The CK2 Phosphorylation of Vitronectin PROMOTION OF CELL ADHESION VIA THE αvβ3-PHOSPHATIDYLINOSITOL 3-KINASE PATHWAY

Phosphorylation of vitronectin (Vn) by casein kinase II was previously shown to occur at Thr50 and Thr57 and to augment a major physiological function of vitronectin-cell adhesion and spreading. Here we show that this phosphorylation increases cell adhesion via the αvβ3 (not via the αvβ5 integrin), suggesting that αvβ3 differs from αvβ5 in its biorecognition profile. Although both the phospho (CK2-PVn) and non-phospho (Vn) analogs of vitronectin (simulated by mutants Vn(T50E,T57E), and Vn(T50A,T57A), respectively) trigger the αvβ3 as well as the αvβ5 integrins, and equally activate the ERK pathway, these two forms are different in their activation of the focal adhesion kinase/phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB) pathway. Specifically, we show (i) that, upon exposure of cells to Vn/CK2-PVn, their PKB activation depends on the availability of the αvβ3 integrin on their surface; (ii) that upon adhesion of the β3-transfected cells onto the CK2-PVn, the extent of PKB activation coincides with the enhanced adhesion of these cells, and (iii) that both the PKB activation and the elevation in the adhesion of these cells is PI3K-dependent. The occurrence of a cell surface receptor that specifically distinguishes between a phosphorylated and a non-phosphorylated analog of Vn, together with the fact that it preferentially activates a distinct intra-cellular signaling pathway, suggest that extra-cellular CK2 phosphorylation may play an important role in the regulation of cell adhesion and migration.

Phosphorylation of vitronectin (Vn) by casein kinase II was previously shown to occur at Thr 50 and Thr 57 and to augment a major physiological function of vitronectin-cell adhesion and spreading. Here we show that this phosphorylation increases cell adhesion via the ␣ v ␤ 3 (not via the ␣ v ␤ 5 integrin), suggesting that ␣ v ␤ 3 differs from ␣ v ␤ 5 in its biorecognition profile. Although both the phospho (CK2-PVn) and non-phospho (Vn) analogs of vitronectin (simulated by mutants Vn(T50E,T57E), and Vn(T50A,T57A), respectively) trigger the ␣ v ␤ 3 as well as the ␣ v ␤ 5 integrins, and equally activate the ERK pathway, these two forms are different in their activation of the focal adhesion kinase/phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB) pathway. Specifically, we show (i) that, upon exposure of cells to Vn/ CK2-PVn, their PKB activation depends on the availability of the ␣ v ␤ 3 integrin on their surface; (ii) that upon adhesion of the ␤ 3 -transfected cells onto the CK2-PVn, the extent of PKB activation coincides with the enhanced adhesion of these cells, and (iii) that both the PKB activation and the elevation in the adhesion of these cells is PI3K-dependent. The occurrence of a cell surface receptor that specifically distinguishes between a phosphorylated and a non-phosphorylated analog of Vn, together with the fact that it preferentially activates a distinct intra-cellular signaling pathway, suggest that extra-cellular CK2 phosphorylation may play an important role in the regulation of cell adhesion and migration.
One of the most important properties of Vn is its ability to promote cell attachment, spreading, and migration (19 -22). In fact, Vn was originally discovered as a "serum spreading factor" (23). The cell adhesion, spreading, and migration activities of Vn are associated with its RGD sequence located near the N terminus of the protein (positions [45][46][47]. This sequence is recognized by the family of receptors known as the integrins: heterodimers composed of ␣ and ␤ subunits (24 -30). There are 17 ␣ and 8 ␤ subunits that heterodimerize to produce 22 different integrins (27,31,32). Several of these integrins, e.g. 6 , and ␣ v ␤ 8 and the platelet-specific ␣ IIb ␤ 3 integrin, are known to recognize and bind Vn.
It is well known that cell adhesion is a complex process that was shown to involve an activation of several Vn receptors and a variety of intra-cellular signaling pathways. For example, the focal adhesion kinase (FAK) was shown to play a central role in mediating the signal from integrins (33). It does so by its autophosphorylation on Tyr 397 upon integrin stimulation. This autophosphorylation leads to the recruitment and activation of intra-cellular mediators such as PI3K, as well as the Src family kinases, by an interaction of their SH2 domain with the autophosphorylated Tyr 397 residue. The PI3K binding to Tyr 397 leads to activation of PKB, whereas the Src family of kinases further phosphorylates FAK on Tyr 925 leading to the recruitment of additional signaling molecules that bring about an activation of the ERK pathway (31)(32)(33)(34)(35)(36)(37)(38).
