| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
(Received for publication, April 6, 1995; and in revised form, June 8, 1995) From the
Aerobic organisms synthesize superoxide dismutases in order to
escape injury from endogenous superoxide. An earlier study of Escherichia coli indicated that intracellular superoxide is
formed primarily by autoxidation of components of the respiratory
chain. In order to identify those components, inverted respiratory
vesicles were incubated with five respiratory substrates. In most
cases, essentially all of the superoxide was formed through
autoxidation of fumarate reductase, despite the paucity of this
anaerobic terminal oxidase in the aerobic cells from which the vesicles
were prepared. In contrast, most dehydrogenases, the respiratory
quinones, and the cytochrome oxidases did not produce any detectable
superoxide. The propensity of fumarate reductase to generate
superoxide could conceivably deluge cells with superoxide when
anaerobic cells, which contain abundant fumarate reductase, enter an
aerobic habitat. In fact, deletion or overexpression of the frd structural genes improved and retarded, respectively, the
outgrowth of superoxide dismutase-attenuated cells when they were
abruptly aerated, suggesting that fumarate reductase is a major source
of superoxide in vivo. Steric inhibitors that bind adjacent to
the flavin completely blocked superoxide production, indicating that
the flavin, rather than an iron-sulfur cluster, is the direct electron
donor to oxygen. Since the turnover numbers for superoxide formation by
other flavoenzymes are orders of magnitude lower than that of fumarate
reductase (1600 min
Synthesis of superoxide dismutase (SOD(
This enzyme maintains intracellular superoxide levels around
10 Recent studies in E. coli have
identified some important physiological targets of superoxide. Among
these are aconitase(6) , dihydroxyacid dehydratase(7) ,
fumarases A and B(8) , and 6-phosphogluconate
dehydratase(9) , all of which belong to a class of dehydratases
whose substrate-liganding iron-sulfur clusters disintegrate upon
oxidation by superoxide. Second-order rate constants for these damaging
reactions can exceed 10 Because superoxide cannot
traverse membranes(11) , the superoxide that damages those
enzymes in SOD-deficient mutants is evidently formed inside the
boundary of the cytoplasmic membrane. The goal of this work is to
identify the predominant mechanisms of its formation. Superoxide is
produced when a single electron is transferred from a donor molecule to
molecular oxygen. The E` The precise sites of electron leakage to oxygen
were not identified in those experiments. Various studies with
mitochondria have implicated autoxidation of ubiquinone and/or
respiratory dehydrogenases(13) . The present work identifies
the respiratory enzyme fumarate reductase of E. coli as a
particularly active source of superoxide.
Uracil (0.2 mM)
was supplemented to anaerobic cultures of quinone-deficient strains in
order to by-pass the requirement for function of the respiration-linked
dihydroorotate dehydrogenase. When vesicles were prepared from aerobic
Ubi
The cytochrome c assay was used to measure superoxide
production by xanthine oxidase both in the presence and absence of
vesicles. There was no difference, even at very high vesicle
concentrations (data not shown). This was also true when respiratory
substrates were included to stimulate electron flow through the
respiratory chain. Thus no evidence was observed to support the
possibility (18) that either reduced or oxidized respiratory
chain components are effective scavengers of superoxide.
To verify the absence of quinones from
vesicles prepared from Ubi
The number of molecules of superoxide produced per
unit of electron flux through the respiratory chain depended upon the
identity of the respiratory substrate. In general, approximately 0.5%
of the respiratory oxygen consumption was due to superoxide formation. With all substrates the rate of superoxide production in vitro was linear with time (Fig. 1A) and increased in
proportion to vesicle concentration (not shown). However, high
substrate concentrations actually inhibited superoxide production when
respiration was driven by succinate (Fig. 1B). This
effect is not due to interference with the superoxide detection system,
since succinate did not affect the ability of cytochrome c to
be reduced by superoxide that was generated by xanthine oxidase. This
phenomenon is explored in further detail below.
Figure 1:
A, time course of superoxide
production by inverted vesicles. Vesicles were prepared from an aerobic
culture of AB1157 and incubated with 0.3 mM succinate. B-D, superoxide production versus succinate
concentration. Vesicles prepared from cultures grown in aerobic
casamino acids medium were incubated at 23 °C for 7.5 min with the
indicated concentrations of succinate. The superoxide yield is
presented per vesicles from 10
Figure 2:
Modular diagram of the respiratory chain. Vertical axis denotes relative free energies of electron
carriers; exact relationships depend upon the relative abundance and
redox status of the carriers. For simplicity a single cytochrome
oxidase and a single
Fumarate reductase (Frd) is a second
succinate-reducible enzyme associated with the respiratory chain. Frd
acts as a terminal oxidase during anaerobic respiration (Fig. 2B), transferring electrons from the
low-potential carrier menaquinone to fumarate, generating succinate as
a final product. However, in vitro Frd can be directly reduced
by succinate in a reversal of its physiological reaction. In fact,
although in their physiological roles Frd and Sdh catalyze the
interconversion of succinate and fumarate in opposite directions, they
are structurally similar isozymes, and the plasmid-driven aerobic
synthesis of fumarate reductase can restore the ability to oxidize
succinate in vivo to mutants that lack succinate
dehydrogenase(26) . Because fumarate-directed respiration is
energetically less profitable than oxygen-directed respiration,
fumarate reductase is generally repressed during
aerobiosis(27) . Thus it was unexpected that Frd might be
present and responsible for superoxide production in these vesicles,
which were prepared from cells that had been grown with vigorous
aeration. Yet vesicles prepared from frd deletion mutants
failed to evolve superoxide when incubated with succinate (Fig. 1D), demonstrating that Frd was indeed the site
of succinate-dependent superoxide production. The addition of cyanide
to the Frd-deficient membranes restored some superoxide formation, and
even this amount was absent from vesicles lacking both Frd and Sdh (not
shown). Thus Sdh transfers electrons to molecular oxygen, producing
superoxide, only when the usual electron flow down the respiratory
chain is blocked; in unblocked vesicles oxidized ubiquinone outcompetes
oxygen for electrons on the reduced enzyme. It therefore seems that
Frd, but not Sdh, may be an important source of superoxide in
vivo. The apparent presence of Frd in membranes prepared from
aerobic cultures appeared to contradict reports that its synthesis is
almost fully repressed during aerobiosis(27) . Measures had
been taken to ensure that the exponential cultures remained
air-saturated until they were harvested and that protein synthesis did
not occur under the microaerobic conditions that occur during
subsequent centrifugation (``Materials and Methods''). Frd
content and superoxide formation were therefore quantitated in
membranes derived from cells grown both in casamino acids medium and in
the glucose media used by other workers (Table 3). It is apparent
that casamino acids promote the aerobic synthesis of Frd, although the
basis of this induction is unknown. These levels were still far below
those which were achieved in anaerobic media, when Frd can fulfill its
physiological role of serving as an alternate terminal oxidase. In all
preparations the rate of succinate-dependent superoxide production
paralleled the Frd content of the membranes (Table 3).
