Differences in eNOS Activity Because of Subcellular Localization Are Dictated by Phosphorylation State Rather than the Local Calcium Environment*

Nitric oxide (NO) produced in the endothelium via the enzyme endothelial nitric-oxide synthase (eNOS) is an important vasoactive compound. Wild-type (WT) eNOS is localized to the plasma membrane and perinuclear/Golgi region by virtue of N-terminal myristoylation and palmitoylation. Acylation-deficient mutants (G2AeNOS) remain cytosolic and release less NO in response to Ca2+-elevating agonists; a disparity that we hypothesized was attributed to the greater distance between G2AeNOS and plasma membrane Ca2+ influx channels. The reduced activity of G2AeNOS versus WT was reversed upon disruption of cellular integrity with detergents or sonication. NO production from both constructs relied almost exclusively on the influx of extracellular Ca2+, and elevating intracellular Ca2+ to saturating levels with 10 μm ionomycin in the presence of 10 mm extracellular Ca2+ equalized NO production. To identify the contribution of calcium to the differences in activity between these enzymes, we created Ca2+/CaM-independent eNOS mutants by deleting the two putative autoinhibitory domains of eNOS. There was no difference in NO production between WT and G2A-targeted Ca2+-independent eNOS, suggesting that Ca2+ was the factor responsible. When eNOS constructs were fused in-frame to the bioluminescent probe aequorin, membrane-bound probes were exposed to higher [Ca2+] in unstimulated cells but upon ionomycin stimulation, the probes experienced equal amounts of Ca2+. The WT and G2A enzymes displayed significant differences in the phosphorylation state of Ser617, Ser635, and Ser1179, and mutating all three sites to alanine or restoring phosphorylation with the phosphatase inhibitor calyculin abolished the differences in activity. We therefore conclude that the disparity in NO production between WTeNOS and G2AeNOS is not caused by different localized [Ca2+] upon stimulation with ionomycin, but rather differences in phosphorylation state between the two constructs.

Nitric oxide (NO) 2 produced in the endothelium via the enzyme endothelial nitric-oxide synthase (eNOS) is an important vasoactive compound (1). Loss of endothelial function, and particularly disturbance of eNOS-derived NO production, has been implicated in a number of cardiovascular diseases, and represents a promising therapeutic tar-get (2,3). The production of NO is a tightly regulated event and is influenced in the short term by a series of post-translational modifications (4). These include modulation of phosphorylation state by kinases and phosphatases (5)(6)(7)(8)(9), protein-protein interactions (4), and availability of requisite substrates and cofactors for eNOS function, including L-arginine, tetrahydrobiopterin (BH 4 ), iron (Fe), FMN, FAD, NADPH, and calcium/calmodulin (1). Finally, eNOS activity can also be affected by the intracellular localization of the enzyme (10).
The subcellular localization of wild-type eNOS (WTeNOS) is dictated by post-translational fatty acylation, specifically myristylation of the glycine at position 2, which then leads to palmitoylation of the cysteines at positions 15 and 26 (11). This fatty acylation has been reported to target eNOS to both the plasma membrane and the perinuclear/Golgi region (12,13). Mutant eNOS constructs, in which the glycine at position 2 is mutated to alanine (G2AeNOS), fail to undergo fatty acylation, and as a consequence remain cytosolic (14). This mistargeting has profound consequences for NO production, as both basal and stimulated NO production by G2AeNOS is markedly reduced compared with WTeNOS (14,15). However, when the WTeNOS and G2AeNOS constructs are purified and examined in vitro with maximal substrate and cofactors, no differences in catalytic activity are observed (14). The mechanisms underlying the difference in catalytic activity in distinct intracellular environments remain unclear.
When inducible NOS (iNOS) is expressed with the targeting sequences for both WTeNOS and G2AeNOS, the production of NO is completely unaffected by intracellular location (16). As iNOS and eNOS require similar amounts of L-arginine, BH 4 , and NADPH for activity, it seems unlikely that local variations in these cofactors are responsible for the altered activity of targeted eNOS constructs. In contrast, iNOS and eNOS are diametrically opposed in their dependence on Ca 2ϩ /CaM-eNOS is highly dependent on increases in intracellular Ca 2ϩ for activity, whereas iNOS is already maximally activated by Ca 2ϩ /CaM even at basal levels of intracellular Ca 2ϩ , and is thus to all practical purposes Ca 2ϩ -independent (17). Therefore, intracellular [Ca 2ϩ ] represents a plausible candidate that mediates the differences in NO production observed between WTeNOS and G2AeNOS.
One of the major sources of Ca 2ϩ entry into non-excitable cells is the opening of non-selective cation channels in the plasma membrane in response to the depletion of intracellular Ca 2ϩ stores, a phenomenon known as capacitative calcium entry (CCE) (18). CCE is the predominate source of agonist-regulated Ca 2ϩ entry into endothelial cells and is also believed to play a major role in shear stress-stimulated Ca 2ϩ entry (19 -21). It is now widely accepted that the influx of external Ca 2ϩ , specifically that provided by CCE, is also the predominate source of the Ca 2ϩ required by eNOS for sustained activation and NO production (22). Elevations in intracellular Ca 2ϩ , particularly in response to Ca 2ϩ influx, are rarely homogeneous throughout the cell. For instance, invok-ing CCE in A7r5 cells produces a rather modest increase in cytosolic Ca 2ϩ , while producing localized [Ca 2ϩ ] up to 300 times greater at the plasma membrane (23). Indeed, it has previously been demonstrated that the peak Ca 2ϩ transient response reported upon invoking of CCE is higher for WTeNOS than a myristylation-deficient eNOS expressed in COS cells (22). However, whether this disparity in local [Ca 2ϩ ] is sufficient to explain the large differences in NO production between WTe-NOS and G2AeNOS over sustained periods of time remains to be elucidated. Indeed, NO production and calcium elevation are elevated above baseline several minutes beyond the initial influx of CCE-driven calcium (24,25). Therefore, the aim of the present study was to determine whether proximity to plasma membrane Ca 2ϩ entry channels and the subsequent differences in localized [Ca 2ϩ ] experienced by the plasma membrane and the cytosol, were the sole determinant of the disparity between the NO production by membrane-bound or cytosolic eNOS constructs.
