INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

A MAJOR FUNCTION of the vascular endothelial cell (EC) monolayer is to serve as a semiselective barrier to fluid and solute flux across the blood vessel wall. The integrity of the EC monolayer is an essential requirement for maintenance of this cellular barrier. Increased endothelial permeability is a prominent feature of acute lung injury and inflammatory lung syndromes and is the result of intercellular gap formation evoked by bioactive agents such as the coagulation protease thrombin. Garcia and Schaphorst (19) and others (34) have employed a working model whereby EC gap formation and barrier function is under close regulation by competing contractile and tethering forces. Isometric force development appears to be controlled, at least in part, by actin-myosin interaction regulated by the phosphorylation status of the 20-kDa myosin "regulatory" light chain (MLC₂₀). Garcia et al. (16) have shown that thrombin-induced EC gap formation and barrier dysfunction require protein kinase (PK) C activation and are coupled to myosin light chain (MLC) phosphorylation catalyzed by a high-molecular-mass Ca²⁺/calmodulin (CaM)-dependent MLC kinase (MLCK) isoform in endothelium. Inhibition of MLCK or activation of cAMP-dependent PKA attenuated MLC phosphorylation and prevented thrombin-induced increases in permeability (16). Changes in the phosphorylation state of MLC are determined by the balance between MLCK and myosin-specific phosphatase (PPase) activities (48). Verin et al. (51) have previously shown that calyculin, a potent type 1 and 2A PPase inhibitor, increased MLC phosphorylation and permeability of EC monolayers and produced gaps in intact and microinjected ECs, suggesting a potential role for PPases in the regulation of EC contractility and barrier function. Although several PPases are capable of dephosphorylating isolated MLC, only type 1 myosin-associated PPase was able to effectively dephosphorylate native smooth and skeletal muscle myosin (1, 7). Type 2A PPase was effective toward isolated MLC but ineffective in dephosphorylation of native myosin (1, 7). Recently, the implied involvement of a Ca²⁺-regulated PPase (2B or calcineurin) in the maintenance of catch tension via MLC dephosphorylation in molluscan muscles (6, 26), as well as the tight association of PPase 2B with the nonmuscle cytoskeleton (13, 38), has raised the possibility that PPase 2B may also be involved in the regulation of contractility in specific cell types. Although Verin et al. (51) recently determined that a type 1 PPase is primarily responsible for the physiological dephosphorylation of myosin in ECs, myosin-specific PPase properties in the endothelium remain incompletely understood. In the present study, we found that, similar to constitutively active myosin-associated PPase 1, thrombin-inducible Ca²⁺/CaM-dependent PPase 2B may be also involved in agonist-mediated EC activation. Furthermore, in contrast to smooth muscle contraction, thrombin-mediated EC contraction-relaxation could require activation of a complex PPase cascade consisting of PPases 1 and 2B.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Materials. Rabbit skeletal muscle glycogen phosphorylase and phosphorylase kinase were purchased from GIBCO (Gaithersburg, MD). Histone type II-S from calf thymus and cAMP-dependent PKA catalytic subunit were obtained from Sigma (St. Louis, MO); CaM was purchased from Ocean Biologics (Edmonds, WA). Purified PPases and polyclonal anti-PPase 1 and anti-PPase 2A antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). Anti-PPase 2B (calcineurin) polyclonal antibody was purchased from either Chemicon (Temecula, CA) or Transduction Laboratories (Lexington, KY). Anti-CS1 antibody was kindly provided by Dr. Anna De Paoli Roach (Indiana University, Indianapolis, IN) and MLC-specific antibody by Drs. James Stull (University of Texas, Dallas, TX) and Susan Gunst (Indiana University, Indianapolis, IN). Deltamethrin was purchased from BIOMOL Research Laboratories (Plymouth Meeting, PA), cyclosporin was obtained from Indiana University Pharmacy (Indianapolis, IN), and calyculin A was obtained from Sigma. Nitrocellulose filters and kaleidoscope-prestained standards were obtained from Bio-Rad Laboratories (Hercules, CA). Prestained molecular-mass standards for SDS-PAGE were purchased from GIBCO (Grand Island, NY). Enhanced chemiluminescence (ECL) developing kit was purchased from Amersham (Little Chalfont, United Kingdom), ¹²⁵I-labeled protein A was purchased from ICN Radiochemicals (Irvine, CA), and [-³²P]ATP was from New England Nuclear (Wilmington, DE). Other reagents were reagent grade and obtained from either Sigma or Bio-Rad.

Bovine pulmonary artery endothelial cell culture. Bovine pulmonary artery endothelial cells (BPAECs) were obtained frozen at 16 passages from American Type Culture Collection (culture line CCL 209; Manassas, VA) and were utilized at passages 19-24 as previously described (16, 51). Cells were cultured in DMEM (GIBCO BRL) supplemented with 20% (vol/vol) colostrum-free bovine serum (Irvine Scientific, Santa Ana, CA), 15 µg/ml of EC growth supplement (Collaborative Research, Bedford, MA), 1% antibiotic-antimycotic solution (10,000 units/ml of penicillin, 10 µg/ml of streptomycin, and 25 µg/ml of amphotericin B; K. C. Biologicals, Lenexa, KS), and 0.1 mM nonessential amino acids (GIBCO) and maintained at 37°C in a humidified atmosphere of 5% CO₂-95% air. The ECs grew to contact-inhibited monolayers with the typical cobblestone morphology. Cells from each primary flask were detached with 0.05% trypsin-EDTA, resuspended in fresh culture medium, and passaged to 100-mm dishes for phosphatase studies, to 60-mm dishes for ³²P labeling, to polycarbonate filters for permeability studies, or into 75-cm² flasks for MLC phosphorylation studies.

