Measurements of nitric oxide on the heme iron and β-93 thiol of human hemoglobin during cycles of oxygenation and deoxygenation

Abstract

Nitric oxide has been proposed to be transported by hemoglobin as a third respiratory gas and to elicit vasodilation by an oxygen-linked (allosteric) mechanism. For hemoglobin to transport nitric oxide bioactivity it must capture nitric oxide as iron nitrosyl hemoglobin rather than destroy it by dioxygenation. Once bound to the heme iron, nitric oxide has been reported to migrate reversibly from the heme group of hemoglobin to the β-93 cysteinyl residue, in response to an oxygen saturation-dependent conformational change, to form an S-nitrosothiol. However, such a transfer requires redox chemistry with oxidation of the nitric oxide or β-93 cysteinyl residue. In this article, we examine the ability of nitric oxide to undergo this intramolecular transfer by cycling human hemoglobin between oxygenated and deoxygenated states. Under various conditions, we found no evidence for intramolecular transfer of nitric oxide from either cysteine to heme or heme to cysteine. In addition, we observed that contaminating nitrite can lead to formation of iron nitrosyl hemoglobin in deoxygenated hemoglobin preparations and a radical in oxygenated hemoglobin preparations. Using ¹⁵N-labeled nitrite, we clearly demonstrate that nitrite chemistry could explain previously reported results that suggested apparent nitric oxide cycling from heme to thiol. Consistent with our results from these experiments conducted in vitro, we found no arterial/venous gradient of iron nitrosyl hemoglobin detectable by electron paramagnetic resonance spectroscopy. Our results do not support a role for allosterically controlled intramolecular transfer of nitric oxide in hemoglobin as a function of oxygen saturation.

How is the activity of the endothelium-derived relaxation factor, nitric oxide (NO), preserved if an abundant scavenger, hemoglobin (Hb), reacts so quickly with it? This question has puzzled many investigators ever since the identification of NO as the endothelium-derived relaxation factor was made (1–4). NO reacts with Hb via the following reactions:

[1]

[2]

where Fe(II) and Fe(III) refer to the ferrous and ferric forms of the iron in the heme group of Hb, respectively. The rate constants for these reactions are 2.6 × 10⁷ M^–1·s^–1 at 20°C (5) and 6–9 × 10⁷ M^–1·s^–1 at 20°C (6, 7) for Eqs. 1 and 2, respectively. The second reaction destroys NO activity through the formation of nitrate. The activity of NO is potentially conserved by the reaction described by Eq. 1, but to function it must come off the heme and get out of the RBC (where the Hb concentration is ≈0.02 M in heme) before undergoing the fast reactions described by Eqs. 1 and 2. Thus, scavenging of NO in blood can be predicted to be a significant NO sink and poses the problem of how NO can function as the endothelium-derived relaxation factor (8).

One possible resolution of this paradox would be that NO produced in endothelial cells acts only locally and is thus immune to scavenging by Hb, but a variety of evidence indicating increased vasoconstriction when free Hb or Hb substitutes are infused argues to the contrary (9–15). Another hypothesis holds that NO bioactivity is preserved because of reduced NO consumption by Hb due to rate-limiting uptake of NO by the RBC. The NO consumption is believed to be reduced because of a combination of factors that limit the rate of diffusion into the RBC (16–21) and an erythrocyte free-zone near the endothelial layer where NO is produced (22–24). The inherent limitation in the rate of uptake by RBCs is caused by a combination of extracellular diffusion (18) and diffusion through the cell membrane itself (19). RBCs thus effectively reduce the apparent rates that NO is consumed by Eqs. 1 and 2 and make Hb in blood a less effective sink than extracellular Hb.

