Catalase-mediated nitric oxide formation from hydroxyurea

Article abstract of DOI:10.1021/jm030547z

Hydroxyurea reduces the incidence of painful crises in patients with sickle cell disease and has recently been approved for the treatment of this condition. A number of in vitro studies show that the oxidation of hydroxyurea results in the formation of nitric oxide, which also has drawn considerable interest as a sickle cell disease therapy. While patients on hydroxyurea demonstrate elevated levels of nitric oxide-derived metabolites, little information regarding the site or mechanism of the in vivo conversion of hydroxyurea to nitric oxide exists. Chemiluminescence detection experiments show the ability of catalase to catalyze the formation of nitrite and nitrate from hydroxyurea. Spectroscopic studies show that the reaction of hydroxyurea and catalase in the presence of a hydrogen peroxide generating system produces a ferrous-NO catalase complex. Trapping studies indicate the intermediacy of a nitroso species during this reaction. The proposed mechanism for this conversion includes initial hydrogen peroxide-dependent oxidation of hydroxyurea by catalase to form the nitroso species, hydrolysis of this nitroso species to produce nitroxyl, and reductive nitrosylation of the ferric heme of catalase by nitroxyl to yield the ferrous-NO catalase complex. Addition of Angeli's salt, a nitroxyl donor, to ferric catalase also produces the ferrous-NO catalase complex. Spectroscopic studies show that the ferrous-NO catalase complex releases nitric oxide as judged by the oxyhemoglobin assay and an NO specific EPR specific trap. These results demonstrate nitric oxide production from the ferric catalase oxidation of nitroxyl and identify a catalase-mediated pathway as a potential source of nitric oxide production from hydroxyurea.

Full text of DOI:10.1021/jm030547z

J . Med. Chem. 2004, 47, 3495-3501
3495
Articles
Ca ta la se-Med ia ted Nitr ic Oxid e F or m a tion fr om Hyd r oxyu r ea
J inming Huang,‡ Daniel B. Kim-Shapiro,§ and S. Bruce King*,‡
Departments of Chemistry and Physics, Wake Forest University, Winston Salem, North Carolina 27109
Received October 28, 2003
Hydroxyurea reduces the incidence of painful crises in patients with sickle cell disease and
has recently been approved for the treatment of this condition. A number of in vitro studies
show that the oxidation of hydroxyurea results in the formation of nitric oxide, which also has
drawn considerable interest as a sickle cell disease therapy. While patients on hydroxyurea
demonstrate elevated levels of nitric oxide-derived metabolites, little information regarding
the site or mechanism of the in vivo conversion of hydroxyurea to nitric oxide exists.
Chemiluminescence detection experiments show the ability of catalase to catalyze the formation
of nitrite and nitrate from hydroxyurea. Spectroscopic studies show that the reaction of
hydroxyurea and catalase in the presence of a hydrogen peroxide generating system produces
a ferrous-NO catalase complex. Trapping studies indicate the intermediacy of a nitroso species
during this reaction. The proposed mechanism for this conversion includes initial hydrogen
peroxide-dependent oxidation of hydroxyurea by catalase to form the nitroso species, hydrolysis
of this nitroso species to produce nitroxyl, and reductive nitrosylation of the ferric heme of
catalase by nitroxyl to yield the ferrous-NO catalase complex. Addition of Angeli’s salt, a
nitroxyl donor, to ferric catalase also produces the ferrous-NO catalase complex. Spectroscopic
studies show that the ferrous-NO catalase complex releases nitric oxide as judged by the
oxyhemoglobin assay and an NO specific EPR specific trap. These results demonstrate nitric
oxide production from the ferric catalase oxidation of nitroxyl and identify a catalase-mediated
pathway as a potential source of nitric oxide production from hydroxyurea.
Sch em e 1
In tr od u ction
Hydroxyurea (1, Scheme 1) finds use in the treatment
of various cancers and sickle cell disease.1-3 Hydrox-
yurea reduces the incidence of painful crises in patients
with sickle cell disease by increasing the level of fetal
hemoglobin (HbF), a genetically distinct hemoglobin
that inhibits polymerization of sickle cell hemoglobin
(HbS).2-4 The ability of hydroxyurea to increase HbF
levels varies among different patients, suggesting that
hydroxyurea may also benefit patients through other
mechanisms, including alteration of red cell-endothelial
cell interactions and improved red cell rheological
properties.4 Hydroxyurea also acts as a source of NO,
which plays an important role in the maintenance of
normal blood pressure and flow, and has drawn con-
siderable interest as a sickle cell disease treatment.5-7
Sickle cell patients demonstrate in vivo NO formation
during hydroxyurea therapy. These patients show sig-
nificant increases in plasma and red cell nitrite, nitrate,
and iron nitrosyl hemoglobin (HbNO) levels one to 2 h
after administration of an average dose of 15 mg/kg of
hydroxyurea.8 Time course studies reveal that these
NO-derived metabolites increase up to 2 h and then
begin to decline.8 Another study on sickle cell disease
patients taking hydroxyurea shows similar trends re-
garding plasma nitrite/nitrate and an increase in plasma
cyclic guanylate monophosphate (cGMP).9 As NO acti-
vates soluble guanylate cyclase (sGC), the enzyme that
catalyzes the cyclization of guanylate triphosphate
(GTP) to cGMP, increased cGMP levels imply an in-
crease in NO.9 In addition, other experiments using both
erythroleukemic cells and primary human erythroblasts
link HbF gene expression to a sGC-cGMP-dependent
protein kinase pathway.10 Recent in vitro work shows
that hydroxyurea and two other mechanistically differ-
ent NO donors increase γ-globin gene expression and
fetal hemoglobin levels in human erythroid progenitor
cells by a sGC dependent pathway.11 Together, these
studies strongly suggest that hydroxyurea stimulates
HbF synthesis by increasing cGMP levels through
metabolism to NO and subsequent sGC activation.
