(Received for publication, May 8, 1997, and in revised form, June 11, 1997)
From the Department of Biochemistry, Medical College of Wisconsin, Milwaukee, Wisconsin 53226, the § Biophysics Research Institute, Medical College of Wisconsin, Milwaukee, Wisconsin 53226, and the ¶ Institute for Enzyme Research, Graduate School and Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin, Madison, Wisconsin 53705
Cellular studies have indicated that some Fe-S
proteins, and the aconitases in particular, are targets for nitric
oxide. Specifically, NO has been implicated in the intracellular
process of the conversion of active cytosolic aconitase containing a
[4Fe-4S] cluster, to its apo-form which functions as an
iron-regulatory protein. We have undertaken the in vitro
study of the reaction of NO with purified forms of both mitochondrial
and cytosolic aconitases by following enzyme activity and by observing
the formation of EPR signals not shown by the original reactants.
Inactivation by either NO solutions or NO-producing NONOates under
anaerobic conditions is seen for both enzyme isoforms. This
inactivation, which occurs in the presence or absence of substrate, is
accompanied by the appearance of the g = 2.02 signals
of the [3Fe-4S] clusters and the g 2.04 signal
of a protein-bound dinitrosyl-iron-dithiol complex in the
d7 state. In addition, in the reaction of cytosolic
aconitase, the transient formation of a thiyl radical,
g
= 2.11 and g
= 2.03, is observed. Disassembly of the [3Fe-4S] clusters of the inactive forms of the enzymes upon the anaerobic addition of NO is also
accompanied by the formation of the g
2.04 species
and in the case of mitochondrial aconitase, a transient signal at g
2.032 appeared. This signal is tentatively
assigned to the d9 form of an iron-nitrosyl-histidyl
complex of the mitochondrial protein. Inactivation of the [4Fe-4S]
forms of both aconitases by either superoxide anion or peroxynitrite
produces the g = 2.02 [3Fe-4S] proteins.
The aconitases (EC 4.2.1.3) are a family of dehyratases that
catalyze the reversible isomerization of citrate and isocitrate via
cis-aconitate (1). These enzymes, of which bovine
mitochondrial aconitase
(m-acon),1 is the most
extensively studied, contain unique [4Fe-4S] clusters in that one of
the irons, Fea, is not ligated to a protein residue but
rather to a hydroxide from solvent (2). This is the same iron to which
substrate binds during turnover. Inactivation of these enzymes occurs
when Fea is lost by oxidation of the Fe-S cluster with the
formation of a cubane [3Fe-4S] cluster that is detectable by electron
paramagnetic resonance spectroscopy (EPR) at very low temperatures at
g 2.02. To date there is no EPR evidence that this
reaction occurs in vivo.
The recent discovery that the apo-form of mammalian cytosolic aconitase
(c-acon), is identical to an RNA-binding protein, iron-regulatory
protein (IRP1), has led to increased interest in the study of this
isoform of the enzyme (3-5). There is a second IRP, IRP2, with
similarities both in structure and function to IRP1 that will not be
dealt with in this study (6, 7). These iron-regulatory proteins bind to
specific stem-loop structures called iron-responsive elements which
occur in the untranslated regions of the mRNA of a number of
proteins involved in iron and energy metabolism (8). When bound to
iron-responsive elements located in the 5-untranslated regions of the
mRNA as occurs with the iron storage protein, ferritin, translation
is blocked (9). In contrast to this, binding to the iron-responsive
elements in the 3
-untranslated regions of the mRNA of the
transferrin receptor stabilizes the message, allowing translation to
occur (10). For example, in iron-deficient cells the binding of IRP1
(apo-c-acon) to these iron-responsive elements would increase the
production of transferrin receptors, which are key for increasing
intracellular iron concentrations, while decreasing the biosynthesis of
ferritin. Of particular interest in this process of cellular iron
regulation is the mechanism of the interconversion between c-acon,
containing a [4Fe-4S] cluster, and the cluster-free IRP1, since it
has been demonstrated that the de novo biosynthesis of
protein is not involved (11). Furthermore, in vitro
experiments have shown that the Fe-S cluster of c-acon is quite stable,
particularly in the presence of substrate (3) which can be assumed to
be present in all cells. Therefore, any mechanism proposed for the
interconversion must take these facts into consideration.