An alternate hypothesis holds that Hb acts to preserve NO activity, rather than destroy it, through an oxygen saturation-dependent, allosteric mechanism that involves participation of the heme irons and β-93 cysteines (25–32). In this scheme, Hb functions as a NO transporter in addition to its function as an oxygen transporter. One of the central observations supporting this hypothesis involving NO transport via Hb is that nitrosation of the β-93 cysteine to form S-nitrosohemoglobin (SNO-Hb) is favored in the R-state, iron nitrosylation is favored in the T-state, and intramolecular transfer between these sites takes place in response to oxygen saturation-dependent conformational changes (25, 26, 31). Thus, NO that is captured by the heme through Eq. 1 has been proposed to be preserved through intramolecular transfer to form SNO-Hb. The transfer between thiol and hemes may be facilitated by the presence of glutathione (29), but it has also been reported to occur in the absence of glutathione (31). According to this model, the Hb-associated NO is delivered to the tissues by transnitrosation from a small fraction of SNO-Hb to other intracellular thiol groups (such as those of the anion exchange protein AE1) (30) upon decreases in Hb oxygen saturation with subsequent export of NO activity. Thus, the model features NO delivery that is coupled to oxygen tension via the allosteric behavior of Hb in a manner that would contribute to vasodilation when it is necessary and thereby improve tissue oxygenation. The total Hb-associated NO has been proposed to be in great excess to that necessary for regulation of vascular tone so that most of the NO released by the β-93 cysteine is proposed to be recaptured by the heme, and only a small portion is transferred to other thiol groups (29). Accordingly, NO would be preserved as iron nitrosyl Hb (HbNO) or SNO-Hb such that the total NO either bound or bonded to Hb would essentially remain constant during changes in oxygen tension, with NO migrating from the heme iron to the β-93 cysteine as oxygen tension increases, and conversely migrating from the cysteine to the heme as oxygen tension decreases (31). Although the NO transport model that includes allosterically controlled intramolecular transfer is attractive, it remains controversial (33–40). In this article, we directly examine one aspect of the model, the ability of Hb to transfer NO from the β-93 cysteine to the heme and vice versa during cycles of oxygenation and deoxygenation. Using whole blood, washed red cells, and Hb we find no evidence for the oxygen-dependent reversible NO transfer.

Materials and Methods

Preparation of Blood, RBCs, and Hb. In all cycling experiments using whole blood and RBCs, venous blood was drawn into tubes containing EDTA and used on the same day for all experiments except for the oxy/deoxy/reoxy cycles, where samples were incubated overnight at 4°C with 10% volume PBS with citrate, dextrose, and adenine (CPD-A solution, Baxter, Covin, CA) and equilibrated with room air (38). RBCs were obtained by washing the cells in 25-fold excess PBS (pH 7.4) at 4°C for 5 min. The cells were spun down by centrifugation at 3,000 × g, and the wash was repeated five times. After the final wash, RBCs were adjusted to ≈50% hematocrit with PBS. Hb was prepared as described (41, 42). No procedures were undertaken to remove naturally occurring organic phosphates.

Preparation of HbNO and SNO-Hb. Samples were deoxygenated by repeated application of a gentle vacuum followed by flowing argon in a round bottomed flask. HbNO was prepared by adding NO-saturated PBS into deoxygenated blood, RBCs, or Hb samples. SNO-Hb in RBCs was made by incubating blood samples with S-nitrosocysteine (CysNO) (10 mM final concentration) on ice for ≈1–2 h as described (38). Blood samples were then washed, and the RBCs, now containing ≈1% SNO-Hb, were diluted with PBS to reach a hematocrit of ≈50%. Purified SNO-Hb was made the same way as for the RBCs except that dialysis against a large volume of PBS (resulting in dilution factor of ≈10⁹ after several changes) was used to remove residual S-nitrosothiols. In both preparations, adding NO buffer or adding CysNO, some SNO-Hb and HbNO always formed as indicated in Figs. 1, 2, 3. Methemoglobin levels were always <3%, ensuring that the R-T transition was possible.

Fig. 1.

EPR and chemiluminescence during cycling of oxygen saturation. (a) HbNO was determined by EPR for a RBC sample that started from an oxygenated sample (pink line), to which CysNO was added. The sample was deoxygenated (blue line) and then reoxygenated (green line). EPR signals were normalized by heme concentrations, which were 3.7 mM for oxy, 6.2 mM for deoxy, and 8.6 mM for the reoxy sample. The corresponding percentage of HbNO (heme basis) was 0.043%, 0.043%, and 0.038%. (b) SNO-Hb was determined by the chemiluminescence () assay for the same sample during cycling. The samples were treated with 5% acidified sulfanilamide (AS)/mercuric chloride (Hg). The difference of peak area between samples treated without AS or Hg (–AS/–Hg) and samples treated with AS and without Hg (+AS/–Hg) gives the amount of nitrite, and the difference between +AS/–Hg and samples treated with both AS and Hg (+AS/+Hg) shows the amount of SNO-Hb. Heme concentrations in the chemiluminescence assay varied because of dilutions during preparation. Heme concentrations for chemiluminescence measurements were 0.205 mM for oxy, 0.386 mM for deoxy, and 0.25 mM for reoxy samples. The corresponding volume of each injection into the chemiluminescence instrument was 70, 150, and 150 ml, yielding corresponding calculated percentages of SNO-Hb of 0.49%, 0.17%, and 0.16%.