Despite these results, little detailed information
regarding the site or mechanism of in vivo NO produc-
tion from hydroxyurea exists. Incubation of human
blood (oxygenated, deoxygenated, or cycled between
oxygenated and deoxygenated) with hydroxyurea (10
mM) does not produce a detectable amount of HbNO
after 2 h as determined by electron paramagnetic
* To whom correspondence should be addressed. Phone: 336-758-
5774, fax: 336-758-4656, e-mail: kingsb@wfu.edu.
‡ Department of Chemistry.
§ Department of Physics.
10.1021/jm030547z CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/03/2004

3496 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Huang et al.
Ta ble 1. Nitrite/Nitrate Production from Hydroxyurea (1 mM)
and Catalase after Incubation at 37 °C for 3 h
-
NO2-/NO3
reaction system
(µM)
hydroxyurea
hydroxyurea/Cat
hydroxyurea/Cat/glucose/glucose oxidase
hydroxyurea/glucose/glucose oxidase
hydroxyurea/Cat(heat inact.)/glucose/glucose oxidase
hydroxyurea/Cat/glucose/glucose oxidase/3-AT
3.8 ( 0.2
4.2 ( 0.4
120.2 ( 8.5
4.7 ( 0.2
5.8 ( 0.7
8.3 ( 0.7
resonance (EPR) spectroscopy.12 Similarly, the NO
producing reactions of hydroxyurea and oxy and deoxy
hemoglobin (Hb) occur at only modest rates that do not
kinetically support the observed increases in NO me-
tabolites that occur within short periods of time in
patients on hydroxyurea therapy.13-15 As such, alterna-
tive mechanisms of in vivo NO release from hydroxyurea
rather than the direct reaction with hemoglobin or other
blood components must be considered.
Previous studies show the reaction of catalase with
hydroxylamine yields a catalase-NO complex, which
competently activates sGC.16,17 We find that reaction of
catalase with the structurally related hydroxyurea
rapidly forms nitrite, nitrate, and a ferrous catalase-
NO complex. More specifically, catalase appears to
metabolize hydroxyurea to nitroxyl (HNO), the one-
electron reduced form of nitric oxide, which reduces
catalase to give the ferrous iron-NO catalase complex.
This complex releases NO that reacts with oxyHb and
an NO-specific EPR trap. These results thus describe a
potential pathway of NO formation from hydroxyurea
that possibly explains in vivo HbNO production and
HbF stimulation in patients taking hydroxyurea.
F igu r e 1. Absorbance spectra of ferric catalase (2 µM, solid
line), the aerobic reaction of catalase (2 µM) and hydroxyurea
(50 mM) in the presence of glucose (10 mM) and glucose
oxidase (15 units) in phosphate buffer (100 mM, pH 7.4) at
room temperature (long dash line), and the reaction of catalase
(2 µM) and excess nitric oxide in phosphate buffer (100 mM,
pH 7.4) at room temperature (dash-dot line). The left panel
shows absorbance in the Soret region and the right panel
shows absorbance between 500 and 700 nm.
air (Figure 1). An increase in absorbance at 545 and 575
nm also occurs during this reaction (Figure 1). Over
time, the absorbance spectrum of the reaction mixture
returns to that of resting catalase. No changes occur in
the absorption spectrum of Cat upon addition of hy-
droxyurea in the absence of glucose/glucose oxidase. The
absorption spectrum of the reaction mixture appears
different from the known spectrum of the ferric cata-
lase-NO complex,21 which shows a characteristic Soret
absorbance at 427 nm and prominent peaks at 540 and
576 nm (Figure 1).
The EPR spectrum of ferric catalase shows three
resonances: a broad resonance at g ) 6.26, a resonance
at g ) 4.28, and a very weak resonance at g ) 1.98
(Supporting Information). In general, this spectra ap-
pears similar to a previously reported EPR spectra of
human erythrocyte catalase.22 The EPR spectrum from
the reaction of hydroxyurea with catalase in the pres-
ence of glucose/glucose oxidase at 130 K shows the
appearance of a new prominent peak (g ) 1.99),
characteristic of a ferrous heme-NO complex (Support-
ing Information).23,24 This spectrum also indicates the
presence of ferric catalase in the reaction mixture.
Expansion of this resonance reveals a three peak
pattern with a hyperfine constant of 17.4 G consistent
with electronic coupling to 14N (Figure 2). This spectra
appears similar to the previously reported spectra of the
catalase-NO complex formed during the reaction of
catalase and hydroxylamine.17 The EPR spectrum of
this reaction mixture also differs from the spectrum of
the addition of NO to ferric catalase (Supporting Infor-
mation). No changes occur in the EPR spectrum of Cat
upon addition of hydroxyurea in the absence of glucose/
glucose oxidase. Room temperature EPR spectroscopy
fails to indicate the intermediacy of the previously
described hydroxyurea-derived nitroxide radical.25
Gas chromatographic analysis of the headspace above
the reaction mixture of hydroxyurea and catalase in the
presence of glucose and glucose oxidase reveals the
presence of carbon dioxide (CO2). These measurements
fail to provide any evidence of nitrous oxide (N2O)
formation during this reaction. Addition of glucose and
Resu lts
Chemiluminescence NO analysis following reaction
mixture reduction with a refluxing vanadium(III) chlo-
ride/HCl solution provides a rapid and accurate mea-
surement of the stable oxidative decomposition products
of NO, nitrite, and nitrate (NO2-/NO3-).18 Table 1
summarizes the production of NO2-/NO3 from the
-
reaction of catalase (Cat) and hydroxyurea. Incubation
of hydroxyurea (1 mM) alone with purified Cat (1.0 µM,
bovine liver) fails to produce a significant amount of
NO2-/NO3-, showing that Cat and hydroxyurea do not
react (Table 1). Addition of glucose (10 mM) and glucose
oxidase (7.5 units), a slow hydrogen peroxide-generating
system,19 to the same mixture of hydroxyurea and
catalase increases the level of NO2-/NO3 28.6-fold
-
(Table 1). Chemiluminescence NO analysis following
reaction mixture reduction with potassium iodide in
acetic acid, which only reduces nitrite to NO, shows that
nitrate (113.0 ( 8.0 µM) constitutes the major species
of this mixture. Incubation of this system with 3-amino-
1,2,4 triazole (3-AT), a known Cat inhibitor,20 or using
-
heat-deactivated Cat abolishes the increase in NO2
/
NO3- (Table 1). In addition, incubation of hydroxyurea
with glucose/glucose oxidase in the absence of catalase
fails to result in the accumulation of NO2-/NO3- (Table
1).
