cortactin by pp60 c-src is dependent on tyrosine phosphorylation.
Cortactin has been previously described as a potent F-actinbinding protein (14). Therefore, we also examined the effect of pp60 c-src on its F-actin binding activity by co-sedimentation at a high centrifugation force of 366,000 ϫ g, at which actin filaments are able to be precipitated. As shown in Fig. 6, the Src treatment resulted in an inhibition of the F-actin binding activity of cortactin. However, the efficiency of the inhibition is apparently less than that for the F-actin cross-linking. At the concentration of 62.5 nM pp60 c-src , the co-precipitated cortactin with F-actin was reduced only from 95 to 85%; at 500 nM pp60 c-src , 42% of cortactin was still bound to F-actin. The mod-erate inhibition by pp60 c-src may be due to the existence of multiple F-actin binding sites in cortactin, which are involved in the F-actin cross-linking activity (30). The lower sensitivity to pp60 c-src could also be the reason for a failure to observe the inhibition of the F-actin binding activity of cortactin in a system using lysates from v-Src-transformed cells (14). However, we cannot rule out the possibility that different phosphorylation sites or additional kinase(s) may be involved in that system.
The dependence on neutral pH for the optimal cross-linking activity of cortactin is uncommon. For example, ␣-actinin has the highest cross-linking activity at pH 6.8 (22). Talin shows a reduced actin cross-linking activity when pH is increased from 6.5 to 7.3, whereas its optimal activity is at pH 6.5 (23). EF1␣  1 and 2), and the other half were further incubated with F-actin for an additional 30 min (columns 3 and 4). All reactions were analyzed by SDS-PAGE followed by autoradiography. The relative levels of tyrosine phosphorylation of cortactin were determined by densitometry analysis. B, cortactin treated with pp60 c-src in either the presence (column 3) or absence (column 4) of Mg 2ϩ was subjected to F-actin cross-linking analysis. Column 1, no cortactin; column 2, cortactin without pp60 c-src treatment.
has high F-actin-cross-linking activity at low pH (6.2-6.5) (24). It has been well recognized that pH is involved in the regulation of the actin cytoskeleton (24) and cell motility (25,26). A recent study also indicates that the induction of Rho on stress actin filaments is dependent on Na ϩ /H ϩ exchange (27). Since both intracellular pH and tyrosine kinase activities are regulated by a variety of extracellular signals (28,29), cortactin could act as an important mediator for ligands to regulate the cytoskeleton reorganization.