Sustained Endothelial Nitric-oxide Synthase Activation Requires Capacitative Ca 2 1 Entry*

Endothelial nitric-oxide synthase (eNOS), a Ca 2 1 /cal-modulin-dependent enzyme, is critical for vascular ho-meostasis. While eNOS is membrane-associated through its N -myristoylation, the significance of membrane asso-ciation in locating eNOS near sources of Ca 2 1 entry is uncertain. To assess the Ca 2 1 source required for eNOS activation, chimera containing the full-length eNOS cDNA and HA-tagged aequorin sequence (EHA), and MHA (myristoylation-deficient EHA) were generated and transfected into COS-7 cells. The EHA chimera was primarily targeted to the plasma membrane while MHA was located intracellularly. Both constructs retained enzymatic eNOS activity and aequorin-mediated Ca 2 1 sensitivity. The plasma membrane-associated EHA and intracellular MHA were compared in their ability to sense changes in local Ca 2 1 concentration, demonstrat-ing preferential sensitivity to Ca 2 1 originating from intracellular pools (MHA) or from capacitative Ca 2 1 entry (EHA). Measurements of eNOS activation in intact cells revealed that the eNOS enzymatic activity of EHA was more sensitive to Ca 2 1 influx via capacitative Ca 2 1 entry than intracellular release, whereas MHA eNOS activity was more responsive to intracellular Ca 2 1 release. When eNOS activation by CCE was compared with that generated a 3 H] L -citrul-line quantified by scintillation spec- troscopy. Three independent done and are the of NOS same cofactors of 5 m M DAF-2. NOC-9 has a half-life of 3 min, and thus, solutions containing DAF-2 and 3 n M to 1 m M NOC-9 were studied 15 mixing in phosphate-buffered saline, representing . release of NO from NOC-9. Consistent with the prior publication of Kojima et al. we found that DAF-2 had a linear response physiological range of M NO

Nitric oxide (NO) 1 is an ubiquitous intracellular signaling molecule, synthesized from L-arginine by nitric-oxide synthase (NOS) in diverse cells and tissues. There are three isoforms of NOS, first identified in neural tissue (nNOS, NOS 1) (1, 2), endothelial cells (eNOS, NOS 3) (3,4), and activated macrophages and hepatocytes (iNOS, NOS 2) (5, 6), respectively. eNOS plays an major role in the control of blood pressure and vascular homeostasis. In the vascular endothelium, production of NO results in vascular smooth muscle relaxation which, in turn, reduces blood pressure. As expected, genetic ablation of the eNOS gene in mice results in systemic and pulmonary hypertension (7,8).
eNOS, a Ca 2ϩ /calmodulin-dependent enzyme, is highly regulated by intracellular Ca 2ϩ . Activation of eNOS is induced by increases in intracellular Ca 2ϩ resulting from the activation of diverse G-protein-coupled cell surface receptors or from mobilization of intracellular Ca 2ϩ stores. Previous studies have shown that thapsigargin (TG), a selective inhibitor of the Ca 2ϩ -ATPase on endoplasmic reticulum and sacroplasmic reticulum, activates NO release in pulmonary artery endothelial cells (PAEC) (9,10). However, little is known about whether this activation results from intracellular Ca 2ϩ release or from storeoperated or capacitative Ca 2ϩ entry (CCE) subsequent to depletion of the intracellular Ca 2ϩ pool. eNOS is predominantly localized to caveolae, a specialized microdomain of the plasma membrane, which serves to compartmentalize signal transduction molecules (3,11,12). The caveolaen-eNOS interaction serves both to partition eNOS in caveolae and inhibit eNOS enzymatic activity. After Ca 2ϩ -calmodulin-dependent phosphorylation, eNOS becomes dissociated from caveolaen and becomes enzymatically activated. We hypothesize that an additional function of eNOS binding to caveolaen is to localize the enzyme in close proximity to CCE channels on the plasma membrane, and that the actual intracellular Ca 2ϩ that is required to activate this enzyme may be significantly different from that released from subcellular compartments as well as average [Ca 2ϩ To test the hypothesis that CCE is essential for sustained activation of eNOS, EHA (eNOS-HA-aequorin) and MHA (myristoylation-deficient EHA) chimeras were generated and expressed in COS-7 and bovine PAEC. The EHA construct retained the ability to bind to caveolaen and target the plasma membrane, whereas the MHA construct was localized to the cytosol, and not to caveolae. Both constructs retained NOS and aequorin activity when expressed in COS-7 and PAEC. Using these two constructs, we investigated the relative contributions of release of intracellular Ca 2ϩ versus CCE to the regulation of [Ca 2ϩ ] in the region of eNOS and activation of NOS enzymatic activity under a variety of physiological conditions. We also measured NOS activation directly with an NO-sensitive fluorophore, comparing the ability to stimulate NO production of CCE induced by thapsigargin to that of generalized increase in [Ca 2ϩ ] i produced by ionomycin.

