Electron Paramagnetic Resonance Analysis of Nitrosylhemoglobin in Humans during NO Inhalation*

The reactions of nitric oxide with hemoglobin play an important role in explaining the vascular biology of this free radical. It is perhaps surprising that the level of nitrosylhemoglobin (HbNO) in which NO is bound to the ferrous hemoglobin heme in whole human blood under basal and stimulated conditions is a matter of some controversy, with measurements ranging from <1 nm to close to 10 μm. In order to examine HbNO levels in human blood by using EPR spectroscopy, we have developed a regression-based spectral analysis technique that has a detection level of about 200 nm HbNO. We have utilized this methodology to detect the level of HbNO under basal conditions and during NO inhalation. The major findings of this study are as follows. (i) HbNO can be accurately detected and quantified in whole blood with a detection limit of ∼200 nm. (ii) By using regression analysis, levels of HbNO as low as 0.5–1 μm can be deconvoluted into component species. (iii) HbNO is present at less than 200 nm at basal conditions in both arterial and venous blood and is formed at a level of 0.5–2.5 μm upon inhalation of 80 ppm NO. (iv) The levels of HbNO detected by EPR are remarkably close (within a factor of 2) to those detected by tri-iodide-based chemiluminescence and much smaller than those detected by photolysis chemiluminescence. (v) The half-time of HbNO in vivo is ∼40 min.

The role of hemoglobin in the biochemistry, physiology, and pharmacology of nitric oxide (NO) 3 and nitrovasodilators is controversial (1)(2)(3). Understanding how hemoglobin can control NO-mediated responses is crucial in order to fully appreciate the physiological limitations of endothelial derived nitric oxide, the sequelae of nitric oxide delivery (either through nitric oxide donors, nitrovasodilators, or nitric oxide inhalation), and the vasoconstrictive propensity of hemoglobinbased blood substitutes. It is generally accepted that oxygenated ferrous hemoglobin (oxy-Hb) can convert nitric oxide to nitrate, through a reductive dioxygenation reaction, to form ferric hemoglobin (met-Hb) (4). This reaction is fast (k ϭ 5 ϫ 10 7 M Ϫ1 s Ϫ1 ) (5), irreversible, and generates an inert product. It is also universally accepted that nitric oxide can bind to unliganded ferrous hemoglobin (deoxy-Hb) to form a nitrosyl derivative (HbNO) at an almost equally rapid rate (k ϭ 2.6 ϫ 10 7 M Ϫ1 s Ϫ1 ) (6). Although HbNO is reasonably stable because of the slow ligand off rate, it can be plausibly argued that the NO has not been destroyed but has been preserved, if a viable mechanism for release and utilization can be established. Two major central areas of controversy are whether sufficient quantities of HbNO can be formed in vivo, and whether the formation of S-nitrosohemoglobin from HbNO represents a viable mechanism for NO reutilization (7)(8)(9)(10).
It has been argued that at oxygen concentrations at which hemoglobin is predominantly oxygenated, the small pool of hemes that are still in the high oxygen affinity state (R-state) but are unliganded will bind NO up to 100 times faster than has been appreciated previously. This would favor the formation of HbNO over met-Hb at oxygen saturation levels less than 99% (11,12). This conclusion has been disputed in several laboratories, and methodological mixing problems with these experiments have been highlighted (13)(14)(15)(16). Gladwin et al. (17) examined NO/hemoglobin reactions in vivo in humans by supplementing endogenous production with inhaled NO. These experiments indicated that the major hemoglobin-derived product formed during NO inhalation was met-Hb, together with a low but significantly increased HbNO concentration and a nonsignificant change in S-nitroso-Hb. These authors detected ϳ0.2-0.4 M HbNO under basal conditions, rising to 1.2-1.4 M upon NO inhalation (calculated by assuming 10 mM Hb in whole blood). Rassaf et al. (2) detected Ͻ1 nM HbNO in humans (below the detection limit of their method), although they were easily able to detect HbNO in rodents. These data have been criticized in editorials, based mainly on the methodological issues concerning the chemiluminescence-based measurements of HbNO, and it has been stated that the levels of HbNO detected by Gladwin et al. (17,18) and Rassaf et al. (2) are 10 -100 times too low, based on other studies using the "gold standard" of EPR (19 -21). Indeed, by using photolysis-based chemiluminescence methodology, McMahon et al. (22) reported HbNO levels of ϳ5 M HbNO in human blood. If this were true it would make the levels of HbNO formed in vivo from NO inhalation approach the levels of met-Hb observed and therefore would provide strong evidence for the conservation of NO by preferential NO binding to unliganded R-state hemoglobin subunits. In addition, it has been argued that EPR methodology is incapable of quantification of low levels of HbNO in blood and is only able to detect the penta-coordinate ␣-chain form of HbNO (23).
