FEBS Letters 586 (2012) 585–595

journal homepage: www.FEBSLetters.org

Review

Superoxide dismutases: Ancient enzymes and new insights

Anne-Frances Miller ⇑
Department of Chemistry, University of Kentucky, Lexington, KY 40506-0055, USA

a r t i c l e i n f o a b s t r a c t

Article history: Superoxide dismutases (SODs) catalyze the de toxification of superoxide. SODs therefore acquired
Received 10 October 2011 great importance as O2 became prevalent following the evolution of oxygenic photosynthesis. Thus
Revised 27 October 2011 the three forms of SOD provide intriguing insights into the evolution of the organisms and organ-
Accepted 30 October 2011
elles that carry them today. Although ancient organisms employed Fe-dependent SODs, oxidation
Available online 10 November 2011
of the environment made Fe less bio-available, and more dangerous. Indeed, modern lineages make
Edited by Miguel Teixeira and Ricardo Louro greater use of homologous Mn-dependent SODs. Our studies on the Fe-substituted MnSOD of Esch-
erichia coli, as well as redox tuning in the FeSOD of E. coli shed light on how evolution accommo-
Keywords:
dated differences between Fe and Mn that would affect SOD performance, in SOD proteins whose
Superoxide dismutase activity is specific to one or other metal ion.
Proton-coupled electron transfer Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
Redox tuning
Iron
Manganese

1. Introduction
Superoxide dismutases, or SODs, are enzymes that play a piv-
otal role in metabolizing O2 , preempting oxidizing chain reac-
tions that cause extensive damage, and forestalling formation of
a cascade of deleterious reactive oxygen species (ROS) including
hydrogen peroxide (H2O2), hypochlorite (OCl), peroxynitrate
(ONO2) and hydroxyl radical (HO). Aerobic metabolism makes
some 18 times more energy available per glucose than does glycol-
ysis, making possible large and complex organisms. Thus, ability to
employ O2 constitutes a decisive evolutionary advantage. Yet even
now, equipped as we are with multiple SODs and numerous sup-
porting antioxidant systems, it is inexorable damage due to ROS
that eventually limits our lives.
Ability to survive O2 is such a stringent selection criterion that
very few organisms lacking SOD survived the transition from
reducing to oxidizing environment brought about by the evolution
of oxygenic photosynthesis some 2.4 billion years ago [1]. The evo-
lutionary pressure to develop protection against superoxide was
sufficiently intense that SODs evolved on at least three separate
occasions. One of these was sufficiently ancient and important that
this enzyme is found in all kingdoms of life, indicating that it
evolved even before differentiation of eubacteria from archaea.
Thus, the enzyme that keeps us young has been a spectator to
much of the evolution of life.
The most primitive versions of SOD employ Fe, consistent with
the prevalence of Fe and its ready availability in the reducing envi-
ronment in which life is believed to have begun. However more
modern organisms employ a version of this enzyme that requires
Mn for activity, consistent with diminished bioavailability of Fe
and increased Fe toxicity as O2 levels rose. Thus, changes in the
inorganic chemical makeup of the environment appear to have dri-
ven evolution of SOD. In what follows, I discuss chemical chal-
lenges inherent in the transition between catalysis based on Fe
and catalysis based on Mn. I then review what has been learned
about the occurrence and phylogeny of modern SODs. A growing
wealth of such information is beginning to reveal the paths by
which we have acquired and inherited the SODs we have today,
from diverse ancestors.
2. Three distinct families of SODs
The name SOD denotes not one, but three unrelated enzymes.
All three earned the name by virtue of ability to convert two mol-
ecules of superoxide to one each of dioxygen and hydrogen perox-
ide, with consumption of two equivalents of H+.
One family of SODs employs a Ni ion to mediate the chemistry,
another uses a Cu ion complexed with a Zn to do the same. The
third family encompasses enzymes that use a Mn or an Fe as well
as enzymes that can use either.1 The different families of SODs differ
not only with regard to the metal ion that supports activity, but also

⇑ Fax: +1 859 323 1069. 1

We refer to enzymes active only with Mn bound as MnSODs, SODs requiring
E-mail address: afm@uky.edu bound Fe for activity as FeSOD and SODs active with either Fe or Mn as Mn/FeSODs.

0014-5793/$36.00 Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.febslet.2011.10.048
586 A.-F. Miller / FEBS Letters 586 (2012) 585–595

Fig. 1. Comparison of the ribbon structures that characterize the three families of SOD. In FeSOD and Cu,ZnSOD one monomer is coloured, for NiSOD two monomers are
coloured. Each structure in the second row displays the view obtained by rotating the structure above by approximately 90° around a horizontal axis in the plane of the page,
tipping what was the top of the figure in the upper row towards the back. The NiSODs are homohexamers of four-helix bundles of total molecular weight  80 kDa [7]. Each
four-helix bundle binds a Ni ion (green sphere) at the N-terminus. However each Ni site also derives supporting hydrogen bonds from residues from the neighboring four-
helix bundle in a reciprocal arrangement, so two four-helix bundles are colored and the other four are shown in greys. The Cu,ZnSODs are generally 32 kDa homodimers (or
dimers of dimers), where each monomer is a flattened eight-strand beta barrel. The Cu and Zn ions (gold and silver spheres, respectively) are bound on the outside of the
barrel by two loops, including a short helix. The Fe or MnSODs are 45 kDa homodimers (or dimers thereof) where each monomer includes an alpha-helical N-terminal domain
and a C-terminal domain comprised of a three-stranded beta sheet surrounded by alpha helices. The Mn or Fe ion (red sphere) is bound between the two domains by two
amino acids from each and a solvent molecule (orange sphere). Cartoons were made using Pymol [8] and the coordinates 1Q0D.pdb for NiSOD [3], 1HL5.pdb chains F and M
[7,9] for Cu,ZnSOD, and 1ISB.pdb for FeSOD [10,11].

