Effects of S-Nitrosation on Oxygen Binding by Normal and Sickle Cell Hemoglobin*

S-Nitrosated hemoglobin (SNO-Hb) is of interest because of the allosteric control of NO delivery from SNO-Hb made possible by the conformational differences between the R- and T-states of Hb. To better understand SNO-Hb, the oxygen binding properties of S-nitrosated forms of normal and sickle cell Hb were investigated. Spectral assays and electrospray ionization mass spectrometry were used to quantify the degree ofS-nitrosation. Hb A0 and unpolymerized Hb S exhibit similar shifts toward their R-state conformations in response to S-nitrosation, with increased oxygen affinity and decreased cooperativity. Responses to 2,3-diphosphoglycerate were unaltered, indicating regional changes in the deoxy structure of SNO-Hb that accommodate NO adduction. A cycle of deoxygenation/reoxygenation does not cause loss of NO or appreciable heme oxidation. There is, however, appreciable loss of NO and heme oxidation when oxygen-binding experiments are carried out in the presence of glutathione. These results indicate that the in vivo stability of SNO-Hb and its associated vasoactivity depend on the abundance of thiols and other factors that influence transnitrosation reactions. The increased oxygen affinity and R-state character that result fromS-nitrosation of Hb S would be expected to decrease its polymerization and thereby lessen the associated symptoms of sickle cell disease.

Hemoglobin (Hb) 1 is of central importance to human health in its role as a respiratory protein. Another chapter in the study of the human health significance of Hb is beginning, focused on NO uptake and delivery by Hb and the role this plays in the control of blood pressure and other NO-dependent reactions. Nitrosation of sulfhydryl groups on the Hb tetramer creates S-nitrosated Hb (SNO-Hb), which has been shown to play an important role in NO uptake and delivery (1). S-Nitrosated forms of proteins such as Hb can be formed via interaction with nitrosating agents formed upon interaction of NO and oxygen (NO x ) and by NO-exchange reactions (transnitrosations) with nitrosated forms of low molecular weight thiols such as cysteine and glutathione. Conversely, the low molecular weight thiols can act as NO acceptors in transnitrosation reactions where NO is donated by S-nitrosated proteins (2,3).
Hb-based NO transport via SNO-Hb is significant because it can greatly extend the range of NO-dependent reactions. Unlike SNO-Hb, free NO is a very reactive molecule, whose lifetime in the complex cellular milieu would be expected to be very short. It is this characteristic of NO that delayed the discovery of NO-dependent reactions in smooth muscle relaxation, platelet inhibition, neurotransmission, and immune regulation (4 -8). What is learned about Hb-based NO transport will have far-ranging applications in these disparate fields.
The studies reported here concern the oxygen binding properties of variably S-nitrosated adult human hemoglobin (Hb A 0 ) and sickle cell hemoglobin (Hb S) that has a Glu 3 Val substitution at ␤6. Although physiological levels of S-nitrosation of Hb are too low for oxygen transport to be significantly affected, the linkage between S-nitrosation and oxygen binding by Hb can affect the stability and subsequent vasoactivity of SNO-Hb. Understanding this linkage is critical for understanding the NO-dependent reactions of Hb in normal and sickle cell erythrocytes, in cell-free Hb-based blood substitutes, or in pharmaceuticals. As will be shown, S-nitrosation of Hb A and unpolymerized Hb S results in increased oxygen affinity via partial stabilization of their high affinity, R-state, conformations. Since R-state-stabilized Hb S does not readily polymerize (9), this finding prompts us to suggest that S-nitrosation of Hb S may be viewed as a possible therapeutic approach to alleviating sickle cell disease.