Absorption measurements show the characteristic
Soret absorbance for ferric catalase at 405 nm shifts to
412 nm during the reaction of hydroxyurea and catalase
in the presence of glucose and glucose oxidase under

Catalase-Mediated Nitric Oxide Formation
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3497
Under the same conditions, the direct reaction of
hydroxyurea with oxyHb to produce metHb does not
occur fast enough (k ) 7.54 × 10-4 ( 2.10 × 10-5
min-1 26
to account for the rapid observed rate of
)
conversion. Appreciable amounts of metHb also do not
form in the absence of hydroxyurea.
Treatment of catalase with the HNO donor, sodium
trioxodinitrate (Angeli’s salt), under anaerobic condi-
tions should provide a method for the preparation of the
ferrous-NO catalase complex. Absorption measure-
ments show the Soret absorbance for ferric catalase at
405 nm gradually shifts to 421 nm upon addition of
Angeli’s salt (Figure 3). Under anaerobic conditions, this
spectrum does not change over time. Increases in
absorbance at 555 and 580 nm also occur after the
addition of Angeli’s salt to ferric catalase (Figure 3).
While this spectrum differs from the spectrum of the
reaction of hydroxyurea with catalase in the presence
of glucose and glucose oxidase (Figure 1), the addition
of Angeli’s salt to this reaction mixture produces an
absorption spectrum nearly identical to the spectrum
from the reaction of catalase and Angeli’s salt (Figure
3). The absorption spectrum of the reaction of catalase
and Angeli’s salt also differs from the spectrum of the
ferric catalase-NO complex (Figures 1 and 3).21 Fitting
the spectrum from the reaction of hydroxyurea and
catalase to the spectra of ferric catalase and the fer-
rous-NO catalase complex reveals the reaction mixture
consists of 25% ferric catalase and 75% of the ferrous-
NO catalase complex (Figure 3). EPR spectroscopy
shows that treatment of catalase with Angeli’s salt
produces a spectrum nearly identical to the EPR spec-
trum from the reaction of hydroxyurea and catalase
with a prominent three-line absorbance at g ) 1.99 (a
) 17.4) (Figure 2).
F igu r e 2. Expanded EPR spectra at 130 K of ferric catalase
(solid line), the reaction of catalase and hydroxyurea in the
presence of glucose and glucose oxidase (long dash line), and
the anaerobic reaction of catalase (0.8 mM) and Angeli’s salt
(0.8 mM) in phosphate buffer (100 mM, pH 7.4) at room
temperature (dash-dot line).
glucose oxidase to a solution of hydroxyurea and cata-
lase in the presence of 1, 3-cyclohexadiene (50 mM)
produces the cycloadduct (3, Scheme 1) after 3 h.
Isolation and characterization by 1H NMR spectroscopy
of this cycloadduct provides clear evidence for the
intermediacy of the C-nitroso compound (2) in this
reaction (Scheme 1).
Absorption measurements reveal the rapid conversion
of oxyHb to metHb during the reaction of catalase and
hydroxyurea in the presence of glucose and glucose
oxidase (Supporting Information). The shift in the Soret
absorbance from 416 to 405 nm and the increases in
absorbance at 500 and 635 nm and the decreases in
absorbance at 540 and 577 nm demonstrate the conver-
sion of oxyHbA to metHbA. Analysis of the change in
absorbance over time reveals an apparent rate constant
Room-temperature EPR experiments using the NO
selective trap, carboxy-PTIO, provide further evidence
of NO release from the ferrous-NO catalase complex.
Carboxy-PTIO produces a five-line EPR spectrum and
reacts with NO to form an imino nitroxide radical, 2-(4-
for this process of k ) 2.37 × 10-2 ( 1.38 × 10-3 min-1
.
F igu r e 3. Absorbance spectra of ferric catalase (solid line), the aerobic reaction of catalase (2 µM) and hydroxyurea (50 mM) in
the presence of glucose (10 mM) and glucose oxidase (15 units) in phosphate buffer (100 mM, pH 7.4) at room temperature (short
dash line), the anaerobic reaction of ferric catalase (2 µM) and Angeli’s salt (20 µM) in phosphate buffer (100 mM, pH 7.4) at room
temperature (dash-dot line), and the addition of Angeli’s salt to the hydroxyurea/catalase reaction mixture in the presence of
glucose and glucose oxidase (long dash line). The left panel shows absorbance in the Soret region and the right panel shows
absorbance between 500 and 700 nm. Inset shows the spectrum from the reaction of hydroxyurea with catalase fit to the spectra
of ferric catalase and the ferrous-NO catalase complex.

3498 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Huang et al.
Sch em e 2
EPR spectroscopy shows that under these conditions
the reaction of hydroxyurea and catalase produces a
mixture of ferric catalase and a ferrous-NO catalase
complex. The appearance of the characteristic triplet
signal at g ) 1.99 in the EPR spectra provides strong
evidence for a ferrous-NO catalase complex.23,24 Both
EPR and absorption spectroscopy confirm that this
complex differs from the ferric-NO catalase complex
formed by the direct addition of NO to ferric cata-
lase.21,28 Also, the EPR spectrum from the mixture of
the HNO donor Angeli’s salt with ferric catalase appears
nearly identical to the reaction spectrum at g ) 1.99.