with regard to the protein fold (Fig. 1). Recent progress in under-
standing the mechanism of Ni-SODs has been described by [2,3],
and ongoing advances in understanding the biological significance
of Cu,Zn SODs are summarized in [4] while the mechanism is re-
viewed in [5,6]. The current review will concentrate initially on
the Fe- and/or Mn-utilizing SODs and then place them in the context
of the occurrence and possible evolution of all three SODs.
3. New SODs from old: differences required for Mn-based SOD
activity vs. Fe-based activity
Diverse primitive organisms employ FeSOD, yet it is apparently
absent from animals and fungi, and present in plants only in the
chloroplasts. This suggests that ancient versions of this enzyme
used Fe but that there was a switch to the use of Mn that occurred
concurrent with evolution of eukaryotes. The switch could have oc-
curred in response to diminished bioavailability of Fe once the
atmosphere became oxidizing and Fe’s tendency to engage in Fen-
ton chemistry with superoxide, especially under conditions of oxi-
dative stress. However the cost of Fe acquisition continues to be
borne for production of hemes and numerous FexSx and Fe-contain-
ing enzymes. A second factor will be that for many Fe-dependent
systems, the chemistry in question simply does not lend itself to
Mn-mediated catalysis. Another factor was likely the ease with
which evolution could execute a transition from Fe-based activity
to activity based on Mn, without ‘interruption of service’. Most
simply, for evolution to be able to effect change, the enzyme
should possess a selectable amount of the destination activity to
serve as the raw material for the action of evolution. SOD meets
both criteria, as witnessed by bacterial and archaeal SODs that
can function with either Fe or Mn. Indeed, these ‘cambialistic’ SODs
are often found in anaerobic organisms that are relatively primi-
tive. Thus modern FeSODs and MnSODs could have evolved from
a common ancestor via SODs able to use either metal.
In order to learn what distinctions might be involved in opti-
mizing Mn-dependent as opposed to Fe-dependent SOD activity,
we compared the MnSOD and FeSOD of Escherichia coli. As for FeS-
ODs and MnSODs in general, those of E. coli share a conserved pro-
tein fold and a dimer interface that involves active site residues as
well as the funnel by which substrate gains access to the catalytic
metal ion (Fig. 2A). Both are dimers; they share 43% amino acid
identity including the three His and one Asp that bind the metal
ion in all Fe- and MnSODs, in addition to the conserved Tyr and a
nearby Gln2 (Fig. 2B). In both, the Fe or Mn ion also binds to a sol-
vent molecule that in turn engages in hydrogen bonds (H-bonds)
with the second sphere Gln. Thus, comparison of the two active sites
does not suggest an obvious explanation for the different metal ion
dependencies of activity. Moreover both sites are able to bind either
metal ion in similar coordination geometries3 [12] with similar elec-
tronic structures [13–15].
The very similar sites provide a conserved context in which to
identify differences that could enable one protein to support Mn-
based activity while the other supports only activity based on Fe.
Fe and Mn have similar ionic radii and ligand preferences, and both
cycle between their 2+ and 3+ oxidation states in the course of SOD
turnover. However the oxidation states with the same charge cor-
respond to different d-electron configurations for Mn and Fe, and
this has consequences for both the ease with which the 3+ ions
can be reduced to the 2+ state, and the tendency of each to coordi-
nate a sixth ligand. Thus, the different metal ions possess different
natural tendencies that may require different modulation and
complementation on the part of the protein.
2
This Gln is replaced by a His in Mycobacterium tuberculosis and its relatives.
3
The active site pK of Fe3+(Mn)SOD is considerably lower than that of Fe3+SOD.
Therefore the structure of Fe-substituted MnSOD (Fe(Mn)SOD) reveals a second
coordinated OH- in one active site that is absent from the structure of FeSOD [12].
 A.-F. Miller / FEBS Letters 586 (2012) 585–595 587

Fig. 2. (A) Overlay of the backbones of FeSOD (orange) and Fe-substituted MnSOD (Fe(Mn)SOD), magenta) and B: detail of the active sites of the overlain proteins. Only one
monomer is shown. Figure was made using Pymol [8] and the coordinates 1ISB.pdb and 1MMM.pdb respectively [10,12]. The amino acid numbering of E. coli FeSOD and
MnSOD are used throughout.

