Estimation of nitric oxide synthase activity via LC-MS/MS determination of ¹⁵N₃-citrulline in biological samples

Abstract

Rationale

We showed that the metabolite peaks of ¹⁵N₃-citrulline (¹⁵N₃-CIT) and ¹⁵N₃-arginine (¹⁵N₃-ARG) could be detected when ¹⁵N₄-ARG was metabolized by nitric oxide synthase (NOS) in endothelial cells. The usefulness of these metabolites as potential surrogate indices of nitric oxide (NO) generation is evaluated.

Methods

A hydrophilic-interaction liquid chromatography electrospray tandem mass spectrometric assay (LC-MS/MS) was utilized for the simultaneous analysis of ¹⁵N₄-ARG, ARG, CIT, ¹⁵N₃-CIT and ¹⁵N₃-ARG. ¹⁵N₃-CIT and ¹⁵N₃-ARG from impurities of ¹⁵N₄-ARG were determined and corrected for the calculation of their concentration. ¹⁵N₄-ARG-derived NO, i.e., ¹⁵NO formation was determined by analyzing ¹⁵N-nitrite accumulation by another LC-MS/MS assay.

Results

After EA.hy926 human endothelial cells were challenged with ¹⁵N₄-ARG for 2 hours, the peak intensities of ¹⁵N₃-CIT and ¹⁵N₃-ARG significantly increased with ¹⁵N₄-ARG concentration and positively correlated with ¹⁵N-nitrite production. The estimated Km values were independent of the metabolite (i.e., ¹⁵N₃-CIT, ¹⁵N₃-CIT+¹⁵N₃-ARG or ¹⁵N-nitrite) used for calculation. However, after correction for its presence as a chemical contaminant of ¹⁵N₄-ARG, ¹⁵N₃-ARG was only a marginal contributor for the estimation of NOS activity.

Conclusions

These data suggest that the formation of ¹⁵N₃-CIT can be used as an indicator of NOS activity when ¹⁵N₄-ARG is used as a substrate. This approach may be superior to the radioactive ¹⁴C-CIT method which can be contaminated by ¹⁴C-urea, and to the ¹⁴N-nitrite method which lacks sensitivity.

1 Introduction

As an endogenous signaling molecule, nitric oxide (NO) is involved in a variety of physiological processes including blood pressure regulation, platelet aggregation/adhesion, neurotransmission as well as cellular defense.^[1] Physiologically, NO is produced by the enzyme nitric oxide synthase (NOS) which converts L-arginine (ARG) into L-citrulline (CIT) and NO. NOS exists in a variety of isoforms including NOS1, NOS2 and NOS3 which are encoded by different genes.^[2] While NOS1 is constitutively expressed in neuronal and certain epithelial cells, NOS2 is inducible with lipopolysaccharide and cytokines in a multitude of different cells. The last isoform, NOS3 (endothelial NOS, eNOS) has been found to be constitutively expressed in endothelial cells. The activities of NOS1 and NOS3 are known to be regulated by Ca²⁺ and calmodulin. In endothelial cells, NO keeps blood vessels dilated, prevents the adhesion of platelets and white cells, and inhibits vascular smooth muscle proliferation.^[2] The loss of adequate production of NO in the endothelium is a major cause of endothelial dysfunction, which is a hallmark of a variety of cardiovascular diseases such as hypertension, atherosclerosis and diabetes.

Due to the pivotal role of NO in physiology, numerous analytical methods have been developed and used to quantify NO production via the NOS-mediated enzymatic conversion of ARG to NO.^[3] Direct measurement of NO in biological fluid is extremely difficult because NO is a short-lived molecule with an estimated in vivo half-life in human blood of 3–5 seconds.^[4] Therefore, oxidation end-products of NO, viz., inorganic nitrite and nitrate ions, are measured in blood and urine, and their accumulation has been used as an index of NO production. Various analytical approaches including colorimetric assay based on Griess reaction,^[5] fluorometric assay^[6] as well as gas chromatography-mass spectrometry^[7] have been applied to measure nitrite and nitrate. However, the extent to which these ion fluxes represent quantitative NO production under diverse experimental conditions is unknown.^[8] In addition, sources of nitrite and nitrate other than from ARG, e.g., dietary intake, would need to be accounted for.^[9]

