Preferential Phosphorylation of Focal Adhesion Kinase Tyrosine 861 Is Critical for Mediating an Anti-apoptotic Response to Hyperosmotic Stress^*

A successful response to an increase in environmental osmolarity is of fundamental importance to cell survival as evidenced by the remarkable conservation of osmotic stress-response pathways from yeast to humans. For example, mammalian p38 mitogen-activated protein kinase and its yeast homologue, Hog1, are similarly activated when mammalian cells and yeast are exposed to hyperosmotic stress (1, 2). Although the conservation of this signaling response to hyperosmotic stress is remarkable, there are compelling reasons to suggest that metazoan cells have evolved additional survival strategies. Metazoan cells, unlike yeast, which have the protection of a rigid cell wall and a conspicuous absence of latent cytoplasmic pro-apoptotic enzymes, will undergo much larger volume contraction and will be at much higher risk of triggering apoptosis under hyperosmotic conditions. In particular, epithelial cells lining mucosal surfaces experience marked hyperosmolar environments under normal and pathological conditions (3, 4). In mammalian cells, hyperosmotic stress induces rapid cortical actin remodeling (5–7). However, the upstream signaling molecules responsible for sensing hyperosmolarity and mounting an anti-apoptotic response to hyperosmotic stress, at least in part by promoting F-actin organization to resist cell shrinkage, remain incompletely characterized.

FAK³ is a nonreceptor tyrosine kinase first identified as one of the proteins phosphorylated on tyrosine in v-Src-transformed chicken embryo fibroblasts (8). FAK signaling has been implicated in the process of focal adhesion turnover required for cell translocation (9–16), but increasingly, it is being shown to be a critical pro-survival signaling node (17–33). Activated FAK blocks mammalian cell apoptosis triggered by several stimuli, including UV irradiation (21) and loss of attachment to extracellular matrix proteins (17, 19). In fibroblasts, FAK also plays a role in preventing apoptosis triggered by hyperosmotic stress (34). However, it is not known whether exposure to hyperosmotic stress preferentially stimulates the phosphorylation of distinct tyrosine residues of FAK. Nor is it known if specific FAK tyrosine phosphorylations play a role in protecting cells from hyperosmotic stress-triggered apoptosis.

Here we determined whether hyperosmotic stress stimulates phosphorylation of specific FAK tyrosine residues in epithelial cells derived from rat intestinal crypts (IEC-18), human colonic carcinoma (Caco2), and human gastric carcinoma (NCI-N87). We found that exposure to hyperosmotic stress induced a striking and selective increase in the phosphorylation of FAK Tyr-861 in all cell types examined. Surprisingly, disruption of F-actin in these cells failed to prevent hyperosmolar-stimulated FAK phosphorylation at Tyr-397, Tyr-576, and most dramatically at Tyr-861. Finally, wild type FAK (FAK WT), but not FAK in which Tyr-861 was mutated to Phe (FAK861F), re-expressed in FAK null cells, promoted cell survival both in the face of hyperosmotic stress and of growth factor depletion. Collectively, these results indicate that FAK activation and specifically phosphorylation of Tyr-861 are sensors of hyperosmotic stress, lying upstream of the actin cytoskeleton. We propose that FAK signaling from phosphorylated Tyr-861 plays a critical role in the response of epithelial cells to withstand hyperosmotic stress.

EXPERIMENTAL PROCEDURES

Cell Culture—IEC-18 cells were maintained in DMEM with 5% FBS between 50 and 80% confluency. For the experiments, 100% confluent monolayer cultures were serumstarved overnight prior to stimulation. Caco2 cells (BBE1 strain) were maintained between 40 and 80% confluency in DMEM supplemented with human transferrin and 10% FBS. NCI-N87 cells were maintained at 40–70% confluency in DMEM/F12 supplemented with 10% FBS and human apotransferrin. Cells were grown to 90% confluency and were serum-starved for 1 h. FAK null (FAK^-/-) (43) and FAK re-expressing (GFP-FAK WT and GFP-FAK 861F) fibroblasts were maintained in DMEM with 4 mm l-glutamine adjusted to contain 1.5 g/liter sodium bicarbonate and 4.5 g/liter glucose with 10% fetal bovine serum in a humidified atmosphere containing 10% CO₂ and 90% air at 37 °C. 1 × 10⁴ cells/35-mm dish were plated, and experiments were performed at 5 days (90% confluency). SYF and YF cells were grown in DMEM with 10% FBS. 1 × 10⁴ cells/35-mm dish were plated, and experiments were performed approximately on day 4 or 5 (90% confluency). All cell lines were obtained from the ATCC.

