Cholecystokinin Activates PYK2/CAKβ by a Phospholipase C-dependent Mechanism and Its Association with the Mitogen-activated Protein Kinase Signaling Pathway in Pancreatic Acinar Cells*

PYK2/CAKβ is a recently described cytoplasmic tyrosine kinase related to p125 focal adhesion kinase (p125FAK) that can be activated by a number of stimuli including growth factors, lipids, and some G protein-coupled receptors. Studies suggest PYK2/CAKβ may be important for coupling various G protein-coupled receptors to the mitogen-activated protein kinase (MAPK) cascade. The hormone neurotransmitter cholecystokinin (CCK) is known to activate both phospholipase C-dependent cascades and MAPK signaling pathways; however, the relationship between these remain unclear. In rat pancreatic acini, CCK-8 (10 nm) rapidly stimulated tyrosine phosphorylation and activation of PYK2/CAKβ by both activation of high affinity and low affinity CCKA receptor states. Blockage of CCK-stimulated increases in protein kinase C activity or CCK-stimulated increases in [Ca2+] i , inhibited by 40–50% PYK2/CAKβ but not p125FAK tyrosine phosphorylation. Simultaneous blockage of both phospholipase C cascades inhibited PYK2/CAKβ tyrosine phosphorylation completely and p125FAK tyrosine phosphorylation by 50%. CCK-8 stimulated a rapid increase in PYK2/CAKβ kinase activity assessed by both an in vitrokinase assay and autophosphorylation. Total PYK2/CAKβ under basal conditions was largely localized (77 ± 7%) in the membrane fraction, whereas total p125FAK was largely localized (86 ± 3%) in the cytosolic fraction. With CCK stimulation, both p125FAK and PYK2/CAKβ translocated to the plasma membrane. Moreover CCK stimulation causes a rapid formation of both PYK2/CAKβ-Grb2 and PYK2/CAKβ-Crk complexes. These results demonstrate that PYK2/CAKβ and p125FAK are regulated differently by CCKA receptor stimulation and that PYK2/CAKβ is probably an important mediator of downstream signals by CCK-8, especially in its ability to activate the MAPK signaling pathway, which possibly mediates CCK growth effects in normal and neoplastic tissues.

Cholecystokinin (CCK) 1 functions as a neurotransmitter in the central nervous system and gastrointestinal tract as well as a hormone in the gastrointestinal tract and has numerous biologic effects (1). In the central nervous system, CCK has such widespread effects as functioning as a potent regulator of satiety and morphine-induced analgesia, modulating dopamine release, stimulating panic attacks, and stimulating vagal afferent transmission (1)(2)(3)(4)(5). In the gastrointestinal tract, CCK is a physiological regulator of pancreatic secretion, gallbladder contraction, gastric emptying, and colonic motility (1, 6 -9). The physiological and cellular basis of action of CCK have been extensively studied in pancreatic acinar cells, where CCK stimulates activation of c-fos, c-jun, and c-myc oncogene expression (10), enzyme secretion, enzyme synthesis (1,6), growth (8), and development (11). Extensive studies show that CCK causes activation of phospholipase A 2 , phospholipase D, and phospholipase C (PLC) (12). The activation of phospholipase C results in generation of diacylglycerol and increased inositol 1,4,5trisphosphate with subsequent activation of protein kinase C (PKC), stimulation of increases in intracellular calcium concentration, and activation of mitogen-activated protein kinase (MAPK) signaling pathways (12)(13)(14)(15)(16). Recent studies have demonstrated that CCK A receptor activation (17)(18)(19), similar to the activation in different tissues by integrins (20,21), bioactive lipids (22,23), various growth factors (19,21,24), and some G protein-coupled receptors (19,(25)(26)(27)(28), causes stimulation of tyrosine phosphorylation of a number of proteins. With other stimulants, activation of this important intracellular pathway has been shown to be particularly important in mediating cellular growth and motility and adapting to cellular stresses (21,25,29).
At present, it is unknown whether CCK A receptor activation stimulates tyrosine phosphorylation of PYK2/CAK␤ or whether its ability to activate one or both limbs of the PLC cascade is important for this tyrosine phosphorylation if it occurs. It is also unknown whether PYK2/CAK␤ activation might be an important intermediate in coupling the CCK A receptor activation to MAPK cascade activation, which is important in mediating the potent effects of this receptor's activation on normal and neoplastic tissues. In the present study, the ability of CCK A receptor activation to stimulate PYK2/CAK␤ tyrosine phosphorylation and its relationship with these other intracellular signaling cascades was explored in pancreatic acini, which are one of the main physiological sites of action of CCK.