Functional and crystallographic studies have shown that the Cys-␤93 residues at which NO is bound as NO ϩ in SNO-Hb are more accessible in the high affinity conformation of oxy (Rstate) Hb than in deoxy (T-state) Hb (10 -12). This conformational sensitivity results in a rate dependence for SNO-Hb formation that mirrors the greater relative exposure of Cys-␤93 in conditions that favor the R-state and was invoked to explain the decreased stability of the deoxy form of SNO-Hb (1). Although deoxy SNO-Hb is less stable than the oxy form, we found that purified deoxy SNO-Hb is sufficiently stable to allow oxygen-binding studies to be carried out over a period of several hours. However, as will be documented, loss of NO from SNO-Hb during a cycle of deoxygenation/re-oxygenation can occur under simulated in vivo conditions where NO acceptors such as glutathione are present.
Because the sulfhydryl groups at Cys-␤93 are in a conformationally sensitive position on the Hb tetramer, the S-nitrosation of Hb would be expected to have heterotropic allosteric effects on ligand binding by the heme groups at the four active sites of the tetramer. The following results document the ex-istence and nature of these thiosteric effects. We show that S-nitrosation promotes increased oxygen affinity and thus acts in opposition to anionic allosteric effectors that decrease oxygen affinity. Hb A 0 and Hb S show similar responses to Snitrosation, as expected based on their similar structures in the region of Cys-␤93 (13). The shift of SNO-Hb forms toward higher oxygen affinity probably involves a regional conformational alteration of the deoxyHb tetramer that prevents His-␤146 from making its normal contribution to T-state stability. This was previously shown to be the case in Hb in which the SH groups at Cys-␤93 were modified by N-ethylmaleimide (11,14).
We previously showed that increasing anion levels can modulate Hb function by decreasing the frequency and extent of conformational fluctuations that control the accessibility of the heme groups where oxygen is bound (15). In this report we show that these same considerations underlie the anion-dependence of oxygen binding to SNO-Hb and its decreased stability in the presence of low molecular weight thiols.

EXPERIMENTAL PROCEDURES
Sample Preparation-Samples of Hb A 0 and Hb S were prepared by using the ammonium sulfate method, stripped of organic phosphate cofactors and purified by chromatography as described previously (16). The amounts of oxidized Hb (metHb), oxygenated Hb (oxyHb), and hemichrome were determined by spectral analysis by methods published by Winterbourn and co-workers (17). Samples that contained any detectable hemichrome or greater than 5% metHb were discarded. For experiments with SNO-Hb, the stock Hb solutions, typically 1-3 mM in heme units (Fe porphyrin units), were made just before use and were never frozen. Other experiments where metHb formation was less critical used stock Hb solutions that were stored in liquid nitrogen prior to use. To generate samples with progressively higher levels of metHb, samples were repetitively deoxygenated and re-oxygenated with intervening periods of exposure to low oxygen levels.
HEPES, 2,3-diphosphoglycerate, and KNO 2 (from Sigma) and KCl (Fisher; Ͼ99%), were dissolved and adjusted to pH values indicated. The metal chelators used were 0.1 mM DTPA or 0.05 mM EDTA, which gave equivalent results in the studies reported here.
Preparation of SNO-Hb-Our protocol for generation of Hb with varied levels of S-nitrosation used purified oxygenated Hbs at 0.5-3 mM in heme in 2% borate buffer at pH 9.2 with 0.1 M DTPA or 0.05 M EDTA. These were kept in the dark at 4°C and exposed to CysNO at varied levels (ratios of 0.5-4 CysNO to heme) and incubated for time periods of 0 -5 min prior to chromatograhy. This range of conditions was found to be adequate for generation of samples with a wide span of S-nitrosation. The low molecular weight NO donor, CysNO, was removed by passage through a 22-cm Sephadex G25 column at 4°C. The elution volumes for free Hb and free CysNO were determined, and controls were run to ensure that all low molecular weight materials were separated from the eluted SNO-Hb sample. The purified SNO-Hb was dialyzed against selected buffers for use in subsequent experiments. The variably S-nitrosated Hb samples were spectrally evaluated after dialysis to the desired pH and anion condition and rejected if they contained detectable hemichromes or more than 5% oxidized Hb A 0 or 9% oxidized Hb S.
