Regulation of Adhesion Dynamics by Calpain-mediated Proteolysis of Focal Adhesion Kinase (FAK)*

The coordinated and dynamic regulation of adhesions is required for cell migration. We demonstrated previously that limited proteolysis of talin1 by the calcium-dependent protease calpain 2 plays a critical role in adhesion disassembly in fibroblasts (Franco, S. J., Rodgers, M. A., Perrin, B. J., Han, J., Bennin, D. A., Critchley, D. R., and Huttenlocher, A. (2004) Nat. Cell Biol. 6, 977–983). However, little is known about the contribution of other calpain substrates to the regulation of adhesion dynamics. We now provide evidence that calpain 2-mediated proteolysis of focal adhesion kinase (FAK) regulates adhesion dynamics in motile cells. We mapped the preferred calpain cleavage site between the two C-terminal proline-rich regions after Ser-745, resulting in a C-terminal fragment similar in size to the FAK-related non-kinase (FRNK). We generated mutant FAK with a point mutation (V744G) that renders FAK resistant to calpain proteolysis but retains other biochemical properties of FAK. Using time-lapse microscopy, we show that the dynamics of green fluorescent protein-talin1 are impaired in FAK-deficient cells. Expression of wild-type but not calpain-resistant FAK rescues talin dynamics in FAK-deficient cells. Taken together, our findings suggest a novel role for calpain proteolysis of FAK in regulating adhesion dynamics in motile cells.

The coordinated and dynamic regulation of adhesions is central for cell migration in both normal and pathological processes (1). Cell migration is initiated by forming protrusions, which are stabilized by integrin-mediated adhesions that establish structural and signaling linkages between the extracellular matrix and the actin cytoskeleton. Subsequently, actomyosin contractility drives forward propulsion, and finally, adhesions are disassembled to allow for directional movement. Thus, it is not surprising that efficient cell migration necessitates precise spatial and temporal control of these events. However, the mechanisms governing the formation and disassembly of adhesions are still not well understood.
A prominent component involved in this regulation is focal adhesion kinase (FAK). 2 FAK promotes cell migration by its capacity to orchestrate signals between integrin and growth factor receptors (2). Downstream of integrin or growth factor stimulation, FAK is phosphorylated at Tyr-397, which is an important binding site for Src family kinases (3). Previous studies have demonstrated a critical role for FAK as a regulator of adhesion dynamics (4 -7). In addition, FAK has been shown to mediate the tyrosine phosphorylation of p190RhoGAP as a negative regulator of Rho activity in focal adhesion turnover and polarity (5,8). Furthermore, microtubule-induced focal adhesion disassembly requires FAK and dynamin (9). Nevertheless, the mechanisms by which FAK regulates the disassembly of focal adhesions remain to be elucidated.
Evidence has emerged supporting the role of the calpain family of intracellular calcium-dependent proteases in regulating cell migration (10 -14). Calpains have been proposed to regulate migration, at least in part, through their ability to modulate the dynamics of adhesions (15). Numerous calpain targets have been identified, some of which are proteins that are present in focal adhesions, including talin, paxillin, and FAK (16). Thus, it is likely that proteolysis of these substrates contributes to the regulation of adhesion dynamics and cell migration.
The involvement of both FAK and calpain in regulating the turnover of adhesions prompted us to investigate the cleavage of FAK by calpain as a possible mechanism by which FAK affects adhesion dynamics. We demonstrated previously that calpain-mediated proteolysis of talin regulates adhesion dynamics (17). Here, we show that FAK also regulates talin dynamics. We have identified the calpain cleavage site of FAK and have generated a mutant form of FAK that is resistant to calpain proteolysis. Expression of wild-type but not calpainresistant FAK restores the adhesion dynamics of talin in FAKdeficient cells. Taken together, our findings suggest a novel role for calpain-mediated cleavage of FAK in regulating adhesion dynamics.
