NaF, a potent G protein activator and Ser/Thr phosphatase inhibitor, significantly increased albumin permeability and decreased transcellular electrical resistance (TER), indicating endothelial cell (EC) barrier impairment. EC barrier dysfunction induced by NaF was accompanied by the development of actin stress fibers, intercellular gap formation, and significant time-dependent increases in myosin light chain (MLC) phosphorylation. However, despite rapid, albeit transient, activation of Ca2+/calmodulin-dependent MLC kinase (MLCK), the specific MLCK inhibitor ML-7 failed to affect NaF-induced MLC phosphorylation, actin cytoskeletal rearrangement, and reductions in TER, suggesting a limited role of MLCK in NaF-induced EC activation. In contrast, strategies to reduce Rho (C3 exoenzyme or toxin B) or to inhibit Rho-associated kinase (Y-27632 or dominant/negative RhoK) dramatically reduced MLC phosphorylation and actin stress fiber formation and significantly attenuated NaF-induced EC barrier dysfunction. Consistent with this role for RhoK activity, NaF selectively inhibited myosin-specific phosphatase activity, whereas the total Ser/Thr phosphatase activity remained unchanged. These data strongly suggest that MLC phosphorylation, mediated primarily by RhoK, and not MLCK, participates in NaF-induced EC actin cytoskeletal changes and barrier dysfunction.
- Rho-associated kinase
- myosin-specific phosphatase
- transendothelial electrical resistance
- actin cytoskeletal rearrangement
- myosin light chain
vascular endothelial cell (EC) monolayers serve as a semiselective barrier to fluid and solute flux across the blood vessel wall. Compromise of EC barrier integrity leads to an increase in vascular permeability, which is a cardinal feature of acute inflammatory lung injury (19,21). Our laboratory and others have previously shown barrier integrity to be critically dependent on the actomyosin-driven contraction, an event that involves the obligatory Ser/Thr phosphorylation of the 20-kDa regulatory myosin light chain (MLC20; see Refs. 16 and 26). For example, the coagulation protease α-thrombin induces EC intercellular gap formation and barrier dysfunction primarily via actomyosin-driven contraction. This process is initiated by myosin light chain kinase (MLCK)-mediated myosin light chain (MLC) phosphorylation, which is tightly linked to actin filament reorganization (16, 48).
At least two responsible classes of MLC kinases are potentially able to phosphorylate MLC in vivo, including the newly cloned endothelial Ca2+/calmodulin (CaM)-dependent MLCK (15, 16, 18, 26,53) and Rho-associated kinase (RhoK, p160 ROCK) (4,22), whose activity is dependent on activation of the small GTP-binding protein Rho. Rho has been implicated in the modulation of the F-actin cytoskeleton, focal adhesion formation, and cell shape (43). Rho GTPases are inactive in the GDP-bound form and are activated by GDP/GTP exchange, a process regulated by guanine nucleotide exchange factors and GTPase-activating proteins (11,47). Recently, several Rho GTPase-binding proteins were isolated. Among them, RhoK appears to function as a Rho effector mediating its action on the cytoskeleton (40). Recent data have indicated that RhoK significantly contributes to thrombin-induced stress fiber formation and EC barrier dysfunction (9, 13). It has been reported that activated RhoK modulates the actin cytoskeleton by LIM kinase-mediated inactivation of cofilin, an actin-severing protein, and by increasing MLC phosphorylation (3, 13, 35) via at least two known mechanisms (4,32). Once activated by Rho-GTP, RhoK phosphorylates the myosin phosphatase target subunit (MYPT1/2), thereby inhibiting myosin phosphatase activity (32). In addition, RhoK can directly phosphorylate MLC (4). Both mechanisms lead to an increase in MLC phosphorylation. Whether RhoK directly modulates the activity of EC MLCK, as suggested in some reports (22), has not been assessed.