This result provides further support for the identity of
the ferrous-NO complex as HNO reductively nitrosy-
lates ferric heme proteins to produce the ferrous-NO
complexes.25,29-32 While the absorption spectra from the
reaction of hydroxyurea and catalase shows some dif-
ferences compared to the spectra from the reaction of
Angeli’s salt and ferric catalase, the addition of Angeli’s
salt to the hydroxyurea reaction mixture produces an
absorption spectra nearly identical to the spectra from
the direct addition of Angeli’s salt to ferric catalase. This
change suggests that the reaction mixture consists of
ferrous-NO catalase and ferric catalase, which reacts
with HNO to form the ferrous-NO complex. The addi-
tion of less than stoichiometric amounts of Angeli’s salt
to ferric catalase also produces absorption spectra which
appear similar to the hydroxyurea and ferric catalase
reaction spectrum indicating a mixture of ferric catalase
and the ferrous-NO complex. The similarity of the
absorption spectrum from the reaction with the fitted
spectrum generated from the spectra of ferric catalase
and the ferrous-NO catalase complex further supports
the assertion that the reaction mixture contains these
two components. While these results differ from a recent
report that states no reaction occurs between HNO
generated from Angeli’s salt and catalase,33 in our hands
the mixture of anaerobic solutions of Angeli’s salt and
ferric catalase reproducibly yields the ferrous-NO
complex of catalase. As dithionite generally fails to
reduce the ferric iron of catalase,21,34 the reductive
nitrosylation of ferric catalase with HNO (from Angeli’s
salt) provides a convenient method for the preparation
of the ferrous-NO catalase complex.
F igu r e 4. Absorption spectrum of the exposure of the fer-
rous-NO catalase complex (10 µM) in phosphate buffer (100
mM, pH 7.4) at room temperature to air. UV-vis measure-
ments were taken every 5 min. Arrows indicate increasing and
decreasing absorbance.
carboxyphenyl)-4,5-dihydro-4,4,5,5,-tetramethyl-1H-imi-
dazolyl-1-oxy (carboxy-PTI), which produces a distinct
seven-line EPR spectrum.27 Addition of carboxy-PTIO
to a solution of the ferrous-NO catalase complex
generated by the anaerobic incubation of ferric catalase
with Angeli’s salt shows the seven-line spectrum of
carboxy-PTI indicating NO transfer from the ferrous-
NO catalase complex to the trap (Supporting Informa-
tion). The seven-line carboxy-PTI signal appears stable
and persists for over 20 min. The addition of Angeli’s
salt to carboxy-PTIO in the absence of catalase does not
produce carboxy-PTI.
Exposure of the ferrous-NO catalase complex to air
results in the formation of ferric catalase as shown by
UV-vis spectroscopy (Figure 4). These measurements
show the Soret absorbance shifts from 421 to 405 nm
of ferric catalase and the decrease of the pronounced
absorbances at 555 and 580 nm (Figure 4). This process
also demonstrates a clean isobestic point at 417 nm
indicating the presence of only two species. Analysis of
these absorbance changes over time yields an apparent
rate constant for this reaction of 5.03 × 10-4 s-1
.
Chemiluminescence detection measurements (after sub-
traction of the nitrite produced during the decomposition
of Angeli’s salt) reveal that the aerobic decomposition
of the ferrous-NO catalase complex (10 µM) yields a
similar amount of nitrate (8.0 ( 1.3 µM, n ) 3).
Discu ssion
Chemiluminescence experiments reveal that catalase
in the presence of a hydrogen peroxide generating
system converts hydroxyurea to nitrite and nitrate and
provide evidence for the intermediacy of NO. These NO-
derived metabolites form from a physiologically relevant
concentration of hydroxyurea (1 mM) and within 3 h, a
time scale compatible with the observed increase in
nitrite/nitrate in patients undergoing hydroxyurea
therapy.8 Nitrite/nitrate could also potentially form from
the reaction of nitroxyl with oxygen. Further experi-
ments with the inhibitor 3-AT and heat-inactivated
catalase demonstrate the requirement of catalase for
nitrite/nitrate formation from hydroxyurea.
Scheme 2 depicts the proposed mechanism for the
formation of a ferrous-NO catalase complex during the
reaction with hydroxyurea. Hydrogen peroxide gener-
ated from the glucose/glucose oxidase reaction activates
ferric catalase to Compound I. Direct two-electron
oxidation of hydroxyurea by the catalase Compound I

Catalase-Mediated Nitric Oxide Formation
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3499
Sch em e 3
nitrate and ferric heme as determined by chemilumi-
nescence detection and absorption spectroscopy, respec-
tively. The autooxidation of iron nitrosyl heme proteins
to nitrate and the ferric heme is a general process and
has previously been noted for the ferrous-NO catalase
complex.34,39 This process occurs at a rate that allows
for reactivation of the enzyme during the reaction with
hydroxyurea and explains the observed mixture of ferric
catalase and ferrous-NO catalase and the preponder-
ance of nitrate. The possible oxidation of nitrite by
catalase to nitrate could also explain the high nitrate-
to-nitrite ratio.35,36
Absorption studies show the conversion of oxyHb to
metHb during the catalase-mediated oxidation of hy-
droxyurea. While hydroxyurea also converts oxyHb to
metHb, the slow rate of this reaction does not account
for the observed conversion of oxyHb to metHb during
the reaction of catalase and hydroxyurea in the presence
of a peroxide generating system.26 These results indicate
that the ferrous-NO complex of catalase formed during
the reaction with hydroxyurea releases NO that reacts
with oxyHb and suggests a weak association of NO to
the ferrous heme iron. The catalase-NO complex de-
rived from the reaction of catalase and azide also
releases NO that reacts with oxyHb.38 To further
illustrate the ability of the ferrous-NO catalase complex
to release NO, EPR experiments show that the ferrous-
NO catalase complex, prepared from ferric catalase and
Angeli’s salt, converts the nitric oxide specific EPR trap,
carboxy-PTIO to carboxy-PTI. Overall, these results
reveal that the combination of ferric catalase and HNO,
derived from either hydroxyurea or Angeli’s salt, rep-
resents a NO donor through the intermediacy of a
ferrous-NO catalase complex.