Cell Culture and Transfection-Human embryonic kidney (HEK) 293 cells were purchased from the American Tissue Culture Collection (Manassas, VA). Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, nonessential amino acids, and antibiotics. Transfections were performed by standard calcium phosphate precipitation methods. Cells were analyzed 24 -48 h post-transfection. Wild-type and calpain 4 knock-out mouse embryonic fibroblasts have been described previously (20). NIH 3T3 control and calpain 2 small interfering RNA (siRNA) fibroblasts were maintained as described previously (16). v-Src-transformed NIH 3T3 cells were generated by retroviral infection of FPGV-v-Src. Briefly, cells were infected at 32°C for 6 h and allowed to recover in growth medium for 24 h before selection with 1 mg/ml G418 (Invitrogen) for 7 days. HEK 293 cells stably expressing mCherry and mCherry-FAK constructs were generated by retroviral infection and sorted by fluorescence-activated cell sorting (University of Wisconsin-Madison Flow Cytometry Facility).
In Vitro Calpain Cleavage Assay-Baculovirus-purified FAK (5 g) was incubated in cleavage buffer (50 mM Tris-HCl pH 7.4, 137 mM KCl, and 1 mM MgCl 2 ) with the indicated concentrations of purified calpain 2 at 30°C for 30 min in the absence or presence of CaCl 2 . Cleavage reactions were stopped by the addition of 6ϫ SDS sample buffer.
Cleavage Site Mapping-FAK-FLAG was expressed in HEK 293 cells. 24 -48 h later, cells were plated on dishes coated with 10 g/ml fibronectin and incubated at 37°C under 5% CO 2 for 1 h. The medium was exchanged with complete medium containing a vehicle control or 1 M ionomycin and 10 mM CaCl 2 , and cells were incubated for 10 min at 37°C under 5% CO 2 . Cell lysates were incubated for 2 h with anti-FLAG M2-agarose beads, which were washed extensively with lysis buffer (20 mM Tris, pH 8.0, 1% Nonidet P-40, 10% glycerol, 137 mM NaCl, 0.2 mM PMSF, 1 g/ml pepstatin, 2 g/ml aprotinin, 1 g/ml leupeptin, and 1 mM sodium orthovanadate). The beads were boiled in nonreducing sample buffer to elute proteins, which were separated on SDS-polyacrylamide gels. Proteins were transferred to polyvinylidene difluoride membrane (Bio-Rad) and stained with Coomassie Blue, and an individual band was subjected to N-terminal sequencing (performed at the Baylor College of Medicine Protein Chemistry/Proteomics Core).
Immunocytochemistry-Glass coverslips were acid-washed, silanized, and coated with 10 g/ml fibronectin. Cells were plated on coverslips in Dulbecco's modified Eagle's medium supplemented with 1% fetal bovine serum and nonessential amino acids and allowed to adhere for 4 h. Cells were then fixed, permeabilized, and stained as described previously (22).
Immunoblot Analysis-Cells were scraped into lysis buffer on ice and clarified by centrifugation. Protein concentrations were determined using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Rockford, IL) according to the manufacturer's instructions. Equal amounts of total protein were loaded onto 6 -20% gradient SDS-polyacrylamide gels and transferred to nitrocellulose. Western blots were imaged with an Odyssey infrared imaging system (LI-COR Biosciences, Omaha, NE).
Live Fluorescence Microscopy-Fluorescence imaging of live cells was performed using a 60ϫ objective on an Olympus IX-70 inverted microscope housed in a 37°C closed system. Glass-bottom dishes were acid-washed and coated with 10 g/ml fibronectin. Cells were plated in Dulbecco's modified Eagle's medium containing 1% fetal bovine serum, nonessential amino acids, and 25 mM HEPES, pH 7.4, and allowed to adhere for 1 h, after which the medium was replaced with Ham's F-12 supplemented with 1% fetal bovine serum, nonessential amino acids, and 25 mM HEPES, pH 7.4. Fluorescent images were then captured every 2 min for 2 h using MetaVue imaging software (Molecular Devices, Downingtown, PA).