Generation of eNOS and Aequorin Chimeric
Constructs-All procedures involving oligonucleotide and cDNA manipulations were performed essentially as described by Sambrook et al. (13). Wild type eNOS (14) and myristoylation-deficient (myr Ϫ ) eNOS (15) cDNA in pBK-CMV vector (Stratagene, La Jolla, CA) were kindly provided by Dr. Thomas Michel (Cardiovascular Division, Brigham and Women's Hospital, Harvard Medical School, Boston, MA). eNOS and ACVI/HA/AEQ containing adenylyl cyclase type VI (ACVI) and hemagglutinin (HA) epitopetagged cytosolic aequorin (16) were used as templates for polymerase chain reaction (PCR) amplification. To generate the eNOS-aequorin chimeric constructs, overlapping PCR was employed using a fourprimer procedure (17). The sequence of the oligonucleotide primers were as follows: primer 1, 5Ј-primer (5Ј-CCGCTCGAGCGGGGCCA-CATG-3Ј) specific to eNOS including the XhoI restriction site (underlined) located at nucleotide 3341 of the bovine eNOS cDNA; primer 2, 3Ј-primer (5Ј-ataatcaggaacatcataGGGGCCGGGGGTGTCTGG-3Ј) specific to the last 18 bp of eNOS coding sequence (uppercase letters) and first 18 bp of HA epitope tag (YDVPDYASL) of ACVI/HA/AEQ (lowercase letters); primer 3, 5Ј-primer (5Ј-CCAGACACCCCCGGCCCCtatgatgttcctgattat-3Ј) specific to 18 bp of eNOS (uppercase letters) and HA tag (lowercase letters), respectively; primer 4, 3Ј-primer (5Ј-CCTCTA-GATTAGGGGACAGCTCCACC-3Ј) specific to the last 15 bp of cytosolic aequorin followed by a stop codon and XbaI site (underlined). The products of PCR1 (generated by primers 1 and 2) and PCR2 (generated by primers 3 and 4) were mixed in equimolar amounts, joined by the 36 bp of overlapping sequence (generated by the internal primers 2 and 3) and filled in by plaque forming unit polymerase (Stratagene); the combined product was amplified using the external primers 1 and 4 to generate the 900-bp PCR3 product. This product was subsequently digested with XhoI and XbaI restriction enzymes (Stratagene). To generate EHA construct, eNOS plasmid was digested with XhoI and XbaI to release 300 bp of the eNOS fragment which was replaced with the 900-bp XhoI/XbaI fragment of PCR3 product. The resulting constructs were sequenced in both directions through the entire sequence generated by PCR; no mutations were detected. Similar to the EHA construct, the MHA construct was generated by replacing the 300-bp XhoI/XbaI fragment of myr Ϫ -eNOS plasmid with the 900-bp XhoI/XbaI fragment from the EHA construct. The first 400 bp of both constructs were sequenced to confirm the point mutations resulting in glycine to alanine substitution at amino acid position 2 of the eNOS sequence in MHA construct. Both expression plasmids were transiently transfected into COS-7 cells and PAEC. The recombinant proteins were identified by immunoblotting with eNOS (Transduction Laboratories, Lexington, KY) or HA (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) antisera.
