Calcium Oscillations Trigger Focal Adhesion Disassembly in Human U87 Astrocytoma Cells*

Integrin-associated intracellular Ca 2 oscillations modulate cell migration, probably by controlling inte-grin-mediated release of the cell rear during migration. Focal adhesion kinase (FAK), via its tyrosine phosphorylation activity, plays a key role in integrin signaling. In human U87 astrocytoma cells, expression of the dominant negative FAK-related non-kinase domain (FRNK) inhibits the Ca 2 (cid:1) -sensitive component of serum-depend-ent migration. We investigated how integrin-associated Ca 2 (cid:1) signaling might be coupled to focal adhesion (FA) dynamics by visualizing the effects of Ca 2 (cid:1) spikes on FAs using green fluorescent protein (GFP)-tagged FAK and FRNK. We report that Ca 2 (cid:1) spikes are temporally correlated with movement and disassembly of FAs, but not their formation. FRNK transfection did not affect generation of Ca 2 (cid:1) spikes, although cell morphology was altered, with fewer FAs of larger size and having a more peripheral localization being observed. Larger sized FAs in FRNK-transfected cells were not disassembled by Ca 2 (cid:1) spikes, providing a possible explanation for impaired Ca 2 (cid:1) -dependent migration in these cells. Stress fiber end movements initiated by Ca 2 (cid:1) spikes were visualized using GFP-tagged myosin light chain kinase (MLCK).

Cell migration is a cyclic process involving initial protrusion of the leading edge, formation of adhesive sites, contraction of the cell body, and release of adhesive sites at the cell rear (1). Adhesive sites are dynamic membrane structures that vary in size and composition during migration. Integrins, actin stress fibers (SFs) 1 and other structural proteins, and regulatory signaling molecules cluster at focal adhesions (2). Focal adhesions (FAs) serve as points of traction for contractile forces underlying forward cell movement and their dynamics are finely regulated. For example, FAs are highly motile in stationary fibroblasts but are largely stationary in migrating fibroblasts, thereby transducing contractile forces into movement (3). This suggests the existence of a molecular clutch that couples cytoskeleton-mediated traction and cell contraction.
Focal adhesion kinase (FAK) is activated and localized at FAs upon cell adhesion to the extracellular matrix (ECM; Refs. 4 and 5). Given the abundance of FAs and the reduced migration of fibroblasts from FAK null mice (6), FAK is likely involved in FA remodeling during migration. FAK-related non-Kinase (FRNK), the non-catalytic C-terminal portion of FAK containing the FA targeting sequence, is also expressed as a separate dominant negative protein (7). The differential expression of FAK and FRNK is transcriptionally regulated, each of these proteins having distinct promoters within the FAK gene (8). Although the function of endogenous FRNK is not clear, FRNK has been used to alter signaling via endogenous FAK. When overexpressed in cells, FRNK acts as a negative regulator of FAK activity, inhibiting phosphorylation of FAK and different FAK-related processes, including cell cycle progression (9,10), cell spreading on fibronectin (7,11), and migration (12,13). This suggests that the inhibitory effects of FRNK in migration might arise from altered FAK localization and phosphorylation.
We and others reported that migration is dependent on Ca 2ϩ signaling in astrocytoma (14), smooth muscle cells (15), neutrophils (16), and neurons (17). In cerebellar granule cells, Ca 2ϩ -dependent migration is correlated with the amplitude and frequency of Ca 2ϩ spikes (17), which may regulate different steps during migration. Disruption of integrin-mediated adhesion involves Ca 2ϩ -sensitive proteins, including calpain (18,19), myosin light chain kinase (MLCK; Ref. 20), and calcineurin (21). These data indicate that Ca 2ϩ signaling may be a component of the molecular clutch regulating transitions between stationary and non-stationary FAs.