Generation of eNOS Constructs-A cytosolic bovine eNOS construct (G2AeNOS) was constructed by mutating the alanine at position 2 of wild-type bovine eNOS (WTeNOS) to a glycine, preventing fatty acylation of the enzyme, as previously described (14). The eNOS constructs with reduced Ca 2ϩ /CaM dependence were based on a Ca 2ϩ /CaM-independent construct (⌬45/⌬14eNOS) of human eNOS first characterized by Chen and Wu (26). Two significant portions of the sequence of human eNOS that were not shared by the Ca 2ϩ -insensitive NOS isoform iNOS, specifically the residues of the two putative autoinhibitory control elements: 594 -606/614 -645 (⌬45) and 1165-1178 (⌬14) were deleted. The constructs for the present study were created by deleting the corresponding residues in bovine eNOS, specifically 596 -608/616 -647 (⌬45) and 1167-1180 (⌬14). cDNA fragments containing the mutated regions were generated by PCR using WTeNOS cDNA as a template, and a pair of primers were used for each region: 5Ј-GAAGC-GGATCTTGTAACTCTTCATCTCCATCAG-GGCAG-3Ј and 5Ј-G-AGTTACAAGATCCGCTTCAGGTTCTGTGTGTTCGGACTGG-GC-3Ј for the ⌬45 mutation, 5Ј-GATGCCGCTCCTGCAGGGAAAA-CGTGAGGCCGAAAATGTC-3Ј and 5Ј-CCTCACGTTTTCCCTG-CAGGAGCGGCATCTG-3Ј for the ⌬14 mutation. Each fragment was then sequentially subcloned into WTeNOS or G2AeNOS cDNA.
Phospho-null eNOS mutants were generated by mutating serines 617, 635, and 1179 to alanine on bovine eNOS using sequential PCR mutagenesis with the primers previously described (5,27), and then subcloning the various fragments into the WT and G2A targeting sequence construct. The sequences of the phospho-null eNOS and the Ca 2ϩ -independent eNOS constructs were confirmed by DNA sequencing at the Genomics Core Facility of the Medical College of Georgia.
NO Release from Intact Cells-Cells were transfected with the constructs of interest, and 48 h later were stimulated with the Ca 2ϩ ionophore ionomycin (1 M) for 30 min. Medium containing NO (primarily NO 2 Ϫ ) was ethanol-precipitated to remove proteins and was refluxed in sodium iodide/glacial acetic acid to convert NO 2 Ϫ to NO. NO was then quantified via specific chemiluminescence following reaction with ozone (Sievers). All NO measurements were corrected by subtracting the signal detected from cells transfected with empty vector (pCDNA3). NO Production in Disrupted Cell Preparations-COS-7 cells transfected with either WTeNOS or G2AeNOS were scraped into 1.5-ml Eppendorf tubes, pelleted, and resuspended at a density of ϳ1 ϫ 10 7 cells/ml in fresh serum-free DMEM supplemented with NADPH (100 M). Cells undergoing sonication were then subjected to 3 ϫ 1-s bursts with a sonic dismembrator (Fisher Model 100, setting: 3), after which the lysates were aliquoted and incubated for 30 min at 37°C. Detergenttreated samples were treated with 1% Triton X-100 and then incubated for 30 min at 37°C. The reaction was terminated by mixing samples with 2ϫ volume of cold absolute ethanol before analysis for NO content (see above).
Generation of Aequorin (AEQ) Constructs-Four AEQ constructs were created; an HA-tagged aequorin for measurement of cytosolic Ca 2ϩ (CYT-AEQ), an AEQ fused in-frame to the transmembrane domain of the T-cell surface glycoprotein CD8 (accession NM_001768) for measurement of Ca 2ϩ at the plasma membrane (PM-AEQ), and AEQ fused in-frame to either WTeNOS (WTeNOS-AEQ) or G2AeNOS (G2AeNOS-AEQ) for measurement of the Ca 2ϩ experienced by each construct (28).
Calibration of Aequorin Constructs to [Ca 2ϩ ]-COS-7 cells were transfected with the AEQ fusion proteins, and 48 h later, the cells were homogenized by sonication in medium of the following composition: 50 mM Tris-HCl, pH 7.5, 0.1 mM EGTA, 0.1 mM EDTA, 10 mM NaF, 1 mM orthovanadate, and 1ϫ protease inhibitor mixture (Sigma). The lysates were then exposed to EGTA-buffered solutions of known Ca 2ϩ concentrations, followed by total discharge of the AEQ with 10 mM CaCl 2 . The ratio of luminescence in the one second immediately after exposure to the Ca 2ϩ solution was calculated as a proportion of the total luminescence emitted over the duration of the experiment (L/L max ) and plotted against each known concentration of Ca 2ϩ to obtain a calibration curve.
