 |
INTRODUCTION |
Human glutaredoxin (thioltransferase)
(GRx,1 EC 1.8.4.2) is known
for its unique properties of specific and efficient catalysis of
deglutathionylation of protein-S-S-glutathione-mixed disulfides (protein-SSG) (1-4). These catalytic properties of glutaredoxin have
identified the enzyme for prominent roles in homeostasis of protein
sulfhydryl groups both in a protective mode under overt oxidative
stress associated with aging and various disease states including
cardiovascular and neurodegenerative diseases, diabetes, AIDS, and
cancer (2, 5, 6) and in a regulatory mode whereby reversible
glutathionylation represents a mechanism of redox-activated signal
transduction (7-11). These physiological roles are supported further
by the documentation that glutaredoxin accounts for essentially all of cellular protein-SSG deglutathionylase activity in
mammalian cells (4, 12, 13) and its inactivation by cadmium is
correlated with inhibition of intracellular deglutathionylase activity
(4, 11).
Although reversible formation of protein-SSG is a prevalent form of
protein sulfhydryl modification, mechanisms of protein-SSG formation
are not resolved. Unless intracellular GSSG concentrations reach
unusually high levels, GSSG is unlikely to be the mediator of
protein-SSG formation based on typical redox potentials for cysteine
residues (14, 15). Consequently, glutathione-thiyl radical and
S-nitrosoglutathione (GS-NO) have been considered as
potential alternative mediators (5, 6, 10). In this context, we
considered the unusually low pKa (=3.5) of the
active site cysteine of glutaredoxin (16, 17) and the selective
stabilization of the glutathionyl moiety in the glutaredoxin-SSG catalytic intermediate (1-3) along with the known ability of thiyl
radicals to be stabilized by the formation of disulfide anion radicals
(18, 19). Accordingly, we reasoned that glutaredoxin might react
preferentially with the glutathione-thiyl radical to form a
glutaredoxin-S-S-glutathione-disulfide anion radical (GRx-SSG
) and that this enzyme intermediate could facilitate transfer of the GS-radical either to form GSSG or protein-SSG adducts.
To test this hypothesis, two GS-radical generating systems were used,
i.e. Fe(II)-ADP/H2O2 + GSH (20, 21)
and HRP/H2O2 + GSH (22). Here we report that
glutaredoxin catalyzes the formation of GSSG in the GS-radical
generating systems. The catalysis of GSSG formation is dependent on
molecular oxygen, and it is distinguished from glutathione
peroxidase-like activity. Moreover, glutaredoxin enhanced the rate of
S-glutathionylation of GAPDH in the presence of GS-radicals.
This model reaction mimics the intracellular formation of GAPDH-SSG
under oxidative conditions where GSSG content is not substantially
changed (23-25). Glutaredoxin also enhanced the rate of formation of
GAPDH-SSG when GSSG or GS-NO were used as the glutathionyl donors, but
these reactions were much less efficient than the GS-radical transfer
reaction. In comparison to GAPDH, both actin and PTP1B were found to be
superior substrates for GRx-facilitated S-glutathionylation
under GS-radical transfer conditions. Both actin and PTP1B are
implicated in redox signal transduction, and both have been shown to be
susceptible to intracellular regulation via reversible
glutathionylation, probably involving glutaredoxin as the
deglutathionylation catalyst (9, 11). Our current findings further
characterize glutaredoxin as a versatile catalyst capable of
facilitating S-glutathionylation of redox signal mediators.
 |
EXPERIMENTAL PROCEDURES |
Materials--
Water was initially purified by reverse osmosis.
It was further purified by a Millipore MilliQ system. Finally, it was
treated with 5 g of Chelex (Bio-Rad)/1 liter to remove potentially
contaminating metals. Ferrous chloride (FeCl2), hydrogen
peroxide, acetic acid, sodium nitrite, sodium dibasic phosphate, and
potassium monobasic phosphate were from Fisher Scientific. ADP,
glutathione reductase from yeast, glutathione peroxidase from bovine
erythrocytes, glucose, trichloroacetic acid, and mono-carboxymethyl-BSA
(BSA-CM) were from Sigma. NADPH, GAPDH, horseradish peroxidase,
acetylated cytochrome c, and glucose oxidase were from Roche
Molecular Biochemicals. [35S]Glutathione was from ICN
Radiochemicals. Glutaredoxin was isolated and purified as described
previously (3). Sodium-potassium phosphate buffer (0.1 M)
was used in all of the assays, diluted from 1 M stock
solution that was prepared by adding 1 M potassium monobasic phosphate to 1 M sodium dibasic phosphate until
the measured pH was 7.5.