Immunoprecipitation-Immunoprecipitation of tyrosine-phosphorylated proteins was performed as described previously (17,56). Dispersed acini from one rat were preincubated with standard incubation solution without or with different inhibitors for 3 hours at 37°C. Aliquots (1 ml) were then incubated at 37°C with different agonists at the concentrations and times indicated. Acinar lysates were obtained as described (17,56). For tyrosine phosphorylation determination, lysates (500 g/ ml) were incubated with 4 g of anti-phosphotyrosine monoclonal antibody (PY20) or 3 g of anti-PYK2/CAK␤ mAb and 3 g of rabbit anti-mouse IgG and 25 l of protein A-agarose overnight at 4°C. For co-immunoprecipitation studies, lysates (500 g) were incubated with 4 g of anti-Crk mAb or 4 g of anti-Grb2 mAb for 2 h at 4°C. Then the immune complexes were incubated with 4 g of rabbit anti-mouse IgG and 25 l of protein A-agarose for 1 h at 4°C. The immunoprecipitates were washed three times with phosphate-buffered saline and further analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting.
Subcellular Fractionation-Acinar cell fractionation was carried out according to the procedure published previously (57) with minor modifications. Briefly, acinar cells were separated by centrifugation, washed with phosphate-buffered saline with 0.2 mM Na 3 VO 4 at 4°C, centrifuged, and resuspended in 1 ml of lysis buffer without Triton and deoxycholate and homogenized with the use of a Polytron homogenizer (Brinkmann Instruments, Westburg, NY) for 20 s at power level 6 at 4°C. Homogenates were centrifuged first at 500 ϫ g for 10 min at 4°C to remove nuclei and debris and then for 30 min at 100,000 ϫ g at 4°C to obtain membrane and cytosol fractions. Precipitates were washed with phosphate-buffered saline with 0.2 mM Na 3 VO 4 at 4°C and resuspended in 0.5 ml of lysis buffer and sonicated for 5 s at 4°C. Lysates were centrifuged at 15,000 ϫ g for 15 min to remove insoluble substances. Protein concentration was measured by the Bio-Rad protein assay reagent.
Western Blotting-Western blotting was performed as described previously (17,56). Immunoprecipitates or subcellular fractions were subjected to 10% SDS-PAGE gels; proteins were transferred to nitrocellulose membranes (0.45 and 0.2 m for proteins higher and lower than 60 kDa, respectively). Membranes were blocked and incubated for 2 h at 25°C with 0.5 g/ml anti-PYK2/CAK␤ mAb, 1 g/ml anti-phosphotyrosine mAb, 0.25 g/ml anti-p125 FAK , 0.5 g/ml anti-Crk mAb, or 0.5 g/ml anti-Grb2 mAb. Membranes were washed twice and incubated for 45 min at 25°C with anti-mouse IgG-horseradish peroxidase conjugate. The membranes were washed, and proteins were visualized using ECL reagents and Hyperfilm ECL. Density of bands was measured using a scanning densitometer (Molecular Dynamics, Inc., Sunnyvale, CA). When necessary, membranes were reprobed as described previously (56).
In Vitro Kinase Assay and Autophosphorylation of PYK2/CAK␤-After incubation with 10 nM CCK-8 for the times indicated above, pancreatic acini were lysed with kinase lysis buffer, which contained 50 mM Tris-HCl (pH 7.4), 1% Triton X-100, 1 mM Na 3 VO 4 , 150 mM NaCl, 5 g/ml leupeptin, 5 g/ml aprotinin, 1 g/ml pepstatin A, and 1 mM phenylmethanesulfonyl fluoride. PYK2/CAK␤ was immunoprecipitated as described above. The immunoprecipitates were washed two times with kinase lysis buffer and two times with kinase buffer (50 mM Tris-HCl, pH 7.4, 5 mM MnCl 2 , 5 mM MgCl 2 ) and resuspended in 20 l of kinase buffer supplemented with 20 M of ATP including 10 Ci of [␥Ϫ 32 P]ATP (3000 Ci/mmol) with (in vitro assay) or without (autophosphorylation assay) 40 g of poly(Glu-Tyr) (4:1). After 10 min at room temperature, the reaction was stopped by the addition of 10 l of 4ϫ SDS sample buffer and boiled for 5 min. Samples were resolved using 4 -20% SDS-PAGE gels. Gels were dried, analyzed by autoradiography, and quantified by densitometry.

RESULTS
When pancreatic acini were incubated with 10 nM CCK-8 for 2.5 min (a concentration and time that causes a maximal stimulation of tyrosine phosphorylation of p125 FAK and paxillin in previous studies (17,18)), immunoprecipitated with anti-phosphotyrosine mAb (PY20), and analyzed by Western blotting with the same antibody ( Fig. 1, left upper panel), an increase in the tyrosine phosphorylation of at least two major components of molecular mass 115-140 kDa and 65-80 kDa was seen. Following stripping of the membrane from anti-phosphotyrosine antibodies, immunoblotting with a specific anti-PYK2 mAb that does not cross-react with p125 FAK revealed that, under basal conditions, very low levels of tyrosine-phosphorylated PYK2/CAK␤ existed; however, 10 nM CCK-8 induced a dramatic increase in the tyrosine phosphorylation of PYK2/ CAK␤ ( Fig. 1, right upper panel). An identical result was obtained when the acinar lysates were first immunoprecipitated with anti-PYK2 mAb and then analyzed with anti-phosphotyrosine mAb (Fig. 1, left bottom panel). Following stripping of this latter membrane from anti-phosphotyrosine antibodies, Western blotting with the specific anti-PYK2 mAb showed that the recovery of PYK2/CAK␤ from cell lysates was not altered by treatment with CCK-8 ( Fig. 1, right bottom panel). These results demonstrate that PYK2/CAK␤ is expressed in rat pancreatic acini and is phosphorylated upon pancreatic acini activation by CCK-8.
