ELSEVIER SCIENCE IRELAND Toxicology 88 (1994) 165-176

TOXiCOlOGY

An impurity present in some samples of SIN-1 oxidizes it to nitric oxide in anaerobic solutions

Ye Wang a, Lori G. Rochelle a, Harriet Kruszyna a, Robert Kruszyna a, Roger P. Smith *a, Dean E. Wilcox b

aDepartment of Pharmacology and Toxicology, Dartmouth Medical School, 7650 Remsen, Room 519, Hanover, NH 03755-3835 USA

bDepartment of Chemistry, Dartmouth College, Hanover. NH 03755-3564, USA

(Received 30 June 1993; accepted 16 August 1993)

Abstract

N-Morpholino-N-nitrosoaminoacetonitrile (SIN-I), a nitrovasodilator metabolite of the drug, molsidomine, is widely used in studies on the pharmacology and toxicology of nitric oxide (NO) because solutions of SIN-1 'spontaneously' release NO in a pathway involving molecular oxygen. Preliminary results, however, suggested that SIN-1 could react with hemo- globin in anaerobic solutions to release NO and form NO-hemoglobin. Electron paramagnetic resonance (EPR) studies showed that heme(III) of methemoglobin was not being reduced, thereby not serving as the oxidant in the reaction generating NO-hemoglobin. When anaero- bic solutions of SIN-I and hemoglobin kept in the light and in the dark were compared, substantially more NO-hemoglobin was eventually generated in the dark, indicating that SIN- 1 did not undergo photochemical decomposition to NO under the conditions used. Solutions of NO-hemoglobin were equally stable under these same conditions of light and dark. The initial pH (7.0) of stirred, unbuffered soluticins of SIN-1 decreased at nearly the same rate whether or not oxygen was present. Anaerobic and aerobic solutions plateaued at the same pH, namely 5.4. Anaerobic solutions of SIN-1 in phosphate buffer, pH 7.4, released NO to the gas phase, where it was identified by trapping it with hemoglobin on agarose beads and deriving the characteristic NO-hemoglobin EPR spectrum. High pressure liquid chromotography revealed the presence of an unknown species with a retention time between that of SIN-I and moisidomine. Samples from two different lots of SIN-1 contained this im- purity which appears to oxidize SIN-1 to products that release NO in the absence of oxygen. This unknown impurity may be unstable toward light.

Key words: Nitric oxide; SIN-l; Molsidomine; Nitrovasodilators; Sydnonimines

* Corresponding author.

0300-483X/94/$07.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0300-483X(93)02683-8

166 K Wang et al./Toxicology 88 (1994) 165-176

I. Introduction

The editorial writers of Science have recently designated nitric oxide (NO) as the 'Molecule of the Year' (Culotta and Koshland, 1992). New biological roles for the endogenously generated mediator are still being discovered even as new explanations for the pharmacological and toxicological actions of exogenous NO-donor xenobiotics are offered. Molsidomine (N-ethoxycar- bonyl-3-morpholinosydnonimine) is a vasodilator, active by virtue of its ability to release NO. It is used clinically in Europe, but it has not yet been approved for use in the United States. The pathway for bioactivation of molsidomine (Fig. 1) to NO begins with hydrolysis mediated by liver ester- ases (Tanayama et al., 1974) resulting in the formation of SIN-I (N- morpholino-N-nitrosoaminoacetonitrile). Subsequently, in aqueous solu- tion, the sydnonimine ring opens in a pH-dependent, base-catalyzed reaction to yield SIN-1A (N-morpholinoiminoacetonitrile). SIN-1A is then believed to react with molecular oxygen to form an unstable intermediate that spon- taneously releases NO (Bohn and Sch6nafinger, 1989).

. .

5. 6.