4. Differences between Fe and Mn to be accommodated by nor or acceptor. This allows them to function as stand-alone
different SOD proteins antioxidant defenses and also to consume two equivalents of
superoxide per turnover.
Fe- and MnSODs alternate between the 2+ and the 3+ state of The Em of hexaaquo Fe is 770 mV, between the Ems of the two
the metal ion in the course of disproportionating superoxide. In SOD reactions, but much higher than the midpoint, so the (Fe)SOD
the equations that follow, M indicates the metal ion and the pro- protein should depress Fe’s Em by some 400 mV for optimal theo-
tonation state of the coordinated solvent molecule is also indi- retical SOD activity (Fig. 3). This is likely accomplished to a signif-
cated, immediately following the metal ion. icant extent by the use of a negatively charged Asp ligand.
3þ
However hexaaquo Mn’s Em is much higher (1510 mV), so I pro-
O þ
2 þ H þ SOD  M  OH ! O2 þ SOD  M2þ  OH2 ð1aÞ posed that for optimal performance, a Mn-specific SOD protein
þ 2þ 3þ 
O
2 þ H þ SOD  M  OH2 ! H2 O2 þ þSOD  M  OH ð1bÞ would have to depress the metal ion’s Em significantly more than
3+
would an Fe-specific SOD protein (Fig. 3). Indeed, I proposed that
In the first half reaction (1a), reduction of Mn involves conver-
sion of a metal ion with 4 d-electrons to a 5 d-electron system, but
reduction of Fe3+ corresponds to accommodation of a sixth d-elec-
tron. A fifth electron is readily accommodated by an empty d-orbi-
tal4 whereas the sixth electron has to share a d-orbital with one of
the pre-existing d-electrons. Thus, reduction of Mn3+ to Mn2+ is
much more favourable than reduction of Fe3+ to Fe2+. The energetics
of reduction are embodied by the reduction midpoint potential
Em = DG°/nF, where F is Faraday’s constant (Coulombs per mole
of electrons) and n is the number of electrons being transferred
(=1 here). Thus, just as a large pKa is indicative of a base that readily
acquires a proton, a positive Em signifies a site that readily acquires
an electron and is a good oxidant. Sites with negative Ems tend to do-
nate electrons and act as reductants (just as compounds with low
pKas release a proton). The greater intrinsic affinity of Mn3+ for an
electron yields Ems that are 300 to 700 mV higher than those of
Fe3+ for analogous simple complexes5 [16,17].
The higher Em of high-spin Mn3+/Mn2+ explains its lower Fenton
reactivity than Fe and makes it a safer metal ion for use under con-
ditions of oxidative stress. However a high Em also makes it ineffi-
cient in SOD activity. The ideal Em for a SOD would be half-way
between the Ems of SOD’s two half-reactions, near 360 mV vs. the
normal hydrogen electrode (NHE) [18]. A metal site with too low
an Em will be able to reduce superoxide (Eq. (1b)) but will not be
able to extract an electron from superoxide (Eq. (1a)). A site with
too high an Em will be able to oxidize superoxide (Eq. (1a)), but will
not be able to pass the gained electron on to a second molecule of
superoxide (Eq. (1b)). Both types of site can consume superoxide
but only stoichiometrically, unless they are regenerated by other
electron transport systems. Yet part of the elegance and biochem-
ical economy of SODs is their independence from any electron do-