MLC phosphorylation in intact endothelium. MLC₂₀ phosphorylation profiles were analyzed by urea-PAGE as previously described by Garcia et al. (16) and Verin et al. (51). Briefly, confluent BPAECs in 75-cm² tissue flasks were incubated with or without phosphatase inhibitors and thrombin. At specified times, the medium was removed, the cells were harvested by scraping into 10% trichloroacetic acid (TCA)-10 mM dithiothreitol (DTT). After centrifugation, the pellets were washed three times with diethyl ether and suspended in 6.7 M urea sample buffer, and ~75 µg protein/lane were run on a 10% polyacrylamide-40% glycerol gel to separate unphosphorylated MLC from phosphorylated MLC, which migrates more rapidly. The proteins were then transferred to nitrocellulose (25 V for 90 min), and both phosphorylated and unphosphorylated MLC₂₀ were detected by immunostaining overnight with an MLC₂₀-specific antibody (1:2,000 dilution) followed by a 1-h treatment with peroxidase-conjugated secondary antibodies (1:3,000 dilution; Bio-Rad). The blot was scanned on a Bio-Rad densitometer, and the percent maximal MLC phosphorylation was expressed by dividing the sum of two times the diphosphorylated MLC area plus the monophosphorylated MLC area by the total of phosphorylated and unphosphorylated areas (i.e., maximum phosphorylation would be 200%). This method is very reproducible because it measures a ratio of peak areas and thus is independent of sample loading.

Endothelial monolayer permeability determination. Macromolecule permeability of EC monolayers was performed as previously described by Patterson et al. (40). Gelatinized polycarbonate micropore membranes (0.8-µm pore size; Corning, Acton, MA) were mounted on the base of plastic cylinders and sterilized with ultraviolet light for 24 h. BPAECs (2 × 10⁵ cells/well) were then seeded on the membrane and grown to confluence. The system consists of two compartments, the upper (luminal) and the lower (abluminal), which are separated by the polycarbonate filter on which the EC monolayer is grown. The lower compartment was stirred continuously and kept at a constant temperature of 37°C by use of a thermally regulated water bath. Medium 199 with 25 mM HEPES in lieu of bicarbonate was used in both compartments. BSA (4% final concentration) complexed to Evans blue dye (EB-BSA) was added to the luminal compartment, and samples were taken from the abluminal compartment every 10 min for 60 min to establish the basal albumin clearance (baseline) and then for an additional 60- to 120-min period after each specific intervention. Transendothelial cell albumin transport was determined by measuring the absorbance of Evans blue dye in abluminal chamber samples at 620 nm in a spectrophotometer (Vmax Multiplate Reader, Molecular Devices, Menlo Park, CA). Albumin clearance rates were calculated by linear regression analysis for control and experimental groups.

PPase activities in EC homogenates. PPase 1 and 2A activities against phosphorylase A were determined as previously described in detail by Verin et al. (51). To measure PPase activity against ³²P-labeled histone II-S (Sigma), EC monolayers were washed two times with phosphate-buffered saline (PBS; 10 mM phosphate buffer, 2.7 mM KCl, and 137 mM NaCl, pH 7.4; Sigma) and two times with ice-cold homogenization buffer (50 mM HEPES and 28 mM -mercaptoethanol, pH 7.4) containing proteinase inhibitors [0.5 mM phenylmethylsulfonyl fluoride, 0.1 mM N-p-tosyl-L-lysine chloromethyl ketone (TLCK), 0.1 mM leupeptin, and 2 mM benzamidine]. Homogenization buffer (200 µl) was added to the EC monolayers, and the plates were scraped and homogenized by passing the cell suspension several times through a 1-ml tuberculin syringe. Aliquots of EC homogenates were diluted two times in assay buffer (50 mM HEPES, pH 7.4, containing 0.6 mM CaCl₂, 50 nM CaM, 28 mM -mercaptoethanol, and 2 mg/ml of BSA). The diluted homogenates (50 µl) containing PPase activity were added to a reaction mixture consisting of 25 µl of ³²P-labeled histone [6 mg/ml; 6,000-9,000 counts · min¹ (cpm) · pmol¹] in HEPES buffer, pH 7.4, containing 2.8 mM -mercaptoethanol. For measuring calyculin-insensitive PPase 2B activity, the assay buffer was supplemented with 30 nM calyculin A. After 30 min at 30°C, the reaction was terminated by the addition of 15 µl of 100% (wt/vol) ice-cold TCA. After 15 min on ice, the suspension was centrifuged for 5 min (IEC Centra-M centrifuge, International Equipment), and aliquots (45 µl for each sample) of supernatant were loaded onto Whatman 3 MM paper, dried for 15 min, and counted in Scintiverse scintillation solution (Fisher Scientific, Fair Lawn, NJ) by beta scintigraphy (Beckman model LS 6000 IC). Reactions were carried out in duplicate, and controls consisted of incubations in which the PPase-containing cell preparation was replaced by assay buffer. To ensure linear rates of dephosphorylation, the extent of dephosphorylation was restricted to <25%. One unit of phosphatase activity was defined as the amount of enzyme that released 1 nmol ³²P_i/min.

