- Email: [email protected]
The PKA Phosphorylation of Vitronectin: Effect on Conformation and Function
Archives of Biochemistry and Biophysics Vol. 397, No. 2, January 15, pp. 246 –252, 2002 doi:10.1006/abbi.2001.2699, available online at http://www.idealibrary.com on
The PKA Phosphorylation of Vitronectin: Effect on Conformation and Function 1 Iris Schvartz, Tamar Kreizman, Vlad Brumfeld, Zeev Gechtman, Dalia Seger, and Shmuel Shaltiel 2 Department of Biological Regulation, The Weizmann Institute of Science, Rehovot 76100, Israel
Received October 25, 2001, and in revised form November 5, 2001
Vitronectin (Vn) stabilizes the inhibitory form of plasminogen activator inhibitor-1 (PAI-1), an important modulator of fibrinolysis. We have previously reported that Vn is specifically phosphorylated by PKA (at Ser378), a kinase we have shown to be released from platelets upon their physiological activation. Here we describe the molecular consequences of this phosphorylation and show (by circular dichroism, and by phosphorylation with casein kinase II) that it acts by modulating the conformation of Vn. The PKA phosphorylation of Vn is enhanced in the presence of either PAI-1, or heparin, or both. This enhanced phosphorylation occurs exclusively on Ser378 as shown with the Vn mutants Ser378Ala and Ser378Glu. The binding of PKA phosphorylated Vn to immobilized PAI-1 and to immobilized plasminogen is shown to be lower than that of Vn. The evidence compiled here suggests that this phosphorylation of Vn can modulate plasminogen activation and consequently control fibrinolysis. © 2002 Elsevier Science Key Words: vitronectin; PAI-1; PKA; phosphorylation; fibrinolysis; glycosaminoglycans.
The cAMP-dependent protein kinase (PKA) was originally considered to be an intracellular enzyme involved in the implementation of hormonal stimuli within the cell (1–3). However, more recent studies indicate that, in addition to its well established localization and function inside the cell, PKA may also be found at the external surface of the cell membranes as an ectoenzyme (4 –7) and has specific extracellular substrates (4, 8). Specifically, we have shown that upon 1
This paper is dedicated to Earl R. Stadtman a distinguished scientist, a superb mentor, and a devoted friend, with best wishes for many more creative years in good health and happiness. 2 To whom correspondence and reprint requests should be addressed. Fax: ⫹972-8-934 2804. E-mail: [email protected]. 246
physiological stimulation of blood platelets with thrombin, the platelets selectively release PKA that specifically phosphorylates one plasma protein, vitronectin (Vn) (4, 8). Vn is present in the circulation as well as in the extracellular matrix (9) and was shown to have a variety of biological activities, including the promotion of cell adhesion and spreading through an Arg-Gly-Asp motif (10, 11), as well as the inhibition of the lytic activities of the complement (12). Vn also binds plasminogen (13, 14) and PAI-1 (15–17). The binding of Vn to PAI-1 stabilizes the inhibitor in its active (inhibitory) conformation (15–17) and, as such, is presumed to be involved in the regulation of plasminogen activation that is implicated in various physiological and pathological processes such as cell migration, fibrinolysis, tissue remodeling, inflammation, and tumor metastasis (18 –21). In circulating blood, Vn is present in two molecular forms: a single chain form with an apparent molecular weight of 75 kDa (Vn 75), and a clipped form composed of two chains (65 and 10 kDa), which are held together by a disulfide bond (Vn 65⫹10) (12, 22). We found that Ser378 is the only target in Vn for its phosphorylation by PKA (4, 8), and that this phosphorylation occurs mainly in the one-chain form of Vn (23). However, the two-chain form of Vn is also phosphorylated if a glycosaminoglycan, such as heparin, is present in the reaction mixture. On the basis of this finding we concluded that the Ser378 in the clipped form of Vn is buried and that upon its binding to the glycosaminoglycan it becomes exposed to phosphorylation by PKA (23). The site of PKA phosphorylation in Vn is located at the carboxy terminal portion of Vn, adjacent to the binding sites of heparin, PAI-1 and plasminogen (14, 24). Thus, it is distal from the Arg-Gly-Asp site localized in the amino-terminus portion of Vn. More recently we showed that Vn can also undergo a phosphorylation by CKII (at Thr50 and Thr57), and by 0003-9861/02 $35.00 © 2002 Elsevier Science All rights reserved.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
PKC (mainly at Ser 361). We also showed that these phosphorylations are associated (respectively) with cell adhesion and spreading (25, 26), and with plasminogen activation (27). In this study we assess the structural and functional consequences of the Vn phosphorylation by PKA, and propose a mechanism through which this phosphorylation can modulate, and possibly control, fibrinolysis, e.g., at the site of a hemostatic event. MATERIALS AND METHODS
Materials Platelet rich plasma and freshly frozen human plasma were obtained from Tel-Hashomer Medical Center (Ramat Gan, Israel). Plasminogen, ATP, PKA inhibitor 6 –22 amide (PKI), heparin, thrombin, PGE 1, soybean trypsin inhibitor, and aprotinin were purchased from Sigma. PAI-1 was obtained from American Diagnostica Inc. (New York). Rabbit anti-human vitronectin antiserum was purchased from Calbiochem. CKII was purchased from Boehringer Mannheim. [␥- 32P]ATP (3000 Ci/mmol) and ECL reagents were obtained from Amersham. Restriction enzymes were purchased from Boehringer Mannheim or Gibco BRL. Taq polymerase was purchased from Promega, and the Baculovirus linear DNA from Clontech.
Methods Purification of vitronectin. Vn was prepared from freshly frozen human plasma using a purification procedure described earlier (28). Where indicated, Vn or PKA phosphorylated Vn (PVn) were purified on a Mono Q column (HR 5/5, 1 ml, Pharmacia, Sweden) connected to HPLC (Hewlett Packard 1050) with a UV detection system at 280 nm as previously described (24). Vn phosphorylation by PKA. The phosphorylation of Vn by the pure catalytic subunit of PKA (29) was assayed in a reaction mixture (final volume 50 l) as previously described (4). The phosphorylation was allowed to proceed for 10 min at 22°C, and arrested by the addition of Laemmli’s sample buffer (30) and boiling for 3 min. The samples were then subjected to 10% acrylamide gel and autoradiography. In order to obtain PVn, the phosphorylation reaction was conducted as described above except that a higher concentration of the ingredients was used as follows: Vn (2–5 M), the catalytic subunit of PKA (30 g/ml), heparin (500 g/ml), and ATP (30 M). The reaction was allowed to proceed for 30 min at 22°C. Binding of Vn to immobilized plasminogen. Plasminogen was diluted in a coating buffer (15 mM Na 2CO 3, 35 mM NaHCO 3, pH 9.5) and added (2 g/well) to a microtiter plate for 24 h at 4°C. Plasminogen was then aspirated and skim milk was added (30 min at 37°C) to block the nonspecific binding sites. Vn or PVn were then diluted in a PBS solution containing 0.05% Tween 20 and added at increasing concentrations for 2 h at 22°C. The binding of Vn was detected as previously described (25). Circular dichroism (CD) spectra. CD spectra were run on a JASCO 500 spectropolarimeter at 25°C. Thirty microliters of either Vn or PVn (250 g/ml) were added to a 0.1 mm demountable cuvette and the spectrum between 190 and 270 nm was recorded. Spectra shown are an average of four scans. Secondary structural estimations were done by the method of Chang et al. using our computer program (31–33). The estimated error is around 3% for helix and sheets and 6% for unordered and turn structures. Construction of S378A and S378E mutants. Fragments carrying the mutations were generated by PCR amplification using Vn cDNA (in pGEX 2T) as templates. The sense oligonucleotide designed for
247
generating the S 378A mutant was 5⬘ TCC CGC CGG CCA GCC CGC GCC 3⬘ and the sense oligonucleotide designed for generating the S 378E mutant was: 5⬘ TCC CGC CGG CCA GAA CGC GCC 3⬘. The antisense oligonucleotide designed from the vector-pGEX 2T was 5⬘ CGT CAG TCA GTC ACG ATG AAT TC 3⬘. Purified PCR amplified fragments digested with NaeI and EcoRI were ligated into VnpGEX-2T digested with NaeI and EcoRI. After confirming the sequence of the constructs they were subcloned into the PVL 1393 transfer vector (BamHI and EcoRI sites). Purification and activation of platelets. Platelet rich plasma (20 ml) was centrifuged in the presence of 1 mM EDTA and 1 M PGE 1 at 600g for 15 min at 22°C. The platelet pellet was resuspended in 20 ml of 10 mM HEPES, pH 6.8, containing 137 mM NaCl, 4 mM KCl, 12 mM NaHCO 3, 1 mM MgCl 2, 0.1% glucose, 1 mM EDTA, 1 M PGE 1, and 0.1% BSA. The platelets were centrifuged as above, resuspended in 2 ml of 10 mM Hepes, pH 7.5, containing 137 mM NaCl, 4 mM KCl, 12 mM NaHCO 3, 10 mM MgCl 2, 0.1% glucose, and incubated (200 l) with [␥- 32P]ATP (5 Ci, 3000 Ci/mmol) in the absence or presence of either thrombin (5 u/ml), PKI (20 M), or both. The phosphorylation reaction was allowed to proceed for 150 min at 37°C, and arrested by the addition of 1 mM ATP and a proteinase inhibitor mixture containing 1 mM PMSF, 50 g/ml soybean trypsin inhibitor, 2 mM EDTA, and 80 u/ml aprotinin. The platelets were then centrifuged as above, and their supernatants were prepared for SDS–PAGE. Procedures carried out as previously described in the literature. Protein concentration was determined by the method of Bradford (34). SDS–PAGE was carried out according to the procedure described by Laemmli (30). Gels assayed for radioactively labeled proteins were dried under vacuum and autoradiographed at ⫺80°C using Agfa X-ray film. In the cases indicated, the bands of such gels were cut out and their radioactivity was determined in a scintillation counter. Alternatively, the autoradiogram was scanned by a densitometer (420 oe, PDI). Western blot analysis of Vn was carried out as described previously (24), except that the detection was done by the ECL reagents. Guanidine HCl or SDS activation of PAI-1 was carried out as described by Stockman et al. (35), and by Hekman and Loskutoff (36). The phosphorylation of Vn by CKII, the adsorption of Vn onto microtiter plates, and the expression of recombinant wildtype Vn in insect cells were conducted as described elsewhere (25). The binding of Vn to immobilized activated PAI-1 was determined as described earlier (37).