4
In both FeSOD and MnSOD, Fe and Mn are high-spin in both oxidation states.
5
When the Em is higher by 300 mV, reduction is more favourable by  30 kJ/mol or Fig. 3. cartoon of redox tuning of the Fe3+/2+ and Mn3+/2+ ions by the (Fe)SOD and
7 kcal/mol. (Mn)SOD proteins (the proteins of FeSOD and MnSOD, respectively).
588 A.-F. Miller / FEBS Letters 586 (2012) 585–595
greater redox depression on the part of the MnSOD protein could
explain the inactivity of Fe-substituted MnSOD [19,20].
4.1. Ems of metal-substituted SODs
In order to determine whether the (Fe)SOD and (Mn)SOD pro-
teins apply different redox tuning to their bound metal ions, we
compared the Ems of Fe in FeSOD with Fe bound in the (Mn)SOD
protein. We also assessed the Ems of Mn bound in each of the
two proteins [14,20]. (In what follows, ‘(Fe)SOD’ and ‘(Mn)SOD’
stand for the proteins of the FeSOD and MnSOD, respectively,
Mn(Fe)SOD indicates Mn-substitutes FeSOD and Fe(Mn)SOD indi-
cates Fe-substituted MnSOD.)
We confirmed that the reduction midpoint potentials of FeSOD
(100 mV) and MnSOD (300 mV) are both intermediate between
the Ems of SOD’s two half-reactions [20,21]. However we found
that the Em of Fe(Mn)SOD was much lower and that of Mn(Fe)SOD
was much higher. While it is difficult to obtain precise Em values
for the SODs, due to the very slow equilibration of the active site
metal ion with mediators, experiments with redox-active dyes
confirmed that when Fe is bound in the (Mn)SOD protein, its Em
is some 300 mV lower than when it is bound in (Fe)SOD protein
[20]. Thus, FeSOD could be reduced by ascorbate (Em  60 mV)
but Fe(Mn)SOD could not. Fe2+(Mn)SOD could reduce superoxide,
demonstrating that the reduced state could bind substrate and
moreover was competent for both the proton transfer and electron
transfer inherent to reaction 1b [20]. However Fe3+(Mn)SOD lacked
Fe3+SOD’s ability to accept an electron from superoxide despite its
greater-than-WT ability to bind small anions [15]. Moreover this
was true even at lower pH, where competitive inhibition by OH
is attenuated [20]. Thus, it appears that the (Mn)SOD protein de-
presses the Em of Fe much more than does (Fe)SOD (Fig. 3).
We showed that the same holds true for Mn. We found that the
Mn(Fe)SOD was fully reduced as-isolated and very difficult to oxi-
dize (Fig. 3). Although the Mn3+ state could be generated using the
very powerful oxidants KMnO4 and K2IrCl6 , milder oxidants able to
oxidize MnSOD were unable to oxidize Mn(Fe)SOD. We were not
able to execute a titration, but based on the Ems of the oxidants
that failed to oxidize Mn(Fe)SOD, its Em must be >870 mV [14].
Thus we found that for Fe, the (Mn)SOD protein produces an Em
some 300 mV lower than that produced by (Fe)SOD, and for Mn,
the (Mn)SOD protein produces an Em at least 500 mV lower than
that produced by (Fe)SOD. These differences are consistent with
the differences between model Mn3+/2+ and Fe3+/2+ complexes.
Moreover our findings suffice to explain the inactivity of E. coli
Fe(Mn)SOD and Mn(Fe)SOD [22]. They also predict that SODs that
are active with either Fe or Mn will apply redox tuning intermedi-
ate between the tuning applied by E. coli’s metal-specific SOD
proteins and that their Fe-versions will have lower Ems than their
Mn-versions. We are testing these predictions.
It has been pointed out that Fe3+(Mn)SOD binds a second OH
more tightly than does Fe3+SOD and that this could be responsible
for the low activity [23,24]. However since Fe3+(Mn)SOD also binds
other small anions much more tightly than does Fe3+SOD [15,24], it
remains unknown whether its affinity for substrate will also be
elevated. Competitive inhibition by OH does not depend only on
the affinity for OH, but on the relative affinity for OH vs. sub-
strate, which is not known. Moreover the restoration of a small
amount of activity at lower pH is exactly what is expected consid-
ering that reduction of FeSOD is accompanied by protonation of the
coordinated solvent [25] (Scheme 1), because the Em of a proton-
coupled-reduction will increase by 60 mV for each pH unit
decrease in pH. Thus lower pH would raise Fe(Mn)SOD’s low Em
towards a catalytically competent value. Thus, while the redox
tuning is related to proton transfer, H-bonding and likely substrate
binding as well, and these features will surely all contribute to the
Scheme 1. Proton-coupled reduction of the metal ion of SODs, where M stands for
Fe or Mn, and the Gln H-bonding to the coordinated solvent is Gln69 of FeSOD or
Gln146 of MnSOD (see Fig. 2B). Details of the coordination are omitted from the
right-hand side for the sake of clarity, since they do not change.
rate of turnover, the Em places a fundamental thermodynamic limit
on what chemistry the enzyme can conduct. The Em measured
for Fe(Mn)SOD is too low for it to serve as an electron acceptor
from superoxide and thus suffices to explain the extremely low
turnover rate of Fe(Mn)SOD at neutral pH [20].
Our findings indicate that in order to optimize SOD’s perfor-
mance with Mn as the catalytic metal ion, evolution would have
had to identify ways of depressing the bound Mn’s Em more than
had been necessary for Fe.
5. A second challenge: different preferences regarding
coordination number
Besides the large difference between the intrinsic Ems of Mn3+/2+
and Fe3+/2+, the different electronic configurations of these ions in
the corresponding oxidation states are also expected to produce
different ligand-binding preferences. As noted above, Fe3+ is a d5
ion (it has 5 d-electrons) and Fe2+ is not, whereas for Mn this prop-
erty is reversed with Mn2+ being a d5 ion and Mn3+ not. For Fe3+
and Mn2+ with one electron in each d-orbital, there is no orbital
with special priority. Ligand binding affects the metal ion’s d-elec-
trons to similar extents regardless of which orbitals the ligand
interacts with. For a d6 ion, one d-orbital has two electrons and
the complex is more stable overall if this orbital does not interact