Glutathionyl radical mediated formation of GSSG--
Stock
solutions of Fe(II)·ADP complex were made by mixing equal
volumes of FeCl2 (20 mM) and ADP (100 mM) solutions at least 1 h before beginning the
experiments. Complete reaction mixtures contained 0.1 M
sodium-potassium phosphate (pH 7.5), various amounts of glutaredoxin,
0.2 mM NADPH, 0.5 mM GSH, 2 units/ml GSSG
reductase, and premixed 0.5 mM FeCl2, 2.5 mM ADP, and 0.05 mM
H2O2. Reactions were initiated with the
Fe(II)·ADP and H2O2 mixture. Controls omitting glutaredoxin and/or Fe(II)·ADP and
H2O2 were performed in parallel. NADPH
oxidation (equivalent to GSSG formation) was monitored at 340 nm using
the extinction coefficient
340 = 6.2 mM
1 cm
1. Factors correcting for
path length in the plate reader were calculated and used to express
rates as nmol GSSG formed per minute. Glutaredoxin enzyme was added as
indicated in the figure legends, and time courses of product formation
and dependence on enzyme concentration were determined.
Oxygen Depletion Experiments--
Experiments were conducted as
described above with the exception that all of the solutions were
bubbled for 15 min with an N2 lance. In some experiments,
glucose (0.1 mM) and glucose oxidase (300 units/ml) were
added to ensure complete oxygen depletion. Experimental results with
N2 purging in the absence or presence of the glucose
oxidase system were indistinguishable.
Concurrent Detection of Superoxide Formation--
Reactions were
conducted as described above with the exception that partially
acetylated cytochrome c (30-90 µM) was added to the initial mixture. Rates of GSSG formation were determined as
above. Changes in absorbance at 550 nm were recorded in a plate reader,
and rates of cytochrome c oxidation were calculated using
550 = 18.5 mM
1
cm
1.
Glutathionyl Radical-mediated S-Glutathionylation of
GAPDH--
Glutathionyl radicals were generated by horseradish
peroxidase utilizing H2O2 and GSH adapted from
the studies of Harman et al. (22) and Mason et
al. (26). The assays measured the incorporation of
[35S]glutathione into GAPDH. Reaction mixtures contained
0.1 M sodium-potassium phosphate, pH 7.5, 0.2 mg/ml HRP,
0.5 mM [35S]GSH (~0.5 nCi/nmol), 28 µM GAPDH, 2 mg/ml BSA-CM). Glutaredoxin was added in
various concentrations as indicated (see "Results"), and
H2O2 was added to initiate the reaction.
Reactions were quenched either immediately after initiation or at
various reaction times by adding trichloroacetic acid (10% final
concentration) to precipitate the proteins. The precipitates were
washed twice with 20% trichloroacetic acid and then solubilized
with scintillation fluid and transferred into vials, and counts/minute
were determined. BSA-CM served as a non-thiol-containing bulk protein
to facilitate quantitative precipitation of GAPDH. The amount of
GAPDH-SSG [35S]mixed disulfide was quantified
according to the specific radioactivity calculated from GSH content and
measurement of radiolabel in each reaction mixture. Incorporated
radioactivity was confirmed to represent GAPDH-SSG formation by
documenting quantitative release of radiolabel with 10 mM dithiothreitol.
Preparation of Other Radiolabeled Glutathione-containing
Oxidants--
Glutathione disulfide (35S-labeled, ~1
nCi/nmol) was prepared by incubating 5 mM GSSG with a trace
amount of [35S]GSH overnight at room temperature.
S-Nitrosoglutathione was prepared as described previously
(27) by combining [35S]GSH with equimolar amounts of
acetic acid and sodium nitrite at room temperature for
5 min. We
confirmed in separate model preparations with unlabeled GSH that this
reaction gives stoichiometric conversion based on the extinction
coefficient of GS-NO at 338 nm (
338 = 980 M
1 cm
1) (28).