In addition to NO, CIT is also produced from ARG by NOS. This amino acid, however, is also formed via other pathways, such as the urea cycle and the dimethylarginine dimethylaminohydrolase-mediated metabolism of asymmetric dimethylarginine (ADMA). Thus, a simple quantitation of total CIT production does not provide an accurate estimation of NO production. In order to discriminate the NOS-derived CIT from other sources, radiolabeled ARG, usually using ³H or ¹⁴C, has been utilized to enable measurement of radioactive CIT.^[10] In this assay, unreacted radiolabeled ARG is removed by retention on a cation exchange column which allows radiolabeled CIT to pass through the column, and NOS activity is determined simply by counting radioactivity in the effluent. However, it was pointed out that, without further chromatographic separation, arginase catalyzed conversion of ¹⁴C-ARG to ¹⁴C-urea would be included erroneously as part of the radioactive signal of ¹⁴C -CIT, as shown in rat liver mitochondria.^[11]

We have reported a liquid chromatography tandem mass spectrometric (LC-MS/MS) assay for the simultaneous bioanalysis of ARG, ¹⁵N₄-ARG, CIT, and methylated arginines.^[12] Because discrete cellular compartments for ARG have been proposed,^{[13, 14]} the use of isotope-labeled ARG, e.g., ¹⁵N₄-ARG, as an exogenous substrate is advantageous since it allows insights into the relative dispositional fates of exogenous vs. endogenous ARG.

When applying this assay to cellular studies, we observed two chromatographic peaks in cell lysate samples after cells were exposed to ¹⁵N₄-ARG.^[12] These peaks were tentatively identified as those of ¹⁵N₃-CIT and ¹⁵N₃-ARG. Since ¹⁵N₃-CIT can be formed with ¹⁵NO from the metabolism of ¹⁵N₄-ARG by NOS, and it can then be further metabolized to ¹⁵N₃-ARG via the urea cycle, monitoring of these peaks may theoretically allow for an estimation of NOS activity. The objective of the present study, therefore, is to evaluate the use of peak intensity of ¹⁵N₃-CIT as surrogate index of NO production when an exogenous ARG source (¹⁵N₄-ARG) is used. This approach was then compared to the extent of ¹⁵N-nitrite production in cultured human endothelial cells.

2 Materials and Methods

2.1 Chemicals and reagents

ARG [as L-arginine HCl], CIT [as L-citrulline], 4-(2-Hydroxyethyl)-1 piperazineethanesulfonic acid (HEPES), CaCl₂·4H₂O, bovine brain calmodulin, (6R)-5,6,7,8-tetrahydro-l-biopterin (BH₄), NADPH, ethylene diamine tetraacetic acid (EDTA), ethylene glycol tetraacetic acid (EGTA), and L-N^G-Nitroarginine methyl ester (L-NAME) were purchased from Sigma (St. Louis, MO). ¹⁵N₄-ARG [as its HCl, (U-¹⁵N₄, 98%)], the internal standard for ARG, ¹³C₆-ARG [as ARG:HCl (U-¹³C₆, 98%)], the internal standard for CIT, D₄-CIT [as L-citrulline (4,4,5,5-D₄, 96.5 %)], and ¹⁵N-nitrite [as NaNO₂ (¹⁵N, 98%+)] were obtained from Cambridge Isotope Laboratories, Inc. These compounds were used without further purification. Cell culture reagents were purchased from Invitrogen.

2.2 In vitro cell study

EA.hy926 human vascular endothelial cells, which is an immortalized cell line derived from human umbilical vein endothelial cells,^[15] were obtained as a gift from the University of North Carolina. Cells were grown in a low glucose-low ARG modified Dulbecco’s modified Eagle’s medium (DMEM) containing 0.9 g·L⁻¹ of glucose and 21 mg·L⁻¹ of ARG^[16] supplemented with 10 % fetal bovine serum, and 100 U·mL⁻¹ penicillin and 100 μg·mL⁻¹ streptomycin at 37 °C in a 5 % CO₂ incubator for 7 days.

For intact cell experiments, cells were washed twice with phosphate-buffered saline and equilibrated in Locke’s solution (LS; 154 mM NaCl, 5.6 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, 10 mM HEPES, 3.6 mM NaHCO₃ and 5.6 mM glucose) for 1 hour before the experiment. Different concentrations of ¹⁵N₄-ARG (0, 20, 50, 100, 200, 500 μM, equivalent to 0, 3.6, 8.9, 17.8, 35.6, 89.1 mg·L⁻¹) were added to the cells. After 2 hours, the cell incubation medium was collected and cells were lysed and collected for analysis. For cell lysates experiments, cells were lysed with lysis buffer before the experiment. Different concentrations of ¹⁵N₄-ARG (0, 5, 10, 20, 50, 100 μM, equivalent to 0, 0.9, 1.8, 3.6, 8.9, 17.8 mg·L⁻¹) were added to the cell lysates. L-NAME (100 μM) was added to the reaction mixture to examine NOS inhibition. After 30 min, cell lysate samples were collected for analysis. Protein concentrations in the cell lysates were determined by Lowry assay.^[17]