Ca 2ϩ Measurements Using Aequorin in COS Cells-COS-7 cells were transfected with one of the four aequorin constructs, and 48 h later, the aequorin was activated by incubating the cells in Ca 2ϩ -free DMEM (BIOSOURCE) containing 5 M coelenterazine (Sigma) for 1 h. Following this, loading medium was replaced with Ca 2ϩ -free DMEM without coelenterazine. The cells were then exposed to Ca 2ϩ -free DMEM, Ca 2ϩ -DMEM (DMEM containing 1.5 mM free Ca 2ϩ ), or Ca 2ϩ -DMEM with 1 M ionomycin. Cells were placed in a luminescence plate reader (Lumistar Galaxy) and after the appropriate treatment time the remaining luminescence (L max ) was discharged by exposing the cells to a lysis buffer consisting of 10 mM Ca 2ϩ and 0.3% Triton X-100 in dH 2 O, and then expressed as a ratio of the total luminescence obtained from untreated cells (t ϭ 0). AEQ and its cofactor coelenterazine react in the presence of Ca 2ϩ to emit a photon. This reaction results in the permanent oxidation of coelenterazine to coelenteramide, and therefore the emission of light in response to Ca 2ϩ is a once-only reaction. The size of the available pool of luminescence therefore decreases constantly over time in direct proportion to the amount of Ca 2ϩ present. We, therefore, interpreted the decrease in L max as being indicative of exposure to Ca 2ϩ over time, with a greater degree of burnout indicating that the AEQ probe had been exposed to more Ca 2ϩ .
Immunoprecipitation and Western Blotting-Cells were washed twice with phosphate-buffered saline and lysed in homogenization buffer of the following composition: 50 mM Tris-HCl, 0.1 mM EGTA, 0.1 mM EDTA, 0.1% SDS, 1% Nonidet P-40, 0.1% deoxycholic acid. Lysates were homogenized and cleared by centrifugation. After determining protein concentration (DC protein assay kit, Pierce), equal amounts of protein were incubated with an excess of eNOS monoclonal antibody for 2 h at 4°C. Protein A-Sepharose was then added, and the mixture incubated for another 1 h at 4°C. The immune complexes were washed three times with lysis buffer and boiled in SDS-PAGE sample buffer (5 min), and then size-fractionated by SDS-PAGE. Blots were transferred to nitrocellulose membrane in standard transfer buffer for 18 h at 4°C and 40 V, and then blocked in nonfat milk for 1 h. After blocking, membranes were incubated with phosphospecific antibodies raised to the individual phosphorylation sites (phospho-116 and phospho-Ser 617 , Upstate; phospho-Thr 497 , Cell Signaling Technologies; phospho-Ser 635 and phospho-Ser 1179 , BD Transduction Laboratories) or with anti-eNOS monoclonal antibody (9D10, Zymed Laboratories) for 2 h, washed, incubated with secondary antibody conjugated to horseradish peroxidase (anti-mouse, Amersham Biosciences), washed again, and then developed using chemiluminescence (Amersham Biosciences).
2Ј,5Ј-ADP-Sepharose Isolation of eNOS Constructs-Cells were washed twice with phosphate-buffered saline and homogenized as for the AEQ calibration experiments (see above). When required, membrane and cytosolic fractions were separated by centrifugation (100,000 ϫ g, 90 min; Beckman ultracentrifuge), and the membrane fraction was resuspended in homogenization buffer supplemented with 1% Nonidet P-40. 2Ј,5Ј-ADP-Sepharose beads (Amersham Biosciences) were then added to the fractions and rotated at 4°C for 3 h. Beads were then washed three times in homogenization buffer before being prepared for Western blotting (see above).
Concentration-Response Curves to Ca 2ϩ /CaM with Isolated eNOS Constructs-COS-7 cells expressing WT and G2A eNOS constructs were homogenized and eNOS affinity-purified with 2Ј,5Ј-ADP-Sepharose beads (as above). Following the 3-h rotation at 4°C, beads were washed twice in HEPES-buffered saline (HBS; 135 mM NaCl, 5.9 mM KCl, 1.2 mM MgCl 2 , 11.6 mM HEPES, 11.5 mM glucose, pH 7.3) then twice in HBS supplemented with EGTA (2 mM) to remove all Ca 2ϩ / CaM from the sample, then washed twice again in HBS. Beads were then resuspended in treatment solutions consisting of HBS supplemented with NADPH (3 mM), BH 4 (30 M), L-arginine (100 M), CaCl 2 (1 mM), and varying concentrations of calmodulin (bovine heart calmodulin, Sigma). Reactions were terminated after 30 min by addition of 2ϫ volume of ice-cold EtOH. The relative production of NO was determined by chemiluminescence as described above.
Confocal Imaging-To visualize the subcellular location of WT and G2A eNOS and eNOS-aequorin fusion protein within cells, we constructed additional fusion proteins encoding WT and G2A eNOS-GFP and WT and G2A-RFP-AEQ. COS-7 cells were then transfected with cDNAs encoding the eNOS fusion proteins of EGFP and monomeric RFP. 24 -48 h later, cells were replated onto glass-bottomed culture dishes (MatTek). All imaging was performed using the LSM 510 Meta 3.2 Confocal Microscope (Zeiss). Magnification power was set at ϫ40 with oil.
Statistical Analysis-NO release data were expressed as mean Ϯ S.E. AEQ burnout data were individually normalized to the L max obtained for each AEQ fusion construct in Ca 2ϩ -free DMEM, then expressed as mean Ϯ S.E. Overall differences between groups were analyzed using a two-way analysis of variance (ANOVA) with Student-Newman-Keuls posthoc analysis for determining differences between the means when more than two groups were compared. An independent Student's t test was used when only two groups were compared. Comparisons between the Ca 2ϩ calibration curves of each AEQ fusion construct were made using 2-way ANOVA. Differences were considered significant if p Ͻ 0.05.

The Disparity in NO Production between WTeNOS and G2AeNOS Is
Abolished by Cellular Disruption-When COS-7 cells were transfected with equal amounts of WTeNOS or G2AeNOS (Fig. 1a) and stimulated for 30 min with three Ca 2ϩ -mobilizing compounds: the Ca 2ϩ ionophore ionomycin (1 M), the SERCA inhibitor thapsigargin (100 nM), or extracellular ATP (100 M), G2AeNOS produced significantly less NO than the WTeNOS (Fig. 1b). However, when cells expressing both constructs were disrupted either by sonication (3 ϫ 1-s bursts, setting 3) or detergent (Triton X-100, 1%) in medium supplemented with NADPH (100 M), G2AeNOS actually produced more NO than the WTeNOS construct (Fig. 1c), suggesting that the disparity in NO production between the constructs was reliant on intracellular localization.