with as many or as strong ligands as the other (singly occupied)
orbitals do. Thus, it is beneficial to avoid coordinating a sixth ligand
and retain an empty coordination site where ligand would have
interacted with the doubly-occupied orbital. By contrast a d4 ion
benefits from asymmetry in the opposite way. Because one d-orbi-
tal is empty, strong ligands can bind the metal ion along the axis
defined by the empty dz2 orbital without incurring a large cost to
the stability of the complex. Meanwhile the axes describing the
occupied d-orbitals are populated by less strong or fewer ligands
for maximum stability of the complex [26]. Thus, Mn3+ SOD can
keep its four d-electrons in the four d-orbitals that have the least
contact with the strong OH ligand that defines the Z axis, and in-
stead populate the orbitals in the X-Y plane defined by the weaker
two His and Asp ligands (1.5 ligands per direction vs. 2 along Z).
This is an example of Jahn-Teller stabilization. d6 Fe2+ benefits from
having only three ligands in the X-Y plane, as it can doubly popu-
late dYZ at lower cost than doubly populating the dz2 which is di-
rected straight at two ligands [26]. Mn2+ or Fe3+ with one
electron in each d orbital cannot exploit an asymmetric ligand
environment and so tend not to favour one: they can accommodate
a sixth ligand and benefit from the electrostatic advantages of
doing so whereas Fe2+ or Mn3+ would sacrifice advantages stem-
ming from a less symmetric coordination environment.
Consistent with the above, small anions bind directly to the Fe3+
of Fe3+SOD, but not the Fe2+ of Fe2+SOD [27]. Indeed, NMR spectros-
copy provided the first positive evidence for substrate analog bind-
ing to the reduced state of Fe2+SOD, indicating that substrate
analogs (and substrate) bind outside the coordination sphere of
 A.-F. Miller / FEBS Letters 586 (2012) 585–595
Fe2+, in an outer-sphere binding site. Similarly, Fee showed that the
pK near 9 of the Fe3+SOD active site corresponds to OH coordinat-
ing to Fe3+ [28]. However for Fe2+SOD we used 13C NMR and 13Cf-
Tyr labelled Fe2+SOD to show that the corresponding active site pK
of Fe2+SOD represents an outer-sphere event: deprotonation of
Tyr34 [29].
In contrast with Fe3+, Mn3+ is a d4 ion and indeed, although
N3 binds to Mn3+SOD, the complex it forms is predominantly out-
er-sphere at room temperature [30]. Thus, the (Mn)SOD protein
provides assistance to substrate analog binding, in the form of an
H-bond from Tyr34 [31]. In accordance with its d5 electronic con-
figuration, Fe3+(Mn)SOD displays considerably higher affinity for
coordinating anions such as N3, F and OH than does Mn3+SOD
and the ions coordinate Fe3+ rather than binding in the outer
sphere as in Mn3+SOD. The fact that Fe3+(Mn)SOD’s affinities for
small anions are also considerably higher than those of Fe3+SOD
suggests that assistance provided by the (Mn)SOD protein is
retained upon Fe-substitution.
In Mn2+SOD, Un and Whittaker have shown that substrate ana-
logs coordinate to Mn2+ consistent with this ion’s d5 configuration,
and in contrast to Mn3+SOD and Fe2+SOD [32,33]. Tabares and Un
showed that Tyr34 favoured coordination of the anion to Mn2+ at
higher temperatures approaching physiological values, whereas
for the Mn3+ state the native Tyr tends to suppress formation of
an inner-sphere complex [24]. This would tend to augment rather
than compensate for the coordination preferences of Mn2+ vs.
Mn3+, however it would have the benefit to the enzyme of favour-
ing dissociation of peroxide formed in reaction 1b by emphasizing
outer-sphere interactions with Mn3+.
Mn2+ should be competent to bind substrate analogs in the in-
ner sphere, however Mn2+(Fe)SOD has much lower inner-sphere
anion binding ability than does Mn2+SOD, since 100 mM N3 or
F did not cause any discernable change in the Mn2+ EPR signal
[14]. The (Fe)SOD protein environment therefore seems to be less
supportive of inner sphere anion binding than the (Mn)SOD active
site, for Mn2+ as for Fe3+.
Thus, we find that the electronic configuration of the metal ion
is the dominant factor in the inner-sphere vs. outer-sphere nature
of anion binding, with the d5 state of both metal ions coordinating
substrate analogs directly, although these correspond to different
oxidation states. It does not appear that the (Fe)SOD and (Mn)SOD
proteins compensate for each metal ion’s different preferences.
Rather, the mechanistic feature exerting the strongest selective
pressure may be ability to dissociate from peroxide.
6. How the SOD proteins tune the Em
The too-low Em of Fe(Mn)SOD can explain its inactivity, but in
turn requires an explanation of its own. The differences between
the Ems for metal bound in (Mn)SOD vs. in (Fe)SOD of -300 mV
for Fe and -500 mV for Mn are very large, corresponding to 5 - 8
pH unit changes in the pKa of an acid. However (Fe)SOD is superfi-
cially very similar to (Mn)SOD (Fig. 2). Therefore, we proposed that
the different Ems could reflect differences in the extent to which
the different active sites favour uptake of the proton that is cou-
pled to metal ion reduction [20,34] (Scheme 1).
Proton-coupled electron transfer is arguably the rule, rather
than the exception, in biochemical redox reactions. Acquisition of
a proton in conjunction with metal ion reduction results in conser-
vation of the same total charge [29,35] and in the case of SOD there
is the additional chemical incentive that two protons are co-sub-
strates in the disproportionation of superoxide. Indeed, reduction
of superoxide to peroxide can occur only in conjunction with or
following protonation, since it is extremely unfavourable to add
an electron to a negatively charged diatomic molecule. SODs are
able to both oxidize and reduce the same substrate precisely
589
because they also protonate the substrate in the second of these
two reactions (1b).
The proton taken up in conjunction with metal ion reduction is
understood to be acquired by the coordinated solvent [25] so that