Furthermore, 7-fold overexpression of Frd from plasmid pH3 caused an
8-fold increase in superoxide production by vesicles prepared from
JI241.
The epinephrine-oxidation and lucigenin assays of O
production quantitatively confirmed the cytochrome c data (Table 4). These results verify that Frd is a highly active
source of superoxide.
In accord with this idea, vesicles that were
prepared from anaerobic cells, which have abundant Frd, generated far
more respiratory O than did vesicles from aerobic cells, no matter what
the respiratory substrate (Table 5). In contrast, vesicles
prepared from Frd-deficient mutants did not generate detectable amounts
of O with
Electrons
are delivered to Frd by the low-potential carrier menaquinone, and
vesicles derived from mutants lacking menaquinone did not generate any
O at Frd when incubated with Thus, when present, Frd was the source
of virtually all of the respiratory O. The tendency of this enzyme to
produce O is evidently not shared by the other respiratory
dehydrogenases or oxidases.
When wild-type
cells that had been cultured in anaerobic glucose minimal medium were
aerated, they continued to grow without significant delay (Fig. 3A). Evidently the concentration of intracellular
O was insufficient to block amino acid biosynthesis. In this
circumstance the O concentration is suppressed primarily by action of
the iron-containing superoxide dismutase, which is synthesized
pre-emptively during anaerobiosis in preparation for re-exposure to
oxygen. In contrast, the growth of sodB mutants, which lack
that enzyme, lags approximately 70 min after re-aeration, finally
resuming when the oxygen-inducible manganese-containing superoxide
dismutase accumulates in sufficient quantity to reduce endogenous O to
subtoxic levels ( (29) and Fig. 3A). To
determine whether Frd was the primary source of the toxic O, strains
lacking both Fe-SOD and Frd were constructed. Upon aeration the sodB frd mutants lagged more briefly than did those containing
Frd. Conversely, strains carrying a multicopy plasmid encoding Frd
exhibited an increase in growth lag (Fig. 3B).
Figure 3:
Correlation between Frd content and oxygen
toxicity upon reaeration of SOD-deficient strains. A, strains
AB1157 (SOD
These
data indicate that Frd may be a significant source of O in
vivo. However, the residual lag of the frd sodB mutant
indicates that sources of O other than Frd may be significant.
Furthermore, the flux of O from Frd is not sufficient to debilitate an
SOD-proficient cell. Possible reasons for this are explored under
``Discussion.''
The
kinetics of O production by Frd and Sdh were investigated in an effort
to identify which of these redox moieties leaks electrons to oxygen.
Excess succinate inhibited superoxide production in vitro by
both Frd and Sdh, although the two enzymes exhibited somewhat different
profiles (Fig. 1). This effect occurred both in respiring
vesicles and when respiration was blocked by cyanide. To verify that
this activity was a behavior of the isolated enzymes, Frd and Sdh were
independently expressed in mutants devoid of quinones, preventing
interaction between the enzymes and other components of the respiratory
pathway. The enzymes recovered in vesicles from such cells continued to
generate O when incubated at moderate, but not high, concentrations of
succinate (Fig. 4). A similar profile was obtained when
superoxide production was quantitated using an epinephrine-oxidation
assay instead of cytochrome c reduction (Fig. 5). Yet
excess succinate did not interfere with superoxide production or
detection when enzymatic sources other than Sdh or Frd were used,
including xanthine oxidase.
Figure 4:
The succinate dependence of superoxide
production by Frd and Sdh in quinone-deficient vesicles. A,
succinate-dependent superoxide production by fumarate reductase.
Vesicles were prepared from JI243 (Ubi
Figure 5:
Kinetics of epinephrine oxidation by
superoxide produced by fumarate reductase. Vesicles were prepared from
strain JI243 (Ubi
The inhibition by concentrated succinate
was immediately reversed when Frd-containing vesicles were diluted into
lower concentrations of succinate, confirming that Frd was not
irreversibly affected (data not shown). Interestingly, the high
succinate concentrations did not suppress the rate of electron transfer
by either enzyme to the artificial quinone plumbagin, and nor did they
inhibit electron transfer from Sdh to endogenous ubiquinone (as
measured by respiration rate) (Fig. 4). In fact, the
superoxide-production optima were approximately coincident with the
apparent K Two hypotheses might explain
this effect. First, because both Frd and Sdh can accommodate a total of
four electrons on their flavin and iron-sulfur clusters, it seemed
possible that the half-reduced form of the enzymes might generate O,
whereas the fully reduced forms could not. (The (4Fe-4S) cluster is not
considered, since its low potential is not within the reach of the
succinate-fumarate couple.) Such behavior is characteristic of xanthine
oxidase, which exclusively forms H The second
possibility was that the binding of excess succinate to the face of the
reduced enzyme might sterically obstruct interaction between oxygen and
the autoxidizable redox center on the enzyme. If so, then the succinate
analogues malonate and malate might also be expected to inhibit
electron transfer from reduced Frd to oxygen. The avidity with which
these analogues bind to the active site of Frd was demonstrated by
their competitive inhibition of its succinate:plumbagin oxidoreductase
activity, with apparent K In
the experiment presented in Fig. 6, Frd was reduced by the
delivery of electrons from
Figure 6:
Structural analogues of succinate inhibit
O production by fumarate reductase. A, experimental design.
Menaquinone delivers electrons from
Because these inhibitors (and succinate) bind to Frd and Sdh across
the flavin, these results indicate that the flavin must be the site of
electron leakage to oxygen. Excess succinate does not block quinone
reduction by either enzyme, because quinones dock at the distant
[3Fe-4S] cluster.