Cell Culture and Transfection-COS-7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin in an atmosphere of 5% CO 2 at 37°C. Bovine PAEC were isolated as described (18) and grown in minimal essential medium (Sigma) supplemented with 20% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin in an atmosphere of 5% CO 2 at 37°C; PAEC used in this study were between passages 3 and 6.
EHA and MHA expression plasmids were transiently transfected into PAEC and COS-7 cells using LipofectAMINE as described by the supplier (Life Technologies, Inc., Grand Island, NY). The transfection efficiency and eNOS expression were assessed by NADPH-diaphorase staining (19,20). Experiments were performed 48 h after transfection.
NOS Enzymatic Activity and Stimulation in Intact Cells-To determine the intracellular localization of EHA and MHA chimera, NOS enzymatic activity was assessed in subcellular fractions of COS-7 cells transfected with EHA or MHA constructs in the presence of excess substrates and cofactors as described previously (21,22). Briefly, 48 h after transfection, cells were incubated with 0.75 Ci/ml [ 3 H]L-arginine (Amersham Pharmacia Biotech) in the absence (basal) or presence of 100 nM TG and extracellular Ca 2ϩ at 37°C for 30 min followed by addition of 1 M trichloroacetic acid, and then lysed by freeze-thaw in liquid nitrogen. The samples were extracted three times with ether and passed through a Dowex AG50W X-8 column (Sigma). The [ 3 H]L-citrulline generated was collected and quantified by liquid scintillation spectroscopy. Three independent experiments were done and results are expressed as the percentage of basal NOS activity in the same plate. The measurements of NOS enzymatic activity in the presence of excess substrates and cofactors reveal total NOS abundance in subcellular compartments.
[Ca 2ϩ ] i Measurements with Aequorin-In vitro [Ca 2ϩ ] calibration curves were determined using the membrane fraction of EHA-transfected cells or cytosolic fraction of MHA-transfected cells as described (16,23). Cells were lysed and cytosolic fraction was separated from membrane fraction by centrifugation. Aequorin was reconstituted with 5 M coelenterazine (Molecular Probes, Inc., Eugene, OR) in the presence of 140 mM ␤-mercaptoethanol for 3 h at 4°C. Aequorin-mediated luminescence was measured with the LS50B Luminescence Spectrophotometer (Perkin-Elmer, Beaconsfield, United Kingdom) in the presence of various amounts of free Ca 2ϩ . The L max was obtained by integrating a continuous recording of aequorin-mediated light emission in the presence of 10 mM Ca 2ϩ .
In vivo measurements of the effect of TG on aequorin in transfected COS-7 and endothelial cells were performed as described previously (16). Cells were loaded with 5 M coelenterazine, washed, and resuspended in nominally Ca 2ϩ -free Krebs buffer (120 mM NaCl  (24). For studies of the effect of histamine, luminiscence measurements were done exactly as described previously by Mersault et al. (25) using a perfusion system. Briefly, the transfected cells, grown on 13-mm coverslips, were placed in a perfused, thermostatted chamber in direct apposition to a low-noise photomultiplier. The photomultiplier has a built-in amplifier-discriminator, the output of which is captured by a Thorn-EMI photon counting board and stored on an IBM-compatible computer for later analysis. Aequorin luminescence data was converted into [Ca 2ϩ ] values using an algorithm described previously (23,26). In all experiments, L max was obtained by integrating a continuous recording of aequorin-mediated light emission in the presence of 0.3% Triton X-100 and 10 mM Ca 2ϩ . Light output from unstimulated transfected cells loaded with coelenterazine was not significantly higher than background. The light emission from EHA-or MHA-transfected cells loaded in the absence of coelenterazine or from cells transfected with vector alone was not detected. To induce CCE, 100 nM TG or 100 M histamine were used as indicated in the figure legends.