Analysis of Extent of S-Nitrosation of Hb Samples-The fraction of S-nitrosation was typically determined by spectral deconvolution analysis as described here. Prior to spectral evaluation, the samples were subjected to Sephadex G-25 chromatography with 2% borate, 0.1 mM DTPA buffer, pH 9.2 (carried out at 4°C) to standardize the sample pH and buffer conditions and to remove any low molecular weight materials. The samples were degassed in tonometers and subjected to spectral analysis before and after treatment with 4 mg/ml dithionite (sodium hydrosulfite, Tech, Acros Chemical Co.). Dithionite addition rapidly removes any residual oxygen, reduces any metHb present and releases NO from the SNO linkage. The NO is effectively captured by the reduced heme, and the spectrum that results is that of partially NO-Hb. NO gas (National Welders, CP grade), further purified by passage through 5 M and then 1 M NaOH, is added to the degassed, dithionitetreated sample, and the spectrum of fully NO-Hb is obtained after a 10-min equilibration period. Comparisons of the partially and fully NO-Hb spectra allow the fraction of S-nitrosation prior to dithionite treatment to be determined. The degree to which the dithionite-treated sample shows the presence of the NO-Hb spectrum is a direct result of NO released from SNO-Hb, since there was no prior NO ligation of the heme and all low molecular weight thiols and nitrogen oxides that could contribute NO have been removed. Accordingly, a sample that has 50% NO-Hb after dithionite treatment corresponds to 100% S-nitrosation of the two ␤93 SH groups. Samples with internal SH groups nitrosated can give higher than 50% NO-Hb readings after dithionite treatment. The dithionite-treated sample was analyzed 5 min after dithionite addition. Since the NO binding to deoxy heme is very fast (diffusion limited) and the NO dissociation rate is extremely slow, the fact that free NO liberated from the NO-Hb complex is slowly reduced by dithionite is not a serious detriment to this spectral assay (18). This somewhat cumbersome protocol avoids experimental difficulties associated with troublesome dithionite reaction products and the pH and aniondependent variation of the Hb-NO spectrum.
Electrospray ionization mass spectrometry with some modifications of the method reported by Ferranti et al. (19) confirmed that the sole effect of the transnitrosation treatment was S-nitrosation of ␤93 SH groups. This methodology also guided our choice of experimental conditions where internal SH groups were neither nitrosated nor cystinylated. The mass spectrometry method was done in aqueous solution using a Micromass Quattro LC triple quadrupole mass spectrometer equipped with Z-spray ion source. The Hb samples were electrosprayed in 0.02 M ammonium bicarbonate, pH 9.04.
Oxygen Binding Measurements-Oxygen binding measurements were performed tonometrically, with a modified spectrophotometric method based on that of Riggs and Wolbach (20). The S-nitrosated samples were kept in the dark except during spectral analysis (Ͻ30 s for data collection with a Hewlett-Packard model 8451A or model 8453 diode array spectrophotometer), and very few data points were collected in each set to minimize the time under low oxygen conditions where Hb oxidation was most pronounced. The total curves presented are composites of data points from multiple experiments, with each experiment limited to about 2 h. Since the presence of metHb can influence the binding curves, experiments that resulted in Ͼ12% metHb A 0 or Hb S at the end of the experiment were not used.

Formation of Variably S-Nitrosated
Hb-Modification of previously published procedures was required for generation of S-nitrosated Hb suitable for measurements of oxygen binding and anaerobic redox reactions. We found that exposing high (Ͼ500 M) concentrations of oxyHb at 4°C with low ratios of CysNO to Hb for short time intervals in the dark in 2% borate-EDTA or DTPA buffer, pH 9.2, largely avoided metHb formation and nitrosation of internal SH groups. These conditions also minimized metal or light-induced release of NO from CysNO and subsequent NO-induced Hb oxidation. Inclusion of a metal chelator in all buffers was essential to prevent the metal-catalyzed formation of disulfides (5). The short half-life usually found for CysNO is due to artifactual contamination of buffers with copper or other redox-active trace metals (21). Free copper and iron in off-the-shelf buffers may frequently be in the nanomolar range, levels sufficient to alter reactions from the pathways they would take in the absence of redox metals.