Quantification of Adhesion Dynamics-Quantification of adhesion dynamics was performed as described previously (4,17). Time-lapse sequences from live fluorescence imaging of GFP-talin1 were first subjected to high-pass filtration based on the "water" algorithm (23) to remove diffuse background fluorescence. Fluorescence intensities of individual adhesions from background-subtracted images were measured over time using MetaVue imaging software. Of the adhesions that formed and disassembled within the period of the movie, 10 -15 randomly selected adhesions in six individual cells/condition over 3 separate days were analyzed. Rate constant measurements for assembly (increasing fluorescence intensity) and disassembly (decreasing fluorescence intensity) of individual adhesions were determined from the slopes of trend lines fitted to semilogarithmic plots of fluorescence intensity ratios over time as described previously (4,17). The duration of an adhesion was determined by the time elapsed between the first and last frames in which an adhesion was observed. Tyr-397 are important for regulating focal adhesion turnover (4,5). However, the contribution of calpain proteolysis of FAK to the regulation of adhesion dynamics has not been defined. To begin to address this, we first tested the effects of FAK depletion on the dynamics of the focal adhesion protein talin. We generated FAK-deficient HEK 293 cells by using siRNA to deplete endogenous levels of FAK by ϳ80% without affecting the expression of the proline-rich tyrosine kinase Pyk2 (Fig. 1A) (24). Immunoblotting of endogenous FAK (ϳ115 kDa) with an antibody specific for the C terminus showed an additional band of ϳ35 kDa in size. We visualized the dynamics of GFP-talin1 in control and FAK-deficient cells by timelapse microscopy ( Fig. 1B and supplemental Movies 1 and 2). Live fluorescence imaging demonstrated that talin-containing adhesions in FAK-depleted cells were extended in duration by ϳ2-fold compared with control cells (Fig. 1C). From plots of GFP-talin1 fluorescence intensities over time, we generated rate constants for net adhesion assembly and disassembly rates. Interestingly, both adhesion assembly and disassembly rates were impaired in FAK-deficient cells. Thus, the extended duration of talin-containing adhesions in FAKdeficient cells resulted from the impaired ability of talin both to incorporate into adhesions and to disassemble from adhesions. Our results are consistent with findings in FAK-null bone marrow macrophages, which exhibit impaired adhesion turnover due to decreased assembly and disassembly rates (25). Taken together, these data indicate that FAK regulates talin dynamics in HEK 293 cells. FAK Is Cleaved by Calpain 2 in Vitro and in Vivo-On the basis of previous studies (16,26,27), we hypothesized that the 35-kDa band present upon FAK immunoblotting was due to calpain proteolysis. However, the specific calpain-mediated FAK proteolytic site had not yet been characterized. We first wanted to confirm the ability of calpain 2 to cleave purified FAK using an in vitro cleavage assay ( Fig. 2A). As expected, in the absence of calpain 2, FAK was not cleaved. The addition of calpain 2 without calcium also did not result in cleavage of FAK. In the presence of 1 mM calcium and low concentrations of calpain 2, immunoblot analysis with N and C terminus-specific antibodies showed a single predominant cleavage band at 80 kDa (  fore, these data demonstrate that FAK is a direct substrate of calpain 2 in vitro.

