Ca2+-independent Phospholipase A2 Is Required for Agonist-induced Ca2+Sensitization of Contraction in Vascular Smooth Muscle*

Excitatory agonists can induce significant smooth muscle contraction under constant free Ca2+ through a mechanism called Ca2+ sensitization. Considerable evidence suggests that free arachidonic acid plays an important role in mediating agonist-induced Ca2+-sensitization; however, the molecular mechanisms responsible for maintaining and regulating free arachidonic acid level are not completely understood. In the current study, we demonstrated that Ca2+-independent phospholipase A2 (iPLA2) is expressed in vascular smooth muscle tissues. Inhibition of the endogenous iPLA2 activity by bromoenol lactone (BEL) decreases basal free arachidonic acid levels and reduces the final free arachidonic acid level after phenylephrine stimulation, without significant effect on the net increase in free arachidonic acid stimulated by phenylephrine. Importantly, BEL treatment diminishes agonist-induced Ca2+ sensitization of contraction from 49 ± 3.6 to 12 ± 1.0% (p< 0.01). In contrast, BEL does not affect agonist-induced diacylglycerol production or contraction induced by Ca2+, phorbol 12,13-dibutyrate (a protein kinase C activator), or exogenous arachidonic acid. Further, we demonstrate that adenovirus-mediated overexpression of exogenous iPLA2 in mouse portal vein tissue significantly potentiates serotonin-induced contraction. Our data provide the first evidence that iPLA2 is required for maintaining basal free arachidonic acid levels and thus is essential for agonist-induced Ca2+-sensitization of contraction in vascular smooth muscle.

Whereas the increase of intracellular free Ca 2ϩ plays a pivotal role in the regulation of smooth muscle contraction, excitatory agonists can induce significant smooth muscle contraction under constant free Ca 2ϩ by increasing the sensitivity of the contractile apparatus to Ca 2ϩ . The latter is called the Ca 2ϩ sensitization mechanism. Ca 2ϩ sensitization plays an important physiological role in the regulation of the tonic phase of contraction induced by various agonists (1). Abnormalities of Ca 2ϩ sensitization and its machinery have been implicated in the pathophysiology of several cardiovascular diseases including hypertension, coronary artery spasms, and restenosis (2)(3)(4)(5)(6). Therefore, the molecular basis underlying Ca 2ϩ sensitization has been extensively studied, and many signaling molecules have been identified that regulate Ca 2ϩ sensitization, such as small G-protein RhoA (7,8) and its effector Rho kinase (2,9), myosin phosphatase (10 -12), the heterotrimeric G protein G 12/13 (13), arachidonic acid (14,15), ZIP-like kinase (16), ZIP kinase (17), and phosphatase inhibitory protein CPI-17 (18 -20).
Several lines of evidence suggest that arachidonic acid contributes to agonist-induced Ca 2ϩ sensitization in vascular smooth muscle. First, many Ca 2ϩ -sensitizing agonists increase arachidonic acid in vascular smooth muscle, including norepinephrine (67), angiotensin II (21), endothelin (22), vasopressin (23), and GTP␥S 1 (15). Second, the time course and the mass of arachidonic acid released by GTP␥S are consistent with the role of arachidonic acid as a messenger in the G-protein-coupled inhibition of phosphatase (15). Third, exogenous arachidonic acid, not its metabolic products, increases 20-kDa myosin light chain phosphorylation and causes the contraction of smooth muscle at constant Ca 2ϩ by dissociating and reducing myosin phosphatase activity (14,15). Fourth, arachidonic acid can stimulate Rho kinase (24,25) in solution, and arachidonic acid-induced Ca 2ϩ sensitization of contraction is partially inhibited by a Rho kinase inhibitor, Y-27632 (25,68). Finally, a PLA 2 inhibitor, ONO-RS-082, concomitantly blocks agonistinduced arachidonic acid release and Ca 2ϩ sensitization of force (26). However, the physiological significance of arachidonic acid in agonist-induced Ca 2ϩ sensitization remains to be established. One major obstacle in demonstrating the physiological importance of arachidonic acid in the regulation of smooth muscle contraction is that, to date, the enzyme responsible for agonist-induced arachidonic acid release and Ca 2ϩ sensitization has not been identified.
