Cytoskeletal Reorganization Evoked by Rho-associated kinase- and Protein Kinase C-catalyzed Phosphorylation of Cofilin and Heat Shock Protein 27, Respectively, Contributes to Myogenic Constriction of Rat Cerebral Arteries*

Background: The myogenic response of cerebral arteries to intravascular pressure regulates blood flow to the brain. Results: Pressurization reduced smooth muscle G-actin and increased phospho-cofilin and -HSP27 content by a mechanism blocked by ROK or PKC inhibitors. Conclusion: ROK- and PKC-mediated control of cofilin and HSP27 contributes to actin polymerization in myogenic constriction. Significance: Knowledge of cytoskeletal dynamics is crucial for understanding myogenic control of cerebral arterial diameter. Our understanding of the molecular events contributing to myogenic control of diameter in cerebral resistance arteries in response to changes in intravascular pressure, a fundamental mechanism regulating blood flow to the brain, is incomplete. Myosin light chain kinase and phosphatase activities are known to be increased and decreased, respectively, to augment phosphorylation of the 20-kDa regulatory light chain subunits (LC20) of myosin II, which permits cross-bridge cycling and force development. Here, we assessed the contribution of dynamic reorganization of the actin cytoskeleton and thin filament regulation to the myogenic response and serotonin-evoked constriction of pressurized rat middle cerebral arteries. Arterial diameter and the levels of phosphorylated LC20, calponin, caldesmon, cofilin, and HSP27, as well as G-actin content, were determined. A decline in G-actin content was observed following pressurization from 10 mm Hg to between 40 and 120 mm Hg and in three conditions in which myogenic or agonist-evoked constriction occurred in the absence of a detectable change in LC20 phosphorylation. No changes in thin filament protein phosphorylation were evident. Pressurization reduced G-actin content and elevated the levels of cofilin and HSP27 phosphorylation. Inhibitors of Rho-associated kinase and PKC prevented the decline in G-actin; reduced cofilin and HSP27 phosphoprotein content, respectively; and blocked the myogenic response. Furthermore, phosphorylation modulators of HSP27 and cofilin induced significant changes in arterial diameter and G-actin content of myogenically active arteries. Taken together, our findings suggest that dynamic reorganization of the cytoskeleton involving increased actin polymerization in response to Rho-associated kinase and PKC signaling contributes significantly to force generation in myogenic constriction of cerebral resistance arteries.

ylation. Inhibitors of Rho-associated kinase and PKC prevented the decline in G-actin; reduced cofilin and HSP27 phosphoprotein content, respectively; and blocked the myogenic response. Furthermore, phosphorylation modulators of HSP27 and cofilin induced significant changes in arterial diameter and G-actin content of myogenically active arteries. Taken together, our findings suggest that dynamic reorganization of the cytoskeleton involving increased actin polymerization in response to Rho-associated kinase and PKC signaling contributes significantly to force generation in myogenic constriction of cerebral resistance arteries.
Blood flow to the brain is regulated through the interplay of multiple mechanisms that affect contractile force generation by vascular smooth muscle cells (VSMCs) 5 within the walls of cerebral resistance arteries and arterioles (1)(2)(3)(4). Flow can be matched to metabolic demand under dynamic physiological conditions because these vessels develop contractile tone in response to intravascular pressure, referred to as the myogenic response (2). Myogenic tone development allows resistance vessels to achieve a state of partial constriction from which they can dilate or further constrict to alter flow in response to factors in the blood or released from surrounding cells, such as endothelial cells, neurons, and astrocytes (1)(2)(3). Our understanding of the molecular basis of myogenic constriction and how it is regulated by these vasoactive factors is incomplete.
A core concept of smooth muscle contraction is that crossbridge cycling and force generation are a consequence of thick filament regulation involving phosphorylation of 20-kDa myosin regulatory light chains (LC 20 ), with the level of phospho-LC 20 dependent on the balance of myosin light chain kinase (MLCK) and phosphatase (MLCP) activities (5)(6)(7)(8). Activation of actomyosin MgATPase activity and cross-bridge cycling in VSMCs is initiated by phosphorylation of LC 20 at Ser-19 by MLCK. MLCK activity is (Ca 2ϩ ) 4 -calmodulin-dependent, and stimulated by a rise in cytosolic free Ca 2ϩ concentration ([Ca 2ϩ ] i ) caused by membrane potential (E m ) depolarization and Ca 2ϩ influx, or release from internal sarcoplasmic reticulum Ca 2ϩ stores (5,9). In contrast, LC 20 dephosphorylation by MLCP is inhibited by Rho-associated kinase (ROK)-mediated phosphorylation of the MLCP targeting subunit MYPT1 (10 -12) and PKC-mediated phosphorylation of the 17-kDa PKCpotentiated protein phosphatase 1 inhibitor protein, CPI-17 (13).