Sodium fluoride (NaF) is a potent, rapid, and reversible activator of the regulatory heterotrimeric GTP-binding proteins in virtually all in vitro systems. It has been reported that the effect of NaF on heterotrimeric G proteins is the result of formation of AlF from fluoride and trace amounts of aluminum, which can come from contamination of glassware (10, 56). The cellular mechanism by which fluoroaluminates activate G proteins is based on the structural similarity of AlF to PO , enabling the former to interact with GDP situated on the α-subunit of the G proteins where it can mimic GTP (5). In C6 rat glioma cells, NaF activates Gsor inhibits Gi to increase the cAMP concentration and intracellular Ca2+ mobilization (31, 38) and enhances cGMP generation via Ca2+/CaM- and nitric oxide synthase-dependent pathways (38). Activation of G proteins by fluoride in smooth muscle cells can initiate a series of events, such as Ca2+ mobilization, phospholipase C activation, and protein kinase C activation (1, 17, 52, 56). Our laboratory has previously shown that, in vascular endothelium, NaF, like α-thrombin, activates a pertussis toxin-insensitive GTP-binding protein, which leads to increases in phosphoinositide hydrolysis, Ca2+ mobilization from intracellular stores, arachidonate release, and prostacyclin synthesis (17). It has also been reported that NaF has a contractile effect on airway smooth muscle mediated by both G protein-dependent and -independent pathways (30, 39). NaF is also a classical Ser/Thr phosphatase inhibitor (45) and is routinely included in extraction buffers to prevent dephosphorylation of proteins on Ser and Thr residues by endogenous phosphatases.
Because of its multiple effects on intracellular signaling, we used NaF as a valuable tool for exploring signaling pathways involved in EC barrier regulation. In the present study, we have demonstrated that NaF induced dramatic EC barrier dysfunction. Our data suggested that NaF-induced EC barrier compromise is tightly linked to RhoK-mediated MLC phosphorylation and actin cytoskeletal alterations. However, RhoK inhibition, which completely abolished MLC phosphorylation and stress fiber formation induced by NaF, significantly attenuated but did not fully prevent NaF-induced barrier dysfunction, suggesting the involvement of both MLC-dependent and -independent mechanisms in NaF-induced EC barrier dysfunction.
Unless specified, reagents were obtained from Sigma Chemical (St. Louis, MO). PBS, Hanks' balanced salt solution without phenol red, DMEM, medium 199, OPTI-MEM, phosphate-free DMEM, lipofectamine, antibiotic and antimycotic solution (10,000 U/ml penicillin, 10 mg/ml streptomycin, and 25 mg/ml amphotericin B), and nonessential amino acids were purchased from GIBCO (Grand Island, NY). EC growth supplement and RhoK inhibitor Y-27632 were purchased from Upstate Biotechnology (Lake Placid, NY). C3 exoenzyme and ML-7 were purchased from Calbiochem (La Jolla, CA). Toxin B was obtained from List Laboratories (Cambell, CA). MLC antibody was produced in rabbit against baculovirus-expressed and purified smooth muscle MLC by Biodesign International (Kennebunk, ME). Anti-phospho-MYPT1(Thr850) was purchased from Upstate Biotechnology. Rabbit anti-MYPT1 antibody was obtained from Covance (Princeton, NJ). Polyclonal rabbit anti-myc antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The transfection reagent fugen 6 was obtained from Roche Molecular Biochemicals (Indianapolis, IN). Protein G-Sepharose was purchased from Pharmacia Biotech (Piscataway, NJ). Protein concentration was determined using BCA protein assay reagent (Pierce, Rockford, IL). Texas red phalloidin and Alexa488anti-rabbit secondary antibody were purchased from Molecular Probes (Eugene, OR). All radioactive reagents were purchased from NEN (Boston, MA). Anti-MLCK D119 antibody was kindly provided by Dr. P. Gallagher (Indiana University, Indianapolis, IN). Diphospho-MLC antibody was kindly provided by Dr. M. Crow (National Institute of Aging, Baltimore, MD). Dominant negative RhoK construct was kindly provided by Dr. K. Kaibuchi (Nara Institute of Science and Technology, Nara, Japan).
Bovine pulmonary artery ECs obtained frozen at 16 passages from American Type Tissue Culture Collection (CCL 209; Rockville, MD) and used at passage 19–24 (16) were cultured in DMEM supplemented with 20% (vol/vol) FBS, 0.1% EC growth supplement, 1% antibiotic and antimycotic solution, and 1% nonessential amino acids. Human pulmonary artery ECs were purchased from Clonetics (Walkersville, MD), cultured in EBM-2 medium (Clonetics) supplemented with 20% (vol/vol) FBS, and used at passages 5–10. The cell cultures were maintained at 37°C in a humidified atmosphere of 5% CO2-95% air and grew to contact-inhibited monolayers. Cells from each primary flask were detached with 0.05% trypsin, resuspended in fresh culture media, and passaged into the appropriate-sized flasks or dishes.
Macromolecular permeability of cultured EC monolayers was performed as described previously (16).
Measurement of transendothelial electrical resistance.