In the current work, we test the hypothesis that the iPLA 2 mediates, at least in part, free arachidonic acid release and Ca 2ϩ sensitization of contraction in vascular smooth muscle. Our results show that iPLA 2 is expressed in vascular smooth muscle tissue, and inhibition of the endogenous iPLA 2 activity by BEL decreases basal free arachidonic acid levels and diminishes phenylephrine-induced Ca 2ϩ sensitization of contraction. Moreover, adenovirus-mediated overexpression of exogenous iPLA 2 in mouse portal vein tissue significantly potentiates agonist-induced contraction. We conclude that iPLA 2 is required for agonist-induced Ca 2ϩ sensitization of contraction in vascular smooth muscle.
Tissue Preparation-Rabbit or mouse portal veins were prepared as previously described (15). Briefly, the adventitia was removed carefully, and under a light microscope, the endothelium cells were denuded by gentle rubbing of the inner surface with a razor blade. The denudation of endothelium was verified by the lose of acetylcholine-induced relaxation in early experiments.
Isometric Tension Measurement-Isometric tension of small strips (3 mm long, 150 -200 m wide, and 75 m thick) was measured with a force transducer (AE801; AME, Horten, Norway) in a well on a "bubble" plate at 24°C. Details of the solution used for studies on intact or ␣-toxin permeabilized were described previously (57). After the steady responses to high [K ϩ ] were observed, the strips were incubated in a Ca 2ϩ -free solution and permeabilized with 17.5 g/ml ␣-toxin for 60 min. To deplete the sarcoplasmic reticulum of calcium, all permeabilized strips were treated with A23187 (10 M; Calbiochem) as described (57).
Western Blot Analysis-Proteins were separated by SDS-polyacryl-amide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Fisher). Nonspecific binding sites on the polyvinylidene difluoride membrane were blocked by 4% nonfat milk in PBST buffer (PBS plus 0.1% Tween 20). iPLA 2 was detected by using anti-iPLA 2 antibodies from Upstate or from Cayman (Ann Arbor, MI). The immunoreactive bands were blotted with horseradish peroxidase-conjugated goat-anti-rabbit antibodies (1:7500 dilution; Jackson ImmunoResearch Laboratories, Inc.) and detected by enhanced chemiluminescence. Phospholipase A 2 Assays-The iPLA 2 activity was assayed using a well established method (58). The reaction was carried out in a Ca 2ϩfree buffer that virtually abolishes sPLA 2 and cPLA 2 activity. 1 mM ATP and 2 mM dithiothreitol were included in order to stabilize the iPLA 2 and inactive sPLA 2 activities, respectively. In preliminary experiments, we found that the majority of the iPLA 2 activity was present in the membrane fraction. Therefore, the membrane fraction was used in the subsequent iPLA 2 assay experiments. The tissue homogenate was separated to the cytosol and membrane fraction by a 100,000 ϫ g centrifugation at 4°C. ϳ100 g of membrane proteins were incubated with mixture of 14 C-labeled and unlabeled DPPC for 90 min at 40°C. iPLA 2 activity was linear with an incubation time of up to 2 h (data not shown). The free fatty acid released was extracted by the modified Dole reagents, and the radiolabeled free fatty acid was quantified by liquid scintillation counting. The iPLA 2 specific activity was expressed as pmol of free fatty acid released/mg of protein in 1 min (pmol/min/mg).
[ 3 H]Arachidonic Acid Release and DAG Production-Rabbit portal vein strips were labeled with [ 3 H]arachidonic acid (2 Ci/ml) in HEPES-buffered Krebs solution overnight at 37°C (15). The strips were washed three times (20 min each time) in Krebs solution containing 0.2% fatty acid-free bovine serum albumin (Sigma) and 10 M BEL or vehicle (Me 2 SO, as controls). The strips were then stimulated with 10 M phenylephrine for 5 min at 24°C. The medium containing released [ 3 H]arachidonic acid was removed and counted by liquid scintillation to determine the amount of arachidonic acid released. The tissue was used to determine the tissue DAG level as previously described (15). Briefly, tissue lipids were extracted twice with chloroform and separated by thin layer chromatography. The bands corresponding to free fatty acid and DAG were scrapped off and counted by liquid scintillation. [ 3 H]Arachidonic acid release and DAG production were normalized as a percentage of the total 3 H counts incorporated into lipids.