Preparation of ³²P-labeled substrates. ³²P-labeled phosphorylase A (1,000-3,000 cpm/pmol) was prepared by phosphorylation of phosphorylase B with phosphorylase kinase as previously described in detail by Verin et al. (51). ³²P-labeled phosphohistone (6,000-9,000 cpm/pmol) was prepared by phosphorylation of histone type II-S (Sigma) with a catalytic subunit of porcine PKA (Sigma) as previously described (11), with some modifications. Briefly, histone (90 mg) was dissolved in 15 ml of 50 mM HEPES buffer, pH 7.4, including 10 mM magnesium acetate and 1 mM DTT. Next, PKA (30 µg) reconstituted in 0.6 ml of water containing 40 mM DTT, 5 µM ATP, and 30 µCi of [-³²P]ATP was added to the histone solution, and the reaction mixture was incubated at 25°C for 1 h in a shaking water bath. The reaction was terminated by adding 100% (wt/vol) TCA to a final concentration of 15%. After centrifugation for 5 min at 10,000 g, the supernatant was discarded, and the pellet was washed two times by resuspension and centrifugation in 15 ml of 15% ice-cold TCA. The pellet was dissolved in 7.5 ml of 0.1 M NaOH and dialyzed first against 3 liters of PBS for 4 h and then against the same volume of HEPES buffer, pH 7.4, containing 2.8 mM -mercaptoethanol overnight. ³²P-labeled substrates were stored at 4°C.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Effect of thrombin on MLC-specific PPase activity in intact endothelium. The dynamics of thrombin-induced MLC phosphorylation-dephosphorylation in BPAECs were monitored by urea-glycerol gel electrophoresis followed by Western blotting with an MLC-specific antibody as previously described by Garcia et al. (16) and Verin et al. (51). Figure 1 shows that MLCs are constitutively phosphorylated, with a stoichiometry of ~0.4 mol phosphate/mol MLC. Thrombin treatment (100 nM) causes a rapid increase in MLC phosphorylation, with maximal phosphorylation at 2 min (~0.9 mol phosphate/mol MLC), which gradually decreased in a time-dependent manner, with MLC dephosphorylation nearly returning to baseline by 60 min, with a stoichiometry of ~0.5 mol phosphate/mol MLC (Fig. 1). Because these results implicated the involvement of PPases in the regulation of thrombin-induced EC MLC phosphorylation, we studied the effect of the PPase inhibitors calyculin A (Fig. 2A) and cyclosporin (Fig. 2B) on the phosphorylation status of MLC in thrombin-stimulated and nonstimulated ECs. Intact EC monolayers were pretreated with vehicle or specific PPase inhibitors for 1 h at 37°C and then stimulated with 100 nM thrombin for 2 and 60 min. The levels of unphosphorylated and mono- and diphosphorylated MLC species were compared both with the basal MLC phosphorylation profile present in ECs and with the level of maximal thrombin-activated MLC phosphorylation (Fig. 2). EC pretreatment with 10 nM calyculin A [equally selective for PPase 1 and PPase 2A; IC₅₀ = 0.3-0.4 nM (24)] significantly increased the level of basal and maximal thrombin-induced MLC phosphorylation but did not completely prevent thrombin-induced MLC dephosphorylation (Fig. 2A). In contrast, 200 nM cyclosporin [selective inhibitor of PPase 2B; IC₅₀ = 7 nM (14)] had no effect on either the basal level or the maximal thrombin-induced MLC phosphorylation but significantly attenuated the thrombin-induced MLC dephosphorylation at 60 min (Fig. 2B). Another specific PPase 2B inhibitor, deltamethrin [0.1-1 µM; IC₅₀ = 0.03 nM (11)] exhibited similar effects in stimulated and control cells (Fig. 3). These data indicate that the activation of both PPase 1 and the Ca²⁺/CaM-dependent PPase 2B may be involved in thrombin-mediated MLC dephosphorylation.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time course of endothelial cell (EC) myosin light chain (MLC) phosphorylation after thrombin activation. Confluent EC monolayers were rinsed 2 times with medium 199 (M199) to remove serum and incubated in M199 with 25 mM HEPES (M199H) in lieu of bicarbonate. Cells were treated with vehicle (Control; M199H) or 100 nM thrombin for indicated time periods, and MLC phosphorylation was monitored by urea gel electrophoresis followed by immunoblotting with anti-MLC antibodies and quantitated by scanning densitometry as described in METHODS. Data are means ± SE from 14 independent experiments. * Significant difference compared with maximal level of thrombin-induced MLC phosphorylation, P < 0.05.

View larger version (19K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of phosphatase (PPase) inhibitors on MLC phosphorylation in EC monolayers. Confluent bovine pulmonary artery endothelial cells (BPAECs) were rinsed as described in Fig. 1, pretreated with either vehicle [0.1% DMSO (A) or M199H (B)] or specific PPase inhibitors [10 nM calyculin A (A) or 200 nM cyclosporin (B)] for 1 h, and then stimulated with 100 nM thrombin for 2 and 60 min. MLC phosphorylation was quantitated as described in METHODS. Data are means ± SE from at least 6 independent experiments. Calyculin significantly increased levels of basal and maximal thrombin-induced MLC phosphorylation but did not completely prevent thrombin-induced MLC dephosphorylation, whereas cyclosporin had no effect on levels of basal and maximal MLC phosphorylation but partially blocked thrombin-induced MLC dephosphorylation. Significant difference (P < 0.05): * compared with basal level of MLC phosphorylation; ** between thrombin-alone and inhibitor-thrombin groups at 2 min; *** between thrombin groups at 2 and 60 min (with and without inhibitor separately); # between maximal level of inhibitor-induced and inhibitor-thrombin-induced MLC phosphorylation; ## between thrombin-alone and inhibitor-thrombin groups at 60 min; ### between inhibitor-alone groups at 2 and 60 min.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effect of deltamethrin on thrombin-induced EC MLC phosphorylation. Confluent BPAECs were pretreated in M199H with vehicle (0.1% ethanol) or 100 nM deltamethrin for 1 h and then stimulated with 100 nM thrombin for 2, 30, and 60 min. MLC phosphorylation was determined as described in METHODS and is expressed as a percentage of maximum response. Data are means ± SE from 3 independent experiments. Significant difference (P < 0.05): * compared with maximal level of MLC phosphorylation; # between control and deltamethrin-pretreated groups. Similar to cyclosporin, deltamethrin did not affect either basal or maximal MLC phosphorylation but significantly attenuated thrombin-induced MLC dephosphorylation.