The cellular barrier properties were measured using the highly sensitive biophysical assay with an electrical cell-substrate impedance sensing system (Applied Biophysics, Troy, NY) described previously (20, 44). Cells were cultured on small gold electrodes (10−4 cm2), and culture media were used as the electrolyte. The total electrical resistance was measured dynamically across the monolayer and was determined by the combined resistance between the basal surface of the cell and the electrode, reflective of focal adhesion, and the resistance between cells. As cells adhere and spread out on the microelectrode, the transmonolayer electrical resistance (TER) increased (maximal at confluence), whereas cell retraction, rounding, or loss of adhesion was reflected by a decrease in TER (25). The small gold electrodes and the larger counter electrodes (1 cm2) were connected to a phase-sensitive lock-in amplifier (5301A; EG&G Instruments, Princeton, NJ) with a built in differential preamplifier (5316A; EG&G Instruments). A 1-V, 4,000-Hz alternating current signal was supplied through a 1-MΩ resistor to approximate a constant-current source. Voltage and phase data were stored and processed with a Pentium 100-MHz computer that controlled the output of the amplifier and relay switches to different electrodes. Experiments were conducted only on wells that achieved >1,000 Ω (10 microelectrodes/well) of steady-state resistance. Resistance was expressed by the in-phase voltage (proportional to the resistance), which was normalized to the initial voltage and expressed as a fraction of the normalized resistance value, as previously described (20).
ECs grown on 60-mm dishes were analyzed for MLC phosphorylation by urea gel electrophoresis, as our laboratory has previously described in detail (16), followed by Western immunoblotting with specific anti-MLC antibody. The blots were scanned on a Bio-Rad densitometer, and the percentage of MLC phosphorylation was determined by dividing the total phosphorylated and unphosphorylated areas. This method takes advantage of the fact that the mono- and diphosphorylated forms of MLC migrate more rapidly than unphosphorylated MLC and are independent of sample loading. Stoichiometry (mol PO4/mol MLC) was calculated by the formula (P1 + 2P2)/(U + P1 + P2), where U is the percentage of unphosphorylated MLC, P1 is the percentage of monophosphorylated MLC, and P2 is the percentage of diphosphorylated MLC. The diphosphorylated MLC (P2) is multiplied by a factor of 2 to reflect the presence of two phosphate groups per molecule. The percentage of MLC phosphorylation was calculated by adding the densitometric values of each phosphorylation state for each isoform [i.e., unphosphorylated (U), monophosphorylated (P1), and diphosphorylated (P2)].
ECs from 100-mm dishes were treated with vehicle, 20 mM NaF, or 10 nM of calyculin for indicated time periods. After treatment, the cells were rinsed with ice-cold TBS (20 mM Tris · HCl, pH 7.6, and 137 mM NaCl) and 50 mM Tris · HCl, pH 7.0, 0.1 mM EDTA, and 28 mM β-mercapthoethanol (buffer A). Buffer A (400 μl) was added to the dishes, and the cells were quickly frozen at −70°C, scraped, and homogenized by passing the cell suspension several times through a 1-ml tuberculin syringe. Myosin-enriched fractions were prepared as our laboratory has described previously (50). Briefly, 500 μl of total cell lysates were incubated with 0.6 M NaCl and 0.1% Tween 20 at 4°C for 1 h with constant agitation followed by low-speed centrifugation (5,000 rpm) at 4°C for 30 min. The supernatant was diluted 10 times withbuffer A and centrifuged at 10,000 rpm at 4°C for 30 min. After centrifugation, the supernatant was removed carefully, and the myosin-enriched pellet was dissolved in buffer A containing 0.6 M NaCl and 0.1% Tween 20. Phosphatase activity was determined in a final volume of 20 μl using a Malachite green 96-well microtiter plate assay, as described previously (12, 28). Total cell lysate (1 μl) or the myosin-enriched fraction was incubated with 0.25 mM of the specific Ser/Thr phosphatase substrate phosphopeptide (KRpTIRR; Upstate Biotechnology) in assay buffer (50 mM Tris · HCl, pH 7.0, 0.1 mM EDTA, 28 mM β-mercapthoethanol, and 0.1 mg/ml BSA) at 30°C for 20 min. As a control, the cell suspension and phosphopeptide were incubated in a assay buffer containing Ser/Thr phosphatase inhibitors (1 mM EGTA and 5 μM okadaic acid). The reactions were terminated by addition of 100 μl of Malachite green solution (Upstate Biotechnology), and the plates were incubated at room temperature for 15 min before measurement of absorbance at 620 nm. The phosphate released in the enzyme reaction was determined by comparing the absorbance over control with the phosphate standard curve. The amount of proteins contained in the cell lysate was assessed by TCA precipitation followed by protein assay using the BCA protein assay kit (Pierce).
Transfection with dominant/negative Rho kinase construct.