RESULTS
Stoichiometric Phosphorylation of Vn by PKA To elucidate the mechanism by which the PKA phosphorylation of Vn at Ser378 affects its structure and function, we needed a stoichiometrically phosphorylated Vn (all molecules in the phospho form), and a nonphosphorylated Vn (all molecules in the nonphospho form). Such Vn preparations could be prepared under the conditions of the experiment depicted in Fig. 1. Phosphorylation of Vn by the catalytic subunit of PKA in the presence of [␥- 32P]ATP and heparin results in the labeling of both the Vn 75 and Vn 65 forms of this protein (Fig. 1, inset). The incorporation of phosphate reaches a plateau at ⬃1 mol of phosphate in Vn 75, and ⬃0.75 mol of phosphate in Vn 65 (Fig. 1, filled and empty circles, respectively). Indeed, when Vn was prephosphorylated under these conditions with non-radioactive ATP (to obtain non radioactive phosphorylated Vn (PVn)), then subjected to phosphorylation with
248
SCHVARTZ ET AL.
PKA Phosphorylation of Ser378 in Vn Affects the Accessibility of the N-Terminus of Vn to Phosphorylation by CKII
FIG. 1. Stochiometry of the PKA phosphorylation of Vn. Vn was phosphorylated by the C-subunit of PKA in the presence of [␥- 32 P]ATP for various periods of time. The reaction was arrested by the addition of Laemmli sample buffer and boiling. Subsequently, the samples were subjected to SDS–PAGE (10% acrylamide gel). The gel was dried, subjected to autoradiography, and the Vn 75 and Vn 65 bands were excised from the gel and counted in a -scintillation counter. The amount of phosphate incorporated into each band was calculated according to the specific activity of [␥- 32 P]ATP. Inset, Biochemical and immunological characterization of Vn and PVn. Vn and PVn were assayed for PKA radiolabeling (a and b, respectively) and for immunoreactivity with anti-Vn antibodies (c and d, respectively) as described under Materials and Methods. Please note that in lane b and d, Vn was fully phosphorylated with nonradioactive ATP prior to its phosphorylation with [␥- 32 P]ATP.
[␥- 32P]ATP there was no additional labeling of either PVn 75 or PVn 65 (Fig. 1, compare lane a with b), indicating that under these experimental conditions the PKA phosphorylation of Vn is complete. Conformational Change upon PKA Phosphorylation of Vn—Evidence from Circular Dichroism To establish whether the PKA phosphorylation brings about a conformational change in Vn, the circular dichroism (CD) spectra of Vn was compared with that of PVn. As seen in Fig. 2, upon phosphorylation of Vn there is a significant decrease of the molar ellipticity of PVn, especially at short wavelengths. The estimated secondary structure of Vn obtained (Fig. 2, inset) is in a good agreement with that estimated by the amino acid sequence of Vn (38), though not as much with that obtained from the CD spectrum performed by Pitt et al. (39). The reason for this difference is not clear. It may be due, however, to a difference in the method of Vn purification. In any event, since the purification of Vn and PVn was similar in the present study, these results suggest that the PKA phosphorylation alters the conformation of Vn by reducing its ␣ helix content and increasing the content of its unordered structure (Fig. 2, inset).
Further support for the conformational change that accompanies the PKA phosphorylation of Vn was obtained by studying its accessibility to phosphorylation by CKII. We have previously shown that Vn undergoes a CKII phosphorylation at two sites (Thr50 and Thr57), and that this phosphorylation increases the Vn cell adhesion capacity (25). Interestingly, the rate of CKII phosphorylation is reduced to half if Vn is prephosphorylated by PKA (Fig. 3), in spite of the fact that the phosphorylation of Vn by PKA occurs at a distal site. Effect of PAI-1, Heparin, and Plasminogen on the PKA Phosphorylation of Vn Work done in several laboratories has shown that Vn is the major binding protein of PAI-1 in the circulation and in the extracellular matrix (15–17). It was also shown that upon thrombin stimulation, both Vn and PAI-1 are released from the ␣ granules of human platelets (7, 40). Previously we showed that thrombin stimulation of human platelets brings about a release of PKA (4). We, therefore, attempted to find out whether the PKA phosphorylation of Vn could take place when Vn is bound to PAI-1. For this purpose, Vn was preincubated with guanidine-activated PAI-1, and thereafter a PKA phosphorylation with [␥- 32P]ATP was conducted in the presence or in the absence of heparin. It was found that a preincubation of Vn with PAI-1 enhances the PKA phosphorylation of Vn (Fig. 4). A similar enhancement was observed when Vn was phosphorylated in the presence of heparin (Fig. 4 and (23)). However, when Vn was preincubated with PAI-1 and
FIG. 2. Circular Dichroism spectra of Vn and PVn. The Circular Dichroism spectra of Vn (circles) and PVn (squares) were run on a JASCO 500 spectropolarimeter at 25°C and recorded between 190 and 270 nm as described under Materials and Methods.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
249
FIG. 5. Effect of PAI-1 on PKA phosphorylation of human plasma Vn or of recombinant Vns. Human plasma Vn or recombinant Vns (0.07 M) were phosphorylated by the C-subunit of PKA in the presence of [␥- 32P]ATP as described under Materials and Methods. The samples were applied on 10% acrylamide gel, the gel was dried and subjected to autoradiography.