Recombinant Adenovirus Construction-Rat iPLA 2 cDNA was cloned as previously described (33). The SpeI/StuI-MluI restriction fragment of pBK-CMV/iPLA 2 (33) was generated by PCR. The SpeI/StuI-MluI fragment of PCR products was ligated into a modified version of pBluescript KS (pFLAGmluI), which contains the FLAG epitope (DYKDDDDK). The SpeI-StuI fragment of pBK-CMV/iPLA 2 and the StuI-NotI fragment of pFLAGmluI/iPLA 2 were ligated into a modified version of an adenoviral shuttle vector (pAdtracgfptre) containing two expression cassettes, one that uses the cytomegalovirus promoter to drive green fluorescent protein (GFP) expression and one that uses the tetracycline response element promoter to drive iPLA 2 expression (59). To generate the pAd-iPLA 2 adenovirus, an adenoviral backbone vector (pAdEasy-1) and the PmeI-linearized iPLA 2 shuttle vector (pAdtracgfptre/iPLA 2 ) were co-transformed into electrocompetent Escherichia coli BJ5183 cells (60). The successful recombination of these two vectors was screened by restriction enzyme analyses. To generate the adenovirus, the identified recombinants were linearized with PacI and transfected into a mammalian packaging cell line (HEK293) by using Lipo-fectAMINE-plus according to the manufacturer's protocol (Invitrogen). Expression of GFP and lysis of the HEK 293 cells were taken as an indication of successful viral production. Moreover, the expression of iPLA 2 was confirmed by Western blot using anti-iPLA 2 and anti-FLAG antibodies. Large quantities of adenovirus were produced by infecting HEK293 cells in 100-mm 2 dishes and purified by cesium chloride gradient ultracentrifugation (61). The physical number of viral particles was determined by optical absorbency.
Adenoviral Infection-The Ad-iPLA 2 and Ad/Tet-on (encoding a tetracycline-regulatory transcription factor) viruses were co-transfected into A10 smooth muscle cells or mouse portal vein tissue. The expression of iPLA 2 was achieved by adding doxycycline to the cell culture medium (10% fetal bovine serum/Dulbecco's modified Eagle's medium; Invitrogen). The adenoviral transfection condition was optimized to obtain maximal expression of GFP, FLAG, and iPLA 2 and to minimize the cytopathic effect of the adenovirus. For cultured A10 smooth muscle cells, a multiplicity of infection of ϳ200 was found to be sufficient to achieve over 90% transfection. For mouse portal vein tissue, however, higher doses of adenovirus (1.77 ϫ 10 9 viral particles in a total of 100 l/one-half mouse portal vein) and a longer incubation time (24 h at 37°C) were applied.
Statistical Analysis-Each experiment was repeated a minimum of three times. Data were expressed as mean Ϯ S.E. Statistical analysis was performed by using an unpaired t test.

iPLA 2 Protein Is Expressed in Vascular Smooth
Muscle Tissues-iPLA 2 protein is expressed in many cell types, including cultured rat vascular smooth muscle cells (62); however, it is not clear whether the iPLA 2 protein is expressed in fully differentiated vascular smooth muscle tissue. To address this question, homogenate from the medium of rabbit aorta, portal vein, or femoral artery were examined by Western blot analysis using two iPLA 2 antibodies that recognize distinct epitopes of iPLA 2 . Both antibodies recognized one major band at the expected molecular mass of about 85 kDa, strongly suggesting that iPLA 2 is expressed in the tissue (Fig. 1). In addition, since iPLA 2 has been shown to be the predominant phospholipase A 2 in the brain, rabbit brain was included as a positive control.

