Rates of release of nitric oxide from HbSNO and internal electron transfer

Article abstract of DOI:10.1016/S0045-2068(02)00509-6

The discovery that hemoglobin (Hb) in erythrocytes contains a fraction of β-Cys-93 thiols as the nitrosylated derivative (HbSNO) led to the suggestion that this species is involved in transporting and releasing nitric oxide, which is the signal for local vasodilation. The release of NO from HbSNO requires an electron transfer to facilitate release and to regenerate the cysteine thiol via one-electron reduction in the absence of added thiols. An alternative mechanism, which has received much attention, transfers the nitrosyl group to an external thiol, which in turn would have to be reduced. The observed first order rate constant for the spontaneous oxidation of the ferrous heme of deoxy HbSNO is 1.0×10^-4s^-1 in the absence of thiols. Under the same conditions, native Hb is stable. The oxidation of HbSNO occurs with the same rate constant that can be derived for the rate reported for the formation of HbNO from HbSNO. These similarities suggest that both processes involve the same reaction: internal electron transfer and direct release of nitric oxide.

Full text of DOI:10.1016/S0045-2068(02)00509-6

BIOORGANIC
CHEMISTRY
Bioorganic Chemistry 31 (2003) 3–10
www.elsevier.com/locate/bioorg
Rates of release of nitric oxide from HbSNO
q
and internal electron transfer
Noam J. Ship, John Paul Pezacki, and Ronald Kluger *
Department of Chemistry, Davenport Chemistry Research Building, University of Toronto, Toronto,
ON, Canada M5S 3H6
Received 7 March 2002
Abstract
The discovery that hemoglobin (Hb) in erythrocytes contains a fraction of b-Cys-93 thiols
as the nitrosylated derivative (HbSNO) led to the suggestion that this species is involved in
transporting and releasing nitric oxide, which is the signal for local vasodilation. The release
of NO from HbSNO requires an electron transfer to facilitate release and to regenerate the
cysteine thiol via one-electron reduction in the absence of added thiols. An alternative mech-
anism, which has received much attention, transfers the nitrosyl group to an external thiol,
which in turn would have to be reduced. The observed first order rate constant for the spon-
taneous oxidation of the ferrous heme of deoxy HbSNO is 1:0 ꢀ 10ꢁ
4
s in the absence of thi-
ꢁ1
ols. Under the same conditions, native Hb is stable. The oxidation of HbSNO occurs with the
same rate constant that can be derived for the rate reported for the formation of HbNO from
HbSNO. These similarities suggest that both processes involve the same reaction: internal elec-
tron transfer and direct release of nitric oxide.
Ó 2003 Elsevier Science (USA). All rights reserved.
1
. Introduction
Hemoglobin is the major constituent of red blood cells and is responsible for
transporting oxygen [1]. It has a tetrameric quaternary structure, with each subunit
q
This paper was intended for inclusion in the issues dedicated to Frank H. Westheimer on the
occasion of his 90th birthday.
*
Corresponding author.
E-mail address: rkluger@chem.utoronto.ca (R. Kluger).
0
045-2068/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0045-2068(02)00509-6

4
N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
having a ferrous heme group that reversibly binds oxygen. The four globin chains of
hemoglobin occur as a pair of ab dimers.
Nitric oxide has been shown to provide a signal to the blood vessel endothelia
that causes the surrounding smooth muscle to relax [2]. Stamler and co-workers
have reported that a small amount of hemoglobin in red cells is nitrosylated
(
HbSNO) specifically, the thiol of one b-Cys-93 is nitrosylated [3,4]. They suggest
that HbSNO is a circulating source of nitric oxide that promotes vasodilation in re-
sponse to reduced oxygen concentration. HbSNO is formed when oxyhemoglobin
reacts with endogenous nitrosylated thiols, such as glutathione S–NO [4]. In the lab-
oratory, HbSNO is synthesized by reacting Hb with S-nitroso N-acetylpenicillamine
(
SNAP) [5,6].
NO is readily transferred to Hb from R–SNO compounds, such as S-nitroso-glu-
tathione (GSNO), in the oxygenated state [7]. The rate of transfer of NO from
HbSNO to GSH increases in the deoxy state [4], and this has been assumed to be
associated with changes in steric effects [8]. The release of NO from HbSNO has been
observed in the absence of added thiols at low oxygen pressures. The process was as-
sumed not to be related to any changes in the state of heme irons [4]. Transfer of NO
to thiols has been assumed to be the major pathway for NO release but has only been
shown to be faster than the release of free NO when there is a 100-fold excess of GSH
over HbSNO [4] (Scheme 1).
In this report we provide evidence for release of nitric oxide occurring as a result
of an internal electron transfer from the heme to the electrochemically labile S–N
Scheme 1.
Scheme 2.

