TOXICOLOGICAL SCIENCES 81, 325–331 (2004)
doi:10.1093/toxsci/kfh211
Advance Access publication July 7, 2004
The Effect of Divalent Cations on Neuronal Nitric Oxide
Synthase Activity
John Weaver,*†jj1 Supatra Porasuphatana,‡ Pei Tsai,†jj Guan-Liang Cao,† Theodore A. Budzichowski,*
Linda J. Roman,§ and Gerald M. Rosen†¶jj
*Department of Chemistry, University of Maryland Baltimore County, Baltimore, Maryland 21250; †Department of Pharmaceutical Sciences, University of
Maryland School of Pharmacy, Baltimore, Maryland 21201; ‡Department of Toxicology, Faculty of Pharmaceutical Science, Khon Kaen University, Khon Kaen
40002, Thailand; §Department of Biochemistry, The University of Texas Health Science Center at San Antonio, San Antonio, Texas 78230;
¶Medical Biotechnology Center, University of Maryland Biotechnology Institute, Baltimore, Maryland 21201; and jjCenter for Low Frequency EPR
for In Vivo Physiology, University of Maryland, Baltimore, Maryland 21201
Received June 7, 2004; accepted June 30, 2004
Neuronal nitric oxide synthase (NOS I) is a Ca21/calmodulin–
binding enzyme that generates nitric oxide (NO) and L-citrulline
from the oxidation of L-arginine, and superoxide (O2) from the
one-electron reduction of oxygen (O2). Nitric oxide in particular
has been implicated in many physiological processes, including
vasodilator tone, hypertension, and the development and properties of neuronal function. Unlike Ca21, which is tightly regulated
in the cell, many other divalent cations are unfettered and can
compete for the four Ca21 binding sites on calmodulin. The results
presented in this article survey the effects of various divalent metal
ions on NOS I–mediated catalysis. As in the case of Ca21, we
demonstrate that Ni21, Ba21, and Mn21 can activate NOS I to
metabolize L-arginine to L-citrulline and NO, and afford O2 in
the absence of L-arginine. In contrast, Cd21 did not activate NOS I
to produce either NO or O2, and the combination of Ca21 and
either Cd21, Ni21, or Mn21 inhibited enzyme activity. These interactions may initiate cellular toxicity by negatively affecting NOS I
activity through production of NO, O2 and products derived
from these free radicals.
Key Words: nitric oxide; superoxide; NOS I; calmodulin; divalent
cations; metal toxicity.
INTRODUCTION
Neuronal nitric oxide synthase (NOS I) belongs to a family of
enzymes, nitric oxide synthases (NOS, EC 1.14.13.39), known
to catalyze the conversion of L-arginine to nitric oxide (NO)
and L-citrulline (Moncada and Higgs, 1993; Nathan and Xie,
1994). NOS I and endothelial NOS (NOS III) are constitutive
isoforms dependent on the transient influx of Ca21 to activate
calmodulin, which binds to the isoform to elicit enzyme activity.
1
To whom correspondence should be addressed at Department of
Pharmaceutical Sciences, University of Maryland School of Pharmacy,
725 W. Lombard Street, Baltimore, MD 21201. Fax: (410) 706–8184.
E-mail: [email protected].
Toxicological Sciences vol. 81 no. 2
#
A third isoform, inducible NOS (NOS II), is a cytokine-inducible
isoform independent of Ca21 in which calmodulin is permanently bound (Nathan and Xie, 1994). These enzymes are composed of an N-terminal oxidase domain containing an iron
protoporphyrin IX (heme), tetrahydrobiopterin (H4B) with a
binding site for L-arginine, and a C-terminal reductase domain
containing the flavin adenine dinucleotide (FAD) and the flavin
adenine mononucleotide (FMN). The two domains are connected by a calmodulin-binding motif to which a transient influx
of Ca21 binds (Abu-Soud et al., 1994; Kobayash et al., 2001;
Matsuda and Iyanagi, 1999; Miller et al., 1999). Besides producing NO, NOS also generates superoxide (O2), the ratio
of these free radicals is dependent on the concentration of Larginine (Pou et al., 1999; Yoneyama et al., 2001). Nitric oxide
from NOS I has been implicated in many physiological
processes, including vasodilator tone and the development
and properties of neuronal function (Moncada and Higgs,
1993; Roskams et al., 1994). In addition, NOS I has been
associated with hypertension (Chrissobolis et al., 2002).
There are regulatory co-factors that dictate whether NOS I
generates NO and/or O2 and H2O2. For instance, in the
absence of L-arginine, O2 (Pou et al., 1999; Yoneyama
et al., 2001) and H2O2 are directly produced; the ratio of
these reduction products of O2 is set by the presence of H4B
(Rosen et al., 2002). Again, the binding of L-arginine to NOS
shifts electron transport from O2 to this amino acid, producing
NO at the expense of O2 (Pou et al., 1999). Another control
element is calmodulin, and, in the presence of Ca21, it binds to
NOS I and enhances electron transport through the reductase
domain to the heme, allowing NO and O2 to be produced [for
a review, see Roman et al. (2002)]. We have recently demonstrated that Pb21 and Sr21 stimulate the NOS I generation of
NO and O2 by binding to and activating calmodulin (Weaver
et al., 2002). Unlike Ca21, which is tightly regulated in the cell
(Clapham, 1995), many divalent cations are unfettered and can
compete for the four Ca21 binding sites on calmodulin. Several
other divalent cations such as Ni21, Ba21, Cd21, and Mn21 have
Society of Toxicology 2004; all rights reserved.