Fig. 2.

Average HbNO and SNO-Hb during oxygen cycling of Hb. The average and standard deviations of the percentages of measured (solid black lines) NO in the form of HbNO (▴) and SNO-Hb (▪) are presented. Data shown in each cycle were normalized to the initial amount of HbNO and SNO-Hb, respectively. The prediction (hashed red lines) from the allosterically controlled NO transfer mechanism is also shown for HbNO (▵) and SNO-Hb (□) and was calculated based on a 30% transfer of NO from HbNO to SNO-Hb during oxygenation steps and 30% of SNO-Hb transferred to form HbNO during deoxygenation steps. (a and b) Deoxy/oxy/redeoxy cycle (labeled, D, O, RD) with Hb (n = 3). HbNO (50 μM) was added to deoxyHb (1 mM). The initial samples contained 50 ± 7 μM HbNO and 3.3 ± 0.9 μM SNO-Hb. (c and d) Oxy/deoxy/reoxy cycle (labeled O, D, RO) with Hb. SNO-Hb was added to 1 mM oxyHb (n = 3). The initial samples contained 31 ± 9 μM SNO-Hb and 8.0 ± 3.3 μM HbNO.

Fig. 3.

Average HbNO and SNO-Hb during oxygen cycling of whole blood and RBCs. The average and standard deviations of the percentages of measured (solid black lines) NO in the form of HbNO (▴) and SNO-Hb (▪) are presented. Data shown in each cycle were normalized to the initial amount of HbNO and SNO-Hb, respectively. The prediction (hashed red lines) from the allosterically controlled NO transfer mechanism is also shown for HbNO (▵) and SNO-Hb (□) and was calculated based on a 30% transfer of NO from HbNO to SNO-Hb during oxygenation steps and 30% of SNO-Hb transferred to form HbNO during deoxygenation steps. (a and b) Deoxy/oxy/redeoxy cycle (labeled, D, O, RD) with whole blood (n = 4). The initial sample was prepared by adding NO buffer to whole, deoxygenated blood and resulted in 0.9 ± 0.3% HbNO and 0.3 ± 0.2% SNO-Hb. (c and d) Deoxy/oxy/redeoxy cycle (labeled, D, O, RD) with washed red cells (n = 4). The initial sample was prepared by adding NO buffer to deoxygenated red cells and resulted in 1.7 ± 1.1% HbNO and 1.1 ± 1.2% SNO-Hb. (e and f) Oxy/deoxy/reoxy cycle (labeled O, D, RO) with washed red cells (n = 3). The initial sample was prepared by adding CysNO to oxygenated red cells and resulted in 0.09 ± 0.06% HbNO and 0.6 ± 0.12% SNO-Hb.

Cycling. All deoxygenation procedures were the same as that described above using a vacuum and argon. Oxygenation was accomplished by gently stirring the sample in room air. For RBC samples, some air-equilibrated PBS was added to increase the rate of oxygenation. Cycling of blood and red cells was aimed at achieving physiologically relevant levels of Hb oxygen saturations during circulation, whereas complete oxygenation and deoxygenation was used for Hb. For blood and red cells, near-IR absorption spectra using a Perkin–Elmer Lambda 9 spectrometer with an integrating sphere detector was used to ensure the desired Hb oxygen saturation, as described (41). The average amount of deoxyHb for all deoxygenation and redeoxygenation steps of blood or RBCs was 66 ± 16% (percentage based on total heme), and the average amount of oxygenated Hb in all oxygenation and reoxygenation steps was 95 ± 10% (heme basis), with the level of metHb always being <3%. For Hb, the percentage of deoxyHb in all deoxygenation and redeoxygenation steps was >99%, and the percentage of oxygenated Hb in all oxygenation and reoxygenation steps was also >99%, as determined by visible absorption spectroscopy.

