INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE TWO CONSTITUTIVELY EXPRESSED ISOFORMS of nitric oxide (NO) synthase (NOS), neuronal (nNOS) and endothelial NOS (eNOS), share large conserved regions of homology and catalyze the production of NO from L-arginine using the same mechanism. However, each isoform also exhibits unique properties that enable it to carry out its specific in vivo function. Both isoforms are activated in the cell by transient increases in intracellular Ca²⁺, which trigger the binding of Ca²⁺/calmodulin and subsequent NO production. An important feature of eNOS is its subcellular localization to specialized regions of the plasma membrane termed caveolae (34). This localization is mediated, according to cell type, by various isoforms of caveolin, which dissociate from eNOS upon activation of the enzyme. This interaction is crucial for the regulation of eNOS activity, because receptor-mediated activation of the enzyme is abolished in a soluble mutant of eNOS or by overexpression of caveolin (8). nNOS has also been shown to associate with caveolin-3 in skeletal muscle (36), and both isoforms are inhibited by this interaction (12). Notably, although both NOS isoforms interact with caveolin using highly homologous domains (36), the primary mechanisms of membrane localization differ. Posttranslational acylation is necessary for the membrane association of eNOS (32), whereas a PSD-95/Discs-large/ZO-1 domain (PDZ domain; for details on terminology, see Ref. 27) mediates the subcellular localization of nNOS in muscle cells and neuronal tissue (3, 4). Another inhibitory interaction that is common to both eNOS and nNOS is the binding to the bradykinin B₂ receptor (13, 15).

Because both the regulation and functions of the constitutive NOS isoforms are largely determined by their subcellular localization, which in turn is cell type specific, an approach was taken that enabled the characteristics of each isoform to be studied in a common cell line in the absence of these cell type-specific effects. To this aim, stably transfected cell lines were generated containing equivalent amounts of each isoform in the human embryonic kidney cell line HEK 293, which lacks endogenous eNOS (33) but may express low levels of another isoform, as shown indirectly by the small increase in cGMP seen upon stimulation with Ca²⁺-mobilizing agents (1). In addition, a soluble mutant of eNOS, which lacks the myristoylation site at Gly², was also studied to enable the importance of membrane association on the functional activity of eNOS to be determined. The results reveal notable differences in the basal and Ca²⁺-stimulated activities of the three proteins despite the common cellular environment.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. HEK 293 cells were obtained from the American Type Culture Collection (Rockville, MD). 2,2-Diethyl-1-nitroso-oxyhydrazine sodium salt (DEA/NO) was purchased from Alexis (Lausen, Switzerland). [-³²P]GTP (>3,000 Ci/mmol) was from Humos Diagnostika (Vienna, Austria). Polyclonal antisera were raised in rabbits against purified bovine eNOS (20) and purified porcine brain NOS (17). All other chemicals were purchased from Sigma (Vienna, Austria). All solutions were prepared in Nanopure water (Barnstead ultrafiltered type I, resistance > 18 m/cm).

Plasmid constructs. The pcDNA3 vector (Invitrogen; Groningen, The Netherlands), which contains the neomycin resistance cassette, was used, containing the cDNAs for human nNOS, human eNOS, and a soluble mutant of human eNOS in which the myristoylation site has been removed by a Gly2Ala mutation. Plasmid DNA was prepared for transfection by linearization with the restriction enzyme PvuI, followed by phenol-chloroform extraction and ethanol precipitation.

Transfection and selection of stable lines. HEK 293 cells were transfected at 40% confluency with 5 µg of linearized plasmid DNA using Superfect (Qiagen; Hilden, Germany) or Tfx-50 (Promega; Mannheim, Germany). The cells were trypsinized, counted 48 h after transfection, and replated at 0.5-1 × 10⁶ cells/plate in 100-mm plates in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, 0.1 mg/ml streptomycin, 1.25 µg/ml amphotericin B, and 1 mg/ml geneticin (antibiotic G418). After 10-12 days, isolated colonies of resistant cells were selected and expanded. In some cases, a limiting dilution was performed to obtain clones arising from single cells. Individual clones were tested for activity (see Determination of NOS activity and [³H]arginine uptake in intact cells), and the following cell lines were maintained in 250 µg/ml G418: cells transfected with wild-type nNOS cDNA (nNOS line); cells transfected with wild-type eNOS cDNA (eNOSwt line); cells transfected with cDNA encoding a soluble mutant of eNOS (eNOSsol line); and cells transfected with vector DNA (Vector line).