Inhibition of iPLA 2 Activity Diminishes PE-induced Ca 2ϩ Sensitization of Contraction in Vascular Smooth
Muscle Tissues-To investigate whether iPLA 2 plays a role in the regulation of smooth muscle contraction, the effect of inhibiting iPLA 2 activity on agonist-induced contractions was determined. BEL was used to selectively inhibit iPLA 2 activity, since it has been shown to be over 1,000-fold more selective for iPLA 2 over cPLA 2 and sPLA 2 (63,64). We determined and compared PE-induced contractions in a single tissue strip in the presence or absence of BEL. In preliminary experiments, the muscle strips were incubated with BEL for 15, 30, and 60 min. We found that 60-min preincubation induced the largest inhibition of PEinduced contractions; therefore, it was used in all of the subsequent experiments. We found that BEL not only significantly inhibits the amplitude of the sustained phase of PE-induced contraction from 46.8 Ϯ 5.38 mg (n ϭ 4) to 26.4 Ϯ 2.74 mg (n ϭ 4, p Ͻ 0.05) but also potently slows down the rate of force development (the t1 ⁄2 from 3.0 Ϯ 0.86 min (n ϭ 4) to 9.0 Ϯ 2.18 min (n ϭ 4, p Ͻ 0.05)) ( Fig. 2A). In contrast, BEL did not significantly affect high K ϩ depolarization-induced contractions (Fig. 2B). The fact that BEL selectively inhibits PEinduced but not K ϩ depolarization-induced contractions suggests that BEL selectively acts on an agonist-activated process. Interestingly, BEL does not completely abolish the PE-induced contraction ( Fig. 2A). In the presence of BEL, the sustained/ slow phase of PE-induced contractions may be mediated by Ca 2ϩ . It may also be mediated by BEL-insensitive phospho-lipase A 2 or due to incomplete inhibition of iPLA 2 activity by BEL.
Various agonists including PE cause smooth muscle contractions by both Ca 2ϩ -dependent and Ca 2ϩ -independent mechanisms. To determine specifically whether iPLA 2 is involved in the Ca 2ϩ -independent Ca 2ϩ sensitization process, we used a well established ␣-toxin permeabilized smooth muscle system (1). In this system, the cytosolic free Ca 2ϩ can be clamped, and the Ca 2ϩ sensitization can be selectively determined. BEL dosedependently inhibited PE-induced Ca 2ϩ -sensitization of contraction. The PE-induced Ca 2ϩ sensitization of contraction (percentage of maximal Ca 2ϩ -induced contraction) was 49.0 Ϯ 3.63% (n ϭ 8) in controls and was inhibited to 21.6 Ϯ 3.20% (n ϭ 5, p Ͻ 0.01) by 3 M BEL and to 12.0 Ϯ 1.01% (n ϭ 3, p Ͻ 0.01) by 10 M BEL (Fig. 3A). In contrast, BEL did not significantly affect the Ca 2ϩ -induced contraction (Fig. 3B). The maximal concentration of Ca 2ϩ -induced contraction (absolute force) was 58.9 Ϯ 4.83 mg (n ϭ 12) in the control and 57.3 Ϯ 4.27 mg (n ϭ 12, p Ͼ 0.05) in the 10 M BEL-treated groups (Fig. 3B). Contraction induced by submaximal concentration of Ca 2ϩ (pCa6.3, expressed as percentage of maximal Ca 2ϩ -induced contraction) was 6.8 Ϯ 3.23% (n ϭ 5) in the control and was 7.4 Ϯ 1.67% (n ϭ 5, p Ͼ 0.05) in the presence of 10 M BEL.