PYK2/CAK␤ tyrosine phosphorylation has been shown to increase in parallel with PYK2/CAK␤ kinase activity (40). Therefore, an assessment of in vitro kinase assay was performed in immunoprecipitates of pancreatic acini treated with 10 nM CCK-8 at different times ( In addition to increasing the production of inositol phosphate, which causes mobilization of cellular calcium, activation of phospholipase C by CCK-8 promotes the hydrolysis of phosphatidylinositol 4,5-bisphosphate, leading to production of diacylglycerol, which in turn activates PKC (12). Results from a number of recent studies in other cell systems (30, 32-34, 58, 64) suggest that increases in cytosolic calcium and activation of PKC may be important for the agonist to stimulate PYK2/ CAK␤ tyrosine phosphorylation and kinase activity (40,55). We next attempted to determine whether CCK-8's increases in intracellular calcium or activation of PKC or both were needed for its ability to cause PYK2/CAK␤ tyrosine phosphorylation in pancreatic acini. To determine whether direct activation of PKC increased the tyrosine phosphorylation of PYK2/CAK␤, pancreatic acini were treated with the phorbol ester, TPA. TPA (1 M) stimulated a rapid increase in PYK2/CAK␤ tyrosine phosphorylation, which reached a maximum at 10 min with a 31 Ϯ 6-fold increase and was maintained for at least 20 min (Fig. 5). To determine whether increased cytosolic calcium either alone or in combination with activation of PKC could alter PYK2/CAK␤ tyrosine phosphorylation, we compared the ability of the calcium ionophore A23187 and thapsigargin, an agent that specifically inhibits the endoplasmic reticulum Ca 2ϩ -ATPase, depletes Ca 2ϩ from intracellular compartments, and increases calcium influx (65,66), to cause PYK2/CAK␤ tyrosine phosphorylation in pancreatic acini when present alone or with TPA ( Fig. 6). Calcium ionophore A23187 (1 M) and thapsigargin (10 M) under conditions previously shown to increase [Ca 2ϩ ] i (66, 67) caused a 14 Ϯ 5-fold increase and a 25 Ϯ 9-fold increase in PYK2/CAK␤ tyrosine phosphorylation, which were 28 Ϯ 9 and 56 Ϯ 7% of that caused by a maximally effective 1 nM concentration of CCK-8, respectively ( CCK A receptor activation results in the release of Ca 2ϩ from intracellular stores followed by the influx of Ca 2ϩ from the medium (12,66). Since calcium ionophore and thapsigargin, under conditions that increase calcium influx, induced the tyrosine phosphorylation of PYK2/CAK␤ in pancreatic acini, we examined the role of extracellular calcium in CCK-8-stimulated PYK2/CAK␤ tyrosine phosphorylation. The responses of pancreatic acini to CCK-8 resuspended in either calcium-free media, with EGTA (5 mM), or calcium-containing media were compared (Fig. 7). Calcium influx was decreased when the pancreatic acini were stimulated by CCK-8 in the absence of calcium in the medium with the result that the [Ca 2ϩ ] i increase by CCK-8 returned to base line sooner (Fig. 7, left panels). However, inhibition of the calcium influx had no significant . Pretreatment of pancreatic acini with thapsigargin or with BAPTA/AM in a calcium-free medium decreased CCK-8-stimulated PYK2/CAK␤ tyrosine phosphorylation by 69 Ϯ 3 and 64 Ϯ 3%, respectively (Fig. 8, right top and bottom  panels). However, neither pretreatment with thapsigargin nor pretreatment with BAPTA/AM had an effect on the increase in p125 FAK tyrosine phosphorylation caused by CCK-8 under identical conditions, demonstrating that neither BAPTA/AM nor thapsigargin pretreatment was having a nonspecific inhibitory effect (Fig. 8, right upper panel).