Appropriate regulation of MLCK and MLCP activities is necessary for the myogenic response. Myogenic constriction is accompanied by: (i) E m depolarization leading to Ca 2ϩ influx through voltage-gated Ca 2ϩ channels (14), increased Ca 2ϩ wave activity (i.e. increased frequency and number of cells exhibiting waves (15), but see (16)), and increased LC 20 phosphorylation (17)(18)(19) that is suppressed by the MLCK inhibitor, ML-7 (17); as well as (ii) increased inhibitory MYPT1-T855 phosphorylation that is suppressed by blockers of ROK activity, including Y27632 and H1152, with an associated loss of myogenic constriction (18,19). Although inhibition of PKC suppresses myogenic constriction (20,21), no change in CPI-17 phosphorylation was detected, excluding a role for PKC-mediated suppression of MLCP activity (18,19).
Although necessary for myogenic constriction, accumulating evidence indicates that mechanisms of MLCK and MLCP regulation are not sufficient (22). Additional mechanisms modulating force development include thin filament regulation (23) and dynamic reorganization of the actin cytoskeleton (24,25). Calponin and caldesmon are thin filament-associated proteins that regulate smooth muscle contractility by directly inhibiting cross-bridge cycling (23,26,27). Phosphorylation of calponin reduces its inhibitory effect on cross-bridge cycling (28) and may account for slow cross-bridge cycling and contraction of smooth muscle cells in the absence of extracellular Ca 2ϩ (29) and when LC 20 phosphorylation is precluded by site-directed mutagenesis (30). Phosphorylation of caldesmon has similarly been linked to force generation in the absence of extracellular Ca 2ϩ (31) and increased LC 20 phosphorylation (32). Whether phosphorylation of calponin and/or caldesmon contributes to the myogenic response is unknown.
Several lines of evidence indicate that the cytoskeleton of smooth muscle cells is not a static structure; rather it is dynamically remodeled during contraction and relaxation (22,24,25,33). Evidence of increased actin polymerization in the myogenic response was obtained for rat cerebral and tail arterial preparations (19, 34 -37). Also, the myogenic response was affected by compounds that disrupt (cytochalasin D), prevent (latrunculin A or B), or enhance (jasplakinolide) actin polymerization (19, 34 -38).
Actin filaments of the contractile apparatus are believed to be anchored to the cytoplasmic tails of integrins by a complex of membrane adhesion proteins and to each other at cytosolic dense bodies within the interior of VSMCs (25). Integrins, membrane adhesion complexes, and dense bodies are connected to each other and reinforced by the cortical actin cytoskeleton, and together these elements are thought to distribute and transmit force generated by the contractile apparatus over the cell membrane and to the extracellular matrix. Dynamic reorganization of the actin cytoskeleton during smooth muscle contraction is thought to involve expansion of adhesion protein complexes, severing of existing filaments to provide nucleation sites, and de novo actin polymerization of globular to filamentous actin (G-and F-actin, respectively) within the cortical actin network beneath the cell membrane (25,39). This is postulated to strengthen the cytoskeleton and enhance force transmission from the contractile apparatus to the cell membrane and extracellular matrix (24,25). Actin dynamics evoked by exposure to vasoconstrictor agonists involves a spatially restricted pool of nonmuscle ␥and ␤-actin within the cell cortex and at focal adhesions and dense bodies, but not smooth muscle ␣-actin that interacts with tropomyosin and forms the contractile actin filaments within the core of VSMCs (40,41). Roles for additional cytoskeletal and membrane adhesion proteins in smooth muscle contraction have also been identified, including proteins such as talin that serve a structural role, and others that facilitate actin dynamics, such as the actin nucleation initiating factor, N-WASp (neuronal Wiskott-Aldrich Syndrome protein), cofilin, HSP20, HSP27, and VASP (vasodilator-stimulated phosphoprotein) (42)(43)(44)(45)(46)(47).
Here, we employed pressurized rat cerebral arteries to assess the role of thin filament regulation and the contribution and regulation of actin polymerization in the myogenic response. No role for calponin or caldesmon regulation was detected. Force generation caused by actin polymerization was detected independent of a change in LC 20 phosphorylation and crossbridge cycling in the myogenic response and in vasoconstrictor-evoked contraction of myogenic arteries, accounting for ϳ30% of myogenic tone at 120 mm Hg. Evidence supporting a role for two regulators of actin dynamics, cofilin and HSP27, that are known to be regulated by ROK and PKC, respectively, was obtained.

EXPERIMENTAL PROCEDURES
Ethical Approval-Sprague-Dawley rats (250 -275 g; Charles River, Montreal, Canada) were maintained and killed by halothane inhalation and exsanguination according to a protocol approved by the Animal Care Committee of the Faculty of Medicine of the University of Calgary and conforming to the standards of the Canadian Council on Animal Care. A total of 210 rats were employed.
The endothelial layer was disrupted by passing a stream of fine air bubbles through the vessel lumen and confirmed by the loss of vasodilatation to 10 M bradykinin. External arterial diameter was measured using edge detection software (IonOptix, Milton, MA). Arteries were equilibrated at 10 mm Hg in warm Krebs' solution (37 Ϯ 0.5°C) for 15 min at the beginning of each experiment prior to an elevation of intraluminal pressure to 60 mm Hg to permit development of myogenic tone within 5-20 min. Arteries were then subjected to two 5-min pressure steps from 20 to 80 mm Hg to evoke a stable level of constriction (vessels exhibiting leaks or a lack of myogenic constriction were discarded). Intraluminal pressure was then set at 10 mm Hg for 10 min prior to the start of each experiment (pressure protocols for individual experiments are given under "Results"). Vessel segments from the same animal were used for control and treatment groups whenever possible to minimize vessel to vessel variability in phosphoprotein and G-actin quantification.