N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
5
bond [9]. This process results in release of NO and an oxidation of the iron in
the adjacent b heme [10], Scheme 2. When S-nitroso thiols are added to deoxygen-
ated hemoglobin, the hemoglobin is partially oxidized and NO gas is released,
Eq. (1) [8].
2þ
þ
3þ
Å
Hb Fe þ RSNO þ H ! Hb Fe þ GSH þ NO
ð1Þ
The rate of release of nitric oxide from HbSNO by an internal electron transfer
mechanism should be first order
Rate ¼ kobs½HbSNO FeðIIÞꢂ:
Our observation that release of NO from HbSNO leads to more rapid oxidation
ð2Þ
at the b heme than at the a supports the second mechanism [10].
Yamamoto [11] reported that the signals from the protons on the heme methyl
1
groups of hemoglobin undergo a paramagnetic shift in the H-NMR when the fer-
rous heme is oxidized and coordinated to a high field ligand, such as azide ion. This
phenomenon results in resolution of the signals from the a and the b ferric methyls in
a region distinct from the rest of the protein (d15–d30). Furthermore, it is known
that in a mixture of native ferrous–heme hemoglobin and hemoglobin with oxidized
dimers, there is a slow equilibration of oxidation states among all the hemes of all
subunits in the solution [12].
HbSNO (with both b subunits nitrosylated), upon spontaneous loss of its NO
groups, undergoes a spectral change that indicates apparent formation of ‘‘half-met-
hemoglobin’’ with half the hemes containing iron in the ferric state [10]. The ob-
served enhanced oxidation of the b hemes [10], should result from the proximity
of the b hemes to the SNO groups at b-Cys-93. The eventual distribution of oxida-
tion of the a and b iron centers will result from the subsequent equilibration of elec-
trons among the subunits [12]. A kinetic analysis of heme oxidation during de-
nitrosylation (in the absence of thiol acceptors and metal contaminants) is necessary
for testing the proposed mechanism of the NO-release reaction. Thus, we measured
the rate of metHb formation from deoxygenated HbSNO. We compare it to the rate
of conversion of HbSNO to Hb derived from data reported by Stamler and cowork-
ers [3,4]. The similar rates for selective heme oxidation and release of nitric oxide are
consistent with a mechanism in which NO release is triggered by an internal electron
transfer.
2
. Materials and methods
Commercial reagents were used without further purification. Water was doubly
distilled and de-ionized. Purified carbonmonoxy human hemoglobin A was obtained
1
from Hemosol (9.6 g/dL). H-NMR spectra were taken at 500 MHz for proteins and
3
00 MHz for all other samples. UV/Vis spectroscopy was performed in quartz cells in
a temperature-controlled sample compartment with gas control ports and an oxygen
sensor. Protein samples were concentrated using centrifugation tubes or stirred-cell
ultra-filtration, using a YM-10 membrane.