NO Production by Both WTeNOS and G2AeNOS Is Ca 2ϩ -dependent, and Raising Intracellular Ca 2ϩ to Saturating Levels Equalizes NO Production-Previous studies have suggested that a contributing factor to the disparity in NO production between membrane-bound WTeNOS FIGURE 1. WTeNOS produces more NO than G2AeNOS in intact cells but not in disrupted cells. a, COS-7 cells were transfected with WT and G2A eNOS and immunoblotted for total eNOS. b, NO release was measured from WT and G2A eNOS-transfected COS-7 cells that were stimulated for 30 min with ionomycin (IM, 1 M), thapsigargin (TG, 100 nM), or extracellular ATP (100 M) by NO-specific chemiluminescence. c, NO release from Triton X-100 (1%) or sonication (3 ϫ 1-s bursts) disrupted COS-7 expressing WT and G2A eNOS. Cell lysates were incubated for 30 min in serum-free DMEM supplemented with 100 M NADPH. Data are presented as means Ϯ S.E. (n ϭ 4 -5). *, p Ͻ 0.05 versus WTeNOS. and the myristylation-deficient, cytosolic G2AeNOS may be the proximity of each construct to store-operated Ca 2ϩ -entry channels (22); a hypothesis supported by our finding that disruption of cellular integrity abolished the differences in NO production. To further investigate this hypothesis, it was first necessary to confirm that the entry of extracellular Ca 2ϩ was the dominant stimulus for NO production by both constructs. Cells transfected with either WTeNOS or G2AeNOS were stimulated with 1 M ionomycin (a concentration that is reported to increase intracellular Ca 2ϩ via activation of CCE and not direct opening of plasma membrane channels, Ref. 29) in serum-free DMEM (1.5 mM free Ca 2ϩ ), in serum-free DMEM after preincubation with the intracel-lular Ca 2ϩ chelator BAPTA-AM (10 M, 20 min), or in serum-free and Ca 2ϩ -free DMEM supplemented with 1 mM EGTA. In the presence of extracellular Ca 2ϩ , G2AeNOS produced approximately half the amount of NO produced by WTeNOS (Fig. 2a). Preincubation with BAPTA-AM significantly decreased NO production by both constructs, but decreased NO production by G2AeNOS to a greater extent than that of WTeNOS (54% reduction in G2AeNOS activity versus only 20% reduction in WTeNOS activity). In contrast, the removal of extracellular Ca 2ϩ reduced the ionomycin-induced NO production by both constructs by greater than 90% (Fig. 2a), demonstrating that both constructs rely almost entirely on the influx of external Ca 2ϩ for sustained NO production.
If the differences in NO production were due solely to differences in Ca 2ϩ interaction with the two constructs and not caused by inferior catalytic activity of the G2AeNOS construct, then elevating cytosolic [Ca 2ϩ ] to saturating levels would be expected to fully activate both constructs and equalize NO production. When cells expressing WTe-NOS or G2AeNOS were stimulated with 1 M ionomycin in the presence of increasing concentrations of external Ca 2ϩ , the NO production by G2AeNOS increased as a proportion of the NO production by WTe-NOS (38% at 1.5 mM Ca 2ϩ and 56% at 10 mM, Fig. 2b) but did not equalize. When the higher concentration of 10 M ionomycin was used to stimulate cells, NO production by G2AeNOS again increased in proportion to that of WTeNOS, and elevating extracellular [Ca 2ϩ ] to 10 mM equalized NO production (58% at 1.5 mM Ca 2ϩ and 97% at 10 mM; Fig. 2c). The ability of saturating concentrations of Ca 2ϩ to equalize the NO production by the two constructs again suggested that differences  in [Ca 2ϩ ] handling, and not an inability of the G2A construct to produce NO per se, might be the key regulatory factor mediating the differences in activity. Additionally, treatment of cells with the superoxide scavenger Tiron (10 mM, 30 min) had absolutely no effect on the NO production by either construct, suggesting that the disparity in NO production was not caused by aberrant production of superoxide (data not shown).
The Activity of Ca 2ϩ /CaM-insensitive Constructs Is Unaffected by Subcellular Localization-Once we had confirmed the dependence of both constructs on extracellular Ca 2ϩ entry, and that NO production could be equalized simply by increasing intracellular Ca 2ϩ to saturating levels, we attempted to demonstrate a direct link between intracellular Ca 2ϩ and the altered NO production by the two constructs. To achieve this we created Ca 2ϩ /CaM-independent deletion mutants of the WTe-NOS and G2AeNOS constructs based on the constructs of Chen and Wu (26). Both deletion mutants were found to be almost entirely Ca 2ϩ / CaM-independent as, in contrast to the native constructs, they displayed extremely high basal NO production. This was unaffected by chelation of extracellular Ca 2ϩ (Fig. 3a) and only a slight (although significant) increase in NO production in response to ionomycin stimulation, which was sensitive to removal of extracellular Ca 2ϩ (Fig. 3b). The WTeNOS and G2AeNOS deletion mutants retained the same membrane/cytosol distribution as the parent enzymes (data not shown). However, in contrast to the original constructs, the WTeNOS and G2AeNOS Ca 2ϩ /CaM-independent mutants produced equivalent amounts of NO both basally and in response to ionomycin stimulation (Fig. 3, a and b). The ability of the Ca 2ϩ /CaM-independent constructs to produce equal amounts of NO regardless of their subcellular localization strongly suggested that the disparity between the NO production of the original WTeNOS and G2AeNOS constructs involved the interaction of the constructs with Ca 2ϩ /CaM.