Glutathionylation of GAPDH by Various Oxidants--
GS-radical,
GS-NO, and GSSG were compared as glutathionyl-donors for their relative
ability to support S-glutathionylation of GAPDH in the
absence or presence of glutaredoxin. Reaction mixtures contained 0.1 M sodium-potassium phosphate, pH 7.5, 2 mg/ml BSA-CM, and
28 µM GAPDH, and one of the GS-donors was added to
initiate the reaction as follows: [35S]GSSG (~1
nCi/nmol; 0.05 or 0.5 mM) or [35S]GS-NO (~1
nCi/nmol; 0.05 or 0.5 mM) or GS-thiyl radical (~0.5 nCi/nmol; estimated as
0.01 mM). The concentration of
GS-radicals probably represents an overestimate based on the amount of
GSSG accumulated (GSSG reductase assay) in 15 min in the HRP GS-radical generating system. The reactions were quenched with trichloroacetic acid, and [35S]GAPDH-SSG was quantified according to the
specific radioactivity of the respective glutathionyl donor as
described above.
 |
RESULTS |
Glutaredoxin Catalysis of GS-Radical Dependent GSSG
Formation--
Production of reactive oxygen species in cells is
broadly implicated in signal transduction mechanisms as well as
pathophysiological changes. However, characterization of the proximal
mediators of these events is still unclear and under active
investigation. To examine the participation of glutathione thiyl
radical and the potential roles of glutaredoxin in regulating these
events, we used the well known system comprised of the Fe(II)·ADP
complex and H2O2 (20, 21, 29) to produce
hydroxyl radicals in the absence or presence of GSH and glutaredoxin.
Hydroxyl radical production in the absence of GSH was documented by the
conversion of salicylate to 2,3-dihydroxybenzoate as reported
previously (30), and the radical transfer reaction of hydroxyl radicals with GSH to produce glutathionyl thiyl radicals (18) was confirmed by
documenting GSH concentration-dependent inhibition of
2,3-dihydroxybenzoate formation in the
Fe(II)-ADP/H2O2 system with salicylate (data not shown). Although glutathionyl radical can form the disulfide anion
radical (GSSG
) and react with oxygen to form superoxide and GSSG (31, 32), this overall reaction is disfavored at neutral pH
(33) (See "Discussion," Scheme 1).
Upon the addition of glutaredoxin, the formation of GSSG is accelerated
over a time course that is linear for at least 5 min, consistent with a
catalytic role for the enzyme. Accordingly, the reaction is enzyme
concentration dependent (Fig. 1), and it
depends on the native enzyme for activity, i.e. boiled
enzyme is inactive (data not shown).

View larger version (23K):
[in this window]
[in a new window]
|
Scheme 1.
Formation and distribution of
glutathione-thiyl radical: Role of glutaredoxin. The scheme
depicts the relative stability of the disulfide anion radicals of GSSG
and GRx-SSG.
|
|

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1.
Dependence of GS-radical-mediated GSSG
formation on glutaredoxin concentration. Glutaredoxin (0-0.88
µM) was preincubated at 30 °C with 0.5 mM
GSH, 0.2 mM NADPH, and 2 units/ml GSSG reductase in 0.1 M sodium/potassium phosphate buffer (see "Experimental
Procedures"). Reactions were initiated by addition of a pre-made
complex (containing 0.5 mM FeCl2 and 2.5 mM ADP) and 50 µM
H2O2. Reactions were monitored continuously at
340 nm, and GSSG formed per minute was calculated from the initial
linear portion of the time courses ( 4 min) according to the
extinction coefficient of NADPH 340 = 6.2 mM 1 cm 1. Inset,
comparison of glutathione peroxidase and glutaredoxin as catalysts of
GSSG formation from GSH and H2O2 in the absence
of GS-radicals. Conditions were as described above with the exception
that the Fe (II)·ADP complex was omitted precluding GS-radical
formation and the H2O2 concentration was varied
as shown. Either glutathione peroxidase (0.002 units/ml,
triangles) or glutaredoxin (1 unit/ml,
rectangles) was added as potential catalysts.
|
|
The accelerated formation of GSSG is distinguished from a glutathione
peroxidase-like activity of glutaredoxin, because glutaredoxin had no
effect on the rate of formation of GSSG with only
H2O2 and GSH in the absence of the Fe(II)·ADP
complex. In contrast, glutathione peroxidase (positive control) added
under the same conditions in the absence of Fe(II)·ADP efficiently
catalyzed H2O2-dependent GSSG
formation as expected. (Fig. 1, inset).