ECs grown on glass coverslips to 50–60% confluence were transfected with plasmids encoding myc-tagged dominant/negative Rho-associated kinase using fugen 6 as a transfection reagent. Briefly, cells were incubated with 1 μg of plasmid DNA and 3 μl of fugen 6 in 1 ml of OPTI-MEM for 4 h. The solution was then replaced by 1 ml of complete culture media and incubated for 24 h. The transfected EC monolayers were stimulated with 20 mM NaF for the indicated time periods and visualized by immunofluorescent microscopy, as described below.
ECs grown on glass coverslips after treatment were fixed in 3.7% formaldehyde solution in PBS for 10 min at 4°C, washed three times with PBS, permeabilized with 0.2% Triton X-100 in PBS-Tween (PBST) for 5 min at 4°C, and blocked with 2% BSA in PBST for 20 min. Incubation with rabbit anti-myc antibody, diluted 1:100 with blocking solution, was performed for 1 h at room temperature. After three washes with PBS, cells were incubated with anti-rabbit secondary antibody conjugated with the fluorescent dye Alexa488 (Molecular Probes) for 1 h at room temperature. Actin filaments were visualized by staining cells with Texas red phalloidin (Molecular Probes) for 1 h at room temperature. The coverslips were mounted and analyzed with a Nikon video-imaging system consisting of a phase-contrast inverted microscope equipped with objectives and filters for immunofluorescence and connected to a digital camera and image processor. The images were recorded and saved on a Pentium II personal computer as TIFF files compatible with Adobe Photoshop 5.0 and National Institutes of Health image analysis programs.
Protein phosphorylation assay.
EC monolayers from 60-mm dishes were preloaded with [32P]orthophosphate (0.2 mCi/ml in phosphate-free DMEM) for 2 h at 37°C and then stimulated with 20 mM NaF for the indicated time periods. Stimulation was stopped by media aspiration, EC monolayers were rinsed two times with PBS, 200 μl 2× SDS sample buffer was added to each dish, cells were scraped, and cell homogenates were transferred to Eppendorf tubes. Cell extracts were boiled for 5 min and microcentrifuged for 5 min, and supernatant was used for 9% SDS-PAGE followed by electrotransfer to a nitrocellulose membrane. Membrane were exposed to X-ray film and then stained with specific anti-MLC antibody.
Permeabilization of ECs.
It is well known that Clostridium botulinum C3 exoenzyme does not readily traverse cell membranes under native conditions. To allow access of C3 exoenzyme to the cytosol, EC monolayers grown to 80–100% confluence were incubated with lipofectamine reagent (20 μg/ml) for 1 h, followed by the addition of C3 exoenzyme (2.5 μg/ml) for an additional 11 h, as described previously (6).
Viability assay was performed using the LIVE/DEAD viability/cytotoxicity kit (Molecular Probes) according to the manufacturer's protocol. This kit provides a two-color fluorescence cell viability assay that is based on the simultaneous dual determination of two recognized parameters of cell viability-intracellular esterase activity (calcein) and plasma membrane integrity [ethidium homodimer-1 (EthD-1)]. Calcein produced an intense green fluorescence in viable cells, whereas EthD-1 only entered cells with damaged membranes and bound to nucleic acids, thereby producing a bright red fluorescence in nonviable cells.
ANOVAs with a Student-Newman-Keuls test were used to compare means of clearance rates in permeability assay. Independent Student'st-test was performed to analyze the means and the differences between the means in MLCK assay, phosphatase assay, MLC phosphorylation assay, and TER experiments. Results are expressed as means ± SE. Differences in two groups are considered statistically significant at P < 0.05.
Effect of NaF on EC barrier dysfunction.
We initially assessed the effect of NaF on EC permeability by measuring the clearance of Evans blue-albumin across EC monolayers (Fig.1 A). Similar to the procoagulant serine protease thrombin, which induces EC shape changes, intercellular gap formation, and barrier dysfunction (21), NaF significantly increased albumin clearance (∼5-fold over control at a time period of 60–120 min). Utilizing a more sensitive index of EC barrier integrity, we next assessed the electrical resistance across EC monolayers grown on gold microelectrodes after NaF treatment (Fig. 1 B). Although 1 mM NaF failed to alter TER, a higher concentration of NaF (20 mM) caused a dramatic decrease in TER that began at ∼10 min of exposure after a brief increase in resistance and reached a maximum decline starting at ∼1 h. Viability assay revealed that cells were 100% viable after 2 h of NaF treatment (Fig.1 C). Increased cytotoxicity was evident only after prolonged incubation with NaF (>4 h). These data demonstrated that NaF induced dramatic EC barrier dysfunction in the absence of cytotoxicity.