To test whether Ca 2ϩ signaling affects FA organization, confocal microscopy was used to visualize simultaneously Ca 2ϩ levels and dynamics of FAs or actin SFs in U87 cells expressing FA (FAK, FNRK) or cytoskeletal (MLCK) proteins tagged with GFP. We report that Ca 2ϩ spikes trigger movement and disassembly of FAs. Although FRNK expression did not suppress Ca 2ϩ -dependent FA disassembly, more Ca 2ϩ -insensitive FAs having a larger surface were observed in FRNK cells. Our results provide an explanation linking FA disassembly to a temporally accurate cellular signal.

EXPERIMENTAL PROCEDURES
Reagents and Cells-Cell culture media (EMEM), fetal calf serum, HEPES, L-glutamine, penicillin, streptomycin, gentamycin, and trypsin-EDTA were from Invitrogen; Fura Red-AM, BAPTA-AM, and pluronic acid were from Molecular Probes; Matrigel and the monoclonal antibody (mAb) against FAK kinase domain were from Interchim; phalloidin-TRITC, mAb against the FAK C-terminal region (amino acids 1039 -1052), and anti-MLCK mAb were from Sigma; the anti-Tyr 397phosphorylated FAK Ab was from BIOSOURCE; secondary horseradish peroxidase-conjugated Abs were from Promega; the FITC-labeled goat anti-mouse (GAM-FITC) Ab was from Zymed Laboratories Inc.. The human astrocytoma U87 cell line was obtained from the ATCC. Cells were maintained at 37°C in a humidified incubator gassed with 5% CO 2 in air on type I collagen (0.06 mg/ml)-coated plastic dishes in EMEM supplemented with 10% heat-inactivated fetal calf serum, 0.6 mg/ml glutamine, 200 IU/ml penicillin, 200 IU/ml streptomycin, and 0.1 mg/ml gentamycin.
Plasmids and Transfection-Fluorescent FA-targeted protein was made by fusion of FAK (human T lymphocyte pCDM8-FAK plasmid; Ref. 40) next to the 3Ј end of "yellow Cameleon-2" (pcDNA3-Ycam2 plasmid; Ref. 22). To allow fusion of FAK in continuity with the Ycam2 reading frame, the stop codon next to EYFP was replaced by a tyrosine codon (QuikChange, Stratagene). FAK cDNA was amplified by PCR using a 5Ј primer containing a MfeI site and a 3Ј primer containing a NheI site. The FAK PCR product was digested with MfeI and NheI and cloned in the corresponding compatible sites, EcoRI and XbaI, located in the multiple cloning site of the newly mutated pcDNA3-YCam2 vector, adjacent to EYFP, to give FAK-Ycam. To create FRNK-Ycam, the FRNK domain was amplified by PCR using pCDM8-FAK as template, a forward primer with an EcoRI site at the 5Ј end of the glutamic acid codon 681 relative to the FAK start codon, and the same reverse primer as for FAK amplification, adding a NheI site. The PCR product was digested with EcoRI and NheI and cloned in-frame with the EYFP coding sequence in the EcoRI/XbaI compatible sites of the mutated pcDNA3-Ycam2 vector. The MLCK-210-GFP construct (24) was in pEGFP vector (CLONTECH). All constructs were verified by sequencing. The plasmids were isolated (JetStar, Genomed) before transfection by electroporation. Cells (5 ϫ 10 6 ) were resuspended at 10 8 cell/ml in EP buffer (in mM: 50 K 2 HPO 4 , 20 CH 3 CO 2 K, 20 KOH, pH 7.4). Plasmidic DNA (2 g of construct-encoding plasmid, 8 g of pBluescript) was diluted in 100 l of EP buffer; 4 l of 1 M MgSO 4 were added and incubated with 50 l of cell suspension for 20 min at room temperature. The cell/DNA mixture was electroporated in a 0.4-cm cuvette (Bio-Rad Gene Pulser; 500 microfarads, 240 V). Cells were then placed in 10 ml of EMEM with 10% fetal calf serum in 80-mm 2 dishes. Cells were selected 24 h later using 800 g/ml G418 (Sigma) and maintained with 400 g/ml G418. Cells were sorted to obtain Ͼ80% expressing cells using a flow cytometer (FACStar, Becton-Dickinson) before use.