[Ca 2ϩ ] i Measurements with Fura-2-Average cellular [Ca 2ϩ ] i in PAEC was fluorometrically measured using a Ca 2ϩ -sensitive fluorescent dye, fura-2. PAEC were loaded with 2 M fura-2 for 30 min, washed, and resuspended in nominally Ca 2ϩ -free buffer. 5 ϫ 10 6 cells were used for each measurement. TG was given as indicated. Fluorescence intensity at emission wavelength of 510 nm in response to excitation wavelengths of 340 and 380 nm was determined with the LS50B Spectrophotometer, and [Ca 2ϩ ] i values were calculated according to the formula of Grynkiewkz et al. (27).
Measurements of NO-To assess NO release from intact cells, NO x (NO 2 Ϫ ϩ NO 3 Ϫ ) in the culture media was measured with a chemiluminescence NO analyzer (Model 205, Sievers Instruments, Inc., Boulder, CO) as described previously by Wang et al. (10). PAEC or transfected COS-7 cells were incubated in the presence or absence of 100 nM TG in Krebs-HEPES buffer (99 mM NaCl, 4.69 mM KCl, 1.87 mM CaCl 2 , 1.2 mM MgSO 4 , 25 mM NaHCO 3 , 1.2 mM K 2 HPO 4 , 11.1 mM glucose, and 20 mM HEPES, pH 7.4) supplemented with 100 M L-arginine at 37°C in 95% O 2 plus 5% CO 2 . At the end of incubation, media was collected and 5-10 l of each sample was used for NO measurement. The amount of NO x was normalized to the protein content determined by Bradford assay.
To directly assess TG-mediated NO production, PAEC or transfected COS-7 cells on 25-mm coverslips were cultured in serum-free media for 24 h and loaded with 10 M DAF-2 DA (Calbiochem), a membranepermeable fluorescent NO indicator (28,29), in Krebs-Ringer phosphate buffer (120 mM NaCl, 4.8 mM KCl, 0.54 mM CaCl 2 , 1.2 mM MgSO 4 , 11 mM glucose, and 15.9 mM sodium phosphate, pH 7.2) at 37°C for 1 h. Cells were washed and placed in 1 ml of Krebs-Ringer phosphate buffer. After 1 min equilibration, TG was added to a final concentration of 100 nM followed by addition of 4 mM Ca 2ϩ . Fluorescence was measured with a fluorescence microscope (Olympus IMT-2, Tokyo, Japan) calibrated for excitation at 485 nm and emission at 520 nm.
The linearity of the response of DAF-2 to NO was assessed in vitro by generating known concentrations of NO with the NO donor NOC-9 (Calbiochem), measuring fluorescence intensity changes in the presence of 5 mM DAF-2. NOC-9 has a half-life of 3 min, and thus, solutions containing DAF-2 and 3 nM to 1 M NOC-9 were studied 15 min after mixing in phosphate-buffered saline, representing Ͼ95% release of NO from NOC-9. Consistent with the prior publication of Kojima et al. (29) we found that DAF-2 had a linear response in the physiological range of 0 -2.4 M NO (data not shown).
Statistics-Comparisons between groups were made using either paired or unpaired students t test, or ANOVA with Fisher post-hoc test. Data are presented as mean Ϯ S.E., with p Ͻ 0.05 accepted as significant.