Glutaredoxin Catalysis of GSSG Formation from GS-Radical:
Dependence on Oxygen--
To determine whether the
glutaredoxin-catalyzed reaction involves the participation of molecular
oxygen, we studied the effect of depleting oxygen. Fig.
2 documents that the enzyme-catalyzed reaction requires O2 for efficient turnover and suggests
coincident formation of superoxide (see "Discussion," Scheme
2).

View larger version (13K):
[in this window]
[in a new window]
|
Fig. 2.
Effect of oxygen deprivation on glutaredoxin
catalysis of GSSG formation from GS-radicals. Reactions were
conducted as in Fig. 1 with the exception that for experiments labeled
" O2," all solutions were pre-bubbled with an
N2 lance for 15 min on ice. Glutaredoxin concentration was
0.1 µM.
|
|
To confirm this interpretation, we measured concomitant reduction of
ferricytochrome c, which is diagnostic for superoxide formation (Fig. 3). Estimation of the
stoichiometry of GSSG formation and superoxide formation (cyt
c reduction) is complicated by the presence of
Fe(III)·ADP, which can compete with cyt c3+
for reaction with superoxide (see "Discussion"). Such competition is indicated by the increasing rate of cyt c reduction that
occurred as the cyt c concentration was increased (Fig.
3).

View larger version (44K):
[in this window]
[in a new window]
|
Fig. 3.
Concomitant cytochrome c
reduction (superoxide formation) during glutaredoxin-catalyzed
formation of GSSG from GS-radicals. Reactions were conducted as in
Fig. 1 with the exception that acetylated cytochrome c
(30-90 µM) was added. Reactions were monitored at 340 and 550 nm. GSSG was calculated as in Fig 1, and superoxide was
measured according to the oxidation of acetylated cytochrome
c using the extinction coefficient 550 = 18.5 mM 1 cm 1. Glutaredoxin
concentration was 0.1 µM. As indicated, the solid
black bars represent the rates of formation of superoxide (cyt
c reduced) at several concentrations of cyt c,
and the cross-hatched bars represent the rates of GSSG
formation.
|
|
Glutaredoxin Acceleration of GS-Radical-mediated Protein-SSG
Formation--
A significant consequence of oxidative stress in cells
is the accumulation of protein-SSG-mixed disulfides, possibly
representing a protective mechanism (2, 5).
S-Glutathionylation of specific proteins has now been
demonstrated also upon activation of signaling cascades in cell culture
as well and probably represents a mode of regulation (9, 11). Although
many key proteins with specific cysteine residues that are subject to
reversible glutathionylation have been characterized (2, 10), the
mechanism(s) of formation of protein-SSG is unclear. One such protein
whose glutathionylation has been studied in various contexts is
glyceraldehyde-3-phosphate dehydrogenase. This abundant cytosolic
protein has been observed to accumulate as GAPDH-SSG in cells after
oxidative insults even though the GSSG concentration did not change
substantially (23-25, 34), suggesting that a form of the glutathione
moiety other than GSSG was the proximal mediator of glutathionylation.
Therefore, we tested the relative effectiveness of GS-radical to serve
as the precursor of protein-SSG formation and examined whether
glutaredoxin would facilitate the reaction. In this case, the
GS-radical generating system was horseradish peroxidase, utilizing
H2O2 as oxidant and GSH as reductant. This
system allowed us to minimize direct reaction of
H2O2 with GAPDH (35), because we could use a
concentration of H2O2 10-fold lower than that
for the Fe(II)·ADP/H2O2 system (described
above) and still generate a significant steady-state concentration of
GS-radical over the reaction time. Therefore, S-glutathionylation of GAPDH could be studied under
conditions more closely related to a cellular signaling environment,
i.e. low H2O2. Using this HRP
GS-radical generating system, glutaredoxin did accelerate the rate of
[35S]GAPDH-SSG formation in a time-dependent
and concentration-dependent manner (Fig.