Determination of NOS activity and [³H]arginine uptake in intact cells. Confluent monolayers of transfected HEK 293 cells in six-well plates were preincubated in nominal Ca²⁺-free incubation buffer [composed of (in mM) 50 Tris · HCl, 100 NaCl, 5 KCl, 1 MgCl₂, and 0.1 EGTA; pH 7.4] for 15 min in the presence or absence of 100 nM thapsigargin or 300 nM ionomycin. [³H]arginine (~1,000,000 dpm) and CaCl₂ (final concentration 0.1-3 mM) were then added, and, after a further 2-min incubation at 37°C, the cells were washed twice with cold wash buffer [composed of (in mM) 50 Tris · HCl, 100 NaCl, 5 KCl, 1 MgCl₂, and 5 EDTA; pH 7.4] before being lysed in 10 mM HCl. An aliquot was retained for determination of [³H]arginine uptake, whereas the remainder was adjusted to pH 5.0 using 200 mM sodium acetate (pH 13) and applied to a column (DOWEX 50W X80-40, Sigma) to separate [³H]citrulline from [³H]arginine. Radioactivity was determined using a scintillation counter, and the results are expressed as the percent conversion of incorporated [³H]arginine into [³H]citrulline, as described previously (28).

Preparation of cell homogenates. Cells were homogenized in 50 mM triethanolamine-HCl buffer (pH 7.4) containing 12.8 mM -mercaptoethanol by three cycles of freeze/thawing. The cytosolic protein fraction was obtained by centrifugation at 10,000 g for 5 min. The pellet was washed twice and then resuspended in homogenization buffer supplemented with 10 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) and 0.5 mM EDTA. After a further three cycles of freeze/thawing, the homogenate was centrifuged to remove insoluble material from the solubilized protein fraction. For preparation of total protein, 10 mM CHAPS and 0.5 mM EDTA were included in the homogenization buffer. Protein concentration was determined using the Bradford assay.

Determination of NOS activity in cell homogenates. Protein (50-200 µg) was incubated at 37°C for 10 min with 10 µM [³H]arginine (~50,000 cpm), 10 µg/ml calmodulin, 0.5 mM CaCl₂, 10 µM (6R)-5,6,7,8-tetrahydro-L-biopterin, 5 µM FAD, 5 µM flavin-adenine mononucleotide, 200 µM NADPH, and 10 mM CHAPS, as described previously (23). [³H]citrulline was isolated using column chromatography and quantitated in a scintillation counter.

Western blot analysis of NOS proteins. Cytosolic or membrane protein (10 µg) was separated using SDS-PAGE (8% acrylamide), transferred to nitrocellulose, and probed with antibodies specific for either nNOS or eNOS using a previously described method (35).

Measurement of cGMP in intact cells. Cells were preincubated for 15 min in nominal Ca²⁺-free incubation buffer containing 1 mM isobutyl-methylxanthine and 10 µM indomethacin and then incubated for 4 min with 100 nM thapsigargin, 300 nM ionomycin, or 1 µM DEA/NO in the presence of 3 mM CaCl₂. Cells were then lysed in 1 ml of 0.01 N HCl for determination of cGMP content by radioimmunoassay (29). When used, the NOS inhibitor N^G-nitro-L-arginine (L-NNA; 300 µM) was added to the culture medium 30 min before preincubation and was present throughout the experiment.