326
WEAVER ET AL.
been shown to bind to and activate calmodulin (Goldstein and
Ar, 1983; Habermann et al., 1993; Mills and Johnson, 1985), and
several of these divalent cations have been associated with NOS,
although the mechanism by which these metals affected NOS is
not well understood (Gupta et al., 2000; Mittal et al., 1995;
Palumbo et al., 2001; Perry and Marletta, 1998; Yamazaki
et al., 1995). Activation of NOS I by divalent cations, other
than Ca21, may initiate toxicity through over-production of
NO, O2, and products derived from these free radicals.
In the present study, we investigated the possible effects of
several divalent metal ions, known to activate calmodulin, on
NOS I-mediated catalysis in the absence and presence of Larginine. We demonstrate that Ni21, Ba21, and Mn21 can activate NOS I to metabolize L-arginine to L-citrulline and NO. In
the absence of L-arginine, these cations also promote NOS I
generation of O2. It is suggested that these divalent
cations, by binding to calmodulin, initiate enzymatic activity.
In contrast, Cd21 did not activate NOS I to produce either NO
or O2.
MATERIALS AND METHODS
Reagents. L-arginine, NADPH, calcium chloride, nickel chloride, magnesium chloride, catalase, xanthine oxidase, ferricytochrome c (bovine), hypoxanthine, and calmodulin were obtained from Sigma-Aldrich (St. Louis, MO).
Cadmium nitrate was obtained from Allied Chemical (Morristown, NJ).
Potassium chloride was obtained from Mallinckrodt Baker, Inc. (Phillipsburg,
NJ). Manganese chloride, potassium nitrate, and barium chloride were obtained
from Fisher Scientific (Fair Lawn, NJ). Superoxide dismutase (SOD) was
obtained from Roche Diagnostics (Mannheim, W-Germany). L-[U-14C]arginine
monohydrochloride ([14C]L-arginine) was purchased from Amersham
Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). 5-tert-Butoxycarbonyl-5-methyl-1-pyrroline N-oxide (BMPO) was synthesized as described
in the literature (Stolze et al., 2003; Tsai et al., 2003). Dowex 50W-X8 cation
exchange resin was obtained from Bio-Rad (Hercules, CA). All other chemicals
were used as purchased without further purification. Buffers used contain less
than 0.01% of inorganic salts, including Ca21.
Determination of nitric oxide production. The initial rate of NO production by purified NOS I was estimated using the hemoglobin assay (Murphy and
Noack, 1994). The reaction was initiated by the addition of NOS I (2.0 mg) to a
cuvette containing HEPES buffer (50 mM, 0.5 mM EGTA, pH 7.4), oxyhemoglobin (10 mM), the divalent cation (500 mM), calmodulin (100 U/ml), NADPH
(100 mM), and L-arginine (100 mM) to a final volume of 500 ml at room
temperature. A UV-Vis spectrophotometer (Uvikon, Model 940, Research
Instruments International, San Diego, CA) was used to monitor the conversion
of oxyhemoglobin to methemoglobin during the course of the reaction.
Specifically, the increase in absorbance at 401 nm was used to quantitate the
reaction, using an extinction coefficient of 60 mM1cm1 at 401 nm.
Spin trapping of NOS I-generated O2 using BMPO. Spin trapping of
O2 by BMPO was conducted in a reaction mixture containing, NOS I (14.2 mg),
various divalent cations (500 mM), and calmodulin (100 U/ml) in phosphate
buffer (50 mM, pH 7.4, 1 mM DTPA, 1 mM EGTA). The reaction was initiated
by the addition of BMPO (50 mM) and NADPH (100 mM) into the reaction
mixture to a final volume of 0.3 ml at room temperature. The reaction mixture
was mixed, transferred into a quartz flat cell and fitted into the cavity of the EPR
spectrometer. EPR spectra were continually recorded at room temperature and
data shown below in Figure 4 were obtained 10 min after the initiation of
the reaction. Instrument settings were as follows: microwave power, 20 mW;
modulation frequency, 100 kHz; modulation amplitude, 0.5 G; sweep time,
12.5 G/min; and response time, 0.5 s. The receiver gain for each experiment
is given in the legend of Figure 4.
Generation of O2 by hypoxanthine/xanthine oxidase. Superoxide was
generated from the action of xanthine oxidase on hypoxanthine (400 mM, final
concentration) in sodium phosphate buffer (50 mM, pH 7.4, 1 mM DTPA, 1 mM
EGTA; Chelexed, pH 7.4). The initial rate of O2 generation was estimated
spectrophotometrically by measuring the SOD-inhibitive reduction of ferricytochrome c (80 mM) at 550 nm using an extinction coefficient of 21 mM1 cm1 at
550 nm (Kuthan and Ullrich, 1982).
Reduction of ferricytochrome c by NOS I. The reduction of ferricytochrome c (80 mM) by NOS I (0.4 mg) was monitored in a reaction containing
Ca21, Ni21, Ba21, Mn21, or Cd21 (500 mM), calmodulin (100 U/ml), and NOS
I in HEPES buffer (50 mM, pH 7.4, 0.5 mM EGTA) at room temperature. The
reaction was initiated by the addition of NADPH (100 mM) to the reaction
mixture to a final volume of 500 ml. The initial rate of ferricytochrome c
reduction in the absence and presence of SOD (30 U/ml) at 550 nm was
estimated spectrophotometrically using an extinction coefficient of 21 mM1
cm1 at 550 nm.
