From the Department of Biochemistry, Duke University Medical Center, Durham, North Carolina 27710
A field of inquiry may be said to have come of
age when conclusions initially viewed as remarkable or even
unbelievable are accepted as commonplace. Study of the biology of the
superoxide anion radical and of related free radicals, and the defenses
thereto, has now reached this happy state of maturity. Superoxide and
even hydroxyl radicals are now known to be produced in living systems, and elaborate systems of defense and repair, which minimize the ravages
of these reactive species, have been described. New members of the
superoxide dismutase, catalase, and peroxidase families of defensive
enzymes are being found, as are new targets that are modified by
O Measurement of O Because of the spin restriction, the univalent reduction of
O2 to O Detection and measurement of fluxes of O Another luminescent method that is misused to measure O One additional artifactual detector of O Is there any method which can reliably be used as a measure of
intracellular O Aconitase can be inactivated by oxidants other than O Oxidants from O Although O There is a mechanism pertinent to living cells, and we may call it the
in vivo Haber-Weiss reaction. It is a process in which O Mutations and Complementations of Superoxide Dismutases The primary defense against the damage that can be caused by
O Support for the free radical theory of senescence was provided by the
shortened lifespan of Drosophila with a mutational defect in
Cu,Zn-SOD (40). These flies were also hypersensitive toward paraquat
and were sterile. A curtailed lifespan was also evident in
Caenorhabditis elegans, which had only half the normal
complement of SOD (41). Mice lacking Cu,Zn-SOD appeared normal while
young but were less able to recover from axonal injury (43) and could not successfully reproduce. They also exhibited a shortened
lifespan.2 Lack of Mn-SOD imposed more
serious consequences (44, 45). These animals lived only a week or two
and exhibited faulty mitochondrial activities in several tissues,
especially the heart.
The familial form of amyotrophic lateral sclerosis, or FALS, has been
shown to be associated with defects in the gene encoding Cu,Zn-SOD
(sod 1) (46-48). That this disease was due to a toxic gain
of function of the mutated protein, rather than to a loss of SOD
activity, was suggested by its genetic dominance and was demonstrated
by the ability of the G93A transgene in mice to impose late onset
progressive paralysis (49, 50). There is evidence for two different
gains of function. One of these is a peroxidase activity (51), and the
other is the ability to catalyze nitration of neurofilaments by
peroxynitrite (52). Yet another novel activity of a FALS-associated
Cu,Zn-SOD (H48Q) has been observed,3 and
that is a superoxide-dependent peroxidase activity.
Cu,Zn-SOD has been reported to catalyze the production of HO·
from H2O2 (53), and the FALS-associated G93A
mutant has been shown to be more active in this regard than the
wild-type enzyme (54). However, reasons for attributing these
observations to the peroxidase activity of the enzyme, rather than to
the production of free HO·, have been delineated (6). It may be
that all of these contribute to the observed pathology in FALS.
Sporadic amyotrophic lateral sclerosis is clinically indistinguishable
from FALS so a related mechanism might be suspected. A recent report
(55) that apparent cases of sporadic amyotrophic lateral sclerosis were
homozygous with respect to Asp90-Ala Cu,Zn-SOD supports
this notion.
Another genetic disease that may be due to faulty superoxide dismutases
is progeria. Polymorphisms in the signal sequence of the gene coding
for Mn-SOD have been reported (56). This led to the proposal that
progeria might be due to faulty localization of Mn-SOD.
Down's syndrome (trisomy 21) is associated with a ~50%
overproduction of Cu,Zn-SOD, since the gene coding for this enzyme is
located on human chromosome 21. It has been suggested that this
overproduction causes the symptoms of this syndrome (57). However, this
seems unlikely. Thus a partial trisomy 21 with overdosages of Cu,Zn-SOD
was associated with none of the usual symptoms (58). This lack of
correlation of symptoms with overdosage has been replicated (59).
Varieties of Superoxide Dismutase The SOD family has been growing. The eukaryotic cytosolic
Cu,Zn-SOD and the mitochondrial Mn-SOD are now joined by an
extracellular Cu,Zn-SOD, which is referred to as EC-SOD. The human
enzyme is a homotetrameric protein with a Mr = 135,000. It shows some sequence homology to the cytosolic Cu,Zn-SOD but
is designed for function in the extracellular spaces and is secreted by
the cells that produce it. It is glycosylated and exhibits affinity for
sulfated polysaccharides, such as heparin or heparan sulfate. Thus,
although detectable in blood plasma, most of it exists bound onto the
extracellular matrix (60-62). Its location positions EC-SOD to
intercept O An entirely new SOD, with nickel at its active site, has been found in
Streptomyces (67). This is a homotetrameric enzyme whose
subunit weight is 13,000. It bears no obvious sequence homology to
known Mn-SODs or Fe-SODs.