Transnitrosation reactions to create SNO-Hb have the net result shown simplistically in the equilibrium representation below, where R represents a low molecular weight thiol.
The extent of S-nitrosation of Hb samples after removal of the low molecular weight NO donor was typically determined by spectral deconvolution analysis under standard conditions as described under "Experimental Procedures." Fig. 1 shows representative spectra for deoxygenated SNO-Hb, the mixture of deoxy and NO-Hb that results from dithionite addition, and the final NO-Hb spectrum obtained after NO addition to fully occupy the heme binding sites. The spectrum of deoxygenated SNO-Hb indicates the presence of a low level of metHb that exceeds that of the deoxygenated control sample, but shows no evidence of NO ligation of the heme. The amount of NO ligan-ded to the heme that appears after the NO-linkage to Cys-␤93 is disrupted by dithionite is used to estimate the extent of S-nitrosation. For some samples the extent of S-nitrosation was also quantified by the Saville reaction (22,23), with a standard curve generated with precisely determined quantities of nitrite. The results obtained were within 10% of the values obtained with our deconvolution assays.
The experimental conditions used in our experiments were established by detailed studies of reactions between CysNO and Hb, in which we used an innovative approach to electrospray ionization mass spectrometry that avoided the use of organic solvents. 2 Using this technique, we obtained spectra of Hb samples after varied periods of exposure to CysNO and chromatographic removal of this low molecular weight NO donor. A representative spectrum is shown in Fig. 2, in which the ␤ chains are about 50% S-nitrosated. Under our reaction conditions, the mass change associated with NO addition to the ␤ chains is the only mass change observed.
Oxygen Equilibria of Hb A 0 with Varied Levels of Nitrosation of Cys-␤93-We determined the effects of S-nitrosation on oxygen binding by Hb A 0 in chloride-free HEPES buffer and in the presence of a 50-fold excess of 2,3-diphosphoglycerate over tetramer concentration. The extent of S-nitrosation of Hb A 0 was measured before and after oxygen equilibria and found to be equivalent in these experiments. Readers are reminded that care was taken to avoid S-nitrosation of buried residues, which can result in shifts in logP 50 larger than those shown. As noted under "Experimental Procedures," these experiments were designed to minimize exposure to light and to time at low O 2 . In the results presented here, the Hb samples had low (Ͻ5%) initial levels of metHb and less than 10% metHb after completion of oxygen equilibria, even at high levels of S-nitrosation.
The results presented in Fig. 3A are for approximately 30%, 40%, and 80% S-nitrosated Hb A 0 and 0% S-nitrosated controls in chloride-free HEPES/EDTA buffer at pH 7.5. Our estimates of the degree of S-nitrosation of a sample varied by Ϯ5%. The changes toward higher affinity are progressive with increases in degree of S-nitrosation. Fig. 3B shows similar data for 0% and approximately 80% S-nitrosated Hb A 0 in the presence of 2 S-Nitrosylation and MetHb Formation-Some metHb is formed during S-nitrosation reactions as a result of NO (liberated from CysNO) interacting with oxygenated Hb. By limiting the levels of CysNO used and time of exposure, it was possible to generate SNO-Hb samples with low levels of metHb for use in oxygen binding experiments. To estimate the effect of increased levels of metHb, we carried out oxygen binding studies with unmodified Hb A 0 with progressively higher levels of metHb. The Hill plots of Fig. 3C show that the R-state shifts induced by metHb formation are similar in character to those associated with S-nitrosation. The shifts associated with Snitrosation are, however, much larger than can be attributed to metHb formation in the SNO-Hb samples. Under our assay conditions, where levels of metHb were minimized, the shifts in logP 50 attributable to the presence of metHb in the S-nitrosated samples are less than 10% of the shifts observed.