FAK Regulates the Adhesion Dynamics of Talin-Previous
We next tested the capacity of several cell lines to generate the proteolytic fragment of FAK. Control mouse embryonic fibroblasts, v-Src-transformed NIH 3T3 fibroblasts, and HEK 293 cells were able to produce a 35-kDa C-terminal cleavage fragment, which was diminished in calpain 2-deficient cell lines (Fig. 2, B-D). In addition, upon exogenous expression of FAK-FLAG (Fig. 2E) or GFP-FAK (Fig. 2F), calpain 2-deficient HEK 293 cells displayed a reduction in the single proteolytic fragment compared with control cells. Similar to our results from in vitro calpain cleavage of FAK ( Fig. 2A), a predominant single band was detected, although we did not observe FAK degradation, indicating that there are likely mechanisms in vivo that limit calpain activity and FAK degradation (reviewed in Ref. 14). Intriguingly, in calpain-deficient mouse embryonic fibroblasts and v-Src-transformed NIH 3T3 fibroblasts, immuno-blot analysis revealed an increase in an additional band of higher molecular mass, which was similar in size to the FAKrelated non-kinase (FRNK) (Fig. 2, B and C) (28,29). Thus, to specifically analyze the effects of calpain-mediated cleavage of FAK separable from the calpain knockdown-mediated enhancement of the FRNK-like protein, we performed subsequent analyses in HEK 293 cells, which do not express FRNK. Taken together, these data indicate that calpain cleaves FAK both in vitro and in vivo, generating a single cleavage fragment in vivo. These findings also indicate that calpain 2-mediated proteolysis of FAK generates a fragment that is similar in size to FRNK. B, FAK-FLAG was transiently transfected into HEK 293 cells, which were treated with vehicle (Ϫ) or 1 m ionomycin ϩ 1 mM CaCl 2 (ϩ). Cell lysates were incubated with anti-FLAG beads, and protein complexes were transferred to polyvinylidene difluoride membrane. The 35-kDa band was analyzed by N-terminal sequencing. The Coomassie Blue-stained polyvinylidene difluoride (PVDF) membrane shown is from one of three independent experiments. C, a schematic of FAK domain organization (4.1, ezrin, radixin, moesin (FERM); kinase; focal adhesion targeting (FAT)) denotes the site of calpain cleavage after Ser-745, between the two C-terminal proline-rich (P) regions. Mapping the Calpain Cleavage Site of FAK-The site of calpain proteolysis of FAK had not been previously determined likely due to the lack of a consensus calpain cleavage site (30) and to the fact that only a small percentage (5-10%) of FAK is cleaved. To enhance calpain-mediated generation of the FAK cleavage fragment, we treated HEK 293 cells with the calcium ionophore ionomycin (Fig. 3A). The ionomycin-induced proteolysis of FAK was diminished in calpain 2-deficient cells. Likewise, treatment of HEK 293 cells ectopically expressing FAK-FLAG with ionomycin resulted in the enhanced generation of the proteolytic fragment. We were able to isolate the in vivo cleavage fragment, which was further analyzed by N-terminal sequencing (Fig.  3B). The cleavage site of FAK occurs after Ser-745, located between the two C-terminal proline-rich regions (Fig. 3C).
Generation of a Calpain-resistant Mutant of FAK-To specifically address the effects of calpain proteolysis on adhesion dynamics, we engineered a calpain-resistant mutant form of FAK. We performed site-directed mutagenesis of GFP-FAK in several residues proximal to the cleavage site (Fig. 4A). A 27amino acid deletion (⌬724 -750) surrounding the cleavage site abolished the cleavage fragment. Furthermore, analogous to the mutations introduced into talin and spectrin that rendered them calpain-resistant (17,31), a point mutation of amino acid 744 at the P2 position from a valine to a glycine (V744G) significantly impaired calpain-mediated proteolysis of GFP-FAK. An additional point mutation (P748G) and a 6-amino acid deletion (⌬743-748) also impaired calpain proteolysis of GFP-FAK, whereas an R724L or S725F mutation did not affect the ability of FAK to be cleaved (data not shown). Furthermore, expression of GFP-FAK and calpain-resistant mutants did not affect the capacity of endogenous FAK to be cleaved. Quantification of the susceptibility to calpain proteolysis demonstrated that GFP-FAK-V744G was 10-fold resistant compared with wild-type GFP-FAK (Fig. 4B). Moreover, we confirmed that GFP-FAK-V744G retained the ability to be phosphorylated at Tyr-397 (Fig. 4C). These data demonstrate, for the first time, the site of calpain cleavage of FAK and the successful generation of a point mutant of FAK that is resistant to calpain proteolysis.
Calpain-resistant FAK Retains Localization and Biochemical Interactions-To further determine whether calpain-resistant GFP-FAK-V744G is functional, we examined the intracellular distribution of GFP-FAK-V744G. We expressed GFP-FAK and GFP-FAK-V744G in HEK 293 cells, followed by immunostaining with an antibody to the adhesion marker paxillin (Fig.  5A). We found that both GFP-FAK and GFP-FAK-V744G localized with paxillin-containing adhesions. Moreover, to determine whether wild-type and calpain-resistant GFP-FAK retained interactions with normal binding partners, we performed co-immunoprecipitation experiments and found that GFP-FAK and GFP-FAK-V744G were able to interact with p130Cas (Fig. 5B) and paxillin (Fig. 5C). Thus, these data indicate that calpain-resistant GFP-FAK-V744G is otherwise functional.