The following data indicated that the
superoxide was generated by the native forms of both enzymes. 1) The
superoxide-generating activity for both enzymes exhibited <10% loss
during week-long storage on ice and was constant for at least 2 h
during in vitro measurements. This stability argues against
the possibility that the superoxide was produced by damaged forms of
the enzymes that accumulate after cell lysis. 2) Superoxide production
by both enzymes was completely inhibited by oxaloacetate (not shown),
demonstrating that the superoxide is not generated by the artifactually
oxaloacetate-bound form of the enzyme that can be recovered during
vesicle preparation. 3) Superoxide production by both enzymes was
prevented when fumarate was available to reoxidize the enzymes,
indicating that the enzymes that generate the superoxide retain their
fumarate reductase capacity. 4) The availability of ubiquinone also
inhibited superoxide production by either enzyme, presumably by
reoxidizing the enzyme before it could directly transfer electrons to
oxygen. Thus the superoxide-producing enzymes retain their
physiological succinate:ubiquinone oxidoreductase capacity. 5) Other
workers have noted that flavin autoxidation can be catalyzed by loosely
bound iron that contaminates biological buffers. That appears not to be
the case for Frd and Sdh. The turnover numbers for superoxide
production by Frd were essentially unchanged whether measured in Tris,
MOPS, or phosphate buffers. Furthermore, the addition of up to 1 mM of the metal chelators EDTA and diethylenetriamine pentaacetic
acid did not inhibit superoxide formation (data not shown).
Collectively, these results argue strongly that the native enzymes are
responsible for the generation of superoxide.
The rates at which the different
respiratory enzymes produce superoxide vary widely. This work has
unambiguously identified only Frd and NADH dehydrogenase II as
significant sources of superoxide in respiring vesicles. L-Lactate dehydrogenase, D-lactate dehydrogenase, and
It is premature to suggest that the autoxidizability of the E.
coli enzyme extends to fumarate reductases from other organisms.
However, Turrens (32) reported that O was generated by the
NADH-fumarate respiration of Trypanosoma brucei.
In
fact, experiments with an SOD-deficient mutant supported the idea that
Frd generates superoxide when anaerobic cells are re-aerated. However,
it was clear that the bolus of superoxide was not sufficient to
incapacitate SOD-proficient cells. This result was expected, since this
facultative organism could not tolerate oxidative poisoning during
every transit from an anaerobic to an aerobic environment. Survival
upon aeration requires the preemptive synthesis of Fe-SOD when E.
coli is anaerobic. Several additional factors may help minimize
the superoxide flux from Frd. First, the resumption of ubiquinone
synthesis upon aeration enables respiratory electrons to by-pass Frd as
flow to cytochrome oxidase is restored. Furthermore, any Frd that
manages nonetheless to receive electrons can be directly re-oxidized by
ubiquinone before the electrons are transferred to oxygen. Second,
the data demonstrated that fumarate also competes with oxygen for the
electrons on reduced Frd. Superoxide production in vitro was
half-inhibited by 95 µM fumarate and 90% inhibited by
about 750 µM. Is there this much fumarate in the
re-aerated cell? When anaerobic cells re-enter an aerobic habitat, the
influx of oxygen will restore aerobic respiration and, by mass action,
the counterclockwise direction of the TCA cycle. The K
Frd, like Sdh, contains a flavin and three iron-sulfur
clusters (35) . When acting to reduce fumarate, electrons are
passed from reduced menaquinone through the [3Fe-4S] and
[2Fe-2S] clusters to the flavin, which then transfers them to
fumarate. The experiments with substrate analogues indicated that the
flavin, rather than an Fe-S cluster, was the site of electron transfer
from Frd to oxygen. Excess succinate had the same effect in Sdh,
implying that superoxide evolved from the flavin of that enzyme as
well. This localization is consistent with similar observations for
xanthine oxidase(36) . The formation of superoxide by a flavin
oxidoreductase of E. coli(37) probably reflects the
oxidation by molecular oxygen of the reduced flavin product. It may
turn out to be generally true that flavoenzymes comprise most of the
superoxide sources of physiological significance. However, it is not
true that all flavoenzymes are similarly proficient at generating
superoxide. Even excluding oxidases and monooxygenases (38) ,
the superoxide turnover numbers for dehydrogenase-type flavoproteins
vary over 5 orders of magnitude (Table 6). Evidently additional
structural or electronic characteristics are critical. The steric
accessibility of the isoalloxazine ring may be variable; for example,
the exposure of the flavin of xanthine dehydrogenase upon sulfhydryl
oxidation is responsible for conferring upon that enzyme its oxidase
activity. A second factor may be that fumarate reductase differs from
simple flavoproteins, which make little superoxide, in that its
redox-active flavin is electronically linked to metal centers. In this
respect it resembles xanthine oxidase, another enzyme that is
particularly proficient at generating superoxide. The correlation
suggests that the secondary redox centers may kinetically promote the
univalent electron transfer from the flavin to oxygen. Two mechanisms
seem plausible: 1) a flavin-linked metal center might unpair the
electrons on the two-electron-reduced enzyme, conferring semiquinone
character to the flavin and improving the likelihood of productive
orbital overlap with triplet oxygen; or 2) after transfer of the first
electron from a dihydroflavin to oxygen, the second might become stably
sequestered on a metal center, allowing the nascent superoxide to
dissociate from the oxidized flavin. The behavior of xanthine oxidase
is most consistent with the second possibility: whereas the fully
reduced enzyme quantitatively reduces oxygen divalently to hydrogen
peroxide, the two-electron-reduced enzyme transfers a single electron (39, 40) . The residual electron is slow to hop onto
oxygen, and it is presumably that delay that enables the previous
superoxide product to diffuse away from the flavin without being
reduced to peroxide. It will be interesting to see whether fumarate
reductase behaves similarly.