BEL Inhibits iPLA 2 Activity and Decreases Basal Free Arachidonic Acid Levels in Portal Vein Smooth Muscle
Tissue-To ensure that the inhibition of Ca 2ϩ sensitization of smooth muscle contraction by BEL results from its inhibition of iPLA 2 activity, we assayed iPLA 2 activity using an exogenous 14 C-DPPC as a substrate in the presence and absence of BEL and/or PE. As shown in Fig. 4A, BEL (10 M, 60 min preincubation) significantly inhibited the iPLA 2 activity from 5.3 Ϯ 0.30 pmol/mg/min (n ϭ 12) to 2.8 Ϯ 0.2 pmol/mg/min (n ϭ 3, p Ͻ 0.01) (Fig. 4A). The iPLA 2 activity was not significantly increased by PE stimulation (10 M, 5 min), although there was a trend of increase (from basal 5.3 Ϯ 0.30 pmol/mg/min (n ϭ 12) to 6.6 Ϯ 0.56 pmol/mg/min (n ϭ 13), p ϭ 0.06). This trend of iPLA 2 activity increase by PE stimulation was abolished by BEL: 2.7 Ϯ 0.19 pmol/mg/min (n ϭ 4) in the absence of PE and 2.8 Ϯ 0.20 pmol/mg/min (n ϭ 3) in the presence of PE.
The Selectivity of BEL's Effect on iPLA 2 and Ca 2ϩ Sensitization of Contraction-In addition to potently and selectively inhibiting iPLA 2 among the classes of phospholipase A 2 , BEL also inhibits Mg 2ϩ -dependent phosphatidic acid phosphohydrolase (PAP-1) (63). PAP-1 can dephosphorylate phosphatidic acid and yield DAG. PE has been shown to increase the cellular DAG level in vascular smooth muscle (15), although the DAG and protein kinase C pathway plays only a minor role in PEinduced Ca 2ϩ sensitization of contraction (19,64). However, to vigorously test our hypothesis, we determined whether BEL diminished DAG levels in rabbit portal vein smooth muscle tissue under our experimental conditions. BEL (10 M) did not , and brain were homogenized in radioimmune precipitation buffer. 10 g of cell lysates from each tissue were analyzed by Western blot using anti-iPLA 2 antibodies from Cayman and Upstate. Note that one major band was identified using either of the two antibodies that recognizes different epitopes on the iPLA 2 protein (Cayman antibody, amino acids 561-575; Upstate antibody, amino acids 681-690 of rat iPLA 2 ␤). Shown in the figure are representative blots of at least three independent experiments. significantly affect the DAG level in the presence of PE (3.7 Ϯ 0.14% (n ϭ 4) versus 4.6 Ϯ 0.62% (n ϭ 4), p Ͼ 0.05) although this concentration of BEL potently inhibited PE-induced Ca 2ϩ sensitization of contraction (Figs. 2 and 3). This suggests that, under our experimental conditions, 10 M BEL selectively inhibits iPLA 2 activity, or PAP-1 does not contribute significantly to DAG level in portal vein smooth muscle tissues.
To further investigate the selectivity and the action site of BEL on the signaling pathways that regulate Ca 2ϩ sensitization of contraction, we examined the effect of BEL on exogenous arachidonic acid-and PDBu-induced contractions. Free arachidonic acid is a product of PLA 2 action; therefore, BEL is not expected to affect exogenous free arachidonic acid-induced contraction. Indeed, BEL (10 M) had no effect on arachidonic acid (50 M)-induced contractions (Fig. 5A). On the other hand, PDBu directly activates protein kinase C to induce Ca 2ϩ sensitization of contraction, and iPLA 2 activity is not required in the pathway downstream of protein kinase C. Therefore, BEL is not expected to affect PDBu-induced Ca 2ϩ -sensitization of contraction. Indeed, BEL (10 M) did not affect PDBu (1 M)induced Ca 2ϩ sensitization of contraction (Fig. 5B).

Overexpression of iPLA 2 in Vascular Smooth Muscle Tissues
Potentiates Agonist-induced Contractions-To further establish the role of iPLA 2 in agonist regulation of smooth muscle contractions, we determined whether overexpression of iPLA 2 in vascular smooth muscle tissue potentiates agonist-induced contractions. A tetracycline-inducible, replication-deficient recombinant adenovirus encoding rat iPLA 2 (33) was constructed and used to infect vascular smooth muscle tissue and overexpress iPLA 2 ( Fig. 6 and Fig. 7).