FIG. 3. The CKII phosphorylation of Vn and PVn. Vn (empty circles) or PVn (full circles) (1 M) were phosphorylated by CKII (2 g/ml) in the presence of [␥- 32P]ATP (10 M; (10 Ci/mmol) in 30 mM Hepes, pH 7.5. Following various periods of time, aliquots (40 l) were removed from the reaction mixture, and prepared for SDS– PAGE as described under Materials and Methods. The gel was dried and subjected to autoradiography (inset). Vn bands were excised from the gel and counted in a -scintillation counter.
the PKA phosphorylation was conducted in the presence of heparin, the Vn phosphorylation was more pronounced than that obtained when Vn was phosphorylated in the presence of each of the proteins alone (Fig. 4). Thus, although one of the PAI-1 binding elements is located at the carboxyl terminal edge of Vn (segment 348 –370) (14, 37) and is adjacent to Ser378, these results suggest that the binding of PAI-1 to Vn exposes either Ser378 or another site that is otherwise not
accessible to the PKA phosphorylation of Vn. Since Vn has also a plasminogen binding domain adjacent to the PAI-1 binding site (14), we examined whether the Vn phosphorylation can occur in the presence of plasminogen. When Vn was preincubated with plasminogen and then a phosphorylation reaction was performed, the phosphorylation of Vn was significantly reduced (Fig. 4), and it was totally impeded in the presence of both plasminogen and heparin (Fig. 4). The increase or decrease of PKA phosphorylation of Vn following PAI-1 or plasminogen binding was not due to a difference in protein concentration since immunoblotting of the transferred proteins with rabbit anti-human Vn antibodies resulted in a similar Vn labeling (data not shown). Therefore, these results demonstrate that while the binding of plasminogen interferes with the PKA phosphorylation of Vn, the binding of PAI-1 contributes to this phosphorylation.
The Enhanced PKA Phosphorylation of Vn That Occurs upon Binding of PAI-1 Is Due to an Increased Exposure of Ser378, not to a Phosphorylation of Another Site in Vn
FIG. 4. The effect of PAI-1, plasminogen and heparin on PKA phosphorylation of Vn. Vn (0.2 M) was preincubated in the absence or presence of either guanidine activated PAI-1 (0.2 M) or plasminogen (0.2 M) in 50 mM Hepes, pH 7.5, for 2 h at 22°C. Thereafter, phosphorylation of Vn by the C-subunit of PKA in the presence of [␥- 32P]ATP and in the absence or presence of heparin was conducted as described under Materials and Methods. The samples were subjected to SDS–PAGE (10%) and autoradiography. Please note that the different phosphate content in the Vn 75 and Vn 65 bands reflects the different ratios of the one-chain and two-chain forms of Vn found in different individuals.
To find out whether upon binding of PAI-1 to Vn the Ser378 residue in Vn becomes exposed or whether this binding unveils another phosphorylation site that is “buried” in the absence of PAI-1, we prepared two single site mutants of Vn: One in which Ser378 was replaced by Alanine (an analog representing the nonphosphorylated Vn), and one in which Ser378 was replaced by Glutamic acid residue (an analog representing the PKA phosphorylated Vn). As seen in Fig. 5, these S378A and S378E mutants did not undergo any phosphorylation by PKA, even in the presence of PAI-1, clearly suggesting that there is no other site in Vn except Ser378 that becomes phosphorylated by PKA.
250
SCHVARTZ ET AL.
lower in the presence of heparin (Fig. 6a), which can be regarded as a representative glycosaminoglycan. To exclude the possibility that the difference observed in the binding of PAI-1 is due to a difference in the binding of Vn vs. PVn to the microtiter plates, the direct adsorption of Vn or PVn to such plates was measured. It was found that both Vn and PVn are essentially identical in their absorption to the plastic plates (Fig. 6b). Since Vn was also shown to contain a plasminogen binding domain (14), we examined the effect of the PKA phosphorylation of Vn on its binding to plasminogen. As seen in Fig. 6c, PVn showed a lower binding to plasminogen in comparison to that observed by Vn. Addition of heparin to the binding reaction did not affect plasminogen binding of either Vn or PVn (Fig. 6c). Taken together, these results indicate that the PKA phosphorylation of Vn reduces its binding to immobilized PAI-1 as well as to immobilized plasminogen. In a previous study we showed that plasma Vn is phosphorylated by PKA released from thrombin stimulated platelets (8). Recently we showed that plasma Vn can also be phosphorylated by an ecto-PKA present on the membrane of platelets and other cells (Belleli and Shaltiel, unpublished results). The question is whether the Vn, released from thrombin stimulated platelets, can be phosphorylated by the PKA that is also released from thrombin stimulated platelets. To test this possibility, washed human platelets were stimulated with thrombin in the presence of [␥- 32P]ATP, and the phosphorylation of “their” Vn was monitored. Indeed, stimulation of platelets with thrombin resulted in a phosphorylation of Vn 75. DensiFIG. 6. The binding of Vn or PVn to immobilized PAI-1 or immobilized plasminogen. (a) SDS activated PAI-1 (2 g/ml; 50 l/well in PBS) was added to 96-well plates as described under Materials and Methods. PAI-1 was then aspirated and Vn or PVn, diluted in PBS containing 0.05% Tween, were added at increasing concentrations in the absence or presence of heparin (50 g/ml) for additional 2 h. To determine bound Vn or PVn, rabbit anti-human Vn antibodies were added (1:1000 dilution), followed by the addition of goat anti-rabbit antibodies-HRP conjugated as described under Materials and Methods. (b) Adsorption to the polystyrene plate. Purified Vn or PVn were added in PBS to 96-well plates at increasing concentrations. Determination of Vn or PVn adsorbed to the plate was conducted as described in the legend to (a). (c) Plasminogen was added (2 g/well) to 96-well plates as described under Materials and Methods. Vn or PVn were added at increasing concentrations in the absence or presence of heparin (50 g/ml). The results are presented as the mean ⫾ SE of three determinations and are representative of three to seven similar experiments.
Does the Phosphorylation of Vn Affect Its Binding to PAI-1 and to Plasminogen? When Vn or PVn are added to immobilized PAI-1, the binding of PVn is significantly lower than that of Vn (Fig. 6a). This reduced binding of PVn was even
FIG. 7. PKA phosphorylation of Vn released from thrombin stimulated platelets. Washed platelets were incubated with [␥- 32P]ATP in the absence or presence of either thrombin (5 u/ml), PKI (20 M), or both for 150 min at 37°C. At the end of the phosphorylation reaction the platelet supernatants were prepared for SDS–PAGE as described under Materials and Methods. The gel was dried and exposed to autoradiography, and the Vn 75 band was excised from the gel and counted in a -scintillation counter.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
tometric analysis of the phosphorylated Vn indicated that addition of thrombin to human platelets increased the Vn 75 phosphorylation by fourfold as compared to the Vn phosphorylation in intact platelets (Fig. 7). Addition of the specific PKA inhibitor, PKI, did not affect the Vn 75 phosphorylation when added to intact platelets (not shown) but reduced the Vn phosphorylation by 50% when added to thrombin-stimulated platelets (Fig. 7). DISCUSSION
The major messages of this paper can be summarized as follows: The phosphorylation of Vn by PKA is stoichiometric and occurs exclusively at Ser378. This phosphorylation affects the conformation of Vn as indicated by circular dichroism (conversion of ⬃7% of the helix structure into an unordered structure), and by the availability of Vn to phosphorylation by CKII. The PKA phosphorylation of Vn is promoted by PAI-1 and by heparin, suggesting that it is targeted preferentially to Vn molecules that are bound to glycosaminoglycans and complexed with PAI-1. Several studies demonstrated the consequences of the VnPAI-1 interaction in relation to the conformation of PAI-1 (20, 21). It was shown, for example, that the binding of Vn stabilizes PAI-1 in its active conformation and the half-life of the inhibitor is increased by two- to fourfold (15). However, the full consequences of the Vn-PAI-1 binding in relation to the Vn structure have not been elucidated yet. It was demonstrated that addition of PAI-1 to nonfractionated plasma or to purified monomeric Vn resulted in the formation of Vn multimers (41). Such conversion of Vn monomers to multimers was shown to be accompanied by changes in Vn function including an increase of PAI-1 binding (42) and of cell binding (43). In the present work we provide evidence showing that following binding to PAI-1, the conformation of Vn is modified in a way that exposes Ser378 to PKA phosphorylation (Figs. 4 and 5). The binding of PVn to immobilized PAI-1 is shown to be lower than the binding of Vn to this inhibitor (Fig. 6a). Heparin impedes the binding of either Vn or PVn to immobilized PAI-1. These findings raise the possibility that the physiological role of the PKA phosphorylation of Vn is to convert the active (inhibitory) PAI-1 (bound to Vn) into its latent, noninhibitory form. This conversion will allow plasminogen activation, and consequently initiate fibrinolysis and blood clot dissolution. The binding of PVn to immobilized plasminogen (Fig. 6c) is also lower than the binding of Vn to this proenzyme, but there is no additional heparin effect in this case. The lowered binding of PVn may indicate that the phosphorylation of Vn favors its detachment from plas-
251
minogen and the consequent conversion of plasminogen to active plasmin, again promoting fibrinolysis. Miles and Plow (44) demonstrated that plasminogen receptors are expressed by a variety of cells, including platelets and endothelial cells. They suggested that the receptors may serve to store plasminogen within thrombi or at sites of injury and inflammation (44). Since platelets contain plasminogen in their granules, and secrete it upon their stimulation (45), the presence of Vn in such sites may probably secure that plasminogen is not converted to plasmin prematurely. On the other hand, the phosphorylation of Vn by PKA may actually serve to initiate fibrinolysis. In line with this suggestion we show that the PKA, released from thrombin-stimulated platelets, can phosphorylate the Vn that is released from such platelets (Fig. 7). In conclusion, the PKA phosphorylation of Vn at Ser378 alters the conformation of Vn and converts it from an antifibrinolytic agent that secures that no fibrinolysis occurs as long as it is not desirable, into a profibrinolytic factor that initiates the solubilization of blood clots and the dissolution of the debris that are not needed any more. The function of this mechanism as a physiological regulatory device has still to be proven in vivo. ACKNOWLEDGMENTS This research was supported by the Israel Science Foundation, Award No. 415/99. S.S. is the Incumbent of the Kleeman Chair in Biochemistry at the Weizmann Institute of Science.