We have previously shown that Vn can be functionally modulated by extra-cellular phosphorylation, making use of the kinase co-substrate ATP found at micromolar levels in the exterior of cells (39). For example PKA, released from platelets upon their physiological stimulation with thrombin (40 -42), selectively phosphorylates Vn, and, as a consequence of this phosphorylation, it reduces its grip on plasminogen activator inhibitor-1 (43). Similarly, PKC phosphorylation of Vn was shown to attenuate its cleavage by plasmin (44). Several laboratories have shown the occurrence of an extra-cellular CK2 activity on a variety of cells. These include epithelial cells (45,46), neutrophils (47,48), platelets (49,50), and endothelial cells (51)(52)(53). Subsequently, we showed that Vn is a substrate for CK2, which phosphorylates Vn at Thr 50 and Thr 57 . Furthermore, we found that this phosphorylation significantly enhances the adhesion and spreading of bovine aorta endothelial cells (BAEC), presumably because the phosphorylated Vn has a higher affinity for ␣ v ␤ 3 (54).
One of the major obstacles in revealing the mechanism of action of CK2-phosphorylated Vn originates from the well known fact that Vn (like other adhesion proteins) can bind to several integrins, including the specific Vn-binding integrin, ␣ v ␤ 5 , and that this family of integrins can, in turn, activate different intra-cellular pathways. Here we extend our studies on the consequences of the CK2 phosphorylation of Vn and * This research was supported in part by the Israel Science Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ show that the enhanced cell adhesion involves ␣ v ␤ 3 (but not ␣ v ␤ 5 ). Furthermore, we show that this enhanced adhesion coincides with a preferential activation of the FAK/PI3K/PKB cascade, rather than the ERK signaling pathway.
Tissue Cultures-HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) heat-inactivated fetal calf serum and glutamine (0.5 mg/ml). H1299 cells were grown in RPMI supplemented with 10% (v/v) heat-inactivated fetal calf serum and glutamine (0.5 mg/ml). The cells were grown in an incubator (37°C) with an atmosphere containing 5% CO 2 . The Sf-9 and High-5 insect cells were maintained in Grace's insect medium (Life Technologies, Inc.) supplemented with 10% (v/v) heat-inactivated fetal bovine serum and grown in an incubator (27°C). For the expression of recombinant Vns, a serum-free medium (Sf-900 II, Life Technologies, Inc.) was used. All media for insect cells were supplemented with 50 g/ml Gentamicin and 12.5 g/ml Fungizone (Life Technologies, Inc.).
Cell Adhesion Assay-Serial dilutions of r-Vns were added to 24-well plates (250 l) for 1.5 h at 22°C to allow coating of the plates. Thereafter the solutions were aspirated, and 0.5 ml of serum free medium containing 1 mg/ml hemoglobin was added for 30 min at 37°C. Confluent cells plated on 10-cm plates were labeled with 30 Ci of [ 35 S]methionine for 3-4 h at 37°C. The cells were collected (using 5 mM EDTA) into serum free medium, centrifuged (5 min at 1200 ϫ g), and resuspended into a serum free medium adjusting their concentration to 10 6 cells/ml. Cell suspensions (250 l) were added to each coated well for 30 min at 37°C. The cells were washed three times with 0.5 ml of PBS, and the adhered cells were treated with 0.5 ml of 1% Triton X-100 in PBS for 5 min. Samples of 0.4 ml were transferred into scintillation vials for counting. The quantitation of cell adhesion is reported as the residual radioactivity (a mean of triplicates in cpm) of the cells tested, after their extensive washing (three times with 0.5 ml of PBS). This comparison was convenient and valid, because each assay was carried out with an identical volume of cell suspension, and an identical number of cells. When cell adhesion assays were performed in 48-well plates, all the components and treatments of the assay were scaled down accordingly.