Effect of PPase 2B inhibition on thrombin-induced EC permeability. Verin et al. (51) have previously shown that PPase 1 is directly involved in EC contractile responses and regulation of EC barrier function. To examine whether PPase 2B is similarly involved, we next determined the effect of the specific PPase 2B inhibitor deltamethrin on basal and thrombin-induced permeabilities of EC monolayers. Treatment of EC monolayers with deltamethrin for 1 h (Fig. 4) did not affect the basal level of EC permeability but significantly enhanced (~1.5-fold) the thrombin-induced increases in EB-BSA flux across EC monolayers. Although speculative, these results suggest that thrombin-inducible type 2B PPase activity may be involved in the recovery of the EC protective barrier after initial disruption by thrombin. The concentration of deltamethrin required was identical for both the enhancement of thrombin-induced EC permeability and the partial prevention of thrombin-induced MLC dephosphorylation.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Effect of PPase 2B inhibition on thrombin-induced EC monolayer permeability. BPAEC monolayers (n = 11) were pretreated at time 0 with either vehicle (0.1% ethanol) or deltamethrin (delta; 100 nM). At 60 min, cells were challenged with either vehicle (M199H alone) or 100 nM thrombin, and permeability was determined from 130 to 180 min as described in METHODS. * and # Significantly different compared with control and thrombin alone, respectively, P < 0.04. Inhibition of PPase 2B by deltamethrin did not significantly alter basal level of EC permeability but significantly enhanced thrombin-induced albumin clearance.

Subcellular localization of PPase 2B in endothelium. To further elucidate PPase 2B involvement in the regulation of EC contractile processes, we studied the distribution of PPases 1, 2A, and 2B in myosin-enriched and myosin-depleted fractions [prepared as previously described by Verin et al. (51); Fig. 5] and in cytoskeletal fractions. For preparation of the cytoskeletal fractions, cytosolic proteins were removed by treatment of BPAEC monolayers with 1% Nonidet P-40-150 mM NaCl extraction buffer (Fig. 5). Dishes were washed two times with the same buffer, and the insoluble proteins remaining on dishes were solubilized in SDS sample buffer. This cell fraction, which is enriched in actin, also includes significant amounts of myosin and has been previously characterized as cytoskeletal (32). Western immunoblotting analysis with specific anti-PPase antibodies revealed that PPase 2B was present in both the cytoskeletal and myosin-depleted cytosolic fractions but not in the myosin-enriched fraction. In contrast, PPase 1 was present in both the myosin-enriched and myosin-depleted cytosolic fractions but was absent in the cytoskeletal fraction (Fig. 6). These data (Fig. 6A) and the previous report by Verin et al. (51) indicate that immunoreactive PPase 2A is absent in both contractile protein-enriched fractions but is present in EC homogenates. Together, these results indicate that only PPase 2B, but not type 1 or 2A, is tightly associated with the actin-enriched cytoskeleton.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Schema for preparation of EC fractions showing algorithm employed to identify myosin-enriched, myosin-depleted, and cytoskeletal BPAEC fractions. For preparation of a cytoskeletal fraction enriched in actin, BPAECs were rinsed with PBS, and cytosolic proteins were extracted with 50 mM Tris · HCl buffer, pH 8.0, containing 1% Nonidet P-40 (NP-40), 150 mM NaCl, and protease inhibitors (see METHODS) for 30 min at 4°C. Extractable proteins were discarded, and remaining detergent-insoluble (cytoskeletal) EC fraction was solubilized in SDS-PAGE sample buffer (31) and used for Western immunoblotting. For preparation of myosin-enriched fractions, low-ionic-strength homogenates obtained from confluent EC monolayers (51) were treated with 0.1% Tween 20-0.6 M NaCl for 1 h at 4°C and centrifuged, and resulting myosin-containing supernatants were further diluted (10-fold) with low-ionic-strength buffer (50 mM Tris · HCl, pH 7.0, 0.1 mM EDTA, 28 mM -mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, and 2 mM benzamidine). Precipitated myosin and associated protein fractions were collected by centrifugation and subjected to PPase activity determination or Western immunoblotting analysis.

View larger version (34K): [in this window] [in a new window]   View larger version (35K): [in this window] [in a new window]   Fig. 6.   Distribution of PPases in EC fractions. PPase content in cytoskeletal fraction (A) and myosin-enriched (M+) and myosin-depleted (M) EC fractions (B) was analyzed by Western immunoblotting. A: cytoskeletal proteins from 100-mm dishes prepared as described in Fig. 5 were solubilized in 3 ml of SDS sample buffer (31), boiled for 5 min, clarified by centrifugation, and subjected to 9% SDS-PAGE followed by Western immunoblotting. Anti-PPase 1, 2A, and 2B antibodies were used for PPase identification. Fifty nanograms of each PPase were used as positive controls. Amount of protein extracts loaded onto gel (~30 µl) was normalized by quantitation of Coomassie-stained gels in preliminary experiments. B: M+ pellet and M supernatant fractions obtained as described in Fig. 5 were subjected to 9% SDS-PAGE. Proteins were visualized by staining gel with Coomassie R-250 (left) or electroblotted from gel onto a nitrocellulose filter (right), and PPases were detected by incubation of nitrocellulose filters with specific anti-PPase 2B antibodies (top) or anti-PPase 1 antibodies (anti-CS1 antibodies; bottom) followed by enhanced chemiluminescence (ECL) detection. Equivalent amounts (~25 µl) of initial homogenate were loaded onto gel for each fraction. Nos. on left, positions of protein molecular-mass markers. These data and previous experiments (50) clearly indicate that PPases 1 and 2B but not PPase 2A are tightly associated with EC fractions enriched in contractile proteins.

Effect of thrombin on subcellular distribution of PPase 2B and MLC in ECs. Because thrombin caused stress-fiber formation, reflecting increased interaction between actin and myosin, we next sought to examine thrombin effects on the subcellular localization of PPase 2B (colocalized with actin) and MLC (a part of myosin). Examination by confocal microscopy of thrombin-challenged BPAEC monolayers (Fig. 7) revealed that, compared with control cells, thrombin produced an increased colocalization (yellow) of MLC (green) and PPase 2B (red), with maximal costaining at 2 min. Although PPase 2B staining was constant, there was a time-dependent decline in MLC staining, likely due to a decrease in MLC antibody accessibility due to increased actin polymerization and actin-myosin interactions. However, even at 10 min, a significant increase of costaining between PPase 2B and MLC was observed compared with control ECs (Fig. 7, A and C), suggesting an increase in the interaction between PPase 2B and MLC after thrombin stimulation.