cDNA Plasmid and Retroviral Transfection Methods—To generate the MSCV-FAK-GFP construct, we first engineered an intermediate vector named pFAK-EGFP. This construct contained the complete coding sequence of mouse FAK cloned upstream from the GFP coding sequence. Briefly, the mouse FAK coding sequence was obtained by PCR using pT7-7-FAK (ATCC number 63207) as template. The forward primer, with the sequence 5′-GGAAGATCTCCGCCACCATGGCAGCTGCTTATCTTGACCCAAAC-3′, contained a BglII restriction site followed by an artificial Kozak sequence and the first 27 nucleotides of the FAK sequence. The reverse primer, with the sequence 5′-CGCGGATCCGCTCCGTGTGGCCGTGTCTGCCCTAGC-3′, contained a BamHI site followed by a mutated stop codon and a sequence complementary to the last 22 coding nucleotides of the FAK sequence. The PCR product was restricted with BglII-BamHI and cloned into the BamHI site of pEGFP-N1 vector (Clontech). Consequently, a DNA fragment bearing the FAK-GFP coding sequence was excised from pFAK-EGFP with KpnI-AflI, Klenow-treated to fill-in the ends, and cloned into the MSCV-IRES-GFP vector whose IRES-GFP sequence was eliminated by restriction with XhoI and bluntended by Klenow treatment.

FAK[Y861F]GFP mutant constructs were generated by site-directed mutagenesis using PCR (QuikChange II XL kit, Stratagene). Using MSCV-FAK-GFP as template, Tyr-861 was converted to phenylalanine by using the mutagenic primers 5′-ACCAACACATCTTTCAGCCTGTGGG (forward) and 5′-TGGTTGTGTAGAAAGTCGGACACCC (reverse). All constructs were verified by DNA sequence analysis, and the products of expression were analyzed by Western blot using murine monoclonal antibodies against the GFP epitope (Clontech) and a polyclonal antibody against FAK. For retrovirus production, logarithmically growing Phoenix ecotropic cells were transfected with MSCV-IRES-GFP (as control), MSCV-FAK-GFP, or mutant MSCV-FAK-861F-GFP, using Lipofectamine Plus following the manufacturer's protocol. Virus-containing supernatants were collected 48 h after transfection and used immediately.

Re-expression of FAK WT and FAK 861F into FAK Null Cells— Logarithmically growing FAK^-/- fibroblasts were incubated with the virus-containing supernatants in the presence of 5 μg/ml Polybrene for 5 h. Cells were collected 48–72 h later, and GFP-positive fractions were FACS-sorted using a FACStar PLUS machine (BD Biosciences). GFP-positive cells were propagated, and multiple aliquots were frozen. A fresh batch of cells was restarted every 2 months. Following sorting, GFP-positive FAK^-/- cells were maintained as described above.

Cell Stimulation with Peptide Agonists—Confluent IEC-18 cells were washed twice with serum-free DMEM and equilibrated for 1 h at 37 °C followed by the addition of angiotensin II in DMEM. The stimulation was terminated by aspirating the medium and lysing the cells in 1 ml of ice-cold RIPA buffer containing 50 mm HEPES, pH 7.4, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 150 mm NaCl, 10% glycerol, 1.5% mm MgCl₂, 1 mm EGTA, 1 mm sodium orthovanadate, 10 mm sodium pyrophosphate, 100 mm NaF, and 1 mm phenylmethylsulfonyl fluoride.

Cell Stimulation by Hypertonicity—Fully confluent IEC-18, Caco-2, or NCI-N87 cells were washed twice with DMEM (IEC-18) or DMEM/F12 (Caco-2 and NCI-N87) and equilibrated in the same media at 37 °C for at least 30 min, and then the media were exchanged for 0.45 m sucrose, 0.6 m sorbitol, or 0.45 m urea in DMEM at 37 °C for the times indicated in figure legends. The stimulation was terminated by rapid rinsing in ice-cold PBS followed by aspirating the medium and solubilizing the cells in 1 ml of ice-cold RIPA buffer as above or in 4× SDS-PAGE sample buffer.

Cell Survival Assay—FAK null fibroblasts from the ATCC were transfected with either WT FAK-GFP or FAK 861F-GFP using a lentiviral transfection system described previously (35). 48 h after infection, comparable transfection efficiencies (25–30%) were confirmed by direct fluorescent microscopy. Transfected cells were further cultured at 37 °C for 10 days without refeeding. Cells were fixed and stained with Diff-Quik cell staining kit from Dade Bhering Inc., Newark, NJ. Remaining cells were counted using a ×10 lens on a Nikon TMS microscope.

