(Received for publication, March 3, 1997, and in revised form, June 18, 1997)
From the Zentrum der Physiologie, The biological signal molecule nitric oxide (NO)
exists in a free and carrier-bound form. Since the structure of
the carrier is likely to influence the interaction of NO with
macromolecular targets, we assessed the interaction of a
dinitrosyl-iron-dithiolate complex carrying different thiol ligands
with glutathione reductase. The enzyme was irreversibly inhibited
by dinitrosyl-iron-di-L-cysteine and
dinitrosyl-iron-di-glutathione in a concentration- and time-dependent manner (IC50 30 and 3 µM, respectively).
Evaluation of the inhibition kinetics according to Kitz-Wilson yielded
a Ki of 14 µM, and a
k3 of 1.3 × 10 Nitrosyl transfer from endogenous nitric oxide
(NO)1 carriers such as
S-nitrosoglutathione (1) and dinitrosyl-iron complex (DNIC)
(2) to macromolecular targets is regarded as one major mechanism of
biological NO signaling (3). Evidence has been provided that endogenous
low mass S-nitrosothiols (predominantly S-nitrosoglutathione) and proteinaceous
S-nitrosothiols (S-nitroso-hemoglobin, S-nitroso-serum albumin, and other yet unidentified
proteins) exist in human erythrocytes (4), plasma (5) and bronchial secretion (6). A further NO adduct with GSH, GSNOH, was recently postulated as another transport form of NO (7).
S-Nitrosation of protein thiols or subsequent reactions such
as ADP-ribosylation (8), formation of protein disulfide (9), and
cysteine sulfenic acid (10) may influence protein function by
allosteric mechanisms. Furthermore, reversible S-nitrosation
of cell membrane-bound proteins may be involved in transmembraneous NO
transport (11).
A concept has been derived from established chemistry to account for
the influence of the redox state of the NO moiety on biological NO
transfer reactions. According to this concept nitrosation of
nucleophilic targets occurs by attack of nitrosonium
(NO+)-like species assumed to be present in "NO
carriers" such as N2O3,
S-nitrosothiols, and certain iron-nitrosyl complexes (3, 12). Less attention has been paid to the influence of the carrier structure on the interaction of NO with macromolecular targets. However, it is conceivable that the NO carrier will direct the NO
moiety specifically to macromolecules recognizing the carrier structure, provided the NO adduct is sufficiently stable and the binding kinetics between the carrier and the macromolecules are rapid
enough to outbalance the decomposition of the NO carrier adduct. There
is also evidence that NO adducts may exhibit intrinsic bioactivity
independent of NO release. Thus, the L-stereoisomer of
S-nitrosocysteine was found to exhibit a significantly
higher blood pressure lowering activity compared with the
D-isomer, suggesting the existence of stereospecific
S-nitroso-L-cysteine receptors in the
cardiovascular system (13).
To assess the influence of the carrier structure on the interaction of
NO with a given macromolecule we chose glutathione reductase (GR; EC
1.6.4.2) as a model target, and low mass dinitrosyl-iron complexes with
L-cysteine and glutathione ligands as NO carriers. These
complexes exhibit S-nitrosating activity toward serum
albumin in vitro (2), and protein-bound forms exist in
vivo in animal tissues expressing inducible NO synthase activity
(14). The flavoenzyme GR catalyzes the NADPH dependent reduction of
oxidized glutathione (GSSG) to maintain a high intracellular level of
GSH. GR carries a redox-active disulfide (Cys-58-Cys-63) in its active site which is reduced by electron transfer from NADPH via the flavin
(15). Recently it has been shown that GR is inhibited by certain NO
carriers (S-nitrosoglutathione, sodium nitroprusside, S-nitroso-N-acetyl-DL-penicillamine)
in millimolar concentrations (16), suggesting that GR is a potential
target for nitrosation reactions.
We show here that low mass dinitrosyl-iron complexes in concentrations
that may be present under pathophysiological conditions irreversibly
inhibit GR, possibly via N-nitrosation. We furthermore demonstrate that the inhibitory potency of the dinitrosyl-iron moiety
increases with structural resemblance of the NO carrier to the natural
GR substrate, GSSG.
