Catalysis of S-nitrosothiols formation by serum albumin: The mechanism and implication in vascular control

Abstract

Nitric oxide (NO^⋅) is a short-lived physiological messenger. Its various biological activities can be preserved in a more stable form of S-nitrosothiols (RS-NO). Here we demonstrate that at physiological NO^⋅ concentrations, plasma albumin becomes saturated with NO^⋅ and accelerates formation of low-molecular-weight (LMW) RS-NO in vitro and in vivo. The mechanism involves micellar catalysis of NO^⋅ oxidation in the albumin hydrophobic core and specific transfer of NO⁺ to LMW thiols. Albumin-mediated S-nitrosylation and its vasodilatory effect directly depend on the concentration of circulating LMW thiols. Results suggest that the hydrophobic phase formed by albumin serves as a major reservoir of NO^⋅ and its reactive oxides and controls the dynamics of NO^⋅-dependant processes in the vasculature.

NO^⋅ is synthesized by various types of cells and involved in numerous biological functions, including vasodilation, platelet aggregation, neurotransmission, and inflammation (1, 2). Heme proteins such as guanylyl cyclase and free radical species—e.g., (^⋅O)—serve as NO^⋅ primary targets (1–3). Another physiologically significant component of NO^⋅ biochemistry involves the formation of thionitrite esters with cysteine (Cys) or Cys residues (S-nitrosothiols; RS-NO). Low-molecular-weight (LMW) RS-NO (e.g., S-nitrosoglutathione) and nitroso-derivatives of proteins such as albumin and hemoglobin exert NO-like activity in vivo. They cause arterial and venous smooth muscle relaxation, inhibit platelet aggregation, and activate guanylyl cyclase (4–8).

Vasoactive S-nitrosothiols are known to be generated in vivo (6–10), although the actual amount of these species in circulation is debated (refs. 11 and 12, and references therein). Because RS-NO are relatively stable and can release NO^⋅ when required, via reactions with transition metal ions or other reducing agents (13–15), they are envisioned as a buffering system that controls intra- and extracellular activities of NO^⋅ and magnify the range of its action. Once formed, circulating RS-NO can deliver NO into cytosol via specific mechanisms (16) or directly transfer the nitrosyl cation (NO⁺) to another thiol via the so-called transnitrosation reaction that ensures the dynamic state of RS-NO in vivo (7, 17). S-nitrosylation of protein free Cys residues modulates activities of various regulatory factors and enzymes and represents a widespread signaling mechanism (refs. 18 and 19, and references therein).

Despite the growing interest in the role of nitrosothiols in biological systems, there is still uncertainty about how they form in vivo. In the absence of an electron acceptor, NO^⋅ is unable to react with nucleophiles under oxygen free conditions, implying that metabolites of NO^⋅ oxidation, such as N₂O₃, are actual nitrosating agents (20–23). However, considering the low concentration and short life span of NO^⋅ in vivo, the third-order reaction of NO^⋅ with O₂ (Eq. 1) (k ≈ 4 × 10⁶ M⁻²⋅sec⁻¹) (20, 21) seems to be too slow to account for any detectable amount of circulating RS-NO.

1

2

3

High instability of N₂O₃ in aqueous solution (Eq. 3) further supports this notion. These apparent theoretical constraints can be resolved, however, owing to remarkable properties of NO^⋅ in multiphase systems (19, 24).

NO^⋅ and O₂ are both hydrophobic molecules, and areas of high hydrophobicity should act to increase their local concentration by sequestering them from the surrounding aqueous phase. Under aerobic conditions, high local concentrations of NO^⋅ and O₂ in the hydrophobic phase—e.g., within lipid membranes or protein interiors—can significantly accelerate NO^⋅ oxidation and N₂O₃ formation (19, 25, 26). In other work, we have shown that such micellar catalysis of NO^⋅ oxidation can be efficiently mediated by serum albumin (19). Albumin is the most abundant transport and depot protein in the mammalian vasculature. Considering the large (≈0.8 mM) concentration of albumin in plasma, we assumed that its hydrophobic core can be the major absorber of free NO^⋅ and catalyst of N₂O₃ (19). Here we designed experiments to estimate the impact of albumin-mediated catalysis of NO^⋅ oxidation on the pool of circulating LMW RS-NO. We suggest the mechanism by which NO⁺ is generated in the albumin hydrophobic interior and transferred to various LMW thiols (RSH), and show how this mechanism can control the vascular tone in mammals.

Experimental Procedures

Chemicals and Reagents.

NO^⋅/H₂O solution was prepared as in ref. 19. BSA (Sigma) was >99% pure and free of fatty acids and globulins. BSA^N₂O₃ was prepared by adding 3.5 ml of aqueous NO^⋅ solution to 10 ml of BSA (17 μM) in 30 mM Tris⋅HCl (pH 7.9) so that the concentration of NO^⋅ was ≈100 μM. As soon as the NO^⋅ concentration decreased from 100 μM to less than 1 μM, 1 ml of 3% ammonium sulfamate was added to remove nitrite.