Apoptosis Assay—FAK null fibroblasts from the ATCC were transfected with either WT FAK-GFP or FAK 861F-GFP using a lentiviral transfection system described previously (35). After FACS sorting for GFP-positive cells, cells transfected with FAK WT or FAK 861F were grown to 90% confluence in DMEM containing 10% FBS. Cells were incubated for 1 h in DMEM containing 450 mm sorbitol. They were then washed five times with DMEM and incubated a further 18 h. Cells were then assayed for apoptotic nuclei (multifragmented by intensely staining) by staining with Hoechst 33342 (Molecular Probes, Eugene OR) for 20 min following the manufacturer's instructions. Imaging was done on a Zeiss Axioskope microscope using the 4,6-diamidino-2-phenylindole filter.

Western Blotting—After SDS-PAGE, proteins were transferred to Immobilon-P (polyvinylidene difluoride) membranes. After transfer, membranes were blocked using 5% nonfat dried milk in PBST, pH 7.2, and incubated overnight at 4 °C with the anti-FAK-Tyr(P)-397 antibody (0.1 μg/ml), anti-FAK-Tyr(P)-576 antibody (0.1 μg/ml), anti-FAK-Tyr(P)-861 antibody (0.1 μg/ml), or anti-FAK C20 antibodies. They were then incubated in horseradish peroxidase-conjugated secondary antibodies (donkey anti-rabbit or sheep anti-mouse) at 1:5000 for 1 h at room temperature. After washing four times with PBST, the immunoreactive bands were visualized using ECL detection reagents and photographic film.

Fluorescence Microscopy—Cells were fixed in 4% paraformaldehyde for 30 min at room temperature. After washing five times in PBS, they were permeabilized with PBS containing 0.2% Triton X-100 for 2 min at room temperature. After washing five times in PBS, cell were blocked in PBS containing 1% bovine serum albumin (fraction V, Sigma) and 2% fetal bovine serum for 2 h at room temperature. Cells were then stained with AlexaFluor-546-phalloidin (Invitrogen) 1:40 in PBS for 30 min at room temperature. After washing six times in PBS, cells were imaged with an epifluorescence microscope (Zeiss Axioskop) and a Zeiss water-immersion objective (Achroplan 40/0.75 w, Carl Zeiss, Inc., Jena, Germany). AlexaFluor-594 signals were observed with HI Q filter sets for rhodamine isothiocyanate (Chroma Technology). For each cell type, the same exposure settings were used for all images. Images were captured as uncompressed 12-bit TIFF files with a SPOT-cooled (-12 °C) single CCD color digital camera (three-pass method) driven by SPOT version 2.1 software (Diagnostic Instruments, Inc., Sterling Heights, MI). Images were processed using Adobe Photoshop CS2.

Materials—Angiotensin II, bradykinin, sucrose (ultra-pure), urea, and sorbitol (ultra-pure) were all obtained from Sigma. Horseradish peroxidase-conjugated donkey anti-rabbit antibody (NA 934) and sheep anti-mouse antibody and ECL reagents were from GE Healthcare. Anti-FAK antibody C20was from Santa Cruz Biotechnology. The polyclonal phospho-specific antibodies to FAK Tyr-397, FAK Tyr-576, and FAK Tyr-861 or total FAK (C20) were obtained from BIOSOURCE. Cytochalasin D and latrunculin A were from Molecular Probes (Eugene, OR). All other reagents used were of the highest grade available.

RESULTS

Hyperosmotic Stress Stimulates Selective FAK Phosphorylation at Tyr-861—To determine whether hyperosmotic stress triggers preferential phosphorylation of FAK tyrosines, we examined basal and hyperosmotic stress-stimulated FAK phosphorylation at Tyr-397, Tyr-576, and Tyr-861 in rat intestinal epithelial IEC-18 cells (36). Confluent cultures of these cells were washed in serum-free DMEM and then incubated at 37 °C for 2 h. The medium was then replaced with DMEM or with DMEM containing 0.45 m sucrose for either 10 or 15 min. Other cultures were left in DMEM alone or in DMEM to which 50 nm angiotensin II was added for 5 min. Cell lysates were analyzed by Western blotting using antibodies that detect the phosphorylated state of FAK at Tyr-397, Tyr-576, and Tyr-861 or with an antibody against the FAK C terminus. The results shown in Fig. 1A (1st lane (-)) show a high level of basal FAK phosphorylation at Tyr-397 whereas that on FAK Tyr-861 is low. Replacement of the medium (Fig. 1A, 4th lane (contr.)) did not affect basal FAK phosphotyrosine levels. Exposure to 0.45 m sucrose for either 10 or 15 min (Fig. 1A, 2nd and 3rd lanes) failed to stimulate any detectable increase over basal high FAK Tyr(P)-397 and caused a modest increase in FAK Tyr(P)-576 in IEC18 cells. In contrast, hyperosmotic stress induced a dramatic increase in the phosphorylation of FAK at Tyr-861 in these cells. For comparison, we show that the G-protein-coupled receptor (GPCR) agonist, angiotensin II at 50 nm (Fig. 1A, 5th lane (AngII)) only stimulated a modest increase in FAK Tyr-861 phosphorylation over base line in confluent IEC-18 cells.