the redox reaction is correctly written M3+OH + e + H+ ?
M2+OH2. Because coordinated OH is a better ligand for Fe3+
(and Mn3+) than for Fe2+ (and Mn2+), I proposed that the active site
of (Mn)SOD might suppress metal ion reduction (and the Em) by
disfavouring protonation of coordinated solvent [20,22,34].6
This model moreover is consistent with phylogenetic analyses,
which find that the largest conserved difference between FeSODs
and MnSODs is the origin of the active site Gln (occasionally His)
which provides the only H-bond between the protein matrix and
the coordinated solvent7 [39,40]. In dimeric FeSODs this Gln derives
from an a helix in the N-terminal domain (positon 69 in E. coli)
whereas in MnSODs it derives from a loop between b strands in
the C-terminal domain (position 146 in E. coli) [10,41,42]. Tetrameric
FeSODs employ a residue from the second of these positions but they
use a His instead of a Gln [40]. Similarly, mutational studies have
demonstrated that residues that can tune the position and polariza-
tion of the active site Gln affect the Em of the bound metal ion as well
as the protein’s relative activity with Mn bound vs. Fe bound [34,43].
Thus the largest difference between FeSODs and MnSODs is in fact
the single residue with the largest influence over the protonation
state of coordinated solvent (Fig. 2B).
We proposed that the lower Ems produced by (Mn)SOD could be
explained by a stronger H-bond donated from Gln to coordinated
OH than in (Fe)SODs, as this would disfavour acquisition of the re-
dox-coupled proton and thus metal ion reduction [13,20]. This was
borne out by 15N NMR direct observation of the H-bonding Gln
Side chain N [34]. We found that the active site Gln of the (Mn)SOD
protein is roughly twice as strongly coupled to Fe2+ as the Gln of
(Fe)SOD, consistent with a considerably stronger hydrogen bond
in the (Mn)SOD protein.
Our model predicts that mutations that stabilize coordinated
OH vs. H2O will lower the Em (and increase the site’s activity with
Mn vs. Fe). To test our proposal, we constructed mutants in which
Gln69 was replaced by either a His or a Glu residue [21]. Whereas
Gln69 of FeSOD is constrained to be an H-bond donor by its teth-
ering H-bond from Trp [10], we postulated that His could be either
a donor or an acceptor, depending on its tautomeric state. In con-
trast, we proposed that Glu would be ionized at neutral pH and act
as an obligate H-bond acceptor. Thus we hypothesized that if
replacement with His weakens H-bond donation to coordinated
solvent then coordinated H2O should no longer be destabilized
and the Em would no longer be actively depressed. Furthermore,
we proposed that if the coordinated solvent were subject to H-
bond acceptance by Glu, favouring coordinated H2O, this would
in turn stabilize the reduced state of the metal ion and raise the
Em in the Q69E mutant.
Our hypotheses were borne out on both counts [21]. The sur-
prise was in the large magnitude of the effect. While the Q69H mu-
tant had an Em some 250 mV above that of WT, the Em of the Q69E
6
Proton transfer is likely also to be coupled to substrate binding and product
release [35]Bull, C. and Fee, J.A. (1985). Steady-State Kinetic Studies of Superoxide
Dismutases: Properties of the Iron Containing Protein from Escherichia coli. J. Am.
Chem. Soc. 107, 3295-3304, [36]Abreu, I.A., Rodriguez, J.A. and Cabelli, D.E. (2005).
Theoretical studies of manganese and iron superoxide dismutases: superoxide
binding and superoxide oxidation. J. Phys. Chem. B 109, 24502-24509, [37]Miller,
A.-F., Yikilmaz, E. and Vathyam, S. (2010). 15N-NMR characterization of His residues
in and around the active site of FeSOD. Biochim. et Biophys. Acta 1804, 275-284,
[38]Maliekal, J. et al. (2002). Comparison and Contrasts between the Active Site pKs of
Mn-Superoxide Dismutase and those of Fe-Superoxide Dismutase. J. Am. Chem. Soc.
124, 15064-15075.
7
We do not count the H-bond between coordinated solvent and the ligand Asp
because this is internal to the metal site.
590 A.-F. Miller / FEBS Letters 586 (2012) 585–595
mutant was at least 660 mV higher than that of WT [21] and the
enzyme was stable in the Fe2+ state. The direction of the changes
supported H-bonding as the dominant mechanism of redox tuning
over electrostatics, since installation of Glu, with its tendency to
negative charge, should have favoured Fe3+ over Fe2+ and lowered
the Em based on electrostatics.
Crystallography showed that the His69 was not constrained to
donate an H-bond the way Gln69 is in WT [21], moreover in the re-
duced state of Q69H FeSOD, 15N NMR spectroscopy revealed that
H69 is an H-bond acceptor instead of a donor [37]. Thus it appears
that in this mutant, residue 69 is able to adapt to the protonation
state of the coordinated solvent as opposed to imposing on it an
H-bond whose polarity is determined by the rest of the protein
[44].
In the Q69E mutant, a combination of density functional theory
(DFT) calculations and X-ray crystallographic structure determina-
tion at 1.1 Å were used to probe the protonation state and H-bond-
ing of residue 69 [44]. Crystallography showed that the overall
structure of the mutant enzyme was essentially superimposable
on that of the WT and active site atom positions differed by at most
0.5 Å [21]. Although the 1.1 Å crystal structure was refined two
ways, one assuming protonated Glu69 and one assuming ionized
Glu69, both converged on the same positions and distances. Simi-
larly, DFT calculations were performed assuming three different
possible distributions of labile protons among the ionizable active
site residues. All three energy minimizations produced the same
result: ionized Glu69 accepting H-bonds from Tyr34 and coordi-
nated solvent in an arrangement with some sharing among all
three residues [44]. The distances obtained from the four crystallo-
graphically independent active sites agreed very well with the dis-