GR from bovine intestinal mucosa, fatty acid-free
bovine serum albumin (BSA), 2,3-diaminonaphthalene,
L-cysteine, glutathione (oxidized and reduced), diethyl
pyrocarbonate (DEPC), and Sephadex G-25 were supplied by Sigma,
Deisenhofen, Germany. 1-Nitroso-2-hydroxynaphthalene-3,6-disulfonic acid and 1-nitrosopyrrolidine were obtained from Aldrich, Deisenhofen, Germany. NO gas was prepared by reaction of FeSO4 (Fluka,
Buchs, Switzerland) with NaNO2 in 5 N HCl and
was purified by low temperature high vacuum (p = 0.01 mm Hg) distillation (2).
Paramagnetic dinitrosyl-iron complexes of the type
((NO)2Fe(RS)2) × 18 RSH
(RS = L-cysteine, or GSH) were synthesized
by mixing evacuated (5 min high vacuum) solutions of FeSO4
(5 mg/ml) and neutralized thiols (72 mM) in a Thunberg-type
reaction vessel under pure NO gas (PNO 500 mm Hg). NO was
added 3 min before mixing. The solution immediately turned dark green
and was evacuated after 1 min for a further 2 min to remove excessive
NO. The solution bearing ((NO)2
Fe(RS)2) in >98% yield with respect to iron
was immediately frozen and stored in liquid nitrogen.
S-Nitroso-L-cysteine and
S-nitroso-BSA were prepared at 4 °C by mixing either
L-cysteine (100 mM) or fatty acid-free BSA (2 mM) for 10 min with an equimolar amount of sodium nitrite
dissolved in 0.5 M H2SO4. The
S-nitrosothiols were frozen and stored in liquid nitrogen.
The yield of both S-nitrosothiols was >90% with respect to
free thiol added (BSA-thiol/BSA = 0.4 ± 0.04), using molar
absorption coefficients of S-nitroso-BSA ( At
concentrations > 1 µM S-nitrosothiols
were assessed by diazotization of sulfanilamide and azocoupling with
N,N-ethylendiamine in the presence and absence of
Hg2+ ions (3 mM) according to Saville (18) as
described recently (2).
In the nanomolar range nitrite and S-nitrosothiols were
quantified by an acid-catalyzed intramolecular diazotization reaction of 2,3-diaminonaphthalene with nitrite forming the highly fluorescent product 2,3-diaminonaphthotriazole (19). The sample (570 µl) was
mixed with 30 µl of 0.1 M potassium Pi
buffer, pH 7.4, and 84 µl of freshly prepared 2,3-diaminonaphthalene
(0.05 mg/ml in 1 M HCl). To release NO from
S-nitrosothiols the buffer contained 20 mM
HgCl2. Hg2+ did not interfere with the
fluorescent assay. After 10 min of incubation at 20 °C in the dark
the reaction was terminated by addition of 42 µl of 4.5 M
NaOH to maximize the intensity of the fluorescent signal (19). The
fluorescence was measured with excitation at 375 nm and emission at 415 nm (Deltascan, Photon Technology). Emitted light was detected by a
photon-counting photomultiplier (D-104, Photon Technology) and the
photomultiplier digital output was collected by an IBM-compatible
computer. The content of S-nitrosothiol was calculated by
the difference of emission readings of Hg2+-containing
versus Hg2+-free samples. A calibration curve
was established in each experiment with freshly synthesized
S-nitroso-L-cysteine and sodium nitrite as
standards (0.02-2 µM). The detection limit was 20 nM.
EPR spectra were recorded on a Bruker EPR
300E spectrometer at 20 °C on solutions (25 µl) filled in a quartz
capillary tube (1 mm, inner diameter). The measurements were performed
with a modulation amplitude of 1 Gauss, a microwave frequency of 9.6 GHz, a microwave power of 20 mW, and a time constant of 0.2 s. The
concentration of dinitrosyl-iron complex was calculated by comparison
with the EPR signal of a standard low molecular mass dinitrosyl-iron
complex based on double integration of the first derivative EPR
signals.