Purification of NOS I. NOS I was expressed and purified essentially
as described in the literature (Roman et al., 1995), with the modification
that the culture volume was 500 ml rather than 1000 ml. The enzyme
concentration was determined by its CO-difference spectrum, as described
by Roman et al. (1995), using an extinction coefficient of 100 mM1cm1 at
De 444–475 nm.
Effect of Divalent Metal Ions on Calmodulin-Dependent
Activation of Purified NOS I
NOS I activation using the [14C]L-citrulline formation assay. The activation of purified NOS I was determined by its ability to catalyze the formation of
L-citrulline from L-arginine as previously reported (Weaver et al., 2002), with
modifications. A cocktail solution ( [14C]L-arginine (0.6 mCi/ml) in the presence
of NADPH (1 mM), L-arginine (100 mM), and calmodulin (100 U/ml) in HEPES
buffer (50 mM, 0.5 mM EGTA, pH 7.4) was prepared. The reaction was initiated
by the addition of the cocktail solution into the reaction mixture containing
purified NOS I (3.7 mg) and Ca21, Ni21, Ba21, Mn21, or Cd21 (from 100
mM to 1 mM) to a final volume of 150 ml. The reaction mixture was incubated
at room temperature for 10 min and terminated with 2 ml of stop solution (20 mM
HEPES, 2 mM EDTA, pH 5.5). The product [14C]L-citrulline was separated by
passing the reaction mixture through columns containing Dowex 50W-X8 cation
exchange resin preactivated with sodium hydroxide (1 M), and radioactivity was
counted using a scintillation counter (Model LS 6500; Beckman Coulter Inc.,
Fullerton, CA). Data were expressed as means and standard deviations of
multiple experiments.
Several divalent cations have been shown to bind and
activate calmodulin (Goldstein and Ar, 1983; Habermann
et al., 1993; Mills and Johnson, 1985). Therefore, we investigated what effect Ba21, Mn21, Ni21, and Cd21, known to
affect other calmodulin processes, may have on the activation of calmodulin-dependent NOS I. NOS I activation was
estimated by measuring the amount of [14C]L-citrulline produced from L-arginine (100 mM) to which [14C]L-arginine
(0.6 mCi/ml) was added in the presence of the different
divalent cations, as described in Materials and Methods.
Maximal activation for each divalent cation, except for
Mn21, was reached at 500 mM (Fig. 1). Ca21- and
Ba21-activated NOS I showed similar dose-response curves,
RESULTS
EFFECT OF DIVALENT CATIONS ON NOS I ACTIVITY
FIG. 1. Effect of divalent cations on the formation of L-citrulline by NOS
I. The formation of [14C]L-citrulline (represented as % control with respect to
Ca21-activated NOS I) from the NOS oxidation of L-arginine. NOS I activity
was assayed in the reaction containing NOS I (3.7 mg), NADPH (1 mM),
calmodulin (100 U/ml), [14C]L-arginine (0.6 mCi/ml), L-arginine (100 mM)
and (&) CaCl2, (m) BaCl2, (3) MnCl2, () NiCl2, or (r) Cd(NO3)2 (various
concentrations) in HEPES buffer (50 mM, 0.5 mM EGTA, pH 7.4). Each point
represents the mean 6 S.D. from three independent experiments on the same
preparation of purified NOS I.
where in up to 1 mM of each divalent cation salt there was no
apparent inhibition in NOS I production of L-citrulline (Fig. 1).
Of note, at maximal activation Ba21-activated NOS I was
70% that of Ca21-activated NOS I. In contrast, NOS I
activation by Ni21 and Mn21 exhibited biphasic response
curves (Fig. 1) for each metal ion salt, but at concentrations
up to 1 mM, NOS I activation decreased by 90% and 60%,
respectively. At maximal response, Ni21 and Mn21 produced
40% and 70%, respectively, of L-citrulline as compared to
Ca21-stimulated NOS I (Fig. 1). Interestingly, Cd21 demonstrated negligible NOS I activation (9%), even at 500 mM.
From these data, it was determined that 500 mM of each
divalent cation would produce respective maximal activation
under these experimental conditions, and this concentration
was used for further experiments.
Next, we examined the effect of various divalent cations on
Ca21-activated NOS I. The reactions contained all components
described in Materials and Methods, except that the divalent
cation, Cd21, Ni21, Ba21, or Mn21 (500 mM), was included
along with Ca21 (500 mM). Activation of NOS I was estimated
following the formation of [14C]L-citrulline from [14C]L-arginine (0.6 mCi/ml) in the presence of L-arginine (100 mM). The
combination of Ca21 and either Cd21, Ni21, or Mn21 resulted in
a diminution in enzyme activation by 90%, 85%, and 40%,
respectively, as compared to Ca21 alone (Fig. 2). Surprisingly,
the combination of Ba21 and Ca21 nearly doubled NOS I
response, indicating a synergistic effect of Ba21 on Ca21activated NOS I.
327
FIG. 2. Effect of divalent cations on the activation of NOS I in the presence
of Ca21. The formation of [14C]L-citrulline from the NOS oxidation of Larginine. NOS I activity was assayed in the reaction containing NOS I (3.7 mg),
NADPH (1 mM), calmodulin (100 U/ml), [14C]L-arginine (0.6 mCi/ml), Larginine (100 mM), CaCl2 (500 mM) and either BaCl2 (500 mM), MnCl2 (500
mM), NiCl2 (500 mM), or Cd(NO3)2 (500 mM) in HEPES buffer (50 mM, 0.5 mM
EGTA, pH 7.4). Each point represents the mean 6 S.D. of the % control from
three independent experiments on the same preparation of purified NOS I.