Migration Assay-A wound-healing migration model was used, as described previously (15). FAK-Ycam or FRNK-Ycam cells (2 ϫ 10 5 cells/ml) were grown to confluence in Matrigel-coated (178 g/ml) Petri dishes. After 24-h serum starvation, a rectangular lesion was made, cells were rinsed and incubated with medium with or without the tested compound. After 24 h of migration, three fields at the lesion border were acquired using a CCD camera (Panasonic) on an inverted microscope (Olympus IMT2, 10ϫ phase objective). In each field, migration distance for the 10 most mobile cells was measured using Image Tool software (available by FTP from maxrad6.uthscsa.edu). For BAPTA experiments, cells were first loaded (45 min) with 20 M BAPTA-AM in the incubator prior to creation of lesions.
Measurement of Intracellular Calcium and Dynamics of Focal Adhesions or Stress Fibers-Intracellular Ca 2ϩ was measured using Fura Red by confocal microscopy (Bio-Rad 1024, krypton-argon laser 488 nm; Nikon Eclipse TE300, 40ϫ oil-immersion CFI Plan-Fluor n.a. 1.3 objec-tive) alone or simultaneously with FA or SF dynamics. Cells (5 ϫ 10 4 cells/ml) expressing FAK-Ycam, FRNK-Ycam, or MLCK-GFP were grown to subconfluence on Matrigel in Petri dishes in which a 2-cm diameter hole had been cut in the base and replaced by a thin (0.07 mm) coverslip (41). After 48 or 72 h, cells in culture medium were loaded with 10 M Fura Red-AM and 0.03% pluronic acid (45 min) in the incubator and then washed two times with Ringer solution (in mM: 140 NaCl, 5 KCl, 2 CaCl 2 , 2 MgCl 2 , 10 HEPES, 11 glucose, pH 7.4). Imaging of single cells in Ringer solution was done at 30°C, at 585 Ϯ 10 nm for Ca 2ϩ alone, at 1-s intervals, usually for 15 min. For simultaneous measurements of Ca 2ϩ (at 585 nm) and FA (Ycam constructs) or SF (MLCK-GFP) dynamics, images at 522 Ϯ 16 nm were acquired simultaneously every 10 s (EYFP) and 15 s (EGFP), usually for 30 min. Images were taken at the bottom cell surface. NIH Image software was used to assess FA surface areas and the dynamics of FAs and SFs. For FA size distributions, FAK-Ycam and FRNK-Ycam cells having similar expression levels were used. Fluorescent FAs were selected by thresholding using NIH Image. The same threshold was applied to all cells, corresponding FA pixel areas were calculated and converted to m 2 . This analysis was done to determine the size of both immobile and motile, Ca 2ϩ -sensitive FAs. FA movement and disassembly triggered by a Ca 2ϩ spike were visualized using a color overlay representation of three sequential images (3).
Immunostaining-FAK-Ycam, FRNK-Ycam, or MLCK-GFP cells were rinsed with PBS and fixed 15 min with 3% paraformaldehye at room temperature. After three washes with PBS, cells were treated for 10 min with 0.2% Triton X-100 in PBS, 0.2% BSA and incubated 30 min with PBS, 3% BSA. Cells were washed three times with PBS, 0.2% BSA and incubated 1 h with either 1 g/ml phalloidin-TRITC, the anti-MLCK mAb K36 (1/10000), or with the anti-FAK mAb (1/1000) in PBS, 0.2% BSA. For actin staining, the incubation buffer was removed and replaced by PBS, 0.2% BSA. For MLCK or FAK staining, cells were washed three times and incubated 1 h with a FITC-labeled goat antimouse Ab (GAM-FITC, 1/100). Fluorescence was observed using a confocal microscope and a 40ϫ objective, as above. FITC and TRITC were excited at 488 and 568 nm, respectively, and fluorescence was collected at 522 and 585 nm.