Intracellular Localization of Recombinant eNOS-To assess
the Ca 2ϩ source for eNOS activation, eNOS-aequorin chimeric constructs were generated by attaching HA epitope-tagged aequorin to the COOH terminus of wild type bovine eNOS (EHA) or myr Ϫ -eNOS (MHA) (Fig. 1A), and the chimera were subsequently transfected into COS-7 cells. The recombinant proteins were identified by both eNOS monoclonal antibodies (Fig. 1B) and HA antibody (data not shown). In contrast to eNOS (M r ϭ 135,000), transfection of COS-7 cells with EHA or MHA resulted in detection of a single product of M r ϭ 150,000. To determine if the recombinant proteins were properly targeted and retained NOS activity, enzymatic activity assays were performed in subcellular fractions of COS-7 cells transfected with wild type eNOS and myr Ϫ -eNOS as well as chimeric EHA and MHA constructs. Equivalent NOS activity was detected in whole cell lysates from cells transfected with all four constructs ( Fig. 2A). Similar to wild type eNOS, NOS activity was 2.5-fold higher in the plasma membrane fraction than that in the cytosol from EHA-transfected cells (Fig. 2B). In contrast, NOS activity was primarily detected in cytosolic and intracellular membrane fractions from either myr Ϫ -eNOS or MHA-transfected cells (Fig. 2B), at a level that was 3-4-fold greater than that in the plasma membrane fraction (Fig. 2C). Overall, these results indicate that targeting of the chimeric EHA and MHA proteins is similar to that of eNOS and myr Ϫ -eNOS, respectively, in transfected COS-7 cells. In addition, NOS enzymatic activity is comparable in EHA-and MHA-transfected cells.
Comparison of EHA and MHA in Transfected COS-7 Cells-To localize aequorin activity in transfected cells, the cytosolic fraction was separated from the cell membrane fraction from cells transfected with EHA or MHA constructs. Consistent with the findings for NOS enzymatic activity (Fig. 2), 75% of the EHA luminescence was detected in the plasma membrane fraction, whereas 92% of the MHA activity was detected in the cytosolic fraction. The luminescence activity of EHA protein was not significantly different from that of MHA or ACVI/HA/AEQ chimera (data not shown), indicating that fusing the eNOS cDNA to the NH 2 terminus of aequorin did not alter the ability of aequorin as a Ca 2ϩ sensor.
In order to measure [Ca 2ϩ ] i with aequorin, an in vitro calibration curve was generated by measuring luminescence activity in response to differing amounts of EGTA-determined free Ca 2ϩ (Fig. 3). The plasma membrane fraction of EHA-transfected cells and the cytosolic fraction of MHA-transfected cells were used for the calibration. The relationship between [Ca 2ϩ ] and log (L/L max ) for EHA was not significantly different from that for MHA (Fig. 3) or from that for ACVI/HA/AEQ, adenylyl cyclase VI-aequorin fusion protein (data not shown). The results indicate that these aequorin constructs measure [Ca 2ϩ ] i in the physiological range of 10 Ϫ7 to 10 Ϫ5 M.
It has been shown that TG activates CCE subsequent to intracellular Ca 2ϩ pool depletion by inhibiting the microsomal Ca 2ϩ -ATPase (31,32). To determine if the membrane-associated EHA can detect the [Ca 2ϩ ] i change resulting from CCE, COS-7 cells were transfected with EHA or MHA constructs and subsequently treated with TG in the absence or presence of extracellular Ca 2ϩ (Table I). The addition of TG resulted in a transient increase in [Ca 2ϩ ] i in cells transfected with EHA or MHA. No change in luminescence was detected in cells transfected with vector alone (data not shown). Table I demonstrates that, similar to the previous finding that membrane-bound aequorin primarily sensed Ca 2ϩ influx while the cytosolic aequorin sensed Ca 2ϩ release (16), MHA tended to detect slightly higher [Ca 2ϩ ] i resulting from TG-mediated intracellular release than EHA (n ϭ 7, p ϭ 0.13), whereas EHA sensed higher [Ca 2ϩ ] in response to the addition of 4 mM extracellular Ca 2ϩ than MHA (n ϭ 7, p ϭ 0.04). For EHA, Ca 2ϩ influx resulted in greater [Ca 2ϩ ] i compared with intracellular release (p Ͻ 0.005). Importantly, the L max value was determined in each experiment and the luminescence value was normalized to L max to control for variable luminescence activity resulting from different levels of protein expression. Therefore, the differences detected between EHA and MHA solely reflects [Ca 2ϩ ] i at the region where the recombinant protein EHA or MHA is distrib- uted, rather than any variation in the level of expression of the two proteins.