To determine whether PKC activation might be involved in mediating CCK-8-stimulated increases in PYK2/CAK␤ tyrosine phosphorylation, we examined the effect of a PKC inhibitor, GF109203X (68). Previously, we have shown that pretreatment of pancreatic acinar cells with 5 M of GF109203X for 2 h caused complete inhibition of p125 FAK and paxillin tyrosine phosphorylation induced by activation of PKC with TPA (17). Pretreatment of pancreatic acinar cells with GF109203X (5 M) for 2 h attenuated the increase in PYK2/CAK␤ tyrosine phosphorylation in response to CCK-8 (1 nM) by 50 Ϯ 7% (Fig. 9,  upper panel, compare lanes 4 and 5, and bottom left panel) but had no effect on CCK-8-stimulated p125 FAK tyrosine phosphorylation in parallel experiments (Fig. 9, upper panel, compare  lanes 4 and 5) as we have shown previously (17).
Since increases in intracellular calcium by the calcium ionophore A23187 and activation of PKC by TPA have synergistic effects on PYK2/CAK␤ tyrosine phosphorylation (Fig. 6), and inhibition of both limbs of the PLC cascade separately attenuate the increase in PYK2/CAK␤ tyrosine phosphorylation in response to CCK-8 by almost 50%, we examined the effect of a combination of inhibition of CCK-8-induced increases in intracellular calcium and PKC activation (Fig. 9). Pretreatment of pancreatic acini with thapsigargin to inhibit mobilization of intracellular calcium and with GF109203X to block PKC, in a calcium-free medium, completely inhibited CCK-8 stimulation of PYK2/CAK␤ tyrosine phosphorylation (Fig. 9, top panel,  compare lanes 4 and 6, and left bottom panel). In comparison, pretreatment with thapsigargin and GF109203X caused almost a 50% decrease in CCK-8-stimulated p125 FAK tyrosine phosphorylation in parallel experiments (Fig. 9, upper panel,  compare lanes 4 and 6, and right bottom panel) as shown previously (17).
Cell activation by growth factors, oncogenes, and neurotransmitter/hormones can result in a redistribution of tyrosinephosphorylated proteins (56,(72)(73)(74)(75). To examine this possibility for PYK2/CAK␤, we determined the distribution of PYK2/ CAK␤ in membrane and cytosolic fractions under basal conditions and under CCK-8 stimulation. In unstimulated pancreatic acini, total PYK2/CAK␤ and total p125 FAK were found in both the cytosolic and the membrane fractions (Fig. 11,  upper left panel, lanes 1 and 3, and upper right panel, lanes 1  and 3). Densitometric analysis of immunoblots indicated that in unstimulated pancreatic acini total PYK2/CAK␤ was largely localized (77 Ϯ 7% of total) in the membrane fraction (Fig. 11,  top left panel, compare lanes 1 and 3, and bottom panel); however, total p125 FAK was largely localized (86 Ϯ 3% of total) in the cytosolic fraction (Fig. 11, upper right panel, compare lanes  1 and 3, and bottom panel). Upon the addition of 1 nM CCK-8 for 2.5 min, there was a 56 Ϯ 9% decrease in the amount of total PYK2/CAK␤ in cytosolic fractions (Fig. 11, left upper panel,  compare lanes 1 and 2, and bottom panel). However, the increase in the total PYK2/CAK␤ in the corresponding membrane fractions did not reach significance after CCK-8 stimulation (Fig. 11, upper left panel, compare lanes 3 and 4, and bottom  panel). In comparison, the addition of 1 nM CCK-8 for 2.5 min caused an 86 Ϯ 32% increase in the amount of total p125 FAK in membrane fractions (Fig. 11, top right panel, compare lanes 3  and 4, and bottom panel). However, no detectable increase in the amount of total p125 FAK in the corresponding cytosolic fractions was found after CCK-8 stimulation (Fig. 11, upper  right panel, compare lanes 1 and 2, and bottom panel).