Two lines of evidence have led investigators to examine whether or not
NO plays a role in the cellular process where c-acon is converted to
IRP. First, it had been known from the work of Hibbs and co-workers
(12, 13) that, when tumor target cells are cocultivated with activated
macrophages, there is a loss of iron from the target cells that is
associated with the inhibition of mitochondrial respiration and DNA
replication as well as with the inactivation of m-acon. Later it was
shown that this process occurs through an
arginine-dependent pathway (14, 15). This pathway is now
widely recognized to result from the production of NO formed by the
action of the inducible isoform of the enzyme, nitric oxide synthase
(16). The second line of evidence implicating NO came from the results
of studies performed prior to the discovery and awareness of the
importance of NO as a biologically active molecule with all its
manifold functions. NO had been used in a variety of research
endeavors, particularly in the area of transition metal chemistry, in
which the addition of NO, with its unpaired electron, often resulted in
products that were amenable to study by EPR. Early on, Vanin and
Commoner (17-19) and their co-workers had described the so-called
"g 2.04" signal observed for complexes generated synthetically by reaction of NO with iron and cysteine or
found in cells treated with exogenous NO. This signal is attributed to
a dinitrosyl-iron-dithiol complex with thiol groups provided by either
low molecular weight compounds or by protein. In addition, it had been
demonstrated by others that this g
2.04 signal can be observed when NO is added to some Fe-S proteins (20, 21). In more
recent cellular experiments, where cells have been induced to produce
NO, it has been shown that the activities of some Fe- or
Fe-S-containing proteins are often decreased and moreover that these
cells, when examined by EPR, display the typical g
2.04 signal (22, 23). Thus, in 1993 there were two reports
demonstrating that the induction of nitric oxide synthase in
macrophages by interferon-
and lipopolysaccharide to produce NO was
accompanied by a loss of aconitase enzyme activity with a concomitant
rise in the level of IRP1 binding to RNA (24, 25). These initial findings, implicating NO as an active agent in the conversion of c-acon
to IRP1, have been further supported by a number of other cellular
investigations (26-29).
We report here results of our in vitro studies on the
reaction of NO with purified (95%) bovine m- and c-acon in both the 3Fe and 4Fe forms by following enzyme activity and by observing EPR
detectable species formed. Contrary to the results reported in previous
publications which stated that activity is not lost when NO is added to
aconitase (30, 31), we demonstrate both loss of activity and evidence
by EPR regarding the disassembly of the Fe-S cluster. As there is
experimental evidence supporting the intracellular formation of
peroxynitrite from NO and superoxide anion (32), we have also
investigated by EPR the effect of peroxynitrite on the active forms of
both enzymes.
Bovine heart mitochondrial
aconitase and bovine liver cytosolic aconitase were purified according
to published procedures (3, 33). Protein purity (95%) was
established by SDS-polyacrylamide gel electrophoresis and the cluster
content was determined from iron, sulfide, and protein analyses by
described methods (34-36). Active enzyme was prepared immediately
before use by anaerobically incubating protein, iron, dithiothreitol,
and dithionite (33). To remove excess activating reagents, the protein
solution was rapidly desalted anaerobically on G-50 Sephadex columns
equilibrated with the appropriate buffer (0.1 M HEPES, pH
7.5, or 0.1 M MES, pH 6.6) (37). Enzyme activity was
determined at 25 °C following the isocitrate to
cis-aconitate reaction (33). Nitric oxide gas was obtained
from Aldrich Chemical Co. and was passed through 10% KOH prior to use.
Saturated NO solutions were prepared by bubbling NO gas through water
or buffer solutions made anaerobic either by equilibration in the
anaerobic chamber or by repeated evacuation followed by the addition of
either argon or nitrogen gas. The NO concentration of these solutions
(~1.8 mM) was determined using an NO selective electrode
(World Precision Instruments Inc., Sarasota FL). NONOates were obtained
from Cayman Chemical Co., Ann Arbor, MI. SpermineNONOate (sperNO)
solutions were prepared anaerobically in 0.1 M TAPS buffer,
pH 8.9, and stored at 77 K. Concentrations were verified
spectrophotometrically (252 nm,
= 8, 500 M
1 cm
1) before use (38). The
rate of decay and concomitant NO release of sperNO was determined by
following the decrease in absorbance at 252 nm using the specific
conditions of the reaction being studied, i.e. buffer, pH,
and additives. Peroxynitrite was prepared as described previously (39)
and stored at 77 K. Prior to use, the solution was diluted with 10 mM NaOH to the desired concentration which was determined
spectrophotometrically (302 nm,
= 1,670 M
1 cm
1). Care was taken that
upon addition of the alkaline peroxynitrite solution the pH of the
sample was not altered significantly. All other chemicals were reagent
grade and were obtained from Aldrich Chemical Co.