species generates the C-nitrosoformamide (2) and trap-
ping studies with 1,3-cyclohexadiene support the inter-
mediacy of 2. Such a conversion finds precedence in the
catalase-mediated two-electron oxidation of a variety of
substrates including ethanol.35 The oxidation of N-
hydroxycyanamide to nitrosyl cyanide by catalase is
particularly similar to the oxidation of hydroxyurea to
the C-nitrosoformamide (2, Scheme 2).36 Unlike the
reaction of hydroxyurea and horseradish peroxidase,25
the failure of room temperature EPR measurements to
identify the nitroxide radical of hydroxyurea argues
against single electron oxidation (or H atom abstraction)
processes. However, the failure to detect the nitroxide
radical does not necessarily preclude two rapid single
electron oxidations. Hydrolysis of C-nitrosoformamide
(2) would yield a carbamic acid that decomposes to
carbon dioxide, ammonia, and nitroxyl (HNO). Reduc-
tive nitrosylation of the ferric heme of catalase by HNO
would form the ferrous NO-catalase complex. The
autooxidation of the ferrous-NO catalase complex
yields nitrate and ferric catalase (vide infra), which can
be activated to Compound I and react with another
hydroxyurea molecule.
Gas chromatographic measurements identify the
production of carbon dioxide during this reaction but
fail to show the formation of nitrous oxide, the dimer-
ization and dehydration product of HNO and a common
marker for the intermediacy of HNO.37 These results
are surprising as nitrous oxide has been identified
during the similar oxidation of cyanamide by catalase
and taken as evidence for HNO formation.19 Control
experiments show that nitrous oxide does not escape
from the reaction vessel during incubation. The limit
of detection of the nitrous oxide assay ensures the
detection of less than 1% of the theoretical amount of
nitrous oxide from this reaction under these conditions.
We propose that the reductive nitrosylation of the ferric
heme of catalase by HNO effectively competes with
HNO dimerization and explains the lack of nitrous oxide
formation. Ferric heme proteins act as effective traps
of HNO that kinetically compete with HNO dimeriza-
tion.30,32 The formation of a nitroxyl-catalase complex
has recently been proposed to explain the inability to
detect HNO electrochemically (after oxidation to NO)
during the oxidation of cyanamide with catalase.38 The
binding or reaction of cyanamide, but not hydroxyurea,
with ferric catalase may effectively compete with the
reductive nitrosylation by HNO allowing HNO dimer-
ization and nitrous oxide formation, explaining the
observed differences in nitrous oxide production in these
reactions. The results of these studies clearly indicate
the need for improved methods of free HNO detection.
Scheme 3 summarizes the reactions of the ferrous-
NO catalase complex. Exposure of the ferrous-NO
catalase complex to air results in the formation of
The formation of a ferrous-NO catalase complex
capable of NO donation from the reaction of hydrox-
yurea and catalase could have important implications
in hydroxyurea based sickle cell disease therapy. Both
animal models and patients receiving hydroxyurea
demonstrate an increase in NO metabolites, indicating
the in vivo ability of hydroxyurea to act as an NO
donor.8,9,40 Catalase-mediated release of NO from hy-
droxyurea could benefit patients through various mech-
anisms such as decreasing platelet adhesion or scav-
enging cell-free hemoglobin.7,41 In addition, recent
experiments indicate that hydroxyurea and other NO
donors increase fetal hemoglobin production at least in
part through a soluble guanylate cyclase-dependent
pathway.11 Complexes of NO and catalase, such as those
formed in the described reaction of hydroxyurea and
catalase, activate soluble guanylate cyclase in a manner
similar to free NO and could ultimately stimulate fetal
hemoglobin synthesis.16,17 While these results clearly
show the ability of catalase, an enzyme found in high
levels in both liver and erythrocytes, to convert hydrox-
yurea to NO, further studies will be required to identify
the role catalase plays during in vivo metabolism of
hydroxyurea.
Su m m a r y
Hydroxyurea is currently used as a therapy for sickle
cell disease, and recent evidence indicates that a portion
of its beneficial effects may be related to hydroxyurea’s
ability to release nitric oxide (NO). While patients

3500 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14
Huang et al.
(15 units) in phosphate buffer (100 mM, pH 7.4, 2.0 mL) at
room temperature. The reaction was monitored by TLC, and
after 3 h the reaction mixture was extracted with EtOAc (2 ×
2.0 mL), dried over MgSO4, and evaporated to give cycloadduct
(3) as a residue: TLC Rf 0.26 (EtOAc); 1H NMR (300 MHz,
CDCl3, TMS internal standard) δ 6.4 (m, 2H), δ 4.7 (s, 1H), δ
4.9 (s, 1H), δ 1.4 (m, 4H).
Rea ction of Sod iu m Tr ioxod in itr a te w ith Ca ta la se. A
degassed solution of sodium trioxodinitrate (Angeli’s salt, 4
µL, 10 mM) in 0.01 M NaOH was added to a degassed solution
of catalase (10 µM) in phosphate buffer (100 mM, pH 7.4, 2.0
mL) at room temperature in a cuvette with a rubber septum
and UV-vis measurements were taken every 5 min. UV-vis
and chemiluminescence nitrite/nitrate measurements were
made on this solution upon exposure to air. Similarly, a
degassed solution of sodium trioxodinitrate (8 µL, 0.1 M) in
0.01 M NaOH was added to a degassed solution of catalase
(0.8 mM) in phosphate buffer (100 mM, pH 7.4, 1.0 mL) at
room temperature and transferred to an EPR tube and frozen
in liquid nitrogen and EPR spectra were taken at 130 K. In
addition, NO saturated buffer (100 µL, ∼2 mM) was added to
a degassed solution of catalase (0.8 mM) in phosphate buffer
(100 mM, pH 7.4, 0.4 mL) at room temperature and transferred
to an EPR tube and frozen in liquid nitrogen.
NO Relea se fr om Ca ta la se (F e2+)-NO a n d Rea ction
w ith Oxyh em oglobin . Catalase (1 µM) was added to a
solution of hydroxyurea (50 mM), glucose (100 µM), glucose
oxidase (2 units), and oxyhemoglobin (20 µM) in phosphate
buffer (100 mM, pH 7.4, 2.0 mL) in a cuvette at room
temperature. The conversion of oxy to methemoglobin was
monitored using UV spectroscopy with measurements being
taken every 5 min.