Inhibition of Cell Adhesion by Function-inhibiting Monoclonal
Antibodies-The monoclonal antibodies used were: P1F6, directed against the integrin receptor ␣ v ␤ 5 ; LM609, directed against the integrin receptor ␣ v ␤ 3 ; and HA, directed against hemagglutinin as control. Plates (24 wells) were coated with 5 g/ml of the Vn to be assayed (250 l) for 1.5 h 22°C, then the nonspecific adsorption sites were blocked with 0.5 ml of serum free medium containing 1 mg/ml hemoglobin (30 min at 37°C). The cells were treated as described above to yield a concentration of 10 5 cells/ml. Before starting the cell adhesion assay, the cells were preincubated with increasing concentrations of monoclonal antibodies (gentle shaking, for 30 min at 22°C). Thereafter, the cells were washed once with 10 ml of serum free medium containing 1 mg/ml hemoglobin and resuspended to yield a concentration of 10 5 cells/ml. An aliquot of this cell suspension (250 l) was added to the Vn-coated wells, and the adhesion assay was allowed to proceed as described above.
FACS Analysis-Confluent cells grown on 10-cm plates were collected as described under cell adhesion and brought to a concentration of 5 ϫ 10 5 cells in 100 l of PBS containing 1% bovine serum albumin and 0.02% sodium azide. The cells were incubated with monoclonal antibodies (final concentration, 4 g/100 l) for 1 h on ice with occasional agitation. They were then washed three times with 1 ml of PBS containing 1% bovine serum albumin, and 0.02% sodium azide using a cooled microcentrifuge (4°C). After the last wash, the cells were resuspended in 100 l of the above-mentioned buffer, supplemented with FITC-conjugated goat anti-mouse IgG (final concentration of 5 g/100 l). The cells were allowed to bind the antibodies during 1 h (on ice) with occasional agitation, then washed as above and resuspended in 0.5 ml of PBS (containing the above constituents) for FACS analysis in a FACScan Becton Dickinson (530 filter). For each antibody, 5000 cells were analyzed. Control cells were incubated with the secondary antibody only.
Expression of the ␤ 3 Integrin Subunit and the r-Vn Mutants-The cDNA encoding the ␤ 3 integrin subunit in pGEM was kindly provided by Dr. P. J. Newman, Blood Research Institute, Milwaukee, WI. The cDNA was digested with DraI and XbaI then treated with Klenow and subcloned into an EcoRV-digested pcDNA3 vector. Transfections of H1299 cells were done using LipofectAMINE according to the manufacturer's instructions (Life Technologies, Inc.). The cells were transfected with the ␤ 3 subunit cDNA in pcDNA3 or, for control, with the empty vector of pcDNA3. Transfected cells were grown on 0.6 mg/ml Geneticin (G418), and single stable clones were isolated. Preparation of the r-Vn mutants and their expression in insect cells was carried as described previously (54).
Preparation of Cell Lysates for the Detection of Activated Kinases (ERK, JNK, p38 MAPK, PKB, and FAK)-Plates (10 cm) were coated with the r-Vns for 1.5 h at 22°C. Thereafter the solutions were aspirated and 3 ml of serum free medium containing 1 mg/ml hemoglobin was added and incubated for 30 min at 37°C. Serum-starved cells were collected (using 5 mM EDTA) into serum free medium containing 1 mg/ml hemoglobin (10 6 cells/ml). The cells were plated on top of the r-Vns and incubated for various time periods at 37°C then washed three times with PBS (ice-cold) and scraped (on ice) into 500 l of a RIPA buffer. The lysates were collected and centrifuged (20,000 ϫ g 15 min at 4°C), and aliquots of the resulting supernatants were assayed for their protein concentration (Pierce protein assay).
Detection of Kinase Activation-Equal amounts of proteins obtained from the cell lysates described above were loaded onto SDS-PAGE, transferred to nitrocellulose paper, and immunoblotted with antibodies exclusively recognizing the active form of the kinase in question (antiactivated ERK, JNK, p38 MAPK, or PKB antibodies). The same samples were also analyzed using anti-total kinase antibodies, which detect the total amount of the kinase in question (activated and non-activated).