Human U87 Astrocytoma Cells Expressing FRNK-Ycam
Have Impaired Calcium-dependent Migration-We reported that migration of U87 cells is associated with Ca 2ϩ oscillations (14), as intracellular Ca 2ϩ buffering by BAPTA partly inhibits serum-dependent migration. To analyze the role of FAK in Ca 2ϩ -dependent migration, U87 cells were transfected with FAK or the dominant negative FRNK, fused to the fluorescent yellow Cameleon-2 (Ycam) tag, a fluorescence resonance energy transfer-based Ca 2ϩ sensor containing CFP and EYFP (22). Expression of FAK-Ycam or FRNK-Ycam did not alter endogenous FAK levels compared with controls (Fig. 1A, left). FRNK-Ycam expression (Fig. 1A, right) was deduced by sub-tracting the endogenous FAK band from the band representing endogenous FAK plus FRNK-Ycam. Expression levels of FAK-Ycam and FRNK-Ycam were 1.5-2 times that of endogenous FAK. Compared with controls, serum-independent migration was unchanged in FAK-and FRNK-transfected cells, while serum-dependent migration decreased by 25% only in FRNKtransfected cells (Fig. 1B, left), in agreement with previous studies (12). Buffering of intracellular Ca 2ϩ with BAPTA did not affect serum-independent migration in control and transfected cells, but inhibited serum-dependent migration by 33% in FAK-Ycam cells. The lack of effect of BAPTA on serum-dependent migration of FRNK-transfected cells together with the clear inhibition by BAPTA in FAK-transfected and control cells indicate that the migration component inhibited by FRNK expression is mainly Ca 2ϩ -dependent.
U87 Astrocytoma Cells Expressing FRNK-Ycam Have Unaltered Calcium Signaling-We found that Ca 2ϩ spikes in U87 cells occur only in the presence of serum and are blocked by inhibitory antibodies against ␤ 1 and ␤ 3 integrin subunits, with the generation of such Ca 2ϩ spikes depending on activation of phospholipase C␥. 2 The FAK tyrosine 397 autophosphorylation site interacts with and activates phospholipase C␥, representing a possible link between integrins and Ca 2ϩ signaling (23). Since FRNK contains the focal adhesion targeting sequence that is necessary and sufficient for FAK recruitment to FAs but lacks the phospholipase C␥-interacting autophosphorylation site, FRNK-Ycam expression might inhibit Ca 2ϩ oscillations. However, the frequency of Ca 2ϩ spikes over 15 min in cells expressing FAK-Ycam or FRNK-Ycam was similar (Fig. 1B,  right) and not different compared with controls (not shown). For most FRNK-Ycam cells, the distribution of endogenous FAK at FAs, evaluated using an Ab against the FAK kinase domain, was unaltered (Fig. 1C, left), as found previously (7). This suggests that sufficient endogenous FAK is expressed in most FRNK-transfected cells (Fig. 1C), perhaps accounting for the unaltered generation of Ca 2ϩ spikes (23). Thus, the Ca 2ϩdependent migration defect of FRNK-Ycam cells is not due to decreased Ca 2ϩ signaling.