Histidine
residues were carboxylated by adding a 100-fold molar excess of DEPC
and subsequent incubation for 5 min at 20 °C (20). Thiol groups were
blocked by incubation of the protein (5 µM protein in 0.1 M potassium phosphate buffer, pH 7.4) for 5 min at 20 °C
with Hg2+ (15 µM) in the presence of NADPH (1 mM). To further demonstrate the involvement of catalytic
site thiols in DNIC-induced inhibition of GR the enzyme (5 µM) was preincubated with Hg2+ (15 µM; 5 min at 20 °C) prior to incubation with DNIC-GSH
(30 µM; 10 min at 37 °C). Controls were performed
without Hg2+ and in the absence and presence of DNIC. The
solutions (100 µl) were desalted at 5 °C by passing over a
Sephadex G50 Nick® column (Pharmacia) equilibrated with assay buffer
(see below) and then incubated with dithiothreitol (DTT) (10 mM, 37 °C) for up to 70 min. After 10, 30 or 70 min of
DTT-treatment the activity of GR was assessed in 1:250 diluted aliquots
as described below.
The
reduction of GSSG by GR was determined at 20 °C by monitoring the
oxidation of NADPH at 340 nm ( Irreversible inhibition
kinetics were assessed according to Kitz and Wilson (22) based on the
following reaction scheme (Equation 1).
Institut für Biochemie II,
3
s
1. A participation of catalytic site thiols in the
inhibitory mechanism was indicated by the findings that only the
NADPH-reduced enzyme was inhibited by dinitrosyl-iron complex and that
blockade of these thiols by Hg2+ afforded protection
against irreversible inhibition. This inhibition was not accompanied by
formation of a protein-bound dinitrosyl-iron complex and/or
S-nitrosation of active site thiols (Cys-58 and Cys-63).
However, one NO moiety exhibiting an acid lability similar to a
secondary N-nitrosamine was present per mol of inhibited monomeric enzyme. These findings suggest specifically
N-nitrosation of glutathione reductase as a likely
mechanism of inhibition elicited by dinitrosyl-iron complex and
demonstrate in general that structural resemblance of an NO carrier
with a natural ligand enhances NO+ transfer to the
ligand-binding protein.
Materials
338 = 870 M
1 cm
1, 1) and
S-nitroso-L-cysteine (
547 = 16.7 M
1 cm
1) (17).
340 = 6200 M
1 cm
1) (21). The enzyme was
diluted in assay buffer (200 mM potassium chloride, 1 mM EDTA, 50 mM potassium phosphate, pH 6.9). To
avoid an interference by NADPH-oxidase activity both reference and
sample cuvettes contained NADPH (0.38 mM) and GR (3.5-20
nM) in a final volume of 1 ml. The reaction was started by
addition of GSSG (1 mM) to the sample cuvette. The enzyme
activity was calculated from the initial rate of the absorbance
decrease at 340 nm during 3 min of incubation. For establishment of
concentration-response relationships, 1-5 µM GR was
preincubated with inhibitors for 30 min and then diluted 300-1000-fold
in the final assay mixture. For characterizing individual residues of
the bovine GR, the numbering system of the well studied human enzyme
(15, 16) was used; the active site thiols are Cys-58 and Cys-63, and
the catalytic imidazole is His-467.
If k3, the rate constant for
transformation of the reversible enzyme-inhibitor complex (E
(Eq. 1)
I) into the inhibited enzyme (E*), is relatively small,
enzyme (E) and inhibitor (I) are in equilibrium with the
reversible inhibitor enzyme complex. Also, in case of irreversible
inhibition k4 can be neglected. The integrated rate law derived from mass balances reads
For [I]
(Eq. 2)
[E0],
or
(Eq. 3)
k3 is the first order rate constant at high
inhibitor concentrations ([I]
(Eq. 4)
Ki). At low
inhibitor concentrations the kinetics are in accordance with a simple
bimolecular mechanism (kapp = (k3/Ki)·I), and the second
order rate constant is
(k3/Ki).
