Hemoglobin is an oxygen-dependent glutathione buffer adapting the intracellular reduced glutathione levels to oxygen availability

Abstract

Fast changes in environmental oxygen availability translate into shifts in mitochondrial free radical production. An increase in intraerythrocytic reduced glutathione (GSH) during deoxygenation would support the detoxification of exogenous oxidants released into the circulation from hypoxic peripheral tissues. Although reported, the mechanism behind this acute oxygen-dependent regulation of GSH in red blood cells remains unknown. This study explores the role of hemoglobin (Hb) in the oxygen-dependent modulation of GSH levels in red blood cells. We have demonstrated that a decrease in Hb O₂ saturation to 50% or less observed in healthy humans while at high altitude, or in red blood cell suspensions results in rising of the intraerythrocytic GSH level that is proportional to the reduction in Hb O₂ saturation. This effect was not caused by the stimulation of GSH de novo synthesis or its release during deglutathionylation of Hb's cysteines. Using isothermal titration calorimetry and in silico modeling, we observed the non-covalent binding of four molecules of GSH to oxy-Hb and the release of two of them upon deoxygenation. Localization of the GSH binding sites within the Hb molecule was identified. Oxygen-dependent binding of GSH to oxy-Hb and its release upon deoxygenation occurred reciprocally to the binding and release of 2,3-bisphosphoglycerate. Furthermore, noncovalent binding of GSH to Hb moderately increased Hb oxygen affinity. Taken together, our findings have identified an adaptive mechanism by which red blood cells may provide an advanced antioxidant defense to respond to oxidative challenges immediately upon deoxygenation.

Graphical abstract

Fig. 1

The impact of high altitude on the selected blood parameters and intraerythrocytic redox state (A) Study design with days of blood sampling. numbers in the arrows are the days of the study counted from ascent to the HA. Colors for the time of blood sampling (red for basal level, blue for HA exposure and grey for the time after descent from HA) are used in all the other plots. (B) Hemoglobin levels, and (C) SO₂ measured by ABL825 FLEX, Radiometer. (D) intracellular GSH and (E) GSSG levels measured by Ellman's reagent. (F) Half-cell redox potential (E_hc) for GSH/GSSG couple. Details may be found in Materials and methods section and in Ref. [21] (G) Association between SO₂ and GSH levels. Blue line is a fit using the method of loess (Local Polynomial Regression Fitting), the grey area shows the confidence interval (95%). Statistics: N = 12, ANOVA with custom contrast: * HA different from pre/sea level; ‡ different HA03 and HA18; † different pre and sea level. HA: High altitude (3500 m), SL: sea level (110 m), GSH: reduced glutathione, GSSG: oxidized glutathione, SO₂: hemoglobin oxygen saturation. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Recycling from GSSG and de novo synthesis do not explain the increase in GSH levels upon deoxygenation. (A) Intraerythrocytic GSH levels as a function of SO₂ in RBC suspension (N = 7). Hemoglobin was pre-equilibrated with 15% O₂ for 30 min and then deoxygenated gradually by switching to 0.5% O₂ for 2 h. Grey curve is for untreated control, and pink curve is for RBC pretreated for 20 min with the inhibitor of de novo GSH synthesis BSO (l-Buthionine-sulfoximine, 1 mM). (B) GSSG levels in RBCs as a function of SO₂ (N = 9). Abbreviations identical toFig. 1. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

The impact of deoxygenation on redox balance in RBCs ex vivo and in vivo. The effect of incubation of RBC suspensions at 1% O₂ (hypoxia) for 3h on: (A) GSH levels measured by Ellman's reagent (N = 6, paired Student's t-test) (B) GSH quantification by ¹H NMR. Region of reference ¹H spectrum for the standard GSH solution corresponding to Cys β-protons (upper panel), and in the same region in spectra of deproteinized RBC samples obtained by consecutive addition of standard GSH solution aliquots (lower panel). Both proton signals are of dd shape (14.2Hz and 5.4Hz; 14.2Hz and 6.7 Hz), but, as the spectrum is not one of the first order, only the central components comprising the most of signal intensity (range between 2.93 and 2.87 ppm) were used for integration, to minimize the contribution of noisy baseline. (C) Quantification of GSH in the samples obtained from oxygenated and deoxygenated RBC by means of ¹H NMR normalized per Hb levels (N = 4, paired Student's t-test) (D) Bulk reduced thiols detected using MBBR staining by means of flow cytometry (N = 6, paired Student's t-test). (E) ROS levels detected using DHR123 (N = 5 paired Student's t-test) and (F) N₂O₃ levels measured using DAF-DA (N = 6, paired Student's t-test). (G) Bulk thiol levels detected using MBBR fluorescence recorded by flow cytometry, and (H) metHb measured using ABL825 FLEX, Radiometer in RBCs of 12 participants of the high altitude exposures study For more details on the HA study design see Fig1A. N = 12, ANOVA with custom contrast: * HA different from pre/sea level; ‡ different HA03 and HA18; † different pre and sea level. MBBR: monobrombimane, NO: nitric oxide, ROS: reactive oxygen species, DHR: dihydrorhodamine, DAF-DA: diaminofluoresceine diacetate.