6
N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
2
.1. S-nitroso N-acetylpenicillamine (SNAP)
ꢁ
4
Sodium nitrite (0.10 g, 14 ꢀ 10 mol) was added to N-acetylpenicillamine (0.10 g,
ꢀ 10 mol) in 5 mL water, acidified with one drop of concentrated hydrochloric
ꢁ
4
5
acid in a 10 mL round bottom flask covered with foil. The mixture was stored at
4
°C for 10 min and then added to a foil covered G-25 Sephadex (2:5 ꢀ 15 cm) col-
umn equilibrated with water to separate the SNAP from excess inorganic nitrite.
The green (SNAP) fraction was collected and lyophilized (0.092 g, 90%).
2
.2. Nitrosylated Hb (HbSNO)
This synthesis was scaled-up based on previous work [6]. A solution of carbon-
monoxy hemoglobin A (10.0 mL, 1:5 ꢀ 10ꢁ mol) was passed through a column con-
taining Sephadex G-25 (5 ꢀ 15 cm) equilibrated with 0.05 M sodium borate, pH 9.0,
and 0.005 M EDTA to chelate any metal contaminants. The eluted protein was con-
verted into oxyhemoglobin by irradiation under flowing oxygen for 2 h at 0 °C. The
5
ꢁ
4
hemoglobin was reacted with SNAP (0.09 g, 4 ꢀ 10 mol) at 4 °C for 15 min in an
opaque (foil-wrapped) round bottom flask. The entire sample was passed through
a foil-wrapped column of G-25 Sephadex (phosphate buffer, pH 7.4, ionic strength
0
.1 M, 5 ꢀ 15 cm) and the hemoglobin fraction was collected.
2
.3. Methemoglobin (metHb)
ꢁ
5
A solution of carbonmonoxy hemoglobin A (12.0 mL, 1:8 ꢀ 10 mol) was re-
ꢁ
4
acted with potassium ferricyanide (0.09 g, 2.70 ꢀ10 mol) for 3 h at room temper-
ature. The sample was passed through a column of Sephadex G-25 equilibrated
with 0.1 M phosphate buffer, pH 7.4. The eluent was concentrated by ultra-filtration,
exchanged with water and lyophilised.
2
.4. UV/Vis
A sample of HbSNO was deoxygenated for 16 h at 25 °C under N pressure and
2
then re-oxygenated for 30 min with air. A parallel control experiment was performed
using native hemoglobin. Absorption spectra of hemoglobin (40 lM) were taken for
solutions in phosphate buffer at 25 °C.
1
2
.5. H-NMR
1
The degree of oxidation of the a and b hemes was monitored by H-NMR. The
nitrosylated sample (HbSNO) was concentrated to 10 mL by ultra-filtration and ex-
changed twice with 30 mL of deuterium oxide. After the sample was concentrated to
1
0 mL, a process which lasted until 3 h from nitrosylation, sodium azide was added
0.005 g, 10-fold excess) and the samples were deoxygenated under argon at 25 °C in
a foil-wrapped flask. H-NMR measurements were taken after 2 and 48 h of deoxy-
genation for the nitrosylated sample. A control was performed whereby the sodium
(
1

N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
7
1
azide was added to the HbSNO moments prior to H-NMR measurements, to assess
the influence of the azide ligand on the rate of oxidation. Measurements on the al-
iquot, sealed in an NMR tube filled with argon, removed after 2 h of deoxygenation,
1
were taken (500 MHz, water suppression, 45 °C) every 30 min for 15 h. A single H-
NMR spectrum was taken of the HbSNO after 48 h as well as of each of the controls.
1
Lyophilized methemoglobin was dissolved in deuterium oxide. The metHb H-NMR
spectrum was recorded (500 MHz, water suppression, 45 °C) for comparison.
3
. Results and discussion
3
.1. Half-metHb after denitrosylation, detected by its absorption spectrum
Two SNO bonds are formed per tetramer, and in theory two equivalents of Fe(II)
are oxidized in an internal electron transfer mechanism. The degree of oxidation for
an extensively deoxygenated sample of HbSNO was shown to be 48 ꢃ 2% methemo-
globin [13], Fig. 1. A mixture of ferrous and ferric hemoglobin can be quantitatively
analyzed where the hemes are oxygentated [13]. Thus, the sample that had been nit-
rosylated was oxygenated prior to analysis. Native hemoglobin and HbSNO are rel-
atively stable when oxygenated. Thus, reoxygenation of a deoxygenating sample
provides a measured of the extent of heme oxidation at the time of oxygenation.
The partial oxidation that we observe confirms that denitrosylation results in heme
oxidation and also that the initial HbSNO sample was fully nitrosylated. The spec-
trum is a composite of those of oxygenated native hemoglobin and methemoglobin
at the same concentration.
Fig. 1. Spectra reveal the extent of heme oxidation upon denitrosylation. Samples of native human hemo-
globin and methemoglobin were briefly oxygenated prior to measurement. S-nitrosylated hemoglobin was
deoxygenated for 16 h at 25 °C, and then reoxygenated for 30 min with air. The extent of heme oxidation in
the previously S-nitrosylated sample is 48 ꢃ 2%.