^WBSA and ^CWBSA were prepared by treating BSA (1 mM) or carboxymethylated BSA (^CBSA, 1 mM, Sigma) with 5 mM N-bromosuccinimide (Aldrich), which selectively modifies Trp residues, for 10 min at 25°C in NaCH₃COO/CH₃COOH (pH 5.0), followed by dialysis (MW cut off 12–14.000 membrane) in 2 liters of 20 mM Tris⋅HCl (pH 7.8), 50 mM NaCl, 0.1 mM EDTA, 1 mM C₂H₅SH, and 5% glycerol. Inability of modified Trp-214 and Cys-34 of ^CWBSA to undergo nitrosation was evaluated spectrophotometrically at 330 nm (Fig. 3B). Specifically, 0.2 mM ^CWBSA or BSA (20 μl) was treated with 0.1 M NaNO₂+0.1 M HCl for 10 min, followed by addition of ammonium sulfamate to 3% and 80 mM Tris⋅HCl (pH 7.9) to remove HNO₂. UV-visible spectra were recorded by using an Ultrospec 3000 spectrophotometer (Pharmacia). Spectra were digitized and analyzed by WIN DIG and ORIGIN 6.0 software (Microcal, Northampton, MA).

Figure 3.

Mechanism of albumin-mediated catalysis of LMW RS-NO. (A) The model. Albumin^N₂O₃ can transfer the nitrosonium cation (NO⁺) to LMW RSH via three pathways: (I) directly, (II) via Cys-34, and (III) via Trp-214. (B) Blockage of albumin Cys-34 and Trp-214 nitrosation by chemical modification. UV-visible absorption spectra of newly generated chromophore during the reaction of BSA or ^CWBSA with 0.1 M NaNO₂/HCl. D, absorbance (OD units) with the maximum at λ = 330 nm for N-Trp-NO and 346 nm for S-Cys-NO (19). (C) Role of albumin Cys-34 and Trp-214 in catalysis of LMW RS-NO. Chemically modified BSA that carry S-carboxymethylated Cys-34 (^CBSA), N-oxidized Trp-214 (^WBSA), or both (^CWBSA) (see Experimental Procedures) were used in the experiment similar to that described in Fig. 2B.

Determination of NO^⋅ Partitioning (Q).

To determine Q_NO for albumin, the ice-cold calibrating solution (10 ml, 0.1 M H₂SO₄ + 0.1 M KI) was degassed in the air-tight device by bubbling with argon for 40 min. BSA, ^CWBSA (0.1g dry powder), or rat plasma (0.1 g dry powder, Sigma) was added to the solution and dissolved in the presence of argon. To minimize protein denaturation, the mixture was prepared initially at 0°C. Temperature was then adjusted to 20°C and the ISO-NO Mark II electrode (WPI Instruments, Waltham, MA) was inserted into the solution under argon and sealed. Released NO^⋅ was detected on injection of 0.4 ml of 50 mM NaNO₂. Q_NO is calculated as described in Fig. 1. Albumin serves as a buffer in the mixture with pH 4.7. The solution remained transparent during the experiment, indicating that no significant protein denaturation and aggregation occurred.

Figure 1.

Micellar catalysis of NO^⋅ autooxidation by serum albumin. (A) The model. If partitioning (Q) of NO^⋅ and O₂ in the albumin/H₂O system is >1, the protein hydrophobic core (dark gray) would increase the local concentration of both gases by sequestering them from the aqueous phase and accelerate NO^⋅ autooxidation and N₂O₃ formation. (B) Determination of Q_NO for albumin. On top, the equation and parameters used for Q_NO calculation: [NO]_H2O, NO^⋅ concentration in the aqueous phase; [NO]_h, NO^⋅ concentration in the hydrophobic phase; [NO], NO^⋅ concentration in the aqueous solution of albumin (4.8 μM); V, volume of the sample (10 ml); V_h, volume of the hydrophobic phase (≈28 × 10⁻³ ml, as determined by WEBLAB VIEWER PRO 3.5 software that calculates an approximate volume of hydrophobic compartments of a protein with known atomic structure). Δ = [NO]_H2O − NO is determined experimentally. Infusion of 100 mg of BSA into 10 ml of NO^⋅ solution under O₂-free argon conditions results in decreasing [NO]_H2O from 6.5 μM (bar 1) to 4.8 μM (bar 2) (Δ = 1.7 μm) as detected electrochemically. Similar Δ values were obtained with rat plasma (10 mg/ml) (bar 3) and chemically modified BSA, ^CWBSA, in which Cys-34 and Trp-214 were covalently blocked (see Experimental Procedures) and resistant to nitrosation (bar 4). Because V_h was not determined experimentally but estimated using computer software, actual Q_NO for BSA may be subject to some variation. (C) Time course of nitrite formation in the absence (black dots) or presence of 0.8 mM BSA (squares) or ^CWBSA (triangles). Nitrite concentration, [NO], was measured by Griess reagent immediately after addition of NO^⋅ to 25 μM. (D) Effect of proteolysis on albumin-mediated nitrite formation. At indicated intervals, proteinase K (0.25 mg/ml; Sigma) was added to the NO^⋅-treated BSA solution (0.2 mM). In each case, rapid increase of [NO] occurred (arrows) as measured by Griess reagent 1 min after the protease challenge. Each time point was processed as an independent experiment. The difference in [NO] between protease-treated and nontreated BSA^N₂O₃ was plotted (squares) and used for calculation of the “half-life” of N₂O₃ (t_1/2 ∼ 7 min) in BSA. Schematic interpretation of results is shown at the bottom. t_1/2(N₂O₃) is much longer in the albumin interior than in the aqueous phase. Enzymatic disruption of the protein hydrophobic core instantly releases N₂O₃ into the aqueous phase, resulting in eruption of nitrite.