In order to determine whether a physiological agonist could give rise to the differences in [Ca 2ϩ ] i reported by the two aequorin chimeras in response to release versus CCE, histamine, an agonist stimulating intracellular Ca 2ϩ release via the generation of inositol 1,4,5-triphosphate (33) was explored. To evoke the release of stored Ca 2ϩ , cells were stimulated with histamine in the absence of external Ca 2ϩ ; this was followed by the addition of external Ca 2ϩ (Fig. 4)   In keeping with these results, NO release was 1.5-fold higher in the presence of TG plus Ca 2ϩ versus TG alone in COS-7 cells transfected with EHA construct: 2.41 Ϯ 0.29 pmol/g of protein versus 1.65 Ϯ 0.21 pmol/g of protein (n ϭ 3, p ϭ 0.04). Alternatively, no difference was evident in cells transfected with MHA: 2.81 Ϯ 1.11 pmol/g of protein versus 2.45 Ϯ 0.87 pmol/g of protein (n ϭ 3, p ϭ 0.14).
Comparison of EHA and MHA in PAEC-The argument can be made that overexpressed eNOS is restricted to a selective domain of the cells which is different from the localization of endogenous eNOS. To address this issue, TG-mediated activation of NOS was tested in bovine PAEC which express endogenous eNOS and PAEC transfected with EHA or MHA constructs. To confirm that cultured PAEC retained TG-mediated CCE, fura-2 was used to assess global [Ca 2ϩ ] i in these cells (Fig. 5). Treatment with TG resulted in a Ca 2ϩ release transient in the range of 60 nM; addition of extracellular Ca 2ϩ between 0.5 and 6 mM induced a rapid Ca 2ϩ entry and elevation in [Ca 2ϩ ] in the range of 40 to 140 nM. Thus, fura-2 results confirm the presence of CCE in cultured bovine PAEC.
To determine if a similar signaling pathway exists for endogenous eNOS, NO production was directly measured in bovine PAEC. Endothelial cells grown in a monolayer produce very low levels of NO which are below the detection range of traditional assays. Therefore, a new fluorescent NO indicator, DAF-2 DA, was employed to measure NO production in real time (Fig. 6). PAEC had detectable basal NO production which increased a small amount (Fig. 7, n ϭ 10, p Ͻ 0.05) after TG treatment in the presence of low extracellular Ca 2ϩ (i.e. 0.54 mM). In contrast, a large increase in NO production (n ϭ 10, p Ͻ 0.001) was detected in the presence of 4 mM extracellular Ca 2ϩ , suggesting that CCE regulates wild type eNOS activity. Use of aequorin chimera to detect subcellular [Ca 2ϩ ] i in PAEC revealed, similar to COS-7 cells, that TG-mediated Ca 2ϩ entry was higher in PAEC transfected with EHA than that with MHA: 5 Fig. 7 demonstrates, the increase in NO production with ionomycin was approximately 10% of that produced by TG, despite comparable increases in average cellular [Ca 2ϩ ] i (n ϭ 10, p Ͻ 0.05 from TG).

DISCUSSION
Due to the key role of Ca 2ϩ in modulating eNOS activity, the current study was undertaken to investigate the effect of CCE on subcellular Ca 2ϩ concentration and eNOS activation. The direct measurement of [Ca 2ϩ ] i concentration in specific subcellular compartments of intact cells has become possible due to cloning of the Ca 2ϩ -sensitive luminescent protein, aequorin (36,37). This protein has been targeted to mitochondria (38), endoplasmic reticulum (26), and nucleus (39,40) by attaching an organelle-specific targeting sequence to aequorin cDNA. A similar strategy was employed here to specifically target aequorin to the plasma membrane where wild type eNOS is located. The eNOS-aequorin constructs, including EHA and MHA, were fully functional in terms of appropriate intracellular targeting and enzyme activity of both eNOS and aequorin. They not only had similar NOS activity as their parental molecules, but they had the same affinity to Ca 2ϩ as the previously reported membrane-targeted ACVI/HA/AEQ chimeric protein (16).