2.3 Rat liver mitochondria study

Mitochondria were isolated from an adult male Sprague Dawley rat liver according to Raha et al.^[18] The isolated liver was finely minced and gently homogenized in homogenization buffer (50 mM HEPES pH 7.4, 70 mM sucrose, 220 mM mannitol, and 1 mM EGTA). Following homogenization the mixture was centrifuged at 3000 × g for 10 min. The supernatant was then centrifuged at 15,000 × g for 10 min. The pellet was gently solubilized in homogenization buffer and centrifugation steps were repeated. The pellet was resolubilized and recentrifuged at 15,000 × g for a final time and the final mitochondrial pellet was resuspended in buffer (0.34 M sucrose, 100 mM KCl, 10 mM Tris-Cl, 1 mM EDTA). The protein concentration was determined and adjusted to 5 mg·mL⁻¹. The cytochrome c oxidase activity of the prepared mitochondria was measured by a colorimetric assay using a Cytochrome c Oxidase Assay Kit (Sigma, MO). The decrease in the absorbance at 550 nm was monitored by a spectrophotometer.

To assess metabolite production from exogenous ARG (¹⁵N₄-ARG) in the mitochondria, reactions were run according to Venkatakrishnan et al.^[11] An aliquot of 150 μg of isolated mitochondria was added to the reaction mixtures which consisted of 50 mM HEPES (pH 7.6), 400 μM NADPH, 400 μM CaCl₂, 5 μM tetrahydrobiopterin, 50 μM ¹⁵N₄-ARG, and 1.5-fold molar excess of calmodulin in a total volume of 0.25 mL. After incubation at 23 °C for 10 min, samples were collected for analysis.

2.4 LC-MS/MS assay for ¹⁵N₃-CIT and ¹⁵N₃-ARG

¹⁵N₃-CIT and ¹⁵N₃-ARG were analyzed by hydrophilic-interaction liquid chromatography electrospray tandem mass spectrometry along with ARG, ¹⁵N₄-ARG, CIT, ADMA, and SDMA.^[12] Briefly, an aliquot of 20 μL of the cell lysates or cell medium was mixed with 3 internal standards (20 μL of ¹³C₆-ARG 1μM, 20 μL of D₄-CIT 1 μM, and 20 μL of D₇-ADMA 250 nM) and 120 μL of mobile phase B (acetonitrile containing 0.5% acetic acid and 0.025% trifluoroacetic acid) was added for protein precipitation. After centrifugation at 13,000 × g for 20 min, the supernatant was collected for analysis.

The liquid chromatography consists of Shimadzu LC-20AD delivery pump, SIL-20AC autosampler and CBM-20A system controller (Shimadzu Scientific Instruments; Columbia, MD). Analytes were separated by liquid chromatography on a 150×2.1 mm Alltima HP HILIC 3 μ column with an isocratic elution with 15 % mobile phase A (water containing 0.5 % acetic acid and 0.025 % trifluoroacetic acid) and 85 % mobile phase B (acetonitrile containing 0.5 % acetic acid and 0.025 % trifluoroacetic acid) at a flow rate of 0.25 mL·min⁻¹ for 6 min.

[M+H]+ ions were analyzed in the multiple reaction monitoring mode of the ABI/Sciex API3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) equipped with an electrospray ion (ESI) source. To identify isotopomeric CIT and ARG metabolites of ¹⁵N₄-ARG, MRM transitions of 179.2→71.1, 178.2→71.0, and 177.2→71.0 were added in the MS setting and monitored. The predicted MRM transitions of ¹⁵N₃-CIT and ¹⁵N₃-ARG are m/z 179.2→71.1 and 178.2→71.0, respectively. Due to the unavailability of authentic standards, ¹⁵N₃-CIT and ¹⁵N₃-ARG concentrations were estimated by assuming identical behaviors of isotopomeric compounds, using established calibration curves for unlabeled CIT and ¹⁵N₄-ARG, respectively. D₄-CIT and ¹³C₆-ARG were used as internal standards, respectively. The MRM transitions of D₄-CIT and ¹³C₆-ARG are m/z 180.2 → 74.1 and 181.2 → 74.2, respectively.