Circumferential Wall Stress Calculation-The circumferential wall stress (CWS) was calculated at radius r within the arterial wall using the following equation, from Coulson et al. (48), developed for a hollow cylinder subjected to uniform pressure where r is the radius, r i is the internal radius, r e is the external radius, and p i is the internal pressure, respectively. Protein Extraction-Vessels were frozen by rapid transfer to an ice-cold mixture of 10% trichloroacetic acid and 10 mM DTT in acetone prior to washing in acetone containing 10 mM DTT, lyophilization overnight, and storage at Ϫ80°C. Prior to protein extraction, the cannulated ends were removed from each lyophilized vessel segment to avoid inclusion of tissue not subjected to test pressures prior to protein extraction in 60 l of sample buffer (4% SDS, 100 mM DTT, 10% glycerol, 0.01% bromphenol blue, 60 mM Tris-HCl, pH 6.8). Each sample contained material from two to four pooled RMCA segments depending on the protein(s) of interest. Samples were heated at 95°C for 10 min and rotated overnight at 4°C prior to gel electrophoresis.
G-actin Determination-G-actin content was determined for individual RMCA segments as previously described (50). Vessels were rapidly transferred to F-actin stabilization buffer (Cytoskeleton, Denver, CO) containing 50 mM PIPES (pH 6.9), 50 mM KCl, 5 mM MgCl 2 , 5 mM EGTA, 5% (v/v) glycerol, 0.1% Nonidet P40, 0.1% Triton X-100, 0.1% Tween 20, 0.1% 2-mer-captoethanol, 0.001% antifoam C and then homogenized in 100 l of stabilization buffer at room temperature. The homogenate was then centrifuged at 155,000 ϫ g for 1 h at 22°C to separate G-and F-actin; 30 l of the high speed supernatant containing G-actin was carefully removed and added to 30 l of 2ϫ sample buffer. Samples were then heated at 95°C for 10 min and stored at Ϫ20°C prior to SDS-PAGE and Western blotting using a standard two-step protocol. SDS-PAGE was carried out in 1.5-mm-thick mini-gels (10% acrylamide in the resolving gel with 4.5% acrylamide stacking gel) at 30 mA for 1.5 h in a Mini Protean Cell (Bio-Rad). Following electrophoresis, proteins were transferred to a 0.2-m nitrocellulose membrane at 100 V for 90 min at 4°C, in transfer buffer containing 25 mM Tris-HCl, 192 mM glycine, and 20% methanol. Blotted membranes were washed in PBS for 5 min, incubated in 0.5% glutaraldehyde in PBS for 15 min to fix proteins on the membrane, and washed (twice for 5 min) with Tris-buffered saline containing Tween 20 (TBST: 20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 3 mM KCl, 0.05% Tween 20). Membranes were blocked with 5% nonfat dried milk in TBS and 0.1% Tween 20 (0.1% TBST) for 1 h. After blocking, each membrane was cut at the 35-kDa molecular mass marker; the high molecular mass proteins were incubated with rabbit polyclonal anti-actin (1:1000 dilution), and the low molecular mass proteins were incubated with goat polyclonal anti-SM22␣ (1:2000 dilution). Both antibodies were incubated overnight at 4°C in 1% dry milk in 0.1% TBST. Membranes were washed (four times for 5 min) in 0.1% TBST and incubated for 1 h in 1% dry milk and 0.1% TBST containing anti-rabbit IgG-HRP-conjugated secondary antibody (1:10,000 dilution) or donkey anti-goat IgG-HRP-conjugated secondary antibody (1:5000 dilution), respectively. After incubation with secondary antibody, membranes were washed (five times for 5 min) with TBST and (once for 5 min) with TBS before chemiluminescence signal detection using Amersham Biosciences ECL advanced Western blotting detection kit. The emitted light was detected and quantified with a chemiluminescence imaging analyzer (LAS3000 Mini; Fujifilm Canada, Mississauga, Canada), and images were analyzed with MultiGauge v3.0 software (Fujifilm Canada). Levels of G-actin were normalized to SM22 content, because SM22 is exclusively retained in the supernatant during high speed centrifugation (50) and expressed as a fraction of the level at 10 mm Hg.