4, inset). No enhancement of
GAPDH-SSG formation occurred when the glutaredoxin enzyme was boiled
before it was added or when either H2O2 or HRP
was omitted from the reaction mixture (data not shown). Therefore, we
conclude that the enhancement of GAPDH-SSG formation requires native
glutaredoxin and it is dependent on GS-radical formation.

View larger version (32K):
[in this window]
[in a new window]
|
Fig. 4.
Time course of glutaredoxin-mediated
GAPDH-glutathionylation with the GS-radical generating system and
dependence on glutaredoxin concentration (inset).
Reaction mixtures contained 0.1 M sodium-potassium
phosphate, pH 7.5, 2 mg/ml BSA-CM, 28 µM GAPDH, 0.5 mM [35S]GSH (~1 nCi/nmol), 0.2 mg/ml HRP,
and 8.6 µM GRx. Reactions were initiated with 0.05 mM H2O2 and allowed to proceed for
up to 15 min. Aliquots (0.02 ml) were withdrawn periodically (as shown)
and added to an equal volume of 20% trichloroacetic acid to
precipitate the proteins. The radioactivity associated with the
precipitated, washed, and resolubilized GAPDH was measured.
Inset, conditions were the same as described above with the
exception that the concentration of GRx was varied as shown (in a total
volume of 0.02 ml), and all of the reactions were allowed to proceed
for 15 min.
|
|
Effect of Oxygen Deprivation on GAPDH-SSG Formation by GS-Radical
in the Absence or Presence of Glutaredoxin--
Fig.
5 shows that removal of oxygen does not
affect net GAPDH-SSG formation from GS-radicals in the presence of GRx
(hatched bars), in contrast to the effect on GSSG formation
(Fig. 2). However, in the absence of GRx, less GAPDH-SSG is formed
under anaerobic conditions. Thus, the enhancement of GAPDH
glutathionylation by GRx is greater in the absence of O2.
This distinction suggests an oxygen-independent rate-limiting step for
the GRx-mediated glutathionylation of proteins (see "Discussion,"
Scheme 4).

View larger version (39K):
[in this window]
[in a new window]
|
Fig. 5.
Effect of oxygen deprivation on GAPDH-SSG
formation from GS-radicals in the absence and presence of
glutaredoxin. Reactions were conducted as in Fig. 4 with
the exception that for experiments labeled Anaerobic, all
solutions were pre-bubbled with an N2 lance for 15 min on
ice. Glutaredoxin concentration was 0.1 µM. The
hatched portion of the right bar in each case
represents the net gain in [35S]GAPDH-SSG because of the
presence of glutaredoxin.
|
|
Acceleration of Protein-SSG Formation by Glutaredoxin: Alternative
GS Donors--
Besides GSSG and the glutathione-thiyl radical,
considerable attention has been focused on GS-NO as a potential
mediator of protein-SSG formation (10, 36). Therefore, we tested the
relative ability of GSSG, GS-radical, and GS-NO to mediate GAPDH-SSG
formation in the absence and presence of glutaredoxin (Table
I). Although both GSSG and GS-NO could
substitute for the GS-radical generating system and glutaredoxin
accelerated GAPDH-SSG formation, the relative efficiency of
these reactions was poor. Comparing the values in the right-most column
of Table I, the GRx-mediated reactions for GS-NO and GSSG at 50 µM (>5× the concentration of GS-radical (<10
µM)) yielded less than half of the amount of GAPDH-SSG
formed in the GS-radical reaction. Only at a very high concentration (500 µM) were GS-NO and GSSG as good or better donors for
GRx-mediated GAPDH-glutathionylation as GS-radical at <10
µM.
View this table:
[in this window]
[in a new window]
|
Table I
Glutaredoxin-mediated GAPDH-S-glutathionylation with various
glutathionyl donors
GS-radical, GS-NO, and GSSG were compared for their relative ability to
support 35S glutathionylation of GAPDH in the absence or
presence of glutaredoxin. Reaction mixtures contained 0.1 M
Na/K phosphate, pH 7.5, 2 mg/ml BSA-CM and 28 µM GAPDH,
and the respective GS -donor (at the concentration indi-cated) was
added to initiate the reaction. The concentration of GS-radicals
probably represents an overestimate based on the amount of GSSG
accumulated (GSSG reductase assay) in 15 min in the HRP GS-radical
generating system. The reactions were quenched with trichloric acid and
[35S]GAPDH-SSG was quantified according to the specific
radioactivity of the respective glutathionyl donor as described under
"Experimental Procedures."