Because Sdh has a structure very
similar to that of Frd, its lesser tendency to generate superoxide
requires additional consideration. One distinction between the two
enzymes may lie in the potentials of their flavins relative to those of
the iron-sulfur clusters. The cluster potentials of Frd (about
-50, -320, and -60 mV for clusters 1 through 3) are
significantly lower than those of Sdh (+10, -175, and
+65 mV)(35) . The flavin potential of Frd, -55
mV(41) , is well above the -219 mV potential of free
flavins and is probably elevated by its covalent linkage to a protein
histidine residue(42) . This high potential relative to those
of clusters 1 and 3 will push the electron density onto the flavin. In
contrast, the higher cluster potentials of Sdh could plausibly tilt the
potentials away from the flavin and push electrons onto the metal
centers. These arrangements are congruent with the physiological
functions of the two enzymes: the electrons of reduced Frd await
transfer from the flavin to fumarate, while those of reduced Sdh await
transfer from the iron-sulfur clusters to ubiquinone. The collateral
effect may be that the high electron density on the flavin accelerates
superoxide formation by Frd, while the shift away from that site slows
superoxide formation by Sdh. Further work will be necessary to test
these ideas.
Volume 270,
Number 34,
Issue of August 25, pp. 19767-19777, 1995
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
), additional steric or
electronic factors must accelerate its autoxidation.
); EC
1.15.1.1; ) is ubiquitous among aerobic and aerotolerant
organisms.
M(1) . Mutant strains of Escherichia coli, yeast, and Drosophila that lack SOD
accumulate superoxide to much higher levels and exhibit severe growth
deficits(2, 3, 4) . In humans, dominant
mutations in the structural gene encoding cytosolic SOD are associated
with the development of familial amyotrophic lateral
sclerosis(5) .
M s
(10) . The inactivation of these
enzymes is responsible for at least some of the phenotypic deficits of
SOD-deficient mutants of E. coli.
for univalent oxygen
reduction is low (-0.16 V), requiring that a donor be a strong
univalent reductant in order to push the reaction forward. Most
biological electron donors do not meet this standard. For example, the
instability of the NAD
species (12) prohibits NADH from
spontaneously transferring electrons to oxygen. On this basis, it is
reasonable to consider flavin and quinone moieties as potential
superoxide sources, since both form stable semiquinone species when
univalently oxidized. Metal centers are similarly facile at
single-electron redox reactions. Because respiration directs a large
electron flux through a series of flavins, quinones, iron-sulfur
clusters, and heme groups, it is widely suspected that the respiratory
chain may be the primary source of superoxide in aerobic organisms.
This expectation was supported by an investigation of superoxide
production by lysates of E. coli, which demonstrated that the
particulate fraction generated more superoxide than did cytosolic
enzymes(1) .
Strains and Media
Strains and plasmids used in this study are described in Table 1. LB medium contained 10 g of bactotryptone, 5 g of yeast
extract, 10 g of NaCl, and 2 g of glucose per liter. Defined media
contained minimal A salts (15) plus carbon sources and
thiamine. Casamino acids medium contained (per liter) 10 g of casamino
acids; glucose/amino acids medium, 5 g of glucose, and 1.5 g of
casamino acids; succinate medium, 25 mM succinate; and
glycerol/fumarate medium, 0.5 g of casamino acids, 40 mM glycerol, and 40 mM fumarate. Plasmid pH3 was maintained
with 60 µg/ml ampicillin; and plasmid pGS133 was maintained with
100 µg/ml kanamycin, since this concentration reportedly assists in
overexpression of the enzyme(14) .
Men
strains, pre-cultures were
grown anaerobically (without 4-hydroxybenzoic acid) in order to avoid
outgrowth of any Ubi
revertants, and cultures were
shifted to air-saturated medium two to three generations prior to
harvesting. The frequency of reversion was checked by streaking the
culture at the time of harvest onto LB medium; revertants, which form
large colonies, always comprised <0.1% of the harvested bacteria.
Genetic Techniques
P1-mediated transduction and plasmid transformation were
conducted according to standard methods(15) . To prevent
reversion of the ubiA420 allele during strain constructions, 1
mM 4-hydroxybenzoic acid was supplemented to the growth medium
in order to allow some ubiquinone biosynthesis(16) . The
presence of the sdhC4 and 
(frdABCD)18 null
alleles, respectively, were confirmed by the inability of such mutants
to grow in aerobic succinate medium or anaerobic glycerol/fumarate
medium, respectively. The sdhC4 null allele in
quinone-deficient strains was verified by enzymatic assay of
succinate:plumbagin oxidoreductase activity (see below), and the
presence of the (frdABCD)18 mutation was demonstrated in
quinoneless backgrounds by retransducing the linked Tet
marker from the putative mutant into AB1157 and confirming that
these secondary transductants had lost the ability to grow in anaerobic
glycerol/fumarate medium.Preparation of Inverted Vesicles
Aerobic cultures were grown with vigorous shaking for at
least five generations before being harvested at approximately 1
10
cells/ml (A =
0.3). Measurements made with a Clarke electrode verified that the
cultures were fully air-saturated at this density. Anaerobic cultures
were grown in an anaerobic chamber (Coy Laboratory Products Inc., Grass
Lake, MI) under 85% N
, 5% CO, 10% H and monitored spectrophotometrically throughout the growth period
to ensure that cultures were growing exponentially up to the point of
harvesting. To avoid induction of enzymes during the harvesting
process, 150 µg/ml chloramphenicol was added 10 min before cultures
were removed from the shaker or anaerobic chamber. Cultures were then
chilled in ice water, centrifuged, washed with ice-cold 50 mM
potassium P
(pH 7.8), resuspended the same buffer at about
0.5% the initial cell density, and lysed by passage through a French
pressure cell. The lysate was clarified by centrifugation at 20,000
g for 20 min. The inverted membrane vesicles were then
pelleted by centrifugation at 100,000
g for 2 h,
resuspended in the same buffer, centrifuged again for 2 h, and finally
suspended into about 0.2% the original culture volume. This procedure
eliminated all detectable superoxide dismutase activity. Vesicles were
stored on ice. With the exception of NADH dehydrogenase I, the
dehydrogenase and superoxide-production activities declined by <10%
over a week. NADH dehydrogenase I is very unstable during storage on
ice (17) and was virtually inactive in the experiments reported
here.
Assays for Superoxide Production
Cytochrome c Reduction
Superoxide formation was
quantitated as SOD-sensitive cytochrome c reduction(1) . Inverted vesicles were incubated in 3 ml of
air-saturated 50 mM potassium P (pH 7.8) at the
indicated temperatures in the presence of 10 µM
ferricytochrome c. Respiratory substrates, also at pH 7.8,
were added; if present, KCN was 3.3 mM. Consequent cytochrome c reduction was quantitated by absorbance measurements at 550
nm with a Beckman DU600 UV-visible spectrophotometer using the relation
cyc
-
cyc
= 21.0. In a
parallel reaction 20 units of manganese-containing superoxide dismutase
(which is not inhibited by cyanide) was included. The superoxide yield
was then calculated by subtraction of the superoxide
dismutase-resistant absorbance change from the total absorbance change.