Determination of soluble guanylate cyclase mRNA expression. Expression of soluble guanylate cyclase (sGC) mRNA was measured using competitive RT-PCR (cRT-PCR). A specific cDNA fragment of 698 bp (positions 66-764) of the -subunit was amplified using the following primers: sGC sense primer, 5'-GAA GAC ATC AAA AAA GAG GC-3'; and sGC antisense primer, 5'-GAG AAG ACA GAC AGA AGG C-3', respectively. An internal deleted sGC cDNA standard of 462 bp was constructed by linker primer PCR using the following linker primer: 5'-GAA GAC ATC AAA AAA GAG GCA CCT CGC CAC CAT CTA C-3'. The linker fragment was subsequently cloned into the pCRII TOPO cloning vector (Invitrogen), and its identity was confirmed by automatic DNA sequencing. The internal deleted sGC cDNA standard was in vitro transcribed into cRNA (RNA Transcription Kit, Stratagene). The cRNA was quantified spectrophotometrically. Increasing amounts of standard RNA (1-500 pg) were mixed with native RNA (500 ng), reverse transcribed into cDNA (Superscript TM preamplification system, GIBCO-BRL), and amplified by PCR. Reaction products were separated by agarose gel electrophoresis. The optical density of each PCR fragment was estimated (Bio-Rad Geldoc 1000, Bio-Rad), and the logarithm of the quotient of normalized standard and sample-specific PCR fragment density was used to calculate the equivalence point (for details, see Ref. 18).

Determination of sGC activity in cell homogenates. Cell homogenates (400 µg total protein) were incubated at 37°C for 20 min in a total volume of 0.2 ml of 50 mM triethanolamine-HCl buffer (pH 7.4) containing 0.1 mM [-³²P]GTP (200,000-300,000 counts/min), 5 mM MgCl₂, 1 mM cGMP, 1 mM isobutyl-methylxanthine, 5 mM phosphocreatine, 152 U/ml creatine phosphate kinase, 2 mM dithiothreitol, and 1 mM EGTA. Reactions were started by adding 10-fold concentrated stock solutions of DEA/NO (final concentration 10 µM) to the assay mixture with subsequent transfer of the samples from 4 to 37°C. Incubations were stopped by ZnCO₃ precipitation, and [³²P]cGMP was isolated by column chromatography as described previously (31). Results were corrected for enzyme-deficient blanks and recovery of cGMP.

Measurement of NO. NO formation was determined using the oxyhemoglobin assay as described previously (30). Briefly, HEK 293 cells cultured in six-well plates were incubated for 10 min at 37°C in incubation buffer containing 3 mM CaCl₂ and 5 µM oxyhemoglobin in the absence and presence of 300 nM ionomycin. The increase in the difference between the absorbances at 402 and 420 nm was monitored, and NO concentrations were calculated using a molar absorption coefficient of 56 mM¹ · cm¹.

Determination of intracellular Ca²+ levels. HEK 293 cells from one confluent petri dish were harvested, suspended in 5 ml DMEM, and incubated with 2 µM fura-AM at 37°C. After 45 min, cells were washed, resuspended in nominal Ca²⁺-free incubation buffer, and transferred into a thermostated cuvette. After equilibration (~5 min), cells were preincubated for 15 min in the absence or presence of 300 nM ionomycin or 100 nM thapsigargin, and CaCl₂ (final concentration 3 mM) was then added. Fluorescence was measured at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm, and the intracellular free Ca²⁺ concentration was calculated using the ratio of fluorescence intensity at 340 to 380 nm.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Stably transfected cell lines were obtained expressing very similar high levels of nNOS, eNOS, or a soluble mutant of eNOS. The specific NOS activities in cell homogenates were <0.1 (Vector), 672 ± 49 (nNOS), 602 ± 18 (eNOSwt), and 640 ± 88 pmol · mg protein¹ · min¹ (means ± SE) (eNOSsol). The majority of eNOS activity was found in the particulate fraction, whereas the mutant eNOS lacking the myristoylation site as well as nNOS were predominantly found in the cytosol (Fig. 1). This distribution was also reflected in the Western blot shown in Fig. 2. Densitometric analysis indicated that 34.8 ± 2.6%, 65.9 ± 4.2%, and 77.0 ± 3.4% of eNOSwt, eNOSsol and nNOS proteins, respectively, were localized in the soluble fractions of the cells.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Distribution of nitric oxide synthase (NOS) activity in homogenates of stably transfected HEK 293 cells. Protein (50-200 µg) was incubated at 37°C for 10 min with 10 µM [3H]arginine (~50,000 cpm), 10 µg/ml calmodulin, 10 µM (6R)-5,6,7,8-tetrahydro-L-biopterin, 5 µM FAD, 5 µM flavin-adenine mononucleotide, 200 µM NADPH, and 10 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS). [3H]citrulline was isolated using column chromatography and quantitated in a scintillation counter. Results are expressed as a percentage of total NOS activity. Data are means ± SE of 3 experiments. eNOSwt, cells transfected with wild-type endothelial NOS (eNOS); eNOSsol, cells transfected with a soluble mutant of eNOS; nNOS, cells transfected with wild-type neuronal NOS.