8
N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
Fig. 2. The rate of NO release from deoxygenated HbSNO. The amount of overall metHb formation (-
was calculated as the sum of the integrations of the signals from the 5-Me protons, normalized to 2.00
-) is consistently below the average oxidation per sub-
ꢄ
-)
at t ¼ 19 h. The integration of the a subunit (-
ꢅ
unit (-Ã-) as the overall oxidation increases, an epiphenomenon of the electron exchange between the
subunits. The rate of appearance of the oxidized b heme is kinetically complex and is not shown.
3
.2. Rate of heme oxidation
The rate of release of NO can be correlated to the extent of heme oxidation in a
deoxygenated sample of HbSNO (Fig. 2, -ꢄ-). The electron exchange among the he-
mes can also be observed as the extent of a heme oxidation (Fig. 2, -d-). The latter
increases steadily until the distribution of Fe(III) is at equilibrium among all species
(
Fig. 2, -Ã-).
The formation of methemoglobin arises from oxidation of the b hemes while the
final distribution of oxidized hemes results from equilibration of electrons among
the subunits [12]. Fitting the rate of overall metHb formation, as observed from inte-
gration of the signal from the protons of the heme 5-Me groups, to the integrated first
order rate expression, gives the observed first order rate constant for b heme oxidation.
kobs ¼ 1:00 ꢃ 0:12 ꢀ 10ꢁ s =tetramer:
4
ꢁ1
The rate constant for a oxidation, an epiphenomenon of the b oxidation by NO
release, is
obs ¼ 0:51 ꢃ 0:03 ꢀ 10ꢁ s =tetramer:
4
ꢁ1
k
The rate of oxidation of the hemes in deoxygenated HbSNO is much greater for
the b hemes than the a hemes. As noted earlier, the apparently slower rate of a heme
oxidation results from the relatively slow exchange of electrons among the hemes
where oxidation occurs in the b hemes [12]. Although oxidation can occur by transfer
of an electron to NO (reacting with oxygen) to give NO [7]. Due to the absence of
oxygen in the experiments and the heme-specificity of the oxidation reaction, we ex-
ꢁ
3

N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
9
Fig. 3. The rate of NO release from deoxygenated HbSNO, by visible spectroscopy measurements of
HbNO recapture (- -), from Stamler and co-workers [4], compared to the rate of heme oxidation, by
ꢄ
1
H-NMR (- -).
ꢅ
clude a mechanism in which an electron is transferred to NO and oxygen is incorpo-
rated to produce nitrate.
To correlate the rate of heme oxidation to NO release, the rate of HbNO forma-
tion by recapture under the same conditions was calculated from the data of Stamler
and coworkers [4]
obs ¼ 0:81 ꢃ 0:15 ꢀ 10ꢁ s :
4
ꢁ1
k
The rate constants for the two processes are the same within experimental error
Fig. 3). Since there were no external thiol acceptors present, an internal electron
(
transfer should be required as in the a transfer from ferrous heme to S–NO [10].
The formation of metHb as NO is released may lead to formation of HbNO (iron-
coordinated NO). Exposure of deoxygenated Hb to GSNO produces metHb
(
detected by EPR) and HbNO (detected by visible spectroscopy), Scheme 2 [5]. Al-
though deoxygenated ferric hemoglobin can bind NO, such binding is slower and
weaker than in the deoxygenated ferrous species [14]. In our case, we do not detect
HbNO from spectroscopic measurements. However, the species may be removed as
oxygen and nitrogen sweep the system.
3
.3. Proposed mechanism
The generation of NO from the nitrosothiol requires addition of an electron to re-
generate the thiol of the cysteine. Our results show that the electron is transferred in-
ternally as a response to deoxygenation of the adjacent heme iron in a hemoglobin
subunit. When this happens in a red blood cell, the resulting oxidized heme can be re-
duced to the active ferrous form by methemoglobin reductase [1], allowing for an effi-
cient recovery system. Thus, we can consider that the combination of the internal
electron transfer and the reductase in the red cell allows hemoglobin to be a specifically