Determination of Thiols, S-Nitrosothiols, and Nitrite.

Total level of thiols in plasma was determined by Ellman's reagent (27). Plasma glutathione (GSH) was detected at 230 nm by HPLC (Waters) by using C18 column with a mobile phase of 99% 0.1 M monochloroacetate (pH 3) and 1% methanol. Total RS-NO was determined in vitro or in plasma by using CuCl₂ (1.5 mM) to displace NO^⋅ from thiol residues. Released NO^⋅ was detected electrochemically with the NO electrode. GS-NO was used as a standard. It was prepared by incubating equimolar amounts of GSH and NaNO₂ in acidified water on ice. LMW RS-NO in plasma was determined by incubating fresh plasma with CuCl₂ for 3 min while conducting electrochemical detection of NO^⋅. To prepare plasma samples (100 μl), 0.5 ml of arterial blood was collected into heparinized plastic tubes and centrifuged at 4,500 rpm for 3 min. Plasma was processed immediately after preparation. In the experiment of Fig. 5A, 20 μl of GSH (100 mM) in 80 mM Tris⋅HCl (pH 7.9) and 10 μl of CuCl₂ (100 mM) were added to 100 μl of plasma, followed by a detection of the level of liberating NO^⋅ with the NO electrode.

Figure 5.

Catalysis of RS-NO in vivo and its effect on systemic blood pressure. (A) Catalysis of GS-NO by rat plasma. Cu⁺⁺-dependant NO^⋅ production was detected electrochemically in 100 μl of fresh plasma on addition of 10 μl 100 mM GSH (dark gray column) or 10 μl buffer (light gray column). The same experiment was done with plasma after 1 h at 25°C in air (right columns). (B) Schematics explaining the vasodilating effect of exogenous RSH. Increased concentration of LMW RSH in rat vasculature leads to a higher production of RS-NO via albumin^N₂O₃. (C) Systemic vascular response to i.v. bolus of GSH produced a dose-dependent decrease in MAP (Left), which correlated with the increasing level of plasma LMW RS-NO (Right). 5 mM GSH, 12 mM GSH, and 20 mM GSH stand for the expected initial plasma GSH concentration upon GSH infusion. (Inset) The change of plasma GSH concentration as function of time as detected by HPLC analysis. The first plasma sample was ready for analysis 5 min after GSH administration. Blood pressure was measured ≈10 min after GSH administration.

Nitrite was detected with the Griess reagent kit at 540 nm as specified by the manufacturer (Cayman Chemical, Ann Arbor, MI). In the case of Fig. 1C, NO^⋅ (25 μM) was added to 0.8 mM BSA or ^CWBSA in 80 mM Tris⋅HCl (pH 7.9). In the case of Fig. 1D, NO^⋅ (25 μM) was added to 0.2 mM BSA in 80 mM Tris⋅HCl. After indicated time intervals, proteinase K (0.25 mg/ml; Sigma) was added for 1 min at 36°C.

Animals.