Stability of SNO-Hb-Prior studies have shown deoxy SNO-Hb to be less stable than oxy SNO-Hb (1). Deoxygenation is a necessary step in the oxygenation studies reported here, a condition that was expected to cause some loss of NO due to instability of deoxy SNO-Hb. Our preparations, however, showed no appreciable loss of NO during oxygen binding experiments. As shown in Fig. 4, the preparations of oxy SNO-Hb used in these studies were stable for over a week without loss of NO when stored at 4°C at high protein concentration. Samples stored in the deoxygenated condition or at lower protein concentration were less stable and had greater loss of NO and more metHb formation. The data shown in Fig. 4 are for aliquots withdrawn from oxy and deoxy samples of about 80% S-nitrosated Hb, 1 mM in heme, held at 4°C. The aliquots were brought to standard conditions for spectral deconvolution as- Low Molecular Weight Thiols Destabilize S-Nitrosated Hb-In contrast to the stability noted above, appreciable diminution of S-nitrosation of Hb A 0 occurs when experiments are carried out in the presence of a 5-fold excess of reduced glutathione (GSH). The bar graph of Fig. 5 illustrates the destabilizing influence of GSH at several levels of S-nitrosation (achieved by exposure to different levels of CysNO for varied times). Previous work demonstrated that GSH can accept NO from SNO-Hb in transnitrosation reactions (1). Fig. 5 shows that SNO-Hb is partly but not completely destabilized when deoxygenation occurs in the presence of this high level of GSH. In vivo conditions may result in more loss of NO during the deoxygenation process. Fig. 5 also shows that the loss of NO from SNO-Hb is roughly the same in the presence and absence of 2,3-diphosphoglycerate. Appreciable increases in metHb levels occur when oxygenation of SNO-Hb is carried out in the presence of glutathione, with metHb levels typically 25-27% after oxygen equilibria, in contrast to Ͻ10% metHb in its absence. This is a consequence of GSH interactions with SNO-Hb, since control experiments showed no effect of GSH on oxygen affinity or metHb levels for unmodified Hb A 0 (data not shown).
The increase in metHb probably results from interactions between oxyHb and free NO that is released from either SNO-Hb or S-nitrosated glutathione during oxygen equilibria.
Oxygen Equilibria of Hb S with Varied Levels of Nitrosation of Cys-␤93-We determined the effects of S-nitrosation on oxygen binding by purified Hb S in HEPES buffer containing EDTA or DTPA and in the presence of a 50-fold excess of 2,3-diphosphoglycerate over tetramer concentration. No aggregation of Hb S was expected or observed upon deoxygenation at the concentrations (60 M in heme) used in these experiments. Aggregation-dependent effects that are evident at much higher protein concentration were not studied. The representative results presented in Fig. 6A are for approximately 80% Snitrosated Hb S and 0% S-nitrosated controls in chloride-free HEPES/EDTA buffer at pH 7.5. Fig. 6B shows a similar pair in the presence of DPG. In these studies, both control and Snitrosated samples of Hb S were more prone to autoxidation than comparable samples of Hb A 0 . The sample handling required for our studies made it difficult to obtain data with low levels of metHb S, and many data sets were not useful due to high metHb levels. The results shown are for control and Snitrosated forms of Hb S samples with relatively low levels of metHb measured before (Յ9%) and after (Յ15%) oxygen binding.