Single Calcium Spikes Trigger FA Movement and Disassembly-Since differences in Ca 2ϩ signaling do not underlie impaired migration of FRNK-Ycam cells, possible effects of Ca 2ϩ spikes on FAs were investigated by simultaneously measuring Ca 2ϩ variations (using Fura Red) and FA dynamics (using EYFP fluorescence of FAK-Ycam and FRNK-Ycam) in migrating cells. Color overlays of three sequential images ( Fig. 2A) distinguish immobile FAs (black) versus motile FAs (rainbow). Most FAs were immobile in FAK-Ycam cells irrespective of Ca 2ϩ oscillations. Motile FAs were present in oscillatory and non-oscillatory cells, being often localized at one edge of a migrating cell (Fig. 2A). FA movements were linear and usually resulted in FA disassembly, with EYFP fluorescence decreasing or disappearing (Fig. 2B), while immobile FAs never disassembled. In many cases, FA movement and subsequent disassembly was triggered by a Ca 2ϩ spike (Fig. 2B). In oscillatory FAK-Ycam cells, such FA disassembly was temporally correlated with Ca 2ϩ spikes for 64% (37/58) of motile FAs. For 19% of motile FAs (11/58), the Ca 2ϩ oscillation frequency was too high to determine clearly a correlation, and for the remaining 17% (10/58), no correlation was found. This indicates that Ca 2ϩ spikes were responsible for the dynamics of a subset of FAs. Somewhat surprisingly, FA disassembly in oscillatory FRNK-Ycam cells (Fig. 2, C and D) was also temporally correlated with a Ca 2ϩ spike for 50% (35/70) of motile FAs; the Ca 2ϩ oscillation frequency was too high to establish a correlation for 30% of FAs (21/70), and there was no correlation for the remaining 20% (14/70).
Preferential Calcium-dependent Disassembly of Small FAs in U87 Cells Expressing FRNK-Ycam-Strikingly, the number, size, and localization of FAs were different in FNRK-Ycam compared with FAK-Ycam cells (Fig. 4, A and B). FRNK-Ycam cells had 3-fold fewer FAs (25 Ϯ 3/cell, n ϭ 4 cells), which were larger (Fig. 4, C and D) and more peripherally located com- pared with FAK-Ycam cells (74 Ϯ 10 FAs/cell, n ϭ 4). Morphometric analysis (ratio of longest to shortest cell lengths) showed that FRNK-Ycam cells were more elongated (3.0 Ϯ 0.2, n ϭ 51) than FAK-Ycam cells (2.3 Ϯ 0.1, n ϭ 69), suggesting that they are more strongly held to the ECM. Comparison of the Ca 2ϩ dependence of FA dynamics revealed that the size distribution of motile Ca 2ϩ -sensitive FAs (Fig. 4E) mirrors that of all FAs in FAK-Ycam cells (Fig. 4C). In contrast, in FRNK-Ycam cells, the size distribution of motile, Ca 2ϩ -sensitive FAs (Fig. 4F) is not the same as for all FAs (Fig. 4D), with smaller FAs being preferentially disassembled by Ca 2ϩ spikes. This may explain impaired Ca 2ϩ -dependent migration in FRNK-Ycam cells, as larger, immobile Ca 2ϩ -insensitive FAs would remain attached to the ECM despite oscillatory Ca 2ϩ signaling.
Calcium Elevation Rapidly Increases FAK Tyrosine Phosphorylation-Since FAK regulates FA dynamics (5,6), and Ca 2ϩ spikes trigger FA disassembly, this may be related to FAK activity. Therefore, FAK tyrosine phosphorylation was evaluated after thapsigargin-induced Ca 2ϩ elevation (14). Increases in Tyr 397 phosphorylation of endogenous FAK and FAK-Ycam were detected within 30 s after 1 M thapsigargin treatment (Fig. 5A). After 1 min, phosphorylation was maximal (73% increase above control; Fig. 5B). Thus, rapid phosphorylation of FAK in response to Ca 2ϩ increases provides a possible link between FA movements and Ca 2ϩ spikes, supporting that these are regulatory events in migration.
Calcium-induced FA Disassembly Correlates with Retraction of Stress Fibers-As FA disassembly involves disruption of ECM-integrin and/or SF-integrin interactions, we simultaneously followed the dynamics of SF ends and Ca 2ϩ signals using MLCK-GFP-transfected cells (Fig. 6). No differences were found in the localization, distribution, and expression of MLCK in MLCK-GFP cells compared with controls (Fig. 6, A  and B), as reported previously (24). Moreover, the MLCK-GFP construct used here has the same activity in phosphorylating the 20-kDa regulatory light chain kinase when compared with endogenous MLCK (25). Evaluation of SF end dynamics (using MLCK-GFP fluorescence; Fig. 6C) and Ca 2ϩ variations (using Fura Red) revealed that SF end movements were triggered by a Ca 2ϩ spike (Fig. 6D).