2.5 LC-MS/MS assay for ¹⁵N-nitrite

¹⁵N-nitrite was determined by an LC-MS/MS assay.^[19] An aliquot of cell lysates and cell medium sample 100 μL was first incubated with 10 μL of 0.5 mM 2,3-diaminonaphthalene (DAN, in 0.5 M HCl) for 10 min at room temperature to convert ¹⁵N-nitrite in the sample to ¹⁵N-2,3-naphthotriazole (¹⁵N-NAT) using the procedure reported by Misko et al.^[6] After addition of acetonitrile for protein precipitation, samples were centrifuged at 13,000 × g for 5 min. An aliquot of 20 μL of the supernatant was then mixed with 5 μL of the internal standard, 1H-naphth[2,3-d]imidazole (5 ng·mL⁻¹), and diluted with 170 μL of water and injected to LC-MS/MS. The liquid chromatography consisted of a Shimadzu LC-20AD delivery pump, SIL-20AC autosampler, and CBM-20A system controller (Shimadzu Scientific Instruments; Columbia, MD). Chromatographic separation was accomplished on Agilent XDB-C18, 2×50mm, 5 μ column by a gradient elution of 5% mobile phase B (0.1 % v/v formic acid in acetonitrile) in mobile phase A (0.1 % v/v formic acid in water), rising linearly to 95 % mobile phase B in 4 min. The flow rate was 0.3 mL/min. [M+H]⁺ ions were analyzed in the multiple reaction monitoring (MRM) mode of the ABI/Sciex API3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) equipped with an electrospray ion (ESI) source. The MRM transitions were m/z 170.1→115.0 for ¹⁴N-NAT, m/z 171.1→115.0 for ¹⁵N-NAT, and m/z 169.2→115.0 for the IS.

2.6 Kinetic calculations

NOS activity was assessed based on the production of various products, by fitting the observed data to the Michaelis-Menten equation.

$$\mathrm{V}=\frac{\text{Vmax}\times \mathrm{S}}{\text{Km}+\mathrm{S}}$$

where V is the reaction rate, Vmax is the maximum rate, S is the concentration of the substrate, i.e., ¹⁵N₄-ARG, and Km is the substrate concentration at which the activity is half-maximal. The kinetic parameters, Vmax and Km were estimated by nonlinear regression of substrate dependence.

V was estimated by the accumulated amounts of either ¹⁵N-nitrite, ¹⁵N₃-CIT, or ¹⁵N₃-CIT + ¹⁵N₃-ARG over the incubation time of 120 min, and corrected for protein content. Individual data were fitted to the Michaelis-Menten equation using WinNonlin 5.0 (Pharsight Corporation, Mountain View, CA).

2.7 Statistical analysis

Statistical analyses were performed using one-way ANOVA. Comparisons among different groups were followed using Tukey post hoc analysis. Differences with p<0.05 were denoted as statistically significant.

3 Results

3.1 Observation of new chromatographic peaks from cell samples after exposure to ¹⁵N₄-ARG

After EA.hy 926 human endothelial cells were exposed to ¹⁵N₄-ARG for 2 hours, cell lysate samples were analyzed to examine peaks representing a series of isotopomeric CIT and ARG compounds in the ion chromatogram. Based on the theoretical molecular masses and fragmentation patterns of possible metabolites, selected MRM transitions of 179.2→71.1, 178.2→71.0, and 177.2→71.0 were performed. Under each MRM transition, peaks with the retention time of CIT, 2.83 min, and that of ARG, 3.33 min were examined.

Fig. 1 shows the observed ion chromatograms with MRM transition of m/z 179.2→71.1(peak A), 178.2→71.0 (peak B), and 177.2→71.0 (peak C) with retention times of 2.83 min (that of CIT) and 3.33 min (that of ARG). The intensity of peak A and peak B increased in a saturable manner when EA.hy926 human endothelial cells were exposed to increasing extracellular ¹⁵N₄-ARG concentrations (Supplementary Fig. S1A and B). However, the intensity of peak C was only marginally dependent on extracellular ¹⁵N₄-ARG concentrations in appearance only, and statistical significance could not be demonstrated (p>0.05, Supplementary Fig. S1C).

Figure 1.

Representative ion chromatograms under MRM transitions of 179.2→71.1 (peak A), 178.2→71.0 (peak B), and 177.2→71.0 (peak C) in a human endothelial cell lysate sample obtained after cells were incubated with 100 μM of ¹⁵N₄-ARG for 2 hours.

3.2 Tentative assignments of observed peaks

Table 1 shows the theoretical MRM transition and chromatographic properties of various CIT and ARG isomers. It is apparent that the observed characteristics of peak A were consistent with that of ¹⁵N₃-CIT, and those of peak B with ¹⁵N₃-ARG. Both of these compounds are generated as a result of metabolism of ¹⁵N₄-ARG to CIT, and the back conversion of CIT to ARG (Fig. 2). Thus, peaks A and B were then tentatively assigned as those for ¹⁵N₃-CIT and ¹⁵N₃-ARG, respectively. The chromatographic characteristics of peak C are consistent with those of ¹⁵N₂-ARG (Table 1). However, formal identity of these chromatographic peaks can only be made when the relevant authentic standard becomes available or when high resolution mass spectrometry technique is applied.^[20]

Table 1.