We then tested the stimulated FAK tyrosine phosphorylation at residues 397, 576, and 861 in response to increasing osmolarity. Confluent IEC-18 monolayers were incubated at 37 °C for 15 min in DMEM alone or DMEM containing increasing concentrations of sorbitol. As shown in Fig. 1B, cell exposure to as little as 100 mm sorbitol for 15 min was sufficient to increase FAK phosphorylation at Tyr-861.

To confirm that the band detected in Western blots of whole cell lysates from hyperosmotic sucrose-treated IEC-18 cells was FAK-phosphorylated at Tyr-861, we first immunoprecipitated, with an anti C-terminal FAK antibody (C20), cell lysates from cultures of these cells exposed to DMEM containing 0.45 m sorbitol for 15 min, and then the immunoprecipitates were Western-blotted with the anti-FAK Tyr(P)-861 antibody or C20 antibody. Fig. 1C substantiates that FAK immunoprecipitated from IEC-18 cells stimulated by hyperosmotic stress shows a dramatic increase in FAK Tyr-861 phosphorylation (lane 2, top panel).

Quantification of phosphorylated FAK species normalized to total FAK from whole cell lysates of IEC-18 cells, under control or after 15 min of hyperosmotic 0.45 m sorbitol (see Fig. 1D), demonstrates that hyperosmotic stress stimulated an ∼5-fold increase in FAK Tyr(P)-861 but did not produce any significant change in FAK phosphorylation at either Tyr-397 or FAK Tyr(P)-576. These data strongly suggest that FAK is phosphorylated preferentially on Tyr-861 under hyperosmotic stress.

Hyperosmotic Stress Stimulates FAK Tyr-861 Phosphorylation in Human Colonic (Caco2) and Gastric (NCI-N87) Epithelial Monolayers—Next, we determined whether the striking stimulation of FAK Tyr-861 phosphorylation in response to hyperosmotic stress in ileal crypt-derived IEC-18 cells also occurs in other epithelial monolayers, including the highly differentiated human colorectal cancer cell line Caco2 (BBE1 strain) or the well differentiated gastric cancer line NCI-N87. Confluent Caco2 cells were treated with DMEM alone or DMEM containing 0.45 m sucrose for 5 or 15 min. For comparison, parallel cultures were treated with DMEM containing 5 ng/ml heparin-bound epidermal growth factor or 50 nm lysophosphatidic acid (LPA) each for 5 min. As shown in Fig. 2A, basal phosphorylation (1st lane) of FAK Tyr-861 (3rd blot) is negligible in Caco2 monolayers. Hyperosmotic stress induced a marked increase in FAK Tyr-861 phosphorylation by 15 min but did not significantly affect phosphorylation of FAK at either Tyr-397 or Tyr-576. Monolayers of NCI-N87 cells were washed twice with DMEM/F12 and then equilibrated at 37 °C for 2 h. They were then incubated a further 30 min in DMEM/F12 alone or DMEM/F12 containing 0.45 m sorbitol. Results shown in Fig. 2B again show that hyperosmotic stress dramatically stimulated FAK Tyr-861 in NCI-N87 cells. Collectively, these results indicate that hyperosmotic stress preferentially stimulates FAK phosphorylation at Tyr-861 in a variety of epithelial cell models.

Hyperosmotic Stress Induces FAK Tyr-861 Phosphorylation via Cell Shrinkage—Studies of tyrosine kinase activation by hyperosmotic stress have demonstrated a requirement for an osmotic gradient between the intracellular and extracellular compartments and concomitant cell shrinkage as a result of water efflux from the cells (37–39). We examined whether cell shrinkage was also necessary for FAK Tyr(P)-861 phosphorylation induced by hyperosmotic stress. We compared FAK tyrosine phosphorylation induced by hypertonic urea, which is membrane-permeable, with that induced by hypertonic sucrose or sorbitol, both of which are membrane-nonpermeable. Monolayers of IEC-18 cells were incubated in DMEM or DMEM containing 100 or 450 mm sucrose, DMEM containing 450 mm urea, or DMEM containing 450 mm sorbitol for 15 min. Fig. 3 again shows that FAK phosphorylation at Tyr-397 and Tyr-576 was only modestly increased by the membrane-nonpermeable solutes sucrose and sorbitol at 0.45 m, whereas 0.45 m membrane-permeable urea did not elicit any increase in the phosphorylation of these sites. Sucrose and sorbitol, both at 0.45 m, strongly stimulated FAK Tyr-861 phosphorylation, whereas 0.45 m urea diminished FAK phosphorylation at this site. These results indicate that hyperosmotic stress induces FAK phosphorylation at Tyr-861 via cell shrinkage.