REFERENCES 1. Krebs, E. G. (1987) The Enzymes, Vol. XVII, Academic Press, New York. 2. Smith, C. M., Shindyalov, I. N., Veretnik, S., Gribskov, M., Taylor, S. S., Ten Eyck, L. F., and Bourne, P. E. (1997) Trends Biochem. Sci. 22, 444 – 446. 3. Taylor, S. S., Radzio-Andzelm, E., Madhusudan, Cheng, X., Ten Eyck, L., and Narayana, N. (1999) Pharmacol. Ther. 82, 133– 141. 4. Korc-Grodzicki, B., Tauber-Finkelstein, M., and Shaltiel, S. (1988) Proc. Natl. Acad. Sci. USA 85, 7541–7545. 5. Kubler, D., Pyerin, W., Bill, O., Hotz, A., Sonka, J., and Kinzel, V. (1989) J. Biol. Chem. 264, 14549 –14555. 6. Kubler, D., Reinhardt, D., Reed, J., Pyerin, W., and Kinzel, V. (1992) Eur. J. Biochem. 206, 179 –186. 7. Morgenstern, E., Gnad, U., Preissner, K. T., Dierichs, R., Belleli, A., Chestukhin, A., Schvartz, I., and Shaltiel, S. (2001) Eur. J. Cell Biol. 80, 87–98. 8. Korc-Grodzicki, B., Tauber-Finkelstein, M., Chain, D., and Shaltiel, S. (1988) Biochem. Biophys. Res. Commun. 157, 1131–1138. 9. Preissner, K. T. (1991) Annu. Rev. Cell Biol. 7, 275–310. 10. Suzuki, S., Oldberg, A., Hayman, E. G., Pierschbacher, M. D., and Ruoslahti, E. (1985) EMBO J. 4, 2519 –2524. 11. Hynes, R. O. (1992) Cell 69, 11–25. 12. Dahlback, B., and Podack, E. R. (1985) Biochemistry 24, 2368 – 2374.
252
SCHVARTZ ET AL.
13. Preissner, K. T. (1990) Biochem. Biophys. Res. Commun. 168, 966 –971. 14. Kost, C., Stuber, W., Ehrlich, H. J., Pannekoek, H., and Preissner, K. T. (1992) J. Biol. Chem. 267, 12098 –12105. 15. Declerck, P. J., De, M. M., Alessi, M. C., Baudner, S., Paques, E. P., Preissner, K. T., Muller, B. G., and Collen, D. (1988) J. Biol. Chem. 263, 15454 –15461. 16. Mimuro, J., and Loskutoff, D. J. (1989) J. Biol. Chem. 264, 5058 –5063. 17. Seiffert, D., and Loskutoff, D. J. (1991) J. Biol. Chem. 266, 2824 –2830. 18. Dano, K., Andreasen, P. A., Grondahl, H. J., Kristensen, P., Nielsen, L. S., and Skriver, L. (1985) Adv. Cancer Res. 44, 139 – 226. 19. Pollanen, J., Stephens, R. W., and Vaheri, A. (1991) Adv. Cancer Res. 57, 273–328. 20. Podor, T. J., Shaughnessy, S. G., Blackburn, M. N., and Peterson, C. B. (2000) J. Biol. Chem. 275, 25402–25410. 21. Podor, T. J., Peterson, C. B., Lawrence, D. A., Stefansson, S., Shaughnessy, S. G., Foulon, D. M., Butcher, M., and Weitz, J. I. (2000) J. Biol. Chem. 275, 19788 –19794. 22. Barnes, D. W., Silnutzer, J., See, C., and Shaffer, M. (1983) Proc. Natl. Acad. Sci. USA 80, 1362–1366. 23. Chain, D., Korc-Grodzicki, B., Kreizman, T., and Shaltiel, S. (1990) FEBS Lett. 269, 221–225. 24. Gechtman, Z., Belleli, A., Lechpammer, S., and Shaltiel, S. (1997) Biochem. J. 325, 339 –349. 25. Seger, D., Gechtman, Z., and Shaltiel, S. (1998) J. Biol. Chem. 273, 24805–24813. 26. Seger, D., Seger, R., and Shaltiel, S. (2001) J. Biol. Chem. 276, 16998 –17006. 27. Gechtman, Z., and Shaltiel, S. (1997) Eur. J. Biochem. 243, 493–501.
28. Korc-Grodzicki, B., Chain, D., Kreizman, T., and Shaltiel, S. (1990) Anal. Biochem. 188, 288 –294. 29. Chestukhin, A., Litovchick, L., Batkin, M., and Shaltiel, S. (1996) FEBS Lett. 382, 265–270. 30. Laemmli, U. K. (1970) Nature 227, 680 – 685. 31. Chang, C. T., Wu, C. S-C., and Yang, J. T. (1978) Anal. Biochem. 91, 13–31. 32. Brumfeld, V., and Miller, I. R. (1988) Biophys. J. 54, 747–750. 33. Brumfeld, V., and Werber, M. (1993) Arch. Biochem. Biophys. 302, 134 –143. 34. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254. 35. Stockmann, A., Hess, S., Declerk, P., Timpl, R., and Preissner, K. T. (1993) J. Biol. Chem. 268, 22874 –22882. 36. Hekman, C. M., and Loskutoff, D. J. (1985) J. Biol. Chem. 260, 11581–11587. 37. Gechtman, Z., Sharma, R., Kreizman, T., Fridkin, M., and Shaltiel, S. (1993) FEBS Lett. 315, 293–297. 38. Preissner, K. T., and Muller-Berghaus, G. (1987) J. Biol. Chem. 262, 12247–12253. 39. Pitt, W. G., Fabrizius-Homan, D. J., Mosher, D. F., and Cooper, S. L. (1989) J. Colloid Interface Sci. 129, 231–239. 40. Preissner, K. T., Holzhuter, S., Justus, C., and Muller-Berghaus, G. (1989) Blood 74, 1989 –1996. 41. Seiffert, D., and Loskutoff, D. J. (1996) J. Biol. Chem. 271, 29644 –29651. 42. Naski, M. C., Lawrence, D. A., Mosher, D. F., Podor, T. J., and Ginsburg, D. (1993) J. Biol. Chem. 268, 12367–12372. 43. Zanetti, A., Conforti, G., Hess, S., Martin, P. I., Ghibaudi, E., Preissner, K. T., and Dejana, E. (1994) Blood 84, 1116 – 1123. 44. Miles, L. A., and Plow, E. F. (1988) Fibrinolysis 2, 61–71. 45. Holt, J. C., and Niewiarowski, S. (1985) Semin. Hematol. 22, 151–163.