EP R Detection of NO Relea se fr om Ca ta la se (F e2+)-
NO Com p lex. A solution of Angeli’s salt (2 µL, 0.1M) in 0.01
M NaOH was added to a degassed solution of catalase (5 µM)
in phosphate buffer (100 mM, pH 7.4, 0.2 mL) at room
temperature. After 10 min, a solution of carboxy-PTIO (20 µL,
1 mM) was added, and the mixture was transferred to a
capillary tube. EPR spectra were taken at room temperature
using 8.5 mW microwave power, 1.25 G modulation amplitude,
and 9.42 GHz microwave frequency.
taking hydroxyurea show measurable increases in a
number of NO-derived metabolites,8,9 the site and
mechanism for the conversion of hydroxyurea to NO
remains poorly undefined. These results identify the
enzyme catalase as a potential candidate for the in vivo
oxidation of hydroxyurea to NO. Spectroscopic and
product analysis studies strongly suggest the initial
conversion of hydroxyurea to nitroxyl (HNO) that re-
ductively nitrosylates the ferric heme of catalase to form
a ferrous-NO catalase complex. This complex releases
NO that reacts with oxyHb and an NO-specific EPR
trap. The formation of such ferrous-NO complexes
could play a role in the beneficial effects of hydroxyurea
in sickle cell disease.
Exp er im en ta l Section
Ma ter ia ls. Hydroxyurea, 30% hydrogen peroxide, 3-amino-
1,2,4-triazole, catalase (Sigma C-40, bovine liver), and glucose
oxidase were purchased from Sigma Chemical Company, St.
Louis, MO. 1,3-Cyclohexadiene and D-glucose were purchased
from Aldrich Chemical Co., Milwaukee, WI. 2-(4-Carboxyphe-
nyl)-4,5-dihydro-4,4,5,5,-tetramethyl-1H-imidazolyl-1-oxy-3-
oxide, potassium salt (Carboxy-PTIO) and sodium trioxodini-
trate (Na2N2O3, Angeli’s Salt) were purchased from Cayman
Chemical Company, Ann Arbor, MI. Normal adult hemoglobin
was isolated and purified as previously described.13
Nitr ite/Nitr a te Mea su r em en ts. Catalase (1.0 µM) was
added to a solution of hydroxyurea (1 mM), glucose (10 mM),
and glucose oxidase (7.5 units) in phosphate buffer (100 mM,
pH 7.4, 0.1 mL). After incubation at 37 °C for 3 h, an aliquot
(5 µL) of the reaction mixture was injected into the reaction
vessel of a Sievers 280 Nitric Oxide Analyzer chemilumines-
cence detector, and nitrite and nitrate were measured as
reported earlier.13 Identical experiments were performed using
heat-inactivated (95 °C, 30 min) catalase or in the presence of
3-amino-1,2,4-triazole (200 mM). Control experiments were
performed in the absence of either glucose/glucose oxidase or
catalase.
Sp ectr oscop y of th e Rea ction of Hyd r oxyu r ea a n d
Ca ta la se. Catalase (2 µM) was added to a solution of hydrox-
yurea (50 mM), glucose (10 mM), and glucose oxidase (15 units)
in phosphate buffer (100 mM, pH 7.4, 2.0 mL) at room
temperature. UV-vis measurements were taken every 5 min
on a Cary 100 Bio UV-visible Spectrophotometer (Varian,
Walnut Creek, CA). Similarly, hydroxyurea (50 mM) was
added to a solution of catalase (0. 8 mM), glucose (10 mM),
and glucose oxidase (7.5 units) in phosphate buffer (100 mM,
pH 7.4, 0.3 mL) at room temperature. After 3 h, this solution
was transferred to an EPR tube and frozen in liquid nitrogen.
EPR spectra were taken at 130 K on a Bruker ER200D
spectrometer using 8.5 mW microwave power, 5.0 G modula-
tion amplitude, and 9.37 GHz microwave frequency with
g-values being determined from a superimposed spectrum of
2,2-diphenyl-1-picrylhydrazyl, g ) 2.0036.
Ga s Ch r om a togr a p h ic An a lysis of th e Rea ction Hea d -
sp a ce. Catalase (10 µM) was added to a solution of hydrox-
yurea (500 mM), glucose (100 mM), and glucose oxidase (15
units) in phosphate buffer (100 mM, pH 7.4, 2.0 mL) at room
temperature in a 20 mL flask with a rubber septum. After 3
h, an aliquot of the reaction headspace (250 µL) was injected
onto a 6890 Hewlett-Packard gas chromatograph equipped
with a thermal conductivity detector and a 6 ft. × 1/8 in.
Porapak Q column at an operating oven temperature of 50 °C
(injector and detector 150 °C) with a flow rate of 16.67 mL/
min (He, carrier gas). The retention time of carbon dioxide was
2.10 min and identical to a known sample. Using a standard
mixture of 10% N2O in helium (Matheson), the limit of
detection of this assay is 9 nmol of N2O. Control experiments
show no loss of N2O over 3 h from the reaction vessel
Cycloa d d u ct (3) Tr a p p in g. 1,3-Cyclohexadiene (19.0 µL,
0.1 mmol, 50 mM) was added to a solution of hydroxyurea (500
mM), catalase (10 µM), glucose (20 mM), and glucose oxidase
Ack n ow led gm en t. This work was supported by the
National Institutes of Health (HL62198, S.B.K.). The
NMR spectrometers used in this work were purchased
with partial support from NSF (CHE-9708077) and the
North Carolina Biotechnology Center (9703-IDG-1007).
Su p p or tin g In for m a tion Ava ila ble: Figures S1-S3.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Refer en ces
(1) Donehower, R. C. Hydroxyurea. In Cancer Chemotherapy and
Biotherapy; Chabner, B. A., Longo, D. L., Eds.; Lippincott-
Raven: Philadelphia, 1996; pp 253-261.