WTeNOS Experiences Higher [Ca 2ϩ ] in Unstimulated but Not Stimulated Cells-As our observations so far had concurred with the hypothesis that the disparity between NO production by WTeNOS and G2AeNOS was caused by differences in localized [Ca 2ϩ ], we attempted to test the hypothesis by measuring the localized Ca 2ϩ experienced by each construct in response to CCE. Aequorin fusion constructs were constructed as discussed under "Materials and Methods" (Fig. 4a) and were evaluated to ensure that none of the modifications altered the sensitivity of the AEQ to Ca 2ϩ (Fig. 4b). To ensure that the targeting of the eNOS constructs was unaffected by the fusion to AEQ, we examined the subcellular distribution of the WTeNOS and WTeNOS-AEQ constructs expressed in COS-7 cells. When cell lysates were separated into cytosolic and membrane fractions by centrifugation, both WTeNOS and WTeNOS-AEQ were found in equal proportion, primarily in the membrane fraction. However, the bulk of the G2AeNOS and G2AeNOS-AEQ protein was found in equal proportion in the cytosolic fraction (Fig. 5a). To further confirm that AEQ did not disrupt eNOS targeting we generated eNOS-GFP and eNOS-AEQ-RFP fusion proteins. In COS-7 cells transfected with WT and G2A variants of eNOS-GFP and eNOS-AEQ-RFP, we performed confocal microscopy to visualize the subcellular distribution of eNOS fusion proteins. As shown in Fig. 5b, G2AeNOS is found distributed throughout the cytosol, colocalizing completely with G2A-eNOS-AEQ. In contrast, the WTeNOS is found at discrete intracellular locations, primarily at the perinuclear Golgi (arrows) with less expression at the plasma membrane (arrowheads) and importantly it also colocalizes with the WTeNOS-AEQ. These results further suggest that the addition of AEQ does not disrupt  the subcellular targeting of WT or G2A eNOS. Finally, the functional effect of AEQ fusion on the eNOS constructs was determined. Although a slight reduction in NO production by both constructs was observed when fused to AEQ, as has been reported for eNOS-GFP fusion proteins (30), the relative differences between WTeNOS and G2AeNOS were unaffected by the addition of AEQ (Fig. 5c). Taken together, these results strongly suggest that the targeting of the two eNOS-AEQ constructs accurately reflects that of the native enzymes. When aequorin was expressed in unstimulated COS-7 cells either as a free cytosolic probe (CYT-AEQ) or fused to the transmembrane domain of CD8 as a plasma membrane probe (PM-AEQ), the localized [Ca 2ϩ ] experienced, as interpreted by the burnout of the coelenterazine pool, was much greater for the plasma membrane probe than the cytosolic probe (Fig.  6a). These findings are consistent with our hypothesis that proximity to plasma membrane Ca 2ϩ entry channels results in substantially higher localized [Ca 2ϩ ]. Likewise, when aequorin was fused to the C terminus of the eNOS constructs in unstimulated cells, the membrane-bound WTeNOS-AEQ construct experienced greater coelenterazine burnout than the cytosolic G2AeNOS-AEQ construct (Fig. 6b).
When COS-7 cells expressing the AEQ constructs were stimulated with ionomycin (1 M) the coelenterazine burnout between the CYT-AEQ and PM-AEQ was not significantly different, suggesting that this stimulation results in both the plasma membrane and cytosolic probes being exposed to equivalent amounts of Ca 2ϩ over time (Fig. 7a). Interestingly, the G2AeNOS-AEQ displayed significantly more coelenterazine burnout than the WTeNOS-AEQ probe (Fig. 7b), suggesting that over the 30 min of ionomycin stimulation the cytosolic G2AeNOS is actually exposed to more Ca 2ϩ than the membrane-bound WTeNOS.
Phosphorylation State Contributes to Differences in NO Production between the Two Constructs-Both WT and G2A eNOS constructs experience similar exposure to intracellular Ca 2ϩ upon ionomycin stimulation. However our earlier observations demonstrated that intracellular Ca 2ϩ was undoubtedly implicated in the phenomenon. If Ca 2ϩ was involved in the phenomenon, but the constructs did not experience different exposure to Ca 2ϩ , then the logical explanation was that the constructs were not equally responsive to Ca 2ϩ . A number of phosphorylation sites on eNOS are believed to increase enzyme sensitivity to intracellular Ca 2ϩ (31)(32)(33). Indeed, when creating our own Ca 2ϩ -independent eNOS mutants we deleted three such sites; specifically Ser 617 , Ser 635 , and Ser 1179 (Fig. 8a). We therefore examined the differences in the phosphorylation state of the two constructs. When the eNOS constructs were immunoprecipitated and probed with phosphospecific antibodies (Fig. 8b), we observed a slight decrease in Ser 635 phosphorylation, a marked decrease in Ser 1179 phosphorylation, and virtually no Ser 617 phosphorylation on the G2AeNOS compared with the WT enzyme, confirming earlier observations (10). As stimulation of eNOS has been reported to alter phosphorylation state (31, 34 -36), we also examined the phosphorylation state of immunoprecipitates after stimulation with ionomycin (1 M, 30 min). We observed an increase in the phosphorylation of Ser 617 , Ser 635 , and Ser 1179 in the stimulated cells  versus the unstimulated cells; however, there was little change in the relative differences between WTeNOS and G2AeNOS after ionomycin stimulation (Fig. 8b). Although it had not been removed in our Ca 2ϩinsensitive mutants, we also examined the phosphorylation state of Thr 497 , a residue whose phosphorylation state has been intimately linked with CaM binding by the enzyme (8,37). The Thr 497 site was phosphorylated in WTeNOS but not G2AeNOS, an observation that was once more unchanged after stimulation with ionomycin. Moreover, the phosphorylation of Thr 497 in the Ca 2ϩ -insenstitive deletion mutants was no different to that of the original constructs (data not shown), suggesting that the equalization of NO production in the deletion mutants was not a result of altered phosphorylation of the Thr 497 site.