In this study we have re-examined the critically important issue of the use of EPR for the quantitative measurement of HbNO in human blood. We show that HbNO can be accurately detected and quantified in whole blood with a detection limit of ϳ200 nM. HbNO is present at less than 200 nM at basal conditions and is formed at a level of 0.5-2.5 M upon inhalation of 80 ppm NO. These levels of HbNO, detected by EPR, are remarkably close (within a factor of 2) to those detected by triiodide-based chemiluminescence and much smaller than those detected by photolysis chemiluminescence. In contrast to these small increases in HbNO, levels of met-Hb rise from basal levels of 10 to ϳ150 M after NO inhalation. Finally, we determine the in vivo half-time of HbNO as ϳ40 min.
These data indicate that EPR is a powerful tool for the detection, quantification, and characterization of HbNO in whole blood and substantially validate previous measurements using tri-iodide-based chemiluminescence methods. In addition, the data suggest that nitric oxide dioxygenation to form nitrate, rather than preservation in the form of HbNO, is the major fate of inhaled NO in the bloodstream.

EXPERIMENTAL PROCEDURES
Materials-All chemicals were purchased from Sigma unless otherwise indicated.
Experimental Protocol-The study protocol was approved by the Institutional Review Board of the NHLBI, National Institutes of Health, and all normal volunteers gave written, informed consent. Briefly, blood was collected from the brachial artery and vein using large bore catheters (18 gauge). Venous and arterial blood were collected directly into vaccutainers, spun at 750 ϫ g and opened, and plasma was removed. A thin plastic tube connected to 3-5 mL syringe was inserted into the bottom of red cell pellet and used to transfer red blood cells directly to EPR tubes, which were immediately frozen in liquid nitrogen. All procedures were done at the bedside.
EPR Analysis-For the detection of HbNO and other features in the g ϭ 2 region, EPR spectra were collected on an X-band Bruker Elexsys spectrometer using the following conditions: microwave power 2 milliwatts; time constant of 40.96 ms; modulation frequency 100 kHz; modulation amplitude 5 G; scan time 84 s; scan width 400 G; and temperature 100 K. For each sample, 10 individual scans were averaged. The spectrum in the g ϭ 2 region was deconvoluted into its component spectra by multiple linear regression, as discussed under "Results." Quantification of HbNO was performed by obtaining the double integrated area of the EPR spectrum of a series of HbNO standards (synthesized by mixing oxy-Hb, sodium dithionite, and nitrite and quantified by visible spectrophotometry, ⑀ ϭ 13.0 mM Ϫ1 cm Ϫ1 at 575 nm) (24) taken using identical EPR parameters. Multiple linear regression of spectra using the basis spectra for the three HbNO subtypes (normalized to a double-integrated area of 1, see Fig. 3) allowed the direct determination of the double-integrated area of the three HbNO subspecies in the samples, which could then easily be converted into concentration with reference to standards. Basis spectra for the three spectrally distinct forms of HbNO (6-coordinate ␤, 6-coordinate ␣, and 5-coordinate ␣) were determined from HbNO standards obtained at varying degrees of NO saturation in the presence and absence of inositol hexaphosphate or N-ethylmaleimide, by a subjective iterative process. The spectra obtained are very similar to those published by Hille et al. (25) and Louro et al. (26). The free radical signals at g ϭ 2 were quantified by incorporating the basis spectra for these radicals (again normalized to an area of 1) in the regression analysis, and by converting the area contributions of these species to concentration with reference to a standard curve generated using the standard nitroxide 3-carboxy-2,2,5,5-tetramethyl-1pyrrolidineyloxy at 100 K.