(2) Charache, S.; Terrin, M. L.; Moore, R. D.; Dover: G. J .; Barton,
F. B.; Eckert, S. V.; McMahon, R. P.; Bonds, D. R. and the
Investigators of the Multicenter Study of Hydroxyurea in Sickle
Cell Anemia. Effect of Hydroxyurea on the Frequency of Painful
Crises in Sickle Cell Anemia. N. Engl. J . Med. 1995, 332, 1317-
1322.
(3) Charache, S.; Barton, F. B.; Moore, R. D.; Terrin, M. L.;
Steinberg, M. H.; Dover: G. J .; Ballas, S. K.; McMahon, P.;
Castro, O.; Orringer, E. P and the Investigators of the Multi-
center Study of Hydroxyurea in Sickle Cell Anemia. Hydrox-
yurea and Sickle Cell Anemia. Medicine 1996, 75, 300-326.
(4) Halsey, C.; Roberts, I. A. G. The Role of Hydroxyurea in Sickle
Cell Disease. Br. J . Haematol. 2003, 120, 177-186.
(5) King, S. B. The Nitric Oxide Producing Reactions of Hydrox-
yurea. Curr. Med. Chem. 2003, 10, 437-452.
(6) Reiter, C. D.; Gladwin, M. T. An Emerging Role for Nitric Oxide
in Sickle Cell Disease Vascular Homeostasis and Therapy. Curr.
Opin. Hematol. 2003, 10, 99-107.
(7) Gladwin, M. T.; Schechter, A. N. Nitric Oxide Therapy in Sickle
Cell Disease. Semin. Hematol. 2001, 38, No. 4, 333-342.

Catalase-Mediated Nitric Oxide Formation
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 14 3501
(8) Gladwin, M. T.; Shelhamer, J . H.; Ognibene, F. P.; Pease-Fye,
M. E.; Nichols, J . S.; Link, B.; Patel, D. B.; J ankowski, M. A.;
Pannell, L. K.; Schechter, A. N.; Rodgers, G. P. Nitric Oxide
Donor Properties of Hydroxyurea in Patients with Sickle Cell
Disease. Br. J . Haematol. 2002, 116, 436-444.
(9) Nahavandi, M.; Tavakkoli, F.; Wyche, M. Q.; Perlin, E.; Winter,
W. P.; Castro, O. Nitric Oxide and Cyclic GMP Levels in Sickle
Cell Patients Receiving Hydroxyurea. Br. J . Haematol. 2002,
119, 855-857.
(10) Ikuta, T.; Ausenda, S.; Cappellini, M. D. Mechanism for Fetal
Globin Gene Expression: Role of the Soluble Guanylate Cyclase-
cGMP-dependent Protein Kinase Pathway. Proc. Natl. Acad. Sci.
2001, 98, 1847-1852.
(11) Cokic, V. P.; Smith, R. D.; Beleslin-Cokic, B. B.; Njoroge, J . M.;
Miller, J . L.; Gladwin, M. T.; Schechter, A. N. Hydroxyurea
Induces Fetal Hemoglobin by the Nitric Oxide-Dependent
Activation of Soluble Guanylyl Cyclase. J . Clin. Invest. 2003,
111, 231-239.
(25) Huang, J .; Sommers, E. M.; Kim-Shapiro, D. B.; King, S. B.
Horseradish Peroxidase Catalyzed Nitric Oxide Formation from
Hydroxyurea. J . Am. Chem. Soc. 2002, 124, 3473-3480.
(26) Huang, J .; Zhou, Z.; Kim-Shapiro, D. B.; Ballas, S. K.; King, S.
B. Hydroxyurea Analogues as Kinetic and Mechanistic Probes
of the Nitric Oxide Producing Reactions of Hydroxyurea and
Oxyhemoglobin. J . Med. Chem. 2003, 46, 3748-3753.
(27) Akaike, T.; Maeda, H. In Methods in Enzymology, Volume 268,
Nitric Oxide Part A: Sources and Detection of NO; NO Synthase;
Packer, L., Ed.; Academic Press: San Diego, 1996; Chapter 20.
(28) Brunelli, L.; Yermilov, V.; Beckman, J . S. Modulation of Catalase
Peroxidatic and Catalytic Activity by Nitric Oxide. Free Radical
Biology and Medicine 2001, 30, 709-714.
(29) Doyle, M. P.; Mahapatro, S. N.; Broene, R. D.; Guy, J . K.
Oxidation and Reduction of Hemoproteins by Trioxodinitrate-
(II). The Role of Nitrosyl Hydride and Nitrite. J . Am. Chem. Soc.
1988, 110, 593-599.
(12) Lockamy, V. L.; Huang, J .; Shields, H.; Ballas, S. K.; King, S.
B.; Kim-Shapiro, D. B. Urease Enhances the Formation of Iron
Nitrosyl Hemoglobin in the Presence of Hydroxyurea. Biochim.
Biophys. Acta 2003, 1622, 109-116.
(13) Huang, J .; Hadimani, S. B.; Rupon, J . W.; Ballas, S. K.; Kim-
Shapiro, D. B.; King, S. B. Iron Nitrosyl Hemoglobin Formation
from the Reactions of Hydroxyurea and Hemoglobin. Biochem-
istry 2002, 41, 2466-2474
(14) Kim-Shapiro, D. B.; King, S. B.; Bonifant, C. L.; Kolibash, C.
P.; Ballas, S. K. Time-Resolved Absorption Study of the Reaction
of Hydroxyurea with Sickle Cell Hemoglobin. Biochim. Biophys.
Acta 1998, 1380, 64-74.
(15) Rupon, J . W.; Domingo, S. R.; Smith, S. V.; Gummadi, B. K.;
Shields, H.; Ballas, S. K.; King, S. B.; Kim-Shapiro, D. B. The
Reactions of Myoglobin, Normal Adult Hemoglobin, Sickle Cell
Hemoglobin and Hemin with Hydroxyurea. Biophys. Chem.
2000, 84, 1-11.