Statistical Analysis-All values are presented as means Ϯ S.E., with n values indicative of the number of vessels, or in the case of phosphoprotein and G-actin analysis, the number of samples (with two vessels pooled or a single vessel per sample, respectively) studied for each treatment. Statistical difference was determined using paired or unpaired Student's t test, analysis of variance, or repeated measures analysis of variance followed by Bonferoni's post hoc test as required. A p value of Ͻ 0.05 was considered to be statistically significant. 20 , Calponin, or Caldesmon Phosphorylation-RMCA constriction under conditions in which a detectable change in the level of LC 20 phosphorylation was not apparent was specifically examined to assess the contribution of thin filament regulatory phosphoproteins and cytoskeletal reorganization to force development. Phospho-LC 20 content of RMCAs pressurized to 100 mm Hg was previously shown to be ϳ50% (18), consistent with the maximal level of 50 -55% detected for rat skeletal muscle arterioles pressurized to between 80 and 140 mm Hg (19). Based on these data, the level of phospho-LC 20 was quantified using RMCA segments flash frozen at 10, 80, and 120 mm Hg once vessel diameter had stabilized at the relevant intraluminal pressure (this required 30 s to 3 min in different vessels). Fig. 1A shows a representative recording of arterial diameter during sequential pressure steps from 10 to 80 and 120 mm Hg. Pressure elevation from 80 to 120 mm Hg was associated with a reduction in the diameter of 17 Ϯ 2.8 m (n ϭ 9 vessels; p Ͻ 0.05). Fig. 1B shows a representative Western blot of phosphorylated and unphosphorylated LC 20 at 10, 80, and 120 mm Hg obtained using Phos-tag TM gels, and mean phospho-LC 20 content was expressed as a percentage of total LC 20 at 10, 80, and 120 mm Hg (Ϯ S.E.; n ϭ 14, 9, and 7 arterial segments, respectively). No difference in phospho-LC 20 content was observed for segments at 80 and 120 mm Hg, with levels of ϳ50% detected at both pressures (p Ͼ 0.05). Fig. 1 (C and D) shows representative Western blots and mean (Ϯ S.E.) levels of calponin and caldesmon phosphorylation at 80 and 120 mm Hg, respectively. No difference in calponin or caldesmon phosphorylation was evident at 80 and 120 mm Hg (n ϭ 3 per group; p Ͼ 0.05 for both). In contrast, a significant reduction in G-actin content was detected. Fig. 1E shows representative blots of G-actin and SM22 content at 10, 80, and 120 mm Hg, as well as the mean level (Ϯ S.E.) of G-actin expressed as a fraction of the content at 10 mm Hg for vessels at 10, 80, and 120 mm Hg. Increasing pressure from 10 to 40 mm Hg did not evoke a significant change in G-actin content (data not shown), but a step from 10 to 80 caused G-actin levels to decline by ϳ66%, and a step from 80 to 120 mm Hg caused a further decline in G-actin content of ϳ21% (n ϭ 8 for 10, 80, and 120 mm Hg, respectively; p Ͻ 0.05). Thus, the increased force generation at 120 compared with 80 mm Hg was associated with a reduction in G-ac-tin content but no change in LC 20 or thin filament regulatory protein phosphorylation.

Decreased G-actin Content Associated with Myogenic-or Agonist-evoked Constriction in the Absence of Alterations in LC
An identical maximal level of LC 20 phosphorylation of ϳ50% was previously detected in RMCAs exposed to 1, 3, or 10 M 5-HT at 10 mm Hg (52). Based on these data, we quantified the changes in LC 20 , calponin, and caldesmon phosphoprotein and G-actin content caused by pressurization from 10 to 80 mm Hg in the presence of 5-HT (1 M), or in the reverse situation, during exposure to 5-HT (1 M) after stable development of myogenic tone at 80 mm Hg. Fig. 2A shows representative recordings of arterial diameter for RMCAs treated with 5-HT at 10 mm Hg in the absence or presence of a subsequent pressure step to 80 mm Hg. Notably, the pressure step from 10 to 80 mm Hg after 5-HT treatment evoked myogenic tone development as indicated by the presence of an initial dilation followed by constriction of 25 Ϯ 5 m (n ϭ 6 vessels; p Ͻ 0.05). No difference in LC 20 , calponin, or caldesmon phosphorylation was detected, but a significant 63 Ϯ 11% decline in the level of G-actin was evident in vessels exposed to 5-HT and then pressurized to 80 from 10 mm Hg (Fig. 2, B (n ϭ 6 vessels/group) and C-E (n ϭ 3 vessels/group)). A similar decline in G-actin content in the absence of a change in phosphoprotein content was also associated with the vasoconstriction of RMCAs treated with 5-HT at 80 mm Hg. Fig. 3A shows representative recordings of RMCA diameter during a step increase in pressure from 10 to 80 mm Hg alone and an identical pressure step followed by 5-HT (1 M). Although vasoconstriction of ϳ35 m (215 Ϯ 8 versus 180 Ϯ 11 m; n ϭ 7 vessels, p Ͻ 0.05) was evident on exposure to 5-HT, no difference in LC 20 , calponin, or caldesmon phosphorylation was detected in the presence compared with the absence of 5-HT at 80 mm Hg (Fig. 3, B (n ϭ 5 vessels/ group) and C and D (n ϭ 3 vessels/group); p Ͼ 0.05 in each case). However, the vasoconstriction to 5-HT at 80 mm Hg was associated with a significant decline of 61.5 Ϯ 5.5% in G-actin content (n ϭ 4; p Ͻ 0.05). Notably, the values for the final diameter prior to freezing and magnitude of decline in G-actin content at 80 mm Hg in the presence of 5-HT were not different regardless of the order of pressurization and agonist exposure (p Ͼ 0.05).