The PKA Phosphorylation of Vitronectin: Effect on Conformation and Function 1 Iris Schvartz, Tamar Kreizman, Vlad Brumfeld, Zeev Gechtman, Dalia Seger, and Shmuel Shaltiel 2 Department of Biological Regulation, The Weizmann Institute of Science, Rehovot 76100, Israel
Received October 25, 2001, and in revised form November 5, 2001
Vitronectin (Vn) stabilizes the inhibitory form of plasminogen activator inhibitor-1 (PAI-1), an important modulator of fibrinolysis. We have previously reported that Vn is specifically phosphorylated by PKA (at Ser378), a kinase we have shown to be released from platelets upon their physiological activation. Here we describe the molecular consequences of this phosphorylation and show (by circular dichroism, and by phosphorylation with casein kinase II) that it acts by modulating the conformation of Vn. The PKA phosphorylation of Vn is enhanced in the presence of either PAI-1, or heparin, or both. This enhanced phosphorylation occurs exclusively on Ser378 as shown with the Vn mutants Ser378Ala and Ser378Glu. The binding of PKA phosphorylated Vn to immobilized PAI-1 and to immobilized plasminogen is shown to be lower than that of Vn. The evidence compiled here suggests that this phosphorylation of Vn can modulate plasminogen activation and consequently control fibrinolysis. © 2002 Elsevier Science Key Words: vitronectin; PAI-1; PKA; phosphorylation; fibrinolysis; glycosaminoglycans.
The cAMP-dependent protein kinase (PKA) was originally considered to be an intracellular enzyme involved in the implementation of hormonal stimuli within the cell (1–3). However, more recent studies indicate that, in addition to its well established localization and function inside the cell, PKA may also be found at the external surface of the cell membranes as an ectoenzyme (4 –7) and has specific extracellular substrates (4, 8). Specifically, we have shown that upon 1
This paper is dedicated to Earl R. Stadtman a distinguished scientist, a superb mentor, and a devoted friend, with best wishes for many more creative years in good health and happiness. 2 To whom correspondence and reprint requests should be addressed. Fax: ⫹972-8-934 2804. E-mail: [email protected]. 246
physiological stimulation of blood platelets with thrombin, the platelets selectively release PKA that specifically phosphorylates one plasma protein, vitronectin (Vn) (4, 8). Vn is present in the circulation as well as in the extracellular matrix (9) and was shown to have a variety of biological activities, including the promotion of cell adhesion and spreading through an Arg-Gly-Asp motif (10, 11), as well as the inhibition of the lytic activities of the complement (12). Vn also binds plasminogen (13, 14) and PAI-1 (15–17). The binding of Vn to PAI-1 stabilizes the inhibitor in its active (inhibitory) conformation (15–17) and, as such, is presumed to be involved in the regulation of plasminogen activation that is implicated in various physiological and pathological processes such as cell migration, fibrinolysis, tissue remodeling, inflammation, and tumor metastasis (18 –21). In circulating blood, Vn is present in two molecular forms: a single chain form with an apparent molecular weight of 75 kDa (Vn 75), and a clipped form composed of two chains (65 and 10 kDa), which are held together by a disulfide bond (Vn 65⫹10) (12, 22). We found that Ser378 is the only target in Vn for its phosphorylation by PKA (4, 8), and that this phosphorylation occurs mainly in the one-chain form of Vn (23). However, the two-chain form of Vn is also phosphorylated if a glycosaminoglycan, such as heparin, is present in the reaction mixture. On the basis of this finding we concluded that the Ser378 in the clipped form of Vn is buried and that upon its binding to the glycosaminoglycan it becomes exposed to phosphorylation by PKA (23). The site of PKA phosphorylation in Vn is located at the carboxy terminal portion of Vn, adjacent to the binding sites of heparin, PAI-1 and plasminogen (14, 24). Thus, it is distal from the Arg-Gly-Asp site localized in the amino-terminus portion of Vn. More recently we showed that Vn can also undergo a phosphorylation by CKII (at Thr50 and Thr57), and by 0003-9861/02 $35.00 © 2002 Elsevier Science All rights reserved.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
PKC (mainly at Ser 361). We also showed that these phosphorylations are associated (respectively) with cell adhesion and spreading (25, 26), and with plasminogen activation (27). In this study we assess the structural and functional consequences of the Vn phosphorylation by PKA, and propose a mechanism through which this phosphorylation can modulate, and possibly control, fibrinolysis, e.g., at the site of a hemostatic event. MATERIALS AND METHODS
Materials Platelet rich plasma and freshly frozen human plasma were obtained from Tel-Hashomer Medical Center (Ramat Gan, Israel). Plasminogen, ATP, PKA inhibitor 6 –22 amide (PKI), heparin, thrombin, PGE 1, soybean trypsin inhibitor, and aprotinin were purchased from Sigma. PAI-1 was obtained from American Diagnostica Inc. (New York). Rabbit anti-human vitronectin antiserum was purchased from Calbiochem. CKII was purchased from Boehringer Mannheim. [␥- 32P]ATP (3000 Ci/mmol) and ECL reagents were obtained from Amersham. Restriction enzymes were purchased from Boehringer Mannheim or Gibco BRL. Taq polymerase was purchased from Promega, and the Baculovirus linear DNA from Clontech.
Methods Purification of vitronectin. Vn was prepared from freshly frozen human plasma using a purification procedure described earlier (28). Where indicated, Vn or PKA phosphorylated Vn (PVn) were purified on a Mono Q column (HR 5/5, 1 ml, Pharmacia, Sweden) connected to HPLC (Hewlett Packard 1050) with a UV detection system at 280 nm as previously described (24). Vn phosphorylation by PKA. The phosphorylation of Vn by the pure catalytic subunit of PKA (29) was assayed in a reaction mixture (final volume 50 l) as previously described (4). The phosphorylation was allowed to proceed for 10 min at 22°C, and arrested by the addition of Laemmli’s sample buffer (30) and boiling for 3 min. The samples were then subjected to 10% acrylamide gel and autoradiography. In order to obtain PVn, the phosphorylation reaction was conducted as described above except that a higher concentration of the ingredients was used as follows: Vn (2–5 M), the catalytic subunit of PKA (30 g/ml), heparin (500 g/ml), and ATP (30 M). The reaction was allowed to proceed for 30 min at 22°C. Binding of Vn to immobilized plasminogen. Plasminogen was diluted in a coating buffer (15 mM Na 2CO 3, 35 mM NaHCO 3, pH 9.5) and added (2 g/well) to a microtiter plate for 24 h at 4°C. Plasminogen was then aspirated and skim milk was added (30 min at 37°C) to block the nonspecific binding sites. Vn or PVn were then diluted in a PBS solution containing 0.05% Tween 20 and added at increasing concentrations for 2 h at 22°C. The binding of Vn was detected as previously described (25). Circular dichroism (CD) spectra. CD spectra were run on a JASCO 500 spectropolarimeter at 25°C. Thirty microliters of either Vn or PVn (250 g/ml) were added to a 0.1 mm demountable cuvette and the spectrum between 190 and 270 nm was recorded. Spectra shown are an average of four scans. Secondary structural estimations were done by the method of Chang et al. using our computer program (31–33). The estimated error is around 3% for helix and sheets and 6% for unordered and turn structures. Construction of S378A and S378E mutants. Fragments carrying the mutations were generated by PCR amplification using Vn cDNA (in pGEX 2T) as templates. The sense oligonucleotide designed for
247
generating the S 378A mutant was 5⬘ TCC CGC CGG CCA GCC CGC GCC 3⬘ and the sense oligonucleotide designed for generating the S 378E mutant was: 5⬘ TCC CGC CGG CCA GAA CGC GCC 3⬘. The antisense oligonucleotide designed from the vector-pGEX 2T was 5⬘ CGT CAG TCA GTC ACG ATG AAT TC 3⬘. Purified PCR amplified fragments digested with NaeI and EcoRI were ligated into VnpGEX-2T digested with NaeI and EcoRI. After confirming the sequence of the constructs they were subcloned into the PVL 1393 transfer vector (BamHI and EcoRI sites). Purification and activation of platelets. Platelet rich plasma (20 ml) was centrifuged in the presence of 1 mM EDTA and 1 M PGE 1 at 600g for 15 min at 22°C. The platelet pellet was resuspended in 20 ml of 10 mM HEPES, pH 6.8, containing 137 mM NaCl, 4 mM KCl, 12 mM NaHCO 3, 1 mM MgCl 2, 0.1% glucose, 1 mM EDTA, 1 M PGE 1, and 0.1% BSA. The platelets were centrifuged as above, resuspended in 2 ml of 10 mM Hepes, pH 7.5, containing 137 mM NaCl, 4 mM KCl, 12 mM NaHCO 3, 10 mM MgCl 2, 0.1% glucose, and incubated (200 l) with [␥- 32P]ATP (5 Ci, 3000 Ci/mmol) in the absence or presence of either thrombin (5 u/ml), PKI (20 M), or both. The phosphorylation reaction was allowed to proceed for 150 min at 37°C, and arrested by the addition of 1 mM ATP and a proteinase inhibitor mixture containing 1 mM PMSF, 50 g/ml soybean trypsin inhibitor, 2 mM EDTA, and 80 u/ml aprotinin. The platelets were then centrifuged as above, and their supernatants were prepared for SDS–PAGE. Procedures carried out as previously described in the literature. Protein concentration was determined by the method of Bradford (34). SDS–PAGE was carried out according to the procedure described by Laemmli (30). Gels assayed for radioactively labeled proteins were dried under vacuum and autoradiographed at ⫺80°C using Agfa X-ray film. In the cases indicated, the bands of such gels were cut out and their radioactivity was determined in a scintillation counter. Alternatively, the autoradiogram was scanned by a densitometer (420 oe, PDI). Western blot analysis of Vn was carried out as described previously (24), except that the detection was done by the ECL reagents. Guanidine HCl or SDS activation of PAI-1 was carried out as described by Stockman et al. (35), and by Hekman and Loskutoff (36). The phosphorylation of Vn by CKII, the adsorption of Vn onto microtiter plates, and the expression of recombinant wildtype Vn in insect cells were conducted as described elsewhere (25). The binding of Vn to immobilized activated PAI-1 was determined as described earlier (37).