View larger version (6K): [in this window] [in a new window]   Fig. 2.   Western analysis of NOS proteins in stably transfected HEK 293 cells. Cytosolic (c) or membrane (m) protein (10 µg) was separated using SDS-PAGE (8% acrylamide), transferred to nitrocellulose, and probed with antibodies specific for either nNOS or eNOS. Std, 50 ng of purified eNOS or nNOS protein; HEK, untransfected control cells. The blot is representative of 3 similar blots.

Having established that the homogenates of each stable cell line exhibited similar NOS activity, experiments were performed to establish whether differences in activity could be found under identical conditions in intact cells. To this end, cells were incubated with [³H]arginine in the presence of increasing concentrations of extracellular Ca²⁺ in the presence and absence of thapsigargin and ionomycin. The former stimulates Ca²⁺ entry by depleting intracellular Ca²⁺ stores, whereas the latter is an ionophore that transports Ca²⁺ across the plasma membrane. In the absence of Ca²⁺-mobilizing agents (Fig. 3A), both the wild-type eNOS- and the mutant eNOS-transfected cells showed a very low level of conversion of [³H]arginine to [³H]citrulline, which increased only marginally in the presence of 3 mM extracellular Ca²⁺. In contrast, the nNOS-transfected cells exhibited a higher basal NOS activity, which increased significantly in the presence of 3 mM extracellular Ca²⁺. In the presence of ionomycin or thapsigargin (Fig. 3, B and C), [³H]arginine conversion increased with increasing extracellar Ca²⁺ in all three lines, although the nNOS cell line, again, exhibited the highest absolute levels of NOS activity. In all cases, the activity of the soluble mutant of eNOS did not differ significantly from that of the wild-type isoform, although both these lines had a greater degree of increase in activity after stimulation compared with the nNOS cells as a result of their much lower background activity. As shown in Table 1, the two eNOS cell lines showed virtually identical intracellular concentrations of free Ca²⁺ upon stimulation by Ca²⁺-mobilizing agents. Slightly higher Ca²⁺ concentrations were measured in nNOS-transfected cells, but it is rather unlikely that this small difference explains the higher NOS activity in the nNOS cell line. Table 2 shows that a similar pattern of NOS activation was achieved with receptor agonists (carbachol, bradykinin, and ATP) elevating the concentrations free intracellular Ca²⁺ in the presence of two different concentrations of extracellular Ca²⁺. Finally, Table 3 demonstrates that the cell transfection did not affect [³H]arginine uptake.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   NOS activity in intact stably transfected HEK 293 cells in the absence (A) or presence of 300 nM ionomycin (B) or 100 nM thapsigargin (C). Cells were preincubated in nominal Ca2+-free incubation buffer for 15 min in the presence or absence of 100 nM thapsigargin or 300 nM ionomycin. [3H]arginine and CaCl2 were then added, and, after a further 2-min incubation at 37°C, the cells were washed twice before being lysed in 10 mM HCl. [3H]citrulline was isolated using column chromatography and quantitated in a scintillation counter. Results are expressed as the percent conversion of incorporated [3H]arginine into [3H]citrulline. Data are means ± SE of 5 experiments.