Epinephrine Oxidation
The oxidation of epinephrine
by superoxide (19) was measured in 3-ml reactions containing
vesicles, respiratory substrate, and 0.5 mM epinephrine.
Reactions were buffered with 50 mM potassium P (pH
7.8), as used in the cytochrome c system above. At this pH the
oxidation of epinephrine is strictly dependent upon an enzymatic
superoxide source. Epinephrine oxidation was monitored by the
absorbance at 480 nm of its product, adrenochrome, and was almost fully
inhibitable by 400 units of SOD. The superoxide yield from respiratory
enzymes was quantitated ( = 4.0)
after calibration of the detection system using xanthine oxidase, whose
superoxide production activity had previously been determined by the
cytochrome c method. The four-electron oxidation of
epinephrine by O occurs by a chain reaction, the length of which
depends upon epinephrine concentration and buffer composition. With
both xanthine oxidase and respiratory enzymes the ratio of
superoxide-dependent cytochrome c reduction to adrenochrome
formation was 1.58 ± 0.04, indicating that the chains are of
identical length and justifying the use of this technique as a
quantitative assay of O.
Lucigenin Luminescence
Luminometry was performed
in a Turner Model 20e luminometer (Turner Designs, Mountain View, CA).
One-ml reactions contained 100 µM lucigenin, vesicles, and
substrate in 50 mM glycine buffer (pH 9.0). Reactions were
initiated by the addition of substrate, and data were collected and
summed over the subsequent 10 s. Light emission with any superoxide
source was completely quenched at pH 7.8, requiring that reactions be
run at the higher, nonphysiological pH.Assays of Respiratory Oxidase and Dehydrogenase
Activities
Oxidase activities were measured in potassium P buffer with a Clarke electrode. Dehydrogenase activities were
assayed using 0.4 mM of the naphthoquinone plumbagin
(5-hydroxy-2-methyl-1,4-naphthoquinone) to mediate electron transfer
from the dehydrogenase to cytochrome c, in the presence of 3.3
mM KCN(20) . Ferricyanide reductase measurements were
made with 1 mM ferricyanide added to vesicles in the presence
of 3.3 mM KCN and saturating reductive substrate (21) using = 1.0 at 420 nm
for ferricyanide. In order to calculate turnover numbers, succinate
dehydrogenase was quantitated by absorbance change of
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide in the
presence of phenazine methosulfate exactly as described(14) .
Fumarate reductase was quantitated for the same purpose using the
phenazine methosulfate-mediated reduction of
dichlorophenolindophenol(22) , with extrapolation to determine
the rate at infinite concentration of the mediator, as
described(21) .
Men
strains, NADH oxidase activity was measured and compared to NADH
dehydrogenase activity. The ratio of NADH oxidase activity per NADH
dehydrogenase capacity in quinoneless vesicles was approximately 1%
that of wild-type vesicles, confirming the virtual or complete absence
of endogenous ubiquinone and menaquinone.
Determination of K
Inverted vesicles were prepared from AB1157 after anaerobic
growth in glycerol/fumarate medium. The succinate:plumbagin
oxidoreductase activity of fumarate reductase was determined at 12
concentrations between 20 µM and 40 mM succinate
as described above, using 0.4 mM plumbagin, which is virtually
saturating even at 80 mM succinate. The effects of competitive
inhibitors were determined at four concentrations of inhibitor, and Kof
Fumarate Reductase for Malonate and Malate
values were calculated by
Lineweaver-Burk analysis. To determine apparent K
values for blockage of O formation, vesicles were incubated
with 20 mM
-glycerolphosphate. In this circumstance
electrons are transferred from the -glycerolphosphate
dehydrogenase through the quinone pool to Frd, which autoxidizes at a
high rate. Rates of O production at the fumarate reductase site were
measured in the presence of inhibitors. Inhibitors did not lessen
-glycerolphosphate:plumbagin oxidoreductase activity, indicating
that they did not inhibit the -glycerolphosphate dehydrogenase.
Frd did not exhibit any malonate:plumbagin or malate:plumbagin
oxidoreductase activity, demonstrating that neither inhibitor was
redox-active with the enzyme.Miscellaneous
Measurements of succinate dehydrogenase behavior were
conducted after a 10 min preincubation with succinate at 37 °C to
displace inhibitory oxaloacetate. Xanthine oxidase was used to generate
superoxide with 50 µM xanthine in the standard potassium
P medium. Xanthine oxidase activity was independently
assayed by monitoring urate production at 295 nm
[(xanthine to urate) =
11.0]. Superoxide dismutase was assayed using the xanthine
oxidase/cytochrome c method(23) . The production of
H
O was quantitated using the horseradish
peroxidase-coupled oxidation of o-dianisidine(24) .
Protein was assayed with the Coomassie Dye reagent (Pierce).Chemicals
B-NADH, succinic acid, DL--glycerolphosphate2Na, D-lactic acid, E. coli manganese-containing superoxide dismutase, xanthine
oxidase, xanthine, horse heart ferricytochrome c (type III),
horseradish peroxidase (type II), plumbagin,
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide,
diethylenetriamine pentaacetic acid, phenazine methosulfate,
2,6-dichlorophenolindophenol, o-dianisidine, fumaric acid,
malonic acid, racemic malic acid, oxaloacetic acid, potassium
ferricyanide, lucigenin (bis-N-methylacribinium nitrate),
Tris, MOPS, chloramphenicol, ampicillin, kanamycin sulfate,
tetracycline hydrochloride, acid-hydrolyzed casamino acids, vitamins,
and uracil were purchased from Sigma. Potassium cyanide,
4-hydroxybenzoic acid, L-lactic acid, and
(R)-(-)-epinephrine were from Aldrich; monobasic and dibasic
potassium phosphate salts, from EM Science. Water used for all in
vitro experiments was house-deionized water further purified by
passage through a Labconco Water Pro PS system.