The initial spontaneous development of myogenic tone by RMCAs during equilibration at 60 mm Hg typically requires 5-20 min (Fig. 4A, left panel). Fig. 4B shows that no detectable difference in the level of LC 20 phosphorylation was apparent when RMCAs were flash frozen prior to tone development (Fig.  4A, right panel) compared with vessels frozen at ϳ20 min when stable myogenic constriction was achieved. In contrast, G-actin content was different before and after tone development at 60 mm Hg (Fig. 4C). Specifically, the level of G-actin detected before tone development was identical to that at 10 mm Hg, but it was ϳ60% lower after stable tone was achieved (p Ͻ 0.05; n ϭ 5 per group).
We next determined whether the pressure-evoked decline in G-actin content of RMCAs was sensitive to modulators of the actin cytoskeleton and/or cross-bridge cycling. Treatment with latrunculin B (10 M) to sequester G-actin did not affect RMCA diameter at 10 mm Hg (data not shown), but it did cause a significant vasodilatation at 80 mm Hg (Fig. 5A). The loss of tone at 80 mm Hg in latrunculin B was not associated with a decline in LC 20 phosphorylation, but a significant ϳ4-fold increase in G-actin content was detected (Fig. 5, B and C). Latrunculin B similarly evoked dilation of vessels treated with 5-HT at 80 mm Hg or pressurized to 80 mm Hg after 5-HT treatment (Fig. 6, B and C), but not in vessels treated with 5-HT at 10 mm Hg (Fig. 6A).
Previous work indicated that microcystin (phosphatase inhibitor)-triggered LC 20 phosphorylation in permeabilized rat  JULY 25, 2014 • VOLUME 289 • NUMBER 30 mesenteric arteries evoked increased actin polymerization, whereas depolymerization was observed if LC 20 phosphorylation was suppressed with the MLCK inhibitor ML-7, suggesting interplay between the extent of LC 20 phosphorylation and the level of actin polymerization (53). To test this possible mechanism in the myogenic response, we used blebbistatin, which potently inhibits the actin-activated ATPase activity of mammalian smooth muscle myosin II without affecting myosin phosphorylation (54). RMCAs were treated with blebbistatin (20 M), before or after induction of the myogenic response, and vessel diameter and G-actin content were measured. The addition of blebbistatin after the pressure jump reversed the myogenic response but had no effect on the pressure-induced actin polymerization (Fig. 7A). When added at 10 mm Hg, prior to increasing intraluminal pressure to 120 mm Hg, blebbistatin had no effect on vessel diameter at 10 mm Hg, and the vessel dilated when the pressure was increased to 120 mm Hg in the presence of blebbistatin, i.e. the myogenic response at 120 mm Hg was abolished (Fig. 7B, top panel). Most importantly, as shown in the middle and bottom panels of Fig. 7B, blebbistatin did not affect the pressure-induced actin polymerization. These results suggest that force generation per se was not responsible for the decline in G-actin content; i.e. there was no evidence of interplay between the extent of LC 20 phosphorylation and the level of actin polymerization.

Cytoskeletal Dynamics in Cerebral Arterial Myogenic Response
Contribution of ROK-and PKC-mediated Cytoskeletal Reorganization Involving Cofilin and HSP27 Phosphorylation in the Myogenic Response of RMCAs-The lack of a direct effect of force generation on G-actin content suggested the possibility that cellular signaling was likely involved in initiating the  remodeling mechanism. For this reason, we assessed the contribution of ROK and PKC signaling pathways and two potential regulators of cytoskeletal reorganization, cofilin and HSP27, that are downstream of ROK and PKC, respectively. Fig. 8 indicates the effect of ROK and PKC inhibition by H1152 (0.5 M) and GF109203X (3 M) treatment, respectively, on the level of G-actin at 120 mm Hg. RMCAs were pressurized to 120 mm Hg, and H1152 or GF109203X was then applied when stable myogenic tone was observed. Both inhibitors caused a loss of myogenic tone and dilatation, as previously reported (18,19,52). G-actin content in the presence of H1152 or GF109203X was significantly elevated compared with that of untreated RMCAs at 120 mm Hg and not different from that detected at 10 mm Hg (p Ͼ 0.05; n ϭ 4 per group), consistent with the view that ROK and PKC mediate cytoskeletal reorganization in the myogenic response.