EPR spectra of met-Hb, in the g ϭ 6 region, were collected using the following conditions: microwave power 2 milliwatts; time constant of 20.5 ms; modulation frequency 100 kHz; modulation amplitude 5G; scan time 42 s; scan width 500 G; and temperature 4.5 K. Quantification of met-Hb was obtained by correlating the sample spectra with that of a standard. This is simply a nonsubjective method of measuring peak height. A standard curve was obtained using authentic met-Hb, generated from the oxidation of oxy-Hb with potassium ferricyanide. The concentration of met-Hb was measured by spectrophotometry (⑀ 630 ϭ 3.6 mM Ϫ1 cm Ϫ1 at pH 7.4). The slope of the correlation plot (multiplied by the concentration of the standard) is a direct readout of the concentration of met-Hb in the sample, and the correlation coefficient (R 2 ) gives a measure of the underlying assumption that the spectral shape is unchanging. In no case in this study did R 2 fall below 0.99, indicating the spectral deviations between samples and standard were small.

RESULTS
Standardization and Characterization of HbNO Analysis-In order to determine the linearity, sensitivity, and reproducibility of the EPR technique for the detection of HbNO, we generated standard curves from synthesized HbNO. Raw data from such a sample set is shown in Fig. 1A (black lines). The area under the absorption spectrum (obtained by integrating the first derivative EPR spectrum twice) is directly pro- A, EPR spectra of HbNO standards (black line) and reconstructed spectra after regression analysis (gray line). B, the area under the absorption spectrum from the data shown in A, and other similar data, plotted as a function of HbNO concentration (mean Ϯ S.E., n ϭ 3). C, spectra of (i) 5-coordinate-␣-HbNO, (ii) 6-coordinate-␣-HbNO, and (iii) 6-coordinate-␤-HbNO used in the regression analysis. D, residual after regression analysis representing subtraction of the gray from the black lines in A.
portional to concentration, and in Fig. 1B we show this area, plotted as a function of HbNO concentration for three identical experiments. As can be seen the area under the absorption spectrum is a linear function of HbNO concentration.
Although integration is possible for more concentrated standard spectra containing little noise, this approach becomes problematic for low concentrations of HbNO, because of difficulties in integrating noisy spectra, and for spectra that contain multiple overlapping paramagnetic species. Both of these issues are relevant to detecting low concentrations of HbNO in whole blood. In order to address this, we have developed a regression approach using basis spectra for the three major HbNO species. The general approach is to use basis spectra for all component species with absorption spectra area normalized to unity. By this method, the results of the regression analysis map directly to the area and can be easily converted to concentration. The basis spectra for the three major HbNO species (5-coordinate ␣-chain, 6-coordinate ␣-chain, and 6-coordinate ␤-chain) are shown in Fig. 1C. Fits of these spectra to the standards are shown in Fig. 1A (gray lines), with the residuals in Fig. 1D. As can be seen, the basis spectra accurately represent the HbNO spectra at all concentrations. Although these three species may not represent all the spectrally distinct subtypes of HbNO, they encompass the vast majority of spectral intensity at 100 K. For 18 HbNO standards, at a range of concentrations obtained at various receiver gains and accumulated scan number, we normalized the absorption spectrum area to a single scan at a receiver gain of 1. From these we have obtained a factor of 438 Ϯ 22 (mean Ϯ S.D., n ϭ 18) representing the area under the absorption spectrum expected for a single scan (1 average) of a 1 M solution of HbNO at a receiver gain of 1. This number was used to calculate HbNO concentrations in later experiments. It should be noted that this number is highly spectrometer-dependent and should be obtained independently by each laboratory.
To examine non-HbNO EPR-detectable species in blood, we examined packed red blood cells from both arterial and venous sources by EPR. As shown in Fig. 2, the EPR spectrum of arterial blood exhibited a broad underlying feature that can be attributed to the presence of copper(II) proteins, such as superoxide dismutase and residual cerulloplasmin, and also an unidentified sharp single line free radical species. In venous blood (Fig. 2), an additional broader free radical species was observed that was more easily examined after subtraction of the arterial sample from the venous sample (Fig. 2). The broader signal was spec-trally similar to the free radical species observed upon addition of hydrogen peroxide to met-Hb and was tentatively identified as a hemoglobin protein radical. The presence of this species has been observed previously in venous blood (27). Based on these observations, for the quantification of HbNO in whole blood, we incorporated three additional basis spectra in the regression analysis, both the broad and sharp radical signals and the copper(II) species. Radical concentration was determined with reference to a 3-carboxy-2,2,5,5-tetramethyl-1-pyrrolidineyloxy standard.