LC-MS/MS properties of stable isotope-labeled isomers of CIT and ARG originating from the metabolism of ¹⁵N₄-ARG.

Compounds		Expected MRM transition (m/z)	Retention time (min)	Results
CIT analogs	15N3-CIT	179.2→71.1	2.83	Peak A in Fig. 1
	15N2-CIT	178.2→71.0	2.83	Not detected
	15N1-CIT	177.2→71.0	2.83	Not detected
ARG analogs	15N3 -ARG	178.2→71.0	3.33	Peak B in Fig. 1
	15N2-ARG	177.2→71.0	3.33	Peak C in Fig. 1

Figure 2.

Schematic representation of ¹⁵N₃-CIT and ¹⁵N₃-ARG formation from ¹⁵N₄-ARG. When ¹⁵N₄-ARG is used as a substrate for NOS, ¹⁵N₃-CIT is produced via NOS and subsequent urea cycle enzymes, argininosuccinate synthase (ASS) and argininosuccinate lyase (ASL), form ¹⁵N₃-ARG. Further metabolism of ¹⁵N₃-ARG by NOS produces unlabeled NO and ¹⁵N₃-CIT.

3.3 Other sources of ¹⁵N₃-CIT and ¹⁵N₃-ARG in samples

To determine possible sources of these compounds from chemical contamination and blank cells, an experiment was carried out to examine their presence using a broad range of ¹⁵N₄-ARG standard solutions (0, 1, 4, 10, and 20 μM, n=2 each) in blank cell lysates without incubation. The concentrations of ¹⁵N₄-ARG standard solutions were chosen to cover the achievable intracellular ¹⁵N₄-ARG concentration range after cells were challenged with 0 – 500 μM of ¹⁵N₄-ARG.

When these samples were analyzed directly without incubation, there was no peak of ¹⁵N₃-CIT found in the ion chromatogram, suggesting that this compound was generated by metabolism, rather than as an impurity. On the other hand, significant ¹⁵N₃-ARG peaks were observed. The peak intensity of ¹⁵N₃-ARG increased proportionally with the added ¹⁵N₄-ARG concentration, with a mean value of 2.4 ± 0.3 % of ¹⁵N₄-ARG (n=8), over the concentration range studied. This percentage was interpreted as arising from the presence of ¹⁵N₃-ARG as an impurity in the ¹⁵N₄-ARG commercial sample. In subsequent data analysis, therefore, all observed ¹⁵N₃-ARG peak intensities were corrected for the presence of this compound as an impurity in ¹⁵N₄-ARG.

3.4 Correlation between ¹⁵N₃-CIT and ¹⁵N₃-ARG vs. ¹⁵N-nitrite

To evaluate whether ¹⁵N₃-CIT and ¹⁵N₃-ARG production can estimate NOS activity, the peak intensities of ¹⁵N₃-CIT and ¹⁵N₃-ARG were compared to another index of NOS activity, ¹⁵N-nitirte production, by using cell samples after EA.hy926 human endothelial cells were exposed to different concentrations of ¹⁵N₄-ARG for 2 hours. After incubation, an aliquot of the sample was used to determine peak intensities of ¹⁵N₃-CIT and ¹⁵N₃-ARG and an aliquot of the same sample was used to determine ¹⁵N-nitrite concentrations and two measurements were compared. The peak intensities of ¹⁵N₃-CIT and ¹⁵N₃-ARG increased with an increase in the observed ¹⁵N-nitrite concentration (Supplementary Fig. S2).

To test whether the combined ¹⁵N₃-CIT + ¹⁵N₃-ARG concentrations would correlate better to ¹⁵N-nitrite than ¹⁵N₃-CIT alone, we estimated the concentrations ¹⁵N₃-CIT and ¹⁵N₃-ARG assuming the calibration curves. Based on this assumption, the contribution of ¹⁵N₃-ARG, after correction as an impurity from added ¹⁵N₄-ARG, was relatively small to the total estimated concentration of ¹⁵N₃-CIT + ¹⁵N₃-ARG. In principle, the metabolism of ARG to CIT and NO should produce a 1:1 ratio of these metabolites. However, results showed that the estimated concentrations of ¹⁵N₃-CIT were 14.5-fold higher than those of ¹⁵N-nitrite (Supplementary Fig. S3). Statistical analyses of the data indicate significant correlations between ¹⁵N₃-CIT vs. ¹⁵N-nitrite (Pearson correlation coefficient, p = 0.862, R² = 0.740), ¹⁵N₃-CIT + ¹⁵N₃-ARG vs. ¹⁵N-nitrite (p = 0.865, R² = 0.744), and ¹⁵N₃-CIT vs. ¹⁵N₃-ARG (p = 0.998, R² = 0.931).