In addition to the conversion of [³H]arginine to [³H]citrulline as a measure of NOS activity, the oxyhemoglobin assay was used to assess NO production by the three cell lines. As seen in the [³H]arginine conversion assay, the nNOS-transfected cells had significantly higher activity than the wild-type eNOS cells both in resting and ionomycin-stimulated cells (Table 4). However, in contrast with the previous assay, the soluble eNOS mutant-transfected cells produced 1.8-fold more NO under basal conditions than cells transfected with wild-type eNOS. As a result of the lower basal NO production, wild-type eNOS showed the greatest percent increase in NO production upon Ca²⁺ stimulation.

HEK 293 cells express sGC (1), which can be stimulated by NO to produce cGMP. In an effort to identify any functional differences in NO produced by the three distinct NOS isoforms, cGMP levels were determined under basal and stimulated conditions. Basal cGMP levels increased in all the NOS-transfected cell lines upon increasing the extracellular Ca²⁺ concentration (Fig. 4A), supporting the oxyhemoglobin assay results, which showed that low levels of NO are continuously produced in NOS-transfected cells in the presence of extracellular Ca²⁺. This was confirmed using the NOS inhibitor L-NNA, which reduced the level of basally produced cGMP to below 2 pmol/mg (data not shown). Basal cGMP levels of cells in the nNOS line were higher than in the eNOSsol cell line, which in turn contained significantly more cGMP than cells in the eNOSwt line. This reflects the relative amounts of NO measured using the oxyhemoglobin assay. In cells treated with Ca²⁺-mobilizing agents, the absolute levels of cGMP were significantly lower in eNOSwt cells than in either eNOSsol or nNOS cells, which, at the highest concentration of extracellular Ca²⁺, had identical cGMP levels (Fig. 4, B and C). However, the eNOSwt cells showed the greatest degree of increase in activity upon stimulation.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   cGMP levels in untreated cells (A) or after stimulation with 300 nM ionomycin (B) or 100 nM thapsigargin (C). Cells were preincubated for 15 min in nominal Ca2+-free incubation buffer containing 1 mM isobutyl-methylxanthine and 10 µM indomethacin and then incubated for 4 min with 100 nM thapsigargin or 300 nM ionomycin in the presence of 3 mM CaCl2. Cells were then lysed in 1 ml of 0.01 N HCl for determination of cGMP content by radioimmunoassay. Data are means ± SE of 3 experiments.

The capacity of the transfected cells to produce cGMP in response to NO was determined by incubation with a maximal concentration of the NO donor DEA/NO. The maximum cGMP levels obtained in pmol/mg were 70 ± 17 (Vector), 46 ± 3 (eNOSwt), 96 ± 9 (eNOSsol), and 99 ± 7 (nNOS). Thus eNOSwt cells produced significantly less cGMP than the vector controls and the other NOS-transfected cell lines. To directly compare cGMP production between the three lines, values were converted to a percentage of the maximum obtained with DEA/NO. This revealed that, although the eNOSwt cell line produced the least cGMP under basal conditions (Fig. 5A), all three cells types attained a similar maximum when treated with Ca²⁺-mobilizing agents in the presence of 3 mM extracellular Ca²⁺ (Fig. 5, B and C).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   cGMP levels in intact cells expressed as a percentage of the maximum cGMP level obtained after stimulation with 1 µM 2,2-diethyl-1-nitroso-oxyhydrazine (DEA/NO). A: untreated; B: 300 nM ionomycin treated; C: 100 nM thapsigargin treated. Data are means ± SE of 3 experiments.