Measurement of SNO-Hb and HbNO in Cycling Experiments. Electron paramagnetic resonance (EPR) spectroscopy was carried out at 132 K with a Bruker 4111 VT controller and ER-200 D ESR spectrometer set at 9.43 GHz, 10 mW, 5 G modulation amplitude, a time constant of 100 ms, and collection time of 100 s per scan, or with similar conditions using a Bruker Elexsys E500 spectrometer. The percentage of HbNO was determined by double-integrating the EPR spectrum and comparing it to a standard as described (43). In a few cases where the signal to noise of the EPR spectrum was poor, it was deconvoluted into basis spectra (43), and the double integral was performed on the fit. The total heme concentration of blood and red cell samples was determined by diluting the sample into oxygenated PBS and measuring the absorbance. The chemiluminescence () assay was performed as described to measure SNO-Hb (38, 39). Although the assay to measure SNO-Hb has been extensively verified in comparison to the Griess–Saville assay (38), we independently verified the methods ourselves. Plotting the measured SNO-Hb by using the assay against that measured with the Griess–Saville assay fit to a line with a slope of 1.12, with an r value of 0.99 (P < 0.001) and had a 5% rms error (six trials, data not shown), verifying the technique.

Measurements on Whole Venous and Arterial Blood. Arterial and venous blood was collected from indwelling catheters in normal volunteer subjects participating in National Institutes of Health protocol 03-H-0020. Blood was directly collected into heparinized vacutainers to preserve in vivo Hb-oxygen saturation, placed on ice, and immediately centrifuged at 750 × g; plasma was removed, flash-frozen, and stored at –80°C. Samples from arterial blood were thawed in air at room temperature (21°C), transferred, and frozen in EPR tubes. The process of thawing and freezing took <5 min. For venous blood samples, the tubes were thawed in a septum-capped flask under argon and transferred to the EPR tubes with a gas-tight syringe.

Results and Discussion

Cycling. We performed Hb oxygen saturation cycling experiments on whole blood, washed RBCs, and Hb, which was either predominantly S-nitrosated or predominantly heme nitrosylated. Typical data from a deoxygenation/reoxygenation cycle are shown in Fig. 1. In this case, RBCs, pretreated with CysNO, were cycled, and at each point in the cycle, samples were taken to measure both HbNO by EPR and SNO-Hb by chemiluminescence. The data were quantified and calculated as fractional change based on the amount of each species present in the original preparation.

HbNO (50 μM) was incubated with deoxyhemoglobin (1 mM) under deoxygenated conditions and then subjected to an oxygenation, redeoxygenation cycle. As shown in Fig. 2 a and b, both the concentration of HbNO and the concentration of SNO-Hb were not affected by this process. In a similar experiment, an oxygenated sample of SNO-Hb was deoxygenated and reoxygenated, and again there was little change in concentration of both species. These data demonstrate that in Hb no oxygen-dependent transfer of NO between hemes and thiols occurs.

To determine whether allosterically controlled heme-to-thiol cycling is an in vivo phenomenon that is lost upon Hb purification, similar experiments were performed with whole blood and washed red cells. These studies are particularly important because glutathione, present in our red cell and blood preparations but not our Hb preparations, may contribute to intramolecular transfer of NO (29). Addition of NO to deoxygenated blood resulted in the formation of HbNO as a major product and SNO-Hb as a minor product. Oxygenation and subsequent redeoxygenation of this preparation resulted in a decrease in HbNO (Fig. 3a) and also a decrease in SNO-Hb (Fig. 3b). A similar experiment performed with washed red cells showed little change in the concentration of either component upon cycling (Fig. 3 c and d). Washed red cells treated with CysNO under aerobic conditions gave SNO-Hb as the major product and HbNO as the minor product. Deoxygenation and subsequent reoxygenation of this preparation resulted in little change of HbNO but a decrease in SNO-Hb (Fig. 3 e and f). The trends showing overall decay in SNO-Hb and HbNO are consistent with a general instability of SNO-Hb in the red cell (37) and reactions between oxygen and HbNO to form metHb and nitrate (44).

Also shown in Figs. 2 and 3 are predictions (hashed red lines) of the allosterically controlled intramolecular transfer mechanism based on 30% of the NO bound to iron being transferred to thiols upon oxygenation and 30% of the NO on thiols being transferred to iron upon deoxygenation. The 30% transfer rate is a conservative estimate compared with where virtually 100% of the NO in Hb undergoing an R/T transition was reported to participate in intramolecular transfer of NO (31). It is clear that in no case does the prediction of allosterically driven intramolecular NO transfer agree with the experimental data. Overall, our studies summarized in Figs. 2 and 3 imply that, under physiological conditions, any transfer of NO between thiols and hemes within Hb is less significant than the overall decay in the HbNO ligation, and any transfers that may occur do so in a manner that is not consistent with allosteric control of intramolecular transfer of NO in Hb.