The Hyperosmotic Stress-induced Phosphorylation of FAK Tyr-861 Is Inhibited by the Selective Src Family Inhibitors PP2 and SU6656—In view of the close functional relationship between Src family kinases and FAK, we determined whether hyperosmotic stress stimulates c-Src kinase activation in IEC-18 cells. Monolayers of these cells were exposed to DMEM alone or DMEM containing various concentrations of sorbitol, sucrose, or urea for 15 min. The cell lysates were then analyzed by Western blot using a phospho-specific antibody to c-Src Tyr-418, which reflects Src kinase activation (40). Fig. 4A, upper panel, shows that 450 mm sucrose or sorbitol, but not urea, induced strong Src Tyr-418 phosphorylation. These results suggest that Src, like FAK, responds to cell shrinkage because of increased extracellular osmolarity.

We next tested the role of Src family kinase activity in hyperosmotic stress-induced FAK Tyr-861 phosphorylation. Confluent monolayers of IEC-18 cells were pretreated for 30 min in the presence or absence of PP2, a selective Src family inhibitor, or SU6656, another structurally distinct Src family kinase inhibitor. The cells were then exposed to DMEM or DMEM containing 0.45 m sucrose for 15 min in the absence or presence of the corresponding Src kinase inhibitor. As seen in the Western blot in Fig. 4B, 5th and 6th lanes, pretreatment of IEC-18 cells with either of the two distinct selective Src family kinase inhibitors markedly attenuated FAK Tyr-861 phosphorylation induced by hyperosmotic sucrose.

Hyperosmotic Stress Induces FAK Tyr-861 Phosphorylation in YF but Not SYF Fibroblasts—To substantiate that Src family kinase activity is required for hyperosmotic stress-induced FAK Tyr-861 phosphorylation, we determined hyperosmotic stress-induced FAK tyrosine phosphorylation in the Src/Yes/Fyn^-/- (SYF) and Yes/Fyn^-/- (YF) fibroblast cell lines (41). SYF cells were washed and incubated in DMEM, DMEM containing LPA for 5 min, or 0.45 m sucrose for increasing times. As seen in Fig. 4C (left panel), 3rd blot (FAK pTyr-861), FAK Tyr-861 phosphorylation is undetectable even after 30 min of sucrose in SYF cells. Parallel hyperosmotic stimulation and analysis of FAK phosphorylation sites in YF cells (Yes/Fyn null cells with endogenous c-Src levels) demonstrate that the presence of c-Src restores FAK Tyr-861 phosphorylation in response to hyperosmotic stress. As seen in Fig. 4C (right panel), 3rd blot, YF cells exposed to 0.45 m for 15 min show markedly increased FAK Tyr-861 phosphorylation. The results shown in Fig. 4 indicate that Src family kinases play a role in mediating hyperosmotic stress-induced phosphorylation of FAK Tyr-861 and imply that hyperosmotic stress induces preferential stimulation of a Src/FAK Tyr-861 pathway.

FAK Tyrosine Phosphorylation Induced by Hyperosmotic Stress Exhibits Unique Independence from Actin Cytoskeleton Organization— Virtually all stimuli tested so far that lead to FAK tyrosine phosphorylation, including adhesion-dependent signals, receptor tyrosine kinase activation by growth factors, and GPCR agonists, require intact cellular F-actin substructure. By utilizing the F-actin-disrupting agents, cytochalasin D and latrunculin A, we determined whether FAK Tyr-861 phosphorylation in response to hyperosmotic environment requires an intact F-actin cytoskeleton. These agents disrupt F-actin by distinct mechanisms, by capping rapidly growing barbed ends in the case of cytochalasin D or by tightly binding monomeric G-actin and thereby removing it from the free actin pool in the case of latrunculin A (42). Confluent monolayers of IEC-18 cells were pretreated with serum-free DMEM for 1 h in the presence or absence of either 2 μm latrunculin A or 2 μm cytochalasin D. The cultures were then exposed to 0.45 m sucrose for 15 min in the presence or absence of the actin polymerization inhibitors. Western blot results using phospho-specific FAK antibodies against FAK Tyr-397, Tyr-576, and Tyr-861, as shown in Fig. 5A, were consistent with our preceding results (1st lane) in that basal phosphorylation of FAK Tyr-397 was high, and it was not significantly increased by angiotensin II treatment. Again, basal FAK Tyr-861 phosphorylation was very low, and angiotensin II stimulation elicited a very modest increase, and hyperosmotic stress stimulated a striking increase in these levels. Latrunculin A and cytochalasin D significantly lowered the modest increase of FAK Tyr-861 induced by angiotensin II (Fig. 5A, compare 3rd lane with 6th and 9th lanes). However, in surprising contrast to all previously published FAK phosphotyrosine-promoting stimuli, which are significantly inhibited by F-actin disruption, hyperosmotic stress-induced phosphorylation of FAK at Tyr-861 was not affected by latrunculin A (Fig. 5A, compare 2nd lane with 5th and 8th lanes). Interestingly, whereas latrunculin A abolished basal FAK tyrosine phosphorylation at FAK Tyr-397 and FAK Tyr-576, it failed to block hyperosmotic stress-induced phosphorylation at these sites. Similarly, cytochalasin D did not block hyperosmotic stress-induced FAK tyrosine phosphorylations.