Effect of NaF on EC MLC phosphorylation and cytoskeletal reorganization.
Several studies, including our own, have previously shown that EC barrier dysfunction is tightly linked to myosin-driven contraction initiated by MLC phosphorylation (16, 54, 55). We next examined the effect of NaF (1 or 20 mM) on EC MLC phosphorylation. At specified times, the NaF treatment was quickly terminated with ice-cold TCA, and the level of MLC phosphorylation was analyzed by urea gel electrophoresis followed by immunoblotting with MLC specific antibody (Fig. 2). NaF (20 mM) caused a time-dependent increase in MLC phosphorylation, with a maximum of an approximately sevenfold increase compared with control values (from a stoichiometry of ∼0.25 mol PO4/mol MLC to ∼1.80 mol PO4/mol MLC at 60 min). In contrast, treatment with 1 mM NaF failed to significantly affect EC MLC phosphorylation.
EC contraction is reflected by changes in the actin cytoskeleton. Figure 3 depicts immunofluorescent images of ECs double stained for F-actin and diphosphorylated MLC using Texas red phalloidin and a diphosphospecific MLC antibody. Under control conditions, ECs form a compact monolayer with the actin filament distributed primarily in a cortical ring with few stress fibers spanning the cells. Treatment with 20 mM NaF caused time-dependent cytoskeletal reorganization characterized by the early (10 min) formation of stress fibers, but, importantly, the majority of F-actin and diphospho-MLC remained in the peripheral cortical actin rings. Subsequent changes in the cytoskeleton (30 and 60 min) revealed dissolution of the cortical rings and prominent formation of stress fibers (Fig. 3, A-E). These significant cytoskeletal rearrangements reflected cell contraction and led to intercellular gap formation (shown by arrows in Fig. 3), indicating EC barrier compromise. In contrast, 1 mM NaF failed to cause any significant changes in the actin cytoskeleton (unpublished observation). Note that the intensity of diphospho-MLC staining was increased over the time that is well correlated with the observation of increased MLC phosphorylation after NaF based on biochemical analysis (Fig. 2). Our data demonstrate that increases in MLC phosphorylation and profound cytoskeletal alterations precede NaF-induced EC barrier dysfunction.
Role of Ca2+/CaM-dependent MLCK and intracellular Ca2+ in NaF-induced EC activation.
MLC phosphorylation status is believed to be determined by the balanced activities of MLC kinases (Ca2+/CaM-dependent MLCK and Rho-associated kinase) and myosin phosphatases. We next examined the role of these enzymes in NaF-induced EC MLC phosphorylation, actin filament reorganization, and barrier dysfunction. First, we determined the effect of NaF on EC MLCK activity present in MLCK immunoprecipitates. NaF caused a transient increase in MLCK activity (∼2-fold over control) at 2 min followed by a significant decrease (∼50% inhibition) after 30 min (Fig.4 A). To further test whether MLCK is involved in NaF-induced EC activation, we employed the specific MLCK inhibitor ML-7 to block MLCK activity and demonstrated that ML-7 failed to significantly alter the MLC phosphorylation induced by NaF, whereas thrombin-induced MLC phosphorylation was abolished (Fig.4 B). Consistent with these data, pretreatment with ML-7 did not affect NaF-induced declines in electrical resistance across EC monolayers but attenuated the effect of thrombin (Fig. 4 C). Immunofluorescent staining for F-actin showed that ML-7 slightly reduced the amount of actin filament in control cells but failed to significantly affect stress fiber formation induced by NaF (Fig.5, A-D). Taken together, these data suggest that Ca2+/CaM-dependent MLCK does not significantly contribute to NaF-induced EC MLC phosphorylation, actin filament reorganization, and barrier dysfunction, despite a rapid and transient activation of MLCK.
Our laboratory has previously shown that NaF increases intracellular Ca2+ in endothelium (17). To test whether other Ca2+-dependent mechanisms are involved in NaF-induced barrier dysfunction, we employed the cell-permeable Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)-AM and demonstrated that blockage of intracellular Ca2+ elevation failed to significantly affect the NaF-induced decrease in TER (Fig. 6). In contrast, the thrombin-induced decrease in TER, which involves Ca2+ signaling (16), was significantly attenuated by BAPTA-AM. Our data suggest that Ca2+signaling does not appear to be a major participant in NaF-induced barrier dysfunction.
Role of Rho and Rho-associated kinase in NaF-induced EC activation.