Adult male Wistar rats (300–400 g) were used. Both femoral arteries and left femoral vein of anesthetized animals (urethane 1.2 g/kg i.p.) were cannulated using polyethylene tubing (PE-50, WPI). Arterial catheter was used for continuous monitoring of aorta blood pressure and heart rate, and for withdrawing blood samples. Venous catheter was used for drug administration. The same infusion rate of drugs or saline (0.2 ml/min) was applied for all animals. Drugs were used as follows: BSA, BSA^N₂O₃, or ^CMBSA, 0.05 g/ml; saline, 0.9% NaCl; N-nitro-l-arginine methyl ester (l-NAME), 20 mg/kg; GSH, 1 ml to the final concentration 5 mM, 12 mM, or 20 mM, based on the total blood volume (Fig. 5C). Before drug administration, mean arterial pressure (MAP) was recorded and the initial level of RS-NO was determined in plasma as described above. MAP and the heart rate were detected with the pressure sensor connected to the amplifier. The signal was digitally converted at the PowerLab 200 workstation (A. D. Instruments, Milford, MA) and transmitted to PC.

Results

Acceleration of NO^⋅ Oxidation by Serum Albumin: Quantitative Assessments.

Our previous study suggests that NO^⋅ oxidation occurs much faster in the hydrophobic interior of albumin than in surrounding aqueous media, leading to effective nitrosylation of the only free thiol, Cys-34, and the only Trp-214 residues of the protein (19). To calculate the maximum acceleration value (H) of NO^⋅ oxidation by the hydrophobic phase of albumin, we first determined the partition coefficient (Q) of NO^⋅ for the albumin/H₂O system (Fig. 1). For the third-order reaction of NO^⋅ with O₂ (1), H = Q × Q_O2—i.e., acceleration of NO^⋅ oxidation is primarily determined by Q_NO. To calculate Q_NO, which is described by the equation shown in Fig. 1B, we used a Clark-type NO^⋅-electrode attached to an airtight gas/liquid mixing device to determine changes of the NO^⋅ concentration (Δ) upon addition of the BSA under O₂-free conditions (Fig. 1B). NO^⋅ was generated in the reaction: NO + I⁻ + 2H⁺ → NO^⋅ + 1/2I₂ + H₂O, as described in Experimental Procedures. Q_NO(BSA) appears to be ≈120 (Fig. 1B), implying that albumin accelerates the reaction in Eq. 1 at least 120² or 1.4 × 10⁴ times. A similar Δ value was obtained for rat plasma (Fig. 1B), suggesting that the hydrophobic phase formed by albumin is a powerful absorbent of free NO^⋅ and a catalyst of its oxidation in the mammalian vasculature.

To provide independent evidence for albumin-mediated NO^⋅ absorption and to estimate the half-life (t_1/2) of N₂O₃ in the protein interior, we used enzymatic proteolysis as a tool for rapid disruption of the protein hydrophobic core and release of N₂O₃ into the aqueous phase (Fig. 1 C and D). In this and other in vitro studies we used a freshly prepared oxygen-free water solution of NO^⋅. Addition of NO^⋅ to an aerobic BSA solution resulted in nitrite formation, which accumulated much slower than in the control, without BSA (Fig. 1C). This result suggests that a significant proportion of N₂O₃ was trapped by albumin and reacted with water only upon slow release. Addition of proteinase K to the BSA probe shortly after it was treated with NO^⋅ caused an immediate nitrite accumulation, almost to the plateau level (Fig. 1D), indicating that most of the trapped N₂O₃ was abruptly released into the aqueous phase. As judged by SDS/PAGE analysis, more than 99% of BSA was degraded within 1 min of proteolysis (not shown) thus eliminating most, if not all, protein hydrophobic pockets. The burst of nitrite in response to the protease declined with time so that after 40 min no detectable nitrite accumulation was observed upon protease challenge. At this time, the overall level of nitrite had already reached its plateau, suggesting that no N₂O₃ remained inside the protein. Based on these data, we estimate t_1/2 (N₂O₃) in BSA to be ≈7 min (Fig. 1D).

Such a dramatic stabilization of N₂O₃ by albumin may be due to its potentially high Q value. Additionally, slow diffusion N₂O₃ within the albumin interior can be attributed to formation of π and σ complexes with aromatic amino acid residues and nucleophiles.

Addition of Cu⁺⁺ to 1.5 mM, which induces cleavage of S-NO bond, to BSA 40 min after it was treated with NO^⋅ results in ≈800 nM NO^⋅ release as detected electrochemically (see Experimental Procedures), indicating that 5–6% of BSA was S-nitrosated during the experiment. The only free thiol of albumin, Cys-34, and Trp-214 were nitrosated as confirmed by spectroscopic measurements (see Fig. 3B). Other amino acid residues undergo nitrosation only to a small extent under these conditions (19, 28). Indeed, NO/H⁺, which is a more aggressive nitrosating agent than NO^⋅, at 1:1 ratio to BSA generated 0.6 nitrosating amino acid residues per BSA molecule (28). Thus, even theoretically, if all NO^⋅ would act as NO/H⁺ (in our experiment it was not the case as the pH was neutral), we could expect no more than one to two nitrosated residues per BSA molecule (most likely Cys and Trp, because they are the strongest nucleophiles).