View larger version (74K): [in this window] [in a new window]   Fig. 7.   Effect of thrombin on intracellular localization of PPase 2B and MLC. BPAECs grown to confluence on glass coverslips were exposed to vehicle (A) or bovine thrombin (100 nM) for 2 (B) or 10 min (C). Cells were fixed with paraformaldehyde and permeabilized with 0.2% Triton as described in METHODS. Cells were incubated with both rabbit anti-smooth MLC antibody (1:50 in Tris-NaCl-1% albumin buffer) and mouse anti-calcineurin antibody (1:50) overnight at 4°C. Washed cells were then secondarily stained with donkey anti-rabbit IgG antibody (fluorescein conjugated; 1:50) and donkey anti-mouse IgG antibody (lissamine rhodamine conjugated; 1:50). Immunofluorescence was recorded by confocal microscopy. Cells show an increase in costaining (yellow) of MLC (green) with PPase2B (red), with maximal costaining at 2 min.

PPase activity in BPAEC homogenates and cytoskeletal fractions. A previous study by Verin et al. (51) has characterized the involvement of a type 1 PPase in the regulation of EC contractility and barrier function. The data described in Effect of thrombin on subcellular distribution of PPase 2B and MLC in ECs strongly implicate the additional involvement of PPase 2B in the regulation of EC contractility after thrombin stimulation. To further evaluate the myosin-specific PPase involved in agonist-inducible EC contractile responses, we studied the effect of thrombin on the PPase activity present in the homogenates and specific EC fractions. For quantification of PPase 1 and 2A activities in EC, ³²P-labeled phosphorylase A, a well-recognized, selective, and specific substrate for PPases 1 and 2A in mammalian tissues, was utilized (25). To separate PPase 1 and 2A activities of EC homogenates, we used 2 nM okadaic acid in the assay buffer, which in the diluted extracts (<0.1 U/ml in the assay) inhibits PPase 2A completely (IC₅₀ ~ 0.1-0.3 nM) but has no effect on PPase 1 (IC₅₀ = 51 nM) (8, 24). Table 1 shows that stimulation of BPAEC monolayers with 100 nM thrombin for 10 min has no significant effect on either PPase 1 or 2A activity in cell homogenates. Increases in either the thrombin concentration (1 µM) or the duration of stimulation (up to 30 min) also failed to alter EC PPase 1 and 2A activities (data not shown). Interestingly, thrombin stimulation actually decreased PPase 1 activity against phosphorylase A (by 30%) in the myosin-enriched fraction (Table 1). Western immunoblotting of control and agonist-stimulated BPAEC fractions shows that this observation can be explained by the partial release of the PPase 1 catalytic subunit from the actomyosin complex, with redistribution from the myosin-enriched to the myosin-depleted (cytosolic) fraction (Fig. 8).

View larger version (59K): [in this window] [in a new window]   Fig. 8.   Effect of thrombin on PPase 1 fractional distribution. M supernatant fractions and M+ pellet fractions were prepared from BPAECs treated with vehicle (M199H) or thrombin (100 nM) for 10 min and subjected to 9% SDS-PAGE followed by electrotransfer to nitrocellulose filters. PPase 1 was detected by incubation of nitrocellulose filters with specific anti-PPase 1 antibodies followed by 125I-protein A treatment. Amount of protein loaded onto gel in each fraction was equivalent to ~25 µl of initial homogenate. Nos. on right, positions of protein molecular-mass markers. Thrombin induced redistribution of PPase 1 catalytic subunit from M+ fraction to M EC fraction, a finding that correlates well with decreased myosin-associated PPase 1 activity noted after thrombin (Table 1).

For quantification of Ca²⁺/CaM-dependent PPase 2B activity, we utilized ³²P-histone II-S (Sigma) phosphorylated by cAMP-dependent PKA as a PPase 2B substrate. It was initially shown that PKA-phosphorylated histone H1 was a substrate for PPases 1 and 2A but not for PPases 2B and 2C (46). However, it was recently demonstrated that PKA-phosphorylated histone II-S (mixed histone fraction prepared from calf thymus by sequential high-salt extraction, precipitation in water, acid extraction, and reprecipitation with alcohol) had a high capacity to be dephosphorylated by PPase 2B (11). In preliminary experiments (data not shown), we demonstrated that under our assay conditions both purified PPases 2A and 2B, but not PPase 1, effectively dephosphorylated PKA-phosphorylated histone II-S. However, only PPase 2B activity is dramatically increased (>8-fold) with increasing Ca²⁺ concentration (from 0.1 mM EGTA to 0.3 mM CaCl₂ in the assay buffer), whereas PPase 2A activity remains unaffected. To eliminate PPase 2A contribution in EC PPase activity against phosphohistone II-S, we measured PPase activity in the presence of 10 nM calyculin, which we have determined to inhibit completely both type 1 and type 2A PPases in EC homogenates (51) but has no effect on PPases 2B and 2C (27, 45). Although the addition of calyculin (10 nM) significantly decreased PPase activity against histone II-S in EC homogenates by ~70% (Fig. 9A), we have assumed that the remaining calyculin-insensitive PPase activity against histone II-S reflects PPase 2B. In contrast to the lack of induced PPase 1 activity after thrombin (Table 1), pretreatment of BPAECs with 100 nM thrombin led to significant increases in calyculin-insensitive EC PPase 2B activity against histone II-S (Fig. 9B) in a time-dependent manner (peaked at 20 min with a 1.7-fold increase). Consistent with these results, BPAEC stimulation with the Ca²⁺ ionophore ionomycin (10 µM) also led to rapid activation of calyculin-insensitive Ca²⁺/CaM-dependent PPase 2B activity in EC homogenates (Fig. 9A).