EPR spectra were recorded with an X-band Varian E112 century series spectrometer at room temperature, 77 K, and 10-20 K using standard methods. A frequency counter (EIP model 548) was used to determine the microwave frequency and the field positions of prominent features in the spectra were determined using a Gauss meter (MH-110R Radiopan, NMR Magnetometer) or by comparison of the magnetic field to positions so determined. Quantitation of spin concentrations were made using a 1 mM copper perchlorate standard by comparing the double integral of the spectra to that of the standard run under similar conditions. All EPR samples run at 77 K or 10-20 K were contained in calibrated quartz EPR tubes. Data manipulation of the final EPR spectra was performed using the graphing and data analysis program SUMSPC92, available from the National Biomedical ESR Center, Milwaukee, WI. To separate the signals of a spectrum (A), the following strategy was employed in obtaining the difference spectra: fractions of spectrum B of a single species are subtracted from spectrum A until negative peaks of the signal for B vanish.
Addition of NO to AconitaseAll manipulations were
performed inside a Coy Anaerobic Chamber equipped with a gas analyzer
sensitive to 2 ppm oxygen using anaerobic reagents, unless otherwise
noted. Initial experiments were conducted using NO solutions which were
added directly to the enzyme samples in 0.1 M
HEPES/K+, pH 7.5. In some samples the order of additions
was reversed. When sperNO solutions were used, the enzymes were in 0.1 M MES, pH 6.6, as preliminary experiments had established
that the rate of NO release at this pH and at room temperature
(t1/2 in the absence of enzyme = 62 min) was
suitable for both EPR and activity measurements. It was also determined
that the rate of NO release from sperNO at pH
8.5 and at
4-6 °C is negligible. For the experiments using sperNO, two
different methods were employed. In method 1, following the addition of
sperNO to the enzyme, EPR samples were taken at different time points,
quenched by rapidly raising the pH to
8.5 upon addition into tubes
which contained the residue of 20 µl of 2.0 M TAPS, pH
8.8, evaporated to dryness. The tubes were placed in an ice bath until
frozen in liquid nitrogen. Similarly, activity measurements were made
on samples that were first diluted into substrate at pH 8.5 and at
4-6 °C. In method 2, after addition of sperNO to the enzyme, the
solution was placed in an EPR flat cell and the reaction followed
either by recording repeated scans or by observing the increase in the
g
2.04 signal. The first time point for these EPR
experiments occurs at ~5 min; the time it takes to fill the flat
cell, remove it from the anaerobic chamber and position it in the
spectrometer. At the end of the reaction, i.e. when there
was no further discernible increase in the EPR signal, the sample was
transferred anaerobically from the flat cell to an EPR tube for
examination and quantitation at low temperatures. Samples were reduced
using aliquots of a 50 mM solution of dithionite in 0.1 M TAPS, pH 8.9.
EPR studies were performed at room
temperature, at 77 K (liquid nitrogen), and at 10-20 K for
identification and detection of species formed from the reaction of NO
and aconitase. Although the [4Fe-4S]2+ cluster found in
native, active aconitases is in the S = 0, EPR silent
state, the loss of Fea by oxidation leads to inactivation of the enzyme with formation of the [3Fe-4S]1+ cluster
which is readily observed by EPR at temperatures below ~30 K, at
g 2.02 (2). Fig. 1
shows the spectra for the two bovine isoforms, c-acon and m-acon, which
are sufficiently different to allow for their identification. In
contrast, the EPR signals for the d7 (g
2.04) and d9 forms of the dinitrosyl-iron-thiol complexes
of the protein can be detected at all temperatures (Fig.
2). The line shape of the g
2.04 signal is invariant at low and high
temperatures as it arises from a protein-bound species, whereas the
signals for the low molecular weight complexes at room temperature
exhibit multiline patterns (17, 40, 41). The isotropic hyperfine
interactions observed for the low molecular weight complexes are
characteristic for rapidly tumbling paramagnetic species.