Subsequently, we will refer to albumin saturated with NO^⋅/N₂O₃ as BSA^N₂O₃ and the fraction that underwent nitrosation as NO-BSA^N₂O₃ or NO-BSA.

Albumin Is a Catalyst of LMW RS-NO Formation.

S- and N-nitrosation of albumin with NO^⋅ is much more efficient than that of LMW thiols or amines (19). This phenomenon is readily explained by the ability of albumin to catalyze and preserve nitrosating species, N₂O₃, in its own hydrophobic compartments (ref. 19; Fig. 1). We reasoned that albumin may also be the micellar catalyst of nitrosation of external molecules such as LMW RSH. To test this hypothesis, we compared the rate of LMW RS-NO formation in the presence or absence of albumin. We chose two bioactive LMW RSH: GSH and l-cysteine (Cys) (Fig. 2). Nitrosothiols were detected by employing Cu⁺⁺, followed by electrochemical detection of released NO^⋅ (see Experimental Procedures). Each experiment began with generation of BSA^N₂O₃, that is saturation of BSA with NO^⋅ under aerobic conditions by using water solution of NO^⋅ (100 μM). As soon as the NO^⋅ concentration in the BSA solution or control probe dropped from 100 μM to ≈1 μM, GSH or Cys were added. In each case, the presence of BSA had a great stimulating effect on the rate and efficiency of LMW RS-NO formation (Fig. 2 A and B).

Figure 2.

Generation of LMW RS-NO by albumin^N₂O₃. (A) BSA^N₂O₃-mediated Cys-NO formation as a function of Cys concentration. (Left) A representative NO^⋅-electrode tracing of Cu⁺⁺-dependant NO^⋅ production after addition of 1 mM Cys to 17 μM BSA^N₂O₃ (“BSA” trace) or just Tris⋅HCl buffer (80 mM, pH 7.9) (“buffer” trace). Cu⁺⁺ was added as the NO^⋅ concentration in the probe reached 1 μM after initial 100 μM. (B) BSA^N₂O₃-mediated GS-NO formation as a function of GSH concentration. (Left) Experiment is the same as in A, except that 2 mM GSH was used instead of Cys. (C) BSA^N₂O₃-mediated GS-NO formation as a function of BSA concentration (experimental design as in B). (D) BSA^N₂O₃-mediated GS-NO formation as a function of time (experimental design as in B, except that 30 μM GSH was added after indicated time intervals). (E) Effect of proteolysis on BSA^N₂O₃-mediated GS-NO formation [experimental design as in B, except that proteinase K (0.25 mg/ml) was added before the addition of NO^⋅ to BSA (Left), or after that, when the NO^⋅ concentration dropped to ≈1 μM (Right)].

The ability of BSA^N₂O₃ to generate GS-NO declined with time (Fig. 2D). The rate of decline (t_{1/2(activity)} ∼ 10 min) was close to what one can expect from the estimated “half-life” of N₂O₃ in albumin (Fig. 1D).

The efficiency of albumin-mediated S-nitrosation also depended on the amount of albumin in the probe. The maximum effect was observed at ≈25 μM BSA (Fig. 2C). This fact is readily explained in terms of micellar catalysis of NO^⋅ oxidation (24) when the relatively small “optimal” volume of the hydrophobic phase (in this case albumin) generates the maximum acceleration of the reaction.

Mechanism of Albumin-Mediated Catalysis of LMW RS-NO Formation.

To confirm that micellar catalysis of ^⋅NO oxidation is responsible for observed results, we studied the effect of proteinase K on albumin-mediated GSH nitrosation. As described above, proteinase K rapidly converts albumin into short peptides incapable of forming appropriate micelles. Addition of this protease to BSA^N₂O₃ causes an immediate release of N₂O₃ into the aqueous phase, resulting in a burst of nitrite formation (Fig. 1D). Fig. 2E demonstrates that protease-treated BSA was unable to potentiate GS-NO formation, apparently because of its inability to accumulate ^⋅NO/N₂O₃. On the other hand, if the protease was added to BSA^N₂O₃, the abrupt accumulation of GS-NO occurred (Fig. 2E), indicating that the large pool of NO⁺ became available for GSH upon proteolysis. GSH itself is resistant to proteinase K. These results directly implicate micellar catalysis of NO^⋅ oxidation as a mechanism for albumin-mediated LMW RSH nitrosation.