To demonstrate that the tetracycline-inducible recombinant adenoviral system produces functional iPLA 2 expression, we used an A10 cell line derived from rat aortic smooth muscle. As shown in Fig. 6B, lane 1, FLAG-tagged iPLA 2 (recombinant iPLA 2 ) was not detectable by Western blot in the absence of doxycycline, indicating that the expression of recombinant iPLA 2 in the smooth muscle cell line was tightly controlled by tetracycline. The addition of different concentrations of doxycycline caused a dose-dependent increase in iPLA 2 expression (Fig. 6B, lanes 2-6). Importantly, overexpression of iPLA 2 increased iPLA 2 activity up to 100-fold when compared with endogenous iPLA 2 activity (Fig. 6C), indicating that the expressed iPLA 2 is enzymatically active.
Adenoviruses have been successfully used to transfer specific genes into cultured vascular cells to enable gene expression and functional activity, but application of this technique to the vascular wall is currently limited by its relatively low efficiency of transfection (65). In order to monitor the transfection efficiency, a GFP whose expression was driven by a cytomegalovirus promoter (not an inducible promoter) was included in the recombinant adenoviral construct (Fig. 6A) (59). Confocal microscopy was used to monitor the expression of GFP as an indication of the efficiency of adenoviral transfection. In addi- tion, in order to maximize the possibility of obtaining a significant functional effect of overexpressing iPLA 2 , we sought to use a tissue with a low level of endogenous iPLA 2 . Through combined analyses of GFP expression and the endogenous iPLA 2 protein expression levels in various vascular tissues (portal veins and femoral arteries from rabbits, rats, and mice), we found that a higher smooth muscle cell transfection rate and a higher recombinant iPLA 2 /endogenous iPLA 2 ratio could be obtained with mouse portal vein tissue. However, PE only induces minimal contractions in mouse portal vein tissue, whereas serotonin initiates a good contractile response. Accordingly, we sought to examine the effect of overexpressing iPLA 2 on serotonin-induced contractions. First, we tested whether serotonin-induced contractions were sensitive to BEL. We found that 3 M BEL very significantly inhibited serotonininduced contractions to 33.1 Ϯ 2.84% of the control level (n ϭ 5, p Ͻ 0.01). Therefore, mouse portal veins were used in subsequent adenovirus experiments. As shown in Fig. 7A, recombinant iPLA 2 protein is readily detected by Western blot using an anti-iPLA 2 antibody and confirmed by an anti-FLAG antibody. The iPLA 2 protein reached about 2.3-fold of the control level with the adenoviral transfection. In the absence of doxycycline, only endogenous iPLA 2 protein is detected. Moreover, expression of GFP was detected by confocal microscopy in about 50% of the smooth muscle cells (Fig. 7B). What is responsible for the mobility shift between the endogenous mouse iPLA 2 and the overexpressed recombinant iPLA 2 (Fig. 7A) remains to be identified; it may relate to the splice variants of iPLA 2 .
Finally, we determined the effect of iPLA 2 overexpression on the dose-response curve of serotonin-induced contractions. The portal vein tissue from each mouse was cut into two equal pieces and randomly divided into the two groups. A total of six mice were used. Both groups were infected with the same concentration of recombinant adenovirus encoding iPLA 2 . One group of portal vein tissue was treated with doxycycline to induce iPLA 2 overexpression, and the other group of portal vein tissue served as the control (without doxycycline). As shown in Fig. 7C, expression of iPLA 2 in mouse portal vein smooth muscle significantly potentiated serotonin-induced contractions in comparison with the controls (0.1 M serotonin, from 12 Ϯ 1.44 (n ϭ 6) to 20 Ϯ 2.74 mg (n ϭ 6, p Ͻ 0.05); 10 M serotonin, from 19 Ϯ 2.11 (n ϭ 6) to 29 Ϯ 3.9 mg (n ϭ 6, p Ͻ 0.05); 100 M serotonin, from 19 Ϯ 2.07 (n ϭ 6) to 30 Ϯ 3.89 mg (n ϭ 6, p Ͻ 0.05)). The EC 50 was not significantly different between the control and doxycycline-treated group (log EC 50 was Ϫ7.3 Ϯ 0.058 versus Ϫ7.4 Ϯ 0.039, n ϭ 6 each, p Ͼ 0.05). However, the high potassium depolarization-induced contractions were not significantly different between the two groups (9 Ϯ 0.66 mg versus 10 Ϯ 1.57 mg, p Ͼ 0.05, n ϭ 6 each). Importantly, doxycycline alone did not significantly affect serotonin-induced contractions. The maximal contraction induced by serotonin was 20 Ϯ 2.45 mg (n ϭ 6) in the absence of doxycycline and 14 Ϯ 4.05 mg (n ϭ 6, p Ͼ 0.05) in the presence