View larger version (14K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Fig. 9.   Effect of agonists on histone PPase activity in EC homogenates. BPAEC monolayers were stimulated with ionomycin (5 min, 10 µM; A) or 100 nM thrombin for indicated time periods (B), and PPase activity against protein kinase (PK) A-phosphorylated histone II-S was determined in presence and absence of 10 nM calyculin A as described in METHODS. A: values are means ± SE for 4 independent experiments. * Significant difference compared with total activity in control monolayers, P < 4 × 103. # Significant difference compared with calyculin-insensitive activity in control monolayers, P = 4.4 × 103. B: BPAECs were treated with thrombin at time 0. PPase activity of EC homogenates from stimulated () and control () cells was sequentially monitored. Values are means ± SE for 8 independent experiments. * Significant difference compared with control activity (either total or calyculin-insensitive activity), P  0.03. Thrombin stimulates calyculin-insensitive PPase 2B activity in a time-dependent manner, with this activation mimicked by Ca2+ ionophore ionomycin.

Because we detected the presence of PPase 2B but not PPase 1 or 2A in the cytoskeleton (Fig. 10), we next determined the effect of thrombin-induced PPase 2B activation on the phosphorylation status of cytoskeletal proteins. We purified cytoskeletal fractions from ³²P-labeled EC monolayers and analyzed SDS-PAGE radioactive protein profiles by autoradiography. Figure 10 demonstrates that EC treatment with the specific PPase 2B inhibitor deltamethrin caused a significant increase in the phosphorylation of several cytoskeletal proteins, indicating basal PPase 2B activity in the cytoskeleton. Thrombin treatment (100 nM, 30 min) caused the significant dephosphorylation of specific proteins in the cytoskeleton, which was completely abolished by pretreatment with deltamethrin. Similar results were obtained with cells treated with ionomycin (data not shown). Taken together, these results demonstrate the existence of thrombin-inducible PPase 2B activity that is tightly associated with the cytoskeleton.

View larger version (70K): [in this window] [in a new window]   Fig. 10.   Effect of thrombin on cytoskeletal protein phosphorylation. Shown is autoradiogram of phosphorylated cytoskeletal proteins obtained from thrombin-treated ECs labeled with 32P. BPAEC monolayers were loaded with [32P]orthophosphate (0.4 mCi/ml) in phosphate-free DMEM in presence of 1% serum for 4 h and then preincubated with vehicle (0.1% ethanol) or deltamethrin (1 µM) for 1 h after thrombin treatment (100 nM, 30 min). +, Presence; , absence. Cytoskeletal fractions were prepared as described in METHODS, subjected to SDS-PAGE (31), and stained for proteins with Coomassie blue R-250. Stained gels were dried and subjected to autoradiography with ECL Hyperfilm. Nos. on left, positions of molecular-mass standards. Arrows, major phosphorylated bands. Thrombin caused a significant decrease in cytoskeletal protein phosphorylation, which was completely attenuated by deltamethrin.

Effect of thrombin on PPase 2B phosphorylation status. Our data indicate that both the Ca²⁺ ionophore ionomycin and thrombin are able to activate Ca²⁺/CaM-dependent PPase 2B in the endothelium (Fig. 9). However, the time course of thrombin-induced PPase 2B activation, which peaks at 20 min (Fig. 9B), correlates poorly with thrombin-induced Ca²⁺ influx, which peaks at <1 min and rapidly declines (18). To explore additional factors that may modulate PPase 2B activation in thrombin-stimulated endothelium, we examined the status of PPase 2B phosphorylation after thrombin stimulation in denaturing immunoprecipitates from ³²P-labeled BPAEC monolayers. Figure 11 demonstrates that thrombin initially produces a rapid PPase 2B dephosphorylation. Although we were not able to detect a rapid transient increase in PPase 2B activity in EC homogenates after thrombin, we believe that the thrombin-induced decrease in PPase 2B phosphorylation may result from Ca²⁺-dependent PPase 2B autodephosphorylation. The subsequent time-dependent thrombin-induced increase in the phosphorylation status of PPase 2B correlated well with the increase in PPase 2B catalytic activity noted in EC homogenates and depicted in Fig. 9. These data suggest that thrombin initiates a sequence of events that lead to time-dependent PPase 2B activation in endothelium, possibly via phosphorylation of the enzyme. Although information is limited about the role of PPase 2B phosphorylation in the regulation of enzymatic activity, the activation of PPase 2B via phosphorylation of a PPase 2B catalytic subunit has been reported (41).

View larger version (28K): [in this window] [in a new window]   View larger version (34K): [in this window] [in a new window]   Fig. 11.   Effect of thrombin on PPase 2B catalytic subunit phosphorylation. BPAEC monolayers from 60-mm dishes were preloaded with [32P]orthophosphate (0.5 mCi/ml in phosphate-free DMEM) for 2.5 h at 37°C, followed by stimulation with either vehicle alone or thrombin (100 nM, various times). Stimuli were then removed; monolayers were rinsed 2 times with 2 ml of medium, further rinsed with 2 ml of PBS, and scraped into 100 µl of SDS-denaturing stop solution; and cell extracts were used for immunoprecipitation under denaturing conditions with PPase 2B-specific antibodies as described in METHODS. Immunocomplexes were then subjected to SDS electrophoresis (31). After electrophoresis, proteins were transferred to nitrocellulose membranes, and 32P signals were detected by autoradiography at 70°C. To identify PPase 2B position, membranes were subsequently stained with specific PPase 2B antibodies. Relative intensities of 32P-labeled PPase 2B were quantified by scanning densitometry. A: Western immunoblot of purified (Pur.) PPase 2B catalytic subunit (50 ng) and PPase 2B immunoprecipitated from BPAECs (EC I/P). Nos. on left, positions of molecular-mass markers. Arrows, positions of PPase 2B and IgG. B: autoradiogram (Autorad) of phosphorylated PPase 2B from thrombin-treated ECs loaded with 32P (top), an immunoblot of the same membrane stained with anti-PPase 2B antibodies (middle), and quantitation of level of PPase 2B phosphorylation (mean ± SE) in 3 independent experiments (bottom). 2', 10', and 60', 2, 10, and 60 min, respectively. Thrombin stimulation results in a time-dependent increase in PPase 2B catalytic subunit phosphorylation in intact BPAECs, with maximal 260% increase at 10 min as detected by densitometric scanning.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