HbNO Detection by EPR in Whole Blood-In order to test the applicability of this method to whole blood, we prepared HbNO in whole blood by exposing deoxygenated blood to the nitric oxide donor PROLI/NO and then serially diluted this sample with whole blood. A comparison between the expected concentration and the measured concentration of HbNO is shown in Fig. 3. As can be seen the area of the EPR spectrum of HbNO is a linear function of HbNO concentration in whole blood, and this method can accurately determine HbNO concentrations in whole blood to submicromolar levels. We estimated the level of detection to be about 200 nM in whole blood. No HbNO was detected in freshly drawn blood (either arterial or venous) indicating that basal levels of HbNO are less than ϳ200 nM.
Detection of HbNO during NO Inhalation-To examine if the level of HbNO could be increased to detectable levels by NO inhalation, we exposed three volunteers to 80 ppm NO inhalation for 4 h, followed by 2 h of recovery. At frequent intervals during the procedure, both arterial and venous blood samples were drawn and centrifuged for 5 min, and the packed red cells were placed in a quartz EPR tube and immediately frozen for later EPR analysis. All samples were analyzed by regression analysis using the six basis spectra described above.
Upon NO inhalation, additional spectral intensity was observed in the g ϭ 2 region (cf. Fig. 4 (left panel) with Fig. 2), suggestive of the formation of HbNO in both arterial and venous samples. Fig. 4 (right panel) shows the spectra from Fig. 4 (left panel) after subtraction of the radical and copper(II) spectra (gray line). Also shown is the fit of the HbNO components to the data. The majority of the newly formed signal can be explained by the formation of HbNO, and in both the venous and arterial sample, the three HbNO species give a good fit to the data. Most . EPR spectra of packed erythrocytes in the g ‫؍‬ 2 region before commencement of NO inhalation. Packed red cells from arterial and venous blood were analyzed by EPR spectrometry at 100 K. Also shown is a subtraction of the arterial signal from the venous signal, and a signal obtained after treatment of met-Hb (500 M) with hydrogen peroxide (2 mM) (shown in much reduced scale). FIGURE 3. Measurement of HbNO in whole blood. Red blood cell HbNO was prepared by the addition of PROLI/NO to whole blood, and HbNO concentration was determined by spectrophotometry. This preparation was serially diluted with whole blood, and the EPR spectrum was taken. EPR spectra were analyzed by regression analysis. ࡗ, concentration of HbNO expected from the dilution factor. f, concentration of Hb calculated from regression analysis. interestingly, the shape of the signal is distinctly different in the venous and arterial samples, indicating that a change in contribution of the three underlying species is occurring during artery to venous transit.
Quantitative analysis of the free radical component of the signal (TABLE ONE) indicates that for all samples analyzed, the sharp radical signal (see Fig. 2) remains relatively unchanged during the artery to venous transit, whereas the broad radical increases about 5-fold. NO inhalation did not affect the increase in the broad radical signal during artery to vein transit.
Quantification of HbNO formation using the factor described above results in the time course shown in Fig. 5A. In all three subjects, the level of HbNO present in both arterial and venous samples at base line was below the detectable level by EPR. Upon NO inhalation, HbNO levels rose to between 0.5 and 2.5 M (depending on the subject) and sustained this level for the period of NO exposure. Most interestingly, in all but one data point, the arterial HbNO levels were higher than the venous HbNO levels. Upon cessation of NO inhalation, the HbNO levels decreased to base line over the next 2 h with an approximate half-time of 43 Ϯ 14 min (mean Ϯ S.D., n ϭ 6 from artery and vein samples from three experiments).
The subdivision of the HbNO spectra obtained during NO inhalation into the component spectra is shown in Fig. 5B. The major change observed during artery to venous transit is an increase in the concentration of 5-coordinate-␣ HbNO at the expense of 6-coordinate-␣ HbNO. This is a typical response during the deoxygenation of partially nitrosylated hemoglobin and has been linked to the R to T transition of this protein. The arterial met-Hb levels detected before and during NO exposure were 23.5 Ϯ 11.45 M (mean Ϯ S.D., n ϭ 4), or 0.12%, before NO inhalation and 165.2 Ϯ 25.8 M (mean Ϯ S.D., n ϭ 12), or 0.83%, during NO inhalation.