We verified that these agents, at the concentrations used, caused F-actin disruption in IEC-18 cells (Fig. 5C, top three micrographs). Latrunculin A effected a marked F-actin disruption that abolished all stress fibers and cortical F-actin arrangements (Fig. 5C, top middle panel). Cytochalasin D pretreatment disrupted all stress fibers, and cortical F-actin structures persisted at cell-cell junctions (see Fig. 5C, top left panel). This is entirely consistent with the known mechanism of action of cytochalasin D, which disrupts rapidly turning over F-actin in stress fibers, but it is not effective at disrupting more slowly remodeling cortical actin structures. As seen in Fig. 5C (middle and right panels), the characteristic F-actin disruption by latrunculin A and cytochalasin D was completely unaffected by the further treatment of these cells with hyperosmotic sorbitol.

To confirm that this unexpected F-actin insensitivity of FAK activation shown in Fig. 5A in IEC-18 cells was not cell type-specific, we also tested the effect of F-actin disruption on FAK phosphorylation by hyperosmotic stress in NCI-N87 cells. After pretreatment for 1 h with 2 μm latrunculin A or 2 h with cytochalasin D, we exposed confluent NCI-N87 cells to DMEM/F12 alone or DMEM/F12 containing 0.45 m sorbitol for 30 min (with continued presence or absence of latrunculin A or cytochalasin D). Western blot results shown in Fig. 5B show that confluent NCI-N87 epithelial cells exhibit a relatively high basal phosphorylation of FAK Tyr-397 and Tyr-576 but low Tyr-861 (left lane). Exposure to hypertonic sorbitol for 30 min caused an increase in the phosphorylation of FAK at Tyr-861. Remarkably, pretreatment with either latrunculin A for 1 h or cytochalasin D for 2 h enhanced the stimulation of FAK Tyr-861 phosphorylation produced by hyperosmotic stress (Fig. 5B, compare lanes 2, 4 and 6). This is consistent with results in Fig. 5A from IEC-18 cells where FAK Tyr-861 phosphorylation stimulated by hyperosmotic stress is completely insensitive to F-actin disruption.

To control for the possibility that hyperosmotic stress exposure interferes with the action of these F-actin-disrupting agents, we examined the F-actin in control, latrunculin A-, and cytochalasin D-pretreated NCI-N87 cells after a further 30-min exposure to 0.45 m sorbitol. As seen comparing the control and sorbitol-treated micrographs of the right two panels of Fig. 5D, the further hyperosmotic exposure does not affect the F-actin alterations elicited by either of these agents.

FAK Tyr-861 Is Required for FAK-promoted Cell Survival of Hyperosmotic Stress—We have previously published that FAK null fibroblasts are markedly more susceptible to apoptosis when exposed to hyperosmotic stress; likewise, SYF cells when compared with YF cells undergo markedly increased rates of apoptosis in response to hyperosmotic stress (34). This background, and the striking differential phosphorylation of FAK in response to hyperosmotic stress demonstrated in this study, prompted us to determine whether the selective phosphorylation of FAK at Tyr-861 plays a role in opposing apoptosis elicited by hyperosmotic stress. FAK null cells stably expressing either GFP-tagged FAK WT or GFP-tagged FAKY861F were grown to 90% confluence and exposed to 0.45 m sorbitol for 1 h. The cells were then washed and further cultured for 18 h at which time they were stained with Hoechst 33342 for 30 min. The striking finding is that the number of cells remaining in the FAKY861F-expressing cells was markedly reduced compared with the number remaining in the FAK WT-expressing cells (15 versus 70%). Additionally, the majority (65%) of the remaining FAKY861F-expressing cells showed characteristically intense Hoechst 33342-stained, pyknotic, and fragmented apoptotic nuclei (Fig. 6, A, right panel, and B, right bar). In sharp contrast, the vast majority of the remaining FAK WT-expressing cells had nonfragmented, diffusely staining nuclei (Fig. 6, A, left panel, and B, left bar). This result implies that FAK requires Tyr-861 phosphorylation to oppose hyperosmotic stress-triggered apoptosis.