Addition of NO Solutions to Aconitase
Initial experiments of
the reaction of NO with aconitase were carried out by adding varying
amounts of a buffered saturated solution of NO to active enzyme, either
aerobically or anaerobically. The results showed a rapid inactivation
of enzyme followed by a slower inactivation over time. In these
experiments substrate afforded some protection but did not prevent
inactivation. For example, when m-acon was incubated anaerobically with
an NO solution at an NO:enzyme ratio of 65:1, in the absence of
substrate, 56% activity remained at 5 min and 25% at 60 min. The
corresponding values in the presence of substrate were 96 and 76%.
This trend was seen for both enzymes under aerobic as well as anaerobic
conditions. The main difference between the two isoforms of the enzyme
is that it takes a higher concentration of NO to bring about a
corresponding inactivation of c-acon when compared with m-acon. When
these samples were examined by EPR, the main species detected at 77 K
was the g 2.04 signal, typical for the
dinitrosyl-iron-thiol complex (DNI-acon) (Fig. 2A) while at
10-20 K, varying intensities of the signal for the 3Fe cluster were
also seen (data not given). For c-acon, a third signal to be discussed
below, appeared in some of the samples. Although we consistently
observed inactivation of the enzyme using this method, we were unable
to obtain reproducible, quantitative results. As an example, the extent
of inactivation of enzyme would differ when enzyme was added to the NO
solution as compared with the reverse order of addition. Furthermore,
adding NO solutions in a single dose does not simulate cellular
conditions where NO can be generated continuously. Therefore, in the
remainder of the experiments to be described NONOates were used as the
source of NO.
The results of experiments using method 2 (see
"Experimental Procedures") to follow the reaction of active
[4Fe-4S] m-acon and NO are shown in Fig.
3. From these results it is clear that there is a loss of activity and that this loss occurs prior to the
appearance of the DNI-acon signal. As an example, the
t1/2 for the loss of enzyme activity in the presence
of 10 mM citrate was 20 min, whereas the time for the
half-maximal height of the g 2.04 signal to develop
was 40 min. It should be pointed out that this signal results from iron
in the d7 state in the DNI-acon complex and that the Fe in
the complex may also exist in the d6 or d8
EPR-silent states. The quantitation of these species by EPR is not
possible and to be able to do so will require other more discriminating techniques such as Mössbauer spectroscopy. The EPR spectra at 77 K and 12 K of the samples obtained at the end of the room temperature reaction were the same as at room temperature, i.e. only the
signal for DNI-acon was seen. Quantitation of this signal for a number of similar experiments gave amounts that were ~40% of the original enzyme concentration, although in one experiment it was as high as
60%. Similarly, when excess dithionite was added to these samples in
an attempt to convert all of the DNI-acon to the d9 state
(Fig. 2, lower spectrum) and when quantitation of the EPR signal was performed, the concentrations again were only ~40% of the
original enzyme solution. The presence of 1.0 mM substrate in these reactions (Fig. 3) was ineffective in protecting the enzyme in
that both the rate of inactivation and the rate of formation of
DNI-acon were essentially the same as in the absence of substrate. A
comment should be made about the apparent anomalous results showing
that both inactivation of the enzyme and formation of the complex were
faster in the presence of 10 mM citrate than in its absence
(Fig. 3, A and B). By comparing the rates of NO release from sperNO in buffer, and buffer plus 10 mM
citrate, this effect was shown to be due to the faster rate of decay of sperNO and NO release in the latter. The increase in the rate of decay
closely parallels the increased rate of inactivation when compared with
the substrate-free reaction. The rate of the decomposition of NONOates
is dependent on a number of factors, not just pH and temperature.
Consecutive Sampling following Addition of SpermineNONOate to Aconitase
A drawback in performing the EPR experiments at room
temperature is that during the course of the reaction we obviously
cannot detect those EPR signals observable only at low temperatures. As
described in method 1 ("Experimental Procedures") experiments were
thus performed using the 4Fe enzymes where samples were taken at
different time points for low temperature EPR measurements while
concomitantly performing activity assays. At 77 K the DNI-acon signal
was seen in all the samples while at 10-20 K, in addition to the
g 2.04 signal, signals of varying intensity for the
cubane [3Fe-4S] cluster were also observed for the early time points (data not given). Upon quantitation, neither signal correlated with the
loss of activity and, furthermore, the concentration of the 3Fe signal
was much lower than could be accounted for from activity measurements.