Ki and k3 were derived from a
double-reciprocal plot of the apparent first order rate constants
kapp versus concentration of
inhibitor [I], i.e. dinitrosyl-iron complex. It should be
noted that Ki differs from the half-maximal
inhibitory concentration (IC50), since
Ki describes the reversible
pre-equilibrium, while the IC50 refers to the subsequent
irreversible reaction.
To assess the
release of bioactive NO from inhibited GR soluble guanylyl cyclase (GC)
purified to apparent homogeneity from bovine lung was used as a
detector system (23). GC activity was measured by the formation of
[32P]cGMP from [-32P]GTP (23). Inhibited
GR- or NO-containing substances (40 µl) were incubated for 10 min at
37 °C in a final volume of 100 µl with GC assay buffer. The final
mixture contained triethanolamine-HCl (50 mM, pH 7.4),
3-isobutyl-1-methylxanthine (0.5 mM),
[
-32P]GTP (0.2 mM; 0.2 µCi), cGMP (0.1 mM), GSH (2 mM), MgCl2 (3 mM), creatine phosphate (10 mM), creatine
phosphokinase (5 units),
-globulin (0.1 mg/ml), superoxide dismutase
(0.3 µM), and 0.8 µg of purified GC. Enzymatic cGMP
formation was stopped by addition of zinc acetate (450 µl; 110 mM) and sodium carbonate (450 µl; 120 mM).
[32P]cGMP was isolated by chromatography on acid alumina
and quantified by liquid scintillation counting.
To study the influence of DNIC on
the activity of isolated GR, the enzyme was incubated with different
concentrations of DNIC-L-cysteine and DNIC-GSH in the
presence of NADPH and substrate (GSSG). The GSSG-driven consumption of
NADPH was monitored continuously by recording the decrease in
absorbance at 340 nm (Fig. 1). During 3 min of reaction at 20 °C, samples containing DNIC-GSH exhibited a
nearly constant rate of decrease in absorbance (data not shown), which
was inversely related to the concentration of DNIC used. In contrast,
NADPH consumption in DNIC-L-cysteine containing reaction mixtures initially exhibited exponential kinetics (Fig. 1, A
and B), indicating that the degree of inhibition of GR by
this DNIC derivative progressively increased with time. At longer
preincubation periods (>20 min) in the absence of GSSG the kinetics of
inhibition by DNIC-L-cysteine became more linear (data not
shown). Thus, it is conceivable that during incubation of
DNIC-L-cysteine with GR in the presence of GSSG DNIC-GSH
was formed by reaction with enzymatically generated GSH. In addition,
during preincubation in the absence of GSSG an intrinsic slow
inhibitory action of DNIC-L-cysteine was revealed (Fig.
1B). Therefore, to establish the dose-response relationship
for inhibition of GR by DNIC, the enzyme (1 µM) was
preincubated at 20 °C in the absence of GSSG with different
concentrations of low molecular mass DNIC (6-200 µM) in
buffer containing 1 mM NADPH. The enzyme activity was then measured after 30 min of preincubation. It decreased in a DNIC concentration-dependent manner (Fig.
2). The GSH complex was about 10-fold
more potent than the L-cysteine complex, although both were
equally efficacious. 50% inhibition was elicited by 3 ± 1 µM DNIC-GSH and by 30 ± 3 µM
DNIC-L-cysteine. No direct oxidation of NADPH or reduction
of NADP+ by DNIC could be observed.
GR carries a redox-active disulfide (Cys-58-Cys-63) at its catalytic site, which is reduced after binding of NADPH. To investigate whether the inhibition by DNIC depends on the redox state of the enzyme, GR (5 µM) was incubated with DNIC-GSH (30 µM) at 37 °C in the presence and absence of NADPH (1 mM). The activity of GR decreased to 5 ± 2% (n = 4) of control in the presence of NADPH, but did not change in its absence (100 ± 3%, n = 4). This finding suggests that only the reduced enzyme is susceptible to inhibition by DNIC. Since the homodimeric form is essential for catalytic function of GR (15) we assessed whether DNIC influenced the aggregation state of the native enzyme by gel permeation chromatography (Superose 200, Pharmacia). DNIC-treated and untreated GR migrated as single peaks at identical positions with an apparent molecular mass of 100 kDa. Therefore DNIC does not inhibit GR by promoting dissociation of the homodimeric enzyme.