|
|
Glutathionylation of Actin and PTP1B--
Changes in the
intracellular S-glutathionylation status of PTP1B and actin in response
to a physiological redox stimulus have been reported recently; however,
the glutathionylation mechanism has yet to be resolved. Therefore, we
also examined the capability of glutaredoxin to mediate
GS-radical-dependent formation of PTP1B-SSG and actin-SSG
relative to GAPDH-SSG (Fig. 6). Both
actin and PTP1B were much more readily glutathionylated by
glutaredoxin-mediated GS-radical transfer compared with GAPDH. The
upper unshaded portions of the bars at the right
in each set in Fig. 6 represent the GRx-mediated protein-SSG formation,
showing that actin and PTP1B are glutathionylated by the GRx and
GS-radical system >12 times and >15 times more, respectively, than
GAPDH.

View larger version (35K):
[in this window]
[in a new window]
|
Fig. 6.
Relative glutaredoxin-mediated
S-glutathionylation of GAPDH, actin, and PTP1B.
GAPDH, actin, and PTP1B were compared as glutathionyl recipients in the
GS-radical generating system in the absence and presence of
glutaredoxin. Reaction mixtures contained 0.1 M
sodium-potassium phosphate, pH 7.5, 2 mg/ml BSA-CM (as co-precipitant),
and GAPDH or actin or PTP1B, each at 28 µM, in the
presence of the HRP GS-radical generating system as described under
"Experimental Procedures." Reactions were initiated with
H2O2 and allowed to proceed for 15 min.
Proteins then were precipitated with trichloroacetic acid, and the
radioactivity associated with the respective trichloroacetic
acid-precipitated, washed, and resolubilized protein pellets was
measured. The left-most solid black bar represents the
overall minus HRP control. The solid gray bars in
each case represent the minus glutaredoxin controls, and the
right bar in each case represents the plus
glutaredoxin results. The upper unshaded portion of the
right-hand bars represents the GRx-mediated portion of each
glutathionylation reaction, i.e. after subtracting the
minus GRx control data.
|
|
 |
DISCUSSION |
Glutaredoxin Catalysis of Radical Scavenging--
A generalized
scheme has been proposed (31, 32, 37) that depicts the central role of
GSH in mediating the dissipation of radicals within cells. The scheme
involves intermediate formation of the GS-thiyl radical and the
GSSG-disulfide-anion radical. Ultimately, the radical is dissipated by
reaction with molecular oxygen to yield the superoxide anion radical,
which is converted finally to H2O and O2 by the
combined actions of superoxide dismutase and catalase and/or
glutathione peroxidase (see Reactions 1-4 below).
Even though the formation of the glutathione-disulfide anion
radical (Reaction 3) is inefficient because of a relatively low second
order rate constant and the fact that GSH is predominantly in the
protonated form at physiological pH (Reaction 2), the overall reaction
leading to formation of superoxide is facilitated by the favorable
second order rate constant for Reaction 4 (33). In this study, we made
the remarkable discovery that the slow step in this reaction sequence
(Reaction 3) could be supplanted by an enzyme-catalyzed reaction
involving glutaredoxin (thioltransferase). The extent of acceleration
of GSSG formation by GRx demonstrates catalytic turnover of the enzyme.
From Fig. 1, the initial linear region of dependence of GSSG formation
rate on GRx concentration gives a turnover of ~7 min
1
(0.1 s
1). At a typical cellular concentration of GRx
(i.e. 1 µM) (29), 10 µM
GS-radical would be converted to GSSG in 1.5 min. In fact, local
dissipation of the GS-radical could occur much more rapidly, because
its reaction with GRx (Scheme 2, step 1) would occur much faster than
the overall reaction. It seems likely that the properties of
glutaredoxin that impart this catalytic activity for scavenging the
GS-radical would be related to those that are responsible for its
efficient catalysis of protein-SSG deglutathionylation, namely
exquisite selectivity for the glutathionyl moiety (1, 3), and unusually
low pKa of the active site cysteine thiol
(pKa of C-22-SH = 3.5) (16, 17). These
properties would favor stabilization of a GRx-SSG
disulfide anion radical intermediate as depicted in Scheme
1.