Expression of Wild-type but Not Calpain-resistant FAK Rescues Impaired Adhesion Turnover of Talin in FAK-deficient Cells-Given our findings that FAK regulates the adhesion dynamics of talin, we wanted to investigate the role of calpain cleavage of FAK in regulating talin dynamics. Accordingly, we generated mCherry fusions of wild-type and calpain-resistant FAK and were able to deplete cells of endogenous FAK without affect- ing the expression of exogenously expressed mCherry-FAK or mCherry-FAK-V744G (Fig. 6A). Using cell lines stably expressing mCherry, mCherry-FAK, or mCherry-FAK-V744G, we transiently depleted cells of endogenous FAK and visualized the dynamics of GFP-talin1 ( Fig. 6B and supplemental . Live fluorescence imaging showed that talin-containing adhesions in FAKdeficient mCherry-expressing cells were extended in duration compared with those in control cells (Fig. 6C). Quantification of adhesion dynamics demonstrated that the extended durations were due to significantly decreased rates of adhesion assembly and disassembly in FAK-deficient cells. Expression of wild-type mCherry-FAK was able to restore adhesion dynamics of talin in FAK-deficient cells. However, expression of calpain-resistant mCherry-FAK-V744G failed to rescue adhesion dynamics, indicating that calpain cleavage of FAK regulates talin dynamics. Taken together, these data indicate that FAK and its cleavage by calpain contribute to the regulation of adhesion turnover.

DISCUSSION
We showed previously that calpain-mediated proteolysis of talin regulates adhesion dynamics (17); however, the contribution of cleavage of other adhesion proteins has not been defined. Here, we have demonstrated that FAK regulates talin dynamics. We have identified the calpain cleavage site of FAK and have generated mutant FAK that is resistant to calpain proteolysis but retains proper localization and biochemical interactions. Expression of wild-type but not calpain-resistant FAK restores the adhesion dynamics of talin in FAK-deficient cells. Collectively, our data highlight a novel function for calpain proteolysis of FAK in regulating adhesion dynamics.
We found that the calpain cleavage site of FAK resides between consensus caspase cleavage sites (32), indicating that this region is particularly susceptible to proteolytic regulation. Caspases have been demonstrated to cleave FAK under conditions of apoptosis (32)(33)(34)(35). Recent evidence has also shown the incorporation of caspase-8 into a complex containing FAK and calpain 2 to promote tumor cell migration and metastasis (36). Alternatively, it has been proposed that calpain FIGURE 6. Expression of wild-type but not calpain-resistant FAK rescues impaired adhesion turnover in FAK-deficient cells. A, HEK 293 cells were transiently cotransfected with control (Ctlsi) or FAK (FAKsi) siRNA and mCherry (mChe), mCherry-FAK, or mCherry-FAK-V744G. Cell lysates were analyzed by immunoblotting and probed for FAK. The immunoblot shown is from one of three independent experiments. C-term, C-terminal. B, HEK 293 cells stably expressing mCherry, mCherry-FAK, or mCherry-FAK-V744G were transiently cotransfected with control or FAK siRNA and GFP-talin1. Cells were plated on fibronectin-coated glass-bottom dishes and analyzed by time-lapse fluorescence microscopy. Time-lapse montages demonstrate representative images of the dynamics of GFP-talin1 over a period of 24 min. Scale bars ϭ 10 m. Arrows indicate a representative adhesion. Representative movies are shown in supplemental Movies 3-6. C, duration was measured as the time elapsed between the appearance and dissolution of an observed adhesion. Rate constants for net adhesion assembly and disassembly were calculated from plots of fluorescence intensities of GFP-talin1 as described under "Experimental Procedures." Data for each condition are the mean Ϯ S.E. from a total of six cells over three independent experiments. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 (by t test) compared with control siRNA. cleavage may play a role in terminating FAK signaling by attenuating its kinase activity (27).