O N+=N-CH 2-CN +NO ~ O , N - N = C H - C N + H + \ /

Fig. 1. Biotransforraation of molsidomine (compound l) according to Bohn and Sch6nafinger (|986). In vivo enzymatic cleavage of the carboxyethyl group yields SIN-I (compound 2). Under neutral conditions the ring opens to yield SIN- IA (compound 3). Diatomic oxygen is involved in the next step to form a cationic radical (compound 4). NO is released spontaneous-

ly, and the unstable cation (compound 5) deprotonates to S[N-1C (compound 6).

enzyme

\ / I G / - N - c ° 2 E t ' / NH N-O [ N--O

OH-

4. 3.

O/'-"N N 02 /---'N +- N - CH 2- CN ~ O N - N - CH 2- CN I k _ _ / I N N II II O O

Y. Wang et al. / Toxicology 88 (1994) 165-176 167

Its essential role in facilitating the release of NO from SIN-1 is easily overlooked or forgotten because biological experiments with SIN-1 are usually conducted in the presence of molecular oxygen. SIN-1 has become an extremely important experimental compound because of its ability to release NO readily, albeit, not 'spontaneously'. It is widely used for that pur- pose in pharmacological and toxicological studies. Technically, however, it is an oxidation product of SIN-1A that spontaneously releases NO (Fig. 1).

Recently, attention has been drawn to a photochemical decomposition reaction of molsidomine (De Muynck et al., 1992). When exposed to sunlight, solutions of molsidomine in polyvinyl infusion bags showed a 90%- decomposition in 1 h. There was no decomposition when the bags were ex- posed only to fluorescent lighting or when the solutions were protected by UV filters. However, no mention was made of whether or not molecular oxy- gen was present during these experiments. We are not aware of previous studies on the photochemical stability of SIN-1 or SIN-1A, which should both be stable under anaerobic conditions (Bohn and Sch6nafinger, 1989; Feelish et al., 1989). However, we describe here a reaction that results in the release of NO from SIN-1 in the absence of molecular oxygen, and at a rate comparable to that of the reaction in the presence of oxygen. Although preli- minary results suggested that SIN-1 might react with hemoglobin, the reac- tion described appears to be mediated by an as yet uncharacterized impurity found in the samples of SIN-1 supplied to us.

2. Materials and methods

The experimental conditions for work with rabbit aortic strips were previously described (Kruszyna et al., 1982). Human blood specimens, drawn for routine hematologic workup, were pooled and washed 3 times with isotonic saline. The red blood cell (RBC) fraction was diluted with an equal volume of distilled water and lysed by freezing twice in ethanol-dry ice mixtures. The RBC lysate was diluted with 0. I M phosphate buffer, pH 7.4, to a final concentration of 1.3 mM in heme. Ten ml aliquots of the lysate were placed in a 250 ml cylindrical tonometer with a stopcock at each end to permit degassing with nitrogen and a sidearm port sealed with a rubber septum in the middle to permit anaerobic addition of reagents or the removal of aliquots. The rotating tonometer at room temperature was purged for 1 h with a nitrogen stream which was originally passed through a vanadium(II) scrubber to remove any traces of oxygen (Meites and Meites, 1948). Nitric oxide gas (Matheson Gas Products, Inc., Twinsbury, OH) was passed through a solution of KOH to remove any NO2 and collected in another tonometer. The NO-containing tonometer was sealed, and a 0.4-ml aliquot

168 Y. Wang et al./Toxicology 88 (1994) 165-176

of the gas was removed anaerobically with a syringe and transferred to the nitrogen-purged lysate.

In other experiments, freshly prepared aqueous solutions of SIN-1 (see acknowledgements) were added to lysates anaerobically through the rubber septum to a final concentration of 1.6 mM. In experiments with carbon monoxide (commercial grade, Union Carbide Corp., Linde Division, Dan- bury CT), 60 ml of the pure gas was collected anaerobically in a syringe and introduced into the tonometer containing nitrogen-purged lysates. It was necessary to open a stopcock briefly to relieve some of the positive internal pressure. Aliquots were removed at periodic intervals under anaerobic condi- tions via the side arm, placed into nitrogen-purged 3-mm quartz electron paramagnetic resonance (EPR) tubes, and were immediately frozen in liquid nitrogen. The EPR analysis for NO-hemoglobin was performed at 77°K in a dewar (Wilmad) as previously described (Wilcox et al., 1990).