Initial Rate of NO Production by Divalent
Cation-Activated NOS I
While the above experiments determined the effect of various divalent cations on NOS-mediated oxidation of L-arginine
to L-citrulline and, by implication, NO, we measured the initial
rate of NO production from NOS I, activated either by Ca21,
Ni21, Ba21, Mn21, or Cd21. Initial rates of NO generation were
estimated using the oxyhemoglobin assay, as described in Materials and Methods. For these experiments, the concentration of Larginine was set at 100 mM, whereas the concentration of Ca21,
Ni21, Ba21, Mn21, or Cd21 was fixed at 500 mM. The initial rate
of NO production catalyzed by Ca21 or Ba21 was approximately the same, at 370 nmoles/min/mg protein (Table 1). In
the presence of Cd21, no measurable rate of NO generation was
observed, as compared to the control, in the absence of
calmodulin.
A correlation between the initial rates of NO generation
and production of L-citrulline was established by converting
the respective values to percent control with respect to Ca21activated NOS I (Table 1 and Fig. 3). The percent control values
show that the initial rates of NO production are comparable to
the L-citrulline assay, each following the same trend of activation in which Mn21 and Ni21 produced the least amount of NO.
Control experiments were also performed to ensure that the
counter ions, Cl or NO3, did not contribute to the observed
results (Nishimura et al., 1999; Schrammel et al., 1998). Substituting KCl (500 mM) or KNO3 (500 mM) for the divalent salts
resulted in no activation of NOS I in the absence of Ca21 or
inhibition in the presence of Ca21 (500 mM) (data not shown). It
was concluded that any effect these counter ions have on NOS I
activity is negligible in this study, and all observations were a
328
WEAVER ET AL.
TABLE 1
Initial Rates of NO Generation from NOS I Induced by
Various Divalent Cations
NO production
Metal salts (500 mM)
CaCl2
BaCl2
MnCl2
NiCl2
Cd(NO3)2
(nmoles/min/mg protein)a
% Control
387 6 28
364 6 10
290 6 47
217 6 36
3.32 6 6.0
100 6 7.3
94 6 3
75 6 10
55 6 9
0.8 6 2
a
Rates are the average of three independent experiments, expressed as the
means and standard deviations.
FIG. 4. Spin trapping of O2 from purified NOS I by BMPO. A
representative plot of the first low-field peak height, arrows in the insert, of the
EPR spectrum of BMPO-OOH from O2 generated from divalent cationactivated NOS I in the absence of L-arginine. This nitroxide is derived from
the reaction of O2 with BMPO. Each point represents the mean 6 S.D. of
the % control from three independent experiments on the same preparation of
purified NOS I. Insert – The EPR spectrum of BMPO-OOH was obtained
10 min after the spin trapping of O2 generated Ca21-activated NOS I in the
absence of L-arginine. The reaction system consisted of NOS I (14.2 mg),
CaCl2 (500 mM), calmodulin (100 U/ml), NADPH (100 mM), and BMPO (50
mM) in phosphate buffer (50 mM, pH 7.4, 1 mM DTPA, 1 mM EGTA).
Receiver gain was 53104.
FIG. 3. A plot of NOS I activation versus ionic radii of the divalent
cations. The formation of [14C]L-citrulline () by the [14C]L-citrulline
formation assay and NO (&) production by the hemoglobin method from
the NOS I oxidation of L-arginine with respect to the ionic radii of each
divalent cation (500 mM) (Huheey et al. 1993). Each point represents the mean
6 S.D. of the % control with respect to Ca21-activated NOS I of three
independent experiments on the same preparation of purified NOS I.
consequence of the respective divalent cations. These data also
further confirmed that Cd21 does not activate NOS I.
Spin Trapping of NOS-Generated O2
NOS I produces O2 in the absence of L-arginine (Pou
et al., 1992; Yoneyama et al., 2001). Of interest is whether Ba21,
Ni21, or Mn21, which activate NOS I to metabolize L-arginine to
L-citrulline and NO, can, in the absence of substrate, generate
O2. Given that ferricytochrome c and its derivatives cannot
report NOS generated O2 (Weaver et al. 2003), we turned to
spin trapping/EPR spectroscopy. For these experiments, BMPO
was chosen as the spin trap, because its reaction with O2 gives
a spin-trapped adduct, BMPO-OOH, that exhibits a long half-life
(Rosen et al. 2002). As depicted in Figure 4, BMPO spin trapped
O2 from NOS I activated by various divalent cations. However, the EPR spectral peak height of BMPO-OOH was much
greater with Ca21 than with Ba21, Ni21, and Mn21 (500 mM for
each ion)-activated NOS I. These data suggested that NOS I
generated considerably less O2 when activated by these
metal ions than when activated by Ca21. In the case of Mn21,
the EPR spectral peak height of BMPO-OOH was diminished by
nearly 80% as compared to control (Fig. 4).
The Effect of Divalent Cations on the Rate of Ferricytochrome
c Reduction by O2
To determine whether the decrease in O2 spin trapping
was attributable to competition for O2 by Ba21, Ni21 or
Mn21, a series of control experiments were performed. A hypoxanthine/xanthine oxidase system was used as a source of O2.