The EHA construct was primarily targeted to the plasma [Ca 2ϩ ] i was determined in aliquots of 5 ϫ 10 6 fura-2-loaded PAEC as described under "Materials and Methods." CCE was induced by depleting intracellular Ca 2ϩ pools with 100 nM TG in nominally Ca 2ϩ -free buffer at 60 s, followed by the addition of various amounts of external Ca 2ϩ as indicated at 240 s, which resulted in a dose-dependent increase of [Ca 2ϩ ] i . membrane while MHA was localized intracellularly, similar to myr Ϫ -eNOS. Previous studies have shown that N-myristoylation is essential for targeting eNOS to the plasma membrane (15,41) and that phosphorylation results in a decrease of NOS activity, which is associated with the translocation of eNOS from the plasma membrane to cytosol (34,35). In both endothelial and COS-7 cells membrane targeting restricts eNOS to caveolae (21). These findings suggest that eNOS is regulated by its intracellular localization and that the site of NO synthesis likely affects the function of NO as a messenger molecule. Thus, EHA and MHA constructs were used to directly measure [Ca 2ϩ ] i in the region of eNOS, testing whether localization to the plasma membrane places eNOS in a microdomain in which CCE modulates its activity.
TG and histamine have been widely used to study the regulation of intracellular Ca 2ϩ pools in many cell types. TG elevates [Ca 2ϩ ] i in PAEC by specifically inhibiting the Ca 2ϩ -ATPase activity of the endoplasmic reticulum without affecting the plasmalemmal Ca 2ϩ -ATPase, ultimately depleting intracellular Ca 2ϩ stores. This, in turn, leads to secondary CCE (42,43). Similar to PAEC, in COS-7 cells TG-induced Ca 2ϩ mobilization involving both release of Ca 2ϩ from intracellular stores and CCE, since the sustained phase of the Ca 2ϩ transient was ablated in the absence of [Ca 2ϩ ] ex . Previous studies have demonstrated that reduction of [Ca 2ϩ ] ex attenuates TG-mediated NO release in PAEC, suggesting a role of Ca 2ϩ influx in NOS activation and a close correlation between cytoplasmic Ca 2ϩ and NO release (10,44,45). Therefore, the presence of extracellular Ca 2ϩ is essential for the refilling of the intracellular pools, as well as for TG-mediated NO production. However, lack of measurement of subcellular [Ca 2ϩ ] i in prior studies made it impossible to directly establish the contribution of CCE to modulation of [Ca 2ϩ ] i in the region of eNOS.
In the present study, quantification of Ca 2ϩ influx was achieved by measuring aequorin-mediated luminescence of specifically targeted EHA. The primarily membrane-bound EHA was exposed to [Ca 2ϩ ] i in the range of 4 M resulting from TG-mediated CCE. EHA detected higher [Ca 2ϩ ] i in response to Ca 2ϩ entry while the cytosolic MHA detected a relatively greater increase due to intracellular Ca 2ϩ release, consistent with the hypothesis that CCE is a major contributor to changes in [Ca 2ϩ ] i in the region of membrane-bound eNOS. Studies using histamine were similar to those with TG, suggesting that these findings are generalizable to physiological stimuli. Similar results were obtained in a previous study in which ACVIaequorin fusion protein, another membrane-associated Ca 2ϩsensitive protein, sensed a much higher [Ca 2ϩ ] i at the plasma membrane than was seen by cytosolic aequorin (16).