(16) Ignarro, L. J .; Adams, J . B.; Horwitz, P. M.; Wood, K. S.
Activation of Soluble Guanylate Cyclase by NO-Hemeproteins
Involves NO-Heme Exchange. J . Biol. Chem. 1986, 261, 4997-
5002.
(17) Craven, P. A.; DeRubertis, F. R.; Pratt, D. W. Electron Spin
Resonance Study of the Role of NO-Catalase in the Activation
of Guanylate Cyclase by NaN3 and NH2OH. J . Biol. Chem. 1979,
254, 8213-8222.
(18) Wink, D. A.; Darbyshire, J . F.; Nims, R. W.; Saavedra, J . E.;
Ford, P. C. Reactions of the Bioregulatory Molecule Nitric Oxide
in Oxygenated Aqueous Media: Determination of the Kinetics
for Oxidation and Nitrosation by Intermediates Generated in
the NO/O2 Reaction. Chem. Res. Toxicol. 1993, 6, 23-27.
(19) Nagasawa, H. T.; DeMaster, E. G.; Redfern, B.; Shirota, F. N.;
Goon, D. J . W. Evidence for Nitroxyl in the Catalase-Mediated
Bioactivation of the Alcohol Deterrent Agent Cyanamide. J . Med.
Chem. 1990, 33, 3120-3122.
(20) Margoliash, E.; Novogrodsky, A. A Study of the Inhibition of
Catalase by 3-Amino-1, 2, 4-Triazole. Biochem. J . 1957, 68, 468-
475.
(21) Hoshino, M.; Ozawa, K.; Seki, H.; Ford, P. C. Photochemistry of
Nitric Oxide Adducts of Water-Soluble Iron(III) Porphyrin and
Ferrihemoproteins Studied by Nanosecond Laser Photolysis. J .
Am. Chem. Soc. 1993, 115, 9568-9575.
(22) Palcic, M.; Dunford, H. B. Spectral Studies of Human Erythro-
cyte Catalase. Can. J . Biochem. 1979, 57, 321-329.
(23) Kon, H. Paramagnetic Study of Nitric Oxide Hemoglobin. J . Biol.
Chem. 1968, 243, 4350-4357.
(24) Yonetani, T.; Yamamoto, H.; Erman, J . E.; Leigh, J . S., J r.; Reed,
G. H. Electromagnetic Properties of Hemoproteins: Optical and
Electron Paramagnetic Resonance Characteristics of Nitric
Oxide Derivatives of Metalloporphyrin-Apohemoprotein Com-
plexes. J . Biol. Chem. 1972, 247, 2447-2455.
(30) Bazylinksi, D. A.; Hollocher, T. C. Metmyoglobin and Methemo-
globin as Efficient Traps for Nitrosyl Hydride (Nitroxyl) in
Neutral Aqueous Solutions. J . Am. Chem. Soc. 1985, 107, 7982-
7986.
(31) Miranda, K. M.; Nims, R. W.; Thomas, D. D.; Espey, M. G.;
Citrin, D.; Bartberger, M. D.; Paolocci, N.; Fukuto, J . M.;
Feelisch, M.; Wink, D. A. Comparison of the Reactivity of Nitric
Oxide and Nitroxyl with Heme Proteins: A Chemical Discussion
of the Differential Biological Effects of these Redox Related
Products of NOS. J . Inorg. Biochem. 2003, 93, 52-60.
(32) Miranda, K. M.; Paolocci, N.; Katori, T.; Thomas, D. D.; Ford,
E.; Bartberger, M. D.; Espey, M. G.; Kass, D. A.; Feelisch, M.;
Fukuto, J . M.; Wink, D. A. A Biochemical Rational for the
Discrete Behavior of Nitroxyl and Nitric Oxide in the Cardio-
vascular System. Proc. Natl. Acad. Sci. 2003, 100, 9196-9201.
(33) Bastian, N. R.; Foster, M. J . P.; Ballantyne, M.; Lu, Y. Nitrogen
Oxides, The Good, the Bad, and the Ugly. Curr. Topics Biophys.
2002, 26, 115-127.
(34) Nicholls, P. The Reactions of Azide with Catalase and their
Significance. Biochem. J . 1964, 90, 331-343.
(35) Keilin, D.; Hartree, E. F. Catalase, Peroxidase and Methemo-
globin as Catalysts of Coupled Peroxidatic Reactions. Biochem.
J . 1955, 60, 310-325.
(36) Shirota, F. N.; Goon, D. J . W.; DeMaster, E. G.; Nagasawa, H.
T. Nitrosyl Cyanide, a Putative Metabolic Oxidation Product of
the Alcohol-Deterrent Agent Cyanamide. Biochem. Pharmacol.
1996, 52, 141-147.
(37) Bonner, F. T.; Hughes, M. N. The Aqueous Solution Chemistry
of Nitrogen in Low Positive Oxidation States. Comments Inorg.
Chem. 1988, 7, 215-234.
(38) Shahidullah, M.; Duncan, A.; Strachan, P. D.; Rafique, K. M.;
Ball, S. L.; McPate, M. J . W.; Nelli, S.; Martin, W. Role of
Catalase in the Smooth Muscle Relaxant Actions of Sodium
Azide and Cyanamide. Eur. J . Pharmacol. 2002, 435, 93-101.
(39) Cooper, C. E. Nitric Oxide and Iron Proteins. Biochim. Biophys.
Acta 1999, 1411, 290-309.
(40) J iang, J .; J ordan, S. J .; Barr, D. P.; Gunther, M. R.; Maeda, H.;
Mason, R. P. In Vivo Production of Nitric Oxide in Rats after
Administration of Hydroxyurea. Mol. Pharmacol. 1997, 52,
1081-1086.
(41) Reiter, C. D.; Wang, X.; Tanus-Santos, J . E.; Hogg, N.; Cannon,
R. O.; Schechter, A. N.; Gladwin, M. T. Cell-Free Hemoglobin
Limits Nitric Oxide Bioavailability in Sickle Cell Disease. Nat.
Med. 2002, 8, 1383-1389.
J M030547Z