Effects of Nitrite. Previous studies that supported allosterically controlled intramolecular transfer of NO in Hb demonstrated loss of iron nitrosyl and formation of a free radical signal upon oxygenation of HbNO (31). This was interpreted to indicate that redox chemistry is involved in the heme-thiol transfer. However, at least some of these studies were performed in the presence of 300 μM nitrite (31). Here, we show that changes that have been attributed to intramolecular transfer of NO can be explained by nitrite chemistry. When oxyHb was incubated with nitrite, a free radical signal was observed (Fig. 4a Inset). This signal was observed by Kosaka et al. (45), is centered at g = 2.005, and has an 18G peak-to-peak width. Signals of this type are most frequently assigned to aromatic amino acid radicals. As demonstrated by Doyle et al. (46), incubation of nitrite with deoxyhemoglobin led to the robust formation of HbNO (Fig. 4a), which was preserved upon reoxygenation. Doyle et al. found that the products of the reaction of nitrite and deoxyhemoglobin, metHb and iron-nitrosyl-Hb, are made at equal rates (2.69 M^–1·s^–1), where 28% of the reacted deoxyHb becomes iron-nitrosyl-Hb (46). To prove that the nitrosyl ligand derives from nitrite and not from SNO-Hb, S¹⁴NO-Hb was incubated in the presence and absence of ¹⁵N-labeled nitrite (300 μM). The hyperfine splitting in an EPR spectrum of HbNO, which arises from the pentacoordinate alphanitrosyl species, will be a triplet for naturally abundant ¹⁴N but will be a doublet for ¹⁵N. A negligible EPR signal was seen in the oxygenated SNO-Hb preparation without nitrite, and (again) the free radical signal was observed upon addition of nitrite (Fig. 4b). When the SNO-Hb sample prepared without nitrite was deoxygenated, there was no increase in HbNO (Fig. 4b), consistent with Fig. 1 and contrary to the allosterically controlled intramolecular transfer mechanism. When the SNO-Hb prepared in the presence of nitrite was deoxygenated, a large increase in HbNO was observed (Fig. 4b).

Fig. 4.

Effects of nitrite on HbNO formation upon cycling. (a) Nitrite (300 μM) was added to 1 mM oxygenated Hb, which resulted in the radical signal shown (Inset). When the sample was deoxygenated, a large HbNO signal appeared with characteristic hyperfine splitting (blue line). When this sample was reoxygenated, the magnitude of the HbNO signal remained unchanged (pink line). Hyperfine structure is indicated by arrows. (b) EPR spectra were obtained for SNO-Hb (1.6 mM total heme, 3.3% S-nitrosation) prepared in the absence (dark blue line) or presence (pink line) of 300 μM ¹⁵N nitrite. Upon deoxygenation, the sample without nitrite gave a very small EPR signal (yellow line), but that containing ¹⁵N nitrite produced a large signal characteristic of HbNO with doublet hyperfine splitting (light blue line), indicated by arrows.

The doublet hyperfine splitting demonstrates that the HbNO formed by deoxygenation is caused by the ¹⁵N-labeled nitrite and not the β-93 S-nitrosocysteine. The total amount of HbNO in the deoxygenated sample, calculated by performing the double integral, was 60 μM. That this is less than the maximum predicted through reduction of the nitrite by deoxygenated Hb (28% of 300 μM) (46) is probably caused by reaction of the nitrite with oxygenated Hb before deoxygenation and/or freezing the sample before the reduction reaction by deoxyhemoglobin is complete. In any case, the amount made supports our contention that there was no significant contribution from intramolecular transfer of NO from SNO-Hb. When we fit the HbNO signal produced by deoxygenation in the presence of nitrite to basis spectra for Hb¹⁴NO (betanitrosyl, hexacoordinate alphanitrosyl, and pentacoordinate alphanitrosyl) (43) and ¹⁵N pentacoordinate α-HbNO, we found 85% of the signal seen is from the ¹⁵N nitrite rather than the ¹⁴N SNO-Hb (calculated as the ratio of ¹⁵N pentacoordinate alphanitrosyl over the total alphanitrosyl). The 15% from ¹⁴N is probably just an artifact of fitting to four rather than three basis spectra, as the improvement in the sum of the squared differences (χ²) was only 8% after the spectrum for ¹⁴N pentacoordinate alphanitrosyl species was added to the fit with the other three spectra. On the other hand, when the spectrum for ¹⁵N pentacoordinate alphanitrosyl species is added to the fit with the other three spectra, the (χ²) improved by 185%. Again, that there is no contribution from SNO-Hb is most clearly demonstrated in Fig. 4b where no detectable HbNO is formed upon deoxygenation in the absence of nitrite, consistent with the data summarized in Figs. 1, 2, 3. These data suggest that previous results demonstrating an NO heme/thiol cycle (31) may have been confounded by the presence of nitrite.