DISCUSSION
Ca 2ϩ sensitization plays a physiological role in mediating agonist-induced smooth muscle contractions (1,66). Multiple lines of evidence suggest that arachidonic acid plays a role in mediating agonist-induced Ca 2ϩ sensitization of contraction in smooth muscle; however, the molecular mechanisms responsible for maintaining and regulating free arachidonic acid level are not completely understood. The major finding of the present study is that an iPLA 2 is required for Ca 2ϩ sensitization of smooth muscle contraction, and such a requirement results from the essential role of iPLA 2 in maintaining the basal free arachidonic acid levels. Several lines of evidence support this conclusion. First, immunodetectable iPLA 2 protein is present in portal vein smooth muscle (Fig. 1). Second, inhibition of iPLA 2 activity with BEL diminishes basal free arachidonic acid levels (Fig. 4), contractions in intact (nonpermeabilized) portal veins (Fig. 2), and Ca 2ϩ sensitization of contraction in ␣-toxin-permeabilized tissue preparations (Figs. 3 and 5). Importantly, BEL does not affect agonist-induced generation of DAG or high K ϩ -, Ca 2ϩ -, PDBu-, or arachidonic acid-induced contractions (Figs. 2 and 5), indicating the selectivity of BEL under our experimental condi-tions. Third, overexpression of iPLA 2 in mouse portal veins by adenovirus-mediated gene transfer leads to a significant increase in serotonin-induced smooth muscle contractions (Figs. 6 and 7).  figure) for 12 h. Forty-eight hours later, the cells were lysed. Expression of iPLA 2 protein was determined by Western blot, and iPLA 2 activity was assayed using the methods described in detail under "Experimental Procedures." A, schematic representation of the adenoviral construct. B, iPLA 2 immunoblot. 10 g of protein was loaded in each lane. Shown is a representative blot from three independent experiments. C, induction of iPLA 2 expression is associated with over a 100-fold increase in iPLA 2 specific activity. n ϭ 3 for each bar. *, p Ͻ 0.05.

FIG. 7.
Overexpression of iPLA 2 in vascular smooth muscle tissue potentiates serotonin-induced contractions. The portal vein from each mouse was cut into two parts of equal size and incubated with Ad-iPLA 2 for 24 h. One part was incubated in the presence of doxycycline (2 g/ml) to induce the expression of iPLA 2 , whereas the other part was incubated in the absence of doxycycline to serve as a control. After incubation with the adenovirus, the portal vein tissues were incubated at 37°C for another 40 h. A, expression of iPLA 2 was analyzed by Western blot using an anti-iPLA 2 antibody that recognizes both endogenous and exogenous iPLA 2 and an anti-FLAG antibody that only recognizes exogenous iPLA 2 . Six mice portal veins were used for Western blotting of iPLA 2 in three independent blots. B, transfection efficiency was determined by confocal microscopy that detects the expression of GFP (left panel). GFP is expressed regardless of the presence or absence of doxycycline. Right, a Nomarski image showing all of the smooth muscle cells at the same cross-layer. Three mice were used for confocal microscopy detection of GFP expression. C, expression of iPLA 2 shifts the serotonin dose-response curve to the left and increases the maximal contraction to 153% of that of the control group. Six mice portal veins were used to determine the serotonin dose-response curve. **, p Ͻ 0.01. iPLA 2 has been implicated in multiple cellular functions in a variety of cell types and tissues. First, iPLA 2 has been proposed to be responsible for membrane phospholipid remodeling by providing lysophospholipids as an acceptor for free fatty acid in murine P388D1 macrophage-like cell line (28,39). Second, iPLA 2 has been proposed to play a signaling role in mediating agonist-induced net free arachidonic acid release in pancreatic