Optimized S-nitrosohemoglobin Synthesis in Red Blood Cells to Preserve Hypoxic Vasodilation Via β Cys93

Abstract

Classic physiology links tissue hypoxia to oxygen delivery through control of microvascular blood flow (autoregulation of blood flow). Hemoglobin (Hb) serves both as the source of oxygen and the mediator of microvascular blood flow through its ability to release vasodilatory S-nitrosothiol (SNO) in proportion to degree of hypoxia. β-globin Cys93Ala (βCys93Ala) mutant mice deficient in S-nitrosohemoglobin (SNO-Hb) show profound deficits in microvascular blood flow and tissue oxygenation that recapitulate microcirculatory dysfunction in multiple clinical conditions. However, the means to replete SNO in mouse red blood cells (RBCs) to restore RBC function is not known. In particular, although methods have been developed to selectively S-nitrosylate βCys93 in human Hb and intact human RBCs, conditions have not been optimized for mouse RBCs that are used experimentally. Here we show that loading SNO onto Hb in mouse RBC lysates can be achieved with high stoichiometry and β-globin selectivity. However, S-nitrosylation of Hb within intact mouse RBCs is ineffective under conditions that work well with human RBCs, and levels of metHb are prohibitively high. We developed an optimized method that loads SNO in mouse RBCs to maintain vasodilation under hypoxia and shows that loss of SNO loading in βCys93Ala mutant RBCs results in reduced vasodilation. We also demonstrate that differences in SNO/met/nitrosyl Hb stoichiometry can account for differences in RBC function among studies. RBCs loaded with quasi-physiologic amounts of SNO-Hb will produce vasodilation proportionate to hypoxia, whereas RBCs loaded with higher amounts lose allosteric regulation, thus inducing vasodilation at both high and low oxygen level. SIGNIFICANCE STATEMENT: Red blood cells from mice exhibit poor hemoglobin S-nitrosylation under conditions used for human RBCs, frustrating tests of vasodilatory activity. Using an optimized S-nitrosylation protocol, mouse RBCs exhibit hypoxic vasodilation that is significantly reduced in hemoglobin βCys93Ala mutant RBCs that cannot carry S-nitrosothiol allosterically, providing genetic validation for the role of βCys93 in oxygen delivery.

Fig. 1.

Stoichiometry of SNO-Hb synthesis using CysNO in human RBC lysates. (A) Quantitation of SNO-Hb. The Saville assay (1.7 ± 0.5 SNO/tetramer; n = 3) and photolysis/chemiluminescence (1.6 ± 0.2 SNO/tetramer; n = 3) yield equivalent results for S-nitrosylated Hb, consistent with SNO residing primarily at βCys93. No significant difference by t test. (B) Synthetic SNO-Hb contains 13% ± 10% metHb (n = 4). (C) The SNO-Hb concentration determined by the Saville assay was used to prepare SNO-Hb standards for comparison with GSNO standards. Both SNO-Hb and GSNO standard curves show equivalent photolysis signal peak areas over a range of concentrations.

Fig. 2.

SNO and heme-NO quantification in human SNO-Hb[Fe]NO. (A) Absorbance scans of serial dilutions of CysNO-treated Hb, representative of 3 assays. (B) Overlay of photolysis signals from the serially diluted Hb samples shown in panel (A). Each trace shows sequential injections of a GSNO standard, followed by SNO-Hb, and SNO-Hb in the presence of 1 mM HgCl₂ (= FeNO concentration). Representative of 3 assays. (C) The calculated Hb tetramer concentration from assays as in panel (A) plotted against the calculated total NO concentration (XNO; SNO+FeNO) from panel (B). SNO-Hb is calculated as the difference between the untreated sample (total NO) and HgCl₂-treated sample (FeNO). SNO-Hb is 80% ± 4% of total Hb-NO at all tested dilutions. n = 3; SNO-Hb and FeNO-Hb correlations p < 0.0001; SNO versus Hb, r² = 0.8832; FeNO versus Hb, r² = 0.8765.

Fig. 3.

SNO loaded on hemoglobin is primarily on Cys93 in humanized-mouse RBC lysates. (A) Representative photolysis signals (of n = 4) of a GSNO standard, and synthetic SNO-Hb in control (C93) and mutant (C93A) RBC lysates ± HgCl₂. (B) Stoichiometry of SNO and FeNO (combined, XNO) per Hb tetramer in control (C93) and βC93A mutant Hb after treatment (n = 4; *, p < 0.05 by t test). (C) Heme concentrations used in assays, and (D) metHb formation after treatment (n = 4; no significant differences by t test).

Fig. 4.

S-nitrosylation of hemoglobin in intact human RBCs and humanized-mouse (C93) RBCs. All treatments were for 5 minutes with EtCysNO in PBS, pH 7.8. (A) Treatment of a 50% RBC suspension with 50 µM EtCysNO shows increased S-nitrosylation in human RBCs (n = 6), but not in humanized-mouse RBCs under the same conditions (n = 6; *, p < 0.05 by t test). (B) Dilution of RBCs from 50% to 6.25% improves SNO yield in human RBCs treated with 50 µM EtCysNO (n = 4), and also in humanized-mouse RBCs but to a lesser degree (n = 6; *, p < 0.05 by t test).

Fig. 5.

Optimizing EtCysNO concentration for S-nitrosylation in humanized-mouse RBCs. RBCs at 6.25% dilution in PBS, pH 7.8 were treated for 5 minutes with the indicated concentrations of EtCysNO (n = 6–14). (A) SNO-Hb formation increased with increasing EtCysNO concentration, but the increase was significantly higher in C93 control than in C93A mutant RBCs at every EtCysNO concentration. *, p < 0.05 by t test. (B) metHb formation increased with increasing EtCysNO concentration, but there was no significant difference between C93 and C93A RBCs at any concentration (by t test).

Fig. 6.

SNO-loaded RBCs induce hypoxia-specific vasodilation, but SNO overloading promotes high metHb formation and loss of hypoxia-specific vasodilation. (A) SNO-Hb formation increases with increasing EtCysNO concentration. (B) MetHb formation increases with increasing EtCysNO concentration, and is higher in mouse than in human RBCs. (C) SNO-loaded RBCs promote vasodilation under hypoxia (1% O₂). Human RBCs produce significantly more vasodilation than mouse RBCs at every EtCysNO treatment dose, and the vasodilation effect plateaus at 200 μM EtCysNO. (D) High SNO loading mediates loss of vasoconstriction by RBCs in room air and induces aberrant vasorelaxation by human and mouse RBCs under normoxia (20% O₂). For all panels, n = 4, except for mouse 2000, 5000 μM-treated RBCs (SNO-Hb and metHb) where n = 2; *, p < 0.05 by t test.

Fig. 7.

SNO-loaded C93 and C93A RBCs differ in hypoxic vasodilation. (A) SNO-Hb formation increases with increasing EtCysNO concentration in C93 and C93A RBCs but is less in C93A RBCs. (B) MetHb formation increases similarly with increasing EtCysNO concentration in both C93 and C93A RBCs, but is excessive above 200 μM. (C) Optimally SNO-loaded RBCs (200 μM ECysNO) promote vasodilation under hypoxia (1% O₂) in proportion to their amount of SNO-Hb. C93A mutant RBCs generate significantly less vasodilation than do C93 RBCs, consistent with lower amounts of SNO. *, p < 0.05 by t test. †, p < 0.05 by one-way ANOVA across doses, with post hoc Šidák multiple comparison test.