To examine whether differences in phosphorylation contributed to the altered NO production between the constructs, we mutated three phosphorylation sites (Ser 617 , Ser 635 , and Ser 1179 ) to alanine to create an eNOS construct that was phospho-null at these sites. Phospho-null constructs expressing either the WT or G2A targeting sequences produced equivalent amounts of NO when expressed in COS-7 cells (Fig. 8c).
To further confirm that phosphorylation state was responsible for the different catalytic activity between the two constructs, we examined the effect of calyculin A, an inhibitor of the protein phosphatases PP1 and PP2A. After incubation with calyculin A, phosphorylation of Ser 617 , Ser 635 , and Ser 1179 was markedly increased (Fig. 9a), and NO production by both constructs was significantly increased (Fig. 9b). Importantly, treatment with calyculin A increased the stimulated NO production by G2AeNOS much more than that of WTeNOS (7-fold versus 2-fold), and in fact after calyculin A treatment stimulated NO production by G2AeNSO was equivalent with that of WTeNOS (Fig. 9b). Calyculin significantly increased basal NO production from both constructs; however, NO production from WT was elevated to a much greater degree than the G2A enzyme (4-fold versus 2-fold). This difference most likely results from the higher Ca 2ϩ experienced by the WT enzyme versus the G2AeNOS under basal conditions (Fig. 6b).
When taken together, the above results strongly suggested that differences in ionomycin-stimulated NO production by WT and G2A eNOS was in fact because of differences in Ca 2ϩ /CaM sensitivity as a result of differences in phosphorylation. We therefore sought to directly examine the Ca 2ϩ /CaM-sensitivity of each WT and G2A eNOS. When both constructs were isolated on ADP-Sepharose beads and exposed to saturating levels of all other cofactors (NADPH, BH 4 , L-arginine, and Ca 2ϩ ), WTeNOS displayed a higher sensitivity to calmodulin than G2AeNOS (Fig. 9c), directly confirming that the two constructs have different sensitivity to Ca 2ϩ /CaM.
Observations in COS Cells Are Mirrored in the More Physiologically Relevant Bovine Aortic Endothelial Cells (BAECs)-To confirm the physiological relevance of our findings in COS-7 cells, we extended some key experiments to primary cultures of BAECs. When either WTeNOS or G2AeNOS were overexpressed in BAECs, the NO production both basally and in response to stimulation was significantly less for cells overexpressing G2AeNOS compared with those overexpressing WTeNOS (Fig. 10b). Likewise when BAECs were transduced with WTeNOS-AEQ or G2AeNOS-AEQ, the WT construct reported higher Ca 2ϩ exposure (based on coelenterazine burnout) than the G2A construct in unstimulated cells (Fig. 10c). When cells expressing the AEQ constructs were stimulated with ionomycin, the constructs did not report significantly different calcium concentrations (unlike the findings in COS cells). However the most significant finding was that WTe-NOS is not exposed to more Ca 2ϩ than the G2AeNOS, results that are consistent with the findings in COS cells (Fig. 10d).
Finally, we sought to examine the phosphorylation state of the eNOS constructs in BAECs. After transducing either WTeNOS or G2AeNOS in BAECs, we then separated the membrane and cytosolic fractions of the cells. Using ADP-Sepharose beads to isolate WTeNOS from membrane fractions and G2AeNOS from cytosolic fractions, we immunoblotted for the relevant phosphorylation sites (Fig. 10e). The phosphorylation state of each site was similar in BAECs as had been the case in COS-7 cells, once more suggesting that our findings in COS-7 cells are physiologically relevant.

DISCUSSION
The reduced activity of myristylation-deficient and thus incorrectly targeted eNOS constructs (G2AeNOS) has been known for some time (14); however, the mechanism behind this phenomenon remains poorly understood. It has long been known that intracellular Ca 2ϩ is an absolute requirement for eNOS activation (38). Furthermore, it is now widely accepted that when sustained NO production is triggered by a number of Ca 2ϩ -elevating agents such as the IP 3 generating agonist histamine, the SERCA inhibitor thapsigargin, or the Ca 2ϩ ionophore A-23187, the major source of this Ca 2ϩ is the influx of extracellular Ca 2ϩ via CCE (22,39,40). The localization of WTeNOS in plasma membrane caveolae, sites where CCE channels have also been proposed to reside (41), raised the tempting hypothesis that the decreased NO production by G2AeNOS upon stimulation of CCE was simply a result of the construct experiencing lower localized [Ca 2ϩ ] than the WTeNOS because of greater spatial separation from Ca 2ϩ influx channels. Indeed, the study of Lin et al. (22) conclusively demonstrated that upon activation of CCE, a WTeNOS construct did experience higher peak [Ca 2ϩ ] transient than a myristylation-deficient construct. Our previous observations that expression of WTeNOS and G2AeNOS targeting sequences fused to the Ca 2ϩ -independent iNOS fail to demonstrate differences in NO production (16) also seemed to support this hypothesis. However, despite the peak differences in calcium inflow, no direct link between has been made between local NO production and the more functionally relevant measurement of compartmentalized increases in intracellular calcium over time. Therefore, the aim of the present study was to directly demonstrate if the disparity in catalytic activity between the two constructs was caused by differences in localized [Ca 2ϩ ] upon stimulation with Ca 2ϩ -elevating agonists, or if other regulatory factors were involved.