It appears that although varying amounts of the cubane 3Fe protein are
formed during the reaction of both m- and c-acon, that it is not an
obligatory intermediate. In support of this observation, some of the
samples taken during the course of the reaction with residual activity,
give signals only for DNI-acon. Since other samples contain both
signals, it can be concluded that the reason for the absence of the 3Fe
signal is not due to the rapid disassembly by NO of cubane 3Fe clusters once formed. This conclusion will be supported by results obtained from
experiments of NO and 3Fe m-acon to be discussed below.
As stated above, when a NO solution was added to active 4Fe
c-acon, a third signal was observed in some samples at both 77 and
10-20 K (Fig. 4A and
uppermost spectrum). Since the g 2.02 signal of the 3Fe cluster is not seen at 77 K it was possible to obtain
g values and the line shape of the third signal by
subtraction of the g
2.04 signal from spectra
obtained at this temperature. The resulting spectrum with
g
= 2.11 and g
= 2.03 (Fig. 4, A and B) has been assigned to the
transient formation of a thiyl radical by comparison to similar signals
described by Nelson and Symons (42). In one experiment the intensity of
the signal was as high as 65% of the signal intensity for DNI-acon.
Although we have not conducted a thorough search for the exact
conditions as to when this signal appears, as demonstrated in Fig. 4,
it can appear early in the reaction.
When a reaction of 4Fe c-acon and sperNO (Fig.
5) was run under the same conditions as
described above for m-acon (Fig. 3) using method 2, a similar pattern
for the formation of the DNI-acon signal was observed. Also, as for
m-acon, the presence of substrate did not substantially alter the rate
of reaction. The primary difference between m-acon and c-acon is that
under these conditions, the extent and pattern of inactivation are
different. For c-acon a plot of activity versus time
indicated a biphasic reaction (data not given) and at the end of the
reaction (Fig. 5) 34 and 41% of the activity remained in the absence
and presence of substrate, respectively, whereas for m-acon under the
same conditions, none remained.
Formation of a Dinitrosyl-Iron-Histidyl Complex with Mitochondrial Aconitase and NO
When sperNO was added to 3Fe m-acon at room
temperature and the spectrometer set to follow increases in the
intensity of the first peak appearing in the spectrum, traces as shown
in Fig. 6A were obtained. It
had been assumed that the spectrometer had been set to record increases
in the peak of the g 2.04 species since a rough
calculation of the g value from the spectrometer settings
indicated this. Furthermore, in trial experiments at room temperature,
only the signal for DNI-acon had been observed. However, it became
apparent from the pattern of the traces obtained from following the
reactions by EPR, that what was being observed was either a variation
in the formation of DNI-acon or the transient formation of another EPR
active species. When the experiment was repeated and the peak height of
the signals was determined from repeated 4-min scans, the plot shown in
Fig. 6B was obtained. It became evident that what was
occurring was the formation and disappearance of a new signal which
overlapped the g
2.04 signal of DNI-acon. The
original traces (Fig. 6A) had been obtained at a position
corresponding to the peak of the new signal which is located on the
right shoulder of the signal for DNI-acon. The new signal reaches a
maximum at ~15 min and is not detectable at 30 min. Scans A and B in
Fig. 7 (23 °C) are for the 15- and 75-min time points, respectively, for the experiment shown in Fig.
6B. The presence of a second signal is obvious (Fig.
7A at 23 °C) and by subtraction for the contribution to
the spectrum of the g
2.04 DNI-acon species (Fig.
7B at 23 °C) two new signals (Fig. 7,
A-B, at 23 °C) are observed: one that is
similar in line shape to that of d9 DNIC-acon (Fig.
2B) but with very different g values, 2.032 and 2.004 versus 2.006 and 1.970, and a second minor species
with g values of 2.05 and 2.01.
To identify these species and also to determine if other EPR active
substances that are seen only at low temperatures were present, samples
were prepared under conditions as in the room temperature experiment
and then rapidly frozen in liquid nitrogen at 10, 15, and 75 min. The
EPR spectrum at 77 K for the 15-min time point (Fig. 7A at
77 K) again shows this second signal in addition to that for
g 2.04 DNI-acon. Subtraction of the latter (Fig.