Superoxide Is Produced during Respiration in
Vitro
Inverted respiratory vesicles were prepared from wild-type E. coli after aerobic growth in casamino acids medium. This
medium was chosen because it promotes synthesis of a wide variety of
respiratory dehydrogenases(25) . Accordingly, when incubated in
air-saturated buffer, the isolated membranes readily oxidized each of
the respiratory substrates NADH, succinate, D-lactate, L-lactate, and -glycerolphosphate. The concomitant
production of superoxide was assayed by its ability to reduce
cytochrome c, as described under ``Materials and
Methods,'' and the calculated superoxide yields are presented as Table 2.
bacteria in order to enable
comparisons among the three strains. Note that the abscissa is
drawn on an exponential scale. Circles, no additions. Squares, 3.3 mM cyanide present. B,
superoxide from vesicles of AB1157 (Sdh
Frd
). Dashed curve, no additions, with
signal expanded 7.5-fold so that the profile can be compared to that
obtained in the presence of cyanide. Triangles, rate of
superoxide production by xanthine oxidase, determined by the cytochrome c method. C, superoxide from vesicles of JI241
(Sdh
Frd
). D, superoxide
from vesicles of JI222 (Sdh
Frd
).
Superoxide Does Not Escape from the Terminal Oxidases
during the Reduction of Oxygen to Water
The purpose of this
study was to identify sites on the respiratory chain that rapidly
generate superoxide. A modular diagram of the aerobic chain is shown in Fig. 2A. Since all of the respiratory substrates
stimulated the production of superoxide, the most economical hypothesis
was that the superoxide was evolved by either the quinones or the
cytochrome oxidases, since electrons pass through these carriers no
matter what the substrate. In particular, one might imagine that
superoxide occasionally escapes from the cytochrome oxidases as a
partially reduced intermediate in the four-electron reduction of
molecular oxygen to water. To test this possibility, cyanide, which
binds to the cytochrome oxidases and prevents their association with
molecular oxygen, was added to the in vitro assays of
superoxide production. The effect of the cyanide was actually to
increase the rate of superoxide production, with each respiratory
substrate (Table 2). These data indicate that superoxide was
formed when electrons were leaked to molecular oxygen by a component
upstream of the cytochrome oxidases; the acceleration by cyanide was
presumably due to the greater electron occupancy on that component,
since outflow was blocked. A similar enhancement of superoxide
production occurs when downstream inhibitors are added to respiring
submitochondrial particles (13) .
-glycerolphosphate dehydrogenase are
depicted, and anaerobic reductases induced by alternative electron
acceptors are omitted. For a complete treatment see (43) and (44) . A, aerobic chain. B, anaerobic chain.
Fumarate is acquired either by import from the growth medium (45) or by reversal of the terminal reactions of the TCA
cycle.
Superoxide Is Formed by Autoxidation of Succinate- and
NADH-reducible Dehydrogenases
Soluble ubiquinone analogues
autoxidize in vitro and in vivo, and some studies of
mammalian mitochrondria have suggested that the spontaneous
autoxidation of reduced ubiquinone might comprise an important source
of superoxide(18) . Membranes were prepared from strains that
lacked quinones due to mutations in both the ubiquinone and menaquinone
biosynthetic pathways. The poor-growth phenotype affected the relative
abundance of some dehydrogenases, reducing the -glycerolphosphate
dehydrogenase and the lactate dehydrogenase activities to virtually
undetectable levels and obviating measurements of superoxide formation
during incubation with these substrates. However, the succinate and
NADH dehydrogenases were present, and superoxide was produced at
undiminished rates when the quinoneless vesicles were incubated with
either of these substrates (data not shown). Since the electrons could
not move beyond the enzymes which succinate and NADH directly reduced,
the superoxide must have been generated by autoxidation of those
enzymes themselves. E. coli expresses two respiratory NADH
dehydrogenases; preliminary data obtained in this laboratory suggests
that NADH dehydrogenase II was responsible for the superoxide detected
in these experiments and leaves open the question of whether NADH
dehydrogenase I produces superoxide. The remainder of this report
focuses upon the site of superoxide production during respiration of
succinate.Fumarate Reductase Is a Major Source of
Superoxide
The experimental results discussed above suggested
that succinate dehydrogenase (Sdh) was the probable source of
superoxide during succinate-driven respiration. To test this
prediction, vesicles were prepared from a mutant strain devoid of Sdh
activity. Surprisingly, these vesicles evolved just as much superoxide
as did those from its Sdh-proficient parent strain (Fig. 1C).
Frd Is Responsible for O Production during Respiration of
Varied Substrates
Because Frd is situated in the respiratory
chain as a terminal oxidase, it receives from the quinone pool
electrons that originated from a variety of respiratory substrates.
Thus, because of its predisposition to transfer electrons to oxygen,
Frd could conceivably have been the source of the superoxide that was
detected when vesicles were incubated with respiratory substrates other
than succinate.-glycerolphosphate, D-lactate, or L-lactate. The essential role of Frd in generating O with
these substrates was also confirmed by the fact that the inclusion in
the assay of fumarate, which re-oxidizes Frd and thereby prevents
electron transfer to oxygen, prevented O formation with these
substrates. Oxygen-directed respiration continued in the presence of
fumarate, so the absence of O production indicates that these
dehydrogenases, the quinones, and the terminal oxidases are not
significant sources of O. The lone exception was NADH, which generated
some O through autoxidation of the NADH dehydrogenase. In anaerobically
derived vesicles, however, even the contribution of NADH dehydrogenase
to superoxide formation was overshadowed by that of Frd.
-glycerolphosphate, L-lactate, or NADH (not shown). Table 5shows that those
substrates whose dehydrogenases are most kinetically competent at
menaquinone reduction, -glycerolphosphate, D-lactate,
and, to a lesser degree, NADH(28) , were those which stimulated
the most O production by Frd.Frd Generates O in Vivo
The abundance of Frd in
anaerobic cells led to the prediction that Frd would produce much
intracellular O when such cells were abruptly shifted into aerobic
medium. Although no physical technique is currently available that can
directly measure intracellular concentrations of O, the presence of
abundant O in E. coli can be inferred from its inhibitory
effects upon growth. In particular, excess O inactivates dihydroxyacid
dehydratase (7) and thereby prevents growth in the absence of
branched-chain amino acid supplements (2) .