NO⁺ originated from N₂O₃ in the albumin hydrophobic core can be transferred to LMW RSH via at least three possible pathways (Fig. 3A). Direct attack on RSH is possible if RSH enters the hydrophobic interior of BSA (pathway I); alternatively, NO⁺ can reach RSH outside the protein hydrophobic core through the intermolecular transport to Cys-34 or Trp-214 followed by transnitrosation (pathways II and III). To determine which pathway plays the most important role in formation of LMW RS-NO, we prepared chemically modified BSA, where either Cys-34 or Trp-214 or both were covalently blocked and unable to undergo S- or N-nitrosation as confirmed by UV-Vis spectroscopy (Fig. 3B). Corresponding derivatives are designated as ^CBSA, ^WBSA, and ^CWBSA. We next compared the ability of BSA^N₂O₃, ^CBSA^N₂O₃, ^WBSA^N₂O₃, and ^CWBSA^N₂O₃ to stimulate nitrosation of RSH. Covalent modification of Cys-34 or Trp-214 or both should exclude pathway II and/or III from LMW RSH nitrosation. The experiment with ^CWBSA^N₂O₃ shows that this derivative potentiates formation of GS-NO with ≈35% efficiency that of native BSA^N₂O₃, implying that pathway I plays a substantial role in LMW RSH nitrosation (Fig. 3C). ^CBSA^N₂O₃ and ^WBSA^N₂O₃ exhibit ≈60% and ≈70% the activity of intact BSA^N₂O₃, respectively (Fig. 3C), suggesting that these pathways contribute roughly 40% and 30% to the full effect of intact BSA^N₂O₃. Notably, at lower, close to physiological GSH concentrations, the efficiency of S-nitrosation mediated by BSA^N₂O₃ and ^CWBSA^N₂O₃ was virtually the same, suggesting that pathway I may be predominant in vivo.

Albumin as a NO^⋅-Sink in Vivo.

To evaluate the ability of hydrophobic compartments of albumin to absorb NO^⋅ in vivo and the physiological role of this phenomenon, we studied the effect of exogenous albumin and albumin^N₂O₃ on MAP in rats. To exclude the potential hemodynamic effect of S- and N-nitroso-derivatives of BSA, we used ^CWBSA. Intravenous infusion of ^CWBSA (≈0.12 g/kg body weight; see Experimental Procedures) was characterized by extensive hypertension (Fig. 4A), whereas the administration of the same amount of ^CWBSA^N₂O₃ had only a slight effect on MAP (Fig. 4B). Moreover, the same amount of ^CWBSA^N₂O₃ induced a substantial decrease of MAP if it was administered after animals were i.v. pretreated with the inhibitor of NO^⋅ synthesis, N-nitro-l-arginine methyl ester (l-NAME) (Fig. 4B). Experiments with native BSA and BSA^N₂O₃ gave similar results (Fig. 4). These observations are consistent with the model (Fig. 4C) in which the hydrophobic phase of plasma protein serves as a NO^⋅/N₂O₃ reservoir. Exogenous albumin, which has not been saturated with NO^⋅, rapidly absorbs free circulating NO^⋅, thus causing the MAP increase. MAP eventually recovers because newly synthesized NO^⋅ apparently restores the equilibrium between the pool of free NO^⋅ and that solubilized by plasma protein. On the other hand, exogenous ^CWBSA^N₂O₃ or BSA^N₂O₃ did not change MAP because they have already been saturated with N₂O₃ so that the pool of NO^⋅/NO⁺ in blood was not significantly perturbed.

Figure 4.

Effect of albumin and albumin^N₂O₃ on MAP of anaesthetized rats. (A) Sample tracing (top) and time course (bottom) of the hemodynamic effect of BSA (light gray column) or ^CWBSA (black column). Infusion of 0.12 g/kg albumin produced a steep increase in blood pressure followed by a gradual and prolong asymptotic recovery. Mean ± SE from ten independent experiments (see Experimental Procedures). (B) Sample tracings (top) and time course (bottom) of the hemodynamic effect of ^CWBSA^N₂O₃ administered after infusion of l-NAME (light gray column) or without l-NAME (control). ^CWBSA^N₂O₃ was prepared as in Fig. 2B and infused at 0.12 g/kg. l-NAME (20 mg/kg) causes prolonged hypertension (compare left and right tracings). In this case, ^CWBSA^N₂O₃ induces a rapid fall in MAP followed by a gradual asymptotic return to the l-NAME-dependent base line (light gray column). If saline is used instead of l-NAME (dark gray column), ^CWBSA^N₂O₃ does not change MAP significantly. Mean ± SE from ten independent experiments. (C) Schematics explaining hemodynamic effects of albumin and albumin^N₂O₃. Exogenous albumin (open circle) rapidly absorbs free NO^⋅ thus causing vasoconstriction (left drawing). On saturation with NO^⋅, albumin generates NO⁺ via micellar catalysis of NO^⋅ oxidation (filled circles). NO⁺ then acts as a vasodilator to recover MAP. Exogenous albumin^N₂O₃ does not change the NO^⋅/NO⁺ balance and thus has no hemodynamic effect (middle drawing). Inhibition of NOS by l-NAME exhausts plasma albumin with NO⁺ causing vasoconstriction (right drawing). In this case, exogenous albumin^N₂O₃ works as a vasodilator by supplying NO⁺.