There is now substantial evidence implicating MLCK-catalyzed MLC phosphorylation as an essential element of agonist-stimulated EC barrier dysfunction. EC contraction and the resultant intercellular gap formation and barrier dysfunction are associated with redistribution of microfilament proteins and are dependent on the level of MLC phosphorylation (43, 44, 52). Recently, it was shown that EC exposure to thrombin results in isometric tension development, which correlates with MLC phosphorylation, and is accompanied by association of myosin with the detergent-insoluble cytoskeleton (21, 36). Consistent with these observations, a recent study by Garcia et al. (16) in intact ECs demonstrated that MLCK inhibition (with either ML-7 or KT-5926) attenuates thrombin-induced MLC phosphorylation and barrier dysfunction. Less clear, however, is the extent and type of involvement of Ser/Thr PPase activities in the regulation of thrombin-mediated EC MLC phosphorylation-dephosphorylation. In nonmuscle cell systems, the addition of specific PPase inhibitors leads to changes such as actin aggregation in the cytoskeleton (30), which implicates the involvement of PPases in the regulation of nonmuscle contraction. In smooth muscle tissues, PPase-mediated MLC dephosphorylation deactivates the actomyosin ATPase, causing relaxation (see Refs. 10, 48 for reviews). For example, the addition of a PPase-enriched protein fraction derived from aortic smooth muscle cells (SMCs) was shown to enhance relaxation in skinned fibers (42) and to decrease isometric force, shortening velocity, and MLC phosphorylation (2, 3). Four different PPases were subsequently identified in gizzard smooth muscle extracts; however, only two PPases were able to dephosphorylate intact myosin (39). A myosin-associated PPase in skeletal, cardiac, and smooth muscles was subsequently identified as a type 1 enzyme that was able to effectively dephosphorylate native myosin, whereas PPase 2A was effective only toward isolated MLC but was ineffective in dephosphorylation of native myosin (1, 7). Recently, it was shown that several nonmuscle cells, similar to SMCs, contain the 130-kDa "targeting subunit" of myosin-associated PPase 1 (37). Microinjection of PPase 1 but not of PPase 2A into mammalian fibroblasts modulated MLC phosphorylation and induced microfilament reorganization (12), which suggests that, similar to SMCs, PPase 1 is involved in the regulation of nonmuscle cell contractility. Using primarily inhibitory analysis, Verin et al. (51) previously reported that, similar to SMCs, the myosin PPase in ECs involved in physiologically relevant MLC dephosphorylation is a type 1 enzyme. Consistent with our data, Shinoki et al. (47) demonstrated that PPase 1 plays a predominant role in sustaining the normal EC cytoskeletal structure.

Although these studies suggest the primacy of a type 1 PPase in MLC dephosphorylation in mammalian muscle and nonmuscle systems, Ca²⁺-dependent PPase 2B purified from scallop smooth muscle was shown to effectively dephosphorylate MLC bound in scallop opaque and aortic smooth muscle myosin (26) and to be involved in the maintenance of catch tension by MLC dephosphorylation in molluscan muscles (6, 26). PPase 2B has been shown to participate in sperm motility and to be tightly associated with mammalian nonmuscle cytoskeleton (13, 38, 49). These reports raise the possibility that PPase 2B may also be involved in the regulation of contractility in specific cell types. The present study focused on the role of Ca²⁺/CaM-dependent PPase 2B in agonist-stimulated EC contractile responses. Our present results clearly demonstrate that thrombin (100 nM) causes significant (>2-fold) and rapid (2-min) increases in MLC phosphorylation, which gradually declined, with dephosphorylation nearly complete after 1 h (Fig. 1). Garcia et al. (16) previously showed that higher doses of thrombin (1 µM) caused a three- to fourfold increase in MLC phosphorylation (~1.2 mol phosphate/mol MLC) above basal constitutive levels, with maximal effect at 1 min, followed by MLC dephosphorylation, which was nearly complete by 30 min. Taken together, these data demonstrate that thrombin has dose-dependent effects on EC MLC phosphorylation-dephosphorylation status and implicates the involvement of Ser/Thr PPases in EC MLC dephosphorylation events after thrombin stimulation. Although we clearly determined a type 1 PPase is involved in thrombin-mediated MLC dephosphorylation, calyculin A, a potent inhibitor of both type 1 and type 2A PPases (24, 27), failed to prevent MLC dephosphorylation after thrombin (Fig. 2A), suggesting that calyculin-insensitive PPase activities may be involved in MLC dephosphorylation after thrombin.

The two major types of calyculin-insensitive PPases that exist in cells are highly dependent on divalent cation availability: Ca²⁺/CaM-dependent PPase 2B and Mg²⁺-dependent PPase 2C (44). Although there is no evidence that agonists such as thrombin affect the intracellular Mg²⁺ concentration in the endothelium, Garcia et al. (18) and others (22) have previously shown that thrombin-mediated EC activation caused rapid (<15-s) transient increases in intracellular Ca²⁺ concentration. Thus it is logical to assume that PPase 2B rather than PPase 2C would be more likely involved in thrombin-mediated EC contractile responses. The specific PPase 2B inhibitors cyclosporin (Fig. 2B) and deltamethrin (Fig. 3), at concentrations sufficient for complete PPase 2B inhibition (11, 14), did not alter the level of either basal or maximal thrombin-induced MLC phosphorylation but significantly attenuated the predicted MLC dephosphorylation after thrombin. Furthermore, inhibition of thrombin-induced MLC dephosphorylation by deltamethrin was accompanied by a subsequent increase in thrombin-induced EC permeability (Fig. 4). Although not definitive, the same concentration of PPase 2B inhibitor was equally effective for both events, suggesting that the effects on MLC phosphorylation and permeability may be linked. Several studies have previously implicated the involvement of PPase 2B in EC activation responses. For example, PPase 2B is involved in endothelin-1 gene regulation in human endothelium (35), and prolonged incubation of human ECs with cyclosporin A (4 days, 0.3-7 µM) results in significant cell detachment (4). To our knowledge, our data represent the first demonstration that PPase 2B may have an important regulatory role in EC contractile and barrier responses.