Two cytoskeleton-associated proteins, cofilin and HSP27, were previously identified as potential mediators of ROK-and PKC-dependent control of actin polymerization in smooth muscle contraction (44,55,56), but their role in the myogenic response of resistance arteries has not been assessed. Fig. 9 shows the pressuredependent changes in, and the effects of ROK and PKC inhibition on, the levels of cofilin-S3 and HSP27-S82 phosphorylation associated with myogenic constriction at 120 mm Hg (n ϭ 6 per group). Pressurization to 120 from 10 mm Hg was associated with ϳ2.7and ϳ2.5-fold increased levels of phospho-HSP27-S82 and phospho-cofilin-S3, respectively. GF109203X suppressed the pressure-dependent increase in HSP27-S82 phosphorylation at 120 mm Hg to a level similar to that in untreated vessels at 10 mm Hg (Fig. 9, A and B). However, GF109203X did not affect phosphocofilin-S3 content (Fig. 9, A and C). In contrast, suppression of ROK activity with H1152 blocked the pressure-dependent increase in cofilin-S3, but not HSP27-S82 phosphorylation at 120  mm Hg. The level of phospho-cofilin-S3 in H1152 was similar to that in untreated vessels at 10 mm Hg (Fig. 9, A and C). The present finding of increased cofilin phosphorylation in the myogenic response differs from previous reports showing decreased levels of phospho-cofilin in tracheal and conduit arterial smooth muscle treated with constrictor agonists (44,56). For this reason, we performed an additional set of experiments in which RMCAs were exposed to 5-HT (1 M) at 10 mm Hg. The level of phospho-cofilin was increased by 2.4 Ϯ 0.4-fold in 5-HT compared with untreated vessels (n ϭ 6 in each group; p Ͻ 0.05; Fig. 10). We also tested the effects of KRIBB3 (10 M), which inhibits PKC-dependent HSP27 phosphorylation, and cyclosporin A (CyA) (10 M), a phosphatase 2B (type 2B protein serine/threonine phosphatase (calcineurin)) inhibitor. Fig. 11 shows that KRIBB3 suppressed the pressure-dependent increase in HSP27-S82 phosphorylation at 60 mm Hg (Fig. 11, A and B), with no effect on the phospho-cofilin-S3 content (Fig. 11, A and C). CyA, on the other hand, enhanced significantly   the pressure-dependent increase in cofilin-S3 but had no effect on HSP27-S82 phosphorylation at this pressure. These modulators induced opposite effects on the myogenic response of RMCAs: addition of KRIBB3 (10 M) elicited a vasodilation of ϳ35 m (Fig.  12A), whereas CyA treatment (10 M) was associated with a reduction in diameter of ϳ50 m (Fig. 12B).
Finally, we assessed the direct contribution of cofilin and HSP27 to the cytoskeletal reorganization of RMCAs during the myogenic response by analyzing the effect of these phosphorylation modulators on the G-actin content at 60 mm Hg (Fig. 13). G-actin content in the presence of KRIBB3 was significantly elevated compared with that of untreated RMCAs at 60 mm Hg, but not different from that at 10 mm Hg, whereas treatment with CyA induced a significant decrease in G-actin content compared with both controls at 10 or 60 mm Hg (p Ͻ 0.05; n ϭ 5 per group).

DISCUSSION
This study examined the role of thin filament regulatory proteins, calponin and caldesmon, and dynamic reorganization of the actin cytoskeleton in the myogenic response and agonistevoked constriction of pressurized rat cerebral arteries. No change in thin filament protein phosphorylation was apparent, but a decline in G-actin content was observed in three situations of steady-state vasoconstriction not associated with a detectable augmentation of LC 20 phosphorylation. A concomitant increase in the levels of cofilin-S3 and HSP27-S82 phosphorylation and a reduction in G-actin content were detected following pressure elevation, and these changes were blocked by inhibitors of ROK and PKC signaling. Blockade of HSP27 phosphorylation reversed the myogenic response and actin polymerization, whereas enhancement of cofilin phosphorylation induced further vasoconstriction and actin polymerization. Based on calculated levels of circumferential wall stress, we conclude that the fractional contribution of cytoskeletal reorganization to myogenic force development may be sub-stantial, accounting for ϳ30% of maximal force generated at 120 mm Hg that is capable of effectively opposing the dilating influence of intravascular pressure to permit myogenic control of arterial diameter.
It is well accepted that LC 20 phosphorylation is required for the activation of actomyosin MgATPase and cross-bridge cycling and that the level of phosphorylated LC 20 determined by the relative activities of MLCK and MLCP is the principal determinant of force generation in smooth muscle (6,57,58). However, it is also apparent that the relationship between LC 20 phosphorylation and force production is variable. MLCK activation, LC 20 phosphorylation, and force are tightly coupled to the elevation in [Ca 2ϩ ] i during the initiation of contraction (59,60), when changes in shortening velocity and force are proportional to the rise in phospho-LC 20 content (57, 58, 61). How-   JULY 25, 2014 • VOLUME 289 • NUMBER 30 ever, the level of LC 20 phosphorylation declines markedly during sustained contraction, but force production is maintained (57,58,61,62). Several mechanisms have been postulated to account for the lack of correlation between phospho-LC 20 content and force in maintained contraction, including: (i) dephos-phorylated "latch bridges" that cycle slowly or not at all (61); (ii) regulation of cross-bridge cycling by thin filament-associated proteins (23); (iii) a slow ADP off rate that is enhanced by the high force-induced strain on smooth muscle cross-bridges (63); and more recently, iv) dynamic reorganization of the actin cytoskeleton to enhance the efficiency of force transmission from the contractile apparatus to the cell membrane and extracellular matrix (24,25,56,62,64). Our findings support the view that intravascular pressure modulates the dynamics of actin polymerization via cellular signaling through ROK and PKC.