To further investigate the differences in capacity for cGMP production between the various lines, the specific activity of sGC was determined in cell homogenates. As shown in Fig. 6, the eNOS-transfected cell lines exhibited reduced sGC activity compared with vector controls or nNOS-transfected cells. The results obtained with eNOSwt and nNOS cell homogenates agree well with the cGMP levels in intact cells, but the eNOSsol homogenates apparently exhibited less sGC activity despite virtually identical capacity of cGMP formation in the intact cells.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Specific soluble guanylate cyclase (sGC) activity in homogenates of stably transfected HEK 293 cells. Protein (400 µg) was incubated at 37°C for 20 min in an assay mixture described in MATERIALS AND METHODS. Reactions were started by adding DEA/NO (final concentration 10 µM) and terminated by ZnCO3 precipitation. [32P]cGMP was isolated by column chromatography. Data are means ± SE of 3 experiments.

As shown in Fig. 7, sGC mRNA expression was significantly reduced in the eNOSwt cells compared with the other cell lines, including vector controls. These results agree with the capacity of the cells to produce cGMP in response to exogenously added NO. However, the data do not explain the apparently reduced specific sGC activity of eNOSsol homogenates.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Expression of sGC mRNA in vector- and NOS-transfected HEK 293 cells. 500 ng of RNA isolated from various cell lines was mixed with increasing amounts of a sGC standard RNA (1-500 pg), reversed transcribed into cDNA and amplified by PCR. Reaction products were separated by agarose gel electrophoresis, and the expression of sGC mRNA calculated using the equivalence point of standard and sample PCR fragment density. Data are the means ± SE of three experiments. Vector, cells transfected with vector only.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Three stably transfected cell lines, expressing nNOS, eNOS, and a soluble mutant of eNOS, were established. Although all three cell lines contained equivalent amounts of NOS activity, the relative distribution of membrane-associated and cytosolic protein varied. As expected, nNOS was largely found in the cytosol, whereas the majority of wild-type eNOS was membrane associated. The soluble mutant of eNOS was found almost exclusively in the cytoplasm. Thus it appears that myristoylation may be essential for mediating the membrane association of eNOS regardless of the cell type (32), although it cannot be excluded that structural changes of the protein rather than myrostoylation deficiency interferred with membrane association.

The availability of these three NOS isoforms in an identical cell background has enabled for the first time a direct comparison of their function in intact cells. With the use of a variety of methods to assess NOS activity, it could be shown that nNOS is more active in resting cells than either wild-type or soluble eNOS, abeit all three cell lines exhibited virtually identical NOS activity in terms of [³H]citrulline production in cell homogenates. This conclusion is based on the nearly twofold higher conversion of [³H]arginine to [³H]citrulline, the detection of more than twice as much NO using the oxyhemoglobin assay, and significantly higher basal levels of cGMP in nNOS-transfected cells. It should be emphasized that the absence in HEK 293 cells of the membrane-associated proteins to which nNOS normally binds (3, 4) may limit the biologically relevant interpretation of data on nNOS regulation obtained with the transfected HEK cells.

The very low basal activity of wild-type eNOS in an identical intracellular environment is evidence for the almost-complete inhibition of this isoform at resting Ca²⁺ concentrations. In vivo, NO is tonically produced by the endothelium as a result of exposure of endothelial cells to shear stress, which results in phosphorylation of a serine residue, leading to Ca²⁺-independent activation of the enzyme (6, 11). In the absence of this mechanism, as in our model, the enzyme remains virtually inactive. The soluble mutant of eNOS exhibited a slightly higher basal activity in terms of both NO and cGMP production, although not L-citrulline production, suggesting that NO is more rapidly inactivated when produced close to membranes, perhaps through the fast reaction with superoxide produced either by endothelial NADPH oxidase (14) or uncoupled eNOS (19).

It has been reported that wild-type eNOS but not the soluble mutant does interact with caveolin-1 in transfected COS-7 cells (8, 24). However, another study (9) from the same group suggests that the mutant does interact with caveolin, albeit with apparently lower affinity compared with the wild-type enzyme. Our results showing that the soluble isoform does not associate with HEK cell membranes in the presence of 10 mM CHAPS seems to agree with these previous findings. However, preliminary results from our laboratory indicate that HEK 293 cells do not express detectable amounts of caveolin-1 and do not contain membrane invaginations visible in electron microscopy (M. Balzer-Geldsetzer, W. Baumgartner, and K. Groschner, unpublished results). Similarly, Rybin et al. (25) recently reported that HEK 293 cells contain neither caveolin-1 nor caveolin-3. Thus the interaction with caveolin does not appear to be required for membrane association of wild-type eNOS in HEK 293 cells.