And the SOD-like property of each of the divalent cations
(500 mM), which was dissolved in the buffer used in the experiments above, was estimated by monitoring the initial rate of
ferricytochrome c reduction as compared to that observed
with SOD (30 U/ml). The control was performed in the absence
of these metal cations. In the absence of divalent cations, 0.613 6
0.17 mM/min of O2 was generated, as measured by the
reduction of ferricytochrome c, and in the presence of Ca21,
Ba21, Ni21, or Mn21, 0.598 6 0.17, 0.547 6 0.18, 0.604 6 0.15,
or 0.540 6 0.16 mM/min of O2, respectively, was observed.
From these data, no significant SOD-like property was attributed
to any of the divalent cations.
EFFECT OF DIVALENT CATIONS ON NOS I ACTIVITY
TABLE 2
Reduction Rates of Ferricytochrome c by NOS I
Reduction rate (mM/min/mg protein)a
Metal salts (500 mM)
CaCl2
NiCl2
BaCl2
MnCl2
Cd(NO3)2
With calmodulin
Without calmodulin
43.1 6 6.2
29.3 6 5.0
37.5 6 8.4
38.8 6 5.2
2.35 6 0.80
5.10 6 0.96
4.25 6 1.7
3.29 6 0.89
3.84 6 0.93
4.10 6 1.7
a
Rates are the average of three independent experiments, expressed as the
means and standard deviations.
Reduction of Ferricytochrome c by Divalent Cation–Activated
NOS I
As reported in the literature, ferricytochrome c is reduced
by NOS I (Heinzel et al., 1992; Mayer et al., 1991; Pou et al.,
1992; Roman et al., 2000; Sheta et al., 1994). This enzymatic
reduction occurs in the absence and presence of calmodulin,
although the reduction in the absence of calmodulin is
10-fold less than in the presence of calmodulin when NOS I
is activated using Ca21/calmodulin (Roman et al., 2000). Therefore, experiments were conducted to determine what effect
Ba21, Mn21, Ni21, and Cd21 had on the rate of NOS I reduction
of ferricytochrome c (Table 2). Each divalent cation–activated
NOS I, except Cd21, was able to reduce ferricytochrome c at a
rate comparable to that of Ca21. Also, each cation-activated
NOS I followed the trend of maximal activation similar to
that seen in previous experiments. Of note, the rates of ferricytochrome c reduction in the presence of Cd21-activated NOS
I, with and without calmodulin, were the same, confirming that
the activation of NOS I as measured by the L-citrulline assay
by Cd21 is negligible. In each case, whether calmodulin was
present or not, SOD (30 U/ml) did not inhibit the reduction of
ferricytochrome c. These data demonstrate that this reduction
was not a consequence of O2, but resulted from direct reduction by NOS I, as shown previously (Pou et al. 1992).
DISCUSSION
Activation of calmodulin occurs when Ca21 occupies all
four EF-hand domains (Kern et al., 2000). The binding of Ca21
to calmodulin produces a conformational change that converts
the protein to an active form, which binds to NOS I, allowing
electron transport through the reductase domain of NOS to the
oxidase domain to facilitate the production of NO, O2, and
H2O2. Several divalent and trivalent cations besides Ca21 can
bind to calmodulin and exert cellular toxicity by altering the
normal homeostasis of the Ca21/calmodulin pathway
(Habermann et al., 1993; Mills and Johnson, 1985; Ozawa et
al., 1999). We explored the effect Ni21, Ba21, Cd21, or Mn21
has on calmodulin-dependent NOS I-generated NO and O2.
329
The results described in this article indicate that Ni21,
Ba , and Mn21 can activate NOS I, producing NO, and
O2 (Figs. 1 and 3 and Table 1). In the presence of Ca21,
formation of L-citrulline from L-arginine was inhibited with
the inclusion of Ni21, Cd21, or Mn21 (Fig. 2). The increase
in L-citrulline formation seen with the simultaneous inclusion of
Ca21 and Ba21 was surprising. We are exploring this synergistic
phenomenon in more detail. Of note, although the trend of
activation is the same, the percent control values for NO generation and L-citrulline production differed by 10% for Ni21
and Mn21, with the values for Ba21 differing by 20%. Because
the trend of activation remained the same independent of the
assay, these differences were disregarded.
In the absence of L-arginine, each divalent cation–activated
NOS I was shown to reduce ferricytochrome c at rates comparable to Ca21-activated NOS I (Table 2). Similarly, ferricytochrome c reduction by NOS I in the absence of calmodulin was
10-fold less than in its presence (Table 2) , confirming an
earlier report (Roman et al., 2000). This finding supports the
notion that the activity associated with these divalent cations is a
result of these metal ions binding to calmodulin.
Depending on the divalent cation used to activate NOS,
differing amounts of O2 were generated. This observation was
similar to that seen with L-citrulline and NO production (Figs. 3
and 4). However, the disparity in NOS production of O2, NO
cannot be overlooked. Several control experiments verified that
the disparity in the EPR spectral peak heights correlated with
variable amounts of NOS-generated O2, and that these
differing EPR spectral peak heights were not attributed to
other reactions taking place in the reaction mixture, such as
variable SOD-like activity of the metal ions or an enhanced
rate of BMPO-OOH decomposition of the BMPO-OOH in
the presence of the various divalent cations.