It should be noted that the solution of albumin that we used was isooncotic. Also, the amount of exogenous albumin was only 1% of the total plasma albumin and the rate of its infusion was slow (0.2 ml/min). This rules out the possibility that exogenous albumin affected MAP by changing the osmotic status.

Micellar Catalysis of LMW RS-NO Formation in Vivo and Regulation of Blood Pressure.

Our in vitro and in vivo results with albumin suggest that plasma protein absorbs NO^⋅ under normal conditions in vivo because of its high Q_NO. Since free O₂ is dissolved in mammalian blood in micromolar concentration, one can expect that because of micellar catalysis, a large portion of circulating albumin exists as albumin^N₂O₃. Such albumin is also expected to serve as a catalyst of LMW RS-NO formation in vivo. To address this issue directly, we first tested rat blood plasma for its ability to potentiate GS-NO formation. An experiment similar to that described in Fig. 2B was performed in which freshly prepared rat plasma was used instead of BSA^N₂O₃ (Fig. 5A). Although we did not add any exogenous NO^⋅ to plasma, a significant amount of GS-NO was formed on addition of GSH. The amount of newly formed GS-NO exceeded the amount of preexisting RS-NO in plasma by more than 2-fold, suggesting that much of the newly formed GS-NO was not a result of transnitrosation. Furthermore, if plasma was preincubated for 60 min before addition of GSH, two times less GS-NO was formed in comparison with the fresh plasma situation (Fig. 5A). A similar rate of time-dependent decline in the ability to induce S-nitrosation was observed in vitro with pure albumin^N₂O₃ (Fig. 2D). These results strongly suggest that plasma is indeed saturated with NO^⋅/N₂O₃ and serves as a catalyst of LMW RS-NO formation in vivo.

To provide further support for this conclusion, we examined the effect of GSH on LMW RS-NO formation and vasodilation in rats. In vitro, the yield of LMW RS-NO in the presence of albumin^N₂O₃ or fresh plasma directly depends on the concentration of LMW RSH (Fig. 2B). To test whether a similar relationship occurs in vivo, we measured the level of circulating LMW RS-NO and MAP in parallel after i.v. administration of various amounts of GSH to rat. As Fig. 5C demonstrates, infusion of 0.4 g/kg GSH resulted in a more than 4-fold increase in the amount of LMW RS-NO in blood. It was accompanied by a substantial decrease of MAP, which then remained at the same low level for more than an hour.

Large doses of GSH were administered to compensate for the highly efficient import of GSH by various tissues (e.g., refs. 29 and 30). Concentrations shown in Fig. 5C (5, 12, and 20 mM) are estimated to be initial concentrations in blood on GSH injection. After that, plasma GSH decreased exponentially (Fig. 5C) so that at the moment of data collection GSH was detected within a micromolar range, yet significantly above the basal level (Fig. 5C). Effect of GSH on MAP and formation of RS-NO was dose-dependant and correlated well with the dose-dependent curve of GS-NO formation in vitro (Fig. 2B). These results further suggest that the micellar catalytic mechanism of S-nitrosylation operates in mammalian circulation and is involved in control of vascular tone.

Discussion

The present study demonstrates an important role of the hydrophobic phase formed by plasma protein in the biochemistry of NO^⋅. A soluble protein molecule, such as albumin, is envisioned as a micelle that is capable of concentrating nonpolar molecules of NO^⋅ and O₂ in a small volume of its hydrophobic compartments. We estimate that the concentration of NO^⋅ inside the albumin globule must be ≈120 times higher than in the surrounding aqueous phase (Q_NO ∼ 120). Even supposing that Q_O2 for albumin is only 1, our results suggest that the acceleration of NO^⋅ oxidation due to the increased local concentration of NO^⋅ by albumin is several thousand-fold. Furthermore, reactive products of this reaction, most likely N₂O₃, appear to be much more stable within the protein interior than in the surrounding aqueous phase (Fig. 1 C and D). Both these features, the micellar catalysis of NO_x formation and preservation of this otherwise short-lived nitrosating agent, make albumin a potent catalyst of nitrosylation of its own Cys and Trp residues (19), as well as of external thiols (Figs. 2 and 3).