Reduced hemoglobin-agarose bead columns for trapping NO were prepared as described by Wennmalm et al. (1990), and assembled as sum- marized below. The bottom of the glass column containing the agarose beads was attached with Tygon tubing to a stopcock, and the top of the column was attached with tubing to the polycarbonate barrel of a 10 ml syringe. The open end of the syringe barrel was sealed with a rubber septum. Just before use, the column was purged with oxygen-free nitrogen by puncturing the rub- ber septum with a hypodermic needle attached with tubing to the nitrogen gas stream. Nitrogen was purged in a tonometer with 10 ml of phosphate buffer (as above) without lysate. Then the top of each column was placed in line with the tonometer through tubing attached to a needle. The needle penetrated the septum covering the syringe barrel above the hemoglobin- agarose bead layer in the column. Nitrogen gas flow through the tonometer was adjusted to 20 ml/min. Nitrogen, passed for 20 min over the buffer and through the hemoglobin-agarose column, served as the control. Freshly prepared aqueous solutions of SIN-1 were then introduced anaerobically into the tonometer and mixed. Nitrogen purging through a replacement hemoglobin-agarose column was continued for 20 min, and then through a third fresh column for 40 min. The columns were removed from the in-line system, and immediately frozen in liquid nitrogen. EPR analysis of the hemoglobin-agarose columns was as described above except that the col- umns were placed in a Bruker variable temperature liquid nitrogen apparatus instead of the 77°K dewar, and the analyses were conducted at 135°K (Kruszyna et al., 1987).

The HPLC analyses were carried out with a Rainin Instrument Co. Microsorb column with 5 mm C-18 resin particles (25 cm × 4.6 mm i.d.) with all other conditions as previously reported (Feelish et al., 1989). The mobile phase consisted of aqueous ammonium acetate buffer, pH 3.1, acetonitrile

Y. Wang et al. / Toxicology 88 (1994) 165-176

Z o CO Z ILl t--

-']1 gm

t, t3 10 15

MINUTES

169

Fig. 2. Rabbit aortic strips with intact endothelial cell layers were suspended in an aerated organ bath. Both strips were contracted at '1' by the addition of 1 x 10 -7 M norepinephrine. In the top tracing, nitroprusside at 1 x 10 -5 M was added at '2'. In the bottom tracing,

1 × 10 -6 M of freshly prepared SIN-I was added at '3'.

3000

~=L 2000 .

1000 '

0

300 •

A

1oc •

• , o

10 20 30 40 50 60 70

M INUTES

B

• , . , • , - , . , . , • ,

10 20 30 40 50 60 70

M INUTES

Fig. 3. In Fig. 3A NO gas was introduced into a tonometer at time = 0. The tonometer contained either a solution of deoxyhemoglobin under nitrogen (O) or a solution of carboxy- hemoglobin under a mixture of carbon monoxide and nitrogen (e). In 3B the experiments were repeated except that SIN-1 was introduced at time zero, into a solution of deoxy- hemoglobin ~ or a solution of carboxyhemoglobin (1~. In each case, the points represent the

mean of 4 replications and the vertical bars show the SE.

170 Y. Wang et al. / Toxicology 88 (1994) 165-176

and methanol (93:6:2) at a flow rate of 1 ml/min. The detection wavelength was set at 290 nm.