RESULTS
Stoichiometric Phosphorylation of Vn by PKA To elucidate the mechanism by which the PKA phosphorylation of Vn at Ser378 affects its structure and function, we needed a stoichiometrically phosphorylated Vn (all molecules in the phospho form), and a nonphosphorylated Vn (all molecules in the nonphospho form). Such Vn preparations could be prepared under the conditions of the experiment depicted in Fig. 1. Phosphorylation of Vn by the catalytic subunit of PKA in the presence of [␥- 32P]ATP and heparin results in the labeling of both the Vn 75 and Vn 65 forms of this protein (Fig. 1, inset). The incorporation of phosphate reaches a plateau at ⬃1 mol of phosphate in Vn 75, and ⬃0.75 mol of phosphate in Vn 65 (Fig. 1, filled and empty circles, respectively). Indeed, when Vn was prephosphorylated under these conditions with non-radioactive ATP (to obtain non radioactive phosphorylated Vn (PVn)), then subjected to phosphorylation with
248
SCHVARTZ ET AL.
PKA Phosphorylation of Ser378 in Vn Affects the Accessibility of the N-Terminus of Vn to Phosphorylation by CKII
FIG. 1. Stochiometry of the PKA phosphorylation of Vn. Vn was phosphorylated by the C-subunit of PKA in the presence of [␥- 32 P]ATP for various periods of time. The reaction was arrested by the addition of Laemmli sample buffer and boiling. Subsequently, the samples were subjected to SDS–PAGE (10% acrylamide gel). The gel was dried, subjected to autoradiography, and the Vn 75 and Vn 65 bands were excised from the gel and counted in a -scintillation counter. The amount of phosphate incorporated into each band was calculated according to the specific activity of [␥- 32 P]ATP. Inset, Biochemical and immunological characterization of Vn and PVn. Vn and PVn were assayed for PKA radiolabeling (a and b, respectively) and for immunoreactivity with anti-Vn antibodies (c and d, respectively) as described under Materials and Methods. Please note that in lane b and d, Vn was fully phosphorylated with nonradioactive ATP prior to its phosphorylation with [␥- 32 P]ATP.
[␥- 32P]ATP there was no additional labeling of either PVn 75 or PVn 65 (Fig. 1, compare lane a with b), indicating that under these experimental conditions the PKA phosphorylation of Vn is complete. Conformational Change upon PKA Phosphorylation of Vn—Evidence from Circular Dichroism To establish whether the PKA phosphorylation brings about a conformational change in Vn, the circular dichroism (CD) spectra of Vn was compared with that of PVn. As seen in Fig. 2, upon phosphorylation of Vn there is a significant decrease of the molar ellipticity of PVn, especially at short wavelengths. The estimated secondary structure of Vn obtained (Fig. 2, inset) is in a good agreement with that estimated by the amino acid sequence of Vn (38), though not as much with that obtained from the CD spectrum performed by Pitt et al. (39). The reason for this difference is not clear. It may be due, however, to a difference in the method of Vn purification. In any event, since the purification of Vn and PVn was similar in the present study, these results suggest that the PKA phosphorylation alters the conformation of Vn by reducing its ␣ helix content and increasing the content of its unordered structure (Fig. 2, inset).
Further support for the conformational change that accompanies the PKA phosphorylation of Vn was obtained by studying its accessibility to phosphorylation by CKII. We have previously shown that Vn undergoes a CKII phosphorylation at two sites (Thr50 and Thr57), and that this phosphorylation increases the Vn cell adhesion capacity (25). Interestingly, the rate of CKII phosphorylation is reduced to half if Vn is prephosphorylated by PKA (Fig. 3), in spite of the fact that the phosphorylation of Vn by PKA occurs at a distal site. Effect of PAI-1, Heparin, and Plasminogen on the PKA Phosphorylation of Vn Work done in several laboratories has shown that Vn is the major binding protein of PAI-1 in the circulation and in the extracellular matrix (15–17). It was also shown that upon thrombin stimulation, both Vn and PAI-1 are released from the ␣ granules of human platelets (7, 40). Previously we showed that thrombin stimulation of human platelets brings about a release of PKA (4). We, therefore, attempted to find out whether the PKA phosphorylation of Vn could take place when Vn is bound to PAI-1. For this purpose, Vn was preincubated with guanidine-activated PAI-1, and thereafter a PKA phosphorylation with [␥- 32P]ATP was conducted in the presence or in the absence of heparin. It was found that a preincubation of Vn with PAI-1 enhances the PKA phosphorylation of Vn (Fig. 4). A similar enhancement was observed when Vn was phosphorylated in the presence of heparin (Fig. 4 and (23)). However, when Vn was preincubated with PAI-1 and
FIG. 2. Circular Dichroism spectra of Vn and PVn. The Circular Dichroism spectra of Vn (circles) and PVn (squares) were run on a JASCO 500 spectropolarimeter at 25°C and recorded between 190 and 270 nm as described under Materials and Methods.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
249
FIG. 5. Effect of PAI-1 on PKA phosphorylation of human plasma Vn or of recombinant Vns. Human plasma Vn or recombinant Vns (0.07 M) were phosphorylated by the C-subunit of PKA in the presence of [␥- 32P]ATP as described under Materials and Methods. The samples were applied on 10% acrylamide gel, the gel was dried and subjected to autoradiography.