DISCUSSION
In this study we have examined the effects of NO inhalation on EPR-detectable species in whole blood taken from both artery and vein. Before NO inhalation, an increase in a broad free radical species was observed upon artery to vein transit, similar to the radical species observed during oxidation of hemoglobin by either hydrogen peroxide or nitrite (28) and previously observed in venous blood (27). Balagopalakrishna et al. (29) observed the formation of a radical signal at g ϭ 2.004 after freezing of partially deoxygenated hemoglo-bin and assigned this signal to a superoxide radical bound within the heme pocket. It is possible that the radicals we observed in both arterial and venous blood may be related to this phenomenon. However, it is possible that the broad radical occurs as a result of oxidative processes within the red blood cells to generate protein radical species. The presence of NO inhalation did not affect the formation of the broad radical signal, suggesting that any such oxidative processes are unaffected by NO.
Although the EPR spectrum in the g ϭ 2 region consists of multiple overlapping species, these can be isolated by multiple linear regression analysis if pure spectra of the individual components can be obtained. We have identified the following six species in the g ϭ 2 region: two free radical species, a copper protein signal, and the three signals associated with HbNO. We have demonstrated by using extensive standardization and by serial dilution in whole blood that regression analysis allows the  quantification of levels of HbNO down to about 200 nM. HbNO was not detected in any of the three subjects before NO inhalation, suggesting that under normal conditions the level of HbNO is below 200 nM. Although it has been realized previously that it is difficult to detect basal HbNO by EPR (9), analytical measurements have revealed highly variable levels. McMahon et al. (22) reported levels of 0.0005-0.002 HbNO/hemoglobin, by photolysis chemiluminescence measurements, corresponding to ϳ1.25-5 M in whole blood. In contrast Gladwin et al. (17) reported levels between 0.00002 and 0.00004 HbNO/heme by triiodide-based chemiluminescence, corresponding to about 0.2-0.4 M in whole blood. Rassaf et al. (2), using similar tri-iodide-based methods failed to detect HbNO in humans and put basal levels at Ͻ1 nM. Datta et al. (30) measured levels of HbNO ranging between 2.1 and 9 M whole blood concentrations using a tri-iodide/ferricyanide electrochemical method. However, this same group, using tri-iodide based chemiluminescence methodology, recently reported revised levels of 500 nM total nitrogen oxides (31).
Kirima et al. (32) detected about 8 M HbNO under basal conditions in rats, appearing at first glance to support the levels observed by McMahon et al. (22) in humans; however, the method of standardizing this signal involved the addition of saturated NO solution to venous blood, assuming all the added NO will become HbNO. The fact that a significant portion of venous hemoglobin is oxygenated (and so will destroy NO) and that it has been detailed that the addition of saturated NO gas solutions to heme proteins will result in the binding of only a fraction of the NO (because of reactions of NO before mixing is complete) (14), all suggest that the quantification of basal HbNO was problematic in this study. In the pig, Aldini et al. (33) reported basal HbNO levels below the EPR detection limit, which they put at 250 nM, similar to that reported here. Although Takahashi et al. (34) reported basal levels of HbNO of ϳ2 M in sheep, they quantified this level by double integration of the entire g ϭ 2 region and made no correction for the presence of the copper(II) signal, which is by far the predominant signal, and no assessment of basal levels of HbNO can be gleaned from this study. If EPR is indeed to be regarded as a gold standard for HbNO detection, as has been indicated previously (21), then clearly photolysis chemiluminescence techniques overestimate HbNO by at least an order of magnitude.
Upon inhalation of NO, the level of HbNO rapidly rose to between 500 and 2200 nM, depending on the subject, in packed red cells and maintained this level during the 4 h of inhalation. This level corresponds to about 0.0025-0.011% NO/heme, in reasonable agreement with Gladwin et al. (17) using tri-iodide-based chemiluminescence techniques. In the only study so far that has directly compared these techniques in an identical experimental paradigm, this indicates that the tri-iodide-based chemiluminescence technique and EPR differ by at most a factor of 2, and not by a factor of 10 -100 as has been claimed in the literature (21). These results clearly undermine editorial suggestions that the tri-iodide assay grossly underestimates NO-modified hemoglobin levels (19,21,23). Very similar EPR data have been reported in a sheep model of NO inhalation (33), where the level of HbNO rapidly rose to steady-state levels between 4 and 6 M over a 4-h duration.