Previous studies in rat pancreatic acini (15, 16, 76) demon-strate activation of the MAPK signaling pathway after CCK A receptor activation by agonists. Moreover, in other cell systems recent studies show that PYK2/CAK␤ activation could be responsible for activation of the Ras/MAPK signaling pathway by G protein-coupled receptors (30,32,34). To explore whether CCK A receptor activation could promote activation of the Ras/ MAPK signaling pathway by PYK2/CAK␤, we examined the interaction between the adapter protein Grb2 with PYK2/ CAK␤ in pancreatic acini stimulated by CCK-8. We assessed the formation of a complex between Grb2 with PYK2/CAK␤ by determining the results of co-immunoprecipitation of PYK2/ CAK␤ with Grb2 after immunoprecipitation with Grb2 (Fig. 12,  right panels). With the addition of 1 nM CCK-8, there was a rapid stimulation of the formation of a PYK2/CAK␤-Grb2 complex with a maximal effect seen at 2.5 min (Fig. 12, right panels). This change was not due to a difference in protein loading, because Western blotting with an anti-Grb2 mAb showed equal loading (Fig. 12, right upper panel). In other tissues, it has been shown that an alternative pathway for activation of the MAPK signaling pathway could be the tyrosine phosphorylation of the adapter protein c-Crk (77-79). To assess whether CCK A receptor activation could promote the association of PYK2/CAK␤ with the adapter molecule c-Crk, we assessed the formation of this complex by determining the results of co-immunoprecipitation of PYK2/CAK␤ with c-Crk after immunoprecipitation with c-Crk mAb (Fig. 12, left panel). CCK-8 stimulation of pancreatic acini caused the rapid formation of a PYK2/CAK␤-Crk complex with a maximal effect seen at 2.5 min (Fig. 12, left panel). The observed change in the amount of PYK2/CAK␤ in c-Crk immunoprecipitates was not due to differential recovery of c-Crk from CCK-8-treated cells because when c-Crk mAb was used for Western blotting after immunoprecipitation with c-Crk mAb, similar amounts of c-Crk were seen in all samples (Fig. 12, top left panel). DISCUSSION A number of recent studies demonstrate that in addition to integrins (31,36), bioactive lipids (32,40), cellular stress (ra-diation, ultraviolet light, osmotic changes) (80), and growth factors (37-39), a number of G protein-coupled receptors ( number of results in our study support the conclusions that activation of both the high affinity and the low affinity states of the CCK A receptor in rat pancreatic acini can cause tyrosine phosphorylation of PYK2/CAK␤ and suggest that PYK2/CAK␤ may have a significant role in agonist-activated CCK A receptormediated intracellular signaling pathways. First, whether an anti-phosphotyrosine mAb is used for immunoprecipitation followed by protein identification using an anti-PYK2/CAK␤ mAb or the same mAbs are used in reverse order, CCK-8 caused a marked increased in the tyrosine phosphorylation of a single PYK2/CAK␤ immunoreactive band. Second, PYK2/CAK␤ tyrosine phosphorylation and activation were a rapid consequence of the addition of CCK-8 to pancreatic acini, causing a maximal increase in both by 1-2.5 min. Third, the dose-response curve for PYK2/CAK␤ tyrosine phosphorylation occurred over the concentration range in which CCK-8 stimulates changes in cellular calcium (13,17); generation of inositol phosphates (17,82); enzyme secretion (82); MAPK signaling pathway activation (15); p125 FAK , p130 Cas , and paxillin tyrosine phosphorylation (17,56); and [ 3 H]thymidine incorporation (83). Fourth, stimulation of pancreatic acini with CCK-JMV, a CCK analogue, caused a clear increase in the PYK2/CAK␤ tyrosine phosphorylation. Moreover, CCK-JMV antagonized the ability of CCK to stimulate PYK2/CAK␤ tyrosine phosphorylation to a greater extent than that seen with CCK-JMV alone. Because CCK-JMV functions as an agonist for the CCK A receptor high affinity state and an antagonist at the low affinity CCK A receptor state in rat pancreatic acini (18,19,61,62), these results demonstrate that 20% of the increase in tyrosine phosphorylation of PYK2/CAK␤ by CCK is due to activation of the high affinity state and 80% is due to activation of the low affinity CCK A receptor state. Last, this conclusion is further supported by the observation that CCK-8 stimulation of PYK2/CAK␤ tyrosine phosphorylation occurred over the concentration range for CCK-8 occupation of both CCK A receptor states, the high affinity state and the low affinity state (17,82).
Activation of the pancreatic CCK A receptor is known to stimulate phospholipase C activity, resulting in the generation of inositol phosphates and diacylglycerol, which in turn mobilize cellular calcium and activate protein kinase C (PKC), respectively (12,13). Recent studies demonstrate that the mechanism of changes in cellular calcium in stimulating PYK2/CAK␤ tyrosine phosphorylation in various cells with different stimuli varies. With T cell receptor activation in Jurkat cells (51), UTP activation of P 2Y2 receptors in PC-12 cells (35) or angiotensin II or platelet-derived growth factor in rat vascular smooth muscle cells (39), EGTA partially inhibited PYK2/CAK␤ tyrosine phosphorylation, suggesting that calcium influx was at least partially involved in mediating the tyrosine phosphorylation. In contrast, stimulation of PYK2/CAK␤ tyrosine phosphorylation with bradykinin or lysophosphatidic acid in PC-12 cells (30,32,51), stem cell factor in CMK human megakaryocytic cells (38), and angiotensin II in GN4 rat liver cells (81) or rat vascular smooth muscle cells (39) is either not altered by EGTA or partially or completely inhibited by BAPTA/AM, suggesting that mobilization of intracellular calcium is playing an important role. Similarly, studies in other cell systems suggest the role PKC activation in modulating PYK2/CAK␤ tyrosine phosphorylation by receptors coupled to PLC may differ in different cells. PKC inhibitors or PKC down-regulation by preincubation with TPA had either minimal or no effect on stimulation of PYK2/CAK␤ tyrosine phosphorylation by bradykinin in PC-12 cells (30) or angiotensin II in cardiac fibroblasts (43), suggesting that PKC activation was not involved. In contrast, PYK2/ CAK␤ tyrosine phosphorylation was either partially or completely inhibited by PKC inhibitors or PKC down-regulation for the effect of thrombin in platelets (64), stem cell factor in megakaryocytes (38), P 2Y2 receptor activation in PC-12 cells (35), or platelet-derived growth factor in rat aortic smooth muscle cells (39), suggesting PKC activation was essential for some or all of the stimulation seen. In rat pancreatic acini, our results support the conclusion that mobilization of intracellular calcium stores by CCK-8, but not calcium influx, is responsible for almost 50% of the maximal stimulation of PYK2/CAK␤ tyrosine phosphorylation and that CCK A receptor ability to stimulate PKC activation is accounting for the other 50% of the PYK2/CAK␤ tyrosine phosphorylation. Therefore, our results demonstrate that in rat pancreatic acini the ability of CCK A receptor activation to stimulate PYK2/CAK␤ tyrosine phosphorylation is completely dependent on activation of both limbs of the PLC cascade.