Detection of FAK Phosphorylation-Protein samples (600 g) obtained from the cell lysates described above were immunoprecipitated using anti-FAK antibodies immobilized on agarose beads (mixing end to end for 2 h at 4°C). The immunoprecipitated samples were washed once with RIPA buffer, twice with 0.5 M LiCl, 0.1 M Tris-HCl, pH 8.0, and finally twice in 50 mM ␤-glycerophosphate, pH 7.3, 1.5 mM EGTA, 1 mM EDTA, 1 mM dithiothreitol, and 0.1 sodium vanadate. After the last wash, the samples were boiled in Laemmli's sample buffer and subjected on SDS-PAGE. The gels were transferred to nitrocellulose membranes and blotted either with antibodies against phosphotyrosine (PY99, to detect phosphorylated FAK), or with antibodies against FAK (to determine the total FAK as a reference value) in each lane.

RESULTS
Comparing the Adhesion of ␣ v ␤ 3 -and of ␣ v ␤ 5 -bearing Cells in Their Response to Vn and to CK2-phosphorylated Vn-We have previously shown (54) that the CK2 phosphorylation of Vn results in a significant enhancement of BAEC cell adhesion (ϳ2.5-fold, average of three experiments), as indicated by the number of cells that adhere to increasing concentrations of immobilized Vn. We also showed that the effect of the CK2 phosphorylation could be reproduced with a mutant Vn(T50E,T57E) (a close analog of CK2-PVn representing the phospho form of Vn), when compared with Vn(T50A,T57A) (a close analog of Vn representing the non-phospho form of Vn).
In the course of our studies we found that BAEC cells do not express ␣ v ␤ 5 (a characteristic binding receptor for Vn (55)); therefore, we considered the possibility that this integrin might be involved in a response to CK2-PVn by cells that do express this integrin. To find out whether this is the case, we used HeLa cells (Fig. 1A) and H1299 cells (Fig. 1B), whose adhesion to Vn was found to be mediated mainly by ␣ v ␤ 5 . In both cases we found an efficient inhibition of cell adhesion by anti-␣ v ␤ 5 but a minor inhibition by anti-␣ v ␤ 3 . A similar inhibition of cell adhesion by both antibodies was also obtained with Vn(T50A,T57A) (not shown), raising the possibility that the adhesion of these cells to both forms of Vn is mediated by ␣ v ␤ 5 . In line with this finding, the adhesion profile of HeLa as well as H1299 cells to immobilized Vn(T50E,T57E) was found to be essentially identical to their adhesion to Vn(T50A,T57A) ( Fig.  1, C and D). In this context it should be noted that (i) the same adsorption profile of the cells was obtained whether Vn(T50E,T57E) or Vn(T50A,T57A) was used as a substratum (54) and (ii) in all experiments comparing Vn(T50A,T57A) with Vn(T50E,T57E) we ran a similar experiment with wild type r-Vn and showed that, within experimental error, it was identical to Vn(T50A,T57A).
Cells Containing ␣ v ␤ 3 Exhibit an Enhanced Cell Adhesion upon Exposure to CK2-PVn-The results presented above, together with our previous findings with BAEC (54), imply that the enhanced cell adhesion onto CK2-PVn is mediated by the ␣ v ␤ 3 receptor. To confirm this suggestion we endowed H1299 cells (which do not exhibit an enhanced cell adhesion in response to CK2-PVn) with a capability to exhibit an enhanced cell adhesion onto Vn(T50E,T57E) and thus to "discriminate" between the phospho-and non-phospho forms of Vn. This was achieved by transfecting H1299 cells with the ␤ 3 subunit. 2 Isolated clones of H1299 cells overexpressing ␣ v ␤ 3 that were identified by immunoblotting with anti-␤ 3 , and subsequently characterized by FACS analysis with anti-␣ v ␤ 3 (Fig. 2, A and  B), were shown to contain high amounts of the ␣ v ␤ 3 integrin on their surface. Quantitation of the FACS analysis indicated that the ␤ 3 -transfected clones we used contained up to ϳ7-fold more ␣ v ␤ 3 than the control vector-transfected clones, whereas the amounts of the ␣ v and of a non-relevant ␣ 3 integrin were very similar to the control. In addition, we observed a ϳ3-fold reduction of ␣ v ␤ 5 in the ␤ 3 -transfected clone, presumably due to competition between ␤ 5 and the excess of ␤ 3 for the limited amount of their common partner, the ␣ v subunit.