The initial stages of oxygenation of unpolymerized Hb S are MetHb levels, measured before and after oxygen binding, were 1.5-3%, 2.9 -4.2%, 5.1-9%, and 4.8 -9.4%, respectively. In B, data obtained with a 50-fold excess of DPG over Hb tetramer are shown for unmodified Hb A 0 (closed circles) and for about 80% Snitrosated Hb A 0 (open triangles). MetHb levels, measured before and after oxygen binding, were 4.7-6.7% and 7.2-9.7%, respectively. Fig. 3C shows effects of increasing levels of metHb in Hb A 0 . MetHb levels, measured before and after oxygen binding, were 1.5-3% (closed circles), 5. most affected by S-nitrosation (⌬logP 10 Ͼ ⌬logP 50 ), resulting in asymmetric shifts away from the control data. The leftward shift of logP 10 is approximately the same in the presence and absence of 2,3-diphosphoglycerate, resulting in equivalent decreases in cooperativity to about two thirds of the normal values. These results are very similar to those described above for S-nitrosation of Hb A 0 .

DISCUSSION
Understanding the properties of SNO-Hb is critical for better understanding of the NO-dependent reactions moderated by Hb. Results presented here show that S-nitrosation stabilizes the high affinity R-state conformation of Hb, as previously reported for Hbs reacted with N-ethylmaleimide, iodoacetamide, and other SH group modifiers (11, 14, 24 -27). Thiosteric effects brought about by S-nitrosation of Cys-␤93 thus act in opposition to the well known allosteric effectors of Hb function (protons, chloride, organic and inorganic phosphates, bicarbonate, and carbon dioxide) that typically decrease oxygen affinity and thereby enhance oxygen delivery to respiring tis-sues. For SNO-Hb, as found for other SH-modified Hbs, it is probable that the Cys-␤93 residues do not assume their normal position in the deoxygenated molecule, with the consequence that formation of the salt bridge between His-␤146 and Asp-␤94 is inhibited. This critical salt bridge is oxygen-linked, and its pH-dependent formation accounts for about half of the normal Bohr effect (11,12).
The conformational sensitivity of SNO-Hb was shown in an earlier publication to be involved in a dynamic cycle in which SNO-Hb was formed in the lungs and decomposed when NO was delivered to the tissues. The greater accessibility of the Cys-␤93 residues in the oxy (R-state) conformation compared with the deoxy (T-state) conformation of Hb was invoked to account for the more facile S-nitrosation of liganded (R-state) Hb and the greater stability of SNO-Hb in the liganded form (1). The results presented here expand upon these earlier findings and illustrate that deoxygenation alone is not sufficient to cause release of NO from SNO-Hb. NO is accommodated in the altered deoxy structure of SNO-Hb, making it possible to deoxygenate the protein without significant loss of NO. However, we show that the presence of glutathione destabilizes SNO-Hb and allows for NO transfer away from SNO-Hb during a cycle of deoxygenation and re-oxygenation. In red cells, where thiols such as GSH can serve as NO acceptors, the conformational sensitivity of SNO-Hb would facilitate NO release upon deoxygenation. These findings are relevant to the increased vasodilatory action reported for oxy SNO-Hb in the presence of glutathione (1).
Differences in thiosteric and anionic mechanisms of control of Hb function are indicated by the results presented here and by previous studies on other SH-modified Hbs. Notably, Snitrosation results in equivalent shifts of logP 10  or absence of 2,3-diphosphoglycerate, the responses of normal and SNO-Hb to 2,3-diphosphoglycerate are equivalent, and 2,3-diphosphoglycerate does not significantly affect the release of NO from SNO-Hb in a cycle of deoxygenation/re-oxygenation in the presence of glutathione. These observations indicate the existence of regional rather than global effects associated with thiosteric and anionic effectors, a conclusion reached earlier by Perutz and co-workers in regard to the anion sensitivity of Hb in which the SH groups at Cys-␤93 were modified by N-ethylmaleimide (11). This regional control of oxygen affinity makes it possible for SNO-Hb forms to exhibit varied oxygen affinities, dependent on the nature of the parent Hb and its environment.