Cytoskeletal Dynamics in Cerebral Arterial Myogenic Response
We failed to obtain evidence supporting a role for thin filament regulation in myogenic or 5-HT-evoked constriction. Calponin-mediated inhibition of actomyosin MgATPase and subsequent reduction in shortening velocity and force generation were previously shown to be prevented by phosphorylation of serine 175 by PKC or CaM kinase II (28,51,65,66). Similarly, suppression of actomyosin MgATPase activity by caldesmon was blocked by phosphorylation of serine 789 by ERK1 in response to PKC activation (23,67). Here, we failed to detect any change in calponin Ser-175 or caldesmon Ser-789 phosphorylation following pressurization (80 -120 mm Hg) or in the presence of 5-HT and myogenic tone at 80 mm Hg.
There is substantial evidence to support the view that smooth muscle contraction and relaxation involve actin polymerization and depolymerization, respectively. For example, decreased G-actin content or a rise in F-to G-actin ratio, consistent with the utilization of G-actin and increased polymerization, was detected in tissues exposed to contractile agonists, elevated extracellular K ϩ concentration, tissue stretch, osmotic volume change, or intravascular pressure (34 -37, 56, 62, 64, 68 -71). Force generation in these conditions was also shown to be affected by compounds that modify actin polymerization, such as cytochalasin D, jasplakinolide, and latrunculin B (19, 34 -37, 68, 71-74). Here, we show that G-actin content was reduced by ϳ87% in a pressure-dependent manner between 10 and 120 mm Hg; no change was detected at 40 mm Hg, but significantly lower levels of G-actin were detected at 60, 80, and 120 mm Hg. Reduced G-actin content was also detected in the presence of 5-HT at 80 mm Hg (regardless of sequence), or during the initial development of myogenic tone at 60 mm Hg during equilibration. These data are consistent with previous reports, but by exploiting experimental conditions of vasoconstriction not accompanied by a detectable increase in LC 20 phosphorylation, it was also possible to assess the relative importance of cytoskeletal reorganization to myogenic force generation.
Our approach, as illustrated in Fig. 14, was to determine the level of circumferential wall stress (force per unit area at midwall) at 10, 80, and 120 mm Hg based on our measurements of inner and outer diameter and Equation 1 (see "Experimental Procedures") from Coulson et al. (48) that considers RMCAs to be thick-walled, cylindrical tubes (i.e. ratio of radius to wall thickness is Ͼ 0.1). The respective levels of circumferential wall stress at 10 and 120 mm Hg approximate the minimum stress in the relaxed state and the maximal level of circumferential stress that can be opposed by myogenic constriction; i.e. force development above ϳ120 mm Hg is insufficient to offset pressure and forced dilation is observed (21). Given the observed lack of  a detectable change in LC 20 phosphorylation between 80 and 120 mm Hg, the difference in circumferential wall stress at these two pressures is indicative of the fractional increase in force that must be developed independent of an increase in LC 20 phosphorylation and cross-bridge cycling to permit myogenic control of diameter. Based on this reasoning, ϳ30% of the total force generated to oppose intravascular pressure at 120 mm Hg may be dependent on cytoskeletal reorganization involving augmented actin polymerization (Fig. 14). The actual contribution to force generation likely exceeds 30% because (i) reorganization is not limited to the range of 80 -120 mm Hg in RMCAs but also occurs between 40 and 80 mm Hg; however, over this lower pressure range, the fractional increase in force caused by reorganization cannot be distinguished from that caused by increased LC 20 phosphorylation and cross-bridge cycling; (ii) force generation may also continue to increase with pressurization beyond 120 mm Hg, but any change in VSM tone is overwhelmed by intraluminal pressure and thus masked by forced dilatation; and (iii) the contribution of cytoskeletal reorganization is only considered at Յ5 min following pressurization. The cytoskeletal remodeling process associated with prolonged agonist-induced vasoconstriction was previously shown to be maintained, leading to a gradual increase in actin polymerization with time and, presumably, a greater fractional contribution to force generation (62,75).