NO-mediated oxidation and S-nitrosation of protein thiols usually is reversible by an excess of low molecular weight thiols (1, 24, 25). To test the reversibility of the GR modification by DNIC, the complex-inhibited GR was treated with 5 mM DTT or dimercaptopropanol (30 min, 20 °C); neither reducing agent was able to restore enzyme activity. Moreover, dilution (1000-fold) of the inhibited enzyme in thiol-free or thiol-containing buffer failed to restore activity even after 24 h. To test the accessibility of low mass thiols to the catalytic site thiols in GR, we analyzed the reversibility of Hg2+-elicited inhibition of GR by dithiols. Hg2+ exhibits a high affinity for thiol groups. GR (5 µM) was inhibited by the treatment with Hg2+ (15 µM) within 60 s at 20 °C. DTT and dimercaptopropanol regenerated the initial activity (5 mM, 10 min, 20 °C) of the enzyme, indicating that the Hg2+-bound thiol groups at the active center were accessible to both agents (see also Williams (15)). Therefore the dithiols used should be able to interact with the thiol groups in the complex-inhibited enzyme without sterical hindrance. Altogether these findings show that inhibition of GR by DNIC is irreversible.
To assess the involvement of cysteine-thiols within the catalytic site in DNIC-induced inhibition of GR we examined whether or not pretreatment of GR by Hg2+, which inhibits GR in a thiol-reversible manner (see above), affords protection against the irreversible inhibition of GR by DNIC-GSH. Since GR contains 3 cysteine thiols per subunit, NADPH-reduced GR was pretreated with a 2-fold molar excess of Hg2+ prior to incubation with a maximally inhibitory concentration of DNIC-GSH (30 µM). Half of the original GR activity was restored within 10 min following incubation of Hg2+/DNIC- and Hg2+- treated GR with DTT. DTT, however, failed to restore catalytic activity to GR treated with DNIC only. The degree of inhibition after 10 min of DTT treatment was: DNIC-treated GR, 85 ± 10%; Hg2+-treated GR, 45 ± 4%; Hg2+/DNIC-treated GR, 52 ± 7% (mean ± S.E.; n = 3). This inhibition was not significantly altered after either 30 or 70 min of treatment with DTT. Thus, Hg2+ pretreatment protects GR from inhibition by DNIC. Consequently, thiols within the catalytic site of GR appear to be the main targets of DNIC and are involved in the irreversible inhibition.
Kinetics of InhibitionTo study the kinetics of GR inhibition
by DNIC the enzyme (1 µM) was exposed at 20 °C to 0, 6, 10, 15, 20, 35, 50, or 200 µM DNIC-GSH in assay buffer
containing 1 mM NADPH. Aliquots were taken and tested for
GR activity after different time intervals (0, 2, 6, 12, 18, 24, and 30 min). The time course of inhibition displayed first order kinetics at
all DNIC concentrations used. In Fig. 3
the natural logarithm of the ratio from the remaining activity
(E) and the initial activity (E0) is
plotted versus the time of incubation yielding straight
lines according to Equations 2-4 (see "Experimental Procedures").
The slopes represent the rate constants kapp for
the inhibition by the corresponding DNIC concentration and were
replotted according to Kitz-Wilson (1/kapp
versus 1/[I]) as shown in Fig. 4.
A linear relationship was apparent which was used to calculate the
kinetic constants by linear regression analysis (r = 0.95). The constant for the conversion of the reversible enzyme-inhibitor-complex to the irreversibly inhibited enzyme (k3) amounted to 1.3 × 103
s
1, the dissociation constant of the reversible complex
(Ki) was 14 µM.