The average latency between a Ca 2ϩ spike and SF end movement was 28 Ϯ 6 s, with a rate of movement of 0.25 Ϯ 0.04 m/min (n ϭ 21 SFs, 6 cells; Fig. 7). These SF kinetic parameters were compared with those for motile Ca 2ϩ -sensitive FAs in FAK-Ycam and FRNK-Ycam cells (Fig. 7C). The latency between Ca 2ϩ spikes and the onset of FA movement was 33 Ϯ 5 s in FAK-Ycam cells (n ϭ 21 FAs, 8 cells) and 31 Ϯ 5 s in FRNK-Ycam cells (n ϭ 35 FAs; 14 cells), with a FA speed of 0.18 Ϯ 0.02 and 0.22 Ϯ 0.02 m/min, respectively. These kinetic parameters were not different, suggesting that Ca 2ϩ spikes trigger coordinated movement of SF-associated FA complexes. This implies that during cell migration, Ca 2ϩ signaling results in disruption of ECM-integrin interactions rather than a decrease in FA-cytoskeletal interactions. DISCUSSION We investigated the implication of intracellular Ca 2ϩ elevations in FA dynamics during migration of U87 astrocytoma cells. During oscillatory Ca 2ϩ signaling, single Ca 2ϩ spikes triggered FA disassembly and subsequent cell edge retraction. In FRNK-Ycam cells, smaller sized FAs were more sensitive to Ca 2ϩ -triggered disruption compared with large FAs, consistent with FA stability being a limiting factor in motility. FAK phosphorylation was rapidly induced by a Ca 2ϩ increase, indicating that FAK is a Ca 2ϩ target during migration. Analysis of FA and SF kinetic parameters suggests that Ca 2ϩ signaling coordinates disruption of ECM-integrin interactions at FAs.
Cell migration includes phases of protrusion, adhesion, and retraction (1), involving, respectively, formation, strengthening, and disassembly of focal contacts. FAK plays an important role in the dynamics of cell adhesion, but is not required for FA formation beneath lamellipodia of migrating cells (26). FAK aggregates to clustered integrin receptors with or without ligand occupancy (27), consistent with FAK localization to FAs being an early or late event in the cascade of interactions, respectively, leading to formation or disruption of integrincytoskeletal linkages. This justifies our choice of FAK-Ycam (22) as a probe for FA dynamics, indirectly allowing detection of clustered integrins. FAK-Ycam and FRNK-Ycam were localized to newly forming, punctate structures in protrusive areas, further supporting their use to follow indirectly integrin association/dissociation in FAs. As suggested previously (28), these punctate structures very probably are focal complexes, putative precursors of FAs.
FAK tyrosine kinase activity is involved in the regulation FA turnover (29). In FAK-deficient cells, reduced motility is accompanied by an increased number of FAs (6). FRNK expression inhibits integrin-stimulated migration and phospho- rylation of endogenous FAK and other FA components such as paxillin and tensin (7,11,13). Thus, FAK phosphorylation likely governs FA dynamics and, hence, motility. In FAK-Ycam cells, FA morphology was similar to controls, and as found in previous studies using FAK-transfected U87 cells (12), migration was unchanged, unlike in Chinese hamster ovary cells where migration increased after FAK overexpression (30). Expression of FRNK-Ycam induced elongated morphology and a sparse and peripheral distribution of enlarged FAs. Much evidence supports that defective FAK signaling leads to enlarged FAs. Cells expressing a kinase-deficient mutant of Src have larger FAs and reduced migration (31). The interaction and activation of Src occurs at the FAK autophosphorylation site, which is absent for FRNK (11). Src kinase activity weakens integrin-cytoskeletal linkages (32) and may stimulate FA turnover by favoring lateral diffusion of integrins away from FAs. Thus, reduced FA dynamics leading to enlarged FAs is consistent with decreased migration of FRNK-Ycam cells. The disassembly of large FAs in FRNK-Ycam cells was Ca 2ϩ -insensitive, possibly because for this subset of FAs, endogenous FAK levels were small compared with FRNK levels. Our data support that FA size affects migration by also reflecting the Ca 2ϩ sensitivity of FAs and, hence, their remodeling.