tances obtained from DFT where the protonation states are known.
Thus, although our crystal structure was not able resolve H-bond-
ing H atoms in the active site, by reference to the DFT we could
conclude that in reduced Q69E-FeSOD, ionized Glu69 accepts an
H-bond from coordinated solvent. Moreover the H-bond is shorter
than in WT FeSOD and therefore likely stronger [44]. Indeed the
Q69E Fe2+SOD is more stable that WT, indicating that this mutation
releases strain in the active site. Thus the reduced state is more
stable.
It also appears that the oxidized state of Q69E-FeSOD is less sta-
ble in the WT-like conformation, as it is isolated with a sixth ligand
bound, most likely OH [44]. The implication is that in the oxidized
state Glu69 is neutral, in a GluH + OH <=> Glu H2O equilibrium
stabilized in the former form by OH‘s ability to coordinate Fe3+.
Thus we infer that Glu69 changes from being protonated in oxi-
dized Q69E-FeSOD to ionized in the reduced state [44]. Since the
MCD and NMR spectra reveal WT-like coordination and electronic
state for Fe in both oxidation states [45], we infer that coordinated
solvent still acquires a proton upon reduction of Fe, but consider-
ing that a proton is lost from Glu, the net reaction is transfer of a
proton from Glu to coordinated solvent. This will be much less
energetically costly than the WT mechanism in which the proton
must be acquired from water. Thus, the greatly increased Em of
Q69E-FeSOD can be explained by the combination of loss of im-
posed H-bond donation to coordinated H2O in the reduced state,
gain of a very favourable H-bond acceptance from coordinated sol-
vent instead in the reduced state, and a subtle but important
mechanistic change affecting the origin of the proton gained by
coordinated solvent [44]. Together, these three contributions are
able to explain the surprisingly large increase in the Em of Q69E-
FeSOD, and the combination of the Q69H- and Q69E-FeSOD mu-
tants demonstrates that manipulation of the energies associated
with protons whose movements are coupled to electron transfer
can indeed alter the Em by several hundreds of mV [22].
Thus we have demonstrated that the position and identity of
the residue that H-bonds to the coordinating solvent molecule is
critical: we showed that the strength and polarity of the H-bond
to coordinated solvent exerts a very strong effect on the Em of
the metal ion, which in turn is directly reflected in catalytic
activity. The stronger hydrogen bond donation we observed in
(Mn)SOD can account for the lowered Ems produced by (Mn)SOD,
and in turn provide a mechanism by which evolution could have
optimized the SOD protein for function with Mn with relatively
minor perturbations to the structure, and none to the residues
directly responsible for binding the metal ion. Our chemical
mechanism explains the conservation across species and even
phyla of different placements of the active site Gln (His) in FeS-
ODs vs. MnSODs [40].
7. Occurrences of SOD
SODs are found in all the kingdoms of life (Fig. 4). The occur-
rences of the different SODs, as well as homologies and phyloge-
netic trees provide evidence as to the evolutionary histories of
the different SODs.
7.1. NiSODs
NiSOD was first discovered in Streptomyces [49]. Genes with
strong homology to NiSOD have been identified in the genomes
of additional genus’ in actinobacteria [50] and NiSOD has been
identified in cyanobacteria [51]. Schmidt et al. found putative Ni-
SOD genes in other families of bacteria and even a potential NiSOD
fusion gene in the eukaryotic green algae Ostreococcus tauri, the
smallest free-living eukaryote known. Nonetheless, they found no
evidence for NiSODs in gram-positive bacteria, archaea or eukary-
otes other than the aforementioned green alga [50].8
Because NiSOD is restricted to relatively few groups, it appears
to have evolved after differentiation of eukaryotes and substantial
diversification among bacteria. Because NiSOD’s active site derives
mainly from nine N-terminal amino acids that are not integral to
the secondary or tertiary structure [3,7], the Ni binding site might
have been a later addition to a pre-existing four-helix bundle
protein (Fig. 1).
7.2. Cu,ZnSOD
Many Gram-negative bacterial pathogens possess periplasmic
Cu,ZnSODs. The bacterial Cu,ZnSODs retain the same monomer
structure as is found in eukaryotes, but different association be-
tween monomers. For example E. coli’s Cu,ZnSOD appears to be
monomeric [52,53] and Salmonella enterica has two genes for
Cu,ZnSOD, one of which yields a dimeric Cu,ZnSOD while the other
yields a monomeric Cu,ZnSOD [54].
14 of 15 archaeal genomes screened lacked a Cu,ZnSOD homo-
logue with over 20% homology to human Cu,ZnSOD, but Banci et al.
identified a gene with 32% identity to that of human Cu,ZnSOD in
Methanosarcina [55]. Unfortunately, catalytic activity was not
tested.
So far, protists appear to lack Cu,ZnSOD [56]. This is fascinating
if true, since plants, animals and fungi all possess Cu,ZnSOD. The
other eukaryotes’ Cu,ZnSOD could have been obtained by gene
transfer from the endosymbiont that gave rise to mitochondria.
However since many endosymbiont genes were lost in the course
of gene redistribution to the nucleus, and some protists even lost
mitochondria, the ancestor of protists could have lost more genes
than other eukaryotes did and this could explain protists’ lack of
a Cu,ZnSOD.
8
Schmidt et al. nonetheless note the presence of a peptide with high homology to
NiSOD’s active site ’nickel hook’ in a Xenopus protein DUF1619.
 A.-F. Miller / FEBS Letters 586 (2012) 585–595
Fig. 4. Distribution of SODs in cells and branches of life. Different colours indicate the different SODs: orange for FeSOD, magenta for MnSOD, blue for Cu,ZnSOD and green for
NiSOD. Quaternary states of SODs are indicated in the cell cartoon but not in the tree. This cartoon summarizes general features only, for details please see the text and Refs.
[46–48]. There is an older report of CuSOD in the nucleus.