Chemical Characterization of DNIC-modified GR
The following
experiments were performed to reveal the molecular mechanism of
inhibition of GR by DNIC. The dinitrosyl-iron moiety of low molecular
DNIC binds to free thiol groups of proteins due to a thiol-ligand
exchange reaction (2, 26). To assess whether or not the dinitrosyl-iron
group was attached to the inhibited protein, GR (5 µM)
was incubated with different concentrations of DNIC-GSH or
DNIC-Lcysteine (5-50 µM) at 20 or 37 °C.
After 3, 10, and 60 min the mixture was analyzed by EPR spectroscopy at
20 °C. Recording EPR signals at room temperature allows
discrimination between DNIC bound to low and high molecular mass
ligands, since the former exhibits an isotropic signal at
gav 2.03 with 13-line hyperfine structure (Fig.
5a), while the latter is
characterized by an anisotropic signal at g 2.04 and
g
2.01 (Fig. 5c). The initial reaction
mixtures exhibited exclusively the EPR signal of the low molecular
weight DNIC (Fig. 5b). After 60 min of reaction this signal
completely disappeared (Fig. 5d) because of decomposition of
low mass DNIC. The characteristic signal of the protein-bound DNIC
(serum albumin-DNIC; Fig. 5c) was not detectable at any
time, though the enzyme was completely inhibited after 30 min of
incubation. These findings show that inhibition of GR does not involve
formation of a stable DNIC-protein linkage.
Since free cysteine thiols within proteins can be nitrosated by low molecular weight DNIC (2), we assessed whether this covalent modification accounts for inhibition of GR by DNIC. Therefore DNIC-inactivated enzyme (5 µM, 95 ± 3% inhibited) was passed through a desalting column (Sephadex G-25) to remove excess inhibitor. The protein was concentrated by centrifugation (Ultrafree 30-kDa cut-off, Millipore) and assayed for NOx and S-nitrosothiol (18). Neither chromatography nor centrifugation reversed inactivation of GR. In the presence of Hg2+ 0.76 ± 0.04 mol nitrite/mol inactive GR was detected (n = 3), but a similar amount of nitrite was found in the absence of Hg2+ (0.75 ± 0.08). This indicates that inhibited GR contains a Griess-reactive NO moiety, which is not bound to a thiol group.
Hence it was investigated whether a N- or C-nitrosation of GR by DNIC accounts for inhibition. N-Nitrosopyrrolidine, a nitrosamine of a cyclic secondary amine, also released NO independently of Hg2+ under the acid conditions of the Griess reaction (0.25 M HCl). After 30 min 50 ± 4 µM nitrite was generated by this agent (100 µM). In contrast, the C-nitroso compound 1-nitroso-2-hydroxynaphthalene-3,6-disulfonic acid (50 µM, 60 min incubation), failed to give a positive Griess reaction, either in the presence or absence of Hg2+. To increase the sensitivity of the Griess reaction the inhibited GR was also assessed for S-nitrosothiols by the 2,3-diaminonaphthalene assay. Under the mildly acid conditions (0.1 M HCl) of this assay neither nitrite nor S-nitrosothiol could be detected. However when the inhibited GR was preincubated in 0.25 M HCl for 30 min at 37 °C and then neutralized with NaOH, 0.83 ± 0.05 mol nitrite/mol enzyme were found by the 2,3-diaminonaphthalene assay. This finding reveals that the NO moiety bound to GR exhibits a peculiar acid lability. A similar acid lability was exhibited by N-nitrosopyrrolidine. The N-NO bond was split in 0.25 M HCl but not in 0.1 M HCl.
To further substantiate that the NO moiety is firmly bound to DNIC-inhibited GR under neutral conditions, purified soluble guanylyl cyclase (GC) was used as a sensitive NO detector. DNIC-inhibited GR was desalted and then incubated with GC for assessment of cGMP formation. The basal GC activity was not influenced by DNIC-inhibited GR (400 nM). In contrast, the reference S-nitroso protein S-nitroso-serum albumin at 10-fold lower concentration (40 nM) stimulated GC activity 17-fold. N-Nitrosopyrrolidine (4 µM) and 1-nitroso-2-hydroxynaphthalene-3,6-disulfonic acid (4 µM) did not enhance GC activity (data not shown). These findings tend to exclude the possibility that DNIC-inhibited GR carries a labile S-nitrosothiol moiety, but favor the concept of the formation of a stable N-nitroso group.