10
N.J. Ship et al. / Bioorganic Chemistry 31 (2003) 3–10
adapted transporter of nitric oxide in circulation. Other components in the cell, such as
those at the surface, do not provide a coupled electron source. Bimolecular processes
such as transfer of NO to glutathione from HbSNO would imply that there can be a
specific interaction between the protein and glutathione and that a source of electrons
other than the heme is available for the regeneration of the cysteine from the resulting
radical. Our discovery of the competent electron transfer internally coupled to NO re-
lease may explain how the problems of a bimolecular process are overcome.
NO that would be released by internal reduction would be able to bind to free fer-
rous heme groups. However, binding and release from the ferrous heme is dynamic
and the overall release from the erythrocytes will be controlled by an equilibrium
among the heme-associating species. When NO is free, it is a dissolved gas and will
diffuse out of the cell. Given that the heme it is closest to upon release is oxidized and
incapable of tight binding, diffusion away from the immediate hemoglobin will be
entropically driven.
The in vivo system is obviously complex and we cannot surmise to what extent
NO released from hemoglobin would bind to the endothelium to produce local va-
sodilation. However, the mechanism provides a basis for further testing and predic-
tions of its physiological consequences.
Acknowledgments
This paper is warmly dedicated to Professor F.H. Westheimer, who pioneered the
mechanistic study of electron transfer reactions [15]. Our work has been supported by
the Natural Sciences and Engineering Research Council of Canada through an Oper-
ating Grant (RK) and a Postgraduate Scholarship (NJS, who is Sumnel Fellow).
References
[
[
[
1] D. Voet, J.G. Voet, Biochemistry, second ed., 1995.
2] L.J. Ignarro, Angew. Chem., Int. Ed. 38 (1999) 1882–1892.
3] J.S. Stamler, L. Jia, J.P. Eu, T.J. McMahon, I.T. Demchenko, J. Bonaventura, K. Gernert, C.A.
Piantadosi, Science (Washington, DC) 276 (1997) 2034–2037.
[
[
4] L. Jia, C. Bonaventura, J. Bonaventura, J.S. Stamler, Nature (London) 380 (1996) 221–226.
5] R.P. Patel, N. Hogg, N.Y. Spencer, B. Kalyanaraman, S. Matalon, V.M. Darley-Usmar, J. Biol.
Chem. 274 (1999) 15487–15492.
[
[
6] J.P. Pezacki, A. Pelling, R. Kluger, J. Am. Chem. Soc. 122 (2000) 10734–10735.
7] T.J. McMahon, A.E. Stone, J. Bonaventura, D.J. Singel, J.S. Stamler, J. Biol. Chem. 275 (2000)
16738–16745.
[
[
8] N.Y. Spencer, H. Zeng, R.P. Patel, N. Hogg, J. Biol. Chem. 275 (2000) 36562–36567.
9] N. Arulsamy, D.S. Bohle, J.A. Butt, G.J. Irvine, P.A. Jordan, E. Sagan, J. Am. Chem. Soc. 121
(1999) 7115–7123.
[
[
[
[
[
[
10] J.P. Pezacki, N.J. Ship, R. Kluger, J. Am. Chem. Soc. 123 (2001) 4615–4616.
11] Y. Yamamoto, T. Nagaoka, FEBS Lett. 424 (1998) 169–172.
12] N. Shibayama, H. Morimoto, S. Saigo, Biochemistry 36 (1997) 4375–4381.
13] C.C. Winterbourn, Methods Enzymol. 186 (1990) 265–272.
14] V.S. Sharma, T.G. Traylor, R. Gardiner, H. Mizukami, Biochemistry 26 (1987) 3837–3843.
15] F.H. Westheimer, N. Nicolaides, J. Am. Chem. Soc. 71 (1949) 25–28.