The cellular processes that orchestrate cytoskeletal remodeling in the myogenic response remain to be defined. Here, we show that the decline in G-actin and myogenic constriction at 120 mm Hg are accompanied by an increase in cofilin and HSP27 phosphorylation and that these changes are suppressed following inhibition of ROK or PKC. Specifically, H1152 and GF109203X reduced the decline in G-actin associated with pressurization, but H1152 blocked cofilin phosphorylation without affecting phospho-HSP27 content and GF109203X suppressed HSP27 phosphorylation but did not alter the level of phospho-cofilin. Cofilin contributes to actin dynamics by binding to and severing actin filaments, which favors depolymerization. This activity also provides soluble G-actin monomers and short, capped actin filaments with barbed ends that are free to act as nucleation sites for new filament growth (76 -78). Cofilin activity is suppressed following phosphorylation of Ser-3 by LIM kinase (79), and LIM kinase activity is in turn dependent on phosphorylation by ROK or p21-activated kinases (PAK1 and 4) downstream of RhoA and Cdc42, respectively (79). In contrast, HSP27 may suppress actin polymerization by directly binding to G-actin (80,81) or by capping the barbed ends of F-actin which prevents filament elongation (82), but the mechanism is controversial (83). Phosphorylation of HSP27 blocks its G-actin binding and capping activity, but phospho-HSP27 may also bind to and stabilize F-actin filaments (84). HSP27 phosphorylation by MAP kinase-activated protein kinase 2 or 3 in response to upstream regulation by p38 MAP kinase is the most widely identified regulatory pathway (85,86), but PKC has also been demonstrated to phosphorylate HSP27 (87,88). Actin dynamics in the myogenic response may therefore require parallel activation of (i) RhoA-ROK to promote actin polymerization by stimulating LIM kinase and subsequent suppression of cofilin-mediated actin severing and depolymerization and (ii) PKC-mediated activation by diacylglycerol generated through pressure-evoked phospholipase C activity (89) and subsequent phosphorylation of HSP27 to promote polymerization. The involvement of ROK in the regulation of cofilin activity represents the second potential role for this kinase in the myogenic response, the first being the phosphorylation of MYPT1 and suppression of MLCP activity leading to Ca 2ϩ sensitization (18,52). That PKC contributes to the regulation of force in the myogenic response is well recognized (20, 21, 90 -93), but the mechanism(s) involved has not been established with certainty. Previous studies postulated that PKC might phosphorylate CPI-17 and/or MYPT1, leading to an inhibition of MLCP activity, Ca 2ϩ sensitization of the contractile filaments, increased LC 20 phosphorylation, and greater force at constant [Ca 2ϩ ] i (6,8,94). However, this is unlikely because no evidence of a pressure-dependent increase in the level of phospho-CPI-17 or PKC-dependent MYPT1 phosphorylation was detected in the myogenic response of rat cerebral or gracilis arteries (18,19,52). That the pressure-dependent increase in HSP27 phosphorylation and decline in G-actin content were blocked by GF109203X treatment in this study indicates that PKC inhibitors may suppress myogenic constriction through inhibition of PKC-mediated HSP27 phosphorylation and actin polymerization.
KRIBB3 and CyA effects on HSP27 and cofilin phosphorylation levels, myogenic constriction, and G-actin content levels of RMCAs serve as additional evidence that these proteins play direct roles in the actin reorganization process in the myogenic response. KRIBB3 inhibits PKC-dependent HSP27 phosphorylation and migration in adenocarcinoma cell lines (95,96). CyA is a phosphatase 2B inhibitor that prevents cofilin dephosphorylation (44,97). Here, KRIBB3 and CyA treatments were used at 60 mm Hg to provide a wider range for potential changes of arterial diameter, G-actin content, and HSP27 and cofilin phosphorylation levels. KRIBB3, by preventing HSP27 phosphorylation and its uncapping from the actin filament, induced a loss FIGURE 14. Estimation of fractional contribution of cytoskeletal reorganization to force generation in the myogenic response based on levels of circumferential wall stress at 10, 80, and 120 mm Hg. Diagrammatic representation and measured wall dimensions of RMCAs were employed to calculate circumferential wall stress (CWS) at 10, 80, and 120 mm Hg and determination of fractional change in circumferential wall stress between 80 and 120 mm Hg after subtraction of basal CWS at 10 mm Hg.
of myogenic tone and blocked the pressure-induced reduction in G-actin content. CyA, on the other hand, prevented phosphatase 2B-mediated cofilin dephosphorylation and its actin severing function, thus inducing contraction of the artery, as reported for other vascular beds (98). It also evoked a greater decrease in G-actin content and an increase in cofilin-S3, but not HSP27-S82 phosphorylation. The lack of effect on phospho-HSP27-S82 is consistent with previous findings that HSP27 is dephosphorylated by PP2A, but not PP2B (99).
Our findings indicate that the regulation of cytoskeletal reorganization in the myogenic response may differ from that in stretch-, agonist-, and KCl-evoked contraction. For example, contraction of portal vein smooth muscle in response to stretch was accompanied by a ROK-mediated increase in cofilin phosphorylation and F:G-actin ratio (100). However, the time required for these changes (Ͼ1 h) was considerably longer than observed here for myogenic constriction (3-5 min). Also, a reduced level of phospho-cofilin, rather than an increase, was found to accompany the rise in F:G-actin ratio in airway and carotid arterial smooth muscle tissues contracted by acetylcholine or elevated external K ϩ concentration (44,56). We assessed this possibility by treating RMCAs with 5-HT at 10 mm Hg, but an increase, not a decrease, in cofilin phosphorylation was detected. The mechanism of HSP27 inhibition in the myogenic response may also differ; specifically, HSP27 phosphorylation in agonist-and/or stretch-evoked contraction of airway smooth muscle, mesenteric arteries, and rabbit facial vein tissues was shown to be sensitive to inhibition of p38 MAP kinase and/or ROK inhibition rather than PKC (46,101,102). The reason(s) for these varied findings is not known, but an intriguing possibility is that cytoskeletal reorganization in the myogenic response is mediated by a different complement of signaling pathway(s) and membrane adhesion proteins compared with those involved in stretch-, agonist-, or KCl-evoked isometric contraction of VSMCs. A better understanding of processes contributing to actin dynamics in the myogenic response is required to permit definitive conclusions regarding the reason(s) for stimulus-dependent differences in the mechanism of cytoskeletal reorganization.