Rho has been reported to regulate cell contraction and actin cytoskeletal changes in a Ca2+-independent manner in smooth muscle and nonmuscle cells (22, 43). We next examined the role of Rho in NaF-induced MLC phosphorylation using toxin B, an inhibitor of the Rho family G proteins (Rho, cdc42, and Rac), and C3 exoenzyme, a specific Rho inhibitor. Both agents significantly attenuated NaF-induced MLC phosphorylation (Fig.7 A), and C3 exoenzyme attenuated the NaF-induced reduction in TER (Fig. 7 B), indicating the involvement of small G protein Rho in EC MLC phosphorylation and barrier dysfunction induced by NaF.
Next we examined the role of RhoK, a downstream effector of Rho and a possible MLC kinase, in NaF-induced EC activation. To evaluate RhoK activation, we immunoprecipitated the myosin phosphatase target subunit (MYPT1), a substrate for RhoK, followed by Western blotting with anti-phospho-MYPT1 antibody directed at Thr850, one of the major sites of phosphorylation by RhoK (14). NaF induced time-dependent phosphorylation of MYPT1 at Thr850 (Fig.8 A), which can be completely blocked by the RhoK inhibitor Y-27632 (Fig. 8 B), suggesting the specificity of the phosphorylation by RhoK. These data indicate the activation of RhoK upon NaF stimulation. Next, we examined the effect of RhoK inhibition by Y-27632 on NaF-induced EC activation. Y-27632 decreased the basal level of MLC phosphorylation and completely abolished NaF-induced MLC phosphorylation (Fig.9 A). Consistent with these data, treatment with Y-27632 significantly attenuated NaF-induced declines in transendothelial electrical resistance (Fig. 9 B). In fact, in the presence of RhoK inhibition, NaF produced sustained (∼1 h) and significant enhancement of barrier function, as reflected by TER, suggesting dual activation of barrier protective signaling pathways. Immunofluorescent staining for F-actin showed that Y-27632 markedly reduced the amount of actin stress fibers and simultaneously increased the amount of cortical actin in both control and NaF-stimulated cells (Fig. 5, A,B, E, and F). Consistent with these data, overexpression of dominant/negative RhoK in transfected cells completely abolished stress fiber formation in control and NaF-challenged cells (Fig. 9 C). Taken together, these data suggest the involvement of RhoK in NaF-induced EC MLC phosphorylation, actin cytoskeletal reorganization, and barrier dysfunction.
Effect of NaF on myosin phosphatase activity.
Myosin phosphatase is regulated by RhoK and is the important determinant of the level of MLC phosphorylation. Therefore, we next examined the effect of NaF on total phosphatase activity in cell homogenates (Fig. 10 A) and on the myosin-specific phosphatase activity in the myosin-enriched fraction (Fig. 10 B). In contrast to calyculin, another Ser/Thr phosphatase inhibitor, NaF did not affect total phosphatase activity but significantly decreased myosin-specific phosphatase activity (∼50% inhibition at 60 min). Consistent with these data, NaF stimulation did not serve to reduce phosphatase activities but rather resulted in overall declines in total protein phosphorylation of EC homogenates with specific time-dependent increases in the phosphorylation of several specific proteins in intact EC, including MLC (Fig. 11). These data suggest that effects of NaF on EC activation are not the result of nonspecific phosphatase inhibition but rather indicate a role of myosin-specific phosphatase inhibition in NaF-induced MLC phosphorylation.
In the present study, we have demonstrated that NaF causes dramatic EC barrier dysfunction, as evidenced by increases in macromolecule permeability and decreases in transendothelial electrical resistance. Our laboratory has previously reported that Ca2+/CaM-dependent MLCK-mediated MLC phosphorylation is critically involved in EC barrier dysfunction induced by permeability producing agonists such as thrombin and diperoxovanadate (16, 18,22, 48). Other models of EC barrier dysfunction exist, however, that occur independently of MLCK-mediated MLC phosphorylation. For example, phorbol esters (16, 49) and pertussis toxin (41) significantly increased EC permeability without a rise in intercellular Ca2+ and without evidence of increased MLCK activity. Here, we now reported yet another model of EC permeability whereby NaF-induced barrier dysfunction is critically dependent on MLC phosphorylation mediated by Rho/RhoK, with an only minor contribution (if any) for increased Ca2+/CaM-dependent MLCK activity. NaF induced significant but transient MLCK activation (peak at 2 min) accompanied by a slight increase in MLC phosphorylation at a very early time point (2 min), a temporal sequence that correlates with our previous reports of NaF-mediated increases in intracellular Ca2+ (onset at 1–2 min, peak at 8 min; see Ref. 17). However, despite sustained increase in Ca2+, MLCK activity declined thereafter, with significant inhibition after 30 min, consistent with prior work that, in ECs, a rise in cytosolic calcium, while necessary, is insufficient to evoke MLCK activation (20). Our data using the specific MLCK inhibitor ML-7 showed that MLCK inhibition failed to significantly affect NaF-induced MLC phosphorylation (even at 2 min when MLCK was maximally activated; unpublished observation), actin cytoskeletal changes, and barrier dysfunction, suggesting the limited role of MLCK in NaF-induced activation of the EC contractile apparatus.