3.5 Comparison of kinetic parameters of NOS activity

Michaelis-Menten plots of estimated NOS activity based on the formation of different metabolic products as a function of ¹⁵N₄-ARG concentration in the intact cells and cell lysates are shown in Fig. 3 and Fig. 4, respectively. In intact cells, the estimated Km values were independent of the metabolic product used. The Km in the intact cells were higher than that in the cell lysates. On the other hand, the estimates of Vmax varied depending on the metabolic product used. The estimated Vmax values were about 10-fold higher using either ¹⁵N₃-CIT or ¹⁵N₃-CIT+¹⁵N₃-ARG, as compared to using ¹⁵N-nitrite as an indicator of metabolite formation in the intact cells (Table 2). Unlabeled nitrite and nitrate were also measured in the cell lysates. However, the ¹⁵N₄-ARG concentration dependent production of unlabeled nitrite and nitrate was not clear, which hampered to estimate kinetic parameters likely due to the high baseline values (Supplementary Fig. S4).

Figure 3.

Michaelis-Menten plots of estimated NOS activity in EA.hy926 cells based on the formation of different metabolic products after exposure to 0 – 500 μM of ¹⁵N₄-ARG for 2 hours (n=6 for each concentration). The drawn lines are fits to the observed data using the Michaelis-Menten equation.

Figure 4.

Michaelis-Menten plots of estimated NOS activity in EA.hy926 cell lysates based on the formation of ¹⁵N₃-CIT after cell lysates were exposed to 0 – 100 μM of ¹⁵N₄-ARG for 30 min (n=3 for each concentration). The drawn lines are fits to the observed data using the Michaelis-Menten equation.

Table 2.

Estimated parameters of NOS activity by using different metabolic products (n=6 for each ¹⁵N₄-ARG concentration).

Indicator	Vmax (pmol·mg−1·min−1)	Km (μM)
	Estimate (CV%)	Estimate (CV%)
Intact cells
15N-Nitrite	0.506 (11.9)	46.6 (29.5)
15N3-CIT	4.22 (12.6)	59.6 (28.1)
15N3-CIT+15N3-ARG	5.11 (13.1)	68.0 (27.6)
Cell lysates
15N3-CIT	3.18 (9.3)	5.33 (44.1)

3.6 Effects of NOS inhibitor on ¹⁵N₃-CIT production

Fig. 5 shows the effects of L-NAME on the production of ¹⁵N₃-CIT in the cell lysates. In the presence of L-NAME at 100 μM in the reaction mixture, ¹⁵N₃-CIT formation was significantly diminished (p <0.05).

Figure 5.

Production of ¹⁵N₃-CIT in the human endothelial cell lysates after cell lysates were incubated with ¹⁵N₄-ARG (100 μM) in the presence of NOS inhibitor, L-NAME (100 μM) for 30 min. Data are presented as the mean ± SD (n=3). *, p<0.05 vs. control (i.e., without ¹⁵N₄-ARG). #, p<0.05 vs. ¹⁵N₄-ARG (i.e., without L-NAME).

3.7 Application of ¹⁵N₃-CIT assay to rat liver mitochondria

Arginase catalyzed conversion of ¹⁴C-ARG to ¹⁴C-urea has been suggested to present a confounding factor in using radioactive ¹⁴C-CIT as an index of NOS activity. This was pointed out by studies that showed apparent ¹⁴C-CIT conversion in rat liver mitochondria, which did not have NOS activity.^{[11, 21]} Using the same rat liver mitochondria preparation, we examined whether ¹⁵N₃-CIT formation can be observed.

Using cytochrome c oxidase activity as an index, we found that the isolated mitochondria fraction had a 5.5-fold enrichment in enzyme activity compared to whole homogenate, and a 17.0-fold enrichment compared to the supernatant fraction. Significant presence of ¹⁵N₃-CIT was found to be formed when the whole liver homogenate was incubated with ¹⁵N₄-ARG (Fig. 6). However, with the isolated rat mitochondria fraction, 2 of the 4 samples did not show any measurable ¹⁵N₃-CIT, while the other two samples revealed marginal peak intensity. The mean ¹⁵N₃-CIT peak intensity found in the mitochondria preparation was not statistically different with or without added ¹⁵N₄-ARG. No appreciable ¹⁵N₃-ARG was found in any of these samples.

Figure 6.

Production of ¹⁵N₃-CIT in the whole rat liver homogenate or purified rat liver mitochondria fraction after incubation with 50 μM of ¹⁵N₄-ARG for 10 min. Data are presented as the mean ± SD (n=4). ***, p<0.001 vs. control (i.e., without ¹⁵N₄-ARG).