Frd
), JI131 (sodB Frd
), JAC4 (sodB frdA26), and YK100 (sodB
frdABCD) were cultured into log-phase in anaerobic
minimal glucose medium. At time 0, the cultures were diluted into
aerobic medium of the same composition, and growth was monitored by
absorbance at 600 nm. The outgrowth of all strains proceeded without a
lag when branched-chain amino acids were provided in the culture medium
(not shown). B, strains AB1157 (SOD Frd
), JI131 (sodB Frd
), JI314 (SOD
with a plasmid
overproducing Frd), and JI316 (sodB with a plasmid
overproducing Frd) were cultured in anaerobic minimal medium and
diluted at time 0 into aerobic medium of the same
composition.
Electron Transfer to Oxygen Occurs at the Flavin Sites of
Both Frd and Sdh
Fumarate reductase and succinate dehydrogenase
are both well characterized flavoenzymes. They each contain
[3Fe-4S], [4Fe-4S], and [2Fe-2S]
clusters. The flavin interacts directly with the succinate-fumarate
couple, the [3Fe-4S] cluster exchanges electrons with the
respiratory quinones, and the [2Fe-2S] cluster apparently
bridges electron flow between them. The catalytic role, if any, of the
very low-potential [4Fe-4S] cluster is not understood.
Men
Sdh
Frd
)
after anaerobic growth on minimal glucose + casamino acids medium.
Rates are each normalized to their maxima. Circles, superoxide
production. Squares, reduction of 0.4 mM plumbagin. Diamonds, reduction of 1.0 mM ferricyanide. Maximum
rates: 96 nmol/min superoxide production, 1.14 µmol/min plumbagin
reduction, and 1.45 µmol/min ferricyanide reduction. B,
succinate-dependent superoxide production by succinate dehydrogenase in
quinone-deficient vesicles. Vesicles were prepared from JI332
(Ubi
Men
Frd
with
a plasmid overproducing Sdh) for assay of superoxide production or
plumbagin reduction, and from AB1157 (Ubi
Men
Sdh
) for assay of
respiration. Circles, superoxide production. Maximum was 6.1
nmol/min/ml of vesicles. Rates below 1 µM succinate were
unreliable due to succinate depletion during the period of measurement. Squares, reduction of 0.4 mM plumbagin. Maximum was
1.10 µmol/min/ml of vesicles. Triangles, respiration. In
AB1157 electron transfer to plumbagin occurred at about 95% the maximal
rate of respiratory oxygen consumption.
Men
Sdh
Frd
). A, time
course of adrenochrome accumulation. Reactions contained 10 µl of
JI243 vesicles and 0.6 mM succinate. B, response to
increasing amount of vesicles. Vesicles were incubated with 0.6 mM succinate for 15 min. C, circles, superoxide
production by Frd as a function of succinate concentration. Each
7.5-min reaction contained 10 µl of JI243 vesicles; the maximum
rate of adrenochrome formation was 58 nmol/min/ml of vesicles. This
rate of superoxide production, calculated after xanthine oxidase was
used to standardize the assay, was within 1% of that determined with
the cytochrome c assay. Triangles, succinate does not
prevent the detection of superoxide generated by xanthine oxidase. The
inhibition evident when xanthine oxidase was incubated in >40
mM succinate was due to inhibition of the enzyme, as evidenced
by a parallel decline in urate production.
values for electron transfer
to quinones. In summary, superoxide was maximally generated when the
enzymes were half-saturated with substrate, and its formation was fully
inhibited when they were saturated.
O instead of
O when the fully reduced enzyme reacts with oxygen. Such a model would
predict that the concentration of succinate optimal for O production
would be lessened when respiratory blocks impeded electron outflow from
Sdh. However, although the magnitude of O formation was affected, the
optimal concentration of succinate was the same in both respiring and
quinone-deficient membranes (Fig. 4). Furthermore, high
concentrations of succinate did not accelerate HO production by these enzymes (data not shown). values of 25
µM and 1.2 mM, respectively. Neither substrate
was redox-active with Frd (``Materials and Methods'').
-glycerolphosphate dehydrogenase via
menaquinone. This route is responsible for the vast majority of O
production during the respiration of -glycerolphosphate (Table 4). The addition of either malonate or malate prevented O
formation (Fig. 6). The kinetics of inhibition suggest
K values of 110 µM for malonate and
2.5 mM for malate. These are somewhat higher than for the
succinate:plumbagin oxidoreductase activity, presumably because of the
greater affinity of these succinate analogues for the oxidized rather
than the reduced forms of the enzyme. The small amount of O that was
produced with saturating inhibitor either represents a slight degree of
continued flavin exposure or some O production at a second site.
-glycerolphosphate
dehydrogenase to the [3Fe-4S] cluster of Frd. Excess
succinate
[OOC-CH
-CH-COO]
or its analogues, malonate
[
OOC-CH
-COO]
and malate
[
OOC-CH
-CH(OH)-COO],
bind opposite the flavin of the enzyme. If excess succinate suppresses
O production by shielding the flavin from oxygen, its competitive
inhibitors may do so as well. Abbreviations:
-glyc-P,
-glycerolphosphate; DHAP, dihydroxyacetone phosphate; Glp, -glycerolphosphate dehydrogenase; MQ,
menaquinone. The respiratory enzymes are shown protruding from the
inner face of the cytoplasmic membrane. B, inhibition of O
production by malonate. C, inhibition of O production by a
racemic mixture of DL-malate.
Superoxide Turnover Numbers for Frd and
Sdh
Tentative approximations of the superoxide turnover numbers
for Frd and Sdh at 37 °C have been made using dye turnover numbers
reported by other workers. The rates did not significantly differ when
they were measured using MOPS or Tris as buffers in place of potassium
P. Kita et al. (14) reported that Sdh
reduces phenazine methosulfate with a turnover number of 1860
min. This assay was used to quantitate Sdh in the
quinoneless vesicle preparation, and the turnover number for superoxide
by the Sdh was then calculated to be approximately 13
min
. Phenazine methosulfate reductase activity was
measured in parallel with superoxide production for Frd, and a similar
calculation was made using a turnover number of 15,850 min
for electron transfer from Frd to phenazine
methosulfate(22) . The turnover number of Frd for superoxide
production was thereby calculated to be 1600 min
.
This substantially exceeds even that of xanthine oxidase (290
min
(30) ); only the NADPH oxidase of
neutrophils is known to generate superoxide so rapidly (Table 4).
It is clear why all of the succinate-dependent superoxide production in
aerobic membranes was due to Frd (Fig. 1), despite the far
greater abundance of Sdh.