7B at 77 K) yields a spectrum (Fig. 7,
A-B, at 77 K) very similar to that found at room
temperature (Fig. 7, A-B, at 23 °C) indicating
that no other species were being detected. Finally, these same samples
were also examined by EPR at 16.6 K. In Fig. 7A at 16.6 K,
the signal for 3Fe m-acon at g
2.02 is now seen superimposed upon the other two signals. Subtraction of the spectrum for 3Fe m-acon (Fig. 7B at 16.6 K) from the 15-min time
point (Fig. 7A at 16.6 K) gives a spectrum (Fig.
7C at 16.6 K) not unlike the 77 K spectrum (Fig.
7A at 77 K). A scan of the sample at 75 min which at higher
temperatures showed only the signal for DNI-acon now has a broad signal
on the high-field side (Fig. 7D at 16.6 K). We assign this
to excess unreacted NO in the sample (43). Finally, if we subtract this
spectrum (Fig. 7D at 16.6 K)) from the previous difference
spectrum (Fig. 7C at 16.6 K), we again obtain a spectrum
(Fig. 7E at 16.6 K) for the new species with g
values as those found at room temperature and 77 K. Analysis of the
results at the three different temperatures all lead to the conclusion
that two new EPR-detectable species are formed in the reaction of 3Fe
m-acon and NO. The appearance of these signals at room temperature led
us to believe that they might be due to the formation of a different
iron-nitrosyl complex. We were aware from earlier work by the Commoner
group (19), that they had synthesized iron-nitrosyl complexes with
amino acid residues other than thiols. More recently, Chasteen and
co-workers (44) in their studies of the reaction of ferritin and NO had observed a signal at gx
= 2.055, gy
= 2.033, and gz
= 2.015 that they assigned to a histidyl-iron-nitrosyl complex (44). Boese et al. (41) have seen a similar signal with iron,
bovine serum albumin, and NO, and have also attributed it to a histidyl complex.
We thus prepared the iron-nitrosyl-imidazole complex by incubating
anaerobically Fe2+ and imidazole with diethylamineNONOate.
Dithionite was added to one of the samples. The resulting spectra are
shown in Fig. 8. Subtraction of the
spectrum of the reduced species (Fig. 8B) from the original
spectrum (Fig. 8A) which is a mixture of two signals, gives
a spectrum (Fig. 8, A-B) that we assign to the d7 state of the imidazole complex. This spectrum resembles
one of the signals seen in all the final difference spectra of Fig. 7, i.e. the signal with g values of 2.050 and 2.17. The spectrum that most closely resembles the new signals observed in
the reaction of [3Fe-4S] m-acon and NO with g values of
2.032 and 2.004 is that of the d9 state of the imidazole
complex obtained upon reduction with dithionite (Fig. 8B).
We thus tentatively assign these transient signals to the
d7 and d9 forms of a
histidyl-iron-nitrosyl-acon complex. The addition of sperNO to 3Fe
c-acon gave no evidence for the formation of any species other than
g 2.04 DNI-acon (data not given).
Optical Spectra of Reactants and Products
A number of
experiments were performed to record changes in the optical spectra
that occur during the reaction of sperNO and the various forms of
aconitase. In Fig. 9 are shown the
optical spectra for 3Fe and 4Fe m-acon and the product of the reaction of sperNO and 4Fe m-acon in which no activity remained. The inactivated enzyme sample, when examined by EPR, gave evidence for only the g 2.04 species. The optical spectrum for the
product of the reaction of 3Fe m-acon and sperNO had features identical
to that recorded for the 4Fe m-acon reaction. When the reactions of
either 3Fe or 4Fe m-acon were followed by repeating 1-min scans from 300 to 700 nm, no isosbestic points were observed. Maximal changes in
the spectra between reactant and products occur at 365 nm (increase) for 4Fe m-acon and 470 nm (decrease) for 3Fe m-acon.