4 Discussion

In the present study, metabolite formation (¹⁵N₃-CIT and ¹⁵N₃-ARG) from stable isotope labeled ARG (¹⁵N₄-ARG) was evaluated as a possible index of NOS activity in endothelial cells and cell lysates. When ¹⁵N₄-ARG is used as a substrate for NOS, ¹⁵N₃-CIT and ¹⁵NO should be produced at equal molar amounts (Fig. 2). After ¹⁵N₃-CIT is produced, two urea cycle enzymes, i.e., argininosuccinate synthase and argininosuccinate lyase, convert it back to ¹⁵N₃-ARG via ¹⁵N₃-argininosuccinate, which couples with an unlabeled nitrogen from aspartate.^{[22, 23]} Further metabolism of ¹⁵N₃-ARG by NOS produces ¹⁵N₃-CIT and unlabeled NO. Rapid equilibrium of the guanidine group possibly generate ¹⁵N₂-CIT and ¹⁵NO from ¹⁵N₃-ARG. However, ¹⁵N₂-CIT was not detected (Table 1), indicating ¹⁵N₃-CIT and unlabeled NO are dominantly produced from ¹⁵N₃-ARG in our experimental system. Thus, after the first metabolic step of ¹⁵N₄-ARG, subsequent urea cycle lead to repeated formation of ¹⁵N₃-CIT and ¹⁵N₃-ARG in a futile cycle (Fig. 2). Note that with the added substrate ¹⁵N₄-ARG, only isotopically labeled ¹⁵N-NO (measured as ¹⁵N-nitrite) is produced. However, through the subsequent urea cycle (CIT→ARG recycling system), the NO comes from the unlabeled nitrogen atom derived from aspartate, and unlabeled NO was formed as a consequence.

While we could not determine the true concentrations of ¹⁵N₃-CIT and ¹⁵N₃-ARG in this study because authentic compounds are not yet available, we estimated their concentrations based on the assumption that the calibration curves of CIT and ¹⁵N₄-ARG could be applied. A partial justification of this assumption is that the LC-MS/MS peak intensity responses of ¹⁴N-ARG, ¹⁵N₄-ARG, ¹³C₆-ARG were identical, and the calibration curves of ¹⁴N-ARG and ¹⁵N₄-ARG gave the same slopes when ¹³C₆-ARG was used as an internal standard.^[12] Similarly, the LC-MS/MS peak intensity responses of ¹⁴N-CIT vs. D₄-CIT were also identical. Thus, isotopic labeling of CIT and ARG did not appear to affect their chromatographic responses dramatically. Naturally, the validity of the assumption used must be confirmed experimentally using authentic compounds of ¹⁵N₃-CIT and ¹⁵N₃-ARG, if and when they become available.

Based on this estimation, ¹⁵N-nitrite production was positively related to either ¹⁵N₃-CIT production alone, or to the combined ¹⁵N₃-CIT + ¹⁵N₃-ARG accumulation (Supplementary Fig. S3). However, ¹⁵N-nitrite levels were 14.5 ± 4.4 and 18.7 ± 5.6-folds less than those of ¹⁵N₃-CIT alone, and combined ¹⁵N₃ -CIT + ¹⁵N₃-ARG, respectively. Higher levels of ¹⁵N-CIT compared to ¹⁵N-nitrite and ¹⁵N-nitrate were also reported when L-[guanidino-¹⁵N₂]-ARG was used as a substrate.^[24] The use of ¹⁵N-nitrite as a surrogate measure of produced ¹⁵NO also assumes that the predominant route of ¹⁵NO disappearance occurs through conversion to nitrite. However, it is possible that ¹⁵NO may be lost via evaporation and/or oxidized to other product, i.e. ¹⁵N-nitrate. Direct nitration of cellular proteins, e.g., to form nitrated tyrosine, is also possible. Thus, ¹⁵N-nitrite accumulation may represent an under-estimate of ¹⁵NO production. It is also possible that potential abundant interference and/or additional pathways are contributory to labeled CIT in cell systems.^[24]

Analysis of NOS activity using isotopic-labeled ARG typically has been obtained via determination of ¹⁴C-CIT, formed from ¹⁴C-ARG. Flam et al.^[25] reported that NO production which was measured by nitrite concentration, exceeded apparent ¹⁴C-CIT formation at least by 8-fold in bovine aortic endothelial cells, indicating that radioactive CIT formation may significantly underestimate the actual NO production. The favored formation of unlabeled nitrite over ¹⁴C-CIT is opposite to our present finding that ¹⁵N-nitrite levels were 14.5-folds lower than that of ¹⁵N₃-CIT (Supplementary Fig. S3). This discrepancy, however, can be explained by tracking the fate of ¹⁴C-CIT and NO in the radioactive assay (Supplementary Fig. S5). Here, since the isotope labeling is made on the carbon atom, rather than the nitrogen atom, metabolism of ¹⁴C-ARG would generate unlabeled NO and urea cycle would form a futile cycle of ¹⁴C-CIT and ¹⁴C-ARG along with new formation of NO. Thus, utilization of ¹⁴C-CIT as an index of NOS production only measures NO production from the added substrate, and not from urea cycle. In contrast, use of unlabeled nitrite measures NO production from both the added substrate alone and from urea cycle.