The involvement of ␣ v ␤ 3 (but not ␣ v ␤ 5 ) in the enhanced cell adhesion is best illustrated in Fig. 3, which shows that the adhesion of vector-transfected H1299 cells is blocked by anti-␣ v ␤ 5 and not by anti-␣ v ␤ 3 (Fig. 3A), whereas the adhesion of ␤ 3 -transfected H1299 cells is blocked by anti-␣ v ␤ 3 but not by anti-␣ v ␤ 5 (B). In line with these findings, the vectortransfected H1299 cells do not discern Vn(T50E,T57E) from Vn(T50A,T57A), whereas cells overexpressing the ␤ 3 subunit exhibit an ability to enhance cell adhesion on the Vn(T50E,T57E) mutant (compare Fig. 3C with Fig. 3D). It should be noted that the occurrence of a relationship between the integrin content of cells, their adhesion, and the ensuing intracellular signaling triggered by Vn were also observed with two additional ␤ 3 -transfected clones (not shown).
An ERK Activation Cannot Account for the Enhanced Cell Adhesion Observed with CK2-PVn-Following the identification of ␣ v ␤ 3 as a CK2-PVn-specific mediator of the enhanced adhesion obtained with this phosphorylation, we attempted to identify an intra-cellular signaling pathway that might be responsible for this enhancement. Because the activation of ERKs in response to the stimulation of cells by ECM proteins was already established (31-38), we first examined the pattern of ERK activation in the stable ␣ v ␤ 3 and ␣ v ␤ 5 expressing clones of the H1299 cells mentioned above. In response to cell adhesion to r-Vns, the ERK activation of ␣ v ␤ 5 -and ␣ v ␤ 3 -containing clones was found to be low and transient (Fig. 4, A and B): It was found to peak within 10 min after plating and to decline thereafter. No significant change in the pattern of ERK activation that could correlate with the enhancement of cell adhesion was observed (Fig. 4C). These results raised the possibility that an alternative signaling pathway(s) (other than the ERK pathway), might be involved in the enhanced adhesion observed with the ␤ 3 -transfected clone.
The Increased Activation of the PKB Pathway Can Account for the Enhanced Cell Adhesion Mediated by ␣ v ␤ 3 -Because we found that the activation of ERK cannot account for the enhanced cell adhesion, we looked into other signaling pathways such as the JNK, p38 MAPK, and PKB pathways that were previously shown to be activated by Vn-binding integrins. Although no adhesion-triggered activation of JNK and p38 MAPK was detected in the various clones we used (data not shown), we found that the activation of PKB in the ␤ 3 -transfected cells (Fig. 5) led to a significantly enhanced activation of this kinase, in comparison to the very low PKB activation in the vector-transfected cells. 3 These results suggested to us that the activation of PKB depends on the availability of the ␣ v ␤ 3 integrin. As such, the extent of PKB activation in the ␤ 3 -transfected cells correlates well with the extent of enhanced cell adhesion onto CK2-PVn. This was demonstrated with ␤ 3 -transfected cells that were plated on Vn(T50E,T57E), whose enhanced adhesion resulted in an increased PKB activation (ϳ30fold over the PDL control), whereas the PKB activation obtained in cells plated onto Vn(T50A,T57A) was found to be only 18-fold over the control (Fig. 5C).
PI3K Is Essential for the Promotion of Cell Adhesion and for the Activation of PKB-PKB was recently implicated as an important downstream target for PI3K (56). To determine whether the PKB activation in our system requires the activation of PI3K (which precedes PKB in several signal transduction processes (cf. Scheme 1), we treated ␤ 3 -transfected cells with wortmannin (a PI3K inhibitor) prior to their stimulation by adhesion to Vn(T50E,T57E). Indeed, wortmannin prevents PKB activation (Fig. 6A), presumably through a PI3K inhibition, indicating that the enhanced adhesion mediated by ␣ v ␤ 3 transmits the signal to PKB via PI3K. In line with this result, the enhanced adhesion of the ␤ 3 -transfected cells was reduced by preincubation with wortmannin (before allowing the cells to adhere) (Fig. 6B) or with another PI3K inhibitor LY294002 (Fig. 6C). As expected, these two inhibitors blocked cell adhesion onto both the phospho-and the non-phospho forms of Vn, because PKB is activated by both forms of Vn. However, the reduction in the elevation in cell adhesion onto CK2-PVn over Vn, illustrated by using PI3K inhibitors, clearly indicates the involvement of this pathway in elevating cell adhesion on CK2-PVn. The specific involvement of the PI3K-PKB pathway in the adhesion of cells onto the phospho and the non-phospho forms of Vn was supported by our finding that the MEK inhibitor PD98059 does not inhibit the cell adhesion onto these two Vns (Fig. 6D). In that context it should be noted that, in our experiments, the PKB activation occurs within 5-10 min after plating the cells (the cells are attached but not spread), whereas the adhesion of the cells proceeds for 30 min when the cells are already adhered and spread. This observation sets the stage for a detailed study aimed at the identification of the sequence of events that lead from cell adhesion to cell spreading, namely at the elucidation of the mechanism by which downstream mediators of PKB influence the cell-spreading process.