We offer several possible explanations for the differences
in L-citrulline, NO, and O2 production by these metal ions. It
has been suggested that cations with a specific range of radii are
required to induce the necessary conformational change for
calmodulin activity (Ozawa et al., 1999). Therefore, a plot of
NOS I activation versus ionic radii was developed (Fig. 3). This
plot confirms that the initial rates of NO generation and total
L-citrulline production follow the same trend, and that a specific
range of radii is necessary for maximal activation with
calmodulin. Divalent cations smaller or larger than Ca21 exhibit
decreases in activation. In contrast, no activation was observed
using Cd21 with ionic radii of 1.09 Å (Huheey et al., 1993), and
previous work observed that Sr21-activated NOS I was similar to
Ca21-activated NOS I, although Sr21(1.32 Å) has a similar ionic
radius to Pb21 (1.33 Å), which showed minimal activation,
(Huheey et al., 1993; Weaver et al., 2002). These findings suggest that other factors are involved for calmodulin activation of
NOS I. In addition, Cd21 has been shown to activate other
calmodulin-dependent processes, although NOS I activation
was not observed in these studies (Habermann et al., 1993;
Ozawa et al., 1999).
21
330
WEAVER ET AL.
A review of hard-soft acid-base (HSAB) concepts may also
provide insight into the activation of NOS I by these various
divalent cations (Huheey et al., 1993). Metal ions most frequently bind to donor ligands according to preferences dictated
by HSAB. Mainly, hard metals prefer hard ligands, and soft
metals prefer soft ligands. Calmodulin binds to Ca21 via
side-chain carboxylates, alcohol, or carboxamide oxygen, or
via backbone carbonyl groups (Lippard and Berg 1994).
These ligands are classified as hard ligands and prefer to bind
to hard metals such as Ca21, Ba21, and Mn21, and they do not
bind favorably with Ni21, classified as a soft metal. The decrease
in NOS I activation by Ni21 supports this theory, whereas Mn21
and Ba21 exhibited near maximal activation compared to
Ca21. However, the biphasic curves observed with Mn21 and
Ni21 suggest that still other factors are involved.
Other studies have supported our observations in which
divalent cations can promote or inhibit NOS activity from homogenate (Gupta et al., 2000; Mittal et al., 1995; Yamazaki et al.,
1995), purified (Perry and Marletta, 1998), or recombinant
(Palumbo et al., 2001) NOSs as well as NO generated from
macrophages (Tian and Lawrence, 1996). For instance, Ni21 has
also been shown to enhance NOS I activity in the brain and
adrenal glands (Gupta et al., 2000), whereas it has been
shown to inhibit NOS I activity in other studies (Mittal et al.,
1995; Palumbo et al., 2001). Several mechanisms have been
suggested for this activation and inhibition of NOS (Gupta
et al., 2000; Palumbo et al., 2001), but the results shown here
suggest that Ni21 and other divalent cations can bind to calmodulin in a manner similar to Ca21, thereby activating NOS
directly. This finding appears to be concentration specific, as
higher concentrations of several metals ions did not promote
activation (Fig. 1). At higher concentrations of these divalent
cations, further alterations in calmodulin conformation may prevent further activation. Our results are also in agreement with
studies that show that high concentrations of specific metals ions
in the presence of Ca21 promote NOS inhibition (Mittal et al.,
1995; Palumbo et al., 2001), possibly by further altering calmodulin, as mentioned above. Cd21 appears not to produce the
desired conformational change of calmodulin for NOS activation in the absence of Ca21, and it may only promote changes in
calmodulin not suitable for NOS activation, as only inhibition
was observed.
As described by Gupta et al. (2000), Ni21 can be accumulated in the brain after exposure. Although it is unclear whether
the concentration accumulated in the brain would be sufficient to
apply to our results, it is noteworthy. Among the many toxic
effects of Ni21 in biological systems is the ability of this divalent
cation to affect Ca21 homeostasis, produce low but measurable
levels of other free radicals, and participate with NO in Ni21induced hyperglycemia (Denkhaus and Salnikow, 2002; Gupta
et al., 2000). Also, NOS I is present in non-neuronal cell types in
addition to the brain and other neuronal cells types [for a review,
see Förstermann et al. (1998)]. Accumulation of Ni21 or other
divalent cations in such tissues may elicit the response observed
in this study. Of note, these results reflect stimulation or inhibition
from the total concentration of metal salts added. The free ion
concentrations would need to be determined to correlate these
findings to possible concentrations of these metal ions in vivo.
The toxic effects of Ba21 include gastrointestinal, cardiovascular, respiratory, and neuromuscular symptoms, and it
directly affects vascular, cardiac, and gastrointestinal smooth
muscle (Jacobs et al., 2002). Hypertension has also been
reported in both laboratory and clinical studies of Ba21 (Jacobs
et al., 2002). Excessive Mn21 intake can result in neurobehavioral deficits, neurological syndrome similar to chronic Parkinson’s disease, and oxidative stress (Chetty et al., 2001; Husain et
al., 2001). Cadmium ion has been associated with hypertension
and myocarditis among its many toxic effects (Kopp et al., 1983;
Thun et al., 1989). We suggest that toxicity related to Cd21 is the
result of its ability to inhibit NOS I catalysis, whereas other metal
ions may exert toxicity by either producing or inhibiting free
radical production dependent on the concentration of metal ion
present.
Results reported here should be interpreted with caution
because the free ion concentration under the conditions presented is not known. It also cannot be stated that these divalent
cations reach concentrations sufficient for calmodulin activation
in biological systems as described in this paper. These results
merely suggest a model by which several divalent cations may
affect NOS I activity. Because NO is involved in many physiological processes (Moncada and Higgs, 1993), toxic effects
could be attributed to the ability of these divalent cations to
affect NOS I activity as demonstrated in this paper.