In agreement with Gladwin et al. (17), the venous levels of HbNO were generally lower than the arterial levels (only one data point indicated a higher level). Once NO inhalation is stopped, the HbNO signal decays back to base line with a half-time of about 40 min, somewhat slower than the 15 min observed previously after NO infusion in pigs (33). There is a clear discrepancy between the loss of a significant portion of HbNO on every arterial to venous transit and the measured half-time. If ϳ20% of the HbNO is lost during each circulation of ϳ1 min, then the half-life of HbNO should be closer to 3 min. Several possibilities exist for this. The first possibility is that HbNO is transiently converted to an "EPR-silent" form upon deoxygenation. This is unlikely because this difference was also observed using non-EPR techniques (17), and no other species were observed to form in compensation for this loss. A second possibility is that HbNO is more unstable in venous than in oxygenated blood, and the difference occurs ex vivo before processing. For example, heme autoxidation can occur more readily in venous blood than in fully oxygenated blood, generating oxidants such as hydrogen peroxide. Here the observation of a protein radical in venous but not arterial blood attests to this autoxidation phenomenon, although it is not clear if this occurs in vivo or during sample processing. However, there are no reports that HbNO is more unstable in venous/ deoxygenated blood than arterial blood, and studies indicate the opposite (9). The third possibility is that HbNO formation does not cease after the cessation of NO inhalation and that additional mechanisms of HbNO formation are present. Such mechanisms potentially include the reduction of nitrite (35,36) or nitrite esters such as S-nitrosothiols (37).
Even at the low of HbNO levels observed in this study, it was possible to delineate the spectral changes associated with arterial to venous transition. The major difference observed in the composition of the HbNO spectrum between arterial and venous blood is the formation of 5-coordinate-␣ HbNO at the expense of 6-coordinate-␣ HbNO. This is the expected behavior upon deoxygenation of singly nitrosylated HbNO tetramers and is thought to be a reporter of the R to T transition. Similar observations have been observed previously in the rat after injection with NO donor (38). It has been stated recently that at low concentrations NO binds preferentially to the ␤-chain of hemoglobin in the human circulation (39). Our data are inconsistent with this notion. NO is found preferentially bound to the ␣-chain with at most 20% found on the ␤-chain. In addition there is no change in the relative proportion of ␤-chain HbNO during the arterial to venous transit.
Although the HbNO levels rose to a maximum of 2.5 M during NO inhalation, met-Hb formation rose by over 150 M. It is clear from this that the major reaction that occurs, at the level of hemoglobin, as a result of NO inhalation is heme oxidation, likely as a result of the reductive dioxygenation of NO by HbO 2 . This suggests that the majority of inhaled NO is destroyed and not preserved by its interaction with hemoglobin.
Rogers et al. (31) recently criticized the tri-iodide-based method for the detection of "red blood cell-bound nitric oxide" based on the fact that ferrous hemoglobin can bind NO in the reaction chamber of the chemiluminescence analyzer and reduce the peak amplitude. However, these investigators report very clearly that hemoglobin concentrations as high as 100 M within the chemiluminescence reaction chamber (levels far higher than those achieved in any previous studies) do not affect the area under the peak of the chemiluminescence signal. As it is the area under the peak and not the amplitude that is proportional to amount of NO, this criticism is somewhat perplexing. This study also contains some anecdotal comments concerning the inability to separate cyanide/ferricyanide from hemoglobin using Sephadex G25 size exclusion chromatography and that the use of acid/sulfanilamide or acid alone decreases hemoglobin/NO signals. We are puzzled by these technical difficulties as the Sephadex G25 separation is an inherent step in the Drabkin assay for hemoglobin and has not been seen to be a problem before. In addition, the intensive and controlled investigations of the use of acidic sulfanilamide by us and others (40 -42), to remove nitrite from S-nitrosothiols before detection by chemiluminescence, have never revealed a problem with this procedure. Unfortunately, as no data or experimental details were provided, it is difficult to ascertain under what conditions these problems were observed.
In conclusion this paper reports the kinetics of in vivo changes to HbNO levels upon inhalation of nitric oxide. The major findings are as follows. (a) HbNO can be accurately detected and quantified in whole blood with a detection limit of ϳ200 nM. (b) EPR measurements substantially agree with tri-iodide-based chemiluminescence techniques but differ substantially from photolysis-based chemiluminescence techniques, suggesting that photolysis-based methods are significantly problematic for the accurate detection of HbNO. (c) By using regression analysis, levels of HbNO as low as 0.5-1 M can be deconvoluted into component species. (d) HbNO is present at less than 200 nM at basal conditions and is formed at a level of 0.5-2.5 M upon inhalation of 80 ppm NO. (e) The half-time of HbNO in vivo is ϳ40 min. ( f ) The majority of inhaled NO is destroyed and not preserved by interaction with hemoglobin.