In conclusion, it is evident from our results (i) that the PKB activation (which occurs upon exposure of cells to Vn/CK2-PVn) depends on the availability of the ␣ v ␤ 3 integrin on the surface of the cells; (ii) that the extent of PKB activation (that takes place upon exposure of the ␤ 3 -transfected cells to CK2-PVn) coincides with the specific enhanced adhesion of these cells upon their binding to CK2-PVn; and (iii) that both the PKB activation and the subsequent enhanced adhesion of the cells are PI3K-dependent, because the inhibition of PI3K (upstream of PKB) prevents the PKB activation and reduces cell adhesion (Scheme 1).
Cells Containing ␣ v ␤ 3 and ␣ v ␤ 5 Differ in Their FAK Phosphorylation Pattern upon Their Adhesion onto Phospho and Non-phospho Forms of Vn-As mentioned above, there is a significant difference in the intensity of the PKB activation upon exposure of ␤ 3 -transfected cells to the phospho and the non-phospho forms of Vn (Fig. 5). To account for this difference in intensity (shown here to be PI3K-dependent (Fig. 6A)), we compared their FAK phosphorylation pattern, i.e. the possible activation of an upstream kinase in this pathway. It is well known that the phosphorylation of FAK is an early event detected in response to integrin stimulation (33). Upon this stimulation, FAK is autophosphorylated on Tyr 397 , creating a high affinity binding site for a variety of kinases containing an SH2 domain, including the PI3K and Src family kinases. Src further phosphorylates FAK on Tyr 925 , leading to the recruitment of GRB2, which is known to activate the ERK pathway (38). Therefore, we monitored the FAK phosphorylation upon attachment of vector/␤ 3 -transfected cells onto the r-Vns mentioned above. As seen in Fig. 7, the FAK phosphorylation is different in these two cell lines. Although there is a gradual activation of FAK that peaks after 20 min in the vector-transfected clone (Fig. 7A), the FAK phosphorylation in ␤ 3 -transfected cells is weaker and transient. It peaks after 5-10 min and declines thereafter (Fig. 7B). The time course of FAK phosphorylation in the ␤ 3 -transfected cells coincides with that of PKB activation (Fig. 5B). Moreover, although no differences in FAK phosphorylation were observed when vector-transfected cells were plated either on Vn(T50A,T57A) or on Vn(T50E,T57E) (Fig. 7, B and C), a preferential increase in FAK phosphorylation was observed when ␤ 3 -transfected cells were plated on the Vn(T50E,T57E) mutant (5 min, Fig. 7B). Although a small increase, this signal is amplified, and a better reflection of it is viewed in the differences observed in the downstream kinase PKB (Fig. 5). The FAK phosphorylation in the vector-transfected cells is significantly more intense at the peak of the response (20 -30 min). This may suggest that in the vector-transfected cells (␣ v ␤ 5 ) another kinase may further phosphorylate FAK, whereas in the FAK autophosphorylation brings about the association with PI3K, which does not further phosphorylate FAK but, rather, specifically activates the PKB pathway. We conclude that the extra-cellular stimulation by CK2-PVn (as represented by Vn(T50E,T57E)) is transmitted via ␣ v ␤ 3 and that the PI3K pathway is involved in the enhanced cell adhesion.