In Vivo Levels of HbNO. The levels of HbNO in vivo are the subject of some debate and have been reported to be ≈5 μM in mixed venous blood that is lowered to ≈2.5 μM in arterial blood, consistent with allosteric control of intramolecular transfer of NO in Hb (31). If these levels are correct, after centrifugation, one would expect there to be ≈10 μM HbNO in packed red cells from venous blood and 5 μM HbNO in packed red cells from arterial blood, concentrations that are large enough to detect with EPR (which is sensitive down to 500 nM). We measured the EPR spectra of red cells obtained from venous and arterial blood while maintaining appropriate oxygen tensions during the procedure. A standard sample of 1 μM HbNO gave a clearly identifiable EPR spectrum (Fig. 5a). The signal observed in venous blood (Fig. 5b) did not resemble HbNO and is probably caused by the copper(II) signal from ceruloplasmin or superoxide dismutase. In contrast to the calculated g value of 2.009 for HbNO (Fig. 5a), we calculated g = 2.052 for the average of the spectra in Fig. 5b, which agrees well with published values for ceruloplasmin of 2.05/2.06 (47) or superoxide dismutase of 2.056 (47). The small difference in the signal from arterial and venous red cells shown in Fig. 5c is most likely caused by alterations in protein copper oxidation state rather than alterations in HbNO. In any case, comparison of the spectra in Fig. 5 b and c with that of a standard 1 μM HbNO sample demonstrates that there is <0.5 μM HbNO in venous blood (consistent with previous reports) (39, 48, 49) and no measurable A-V gradient.

Fig. 5.

Measurements of HbNO in RBCs prepared from arterial and venous blood. (a) EPR spectrum of a standard Hb (1 mM) sample with an HbNO content of 1.1 μM. The spectrum represents the average of 10 100-s scans. (b) EPR of red cells obtained from venous blood of four different volunteers. The cells were processed so as to avoid exposure to additional oxygen (see Materials and Methods). Each spectrum represents the average of 10 100-s scans. A 10 μM HbNO signal would have a peak-to-peak height of ≈1,500. (c) EPR spectra of red cells from venous blood of four volunteers minus that from cells prepared from their arterial blood. Each spectrum for each volunteer is the average of 20 100-s scans. A 5 μM HbNO signal would have a peak-to-peak height of ≈750.

Summary

If the quaternary state of Hb controls intramolecular transfer of NO, it would be expected that SNO-Hb would be converted to HbNO upon deoxygenation and HbNO to be converted to SNO-Hb upon oxygenation. Of all our many experiments performed on Hb, red cells, and whole blood, we did not observe any NO transfer that could be construed as being consistent with intramolecular transfer of NO linked to the allostery of Hb (Figs. 1, 2, 3). Moreover, we have shown that results previously attributed to intramolecular transfer of NO may have been caused by the presence of exogenous nitrite (Fig. 4). Finally, we have shown that HbNO is not present in venous blood at the levels reported (31), and there is no A-V gradient for HbNO >0.5 μM as predicted by the allosterically controlled NO transfer mechanism (Fig. 5).

The concentration of NO used in our experiments in blood and RBCs was much less than the heme concentration (NO to heme ratio of ≈1:100) but much larger than found in vivo (Fig. 5). However, for these differences in concentration of NO to be important, one must posit some chemistry that occurs at an NO to heme ratio of <1:4,000 but not at 1:100. Our results imply that there is no significant, reversible oxygen saturation-dependent NO transfer between the thiols and hemes within Hb in vitro or in vivo. Thus, we find that NO activity is not likely to be preserved by Hb in this manner.