Reaction of Peroxynitrite and Aconitase
Addition of
peroxynitrite to the 4Fe forms of both c-acon and m-acon resulted in
inactivation of the enzymes and the appearance of the 3Fe
g 2.02 signal in the EPR spectra obtained at 10-20 K (Fig. 1). No other signals were detected. Interestingly, in contrast
to the reaction of NO and aconitase, substrate afforded protection to
the protein. The aerobic addition of peroxynitrite to both c-acon and
m-acon, at a ratio of 133:1, resulted in a 99% loss of activity for
m-acon and 61% for c-acon in less than 3 min in the absence of
substrate. In the presence of substrate both enzymes were completely
protected with no loss of activity. As observed with the reaction of NO
with the enzymes, c-acon requires higher concentrations of
peroxynitrite to attain a similar level of inactivation as seen with
m-acon. For example, the anaerobic addition of peroxynitrite to m-acon
in a ratio of 50:1 resulted in a 32% decrease in activity at 1.5 min
in the absence of substrate. Under identical conditions, for a ratio of
peroxynitrite to enzyme of 100:1, c-acon lost 48% of its activity. It
thus appears that peroxynitrite functions as other oxidants,
inactivating the enzyme by oxidation of the [4Fe-4S]2+
cluster with loss of Fea and formation of the
[3Fe-4S]1+ cluster. However, whereas the oxidative
reaction of ferricyanide and m-acon is stoichiometric (45), this is not
so with peroxynitrite. Apparently the rate of decomposition of
peroxynitrite to form nitrate, t{ifrax,1,2] ~ 1 s
(46), is faster than the rate of oxidation of the cluster. Finally, a
limited number of experiments on the reaction of O
2 and m-acon
were performed. Inactivation of the enzyme occurred in the absence or
presence of either superoxide dismutase or catalase and as reported
previously by Flint et al. (47), is accompanied by the
formation of the cubane 3Fe cluster.
It is apparent from the preceding sections that the number of conditions which can affect the course of the reactions that we studied is so large, that we had to limit ourselves and chose the conditions described. Of particular importance in this study is protein purity, and we thus recommend caution in comparing data obtained here to previous reports (30, 31).
Dinitrosyl-Iron-Thiol-AconitaseA common feature of the
reactions of NO and the various forms of m-acon and c-acon, is the
appearance of the g 2.04 EPR signal assigned to an
iron-nitrosyl-thiol-acon complex. Similar signals have been observed
for synthetic iron-nitrosyl-cysteine complexes as well as for cellular
preparations which have been exposed to added NO or NO produced
endogenously (17-19, 22, 24). In a recent publication, Vanin (48) has
clearly shown that the g
2.04 signal arises from
iron in the d7 state, in contradiction to a report which
attributed these signals to iron with a d9 configuration
(49). Our results agree with those of Vanin, as reduction with
dithionite of aconitase samples exhibiting the g
2.04 signal yielded EPR spectra (Fig. 2B) similar to those obtained by him for the d9 state of the cysteine complex.
Many of the EPR spectra of cellular samples exhibiting the
g
2.04 signal also contain signals suggestive of the d9 species (15, 22, 23).
Despite previous claims to the contrary (30, 31), we have
clearly shown that NO is capable of inactivating both c-acon and m-acon
when added anaerobically or aerobically to the active [4Fe-4S] form
of either enzyme. The mechanism of inactivation does not appear to
involve the obligatory formation of the cubane 3Fe cluster as is
observed when a variety of different oxidants are added to active
enzyme. For instance, the active 4Fe form of either enzyme, determined
by activity, can exist simultaneously with both the inactive 3Fe form
and the DNI-acon complexes as shown by EPR. In addition, there was no
correlation between loss of activity and appearance of the 3Fe
g 2.02 signal. The latter was always much less than
could be accounted for by loss of activity. More importantly, when NO
is added to the 3Fe form of m-acon, new transient signals appear during
the early phases of the reaction and prior to the formation of DNI-acon
(Fig. 7) that are not seen in the reaction of 4Fe m-acon. Also, as
shown in Fig. 2, the pattern of the appearance of the DNI-acon signal
for 4Fe m-acon is quite different from that seen under similar
conditions for 3Fe m-acon (Fig. 6). We have tentatively assigned the
signals for these transient species seen in the 3Fe m-acon reaction to
the d7 and d9 forms of an
iron-nitrosyl-histidyl-acon complex, with the d9 form
predominating. The active site of m-acon, as determined by x-ray
crystallography, contains three histidine residues in close proximity
to the Fe-S cluster (50). As iron is being released from the cluster by
the action of NO, it could ligate to these histidines and react with NO
with formation of the transient complex. As further disassembly of the
cluster occurs, it is possible that two of the three cysteines that are
ligands to the Fe-S cluster, form an iron-thiol-nitrosyl-acon complex.