3. Results

In Fig. 2 the responses ofprecontracted rabbit aortic strips in aerated solu- tions to sodium nitroprusside and to SIN-1 are contrasted. Like nitroglycerin (not shown), the addition of nitroprusside results in an immediate response, whereas the addition of SIN-1 produces a delayed relaxation. These observa- tions suggest that nitroprusside reacts extracellularly with a component of the vascular wall to release NO, which immediately exerts its biological effect. In contrast, the delayed response to SIN-1 could be due to its extra- cellular decomposition to NO (Fig. 1) in the aerated incubation medium (Tanayama et al., 1974; Culotta and Koshland, 1992), and NO diffusion through the endothelial cell layer into the smooth muscle to effect relaxation.

Fig. 3A shows the rate of NO-hemoglobin formation when NO gas was introduced into a tonometer containing either a solution of deoxyhemo- globin under nitrogen or a solution of carboxyhemoglobin under a mixture of nitrogen and carbon monoxide. The formation of NO-hemoglobin from deoxyhemoglobin represents simply the binding of NO to unoccupied heme(II), and complex formation was virtually complete 5 min, or less, after the introduction of NO gas. In the case of carboxyhemoglobin, however, the NO had to compete with CO already bound to heme(II), which greatly slow- ed the rate of formation of NO-hemoglobin. The final concentration of NO- hemoglobin reflects both the relative partial pressures of the two gaseous ligands and their relative affinities for heme(II). In Fig. 3B, the same experi- ment has been repeated using SIN-1 instead of NO. The rates of formation of NO-hemoglobin are very much slower with SIN-1 compared to NO, but clearly NO is being released from SIN-1 in the absence of molecular oxygen. These results suggest that SIN-1 reacted with hemoglobin to result in the release of NO, which was then trapped by heme(II) groups.

Heme(III) also can be quantified by EPR spectroscopy (Kosaka et al., 1981) by virtue of its signal at g = 6.0 (Fig. 4). Accordingly, lysates contain- ing either 7% autoxidized heme(III) or enriched to 59% in heme(III) with ferricyanide-oxidized hemoglobin were prepared and these were incubated with SIN-1. Aliquots were withdrawn at periodic intervals and the amount of heme(III) was determined by its signal at g = 6 and NO-hemoglobin was quantified in the same aliquot by its signal at g = 2.02. The NO-hemoglobin signal increased with time as shown in Fig. 4B and 4D, but there was no change in the g = 6 signal with time (Fig. 4A and 4C), whether the original mixture contained 7% or 59% heme(III). These results indicate that SIN-1 does not react with heme(II or III). The three line hyperfine features are

Y. Wang et al. / Toxicology 88 (1994) 165-176 171

A l a

C

l b

ks

§

D

2 a

l a l b 2 a 2 b . Aj

Fig. 4. The changes with time of exposure of hemoglobin to SIN-1 in forming NO-hemoglobin (Panels B and D) are compared with the lack of change in heme(III) (Panels A and C). In Panel A, 'la' designates the g = 6 EPR signal for heme(III), which is 59% of the total heine (predetermined spectrophotometrically), 5 min after the introduction of SIN-1. The g = 6 sig- nal 60 rain after adding SIN-1 is shown by 'lb'. The ordinate is 35 x that in Panel C. In Panel B, '2a' designates the g = 2 EPR signal for NO-hemoglobin 5 min after adding SIN-1 and '2b' shows the g = 2 signal 60 min after adding SIN-I in the same frozen sample as in Panel A. In Panel C, the designations 'la' and 'lb' are as in Panel A, but the heme(III) is only 7% of the total. In Panel D, the designations '2a' and '2b' are as in Panel B and the ordinates are on the same scale. Four replications of this experiment at a heme(III) concentration of 14% of the total gave the same results, namely no change in the g -- 6 heme(III) signal, and a

dramatic increase in the g = 2 NO-hemoglobin signal.