To further elucidate PPase 2B involvement in the regulation of EC contractile processes, we studied the distribution of this PPase in different EC fractions. Our results indicate that immunoreactive PPase 2B is present in a cytoskeletal actin-enriched fraction (which includes significant amounts of myosin) and in the myosin-depleted cytosolic fraction. In contrast, PPase 1 is present in the myosin-enriched and myosin-depleted fractions but is entirely absent in the detergent-insoluble cytoskeleton. Taken together with previous observations by Verin et al. (51), these results suggest that PPase 2B but not PPase 1 or 2A is tightly associated with the EC cytoskeleton. In agreement with our data, highly active PPase 2B was recently purified from the detergent-insoluble adrenal cell cytoskeleton (38), and neuronal cell treatment with cytochalasin D resulted in dissociation of PPase 2B from disrupted actin filaments (13). These observations suggest that the association of PPase 2B with the cytoskeleton depends on intact actin filaments (13) rather than on filamentous myosin, although we cannot exclude the possibility that PPase 2B can be specifically associated with a specific myosin heavy chain II isoform (A or B) that is present in nonmuscle cells (28).

It has been recently reported that PPase 1 is also associated with the detergent-insoluble cytoskeletal fraction from cerebral cortex (9). We did not find immunoreactive PPase 1 associated with the EC cytoskeleton; however, previously published observations by Verin et al. (51) indicated that ~10% of total phosphorylase PPase activity (combined PPase 1 and 2A activity) resides in the cytoskeletal fraction. Western blotting analysis of EC fractions with anti-PPase 1 antibodies produced against the type 1 catalytic subunit- (CS1) (Fig. 7A) and anti-CS1 isoform antibodies (data not shown) did not detect any CS1 protein band in the cytoskeletal fraction. Therefore, these data suggest that the PPase 1 activity associated with the EC cytoskeleton may be attributable to a CS1 isoform that is immunologically distinct from the CS1 and - isoforms.

Our data indicate that PPase 1 and 2A activities in EC homogenates are not affected by thrombin stimulation (Table 1). Nevertheless, thrombin caused a partial translocation of the EC CS1 from the myosin-enriched to the myosin-depleted fraction (Fig. 8), which was accompanied by decreasing myosin-associated PPase 1 activity (Table 1). These data are consistent with the previously published observations (45) that thrombin inhibits calyculin-sensitive MLC dephosphorylation in the endothelium. The mechanism responsible for thrombin-induced CS1 translocation is unclear; however, dissociation of CS1 from its targeting subunit was proposed as a mechanism of regulation of the glycogen-associated form of PPase 1 (46). In SMCs, arachidonic acid, an intermediate product of G_s protein activation, caused dissociation of CS1 from the myosin-associated PPase 1 holoenzyme, which may account for the G protein-mediated sensitization of the SMC contractile apparatus to Ca²⁺ (23, 29). Garcia et al. (17) previously showed that G protein-mediated EC activation by thrombin results in Ca²⁺ mobilization; phospholipase A₂, C, and D activation; and release of arachidonic acid (17). Although speculative, it is possible that, similar to SMCs, arachidonate generated by thrombin-stimulated phospholipase A₂ and D activity can cause CS1 translocation to the cytosol and decrease myosin-specific PPase 1 activity. Nevertheless, it is clear that the observed decrease in myosin-associated PPase 1 activity cannot explain EC MLC dephosphorylation after thrombin stimulation.

Our results demonstrated that thrombin caused dephosphorylation of several proteins in the cytoskeletal fractions (Fig. 10) and that this reaction was totally blocked by the specific PPase 2B inhibitor deltamethrin. These data are consistent with the existence of a thrombin-inducible PPase 2B, which is tightly bound to the contractile apparatus. As evidenced by immunofluorescence experiments (Fig. 7), thrombin also increased colocalization of PPase 2B and MLC, thereby enhancing MLC substrate accessibility to PPase 2B. Finally, thrombin significantly increased calyculin-insensitive PPase 2B activity in EC homogenates in a time-dependent manner. This activation was evident after 10 min of stimulation and was maximal after 20 min of thrombin treatment (Fig. 9B), a time frame strongly correlated with MLC dephosphorylation induced by thrombin (Fig. 2). EC monolayers exposed to the Ca²⁺ ionophore ionomycin (5 µM for 15 min) also significantly increased PPase 2B activity in EC homogenates (Fig. 9A), consistent with the involvement of Ca²⁺ influx in thrombin-induced EC PPase 2B activation. Of potential concern is the discrepancy in the time course of thrombin-induced PPase 2B activation (Fig. 9) with the time course of the previously reported increase in thrombin-mediated intracellular Ca²⁺ concentration (18). However, our data indicate a strong correlation between PPase 2B catalytic subunit phosphorylation induced by thrombin and PPase 2B activity (Fig. 11). These results indicate that thrombin induces a sequence of biochemical events, including Ca²⁺ mobilization and PK(s) activation, which precedes and may initiate Ca²⁺/CaM-dependent PPase 2B activation in the endothelium. Taken together with the role of PPase 2B in MLC dephosphorylation in intact EC monolayers, these experimental data strongly implicate the involvement of inducible cytoskeletal-associated PPase 2B activity in the regulation of thrombin-mediated EC contractility and barrier function.