To distinguish whether the putative glutaredoxin-glutathione disulfide
anion radical intermediate might be turned over directly by reacting
with a second GS-radical or involve participation of molecular oxygen
(Scheme 2), we studied the effect of
depleting oxygen. The overall data for GRx catalysis of GS-radical
scavenging are consistent with the solid-line portion of
Scheme 2, which is analogous to the reaction scheme we established for
GRx catalysis of reduction of glutathionyl-mixed disulfides (Scheme
3) (2). Catalytic formation of GSSG is
dependent on conditions that generate the glutathione-thiyl radical,
and catalysis is abolished by the removal of molecular oxygen. With
O2 present, concomitant formation of superoxide along with
the GSSG was shown by reduction of cytochrome c. Less than
stoichiometric reduction of cyt c and GSSG formation was
observed, probably because of competition between cytochrome c and Fe(III)·ADP (and GSH) for reaction with the
superoxide. Accordingly, increasing the concentration of cytochrome
c led to increased rates of reduction (Fig. 3).

View larger version (17K):
[in this window]
[in a new window]
|
Scheme 2.
Catalytic mechanism of GS-radical
scavenging by glutaredoxin. The scheme displays the proposed
mechanism of oxygen-dependent glutaredoxin catalysis of
GS-radical scavenging.
|
|

View larger version (16K):
[in this window]
[in a new window]
|
Scheme 3.
Catalytic mechanism of thiol-disulfide
exchange by glutaredoxin. The scheme displays the two-step
glutathione-specific mechanism of thiol-disulfide exchange typically
catalyzed by glutaredoxin (2).
|
|
Thus, Scheme 2 depicts a novel mechanism for scavenging radicals within
cells where GSH is present in abundance and serves as the proximal
acceptor, forming the GS-radical. Then glutaredoxin captures the
GS-radical as the stabilized GRx-SSG
disulfide anion
radical intermediate that can react readily with molecular oxygen,
resulting in net conversion of the original radical to superoxide anion
radical. The superoxide radical would then be dissipated by the coupled
actions of superoxide dismutase and catalase and/or glutathione
peroxidase or the thioredoxin peroxidase system to regenerate molecular
O2 and water as the final products of the scavenging
cascade. Thus, more deleterious radical species (e.g.
hydroxyl radicals) would be eliminated by GSH coupled to the
synergistic catalytic actions of glutaredoxin and superoxide dismutase.
An analogous oxygen-dependent effect of thioredoxin on
turnover of phenoxy-radicals was reported previously (38).
Glutaredoxin-mediated Protein S-Glutathionylation--
There is a
growing interest in reversible S-glutathionylation as a
modulatory mechanism of signal transduction (6, 7, 10, 39). Although
much circumstantial evidence favors this mechanism, there are many
unanswered questions. Perhaps the most prominent question is whether
there are enzymatic mechanisms for controlled formation of specific
protein-SSG adducts as signaling intermediates. This study provides
data pertinent to this consideration and offers an explanation for the
presence of intracellular protein-SSG adducts under conditions where
GSSG does not accumulate sufficiently to support simple thiol-disulfide
exchange consistent with thermodynamic equilibria. Instead,
glutathione-thiyl radical has been suggested as an alternative GS donor
(5, 10) and, consistent with this hypothesis, intracellular GS-thiyl
radicals have been trapped and identified by electron spin resonance
spectroscopy (40, 41). In this study, GAPDH was used as the
primary model to examine whether glutaredoxin catalyzes
GS-radical-mediated protein-SSG formation, and GS-NO and GSSG were
tested as alternative GS donors.
As shown in Table I, GRx-mediated formation of GAPDH-SSG was most
efficient when GS-radical was the donor of the GS moiety, GS-radical
GSSG > GS-NO. Although the mechanisms by which GRx facilitates GAPDH-SSG formation with any of these GS donors are unknown, the comparison among GS donors indicates that a reaction involving radical intermediates is favored and that direct production of GS-radical (by the HRP system) is more efficient than homolytic cleavage of GS-NO either in the absence or presence of glutaredoxin.