Fig. 4

Discrimination between oxygen-dependent covalent interaction of GSH with Hb (Hb S-glutathionylation) and non-covalent docking of GSH to Hb: (A) In silico modeling of the possible changes in availability of Cys residues αCys104, βCys93, and βCys112 for S-glutathionylation. Shown is a superposition of oxy-Hb(blue) and deoxy-Hb(red) structures with a blow-up of the areas where Cys are localized. While thiol group of βCys93 is facing outwards in the deoxy-Hb, it is turned inwards into the Hb molecule in the oxy-Hb state. Position of the thiol groups of αCys104 and βCys112 is not altered by oxygenation-deoxygenation. (B) S-glutathionylation of Hb in RBC suspension that were gradually deoxygenated in a tonometer in the atmosphere of 100% N₂. Aliquots were taken after 2, 5, 10, and 20 min and SO₂ shown above the plot was measured by ABL825 FLEX, Radiometer. RBC were lysed in non-reducing lysis buffer supplemented with NEM (25 mM) and S-glutathionylation of Hb detected using specific antibodies and normalized to the signal for β-globin as a loading control. Representative blots obtained for one of the RBC samples is shown at the upper panel. Densitometry for the S-glutathionylated form of Hb monomers normalized to that for the β-globin is shown at the lower panel. Paired Wilcoxon test was used for analysis. N = 8. (C) Similar approach was used to assess S-glutathionylation of Hb in blood samples of participants from the HA study. Representative immunoblot is shown for one participant before, during and after the staying at HA. Statistics: ANOVA with custom contrast * different from pre/sea level and high altitude. N = 12 (D) Schematic representation of NEM experiment. GSH levels were detected in control untreated oxygenated RBCs that were sham-washed and those oxygenated cells that were pre-treated with 20 mM NEM for 20 min at 21% O₂ and the washed free from the NEM that did not interact with the targets (red circles). Thereafter, both NEM-treated and control non-treated RBCs were deoxygenated with 100% N₂ for 30 min (blue circles) and GSH detected once again. Further explanations to the scheme may be found in the text. GSH (E) and GSSG (F) were detected using Ellmann's reagent and metHb levels (G) was measured by ABL825 FLEX, Radiometer. Paired Wilcoxon test. N = 7. NEM: N-ethyl malemide. RBC: red blood cell, DTT: 1,4-Dithiothreitol, HBB: Hemoglobin subunit beta, GSH: reduced and GSSG: oxidized Glutathione. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Noncovalent binding of GSH to Hb is O₂-dependent: ITC titration curve (upper panel) and binding isotherm (lower panel) for GSH interaction with (A) oxy-Hb (21% O₂) and at (B) deoxy-Hb (1% O₂) at 25 °C. (C) Thermodynamic parameters of GSH binding to oxy-Hb and deoxy-Hb determined by isothermal titration calorimetry. Shown are the number of binding sites for GSH per 1 molecule Hb (N), and association and dissociation equilibrium constants K_a and K_d, and the thermodynamic parameters of Hb:GSH interaction ΔH, -TΔS and ΔG. (D) Best Hb:GSH docking models produced by Vina Autodock docking. Upper panel shows the front and side views for the best four affinity GSH binding sites in the oxy-Hb, and lower panel shows the same for the deoxy-Hb. Beta chains are shown in brown, and the alpha chains are in blue. Shown in pink are GSH molecules occupying the sites 1–4 and in oxy-Hb. The GSH molecules in green interact with the sites 3 and 4 in deoxy-Hb. ΔH: the changes in enthalpy, -TΔS: the changes in entropy, and ΔG: the changes in Gibbs energy. ITC: isothermal titration calorimetry. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

The interplay between GSH and BPG binding to Hb and regulation of O₂affinity of Hb. The experimental design to assess the possible competition between BPG and GSH for binding to Hb (A) and the changes in free GSH in hemolysates that were produced by deoxygenation in pure N₂, reoxygenation with air, supplementation of 2 mM BPG followed up by repeated deoxygenation (B). Dashed lines stand for the basal GSH levels in hypoxic and normoxic hemolysates. Red and blue bars stand for the free GSH levels in hypoxic BPG-treated and control hemolysates respectively. Grey and red bars show free GSH levels in reoxygenated control and BPG-treated hemolysates. N = 3. (C) Models showing interactions of GSH with BPG-Hb complex (front and side views). Alpha chains are shown in light grey, and β chains are in dark grey. GSH molecules are highlighted in red, and BPG molecules are in green. (D) ITC titration curve (upper panel) and binding isotherm (lower panel) for GSH binding to deoxy-Hb:BPG complex (E) interaction of BPG with deoxy-Hb in the absence or (F) in the presence of GSH at 25 °C. Thermodynamic parameters of these interactions determined by isothermal titration calorimetry are shown in Table S3 (G)P50 values obtained for hemolysates in the absence or presence of 5 mM GSH measured by Hemox analyser. (H) Schematic representation of a possible mechanism of changes in docking partners for Hb in RBCs during deoxygenation-reoxygenation (for details, see the text). BPG: 2,3-bisphosphoglycerate, P50: the oxygen tension when hemoglobin is 50% saturated with oxygen. ΔH: the changes in enthalpy, -TΔS: the changes in entropy, and ΔG: the changes in Gibbs energy. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)