The micellar catalytic mechanism of albumin-mediated NO^⋅ oxidation may resolve an apparent paradox between the substantial amount of S-nitrosothiols observed in vivo (5–10) and limited concentration, life span, and reactivity of circulating NO^⋅. Reaction of NO^⋅ with O₂ is third order (k ∼ 4 × 10⁶ M⁻²⋅sec⁻¹) (20, 21), and under normal noninflammatory conditions, formation of nitrosating species, such as N₂O₃, would be too slow to account for any traceable amount of S-nitrosothiols (31). However, in this study we detect the basal level of RS-NO in rat plasma ranging from 50 nM to ≈1 μM, which is consistent with earlier related estimates (5–10). The actual amount of circulating RS-NO is apparently higher, because some LMW RS-NO—e.g., S-nitroso-Cys and S-nitroso-α-lipoic acid—are highly unstable and at least partially decompose while blood samples were processed (O.R. and E.N., unpublished observations).

Our in vitro and in vivo results correlate well with each other and support the model, in which plasma protein, predominantly albumin, is normally saturated with N₂O₃ (NO⁺-NO) and other NO_x species in vivo and transfers NO⁺ to LMW RSH. We show that the level of circulating RS-NO can be increased severalfold in response to iv infusion of GSH (Fig. 5). A reasonable explanation for this result is that endogenous and exogenous GSH acquire NO⁺ not only via transnitrosation from albumin S-NO-Cys-34, which is limited to a basal level of detected RS-NO, but also from albumin N-NO-Trp-214 (ref. 32; Fig. 3) and directly from the pool of NO_x in the protein interior (Fig. 3). This conclusion is further supported by our experiment in which fresh rat plasma is shown to possess a strong nitrosating activity toward exogenous GSH (Fig. 5A). Freshly prepared plasma was able to produce GS-NO to the level that is significantly higher than that of preexisting RS-NO in plasma, indicating that there was a larger source of NO⁺ in vivo unrelated to RS-NO. The nitrosation activity of fresh plasma lasts for about as long as that of albumin^N₂O₃ (Fig. 2D) and correlates well with the estimated half-life of N₂O₃ in the albumin interior (Fig. 1D). Additionally, Q_NO for plasma and pure albumin are virtually identical (Fig. 1B). Taken together, these data argue that much of NO⁺ in circulation originates directly from N₂O₃ that accumulates in hydrophobic compartments of albumin.

Another major source of NO⁺ in blood is S-nitroso-hemoglobin (SNO-Hb) (8). Like albumin, Hb is a globular protein, implying that its hydrophobic core is likely to be a micellar catalyst of N₂O₃ formation and S-nitrosylation. Unlike albumin^N₂O₃, which rapidly and directly transfers NO⁺ to plasma thiols, SNO-Hb employs a relatively slow system, which relies on the membrane anion-exchange protein 1, to “pump” its NO⁺ out of red blood cells (33). Additionally, the ability of Hb to exchange its NO⁺ is highly sensitive to pO₂ (34). Thus, Hb and albumin may represent two major generators and transporters of NO⁺ but for different functioning. Whereas NO-derived vasoactivity of Hb matches ventilation to perfusion in lungs (35), such an activity of albumin may be of a more general nature in that it continuously adjusts the vascular response to rapidly changing NO^⋅ concentrations, causing the systemic blood pressure to change smoothly.

Formation of vasoactive nitrosothiols in response to i.v. infusion of GSH apparently explains dose-dependent hypotention induced by GSH (Fig. 4). This result suggests that one can regulate the level of circulating RS-NO and vascular tone simply by administering LMW RSH. We believe that this approach may have pharmacological implications. Because nitrosothiols are potent vasodilators and antiplatelet agents, they are considered promising therapeutics for various cardiovascular complications (36, 37). However, one major drawback of nitrosothiols as a commercial drug is their instability in physiological vehicles. Use of stable and nontoxic RSH, even at higher concentrations, to achieve the same effect as that of corresponding RS-NO seems to be advantageous.

Another approach to control blood pressure can be envisioned based on our studies. If hydrophobic compartments of albumin serve as a powerful NO^⋅ sink and reservoir for NO_x, one can control vascular tone by changing the volume of circulating hydrophobic phase—e.g., filling of albumin with NO^⋅/NO_x. Indeed, “empty” albumin (1% of total rat plasma albumin) administered i.v. to rats works as a vasoconstrictor, apparently because of its ability to sequester free circulating NO^⋅, whereas “full” albumin was neutral in this sense (Fig. 4). In those experiments, chemically modified albumin, which could not undergo S- and N-nitrosation, produced the same hemodynamic response as that of unmodified albumin, arguing that Q-driving NO^⋅ adsorption, rather than transnitrosation, had the major impact on vascular tone in this case.

Apart from albumin, there are many natural and synthetic compounds that can form appropriate micelles in blood with high Q for NO^⋅ and thus potentially serve as hemodynamic regulators. Exploring these as a tool to manipulate NO^⋅ bioactivity should open a way for conceptionally new and exiting therapies in vascular pathophysiology.

Abbreviations

GSH

glutathione

RSH

thiols

LMW

low-molecular-weight

MAP

mean arterial pressure