Cu,Zn-SODs have been reported in a few species of bacteria and were
considered unusual. However, such Cu,Zn-SODs have been described with
increasing frequency (68-72), and it is now likely that they occur in
most Gram-negative bacteria where they are apt to be localized to the
periplasm (68, 69). The E. coli Cu,Zn-SOD appears to be
unusual in that it is monomeric (70), rather than being dimeric as are
other Cu,Zn-SODs. Null mutants in Cu,Zn-SOD have been prepared; the
Caulobacter crescentis null exhibited increased sensitivity
toward citrate and toward deficiencies in Ca(II) or Mg(II) (71),
whereas the Legionella null showed decreased survival in
stationary phase (72). Sequence analysis and comparisons make it clear
that Cu,Zn-SOD arose prior to the divergence of eukaryotes and
eubacteria (73). It may be that the periplasmic Cu,Zn-SOD serves as a
defense against exogenous O Nothing illustrates the many faceted nature of the defense against
oxidative stress more than does the SoxRS regulon. This group of
coordinately regulated genes is turned on by any condition that
increases O Much remains to be told of that which is already known and much remains
to be learned. I apologize to those workers who may feel slighted
because their important contributions do not appear in the list of
references and to the readers whose special interests are not
discussed. The impossibility of fairly treating the field and those
whose work created it in a minireview may be taken as another
indication of the rapid growth of knowledge in this area.
2. In addition, the involvement of O
2 in both physiological and pathological processes is being established. A
weighty tome would be needed to encompass a comprehensive coverage of
this field of study. This review will describe only aspects of the
biology of oxygen radicals that currently engage the interest of the
writer. Hopefully they will also be of interest to the reader. Other
recent reviews may serve to fill the gaps in this one (1-6).
2 in Vitro and in Vivo
2 is a facile process, and O
2
production by spontaneous as well as enzyme-catalyzed reactions has
been demonstrated. The instability of O
2 in aqueous solutions
is a hindrance to its detection and measurement. This has been
circumvented by exploiting its reaction with various "detector"
molecules such as ferricytochrome c or spin-trapping agents.
These agents are not specific for O
2. Thus, reductants other
than O
2 can reduce ferricytochrome c, and oxidants
other than O
2 can convert the spin traps to their epr-detectable hydroperoxy derivatives (7, 8). Inhibition by
SOD1 is used to lend specificity to these
methods.
2 within cells is a
goal as difficult as it is desirable. Unfortunately enthusiasm for
achieving this goal has led many investigators to use flawed methods.
An example is the luminescence that can be elicited from lucigenin.
Early studies of the lucigenin luminescence elicited by the xanthine
oxidase reaction led to the realization that the lucigenin dication
must be univalently reduced to the corresponding monocation before
reacting with O
2 in the process that leads to luminescence (9,
10). The chemistry involved was discussed in a more recent review (11).
Nevertheless the use of lucigenin as a "specific" detector of
O
2 continues. Quite recently it was shown that the lucigenin
monocation radical can autoxidize and thus produce O
2, even in
cases where no O
2 was being produced in the absence of
lucigenin (12). In studies with Escherichia coli, lucigenin
was shown to function, much as does paraquat, to increase intracellular
O
2 production (13). Hopefully the widespread but inappropriate
use of lucigenin luminescence as a measure of O
2 will
stop.
2 is
based on luminol. In this case the compound must be univalently oxidized to a luminol radical, which reacts with O
2 before the light-emitting pathway is entered upon. The problem in this case is
that the luminol radical can spontaneously reduce O2 to
O
2. Luminol luminescence thus can be caused by a variety of
oxidants and in all cases SOD inhibits (14, 15). Here again the
detector is acting as a source of O
2.
2 needs to be
mentioned because of its widespread misuse, and that is nitroblue
tetrazolium. Many enzymes can cause the reduction of tetrazolium salts
to the corresponding formazans. Reduction of nitroblue tetrazolium to the monoformazan requires two electrons and to the diformazan four
electrons. When proceeding by a univalent pathway, which is usual, one
encounters tetrazoinyl radical intermediates (16), which reduce
O2 to O
2 in a reversible process. SOD by removing O
2 displaces this oxidation to the right and thus prevents
production of the formazan. For this reason many aerobic tetrazolium
reductions are inhibitable by SOD even though O
2 was not being
produced in the system in the absence of the tetrazolium (17).
2? There is and it is based on the rapid
inactivation of [4Fe-4S]-containing dehydratases, such as aconitase.
O
2 oxidizes the clusters of these dehydratases resulting in
loss of Fe(II), and that is reversible by reduction and reincorporating
of Fe(II). The balance between these opposing processes can be used
as a measure of O
2 and has been so used in E. coli
(18) and in mammalian cells (19). The inactivation of aconitase by
O
2 provides an explanation for previously inexplicable
observations. Thus high pO2 (291 mm) increased both glucose
utilization and lactate production, by 4-6-fold, in WI38 cells (20).