much more dramat ic in Fig. 4D than in 4B indicat ing that a larger f ract ion o f the heme(II) was present as the species (NOot2+/33+)2 in 4D than in 4B; this may reflect differences in p roduc ts as a result o f au tox ida t ion compared to ferr icyanide-mediated oxida t ion (Kruszyna et al., 1988). The results o f others (Perrella et al., 1993) suggest tha t 4B and 4D should be reversed. We cannot explain this discrepancy, but there can be no d o u b t tha t b o th 4B and

4D represent forms o f NO-hemoglob in . It seemed possible tha t SIN-1 or SIN-1A might par t ic ipate in a pho to -

chemical react ion to l iberate N O because mols idomine undergoes pho to -

chemical decompos i t ion (De M u y n c k et al., 1992). Accordingly , lysates were incubated with SIN-1 under n i t rogen and ei ther ambien t l abo ra to ry condi- t ions o f i l luminat ion or inside a double , black, plastic, film bag. As shown

8O0

600

400

200

0

Y. Wang et al./ Toxicology 88 (1994) 165-176 172

0 100 200

MINUTES

Fig. 5. Identical lysates under nitrogen were incubated with SIN-i in a tonometer exposed to ordinary laboratory illumination (O) or in a double, black, plastic, film bag (O).

in Fig. 5 there was an effect o f ambien t l ab o ra to ry i l luminat ion in tha t more NO-hemog lob in was fo rmed in the da rk than in the light; however , this result was the opposi te o f tha t ant ic ipa ted for a l ight-facil i tated decompos i t ion o f

SIN-1 to NO. As a control , a l iquots o f a solut ion o f N O - h e m o g l o b i n were incubated for 3 h in the da rk or in the light, and no difference in their N O -

7

"I- Q.

6

5 ~ 0 100 200 300

MINUTES Fig. 6. Changes in pH with time when identical 1.6 mM solutions of SIN-1 in distilled water

were bubbled with air (O) or with nitrogen (O).

K Wang et al./Toxicology 88 (1994) 165-176 173

hemoglobin content was observed. Thus, light did not facilitate the decom- position of NO-hemoglobin under these conditions.

However, SIN-1 might react with some functional group on hemoglobin, other than the heme iron, which would not be detected by EPR. We, therefore, exploited the fact that the oxidative decomposition of SIN-1 results in the release of protons (Fig. 1). Aerated solutions of SIN-1 become progressively more acidic until an equilibrium is reached between the rate of the base-catalyzed ring-opening step and the proton-releasing, oxidation step (Bohn and Sch6nafinger, 1989). As shown in Fig. 6, identical solutions of SIN-1 in distilled water became acidic whether they were bubbled with air or with nitrogen. The sample incubated under air decayed with an initial rate about 3 times faster than the sample under nitrogen, but they both reached the same equilibrium.

To show that this anaerobic decomposition reaction of SIN-1 released NO in the absence of hemoglobin, 1.6 mM SIN-1 in phosphate buffer was placed in a deaerated tonometer, and a nitrogen stream directed over the surface of the solution was passed into a column containing hemoglobin on agarose beads (see Materials and methods). The hemoglobin-agarose column was then frozen and screened by EPR; it gave a typical spectrum for NO- hemoglobin (not shown, but the same as in Fig. 4B).

I O B ~

900 '

0 O O '

7 0 0 "

600"

5 0 0

4 ~ O "

30~1'

2OE1"

l O B "

O O

S I lq 1

.| . . . , , . . . .

iia

Ti m e

PIO L

. . . . L • " • - .

20 qO

(rnin.)

41B

Fig. 7. HPLC traces of a solution of SIN-1 and of a solution of molsidomine which have been superimposed for convenience. From published retention times (Bohn and Schbnafinger, 1986) the peaks, from left to right, are inferred to be SIN-l, unknown impurity and

molsidomine.