ACKNOWLEDGMENTS
This research was supported in part by grants from the National Institutes of
Health, EB-2034 (G.M.R., J.W.), R25-GM55036 (J.W.), AG-20445 (P.T.) and
GM52419, and from the Robert A. Welch Foundation, grant AQ1192 (L.J.R.).
REFERENCES
Abu-Soud, H. M., Yoho, L. L., and Stuehr, D. J. (1994). Calmodulin controls
neuronal nitric oxide synthase by a dual mechanism. J. Biol. Chem. 269,
32047–32050.
Chetty, C. S., Reddy, G. R., Suresh, A., Dessaih, D., Ali, S. F., and Slikker, K. J.
(2001). Effects of manganese on inositol polyphosphate receptors and nitric
oxide synthase activity in the rat brain. Int. J. Toxicol. 20, 275–280.
Chrissobolis, S., Ziogas, J., Anerson, C., Chu, Y., Faraci, F., and Sobey, C.
(2002). Neuronal NO mediates cerebral vasodilator responses to K1 in hypertensive rats. Hypertension 39, 880–885.
Clapham, D. (1995). Calcium signaling. Cell 80, 259–268.
Denkhaus, E., and Salnikow, K. (2002). Nickel essentiality, toxicology, and
carcinogenicity. Crit. Rev. Oncol. Hematol. 42, 35–56.
Förstermann, U., Boissel, J.-P., and Kleinhert, H. (1998). Expressional control of
the ‘‘constitutive’’ isoforms of nitric oxide synthase (NOS I and NOS III).
FASEB J. 12, 773–790.
EFFECT OF DIVALENT CATIONS ON NOS I ACTIVITY
331
Goldstein, G. W., and Ar, D. (1983). Lead activates calmodulin sensitive
processes. Life Sci. 33, 1001–1006.
binding and Ca21/calmodulin-dependent enzyme activation. Biochem.
Biophys. Res. Commun. 285, 142-146.
Gupta, S., Ahmad, N., Hussain, M., and Srivastava, R. (2000). Involvement of
nitric oxide in nickel-induced hyperglycemia in rats. Nitric Oxide 4, 129–138.
Perry, J. M., and Marletta, M. A. (1998). Effects of transition metals on nitric
oxide synthase catalysis. Proc. Natl. Acad. Sci. U.S.A. 95, 11101–11106.
Habermann, E., Crowell, K., and Janicki, P. (1993). Lead and other metals can
substitute for Ca21 in calmodulin. Arch. Toxicol. 54, 61–70.
Pou, S., Keaton, L., Surichamorn, W., and Rosen, G. M. (1999). Mechanism of
superoxide generation by neuronal nitric oxide synthase. J. Biol. Chem. 274,
9573–9580.
Heinzel, B., John, M., Klatt, P., Bohme, E., and Mayer, B. (1992). Ca21/Calmodulin-dependent formation of hydrogen peroxide by brain nitric oxide
synthase. Biochem. J. 281, 627–630.
Huheey, J., Keiter, E., and Keiter, R. (1993). Inorganic Chemistry Principles of
Structure and Reactivity (4th edition). HarperCollins College Publishers,
New York.
Husain, M., Khanna, V. K., Roy, A., Tandon, R., Pradeep, S., and Set, P. K.
(2001). Platelet dopamine receptors and oxidative stress parameters as markers of manganese toxicity. Hum. Exp. Toxicol. 20, 631–636.
Jacobs, I. A., Taddeo, J., Kelly, K., and Valernziano, C. (2002). Poisoning as a
result of barium styphnate explosion. Am. J. Ind. Med. 41, 285–288.
Kern, M., Wisniewski, M., Canell, L., and Audesirk, G. (2000). Inorganic lead
and calcium interact positively in activation of calmodulin. Neurotoxicology
21, 353–364.
Kobayash, K., Tagawa, S., Daff, S., Sami, I., and Shimizu, T. (2001). Rapid
calmodulin-dependent interdomain electron transfer in neuronal nitric-oxide
synthase measured by pulse radiolysis. J. Biol. Chem. 276, 39864–39871.
Kopp, S. J., Perry, H. M., Perry, E. F., and Erlanger, M. (1983). Cardiac
physiologic and tissue metabolic changes following chronic low-level cadmium and cadmium plus lead injestion in the rat. Toxicol. Appl. Pharmacol. 69,
149–160.
Kuthan, H., and Ullrich, V. (1982). A quantitative test for superoxide radicals
produced in biological systems. Biochem. J. 203, 551–558.
Lippard, J., and Berg, J. (1994). Principles of Bioinorganic Chemistry.
University Science Books, Sausalito, California.
Matsuda, H., and Iyanagi, T. (1999). Calmodulin activates intramolecular
electron transfer between the two flavins of neuronal nitric oxide synthase
flavin domain. Biochim. Biophys. Acta 1473, 345–355.
Mayer, B., John, M., Heinzel, B., Werner, E. R., Wachter, H., Schultz, G.,
and Böhme, E. (1991). Brain nitric oxide synthase is a biopterin- and flavin-containing multi-funcitonal oxido-reductase. FEBS Lett. 288, 187–191.
Miller, R. T., Martasek, P., Omura, T., and Masters, B. S. S. (1999). Rapid kinetic
studies of electron transfer in the three isoforms of nitric oxide synthase.
Biochem. Biophys. Res. Commun. 265, 184–188.
Mills, J., and Johnson, D. (1985). Metal ions as allosteric regulators of
calmodulin. J. Biol. Chem. 260, 15100–15105.