Upon stimulation with Ca²⁺-mobilizing agents, nNOS showed the greatest activity in terms of [³H]arginine conversion and NO production as well as cGMP production at submaximal extracellular Ca²⁺ concentrations. Hence, even in the presence of high intracellular Ca²⁺, eNOSwt is not as active as nNOS in transfected HEK 293 cells. Although NO has been implicated in the regulation of Ca²⁺ entry (7), these differences in activity cannot be attributed to isoform-dependent alterations in the intracellular Ca²⁺ levels, which were found to be identical in the eNOS cell lines and only slightly increased in nNOS-transfected cells.

Despite the slight differences in basal activities, eNOSwt and eNOSsol cells both produced similar amounts of NO or L-[³H]citrulline after stimulation by ionomycin or thapsigargin as well as several receptor agonists, showing that membrane association is not necessary for efficient activation of eNOS by Ca²⁺. Previously published comparisons of wild-type and soluble eNOS isoforms have described a diminution in activity of the mutant in stably transfected HEK 293 cells (21, 33), transiently transfected COS-7 cells (26), and adenovirus-transfected neurons (16) despite unchanged catalytic properties of the enzyme (32). Feron et al. (8) found no difference in activity between wild-type and soluble eNOS isoforms in Ca²⁺ ionophore-stimulated cardiac myocytes, but the soluble mutant was unable to respond to receptor-mediated regulation in these cells. Thus it appears that eNOS activity is tightly regulated by the local cellular environment and possibly other factors not yet clarified.

Another novel observation to emerge from comparison of the NOS cell lines is the isoform-dependent alteration in sGC expression. Transfection with wild-type eNOS resulted in a pronounced downregulation of sGC mRNA expression, whereas sGC mRNA levels of eNOSsol and nNOS cells were not significantly different from vector-transfected controls. It has been reported that sGC expression in pulmonary artery smooth muscle cells is downregulated by NO via a decrease in mRNA stability (10). In addition, sGC may become desensitized upon continuous exposure to NO. Such a desensitization of sGC has been demonstrated using isolated cardiomyocytes exposed to S-nitroso-N-acetyl-penicillamine for up to 24 h (5) and by comparing the NO sensitivity of aortas from wild-type and eNOS knockout mice (2). However, because eNOSwt cells exhibited the lowest basal NO production, NO was probably not directly involved in the observed downregulation of sGC expression unless the site of NO production was crucial for its effect. Alternatively, NO produced under basal conditions does indeed reduce sGC mRNA expression, and this effect is counterbalanced in eNOSsol and nNOS cells by other products of NOS catalysis, e.g., superoxide and/or peroxynitrite, which are formed in the course of the partially uncoupled reaction that takes place in conditions of reduced L-arginine or tetrahydrobiopterin availability (22). In any case, after correcting for the differences in sGC content by converting the absolute results to a percentage of the maximum obtained with DEA/NO, the cGMP levels in unstimulated cells correlated well with NO production. Furthermore, the similarity in the amounts of cGMP produced by the three cell lines after stimulation with Ca²⁺-mobilizing agents in the presence of 3 mM Ca²⁺ showed that all the NOS isoforms activate sGC to the same extent.

In conclusion, this study has identified several differences in the functional characteristics of nNOS, eNOS, and a soluble mutant of eNOS when overexpressed in HEK 293 cells. The eNOS isoform is the most efficiently regulated, as shown by the lowest basal activity and hence the largest degree of increase upon stimulation, whereas the soluble eNOS mutant is slightly more active under basal conditions, presumably as a result of the loss of inhibitory membrane interaction. However, membrane association does not appear to be necessary for efficient Ca²⁺ activation of eNOS by receptor agonists and other Ca²⁺-mobilizing agents.