Previous studies have shown that both PYK2/CAK␤ and p125 FAK in the same tissues (64,84) may become tyrosine phosphorylated in response to similar stimuli. However, some recent studies (37)(38)(39) suggest that with activation of some receptors PYK2/CAK␤ and p125 FAK tyrosine phosphorylation can be regulated differentially. For example, p125 FAK tyrosine phosphorylation is either not stimulated or minimally stimulated, whereas PYK2/CAK␤ tyrosine phosphorylation is markedly stimulated by stem cell factor in megakaryocytes (38) and by platelet-derived growth factor and angiotensin II in rat smooth muscle cells (37,39). In contrast, p125 FAK , but not PYK2 tyrosine phosphorylation, is stimulated by platelet aggregation (64), by activation of glutamate receptors in rat hippocampal slices (84), and by adhesion to fibronectin in rat aortic smooth muscle cells (37). Previous studies have reported that PYK2/CAK␤ has a more restricted tissue distribution than p125 FAK (41,42,85), suggesting that both kinases are likely to be cell type-specific functional equivalents (36,42,86). Furthermore, in recent studies in some cells expressing both PYK2/CAK␤ and p125 FAK it was proposed that PYK2/CAK␤ and p125 FAK , albeit highly homologous in primary structure, appear to have different functions (37,84,86,87). Because pancreatic acinar cells express both endogenous PYK2/CAK␤ and p125 FAK and tyrosine phosphorylation of both is stimulated by CCK, they provide an ideal system to compare regulation of agonist-activated G protein-coupled receptor stimulation of PYK2/CAK␤ and p125 FAK tyrosine phosphorylation in the same cell. As shown in the present study in parallel experiments and in a previous study (17), a number of results suggest that in rat pancreatic acini, tyrosine phosphorylation of PYK2/CAK␤ and p125 FAK after CCK A receptor activation is differentially regulated, and this could account for a different function of these kinases in rat pancreatic acini. First, CCKstimulated PYK2/CAK␤ tyrosine phosphorylation was more rapid and of greater magnitude than previously reported (17) for CCK stimulation of p125 FAK tyrosine phosphorylation. Second, stimulation of PYK2/CAK␤ tyrosine phosphorylation by CCK was reduced by 50% after the inhibition of increases in [Ca 2ϩ ] i or PKC inactivation, but CCK-stimulated p125 FAK tyrosine phosphorylation was not affected. Third, inhibition of both increases in [Ca 2ϩ ] i and PKC activation completely inhibited CCK stimulation of PYK2/CAK␤ tyrosine phosphorylation but caused only a 50% decrease in CCK-stimulated p125 FAK tyrosine phosphorylation. These results show p125 FAK tyrosine phosphorylation with CCK A receptor activation is controlled by both PLC-dependent and -independent mechanisms, whereas PYK2/CAK␤ tyrosine phosphorylation is completely controlled by PLC-dependent mechanisms. Similar to our results, a recent study in rat hippocampal slices (84) demonstrated that PYK2/ CAK␤ and p125 FAK tyrosine phosphorylation were regulated differently by changes in PKC or [Ca 2ϩ ] i . However, in contrast to our results, in rat hippocampal slices (84) ionomycin stimulated p125 FAK but not PYK2/CAK␤ tyrosine phosphorylation, whereas PKC activation by TPA caused the inverse. These results suggest that the regulation of these two structurally similar kinases by PLC-activated cascades appears to vary markedly in different cells.
Numerous recent studies demonstrate that the integrity of the actin cytoskeleton is important for various neuropeptides as well as bioactive lipids such as lysophosphatidic acid and sphingosylphosphocholine to stimulate tyrosine phosphorylation of the PYK2/CAK␤ structurally related kinase, p125 FAK (69, 88 -90). In rat pancreatic acini CCK-stimulated PYK2/ CAK␤, tyrosine phosphorylation was completely inhibited by disruption of the actin cytoskeleton with cytochalasin D, but not disruption of the microtubule network. These results are consistent with other reports indicating that PYK2/CAK␤ tyrosine phosphorylation after activation of G protein-coupled receptors by thrombin or angiotensin II (39,64,81), activation of growth factor receptors by platelet-derived growth factor or stem cell factor (38,39), or after integrin stimulation (36,91) was dependent on the integrity of the actin cytoskeleton.