FIG. 3. The CKII phosphorylation of Vn and PVn. Vn (empty circles) or PVn (full circles) (1 M) were phosphorylated by CKII (2 g/ml) in the presence of [␥- 32P]ATP (10 M; (10 Ci/mmol) in 30 mM Hepes, pH 7.5. Following various periods of time, aliquots (40 l) were removed from the reaction mixture, and prepared for SDS– PAGE as described under Materials and Methods. The gel was dried and subjected to autoradiography (inset). Vn bands were excised from the gel and counted in a -scintillation counter.
the PKA phosphorylation was conducted in the presence of heparin, the Vn phosphorylation was more pronounced than that obtained when Vn was phosphorylated in the presence of each of the proteins alone (Fig. 4). Thus, although one of the PAI-1 binding elements is located at the carboxyl terminal edge of Vn (segment 348 –370) (14, 37) and is adjacent to Ser378, these results suggest that the binding of PAI-1 to Vn exposes either Ser378 or another site that is otherwise not
accessible to the PKA phosphorylation of Vn. Since Vn has also a plasminogen binding domain adjacent to the PAI-1 binding site (14), we examined whether the Vn phosphorylation can occur in the presence of plasminogen. When Vn was preincubated with plasminogen and then a phosphorylation reaction was performed, the phosphorylation of Vn was significantly reduced (Fig. 4), and it was totally impeded in the presence of both plasminogen and heparin (Fig. 4). The increase or decrease of PKA phosphorylation of Vn following PAI-1 or plasminogen binding was not due to a difference in protein concentration since immunoblotting of the transferred proteins with rabbit anti-human Vn antibodies resulted in a similar Vn labeling (data not shown). Therefore, these results demonstrate that while the binding of plasminogen interferes with the PKA phosphorylation of Vn, the binding of PAI-1 contributes to this phosphorylation.
The Enhanced PKA Phosphorylation of Vn That Occurs upon Binding of PAI-1 Is Due to an Increased Exposure of Ser378, not to a Phosphorylation of Another Site in Vn
FIG. 4. The effect of PAI-1, plasminogen and heparin on PKA phosphorylation of Vn. Vn (0.2 M) was preincubated in the absence or presence of either guanidine activated PAI-1 (0.2 M) or plasminogen (0.2 M) in 50 mM Hepes, pH 7.5, for 2 h at 22°C. Thereafter, phosphorylation of Vn by the C-subunit of PKA in the presence of [␥- 32P]ATP and in the absence or presence of heparin was conducted as described under Materials and Methods. The samples were subjected to SDS–PAGE (10%) and autoradiography. Please note that the different phosphate content in the Vn 75 and Vn 65 bands reflects the different ratios of the one-chain and two-chain forms of Vn found in different individuals.
To find out whether upon binding of PAI-1 to Vn the Ser378 residue in Vn becomes exposed or whether this binding unveils another phosphorylation site that is “buried” in the absence of PAI-1, we prepared two single site mutants of Vn: One in which Ser378 was replaced by Alanine (an analog representing the nonphosphorylated Vn), and one in which Ser378 was replaced by Glutamic acid residue (an analog representing the PKA phosphorylated Vn). As seen in Fig. 5, these S378A and S378E mutants did not undergo any phosphorylation by PKA, even in the presence of PAI-1, clearly suggesting that there is no other site in Vn except Ser378 that becomes phosphorylated by PKA.
250
SCHVARTZ ET AL.
lower in the presence of heparin (Fig. 6a), which can be regarded as a representative glycosaminoglycan. To exclude the possibility that the difference observed in the binding of PAI-1 is due to a difference in the binding of Vn vs. PVn to the microtiter plates, the direct adsorption of Vn or PVn to such plates was measured. It was found that both Vn and PVn are essentially identical in their absorption to the plastic plates (Fig. 6b). Since Vn was also shown to contain a plasminogen binding domain (14), we examined the effect of the PKA phosphorylation of Vn on its binding to plasminogen. As seen in Fig. 6c, PVn showed a lower binding to plasminogen in comparison to that observed by Vn. Addition of heparin to the binding reaction did not affect plasminogen binding of either Vn or PVn (Fig. 6c). Taken together, these results indicate that the PKA phosphorylation of Vn reduces its binding to immobilized PAI-1 as well as to immobilized plasminogen. In a previous study we showed that plasma Vn is phosphorylated by PKA released from thrombin stimulated platelets (8). Recently we showed that plasma Vn can also be phosphorylated by an ecto-PKA present on the membrane of platelets and other cells (Belleli and Shaltiel, unpublished results). The question is whether the Vn, released from thrombin stimulated platelets, can be phosphorylated by the PKA that is also released from thrombin stimulated platelets. To test this possibility, washed human platelets were stimulated with thrombin in the presence of [␥- 32P]ATP, and the phosphorylation of “their” Vn was monitored. Indeed, stimulation of platelets with thrombin resulted in a phosphorylation of Vn 75. DensiFIG. 6. The binding of Vn or PVn to immobilized PAI-1 or immobilized plasminogen. (a) SDS activated PAI-1 (2 g/ml; 50 l/well in PBS) was added to 96-well plates as described under Materials and Methods. PAI-1 was then aspirated and Vn or PVn, diluted in PBS containing 0.05% Tween, were added at increasing concentrations in the absence or presence of heparin (50 g/ml) for additional 2 h. To determine bound Vn or PVn, rabbit anti-human Vn antibodies were added (1:1000 dilution), followed by the addition of goat anti-rabbit antibodies-HRP conjugated as described under Materials and Methods. (b) Adsorption to the polystyrene plate. Purified Vn or PVn were added in PBS to 96-well plates at increasing concentrations. Determination of Vn or PVn adsorbed to the plate was conducted as described in the legend to (a). (c) Plasminogen was added (2 g/well) to 96-well plates as described under Materials and Methods. Vn or PVn were added at increasing concentrations in the absence or presence of heparin (50 g/ml). The results are presented as the mean ⫾ SE of three determinations and are representative of three to seven similar experiments.
Does the Phosphorylation of Vn Affect Its Binding to PAI-1 and to Plasminogen? When Vn or PVn are added to immobilized PAI-1, the binding of PVn is significantly lower than that of Vn (Fig. 6a). This reduced binding of PVn was even
FIG. 7. PKA phosphorylation of Vn released from thrombin stimulated platelets. Washed platelets were incubated with [␥- 32P]ATP in the absence or presence of either thrombin (5 u/ml), PKI (20 M), or both for 150 min at 37°C. At the end of the phosphorylation reaction the platelet supernatants were prepared for SDS–PAGE as described under Materials and Methods. The gel was dried and exposed to autoradiography, and the Vn 75 band was excised from the gel and counted in a -scintillation counter.
PKA AFFECTS THE STRUCTURE AND FUNCTION OF VITRONECTIN
tometric analysis of the phosphorylated Vn indicated that addition of thrombin to human platelets increased the Vn 75 phosphorylation by fourfold as compared to the Vn phosphorylation in intact platelets (Fig. 7). Addition of the specific PKA inhibitor, PKI, did not affect the Vn 75 phosphorylation when added to intact platelets (not shown) but reduced the Vn phosphorylation by 50% when added to thrombin-stimulated platelets (Fig. 7). DISCUSSION
The major messages of this paper can be summarized as follows: The phosphorylation of Vn by PKA is stoichiometric and occurs exclusively at Ser378. This phosphorylation affects the conformation of Vn as indicated by circular dichroism (conversion of ⬃7% of the helix structure into an unordered structure), and by the availability of Vn to phosphorylation by CKII. The PKA phosphorylation of Vn is promoted by PAI-1 and by heparin, suggesting that it is targeted preferentially to Vn molecules that are bound to glycosaminoglycans and complexed with PAI-1. Several studies demonstrated the consequences of the VnPAI-1 interaction in relation to the conformation of PAI-1 (20, 21). It was shown, for example, that the binding of Vn stabilizes PAI-1 in its active conformation and the half-life of the inhibitor is increased by two- to fourfold (15). However, the full consequences of the Vn-PAI-1 binding in relation to the Vn structure have not been elucidated yet. It was demonstrated that addition of PAI-1 to nonfractionated plasma or to purified monomeric Vn resulted in the formation of Vn multimers (41). Such conversion of Vn monomers to multimers was shown to be accompanied by changes in Vn function including an increase of PAI-1 binding (42) and of cell binding (43). In the present work we provide evidence showing that following binding to PAI-1, the conformation of Vn is modified in a way that exposes Ser378 to PKA phosphorylation (Figs. 4 and 5). The binding of PVn to immobilized PAI-1 is shown to be lower than the binding of Vn to this inhibitor (Fig. 6a). Heparin impedes the binding of either Vn or PVn to immobilized PAI-1. These findings raise the possibility that the physiological role of the PKA phosphorylation of Vn is to convert the active (inhibitory) PAI-1 (bound to Vn) into its latent, noninhibitory form. This conversion will allow plasminogen activation, and consequently initiate fibrinolysis and blood clot dissolution. The binding of PVn to immobilized plasminogen (Fig. 6c) is also lower than the binding of Vn to this proenzyme, but there is no additional heparin effect in this case. The lowered binding of PVn may indicate that the phosphorylation of Vn favors its detachment from plas-
251
minogen and the consequent conversion of plasminogen to active plasmin, again promoting fibrinolysis. Miles and Plow (44) demonstrated that plasminogen receptors are expressed by a variety of cells, including platelets and endothelial cells. They suggested that the receptors may serve to store plasminogen within thrombi or at sites of injury and inflammation (44). Since platelets contain plasminogen in their granules, and secrete it upon their stimulation (45), the presence of Vn in such sites may probably secure that plasminogen is not converted to plasmin prematurely. On the other hand, the phosphorylation of Vn by PKA may actually serve to initiate fibrinolysis. In line with this suggestion we show that the PKA, released from thrombin-stimulated platelets, can phosphorylate the Vn that is released from such platelets (Fig. 7). In conclusion, the PKA phosphorylation of Vn at Ser378 alters the conformation of Vn and converts it from an antifibrinolytic agent that secures that no fibrinolysis occurs as long as it is not desirable, into a profibrinolytic factor that initiates the solubilization of blood clots and the dissolution of the debris that are not needed any more. The function of this mechanism as a physiological regulatory device has still to be proven in vivo. ACKNOWLEDGMENTS This research was supported by the Israel Science Foundation, Award No. 415/99. S.S. is the Incumbent of the Kleeman Chair in Biochemistry at the Weizmann Institute of Science.