To determine the effect of CCE on eNOS function, NO production was assessed. CCE stimulated an increase in NO production in COS-7 cells transfected with EHA, but not in cells transfected with MHA. In contrast, Ca 2ϩ release from intracellular stores had a greater effect on the cytosolic MHA than the wild-type EHA. The results of this study provide the first direct evidence that membrane-associated eNOS is different from cytosolic myr Ϫ -eNOS in its sensitivity to intracellular Ca 2ϩ originating from CCE versus internal storage pools. Furthermore, CCE not only regulates local [Ca 2ϩ ] i but activates NOS enzymatic activity of membrane-bound EHA in transfected COS-7 cells.
Cultured bovine PAEC were studied to determine the relevance of observations in COS-7 cells to the principle cell expressing eNOS in the circulation, the endothelial cell. Measurements of global [Ca 2ϩ ] i using fura-2 confirmed the presence of CCE in cultured bovine PAEC. In order to measure subcellular [Ca 2ϩ ] i in PAEC, EHA and MHA chimera were transfected and luminescence measured. Membrane-associated EHA reported regional [Ca 2ϩ ] in the range of 6 M after TG in the presence of extracellular Ca 2ϩ , significantly greater than the 2 M concentration to which cytosolic MHA was exposed. Consistent with a role for CCE in stimulating eNOS activity in PAEC, direct NO measurement with DAF demonstrated that NO production was 4-fold greater after TG in the presence of extracellular Ca 2ϩ versus Ca 2ϩ free conditions.
To summarize these results, we found that upon activation of CCE by TG or histamine the membrane-targeted EHA construct was stimulated to produce increased NO. In contrast, the MHA construct, which does not target the plasma membrane, was less sensitive to generalized elevation in [Ca 2ϩ ] i . This suggests that membrane targeting results in localization to a unique microdomain, presumably caveolae, in which eNOS is exposed to higher regional [Ca 2ϩ ]. Our measurements of [Ca 2ϩ ] using the aequorin moiety of the fusion proteins comfirms this hypothesis. The EHA construct reported regional [Ca 2ϩ ] of 4-6 M, whereas average [Ca 2ϩ ] i estimated by fura-2 fluorescence was in the submicromolar range. In keeping with this hypothesis, the MHA construct sensed lower regional [Ca 2ϩ ] during the sustained phase of the TG or histamine transient than did the EHA construct.
Taken together, these results support the hypothesis that CCE, stimulated by either TG or histamine, is the stimulus for eNOS activation. However, the difference between NOS activation in the presence versus absence of extracellular Ca 2ϩ in TG-stimulated PAECs could also be explained simply by a lack of sustained Ca 2ϩ influx from any source, and not uniquely to CCE. To address this point, we compared the ability to stimulate eNOS activity of TG to that of ionomycin. The reasoning behind this experiment was that in contrast to TG or histamine, ionomycin produces an increase in Ca 2ϩ influx which is not localized to a specific plasmalemmal microdomain, and stimulates much less CCE (46 ] i , NO production in response to TG-mediated CCE was 10-fold greater than that produced by ionomycin. Thus, a comparable increase in Ca 2ϩ influx and [Ca 2ϩ ] i was not sufficient to recapitulate the effect of CCE stimulated by TG or histamine. This finding is consistent with the prior report of Nakahashi et al. (16) with ACVI/AEQ in HEK293 cells, demonstrating that much higher levels of [Ca 2ϩ ] i in response to ionomycin are required to stimulate comparable amounts of CCE and inhibition of ACVI activity.
In conclusion, the current study demonstrates that membrane-associated wild type eNOS is primarily responsive to Ca 2ϩ entry, whereas myr Ϫ -eNOS is more responsive to intracellular Ca 2ϩ release. In intact endothelial cells, NO production is preferentially stimulated by CCE, and relatively insensitive to the non-localized influx of Ca 2ϩ produced by ionomycin. These results strongly support the CCE model proposed by Putney (30,47) and suggest that the localization of wild type eNOS to plasma membrane caveolae places eNOS in close proximity to CCE channels. This co-localization appears to be critically involved in the rapid stimulation of the enzyme.