There is now considerable evidence suggesting the important role of Rho/RhoK in smooth muscle and nonmuscle cell contraction that is initiated by RhoK-mediated MLC phosphorylation (3, 4, 13,35). We used several complementary approaches to demonstrate that NaF-induced EC barrier dysfunction is tightly linked to Rho/RhoK-mediated MLC phosphorylation and actin cytoskeletal alterations. For example, both toxin B, an inhibitor of small G proteins of the Rho family (Rho, cdc42, and Rac), and C3 exoenzyme, a specific Rho inhibitor, significantly attenuated NaF-induced MLC phosphorylation. The specific pharmacological inhibitor RhoK (Y-27632) completely abolished NaF-induced MLC phosphorylation and actin cytoskeletal changes and significantly attenuated NaF-induced declines in transendothelial electrical resistance. Furthermore, EC transfection with dominant/negative RhoK prevented the formation of stress fibers elicited by NaF.
RhoK increases MLC phosphorylation by two known mechanisms. One mechanism involves RhoK-induced phosphorylation of the myosin phosphatase target subunit MYPT1 that inhibits myosin phosphatase activity and thereby ceases MLC dephosphorylation (32). Rho kinase can also directly phosphorylate MLC, which has been shown to occur at the same sites phosphorylated by MLCK (4). To demonstrate RhoK activation in NaF-stimulated endothelium, we employed antibody specific to Thr850 phosphorylated by RhoK in the structure MYPT1. NaF induced time-dependent phosphorylation of MYPT1 at Thr850 that occurred as early as 5 min. Note that RhoK phosphorylates two sites (Thr695 and Thr850 of MYPT1). However, phosphorylation of Thr695 but not Thr850 was shown to be responsible for the myosin phosphatase inhibition (14). This could explain the apparent discrepancy between the early MYPT phosphorylation (at Thr850) and the late myosin phosphatase inhibition (at 60 min) demonstrated by biochemical assays. Thus MLC phosphorylation at the early time point appears to be linked to the RhoK activity directed toward MLC, whereas the RhoK-mediated phosphatase inhibition provides an additional mechanism leading to an increase in MLC phosphorylation at the later time point (60 min).
In our model of vascular barrier regulation, NaF induces an initial increase (maximum of 10–15% increase at 5–10 min) in TER followed by a dramatic decrease (∼50–60%) that begins at ∼10 min and reaches a maximum decline at ∼1 h. Studies on the subcellular distribution of phosphorylated MLC revealed that at 10 min the intensity of diphosphorylated-MLC staining is increased, consistent with our data obtained by biochemical assays, with the majority of actin and as diphosphorylated-MLC confined to the cortical rings. However, after 30 min, the cortical rings completely disappeared, and phosphorylated MLC colocalized with the newly developed stress fibers. These data are consistent with our observation that increases in MLC phosphorylation may regulate transendothelial electrical resistance (increases and decreases), depending on the spatial localization of MLC phosphorylation. For example, our unpublished data revealed that endothelial differentiation gene receptor ligation by sphingosine 1-phosphate induces increased transendothelial electrical resistance that is associated with increased amounts of actin and phosphorylated MLC in cortical rings. In contrast, thrombin caused a rapid and dramatic decrease in resistance, loss of cortical actin, and significant cytoplasmic stress fiber formation that is colocalized with phosphorylated MLC. Delayed contraction and stress fiber formation in the case of NaF may be explained by alternate effects of NaF. It has been reported that fluoride is an inhibitor of enolase, an enzyme in the glycolysis pathway leading to phosphoenolpyruvate (37). Because both MLC phosphorylation and actin-mediated activation of myosin ATPase are dependent on ATP, a product of glycolysis and oxidative metabolism, inhibition of this enzyme would be expected to reduce glycolytic ATP production and impair cell contraction. Zhao and Guenard (57) demonstrated that NaF inhibited carbachol-induced bovine bronchial contraction via its action on glycolysis, although NaF alone induced delayed (11.3 ± 0.7 min) contraction. Thus NaF appears to induce either contraction or relaxation of smooth muscle, depending on the particular conditions. This provides a potential explanation for the delayed NaF-induced contraction noted in our model that may occur later than MLC phosphorylation.