According to the tertiary structure of GR derived from x-ray diffraction analysis, a histidine residue (His-467) is located in the direct vicinity of Cys-58 (35). During catalysis of GSSG reduction His-467 takes a proton from Cys-58, thereby facilitating the nucleophilic attack of the resulting thiolate anion on one sulfur atom of the substrate GSSG (27). Consequently one imidazole-nitrogen of His-467 could function as an NO+ acceptor. To demonstrate the key function of His-467 in catalysis, the histidine residues of GR (2 µM) were carboxylated by DEPC (200 µM) at 20 °C. DEPC treatment completely inactivated GR within 60 s (n = 3; data not shown). No direct oxidation of NADPH or reduction of NADP+ by DEPC could be observed, indicating that the modification of histidine residues accounted for inhibition of the enzyme. This finding supports the notion that covalent modification of a histidine, presumably His-467, by N-nitrosation at an imidazole nitrogen could account for inhibition of GR by DNIC.
One objective of the present study was to demonstrate that the structure of NO-carrier complexes influences the interaction of NO with macromolecular targets. Since GR was previously shown to be inhibited by S-nitrosothiols (16), we chose this enzyme as a model target to study its interaction with another class of biological NO carriers, DNIC. Two different types of DNIC were used with L-cysteine and GSH as thiol ligands, and the inhibitory reaction of GR with these compounds was studied in detail to reveal the molecular mechanism underlying the inhibition of GR by NO carriers.
Inhibition of GR by DNICBoth types of DNIC readily
accomplished inhibition of GR in the presence of NADPH. The potency of
the inhibitors depended on the nature of the thiol ligand, DNIC-GSH
being a more potent inhibitor (IC50 = 3 µM)
than DNIC-L-cysteine (IC50 = 30 µM). In comparison, the IC50 of other NO
donors such as sodium nitroprusside, S-nitrosoglutathione,
or
S-nitroso-N-acetyl-DL-penicillamine
are at least 100-fold higher (16), and also the nitrosourea diethyl [1-[3-(2-chloroethyl)-3-nitrosoureido]ethyl]phosphonate
(Fotemustine) exhibits an IC50 of 1.5 mM (28).
Therefore, DNIC are the most potent inhibitors of GR reported to date.
Assuming that these complexes attack the catalytic center of GR, the
higher inhibitory potency of DNIC-GSH compared with
DNIC-L-cysteine may be explained by its similarity to the
natural substrate GSSG (see Structure 1),
which favors the interaction of DNIC-GSH with the active center of the
enzyme. Since inhibition of the enzyme was not reversible by dilution
(1000-fold) and progressed exponentially with time, inhibition kinetics
could be evaluated according to Kitz and Wilson (22). The apparent rate
constants of inhibition correlated with the concentration of DNIC-GSH
applied. A dissociation constant of the reversible complex
Ki of 14 µM and a rate constant of the
conversion of the reversible enzyme-inhibitor complex to the
irreversibly inhibited enzyme k3 of 1.3 × 103 sec
1 were derived from the Kitz-Wilson
replot. Thus DNIC-GSH was bound with high affinity in a rapid
equilibrium preceding the irreversible inhibitory reaction.
Chemical Characterization of DNIC-inactivated GR
The irreversible inhibition of GR suggests a covalent modification of the enzyme by DNIC. A likely target site is the catalytic center of GR, which carries a redox-active disulfide (Cys-58-Cys-63). The requirement for reduced thiols at the catalytic site was evidenced by the finding that DNIC led to inhibition of GR only in the presence of the coenzyme NADPH, and that GR pretreated with Hg2+ was protected against inhibition by DNIC. Other agents which inhibit GR by carbamoylation (1,3-bis(2-chloroethyl)-1-nitrosourea) or alkylation (1-(2-chloroethyl)-3-(2-hydroxyethyl)-1-nitrosourea) of Cys-58 also influence the enzyme activity only after two-electron reduction of the enzyme by NADPH (28, 29). Therefore inhibition of GR by DNIC could involve a stable attachment of the dinitrosyl-iron moiety to one or both cysteines via a ligand exchange reaction (30) or a S-nitrosation of one or both cysteines, as has been shown for the interaction of DNIC with serum albumin (2).