Our findings suggest that calpain proteolysis of FAK regulates the dynamic turnover of talin at adhesions. This is interesting in light of our previous study (17) that demonstrated that calpain proteolysis of talin is necessary for its own disassembly from focal adhesions as well as for affecting the dynamics of other adhesion components. We now provide evidence to suggest that calpain cleavage of FAK can also affect talin dynamics, indicating that proteolysis of multiple substrates rather than one specific substrate at adhesion sites is likely important for regulating the dynamic turnover of adhesions. However, how calpain cleavage of FAK regulates the adhesion dynamics of talin has yet to be defined. Also, at present, our data cannot determine whether the simultaneous or sequential cleavage of talin and/or FAK is required for regulating adhesion dynamics. Interestingly, similar to our findings with talin proteolysis (data not shown), we did not find a significant difference in migration in FAK-deficient cells re-expressing wild-type or calpain-resistant FAK, suggesting that proteolysis of multiple adhesion proteins may be necessary to affect cell migration. Future studies should address whether the inability to cleave both FAK and talin leads to a synergistic impairment of adhesion dynamics.
We cannot exclude the possibility that the effects of calpainmediated proteolysis may depend on cellular context. Several oncogenic programs have been shown to exhibit differential cleavage of adhesion proteins (37). For example, in Src-transformed cells, FAK is extensively cleaved compared with other adhesion complex proteins (38,39). In addition, human tumors overexpressing cyclin E have been associated with increased proteolysis of FAK (40). By contrast, in myoblasts overexpressing calpastatin, the proteolytic patterns of talin and FAK are unchanged, whereas an accumulation of MARCKS (myristoylated alanine-rich C kinase substrate) has been observed, indicating that calpain may preferentially cleave MARCKS to affect myoblast migration (41,42).
It is intriguing that calpain proteolysis of FAK generates an FRNK-like fragment, suggesting that calpain proteolysis can generate a functional FRNK-like protein temporally and spatially within cells. Interestingly, we detected an increase in expression of an FRNK-like protein in calpain-deficient mouse embryonic fibroblasts and v-Src-transformed NIH 3T3 fibroblasts. FRNK expression is under the control of an alternative promoter that resides within an intron of FAK (29). These findings suggest that increased FRNK may compensate for the absence of calpain-mediated generation of the 35-kDa C-terminal fragment in calpain-deficient cells. Previous studies have shown that adhesion disassembly is disrupted but not abrogated in cells lacking calpain (15,16). It is possible that, in cells deficient in calpain activity, the enhanced generation of FRNK functions as a compensatory mechanism to regulate adhesion turnover. Indeed, exogenously expressed FRNK has been demonstrated to function as a dominant-negative inhibitor of FAK (2,43,44). Intriguingly, a parallel mode of regulation has been observed in hippocampal neurons, in which the expression of protein kinase M is directed by an intronic promoter within the protein kinase C gene (45). The calpain-dependent proteolytic fragment of protein kinase C is similar in size to the protein generated from the alternative gene transcript and may be important in memory formation (46). Thus, it is very intriguing to speculate that calpain may play a role in mediating cross-talk between transcriptional and post-translational mechanisms of signaling.
In summary, our findings demonstrate a novel role for calpain-mediated limited proteolysis of FAK in affecting talin dynamics and support the notion that proteolysis of adhesion proteins is an important mechanism for regulating the turnover of adhesions. It is interesting to speculate that FAK may modulate proteolysis-mediated effects on adhesion dynamics both by functioning as an adaptor protein through its capacity to interact with calpain 2 (47) and by its direct proteolysis by calpain 2. With our successful generation of a calpain-resistant FAK mutant, future studies should investigate the interplay between FAK and calpain 2 binding and cleavage in modulating adhesion dynamics and cell migration. A challenge will be to determine how the proteolysis of different substrates temporally and spatially within cells affects adhesion turnover and cell motility.