The regional effects associated with SH group modification have been shown to result, in large part, from disruption of the normal salt bridge between His-␤146 and Asp-␤94, which in turn decreases the proximal-side pull on the heme-iron that normally confers T-state character on deoxygenated Hb (11,12,29). It does not appear that SH group modification by S-nitrosation appreciably alters the ␤ chain anion-binding site in Hb A 0 or in Hb S, since in both proteins the magnitude of the 2,3-diphosphoglycerate effect is equivalent for the normal and S-nitrosated forms. The similarity of responses of Hb A 0 and Hb S to S-nitrosation mirrors their structural similarities as shown by x-ray crystallography (13).
Anion-induced shifts in the Hill plot asymptotes are not interpretable as solely due to their preferential binding to the T-state, a feature that led Minton and Imai (30) to suggest that a minimum of three states was required to describe Hb function. We recently advanced a new paradigm of Hb function in which anion-induced shifts in the apparent T-state are explained by anion-dependent alterations in the conformational fluctuations that expose "buried" sites (15). The functional properties of SNO-Hb as described above are supportive of this paradigm. This paradigm also rationalizes the absence of a 2,3-diphosphoglycerate effect on the loss of NO from SNO-Hb in the presence of glutathione, since Cys-␤93 in the altered deoxy state of S-nitrosated Hb is relatively free of steric hindrance and is thereby insensitive to anion-induced decreases in conformational fluctuations.
Differences in thiosteric and anionic mechanisms of control of Hb function may be of considerable significance for understanding the linkage between the oxygen affinity and polymerization of Hb S and for designing better treatments to alleviate sickle cell disease. Treatments that increase the oxygen affinity of Hb S generally decrease polymerization and the associated red cell morphological changes associated with sickle cell disease (9). Significantly, Hb S derivatives with SH groups modified with glutathione (26) or other thiol reagents (25,27) have decreased tendencies to polymerize. Accordingly, the results reported here lead us to anticipate that new approaches to sickle cell therapies may involve ways to increase the S-nitrosation of Hb S and thereby reduce its tendency to form polymers.
At the relatively low protein concentrations used in the experiments reported here, Hb A 0 and Hb S exhibit similar oxygen affinities and similar responses to S-nitrosation. In contrast, at protein concentrations like those in red cells, the protein concentration-dependent formation of Hb S polymers is accompanied by decreases in its oxygen affinity that may result in distinctive responses to S-nitrosation. The lowering of oxygen affinity associated with polymerization of Hb S (31) and with crystallization of Hb A 0 (32) may reflect decreases in the conformational fluctuations that expose the sterically restricted active sites. Readers will recognize this interpretation as another facet of the conformational fluctuations paradigm described above.
Inhalation of low concentrations of NO was recently reported to increase the oxygen affinity of sickle erythrocytes in vitro and in vivo (33). After treatment, the oxygen affinity of the sickle erythrocytes was approximately equal to that of the normal controls, where polymerization does not occur. This result implies that the NO treatment brought about a reduction in the polymerization of Hb S, as would be expected to occur if the non-polymerizing R-state of Hb S was partially stabilized, as it is in SNO-Hb. Partial heme ligation by NO or heme oxidation induced by NO would also result in higher oxygen affinity. Accordingly, we have initiated studies to determine if Hb S gelation is significantly inhibited by NO ligation or S-nitrosation at levels that might be achievable in vivo.
The data thus far available on the vasodilatory effects of SNO-Hb show that oxidation of the metal center affects NO release and thereby the NO-dependent physiological reactions associated with SNO-Hb (1). In the studies reported here, Hb S oxidation was much more pronounced than that of Hb A 0 , suggestive of altered NO uptake and delivery for this variant form of Hb even in its unpolymerized state. The greater tendency of Hb S to autoxidize has been noted previously, and the resistance to malaria associated with Hb S has been postulated to arise in part as a consequence of greater oxidative events in red cells containing Hb S than for red cells containing only Hb A 0 (34,35). Consequently, the R-state shifts in Hb S induced by S-nitrosation and by oxidation may be joint determinants of the role played by Hb S in malarial resistance, Hb-linked vasodilation, and sickle cell disease.