Low molecular weight DNIC did not react with GR to form a protein-bound DNIC, as assessed by EPR spectroscopy (Fig. 5). This finding excludes that GR is inhibited by a linkage of the dinitrosyl-iron group to the active site cysteine residues. We next assessed the formation of a S-nitroso group in DNIC-inhibited GR by the Saville reaction. No S-nitrosation was detectable, although 1 mol of the DNIC-inhibited enzyme contained about 0.8 mol/subunit Griess-positive NOx released from the enzyme by 0.25 M HCl, but not by 0.1 M HCl. Similar acid lability of the NO moiety was observed with N-nitrosopyrrolidine, a nitrosamine of a cyclic secondary amine, whereas the C-nitroso bond of 1-nitroso-2-hydroxynaphthalene-3,6-disulfonic acid was acid-resistant. This result indicates that DNIC-inhibited GR bears a N-nitroso moiety.
A similar conclusion was derived from the comparison of the guanylyl cyclase-stimulating activity of DNIC-inhibited GR, S-nitrosoalbumin, N-nitrosopyrrolidine, and 1-nitroso-2-hydroxynaphthalene-3,6-disulfonic acid. Only S-nitrosoalbumin was able to enhance guanylyl cyclase activity.
Since in vitro a nitrosation of cyclic secondary amines by S-nitrosothiols has been described (31), and a histidine residue (His-467) located in the vicinity of Cys-58 participates in the catalytic reduction of GSSG, we considered N-nitrosation of this residue as a likely mechanism of GR inhibition by DNIC. In fact, GR activity was also lost after N-carboxylation of histidine residues by DEPC. This finding implies that blocking of an imidazole nitrogen could account for inhibition of GR activity by DNIC.
Our results are consistent with the following hypothetical mechanism of
GR inhibition by DNIC (Fig. 6). In a
rapid pre-equilibrium DNIC interacts reversibly with active site
thiols. DNIC then S-nitrosates Cys-58, which is
characterized by a higher electronegativity than Cys-63 (32) and is
located at the entrance of the active site. This reaction is
immediately followed by an intramolecular trans-nitrosation yielding a
N-nitrosamine, presumably at the imidazole nitrogen of
His-467. In this context it should be noted that the covalent modification achieved by bichloroethylnitrosourea is stabilized by a
hydrogen bond between the carbamoyl oxygen and the His-467 (29). This
implies that a trans-nitrosation between Cys-58 and His-467 should be
sterically feasible. However, the involvement of His-467 in the
inhibitory mechanism and its N-nitrosation by DNIC has to be
proven by x-ray diffraction analysis. Since minor species differences
in the protein structure of GR have a great influence on GR's
susceptibility to different kinds of inhibitors, the scheme proposed
here may be valid only for GR from bovine intestinal mucosa.
In conclusion, we have shown that low mass DNIC irreversibly inactivate GR and exhibit a much higher inhibitory potency than other known inhibitors of GR. From our study two biochemical principles might emerge. First, N-nitrosation is a novel mechanism by which proteins may be post-translationally modified and which broadens the spectrum of NO-mediated signal transduction. Adjacent cysteine and histidine residues would be expected to be susceptible to this modification. As a consequence of its size and charge distribution, DNIC should nitrosate proteins more selectively than NO2/N2O3, as accessibility to DNIC is determined by protein three-dimensional structure in the immediate vicinity of the target thiol. Second, we have shown that the thiol ligand structure influences the interaction of DNIC with macromolecular targets. A similar effect of the NO carrier structure on the interaction of NO with proteins has been reported. For instance, human glutathione peroxidase specifically liberates NO from S-nitrosoglutathione, but not from S-nitroso-L-cysteine (33) or dinitrosyl-iron complex.2 Furthermore, S-nitrosoglutathione exhibits specific toxicity against S. typhimurium, because the peptide is taken up into the microorganism by a specific transporter system (34).