We first examined the importance of cellular integrity to the phenomenon. Three different Ca 2ϩ -elevating agonists (ionomycin, thapsigargin, and ATP) all produced greater activation of WTeNOS than G2AeNOS in intact COS-7 cells. However when the cells were disrupted either by mechanical (sonication) or chemical (Triton X-100) means, G2AeNOS was significantly more active than WTeNOS. These findings confirmed that the relative activity of eNOS depends on factors present within the intracellular environment that each construct resides in, rather than a permanent change in enzyme activity.
Our hypothesis relied on the influx of extracellular Ca 2ϩ as the major source of eNOS activation. As eNOS has been suggested to undergo Ca 2ϩ -independent activation in response to certain stimuli (6, 33), we needed to ensure that all NO production by our constructs was dependent on Ca 2ϩ influx. Removal of extracellular Ca 2ϩ almost abolished NO production by both constructs, confirming that both relied on the influx of external Ca 2ϩ . Interestingly, intracellular chelation of Ca 2ϩ using BAPTA-AM failed to completely inhibit the activity of either construct, even though all such activity was dependent on Ca 2ϩ influx. The inability of BAPTA-AM to affect near membrane Ca 2ϩ responses has been previously reported (42), and possibly reflects the inability of BAPTA to immediately access and chelate all of the Ca 2ϩ entering the cytosol through Ca 2ϩ influx channels. Thus, the membrane-localized eNOS is likely to be relatively unaffected by BAPTA-AM even though cytosolic Ca 2ϩ is significantly reduced. Further supporting this hypothesis is the observation that G2AeNOS, the construct furthest from the plasma membrane, was much more effectively inhibited by BAPTA-AM than WTeNOS. Another central tenet of our hypothesis was that if the two constructs were producing disparate amounts of NO because of their exposure to different levels of Ca 2ϩ , then raising intracellular Ca 2ϩ to saturating levels should equalize the NO production by both constructs. Cells with 1 M ionomycin failed to display equalized NO production even in the presence of 10 mM extracellular Ca 2ϩ (probably because of the fact that at this concentration ionomycin is thought to act almost exclusively by activating CCE (29), a Ca 2ϩ entry pathway that can only produce Ca 2ϩ influx as rapidly as plasma membrane Ca 2ϩ channels will allow). However when the cells were stimulated with a 10-fold higher concentration of ionomycin (a concentration that facilitates Ca 2ϩ entry independently of plasma membrane Ca 2ϩ channels), NO production was equalized when extracellular Ca 2ϩ was raised to sufficiently high concentration.  ϭ 4 -9). * , p Ͻ 0.05 versus construct with wild-type targeting sequence, #, p Ͻ 0.05 versus same construct without calyculin pretreatment.
As our hypothesis was still supported by all observations so far, we then attempted to more directly determine the contribution of intracellular Ca 2ϩ to the disparity in NO production between the constructs. To achieve this, we created Ca 2ϩ /CaM-independent mutants based on those described by Chen and Wu (26). Interestingly, the WTeNOS and G2AeNOS Ca 2ϩ /CaM-independent mutants showed absolutely no differences in NO production under any stimulation conditions observed, confirming our previous findings with targeted iNOS constructs (16). This constitutes the first direct evidence that Ca 2ϩ or CaM alone is responsible for the difference in activity.
After confirming that interaction of the constructs with Ca 2ϩ was involved in the phenomenon, we next determined the localized [Ca 2ϩ ] experienced by each construct both basally and upon stimulation by fusing in-frame the bioluminescent Ca 2ϩ probe aequorin (AEQ). Although Lin et al. (22) had demonstrated that the peak [Ca 2ϩ ] experienced after CCE stimulation was higher for WTeNOS than the G2A construct, this response was a transient (Ͻ30s) peak response to a protocol designed to maximally activate CCE. As we stimulated our cells with ionomycin for 30 min to measure NO production, we felt it would be more informative to measure the Ca 2ϩ exposure during the 30-min stimulation with ionomycin rather than an immediate peak response to deliberate CCE activation. We therefore attempted to devise a method to observe Ca 2ϩ exposure over more sustained time periods.
The reaction between AEQ and its cofactor coelenterazine in the presence of Ca 2ϩ and molecular oxygen results in the emission of one photon and the irreversible oxidation of coelenterazine to coelenteramide. The emission of light in response to Ca 2ϩ is thus a once-only reaction (43), and the size of the available pool of luminescence decreases constantly over time in direct proportion to the amount of Ca 2ϩ present. The relative amounts of localized [Ca 2ϩ ] that each construct is exposed to over time can therefore be estimated by measuring the decrease in the size of the releasable pool of luminescence (L max ), as determined by discharging the remaining luminescence with lysis buffer. Using this technique, we observed that in unstimulated cells the PM-AEQ probe displayed a faster rate of burnout than CYT-AEQ. Likewise, WTeNOS-AEQ clearly displayed a faster rate of burnout than G2AeNOS-AEQ, indicating that the WTeNOS construct is exposed to higher levels of Ca 2ϩ than the G2AeNOS construct in unstimulated cells. This probably reflects the constant transmembrane flux of Ca 2ϩ , which enters the cell because of the spontaneous opening of Ca 2ϩ influx channels. It is rapidly pumped back out of the cytosol by plasma membrane Ca 2ϩ -ATPases before having an opportunity to significantly elevate [Ca 2ϩ ] in the cytosol (23).