Superoxide Is Produced by the Native Forms of Frd and
Sdh
Efforts were taken to establish that the superoxide is
generated by native, undamaged forms of both Sdh and Frd. This
consideration motivated the decision to study the enzymes in
quinoneless membranes rather than in purified form, since inadvertent
damage during purification or the exposure of Fe-S clusters normally
buried in the lipid bilayer might cause the enzymes to acquire spurious
autoxidation tendencies.
Sites of Superoxide Production
Most theories of
oxygen toxicity assert that cell damage is caused by partially reduced
oxygen species, either by superoxide directly or else by the
HO and hydroxyl radicals that derive from it.
Although superoxide has been presumed to arise during aerobic
metabolism by autoxidation of reduced enzymes or electron carriers, it
has been difficult to identify sources that are capable of generating
superoxide at a substantial rate. Earlier studies indicated that the
major source of endogenous superoxide in E. coli was the
respiratory chain(1) . The results reported here indicate that
superoxide evolves from the chain primarily by autoxidation of
respiratory dehydrogenases, rather than as a by-product of terminal
oxygen reduction or by quinone autoxidation. The quinone pool, in fact,
has a antioxidant effect due to its ability to quickly remove electrons
from the dehydrogenases before they are leaked to oxygen. The
involvement of quinones in generating superoxide in mitochondria has
been controversial(18) , in part due to the harshness of the
treatments necessary to extract quinones from submitochondrial
particles. However, most data indicate that at least some of the
mitochondrial superoxide arises from the respiratory
dehydrogenases(31) .-glycerolphosphate dehydrogenase generated little or no superoxide
directly. Interestingly, despite its similarities in structure and
function to Frd, Sdh produced substantial superoxide only when
respiration was artificially blocked, and even then only at a
comparatively low rate. Thus the propensity of Frd to generate
superoxide contrasts markedly to the behavior of most redox enzymes. The Physiological Impact of Superoxide Production by
Fumarate Reductase
When E. coli is grown in most
aerobic media, very little Frd is made, and so it cannot comprise a
significant source of superoxide. Instead, one might anticipate that
superoxide production by Frd would acquire significance in the
particular circumstance when anaerobic cells, which are laden with Frd,
enter an air-saturated environment. This situation is an integral part
of the lifestyle of facultative bacteria such as E. coli,
occurring when the bacteria are excreted from the anaerobic colon into
oxygenated surface waters. The abrupt exposure to oxygen of pre-formed
Frd would be expected to cause a large flux of superoxide into the
cytosol. Because electrons from any respiratory substrate can be
directed from the quinone pool to Frd, this scenario would not be
constrained to cells oxidizing succinate. Toxicity would arise because E. coli contains several critical enzymes that are rapidly
inactivated by superoxide. Furthermore, in the presence of hydrogen
peroxide, its dismutation product, superoxide accelerates the
production of DNA damage. Thus the consequence of the autoxidizability
of Frd might be that an oxidative crisis is forced upon the cell
whenever it transits from an anaerobic to an aerobic habitat. for fumarate of fumarase A is 600
µM(33) , which, if taken as an indication of the
intracellular fumarate level, would by itself indicate that superoxide
production by Frd would be largely inhibited. Particularly striking,
however, is the fact that both fumarases A and B (the anaerobic isozyme
that would be carried over from anaerobic growth) are among the handful
of enzymes known to be rapidly inactivated by superoxide(34) .
It is plausible that the excess superoxide production by Frd would
therefore diminish fumarase activity, forcing an increase in
steady-state fumarate concentration and thereby suppressing superoxide
production by Frd. Perhaps this effect upon superoxide production
favored the retention in this facultative bacterium of
superoxide-sensitive fumarases, in contrast to mammalian cells, which
have no Frd and express a structurally dissimilar fumarase that is
unaffected by superoxide.
Mechanism of Superoxide Production by Frd and
Sdh
The single-electron reduction potential of molecular oxygen
is -0.16 V; thus it is only likely to be reduced to superoxide by
particularly good univalent electron donors. Most biomolecules resist
the loss of a single electron and therefore are not sources of
superoxide. Fumarate reductase is evidently exceptional, generating
superoxide far more rapidly than any other intact metabolic enzyme yet
known. Its example demonstrates that metabolic enzymes that are not
damaged (unlike xanthine oxidase) and that have not evolved to this
purpose (unlike the NADPH oxidase of neutrophils) may nonetheless have
the ability to produce large fluxes of superoxide. Frd may provide an
opportunity to identify the characteristics that dispose redox enzymes
to generate superoxide, potentially facilitating the search for other
such enzymes.
)
I am grateful to Kay Keyer and Kevin Messner for
assistance with some of the experiments reported here and to Gary
Cecchini and Bob Gennis for the provision of plasmids. I also thank
Irwin Fridovich and Dennis Flint for provocative conversations about
this work.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Bellomio, P. A. Vincent, B. F. de Arcuri, R. N. Farias, and R. D. Morero Microcin J25 Has Dual and Independent Mechanisms of Action in Escherichia coli: RNA Polymerase Inhibition and Increased Superoxide Production J. Bacteriol., June 1, 2007; 189(11): 4180 - 4186. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Partridge, C. Scott, Y. Tang, R. K. Poole, and J. Green Escherichia coli Transcriptome Dynamics during the Transition from Anaerobic to Aerobic Conditions J. Biol. Chem., September 22, 2006; 281(38): 27806 - 27815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. L. Seib, H.-J. Wu, S. P. Kidd, M. A. Apicella, M. P. Jennings, and A. G. McEwan Defenses against Oxidative Stress in Neisseria gonorrhoeae: a System Tailored for a Challenging Environment Microbiol. Mol. Biol. Rev., June 1, 2006; 70(2): 344 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Kussmaul and J. Hirst The mechanism of superoxide production by NADH:ubiquinone oxidoreductase (complex I) from bovine heart mitochondria PNAS, May 16, 2006; 103(20): 7607 - 7612. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Wainwright, K. T. Elvers, S. F. Park, and R. K. Poole A truncated haemoglobin implicated in oxygen metabolism by the microaerophilic food-borne pathogen Campylobacter jejuni Microbiology, December 1, 2005; 151(12): 4079 - 4091. [Abstract] [Full Text] [PDF]
|
|