Mittal, C. K., Harell, W. B., and Mehta, C. S. (1995). Interaction of
heavy metal toxicants with brain nitric oxide synthase. Mol. Cell. Biochem.
149/150, 263–265.
Moncada, S., and Higgs, A. (1993). The L-arginine-nitric oxide pathway.
N. Engl. J. Med. 329, 2002–2012.
Murphy, M. E., and Noack, E. (1994). Nitric oxide assay using hemoglobin
method. Methods Enzymol. 233, 241–250.
Nathan, C., and Xie, Q.-W. (1994). Nitric oxide synthases: Roles, tolls and
controls. Cell 78, 915–918.
Nishimura, J., Narayanasami, R., Miller, R., Roman, L., Panda, S., and Masters,
B. (1999). The stimulatory effects of Hofmeister ions on the activities of
neuronal nitric-oxide synthase. J. Biol. Chem. 274, 5399–5406.
Ozawa, T., Sasaki, K., and Umezawa, Y. (1999). Metal ion selectivity for the
formation of calmodulin-metal-target peptide ternary complex studied by surface plasmon resonance spectroscopy. Biochim. Biophys. Acta 1434, 211–220.
Palumbo, A., Astarita, G., Picardo, M., and d’Ishchia, M. (2001). Ni21, a doubleacting inhibitor of nueronal nitric oxide synthase interfering with L-arginine
Pou, S., Pou, W. S., Bredt, D. S., Snyder, S. H., and Rosen, G. M. (1992).
Generation of superoxide by purified brain nitric oxide synthase. J. Biol.
Chem. 267, 24173–24176.
Roman, L. J., Martasek, P., and Masters, B. S. S. (2002). Intrinsic and extrinsic
modulation of nitric oxide synthase activity. Chem. Rev. 102, 1179–1189.
Roman, L. J., Martásek, P., Miller, T. R., Harris, D. E., de la Garza, M. A.,
Shea, T. M., Kim, J. P., and Masters, B. S. S. (2000). The C terminal of constitutive nitric-oxide synthases control electron flow through the flavin and
heme domains and affect modulation of calmodulin. J. Biol. Chem. 275,
29225–29232.
Roman, L. J., Sheta, E. A., Martasek, P., Gross, S. S., Liu, Q., and Masters, B. S. S.
(1995). High-level expression of functional rat neuronal nitric oxide synthase
in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 92, 8428–8432.
Rosen, G. M., Tsai, P., Weaver, J., Porasuphatana, S., Roman, L. J., Starkov, A. A.,
Fiskum, G., and Pou, S. (2002). The role of tetrahydrobiopterin in the
regulation of neuronal nitric oxide synthase-generated superoxide. J. Biol.
Chem. 277, 40275–40280.
Roskams, A. J., Bredt, D. S., Dawson, T. M., and Ronnett, G. V. (1994). Nitric
oxide mediates the formation of synaptic connections in developing and
regenerating olfactory receptor neurons. Neuron 19, 893–903.
Schrammel, A., Gorren, A., Stuehr, D., Schmidt, K., and Mayer, B. (1998).
Isoform-specific effects of salts on nitric oxide activity. Biochim. Biophys.
Acta 1387, 257–263.
Sheta, E. A., McMillan, K., and Masters, B. S. S. (1994). Evidence for a bidomain
structure of constitutive cerebellar nitric oxide synthase. J. Bol. Chem. 269,
15147–15153.
Stolze, K., Udilova, N., Rosenau, T., Hofinger, A., and Nohl, H. (2003). Synthesis
and characterization of EMPO-derived 5,5-disubstituted 1-pyrroline N-oxides
as spin traps forming exceptionally stable superoxide spin adducts. Biol. Chem.
384, 493–500.
Thun, M. J., Osorio, A. M., and Schober, S. (1989). Nephropathy in cadmium
workers: Assessment of risk from airborne occupational exposure to cadmium.
Br. J. Ind. Med. 46, 689–697.
Tian, L., and Lawrence, D. A. (1996). Metal-induced modulation of nitric oxide
production in vitro by murine macrophages: lead, nickel, and cobalt utilize
different mechanisms. Toxicol. Appl. Pharmacol. 141, 540–547.
Tsai, P., Ichikawa, K., Mailer, C., Pou, S., Halpern, H. J., Robinson, B. H.,
Nielsen, R., and Rosen, G. M. (2003). Esters of 5-carboxy-5-methyl-1pyrroline N-oxide: A family of spin traps for superoxide. J. Org. Chem. 68,
7811–7817.
Weaver, J., Porasuphatana, S., Tsai, P., Cao, G.-L., Budzichowski, T. A.,
Roman, L. J., and Rosen, G. M. (2002). The activation of neuronal nitricoxide synthase by various divalent cations. J. Pharmacol. Exp. Therap. 302,
781–786.
Weaver, J., Tsai, P., Cao, G.-L., Roman, L. J., and Rosen, G. M. (2003). Modified
ferricytochrome c can be used to estimate nitric oxide synthase generation of
superoxide. Anal Biochem. 320, 141–143.
Yamazaki, J., Ohara, F., Harada, Y., and Nagao, T. (1995). Barium and Strontium
can substitute for calcium in stimulating nitric oxide production in the endothelium of canine coronary arteries. Jpn. J. Pharmacol. 68, 25–32.
Yoneyama, H., Yamamoto, A., and Kosaka, H. (2001). Neuronal nitric oxide
synthase generates superoxide form the oxygenase domain. Biochem. J. 360,
247–253.