DISCUSSION
Intra-cellular protein phosphorylation is now well established as a central regulatory mechanism. In the last few years, several reports provided evidence for the occurrence of protein kinases outside the cell, raising the possibility that protein phosphorylation may also regulate extra-cellular processes (40,41,(45)(46)(47)(48). This possibility was supported by the identification of specific target substrates for the kinases in the cell exterior. Some reports further indicated that the physiological function of such specific substrates is modulated upon their phosphorylation (for a review see Ref. 42). For example, it was shown that Vn is functionally modulated by PKA, a kinase released from platelets upon their physiological stimulation with thrombin (40 -42). Similarly, a PKC phosphorylation of Vn was shown to attenuate its cleavage by plasmin (44).
In addition to PKA and PKC, Vn was recently shown to be a substrate for CK2, which was found to single out and selectively phosphorylate Vn at Thr 50 and Thr 57 to bring about a significant enhancement of one of Vn's well known physiological functions: cell adhesion and spreading (54). The clinical importance of this modulation is evident in view of the fact that invasive metastasis involves an enhanced adhesion of tumor cells to the ECM (6) by binding to integrins, in particular ␣ v ␤ 3 . In fact, this integrin has been implicated in the acquisition of analog (T50A,T57A). A, inhibition of vector-transfected H1299 cell adhesion to r-Vns by antibodies raised against ␣ v ␤ 3 or ␣ v ␤ 5 . B, inhibition of ␤ 3 -transfected H1299 cell adhesion to r-Vns by antibodies raised against ␣ v ␤ 3 or ␣ v ␤ 5 . C, adhesion of vector-transfected H1299 cells to r-Vns. D, adhesion of ␤ 3 -transfected H1299 cells to r-Vns. The assay of cell adhesion and its inhibition were carried out as described in the legend to Fig. 1. metastatic invasiveness (57). In melanoma, for example, the expression of ␣ v ␤ 3 was shown to correlate with invasiveness (58) and with tumorigenic capacity (57,59). In the case of Vn, the specificity in the recognition of its CK2-phosphorylated form may have a special importance in cancer, because Vn seems to be an important ligand in the ␣ v ␤ 3 -mediated adhesion of tumor cells. In line with this fact, human melanoma cells derived from lymphatic metastases were shown to use ␣ v ␤ 3 to adhere to lymph node Vn (7), in a Vn-mediated manner, as indicated by the fact that the replacement of Vn by fibronectin had no effect on invasion (60).

FIG. 3. Transfection with ␤ 3 endows H1299 cells with the capability to discriminate between the CK2-PVn analog Vn(T50E,T57E) and the non-phosphorylated Vn
A major implication of the findings presented in this report is that the CK2 phosphorylation of Vn enhances cell adhesion via ␣ v ␤ 3 and not via ␣ v ␤ 5 . This can be deduced from our finding that cells that adhere mostly via ␣ v ␤ 5 , (e.g. HeLa cells, or the H1299 lung carcinoma cells) do not distinguish between the mutant Vn(T50E,T57E) and Vn(T50A,T57A). Furthermore, we report here that the enhanced cell adhesion can be quantitatively accounted for by assuming that the integrin ␣ v ␤ 3 alone is involved in the preferential recognition of the Vn analog Vn (T50E,T57E), i.e. in the specific response to CK2-PVn. In line with this conclusion the H1299 lung carcinoma cells (whose adhesion to Vn is mediated by ␣ v ␤ 5 ) are not able to discriminate between Vn(T50E,T57E) and Vn(T50A,T57A) but gain the ability to discriminate between these two mutants upon their stable transfection with the ␤ 3 integrin subunit.
In view of our finding that the enhanced cell adhesion onto Vn(T50E,T57E) is mediated by ␣ v ␤ 3 and our previous observation that this enhancement is due to an increased affinity toward this integrin, we undertook to identify the signaling pathway that is involved in this increased affinity. Several signaling cascades were previously shown to be activated by the integrin family of receptors (31-34, 36 -38, 61, 62). Having in our hands cells that use ␣ v ␤ 3 to adhere onto Vn, and essentially identical companion cells that use ␣ v ␤ 5 to adhere to this integrin ligand, enabled us to identify a signaling pathway, which is differentially activated upon adhesion of these cells to CK2-phosphorylated and non-phosphorylated Vn analogs. Specifically, we were able to show that the phospho and nonphospho forms of Vn trigger both ␣ v ␤ 3 and ␣ v ␤ 5 , leading to a similar activation of ERK. These results suggest that the activation of ERK occurs via the ␣ subunit (61), which has not been modified in these cells. The fact that the activation of ERK is not influenced by the introduction to the ␤ 3 subunit supports this suggestion. In contrast, the PKB activation seems to depend on the availability of the ␤ 3 subunit, and therefore, is preferentially activated by the phospho form of Vn. We pre-sume that this enhanced activation of PKB, which is ␣ v ␤ 3 -and PI3K-dependent, results in the enhanced cell adhesion by the CK2-PVn analog (Vn (T50E,T57E)).