The early upstream signaling events by which NaF induces Rho/RhoK activation and subsequent MLC phosphorylation are unknown. Many of the cellular actions of NaF are thought to be attributed to fluoride complexing with aluminum to form fluoaluminates, which activate heterotrimeric G proteins (56). However, in our studies, incubation with deferoxamine, a cation chelator, at concentrations (100 μM) that completely chelate contaminating aluminum, did not change the pattern of NaF-induced declines in transendothelial electrical resistance (unpublished observation). These data do not exclude NaF-mediated G protein activation as the cause of NaF-induced EC barrier dysfunction but rather call into question the role of fluoroaluminate-mediated G protein activation. It has been reported that NaF inhibits tyrosine kinase activity of insulin receptors purified from rat skeletal muscle and human placenta by direct binding to the receptors (51). Our laboratory has previously reported that NaF increases intracellular Ca2+ and prostacyclin synthesis via the activation of pertussis toxin-insensitive GTP-binding proteins (17). The exact G proteins involved in NaF-mediated barrier regulation are unknown; however, G proteins of the G12/13 family have been shown to regulate Rho/RhoK-dependent signaling processes in smooth muscle and nonmuscle cells, including endothelium (8, 27). Further work is required to determine whether NaF binds to receptors or directly activates G12/13 to exert its cellular actions.
Our data suggest that the RhoK/MLC-dependent signaling pathways do not fully explain NaF-induced EC barrier dysfunction, since Y-27632 significantly reduced but did not fully prevent NaF-induced EC barrier dysfunction despite complete attenuation of NaF-induced MLC phosphorylation. There is cumulative evidence which suggests that other cytoskeletal proteins may participate in cell contraction. In both smooth muscle and nonmuscle cells, caldesmon, an actin-, myosin-, and Ca2+/CaM-binding protein, regulates actomyosin interactions (2, 23, 36, 46). In vitro studies indicate that, at low Ca2+ concentrations, caldesmon binding to actin filaments inhibits both myosin ATPase activity and actin-myosin interaction that can be reversed by either increased Ca2+/CaM availability or caldesmon phosphorylation (2, 24, 33, 36, 42, 46). Another actin-binding protein, filamin, has been implicated in regulation of cell shape, locomotion, and actin cytoskeleton organization (29). Our recent data have indicated that filamin is involved in thrombin-induced cytoskeletal reorganization and barrier dysfunction (7). It seems plausible that filamin, caldesmon, or other regulatory cytoskeletal proteins may provide attractive alternate mechanisms for MLC-independent regulation of contractile forces. Besides, the regulation of cell-cell and cell-matrix tethering forces, whose impairment also leads to increases in macromolecular EC permeability (7, 34, 44), can contribute to NaF-induced barrier dysfunction.
In summary, the effects of NaF on endothelial barrier function and the contributory signaling pathways were investigated. Our data suggest that NaF-induced EC barrier dysfunction is tightly linked to actin cytoskeletal changes and MLC phosphorylation. Unlike thrombin, MLCK activity is not a significant contributor to NaF-induced MLC phosphorylation, actin stress fiber formation, and barrier dysfunction. In contrast, NaF-mediated MLC phosphorylation was dramatically reduced by both Rho inhibition and RhoK inhibition, which also significantly attenuated NaF-induced actin cytoskeletal changes and EC barrier dysfunction. In fact, in the presence of RhoK inhibition, NaF produced sustained (∼1 h) and significant enhancement of barrier function, as reflected by TER, suggesting dual activation of barrier protective signaling pathways. Our data also suggest that effects of NaF on EC activation are not the result of nonspecific phosphatase inhibition but rather indicate a role of myosin-specific phosphatase inhibition in NaF-induced MLC phosphorylation. Taken together, our data strongly suggested that Rho/RhoK- but not MLCK-mediated MLC phosphorylation participates in NaF-induced EC barrier dysfunction.
We thank Lakshmi Natarajan for technical assistance. Special appreciation is extended to Drs. K. Kaibuchi (Nara, Japan), M. Crow (Baltimore, MD), and P. Gallagher (Indianapolis, IN) for providing important reagents for this work.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-44746, HL-50533, HL-58064, HL-67307, and HL-68062 and by the American Heart Association.
Address for reprint requests and other correspondence: J. G. N. Garcia, Johns Hopkins Asthma and Allergy Center, 5501 Hopkins Bayview Circle, 4B.77, Baltimore, MD 21224-6801 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society