islet ␤-cell and submandibular gland ductal cell (30,33,41). Third, proteocleavage by caspase-3-induced iPLA 2 activation has been proposed to be required for the execution of apoptosis in U937 cells (42,43). Fourth, iPLA 2 has also been suggested to play a role in cell proliferation and cell spreading (47,53). With regard to the function of iPLA 2 in smooth muscle, Gross and co-workers (23) demonstrated that iPLA 2 activity is largely responsible for [Arg 8 ]vasopressin-induced net arachidonic acid release in a cultured A10 rat aortic vascular smooth muscle cell line. However, little is known about the physiological role of iPLA 2 in the regulation of mature vascular smooth muscle function. In particular, it is unknown whether iPLA 2 is involved in the regulation of mature smooth muscle contractions. In the present study, we provide several lines of evidence strongly suggesting that iPLA 2 is required for Ca 2ϩ sensitization of smooth muscle contractions. Our evidence suggests an essential role of iPLA 2 in maintaining basal free arachidonic acid levels (Fig. 4). Interestingly, agonist stimulation only induces a small and statistically insignificant increase of iPLA 2 activity that is not proportional to the significant net increase in free arachidonic acid level. In addition, the phenylephrineinduced net increase in free arachidonic acid is not significantly different in the presence versus in the absence of BEL. These findings suggest that iPLA 2 plays only a minor role in the phenylephrine-induced net increase of free arachidonic acid, and a phospholipase A 2 other than the BEL-sensitive iPLA 2 is mainly responsible for the agonist-induced net increase in free arachidonic acid. This is consistent with the notion that multiple phospholipase A 2 activities, including Ca 2ϩ -dependent and Ca 2ϩ -independent, are present in vascular smooth muscle tissues (25).
Free arachidonic acid released by iPLA 2 may mediate agonist-induced Ca 2ϩ -sensitization of contraction by multiple molecular mechanisms. Inhibition of the myosin phosphatase that dephosphorylates the 20-kDa myosin light chain is the downstream mechanism mediating Ca 2ϩ sensitization of smooth muscle contractions (12). There are at least two pathways that can couple agonist stimulation to inhibition of myosin phosphatase. One is the Rho/Rho kinase pathway, and the other is the protein kinase C/CPI-17 pathway. Arachidonic acid may interact with both pathways or either pathway to contribute to Ca 2ϩ sensitization of contraction. For example, arachidonic acid may interact with the Rho/Rho kinase pathway by activating Rho kinase. Arachidonic acid has been shown to activate Rho kinase in solution (24), and exogenous arachidonic acid-induced Ca 2ϩ sensitization of contraction is partially reversed by inhibiting Rho kinase (68). Arachidonic acid may interact with the protein kinase C/CPI-17 pathway by activating certain isoforms of protein kinase C (26). In addition, arachidonic acid can directly inhibit phosphatase by dissociating the phosphatase subunits (12). Interestingly, the dissociation of the phosphatase subunits was reported to occur in isolated smooth muscle cells in an agonist-specific manner (69). In addition, multiple physiological agonists can, with different potency and time course, induce vascular smooth muscle contractions and Ca 2ϩ sensitization (1). These agonists couple to divergent pathways and, therefore, may initiate Ca 2ϩ sensitization by turning on different signals. Arachidonic acid may act through different mechanisms when stimulated by different agonists. These various possibilities by which arachidonic acid induces Ca 2ϩ sensitization of contraction are currently under active investigation in our laboratory.
In summary, we used pharmacological and adenovirus-mediated gene transfer approaches to either inhibit or promote iPLA 2 activity in vascular smooth muscle cell tissues. Our results establish the importance of iPLA 2 in maintaining basal free arachidonic acid levels, contractions, and Ca 2ϩ sensitization in vascular smooth muscle tissues. Our findings may provide a molecular basis for developing new therapeutic agents for cardiovascular diseases associated with Ca 2ϩ sensitization.