Based on current data, Scheme 4 is a
plausible working model for how GRx may facilitate GS-radical-mediated
protein-SSG formation; however, additional studies are necessary to
delineate the mechanism. Under the particular conditions used for the
HRP-generating system, the formation of GS-radical probably is rate
limiting for the overall reaction. Accordingly, glutaredoxin would
react very quickly with the GS-radicals to form the GRx-SSG
disulfide anion radical intermediate (step 1). This intermediate then
could abstract a hydrogen atom from the target protein-SH group (step
2) to form protein-S· thiyl radical and GSH (step 2). Completion
of the reaction would involve a radical recombination reaction (step
3), forming the protein-SSG and recycling the glutaredoxin enzyme. The
net reaction would involve two GS-radicals for each protein-SSG product
formed.

View larger version (27K):
[in this window]
[in a new window]
|
Scheme 4.
Hypothetical mechanism of GS-radical
transfer by glutaredoxin. The main scheme (in
box) displays a working hypothesis for
glutaredoxin-facilitated S-glutathionylation of proteins by
GS-radicals. An alternative less likely scheme is shown in
brackets (see "Discussion").
|
|
Alternatively, the target protein-S
thiolate might react
directly with the GRx-radical intermediate to form
protein- SSG
(step 2') and then transfer the radical to
molecular oxygen to give protein-SSG and superoxide (step 3'). This
alternative sequence, however, would not be favorable unless the target
protein had properties analogous to glutaredoxin (low
pKa for the SH moiety and stabilization site for the
glutathionyl moiety).
Using GAPDH as a model, the Grx-dependent rate of formation
of protein-SSG was enhanced under anaerobic conditions (Fig. 5, upper shaded portion of right-hand
bars). This effect is consistent with a higher steady-state
amount of the GRx-SSG
intermediate available for reaction
with the target protein-thiol moiety when the competing reaction with
O2 is eliminated.
Net accumulation of protein-SSG would also be affected by GRx catalysis
of deglutathionylation. To confirm this interpretation with the GAPDH
model, reaction mixtures containing glutathionylated GAPDH were
concentrated and utilized as substrates in a typical deglutathionylation assay for GRx in the absence of the GS-radical generating system. These samples were deglutathionylated (release of radiolabel from [35S]GAPDH-SSG) in a
time-dependent manner at rates proportional to the GRx
concentration (data not shown). Therefore, as the concentration of
protein-SSG accumulated in the complete reaction mixture, its rate of
deglutathionylation would increase until a steady-state was reached. In
this regard, we also investigated the relative reactivities of actin
and PTP1B as targets of S-glutathionylation by the
glutaredoxin-mediated GS-radical system (Fig. 6), because both of these
proteins have recently been reported to have their intracellular
activities modulated by S-glutathionylation (9, 11). Since
considerable quantities of GAPDH-SSG, actin-SSG, and PTP-1B-SSG
accumulated, this suggests that the GRx-mediated S-glutathionylation of these proteins under the limited
GS-radical-generating conditions of our experiments was efficient
enough to overcome the competing deglutathionylation reaction in each
case. Thus, the HRP-GSH-H2O2 system as adjusted
for the current experiments may appropriately simulate intracellular
conditions where a continuous production of a low level of GS-radicals
may occur, e.g. when a redox signaling pathway is activated.
Under equivalent conditions, PTP1B was glutathionylated by GRx most
extensively, i.e. PTP1B-SSG > actin-SSG > GAPDH-SSG. Although the reactivity of PTP1B and GAPDH with the GRx-SSG
anion radical intermediate could be ascribed to their low
pKa thiolates, the same interpretation cannot be
applied to actin whose most accessible cysteine residue has a normal
pKa (11). As noted above, the relative accumulation
of each of the S-glutathionylated proteins reflects their
relative efficiencies as substrates for the two different reactions
catalyzed by glutaredoxin, GS-radical-dependent S-glutathionylation versus
GSH-dependent deglutathionylation.
It remains an open question how protein S-glutathionylation
actually takes place intracellularly and whether it is in fact enzyme-mediated; however, the possibility that glutaredoxin (or other
enzyme(s) with comparable properties) catalyzes the formation of
specific protein-SSG adducts via transfer of the glutathione-thiyl radical is supported by the current studies that provide proof of
principle. The additional catalytic properties of glutaredoxin revealed
by the current studies characterize it as a versatile enzyme important
for a variety of cell functions, reinforcing the concept that
glutaredoxin plays a vital role in sulfhydryl homeostasis and redox
signal transduction.