To test whether the FAK Tyr-861 requirement for blocking hyperosmotic stress-triggered apoptosis could be extended to another apoptosis-triggering stimulus, we examined cell survival after growth factor and nutrient depletion in FAK null cells re-expressing either GFP-FAK WT or GFP-FAK 861F. The GFP-FAK constructs were introduced into 70% confluent FAK null cells via lentivirally mediated gene transfer, and after 48 h of incubation the cultures were examined for levels of construct expression. Both GFP-FAK WT and GFP-FAK 861F inoculated cultures at 48 h showed approximately the same amount of GFP-FAK-expressing cells (25–30%) as assessed by visualizing GFP fluorescence (not shown). After levels of expression were established, cells were allowed to grow without refeeding for 10 days. Microscopic inspection of the cultures revealed that cells expressing GFP-FAK WT or GFP-FAK861F achieved confluence and even overgrew, with many cells floating in the supernatant. However, only in the cells in which GFP-tagged WT FAK was introduced did significant cell numbers survive as growth factors were depleted. As quantified in Fig. 6C, and seen directly in Fig. 6D, substantial numbers of cells remain on dishes after 10 days growth without refeeding in GFP-FAK WT (Fig. 6D, left 10 dishes), In contrast, virtually all of the cells expressing GFP-tagged FAK861F (Fig. 6D, right 10 dishes) died after 10 days without refeeding. These experiments show that WT FAK promotes mammalian cell survival of hyperosmotic stress-triggered apoptosis as well as of growth factor depletion-induced apoptosis, whereas FAK 861F does not. These results indicate that the phosphorylation of Tyr-861 is critical for anti-apoptotic signaling from FAK.

DISCUSSION

Hyperosmotic stress is a ubiquitously experienced environmental condition that cells have been required to withstand throughout evolution. In this context, epithelial cell sheets provide a barrier to luminal hyperosmotic environments in both the renal and gastrointestinal tracts. Hyperosmolarity in the form of a high sodium diet has been implicated in stomach cancer progression both in human populations and in animal models (43–46), but few investigations into the basic mechanisms underlying this link have been published.

Furthermore, in contrast to multiple extensive studies in yeast, pro-survival signaling pathways in metazoan cells responding to hyperosmotic stress have not been well characterized. Although mammalian cells, in response to hyperosmotic stress, activate evolutionarily conserved pro-survival signaling cascades (common to yeast), it is not known if they activate additional pro-survival signaling pathways. We hypothesize the existence of additional survival strategies may have evolved in cells both when the protection of a rigid cell wall was lost and also with the evolution of pro-apoptotic enzyme cascades. Without a rigid cell wall, cells will undergo much larger volume losses for the same level of hyperosmotic stress. The presence of latent cytoplasmic pro-apoptotic enzyme cascades places cells at increased risk of apoptosis on significant cell shrinkage, for example by increased local concentration of apoptosis-initiating caspase recruitment domain-containing proteins (47).

We have previously reported that the presence of FAK and c-Src can protect cells from hyperosmotic stress-triggered apoptosis (34), and here we extend these studies and now report three novel findings. First, hyperosmotic stress stimulates a dramatic and preferential increase in phosphorylation of FAK at Tyr-861 in multiple epithelial cell lines (including gastric epithelial cells) as well as in FAK null cells expressing wild type FAK and in SYF cells expressing c-Src. Second, hyperosmotic stress-stimulated FAK phosphorylation is insensitive to F-actin disruption induced by either cytochalasin D or latrunculin A, in contrast to results from numerous studies examining FAK tyrosine phosphorylation stimulated by multiple stimuli, including GPCR agonists, receptor tyrosine kinase ligands, and integrin clustering. Finally, we demonstrate that FAK Tyr-861 phosphorylation is required for FAK to block apoptosis in mammalian cells undergoing hyperosmotic stress or growth factor depletion.

Although several reports implicate FAK in pro-survival signaling pathways elicited in the face of pro-apoptotic triggers, the molecular details of this anti-apoptotic function of FAK remain obscure. In fibroblasts, the presence of endogenous levels of WT FAK, and separately the presence of c-Src reduced apoptosis triggered by hyperosmotic stress (34). This study strongly suggests that the FAK-c-Src bipartite enzyme system is critical for a metazoan cell anti-apoptotic response to hyperosmotic stress. However, it was unclear whether specific FAK phosphorylation sites or specific domains were required. The results presented here show a robust and preferential increase in FAK phosphorylation at Tyr-861 in response to hyperosmotic stress in several epithelial cell types, including IEC-18 (modest trans-epithelial resistance), Caco2-BBE strain (medium trans-epithelial resistance), and NCI-N87 (high trans-epithelial resistance), and also in several fibroblast cell lines. The response is rapid and sustained, with increases in FAK Tyr-861 phosphorylation detected on only 100 mm increases in extracellular osmolarity. We demonstrate using two structurally distinct Src family kinase inhibitors, PP2 and SU6656, as well as SYF (Src/Yes/Fyn triple null) cells, that hyperosmotic stress-induced increase in FAK Tyr-861 phosphorylation is dependent on Src family kinase activity. We also show that the extracellular to intracellular osmolar gradient, leading to cell shrinkage, is critical to increase Src and FAK Tyr-861 signaling.