Raising pO2 would increase O
2 production, and
inactivation of aconitase by O
2 would force the cells to rely
on fermentation of glucose for energy. In another example
Aspergillus niger was reported (21) to accumulate less citrate in the medium when supplied with 0.1 mg/liter Mn(II). In this
instance enrichment of the medium with Mn(II) would increase the level
of Mn-SOD in the mitochondria, and that in turn would decrease
O
2 and thus raise aconitase. The final effect would be
increased metabolism of citrate via the Krebs cycle and less citrate
excretion. Near UV irradiation of E. coli inhibited growth on succinate more than growth on glucose (22) and inhibited respiration
(23). Both of these effects can be explained by the inactivation of
aconitase by photosensitized production of O
2.
2. Of
these peroxynitrite is particularly relevant to biology, but in NO-producing cells the level of peroxynitrite is itself dependent upon
O
2 production. Inactivation by H2O2 is
relatively unimportant.
2 in Vitro and in Vivo
2 can initiate and propagate free radical
oxidations of leukoflavins, tetrahydropterins, catecholamines, and
related compounds and can inactivate [4Fe-4S]-containing
dehydratases, it does not significantly attack polyunsaturated lipids
or DNA. Yet defects in SODs, which would have the effect of raising
intracellular [O
2], do lead to cell envelope damage (24) and
to enhanced mutagenesis (25). Mechanisms by which O
2 can give
rise to more potent oxidants could explain these seeming anomalies, and
there are several such mechanisms. The simplest of these is protonation to hydroperoxyl radical, whose pKa is 4.8 and which
is a much stronger oxidant than is O
2. Association of
O
2 with other cationic centers such as vanadate (26) or Mn(II)
(27) also have this effect, but these mechanisms are unlikely to apply
generally within cells.
2 increases "free" iron by oxidizing the [4Fe-4S]
center of dehydrases such as dihydroxy acid dehydrase (28),
6-phosphogluconate dehydrase (29), fumarases A and B (30), and
aconitase (31, 32). The released iron is kept reduced by cellular
reductants, and the Fe(II) reacts with H2O2, as
in the Fenton reaction, to yield Fe(III) + HO· or its formal
equivalent, Fe(II)O. This was proposed (33) to provide an explanation
for the enhanced O2-dependent mutagenesis exhibited by sodA sodB E. coli, and it has been
experimentally verified (34, 35). The importance of release of iron by
O
2 from [4Fe-4S] clusters of dehydrases was underscored by
the recent observation of the complementation of sodA sodB E. coli by insertion and overexpression of rubredoxin reductase (36).
Rubredoxin reductase may play a role in reconstitution of the
oxidatively disassembled [4Fe-4S] clusters of dehydrases and thereby
lower the "free" iron in aerobic SOD-null E.
coli, or it may somehow scavenge O
2 within
cells.
2, and by its reactive progeny, is the SODs. The importance of these enzymes has been clarified by the phenotypic deficits of
mutants defective in their production and in a number of cases by the
complementing effects of homologous or heterologous SODs. These
demonstrations have been achieved in bacteria (25, 37), yeast (38, 39),
Drosophila (40), a nematode (41), Neurospora (42), and even mice (43-45). The consequences of a lack of both the
constitutive Fe-SOD (SOD B) and the inducible Mn-SOD (SOD A) in
E. coli include oxygen dependent decrease of growth rate, nutritional auxotrophies, hypersensitivity toward redox cycling compounds such as paraquat and quinones, and an increase in the rate of
spontaneous mutagenesis. In yeast similar problems were seen in strains
lacking either the cytosolic Cu,Zn-SOD or the mitochondrial Mn-SOD.
Envelope damage was made evident by the ability of osmolytes to
facilitate the aerobic growth of the sodA sodB E. coli and
to partially suppress the amino acid requirements (24).
2 released by phagocytic leukocytes and other
cell types. It may be important for decreasing the very rapid reaction
of O
2 with NO (63) and thus for increasing the lifetime of
NO while decreasing net production of the powerfully oxidizing
peroxynitrite (64). Knockout mice lacking EC-SOD develop normally but
were found more susceptible to the toxic effects of 1 atm of
O2 (65). Extracellular superoxide dismutase has been
reported in a variety of organisms including plants, bacteria,
Onchocerca, parasitic nematodes, and Schistosoma.
In the case of Nocardia asteroides, the extracellular SOD
has been shown to be a pathogenicity factor (66).
2.
2 production in E. coli (74, 75). Its
products include: Mn-SOD, to eliminate O
2; glucose-6-phosphate
dehydrogenase, to ensure a supply of NADPH; endonuclease IV, to repair
oxidatively damaged DNA; the stable fumarase c, to replace
the O
2-sensitive fumarases a and b;
ferredoxin reductase, to reductively repair oxidatively disassembled
[4Fe-4S] clusters; micF, to decrease the porosity of the outer
membrane; aconitase A, to replace the aconitase inactivated by
O
2; and others as well. SoxR serves as the redox sensor, which
activates production of SoxS, which then turns on the entire regulon.
SoxR contains a [2Fe-2S] cluster and is a transcriptional activator
only in its oxidized state. The SoxRS regulon is itself only half of
the orchestrated defense against oxidative stress. There is another
independent regulon turned on in response to
H2O2 and referred to as the oxyR regulon (76).