Our results also indicate that the apparent production of ¹⁵N₃-ARG, after correction for its presence as a chemical contaminant of ¹⁵N₄-ARG, was only a marginal contributor for the estimation of NOS activity (Supplementary Fig. S3). Unlike ¹⁵N₃-CIT, which has only one major metabolic pathway operating through urea cycle, ¹⁵N₃-ARG can be metabolized by several other enzymes besides NOS, viz., arginine decarboxylase, arginine:glycine amidinotransferase and arginase.^[26] Thus, ¹⁵N₃-ARG may be partially removed from the futile ¹⁵N₃-CIT→¹⁵N₃-ARG cycle through these enzymes.

Kinetic analysis using the various metabolic products of ¹⁵N₄-ARG showed that the Km values estimated from ¹⁵N₃-CIT or ¹⁵N₃-CIT +¹⁵N₃-ARG were similar to those from ¹⁵N-nitrite, suggesting that all these metabolites utilize the same enzyme. The magnitude of Km we observed in the intact endothelial cells is consistent with a value of 29 ± 6 μM in bovine aortic endothelial cells, using radiolabeled ARG to CIT conversion as an index.^[27] The Km estimates in the cell lysates in the present study were also comparable with the literature reports of ~ 3 μM which was determined in the isolated enzyme preparation.^[27] The higher Km in the intact cells compared with that in the cell lysates may be contributed to the ¹⁵N₄-ARG transport into the cell before its metabolism. As discussed, the different Vmax estimated values obtained likely reflected the underestimation of ¹⁵NO production when ¹⁵N-nitrite was used as the metabolite index.

Finally, we tested the application of this LC-MS/MS method to estimate NOS activity in the rat liver mitochondria preparation.^{[11, 21]} The presence of mitochondrial NOS has long been disputed. Several investigators have pointed out errors in methodologies used in documenting this activity^[28–30] including the presence of apparent activity which could not be blocked by NOS inhibitors. Recently, it was shown that the apparent ¹⁴C-CIT radioactivity signal may have included ¹⁴C-urea, which is formed via arginase-mediated metabolism of ARG.^{[11, 21]} Here, we showed that insignificant ¹⁵N₃-CIT formation in rat liver mitochondria, while significant accumulation of this metabolite can be observed in rat liver homogenates.

The use of ¹⁵N₃-CIT as an index of NOS activity in cell culture studies may possess several technical advantages over that of ¹⁴C-CIT. First, stable isotopes are used in place of radioactive compounds, enabling better safety and easier disposal of experimental waste. Second, the LC-MS/MS method offers additional assay capabilities as other ARG-related compounds such as the dimethylarginines can also be simultaneously determined. Third, the identity of the studied compounds can be made with better certainty because the LC-MS/MS method incorporates chromatographic separation and mass identification along with fragmentation patterns.

Compared to the use of inorganic nitrite as an index of generated NO activity, the present method may also possesses several advantages. First, endogenous (unlabeled) nitrite levels in biological systems are high (in μM’s), so the detection of generated NO (as measured by nitrite) lacks sensitivity. Second, as mentioned previously, the conversion of NO to nitrite in biological systems may not be complete because of NO evaporation and other reactions. The present method suggests that the sum of ¹⁵N₃-CIT and ¹⁵N₃-ARG would represent the total NO produced from the added ¹⁵N₄-ARG substrate.

However, the current method also carries some disadvantages. First, an authentic standard of ¹⁵N₃-CIT is not yet available to allow for unambiguous estimation of the concentrations of the compound. Second, LC-MS/MS methods are more costly to run, and require more extensive instrumentation and resources. However, the present study points out the potential utility of LC-MS/MS methodology as an alternative for the existing ¹⁴C-CIT assay in determining NOS activity.

5 Conclusions

When ¹⁵N₄-ARG is used as a substrate in cell culture studies, metabolite formation of ¹⁵N₃-CIT may be used as a surrogate indicator of NO generation and assessment of NOS activity. Determination of this isotopic metabolite may allow for the simple and sensitive examination of NO production from the exogenous ARG.