The question remains as to why on reaction with NO, 3Fe c-acon does not
form a similar histidyl complex, since both aconitases contain the same
active site amino acid residues (51). It should be noted that it is not
certain whether this histidyl complex contains two histidyl ligands or
is a mixed complex with ligands from both histidine and cysteine
(41).
Another notable difference between the two aconitases is the appearance
of a thiyl radical in the reaction of NO and 4Fe c-acon. Although we
have not made a thorough study of its formation and do not see it in
all the samples examined, we can say it forms early in the reaction. We
have observed it in at least four samples where 75% or more of the
original enzyme activity remains. As shown in Fig. 4, uppermost
spectrum, at temperatures below 20 K the EPR spectrum contains a
mixture of three signals, including signals for 3Fe c-acon
(g 2.02), the thiyl radical and g
2.04 DNI-acon. At 77 K only the last two signals are seen. However, as with 4Fe m-acon, while the reaction progresses, only the
g
2.04 signal is observed. This was true for the
last time points of the reaction shown in Fig. 5 even though 30-40%
of the original 4Fe concentration remained, as was indicated from
activity assays. The pattern of inactivation of c-acon in the presence
of sperNO did not follow a simple decay as observed for m-acon (Fig.
3B) but exhibited a biphasic reaction. Also, as when NO
solutions were used, it takes a larger molar ratio of sperNO to enzyme
to inactivate c-acon in comparison to m-acon.
In Fig. 9 are shown the optical spectra of
3Fe and 4Fe m-acon and the spectrum of the product of the reaction of
NO with either form. The EPR spectra for the NO samples revealed only
the g 2.04 signal and concentration of the
solution, using an ultrafiltration unit, demonstrated that the
chromophore was bound to protein. Exposure to air, or evacuating for 30 min at <0.1 atmosphere, resulted in no change in the optical spectrum.
We would like to point out the difference between the spectrum of the
purified 4Fe active m-acon shown here (Fig. 9) and that reported
previously in one of the earlier studies claiming that NO does not
inactivate the enzyme. These authors attributed the bands at 535 and
575 nm seen in the optical spectrum of the aconitase preparation used by them, to the active form of the enzyme (31). These bands are absent
in the spectrum of 4Fe m-acon (Fig. 9). Having analyzed aconitase
preparations similar to those used by the authors, i.e. obtained from Sigma, we assign these bands to contaminating heme compounds. This was confirmed by EPR where signals for heme were observed as well as by the presence of a very strong band at ~408 nm
in the optical spectrum that shifted to ~424 nm on reduction with
dithionite, behavior typical of the Soret band of heme (data not
given).
Inactivation of
both aconitases by peroxynitrite is accompanied by the appearance of
the g 2.02 EPR signal for the 3Fe clusters (Fig.
1). The limited extent of inactivation by peroxynitrite was surprising.
It has been shown, for example, that 4Fe m-acon can be inactivated by
titration with ferricyanide producing the 3Fe cluster stoichiometric to
the amount of ferricyanide added (45). With peroxynitrite, the molar
ratio of added solution to enzyme inactivated ranged from 22:1 for
m-acon to 85:1 for c-acon. Also, substrate was very effective in
protecting the enzyme. How to relate these results to what may occur
in vivo is difficult for many of the reasons stated above.
It can at least be stated that the source of the g
2.04 signal observed in many of the cellular studies involving NO, is
not due to a reaction of aconitase with peroxynitrite.
The finding that both aconitases are susceptible to inactivation by NO in the presence or absence of substrate and furthermore, that the rates of these reactions are very similar, supports the significance of the results obtained from cellular studies. Inactivation of the mitochondrial enzyme can have profound effects on cellular energy production. The results of this study cannot answer questions concerning the role of NO in iron metabolism. At this point we do not know whether the species arising on reaction of NO with c-acon, which we have detected by EPR, are active as IRP; our quantitative evaluations leave room for the formation of EPR undetectable species, such as intact apoprotein that are active as IRP (4). Further work will be required to establish the quantitative relationships between the products of NO and aconitase, whether or not they are EPR-detectable. The aim of the present work was to establish beyond doubt that NO does react with the Fe-S cluster of aconitase.
We acknowledge the use of the facilities of the National Biomedical ESR Center at the Medical College of Wisconsin where all of the EPR spectra were recorded (supported by National Institutes of Health Grant RR01008).