To date, studies examining signaling pathways that lead to increased FAK tyrosine phosphorylation have shown a universal requirement for an intact actin cytoskeleton arrangement, as they have invariably been sensitive to the actin-depolymerizing macrolide cytochalasin D (48–51). This requirement for actin cytoskeletal organization is thought to reflect a requirement for active FAK clustering at focal adhesions, which are disrupted in the presence of cytochalasin D. Surprisingly, we show here for the first time that hyperosmotic stress activation of FAK tyrosine phosphorylation, most potently at Tyr-861, is insensitive to F-actin disruption. The resistance of hyperosmotic stress-induced FAK tyrosine phosphorylation to actin depolymerization is unique and implies that FAK may be a component of the osmosensing signaling pathway upstream rather than downstream of F-actin reorganization in response to this stress.

The preceding findings prompted us to hypothesize that a marked stimulation of FAK Tyr-861 phosphorylation in response to hyperosmotic stress plays a critical role in FAK-mediated anti-apoptotic signaling. Although much has been published on the role of FAK Tyr-397 (in terms of c-Src interactions), FAK Tyr-576/577 (in terms of activation loop phosphorylation), and FAK Tyr-925 (in terms of Grb2/Erk1/2 interactions), much less is known about FAK Tyr-861 regulation and its function. Vascular endothelial growth factor stimulation of endothelial cells leads to increased FAK Tyr-861 phosphorylation by c-Src (25). Both in vascular endothelial growth factor-stimulated endothelial cells and in epidermal growth factor-stimulated epithelial cells the c-Src phosphorylation of FAK Tyr-861 is required for the formation of a growth factor-stimulated FAK-αvβ5 signaling complex (52). When FAK Tyr-861 phosphorylation in human prostate cancer cell lines with high metastatic potential and spontaneous migratory capacity (PC3 and DU145) are compared with those with low metastatic potential and no spontaneous migratory capacity (LNCaP), constitutive, anchorage-independent FAK Tyr-861 phosphorylation is present only in cells of high metastatic potential (53). Thus FAK Tyr-861 phosphorylation links signaling from polypeptide growth factors to that from specific cell surface integrins and may contribute to anchorage-independent cell survival. In H-Ras (G12R)-transformed fibroblasts, FAK Tyr-861 phosphorylation is critical to pro-survival signals that are essential for the phenotype of invasiveness, focus formation, and anchorage-independent growth (54, 55). These studies suggest the possibility that FAK Tyr-861 phosphorylation promotes cell survival during tumor metastasis.

To test whether FAK Tyr-861 is critical for FAK anti-apoptosis signaling, we compared hyperosmotic stress-triggered apoptosis levels in FAK null cells re-expressing wild type FAK with those re-expressing FAK 861F. We found that cells re-expressing FAK 861F were highly susceptible to hyperosmotic stress-triggered apoptosis as compared with cells re-expressing wild type FAK. Collectively, our results indicate that an increase in the phosphorylation of FAK Tyr-861 is required for mammalian cells to survive hyperosmotic stress. The subsequent down-stream events leading to enhanced survival in the face of hyperosmotic stress remain incompletely characterized, although most interestingly, it has been shown that FAK Tyr-861 phosphorylation may recruit the Src homology 3-containing scaffolding molecule, p130^CAS (56). Both FAK and Src have been shown to phosphorylate p130^CAS on multiple tyrosines (57, 58). Recruited phosphorylated p130^CAS could stimulate the assembly of F-actin-remodeling complexes (59–62) that promote cell survival, in part, by limiting cell shrinkage.

In conclusion, our results are consistent with a model that envisages FAK/c-Src bipartite enzyme as an upstream osmosensor that on cell shrinkage, selectively leads to FAK Tyr-861 by c-Src without the requirement for intact actin cytoskeleton. In turn, FAK Tyr-861 phosphorylation initiates anti-apoptotic signaling pathways, including those leading to local F-actin remodeling. This subsequent F-actin reorganization may limit hyperosmotic stress-induced cell shrinkage and thereby promote cell survival.