REFERENCES 1. Krebs, E. G. (1987) The Enzymes, Vol. XVII, Academic Press, New York. 2. Smith, C. M., Shindyalov, I. N., Veretnik, S., Gribskov, M., Taylor, S. S., Ten Eyck, L. F., and Bourne, P. E. (1997) Trends Biochem. Sci. 22, 444 – 446. 3. Taylor, S. S., Radzio-Andzelm, E., Madhusudan, Cheng, X., Ten Eyck, L., and Narayana, N. (1999) Pharmacol. Ther. 82, 133– 141. 4. Korc-Grodzicki, B., Tauber-Finkelstein, M., and Shaltiel, S. (1988) Proc. Natl. Acad. Sci. USA 85, 7541–7545. 5. Kubler, D., Pyerin, W., Bill, O., Hotz, A., Sonka, J., and Kinzel, V. (1989) J. Biol. Chem. 264, 14549 –14555. 6. Kubler, D., Reinhardt, D., Reed, J., Pyerin, W., and Kinzel, V. (1992) Eur. J. Biochem. 206, 179 –186. 7. Morgenstern, E., Gnad, U., Preissner, K. T., Dierichs, R., Belleli, A., Chestukhin, A., Schvartz, I., and Shaltiel, S. (2001) Eur. J. Cell Biol. 80, 87–98. 8. Korc-Grodzicki, B., Tauber-Finkelstein, M., Chain, D., and Shaltiel, S. (1988) Biochem. Biophys. Res. Commun. 157, 1131–1138. 9. Preissner, K. T. (1991) Annu. Rev. Cell Biol. 7, 275–310. 10. Suzuki, S., Oldberg, A., Hayman, E. G., Pierschbacher, M. D., and Ruoslahti, E. (1985) EMBO J. 4, 2519 –2524. 11. Hynes, R. O. (1992) Cell 69, 11–25. 12. Dahlback, B., and Podack, E. R. (1985) Biochemistry 24, 2368 – 2374.
252
SCHVARTZ ET AL.
13. Preissner, K. T. (1990) Biochem. Biophys. Res. Commun. 168, 966 –971. 14. Kost, C., Stuber, W., Ehrlich, H. J., Pannekoek, H., and Preissner, K. T. (1992) J. Biol. Chem. 267, 12098 –12105. 15. Declerck, P. J., De, M. M., Alessi, M. C., Baudner, S., Paques, E. P., Preissner, K. T., Muller, B. G., and Collen, D. (1988) J. Biol. Chem. 263, 15454 –15461. 16. Mimuro, J., and Loskutoff, D. J. (1989) J. Biol. Chem. 264, 5058 –5063. 17. Seiffert, D., and Loskutoff, D. J. (1991) J. Biol. Chem. 266, 2824 –2830. 18. Dano, K., Andreasen, P. A., Grondahl, H. J., Kristensen, P., Nielsen, L. S., and Skriver, L. (1985) Adv. Cancer Res. 44, 139 – 226. 19. Pollanen, J., Stephens, R. W., and Vaheri, A. (1991) Adv. Cancer Res. 57, 273–328. 20. Podor, T. J., Shaughnessy, S. G., Blackburn, M. N., and Peterson, C. B. (2000) J. Biol. Chem. 275, 25402–25410. 21. Podor, T. J., Peterson, C. B., Lawrence, D. A., Stefansson, S., Shaughnessy, S. G., Foulon, D. M., Butcher, M., and Weitz, J. I. (2000) J. Biol. Chem. 275, 19788 –19794. 22. Barnes, D. W., Silnutzer, J., See, C., and Shaffer, M. (1983) Proc. Natl. Acad. Sci. USA 80, 1362–1366. 23. Chain, D., Korc-Grodzicki, B., Kreizman, T., and Shaltiel, S. (1990) FEBS Lett. 269, 221–225. 24. Gechtman, Z., Belleli, A., Lechpammer, S., and Shaltiel, S. (1997) Biochem. J. 325, 339 –349. 25. Seger, D., Gechtman, Z., and Shaltiel, S. (1998) J. Biol. Chem. 273, 24805–24813. 26. Seger, D., Seger, R., and Shaltiel, S. (2001) J. Biol. Chem. 276, 16998 –17006. 27. Gechtman, Z., and Shaltiel, S. (1997) Eur. J. Biochem. 243, 493–501.
28. Korc-Grodzicki, B., Chain, D., Kreizman, T., and Shaltiel, S. (1990) Anal. Biochem. 188, 288 –294. 29. Chestukhin, A., Litovchick, L., Batkin, M., and Shaltiel, S. (1996) FEBS Lett. 382, 265–270. 30. Laemmli, U. K. (1970) Nature 227, 680 – 685. 31. Chang, C. T., Wu, C. S-C., and Yang, J. T. (1978) Anal. Biochem. 91, 13–31. 32. Brumfeld, V., and Miller, I. R. (1988) Biophys. J. 54, 747–750. 33. Brumfeld, V., and Werber, M. (1993) Arch. Biochem. Biophys. 302, 134 –143. 34. Bradford, M. M. (1976) Anal. Biochem. 72, 248 –254. 35. Stockmann, A., Hess, S., Declerk, P., Timpl, R., and Preissner, K. T. (1993) J. Biol. Chem. 268, 22874 –22882. 36. Hekman, C. M., and Loskutoff, D. J. (1985) J. Biol. Chem. 260, 11581–11587. 37. Gechtman, Z., Sharma, R., Kreizman, T., Fridkin, M., and Shaltiel, S. (1993) FEBS Lett. 315, 293–297. 38. Preissner, K. T., and Muller-Berghaus, G. (1987) J. Biol. Chem. 262, 12247–12253. 39. Pitt, W. G., Fabrizius-Homan, D. J., Mosher, D. F., and Cooper, S. L. (1989) J. Colloid Interface Sci. 129, 231–239. 40. Preissner, K. T., Holzhuter, S., Justus, C., and Muller-Berghaus, G. (1989) Blood 74, 1989 –1996. 41. Seiffert, D., and Loskutoff, D. J. (1996) J. Biol. Chem. 271, 29644 –29651. 42. Naski, M. C., Lawrence, D. A., Mosher, D. F., Podor, T. J., and Ginsburg, D. (1993) J. Biol. Chem. 268, 12367–12372. 43. Zanetti, A., Conforti, G., Hess, S., Martin, P. I., Ghibaudi, E., Preissner, K. T., and Dejana, E. (1994) Blood 84, 1116 – 1123. 44. Miles, L. A., and Plow, E. F. (1988) Fibrinolysis 2, 61–71. 45. Holt, J. C., and Niewiarowski, S. (1985) Semin. Hematol. 22, 151–163.