Based on the results presented here, we suggest that although both ␣ v ␤ 3 and ␣ v ␤ 5 share common structural elements that recognize and bind equally well the core protein shared by Vn and PVn, ␣ v ␤ 3 contains additional recognition elements that bind the two phosphate groups specifically introduced in Vn by its CK2 phosphorylation. Specifically, this result raises the possibility that the ligand binding site of ␣ v ␤ 3 possesses recognition elements to CK2-PVn that are not present in ␣ v ␤ 5 . This suggestion, which is supported by additional experimental evidence using a set of RGD-containing peptides as inhibitors, 4 can account for the distinct behavior of the ␣ v ␤ 3 and ␣ v ␤ 5 integrins and specifically for the ␣ v ␤ 3 -mediated enhanced activation of the PI3K/PKB pathway that correlates with the increased cell adhesion.
One of the important messages reported here lies in the fact that it identifies two intracellular signaling pathways that are unequally activated upon binding of Vn and CK2-PVn, at least to the cells we tested in this study. Both pathways (one functioning via ␣ v ␤ 3 and ␣ v ␤ 5 and one via ␣ v ␤ 3 (Scheme 1)) are activated upon adhesion of the cells onto Vns. However, although the activation of ERK (triggered by both ␣ v ␤ 3 and ␣ v ␤ 5 ) was not modified upon cell adhesion onto CK2-PVn, the activation of PKB (triggered by ␣ v ␤ 3 but not by ␣ v ␤ 5 ) is elevated upon adhesion to CK2-PVn. It is this elevation that is correlated with the enhanced cell adhesion. Therefore, we propose that the PI3K/PKB pathway (and not the ERK pathway) reflects the ␣ v ␤ 3 -mediated enhanced cell adhesion (Scheme 1). In line with this proposal, we found that blocking the activation of ERK by an MEK inhibitor did not have an effect on cell adhesion. In contrast, the blocking of PKB by PI3K inhibitors reduced cell adhesion.
Taken together, the results presented here together with our results reported earlier (54) indicate the occurrence of a cell surface receptor (␣ v ␤ 3 ) and an intracellular signaling pathway that distinguish between a CK2-phosphorylated and a nonphosphorylated form of Vn. These results are based on a CK2phospho and a non-phospho form of Vn, two mutant analogs of Vn, three different cell lines, and four independent cell clones. We believe that these findings indicate that the extra-cellular phosphorylation of Vn by CK2 may well be a physiological process with a distinct regulatory role in the control of cell adhesion and spreading. FIG. 7. FAK phosphorylation triggered by r-Vns in the stably transfected H1299 cells. Plates were coated with the two r-Vns or with the non-integrin adhesive molecule, poly-D-lysine (PDL). Cells were plated on top of the coated plates for the time indicated, or kept in suspension (Sus). Thereafter, cells were harvested in RIPA buffer and the soluble fractions were collected after centrifugation. Protein samples (600 g) obtained from the cell lysates described above were immunoprecipitated using anti-FAK antibodies immobilized on agarose beads as described under "Experimental Procedures." The samples containing the immunoprecipitated FAK were boiled in Laemmli's sample buffer and subjected on SDS-PAGE. The gels were transferred to nitrocellulose membranes and blotted with antibodies against phosphotyrosine (PY99) or with antibodies against FAK. Extracts obtained from the vector-transfected cells (A), or from the ␤ 3 -transfected cells (B). The FAK phosphorylation was quantitated by densitometry. The data illustrated represent the average of three separate experiments (C; open symbols are for cells transfected with the vector, and filled symbols are for cells transfected with ␤ 3 ).