Upon 30 min of stimulation with ionomycin, the burnout between PM-AEQ and CYT-AEQ was not significantly different, suggesting that ionomycin stimulation elevates cytosolic [Ca 2ϩ ] to the point where both plasma membrane and cytosolic probes are equally exposed to Ca 2ϩ . When cells expressing the eNOS-AEQ fusion constructs were stimulated with ionomycin for 30 min, not only did the WTeNOS no longer display greater burnout than G2AeNOS, but in fact the membranebound WT probe displayed significantly less burnout than cytosolic G2A probe, suggesting that the G2A probe had been exposed to more Ca 2ϩ over the stimulation period than the WT probe. This suggests that, even though the [Ca 2ϩ ] increase because of maximally activated CCE is transiently greater at the plasma membrane (as reported by Lin et al.,Ref. 22), upon sustained (30 min) elevation of cytosolic Ca 2ϩ (such as during ionomycin stimulation) the Ca 2ϩ gradient between plasma membrane influx channels and the cytosol may be markedly reduced (or even non-existent) when observed as a Ca 2ϩ exposure over time phenomenon (such as in these coelenterazine burnout experiments). This also supports the recent findings of Jobin et al. (44) using a FRET-based Ca 2ϩ sensor fused to eNOS constructs.
Based on the AEQ burnout data, our hypothesis that the disparity in NO production was simply because of proximity to plasma membrane Ca 2ϩ influx channels seemed insufficient to explain the phenomenon. The preceding data, however, strongly suggested that Ca 2ϩ plays an important role in the phenomenon, and we therefore considered the possibility that the two constructs were differentially sensitive to Ca 2ϩ .
As our Ca 2ϩ -independent deletion mutants (the only modification of the constructs, which had successfully equalized NO production) also deleted three key phosphorylation sites involved in sensitizing eNOS to calcium/calmodulin we next addressed whether this could explain the functional differences between the enzymes. Compared with the WT enzyme, G2A eNOS exhibited slightly reduced Ser 635 phosphorylation, marked reduction of Ser 1179 phosphorylation, and virtually no Ser 617 phosphorylation. The phosphorylation of these residues is associated with reduced dependence on intracellular calcium (33), increased calcium sensitivity (31) and reduced dissociation of calcium-bound calmodulin (32). To directly test whether phosphorylation of these residues mediates the functional differences between these enzymes, we mutated all three residues to the nonphosphorylatable analogue, alanine. The marked difference in activity between the native WT and G2A enzymes was abolished in the equivalent constructs with triple (S617A, S635A, S1179A) mutations, suggesting that indeed a difference in phosphorylation mediates the functional differences in calcium sensitivity between these enzymes.
We then attempted the converse experiment, by equalizing the phosphorylation of WT and G2A eNOS using the phosphatase inhibitor, calyculin A. Treatment of cells with calyculin produced a marked increase in phosphorylation and an increase in both basal and stimulated NO production, for both constructs. Calyculin treatment equalized eNOS phosphorylation and stimulated production of NO between the WT and G2A eNOS. These results are consistent with the data obtained with phospho-null mutants and strongly support the contention that phosphorylation mediates the difference between these enzymes. Interestingly, phosphatase inhibition failed to equalize basal NO production between the constructs even though stimulated NO production was equalized. We believe that this reflects the differences in basal Ca 2ϩ exposure between the two constructs that were observed using eNOS-aequorin fusion proteins.
As the two constructs differed in phosphorylation state, and either increasing or abolishing phosphorylation of the two constructs equalized NO production, it seemed likely that our previous results could be explained by these phosphorylation differences endowing the two constructs with different Ca 2ϩ /CaM sensitivity. We therefore sought direct evidence that the two constructs were differentially sensitive to Ca 2ϩ / CaM. Performing a concentration-response curve to CaM on isolated eNOS constructs in the presence of saturating Ca 2ϩ conclusively demonstrated that the WT construct is more sensitive to Ca 2ϩ /CaM than the G2A construct, supporting our conclusions.
Finally, we sought to confirm that our observations in COS-7 cells were mirrored in BAECs, a cell type which expresses endogenous eNOS and is thus a more physiologically relevant model. When BAECs overexpressing WTeNOS or G2AeNOS were stimulated, G2AeNOS-expressing cells produced significantly less NO than those expressing WTeNOS. When the WTeNOS-AEQ and G2AeNOS AEQ fusion constructs were expressed in BAECs, the basal burnout of WTeNOS-AEQ was once again greater than that of G2AeNOS-AEQ, confirming our results in COS-7 cells. Upon stimulation of BAECs expressing the AEQ fusion constructs, the WTeNOS and G2AeNOS constructs reported no significant difference in coelenterazine burnout. This differs slightly from the results obtained in COS-7 cells (where G2AeNOS displayed significantly higher burnout than WTeNOS) and may reflect either dimerization issues (G2AeNOS-AEQ being taken to the membrane after dimerizing with endogenous eNOS) or possibly some subtle differences in Ca 2ϩ signaling between endothelial cells and COS-7 cells. Nonetheless, the pertinent finding; that stimulation with ionomycin does not produce greater burnout of the membrane-bound probe than the cytosolic probe, was confirmed in BAECs as well as COS-7 cells. In addition, immunoblotting demonstrated that the differences we observed in Ser 617 , Ser 635 , and Ser 1179 phosphorylation of WTeNOS and G2AeNOS in COS-7 cells are preserved in BAECS, further supporting the physiological relevance of our findings in COS-7 cells.
In conclusion, we have presented direct evidence that the functional differences between the WT and myristylation-deficient, cytosolic eNOS constructs are due exclusively to one variable, calcium/calmodulin. We have measured local calcium levels using targeted aequorin fusion proteins and have shown that while basal calcium levels are clearly different between the two enzymes, stimulation of cells provides equivalent levels of calcium to both enzymes. Phosphorylation of eNOS at serines 617, 635, and 1179 provides a mechanism for enhancing the interaction between eNOS and calcium/calmodulin and deletion or mutation of these phosphorylation sites, as well as stimulation of phosphorylation by phosphatase inhibitors, is sufficient to normalize activity between the WT and G2A enzymes. These data are sufficient to explain the paradox between NO release and calcium that underlies the functional differences in membrane-anchored and cytosolic eNOS.