174 K Wang et al. / Toxicology 88 (1994) 165-176

At this point, the only remaining possibility was that the samples of SIN-1 supplied to us contained an impurity that oxidizes SIN-1A in solution resulting in the release of NO in the absence of molecular oxygen. Fig. 7 shows the results of a composite HPLC analysis of solutions prepared from samples of SIN-1 and of molsidomine. The former shows an unknown peak, presumably representing the impurity with a retention time between that of SIN-1 and molsidomine, which correspond to their published retention times (Feelish et al., 1989); the retention time of the unknown peak in the SIN-1 solution appears to rule out the impurity as being SIN-1A or SIN-1C (N- morpholinoiminoacetonitrile), which have retention times of 8.6 and 10.6 min, respectively (Feelish et al., 1989).

4. Discussion

The results in Fig. 3B suggest that either SIN-1 or SIN-IA is capable of reacting with hemoglobin under anaerobic conditions to release NO. Indeed, all NO-vasodilator compounds that we have tested undergo such a reaction (Kruszyna et al., 1987). However, all of the other compounds are reduced to liberate NO, concomitant with the oxidation of heme(II) to heme(III). The proposed mechanism for the liberation of NO from SIN-1A (Bohn and Sch6nafinger, 1989; Feelish et al., 1989) involves an oxidation step. Potas- sium ferricyanide, for example, oxidizes SIN-1 in solution (Tanayama et al., 1974). If hemoglobin is involved in the decomposition of SIN-1A, heme(III) must be reduced to heme(II). During preparation of our lysates some heme(II) necessarily undergoes autoxidation to heme(III), but lysates were discarded if more than 10% of the total heme was oxidized. The results in Fig. 4 show that whether the amount of heme(III) was low as a result of autoxidation or whether it was high as a result of enrichment with ferri- cyanide-oxidized heme(III), the formation of NO-hemoglobin from SIN-1 was not accompanied by reduction of the heme iron. Fig. 4, however, does not rule out an oxidative reaction of SIN- 1 with some other part of the hemo- globin molecule such as functional groups on the globin chain.

In order to eliminate any possibility that a reaction between SIN-1 and he- moglobin results in the release of NO, the breakdown of SIN-1 in distilled water to release protons was followed over time (Fig. 6). The results show that the decomposition of SIN-1 in the absence of hemoglobin, or any other added material, occurred at nearly equivalent rates under both aerobic and anaerobic conditions. Both reactions arrived at the same equilibrium value of pH. Our results with columns of hemoglobin to trap released NO from a nitrogen stream passed over the surface of a solution of SIN-1 proved that the proton-releasing reaction is also involved in the release of NO.

The results in Fig. 5 which tested for a photochemical influence on the

Y. Wang et al./Toxicology 88 (1994) 165-176 175

reaction were quite unexpected. They suggest that light does not facilitate the decomposition of SIN-1 to NO. However, light may catalyze the decomposi- tion of the impurity that appears to be responsible for the oxidative breakdown of SIN-1. Since molsidomine is known to be unstable in daylight (De Muynck et al., 1992), it is suspect as the impurity, but that possibility is ruled out by the results in Fig. 7. The impurity has a very different reten- tion time from that of molsidomine. The HPLC peak area of the impurity (Fig. 7) is quite small, but its extinction coefficient at 290 nm is unknown, so its concentration in the mixture cannot be estimated. In view of the results in Figs. 3, 5 and 6, it is possible that the impurity is present in a substantial concentration.

The rate of NO-hemoglobin formation from deoxyhemoglobin is very much slower, i.e., more than 100-fold, with SIN-1 than with NO (Fig. 3A and 3B), and this difference represents a method for estimating indirectly the rate of decomposition of these SIN-1 samples under anaerobic conditions. We conclude that our samples of SIN-1 contain substantial amounts of an im- purity that oxidizes it to NO in anaerobic solutions, and that neither SIN-1 nor SIN-1A react directly with hemoglobin. Further studies are underway to identify this impurity which may adversely affect pharmacological or toxico- logical studies of SIN-1 and NO.