FA disassembly involves disruption of ECM-integrin and/or integrin-cytoskeleton interactions, which are regulated by several calciproteins. For instance, contractile force activation via Ca 2ϩ /calmodulin-dependent MLCK may strengthen integrin-cytoskeletal linkages (26) and is necessary for Ca 2ϩ -dependent migration of neutrophils (20). Conversely, integrin-cytoskeletal linkages are disrupted by calpain, a Ca 2ϩ -dependent protease, and calpain inhibition or disruption mimics the effects of FRNK expression on FA morphology (18,19). Calreticulin appears to be essential for integrin-mediated Ca 2ϩ signaling and adhesion (33,34). Calcineurin, a protein phosphatase, is involved in integrin recycling to the front of migrating neutrophils (21), probably via affinity modulation. We show that Ca 2ϩ spikes trigger FA disassembly and propose that ECM-integrin linkages are disrupted, given the identical kinetic parameters for FA and SF dynamics, which suggests coordinated FA/SF movement. In agreement, in migrating fibroblasts, FAs move with a similar speed of 0.12 Ϯ 0.08 m/min and remain associated with SFs in retractile edges (3). Our observation of linearly disassembling FAs suggests that FAK-Ycam and FRNK-Ycam remained linked to SF ends until complete disassembly. However, because integrin-cytoskeleton and integrin-ECM interactions are intimately related, linkage perturbations initiated at either side of these molecular complexes could be transmitted via modulation of integrin affinity/avidity to both intracellular and extracellular partners. For example, integrin ligation initiates recruitment of a cytoskeletal actin component (27) and linkage with the forward moving actin cytoskeleton (35). Conversely, detachment of integrins from the actin cytoskeleton may induce decreased integrin affinity/avidity for their ECM ligands (36,37).
The cell migration process implies asymmetric signal transduction that directs cell polarization (1). In U87 astrocytoma cells, we observed spatially localized disassembly and formation of FAs, illustrating this asymmetry. Since FA disassembly but not formation was temporally correlated with Ca 2ϩ spikes, Ca 2ϩ signaling and/or regulatory Ca 2ϩ target proteins may be spatially restricted to discrete subcellular compartments (38,39). We are currently investigating compartmentalized Ca 2ϩ signaling in migrating cells using our Ycam constructs as local detectors of Ca 2ϩ near FAs.
That a single Ca 2ϩ spike is sufficient to trigger FA disassembly agrees with initiation of a regenerative process leading to irreversible FA disruption. The Ca 2ϩ -triggered event might transiently increase FAK signaling, perhaps followed by a Ca 2ϩ -independent process. This implies a latency between the Ca 2ϩ spike and cell edge retraction resulting from FA disassembly. Consistent with our data, Ca 2ϩ elevation and increased migration speed were positively correlated in neutrophils (16), but since there was a 20-s delay between Ca 2ϩ elevation and increased motility, Ca 2ϩ was proposed not to be the immediately causal signal. However, we show that FA disassembly begins 30 s after a Ca 2ϩ spike, and once triggered, several minutes are required for complete disassembly. This supports that Ca 2ϩ elevation is a proximal signal leading to increased motility. Finally, rapid thapsigargin-induced FAK Tyr 397 phosphorylation agrees with Ca 2ϩdependent FAK activation being an initial event associated with FA disassembly.