A number of studies have investigated the distribution of PYK2/CAK␤ under basal conditions using immunocytochemical methods (31,41,86). In chicken embryo cells (86) PYK2/ CAK␤ was diffusely distributed throughout the cell with a small fraction in focal adhesions that are attached to the plasma membrane. In another study (31), in the CMK megakaryocyte cell line both PYK2/CAK␤ and p125 FAK were found in focal adhesions, whereas in COS-7 cells expressing PYK2/CAK␤ it was found only in cell to cell contacts but not in focal adhesions where p125 FAK was localized (41). In the present study, under basal conditions the majority (i.e. 77%) of PYK2/CAK␤ was localized in the membrane fraction. This was the opposite pattern to p125 FAK localization in which the majority (i.e. 86%) was localized in the cytosol. Thus, our results demonstrate a different localization of PYK2/CAK␤ and p125 FAK in rat pancreatic acini under basal conditions, which provides a foundation for a compartmentalization of these kinases, and this may contribute to their differential regulation after CCK A receptor activation in these cells. Moreover, the compartmentalization of different kinases has been proposed to be helped by formation of a cytoskeletal complex, which provides a foundation for the interactions of kinases with different regulators and substrates (36,64). This proposal is consistent with our results that show a clear relationship between the actin cytoskeleton and the tyrosine phosphorylation of both kinases, PYK2/CAK␤ and p125 FAK .
Tyrosine-phosphorylated proteins such as p125 FAK , p130 Cas , and p60 src have been shown to alter their cellular localization after cell activation (56,(72)(73)(74)(75). Our results show that in rat pancreatic acini after CCK A receptor activation, both p125 FAK and PYK2/CAK␤ undergo a change in cell distribution. There was a decrease in the amount of total PYK2/CAK␤ in cytosolic fractions after CCK A receptor activation, whereas with total p125 FAK , there was an increase in the total amount in the membrane fraction. Although a reciprocal increase in the amount of total PYK2/CAK␤ detectable in the corresponding membrane fractions or a decrease in the amount of total p125 FAK in cytosolic fractions was not detected, these changes in the smaller fraction probably represent a redistribution. This is the first report showing that PYK2/CAK␤ can alter its cellular localization after cell activation. What role the redistribution plays in the cellular function of PYK2/CAK␤ is at present unclear.
Recent studies with lysophosphatidic acid (32), growth factors such as colony-stimulating growth factor (92), and some G protein-coupled receptors such as angiotensin II (43,44), UTP (33,35), bradykinin (30,32), and ␣and ␤-adrenergic agents (34) demonstrate that activation of PYK2/CAK␤ tyrosine kinase is an intracellular regulator of the ability of these stimuli to activate the MAPK signaling pathway (30, 32-35, 51, 92). The MAPK signaling pathway represents an important point of convergence of cell signaling, especially in the regulation of cell division and growth (21,93,94). This cascade is regulated by protein phosphorylation, with the MAPKs themselves being serine/threonine kinases and including p38 MAPK , the Jun Nterminal kinase/stress-activated protein kinase, and the extracellular signal-regulated kinases 1 and 2 (21,93,94). Previous studies in rat pancreatic acini (15,16,76) demonstrate that activation of the CCK A receptor causes Ras/MAPK signaling pathway activation via a mechanism involving protein kinase C, calcium mobilization, and tyrosine kinases. However, the full mechanism by which CCK activates Ras/MAPK signaling pathway in rat pancreatic acini is still under study, and the tyrosine kinases involved in this process are still unidentified. A recent study in rat pancreatic acini (14) shows that CCK stimulates formation of a Shc-Grb2-Sos complex, which plays a central role in mediating Ras/MAPK signaling pathway activation (93), by a PKC-dependent mechanism. However, it is unknown how CCK A receptor activation leads to tyrosine phosphorylation of Shc and stimulates the formation of the Shc-Grb2-Sos complex. With other stimuli, recent studies show that PYK2/CAK␤ activation may lead to activation of the MAPK cascade by a number of different mechanisms. These mechanisms include through a Src-mediated pathway (32,34), by transactivation of the epidermal growth factor receptor (35,44), by stimulation of the c-Jun N-terminal kinase pathway (40,80), or by Shc tyrosine phosphorylation and Shc-Grb2-Sos complex formation (30,32,34,43). A number of results in our study suggest that PYK2/CAK␤ tyrosine phosphorylation and activation might be an important intermediate that translates increases in intracellular calcium and PKC activation induced by CCK A receptor occupation into activation of the Ras/MAPK signaling pathway. First, we found that PYK2/CAK␤ associated with Grb2 in a time-dependent manner. Previous studies in other tissues have demonstrated that tyrosine-phosphorylated PYK2/CAK␤ could directly, by means of its SH2 domain, or indirectly, by means of tyrosine phosphorylation of Shc, recruit Grb2, bind Sos by means of its SH3 domains, and induce activation of the Ras/MAPK signaling pathway (30,32,34,60). Second, in the present study, we demonstrated by co-immunoprecipitation that in rat pancreatic acini, CCK A receptor activation results in a rapid formation of a complex between PYK2/ CAK␤ and c-Crk. c-Crk belongs to a family of adapter proteins that consist almost entirely of SH2 and SH3 domains and include the oncogene v-Crk, two forms of c-Crk (Crk-I and Crk-II) and a Crk-like protein (CrkL). Crk phosphorylation results in coupling by its SH3 domains to various guanine nucleotide exchange factors including Sos (79,95,96), activation of which stimulates the MAPK cascade (21,93). Therefore, tyrosine phosphorylation of c-Crk has been proposed as an alternative pathway to activate the Ras/MAPK signaling pathway (77)(78)(79). Third, we found that both the time course of stimulation of PYK2/CAK␤ tyrosine phosphorylation and its activation by CCK and its stoichiometry were consistent with the ability of CCK to stimulate the MAPK signaling pathway as previously reported in pancreatic acini (14,16,76). Specifically, CCK-induced PYK2/CAK␤ tyrosine phosphorylation and stimulation of its kinase occurs rapidly (1-2.5 min) and precedes the peak increase in Shc tyrosine phosphorylation and MAPK activity caused by CCK that was reported to occur at 5 min (14,16,76). In addition, the dose-response curve for CCK-8 stimu-lation of PYK2/CAK␤ tyrosine phosphorylation we found in the present study and that for MAPK signaling pathway activation reported in pancreatic acini previously (16,76) are similar and over the same range in which CCK caused changes in [Ca 2ϩ ] i (17). Last, we found a similar calcium and PKC dependence of CCK-stimulated PYK2/CAK␤ tyrosine phosphorylation to that previously reported for CCK-stimulated Shc tyrosine phosphorylation (14) and MAPK signaling pathway activation (16,76). Specifically, in pancreatic acini, activation of PKC or calcium mobilization induced PYK2/CAK␤ tyrosine phosphorylation and MAPK signaling pathway activation (16,76). Furthermore, simultaneous activation of both processes, PKC activation and calcium mobilization, caused an augmented effect on PYK2/CAK␤ tyrosine phosphorylation in the present study and also for MAPK signaling pathway activation in a previous study (76). Moreover, in rat pancreatic acini we found that inhibition of PKC activation by GF109203X decreased CCKstimulated PYK2/CAK␤ tyrosine phosphorylation similar to that reported previously for activation of the MAPK signaling pathway by CCK (76) and CCK stimulation of Shc tyrosine phosphorylation (14). These data suggest that, in rat pancreatic acini, PYK2/CAK␤ might be one of the tyrosine kinases that act upstream to activate Ras/MAPK signaling pathway after CCK A receptor activation. Therefore, in pancreatic acini we propose the model shown in Fig. 13 for the ability of CCK A -R activation to activate the MAPK signaling cascade. As discussed above, this model is consistent with the known ability of CCK A -R activation to stimulate both cascades of the PLC pathway (i.e. increases in [Ca 2ϩ ] i and PKC activation) and, as shown in the present study, are essential for CCK A -R activation of PYK2/CAK␤. As we demonstrate in the present study, the activation of PYK2/CAK␤ results in the formation of both a PYK2/CAK␤-Grb2 complex and a PYK2/CAK␤-Crk complex, both of which could activate the Ras/MAPK signaling cascade. Whether c-Src activation is involved in activation of the MAPK signaling cascade by PYK2/CAK␤, as reported in other tissues (32,34,45,46), is at present unclear. FIG. 13. Proposed model of CCK A receptor-mediated MAPK signaling pathway activation in rat pancreatic acini. Stimulation of CCK A receptor leads to phospholipase C activation, which results in increases in intracellular calcium and activation of PKC. Simultaneous activation of both of these PLC cascades results in tyrosine phosphorylation and activation of the nonreceptor tyrosine kinase PYK2/CAK␤. The activation of PYK2/CAK␤ results in the recruitment of the Grb2-Sos or Crk-Sos complex to the membrane, activation of Ras through guanine nucleotide exchange, and subsequent activation of Raf, which initiates the MAPK signaling pathway. PYK2/CAK␤ could activate the recruitment of the Grb2-Sos complex directly or possibly by c-Src activation, which results in phosphorylation of the Shc adapter protein. This latter cascade is shown with a question mark because there is at present no evidence that CCK-stimulated changes caused by PYK2/ CAK␤ use this pathway. MEK, mitogen-activated protein kinase/extracellular signal-regulated kinase kinase.