5. Acknowledgments

This research was supported by NIH Grant HL 14127 from the National Heart, Lung and Blood Institute (RPS) and an award from the Air Force Of- fice of Scientific Research sponsored by the Society of Toxicology (LGR). The Bruker ESP-300 spectrometer was purchased with NSF Grant CHE- 8701406. Taken from a thesis submitted by Y. W. to Dartmouth College in partial fulfillment of the requirements for the MS degree. We thank Karen E. Wetterhahn and Susan C. Rossi of the Dartmouth College Department of Chemistry for assistance in the HPLC analyses. The samples of SIN-1 from lots RC 3704 and RC 5573 were a gift from Hoechst-Roussel Phar- maceuticals, Inc., Somerville, NJ.

6. References

Bohn, H. and Sch6nafinger, K. (1989) Oxygen and oxidation prompt the release of nitric oxide from sydnonimines. J. Cardiovasc. Pharmacol. 14, Suppl. I1, $6-S12.

Culotta, E. and Koshland, D.E., Jr. (1992) Molecule of the year. NO news is good news. Sci- ence 258, 1862.

De Muynck, C., Vandenbossche, G.M.R., Remon, J.P. and Colardyn, F. (1992) Rapid breakdown during exposure to daylight of molsidomine when administered by continuous intravenous infusion. Am. J. Cardiol. 70, 836-837.

176 Y. Wang et al./Toxicology 88 (1994) 165-176

Feelish, M., Ostrowski, J. and Noack, E. (1989) On the mechanism of NO release from syd- nonimines. J. Cardiovasc. Pharmacol. 14, Suppl. 11, S13-$22.

Kosaka, H., Imaizumi, K. and Tyuma, I. (1981) Mechanism of autocatalytic oxidation of ox- yhemoglobin by nitrite, an intermediate detected by electron spin resonance. Biochim. Biophys. Acta 702, 237-241.

Kruszyna, H., Kruszyna, R. and Smith, R.P. (1982) Nitroprusside increases cyclic guanylate monophosphate concentrations during relaxation of rabbit aortic strips and both effects are antagonized by cyanide. Anesthesiology 57, 303-308.

Kruszyna, H., Kruszyna, R., Smith, R.P. and Wilcox, D.E. (1987) Red blood cells generate nitric oxide from directly acting, nitrogenous vasodilators. Toxicol. Appl. Pharmacol. 91, 429-438.

Kruszyna, H., Kruszyna, R., Smith., R.P. and Wilcox, D.E. (1988) Generation of valency hy- brids and nitrosylated species of hemoglobin in mice by nitric oxide vasodilators. Toxicol. Appl. Pharmacol. 94, 458-465.

Meites, L. and Meites, T. (1948) Removal of oxygen from gas streams. Anal. Chem. 20, 984-985.

Perrella, M., Shrager, R.I., Ripamonti, M., Manfredi, G., Berger, R.L. and Rossi-Bernardi, L. (1993) Mechanism of the oxidation reaction of deoxyhemoglobin as studied by isolation of the intermediates suggests tertiary structure dependent cooperativity. Biochemistry 32, 5233-5238.

Tanayama, S., Nakai, Y., Fujita, T. and Suzuoki, Z. (1974) Biotransformation of molsido- mine (N-ethoxycarbonyl-morpholinosydnonimine), a new anti-anginal agent, in rats. Xenobiotica 4, 175-191.

Wennmalm, A., Lanne, B. and Petersson, A.-S. (1990) Detection of endothelial-derived relax- ing factor in human plasma in the basal state and following ischemia using electron para- magnetic resonance spectroscopy. Anal. Biochem. 187, 359-363.

Wilcox, D.E., Kruszyna, H., Kruszyna, R. and Smith, R.P. (1990) Effect of cyanide on the reaction of nitroprusside with hemoglobin. Relevance to cyanide interference with the bio- logical activity of nitroprusside. Chem. Res. Toxicol. 3, 71-76.