Yeast Cu,ZnSOD has been found not only in the cytoplasm, but
also in the mitochondrial intermembrane space [57]. Indeed,
sequencing of the Candida genome revealed two genes for Cu,Zn-
SOD (reviewed by [58]).
Plants contain Cu,ZnSOD in the cytosol and in the chloroplast.
There may moreover be multiple Cu,ZnSODs expressed from sev-
eral different genes [59]. Plant Cu,ZnSODs are a highly homologous
group but there are a few features that distinguish cytosolic from
chloroplastic Cu,Zn-SODs, including the numbers and positions of
introns. A notable exception is the (presumed) cytoplasmic
Cu,ZnSOD of the liverwort (Marchantia paleacea) which more clo-
sely resembles chloroplast Cu,ZnSODs. This Cu,ZnSOD has been
proposed to resemble the last common ancestor of plant cytoplas-
mic and chloroplast Cu,ZnSODs [59]. Cu,ZnSOD has also been found
in plant peroxisomes [60].
Animal cells possess dimeric Cu,ZnSOD in the cytoplasm [61],
and also export a tetrameric glycosylated Cu,ZnSOD [62], each
based on a separate gene. The extracellular Cu,ZnSODs of
Schistosoma improve the parasites’ ability to survive the oxidant
attacks of phagocytes [63]. Cytoplasmic Cu,ZnSOD has also been
found in the nucleus [64], the intermembrane space of mitochon-
dria [57,65] and peroxisomes [66,67]. However since the precise
intracellular location of Cu,ZnSOD is responsive to the metabolic
state of cells and tissues [68] the situation is complicated and
our knowledge is incomplete. Analysis of the genes for Cu,ZnSODs
indicates that the intracellular and extracellular Cu,ZnSODs should
collectively be regarded as distinct from the bacterial Cu,ZnSODs
[69]. The extracellular SODs from mammals appear to be more clo-
sely related to fungal Cu,ZnSODs [59], raising the possibility that
the extracellular version is the more ancient one.
Cu,ZnSOD’s distribution among organelles must be interpreted
with caution, as this SOD has been found in the genomes of several
viruses where it could be attributed to gene transfer from a host
591
[59,70]. Given the relative portability of this gene, it is possible that
it has been transferred from bacteria to eukaryotes independently
of endosymbiotic events [47]. Virus-mediated gene transfer could
also explain the presence of a Cu,ZnSOD gene in just one species
of archaeon (so far) [47,55], and would be consistent with this
gene’s occasional lack of amino acid residues that are normally
responsible for metal ion binding, and thus critical for catalytic
activity [55]. Fink and Scandalios suggest that plant chloroplast
and cytosolic Cu,ZnSODs evolved from a common ancestor [59].
The fact that Cu,ZnSOD is absent from protists as well as from such
primitive plants as Chlamydomonas and moss [48] raises the ques-
tion of how it was acquired by higher eukaryotes. The large differ-
ences between eukaryotic and bacterial Cu,ZnSODs do not suggest
an obvious link between the two.
7.3. Fe-, Mn- and Fe/MnSODs
MnSODs and FeSOD are so similar that they are generally ac-
cepted to have arisen from a common ancestor [47,59] (see
Fig. 4). Since then, FeSODs and MnSODs have significantly diverged
from one another, as their amino acid sequences cluster in two
separate groups in phylogenetic trees [47,59]. Thus the presence
of FeSOD and MnSOD in different organelles and in different clas-
ses of organisms provides a fascinating perspective on the evolu-
tion of life [47].
7.3.1. Prokaryotes
FeSOD is found in both aerobic and anaerobic bacteria [71]. It is
also found in oxygenic as well as anoxygenic photosynthetic bacte-
ria [72,73]. Pathogens such as Helicobacter pylorii have FeSOD,
which may contribute to their ability to survive host defenses [74].
MnSOD is found in aerobic bacteria and its transcription is up-
regulated upon exposure to oxygen, suggesting that it represents a
592 A.-F. Miller / FEBS Letters 586 (2012) 585–595
later elaboration [75]. Cyanobacteria are shown to be unique in
containing membrane-bound manganese superoxide dismutases
(MnSOD) [76].
Some archaea such as Aeropyrum pernix and Pyrobaculum calid-
ifontis possess Mn/FeSODs that are active with either metal ion but
tend to emphasize Mn incorporation more under aerobic condi-
tions and Fe under anaerobic conditions [75,77]. The aerobic archa-
eon Sulfolobus solfataricus and facultative aerobe Acidianus
ambivalens produce SODs that are FeSODs [78,79]). Even strictly
anaerobic methanogens have been found to have SODs. Methanob-
acter bryantti and M. thermoautotrophicum have tetrameric FeSODs
whose amino acid sequence and susceptibility to inhibition bears
stronger similarity to the MnSODs’ than the bacterial FeSODs
[80,81].
The SODs of aerobic halophyles are MnSODs [53,82] which have
greater resemblance to the MnSODs of mitochondria than the
MnSODs of bacteria [59].
7.3.2. Eukaryotes
FeSOD has been found in protists ranging from the amitochond-
riate Entamoeba histolica [83], trypanosomes [56], plasmodia [84]
and Perkinsus marinus [85], and has been shown to provide these
parasites with protection against oxidative stress unleashed by
macrophages or produced by many anti-malarial and anti-trypan-
osomal drugs [56]. Some of these FeSODs are cytoplasmic and oth-
ers are mitochondrial, based on work in Plasmodium [86].
MnSOD has been reported tightly bound to the thylakoid mem-
branes of the photosynthetic protist Euglena [87]. This is reminis-
cent of the membrane-bound MnSOD of filamentous
cyanobacteria.
Numerous fungi have been found to have MnSOD.9 Among
pathogens, mitochondrial MnSOD is important for extended survival
and virulence [88]. Several fungi are reported to have cytosolic
MnSOD. For example, Candida albicans has a dimeric cytoplasmic
MnSOD in addition to the expected tetrameric mitochondrial
MnSOD (reviewed by Fréalle [58]).
Plants have long been known to have FeSODs in their chloro-
plasts [89]. Recently however FeSOD has also been found in the
cytoplasm of cowpea [90]. Some plants have several FeSOD genes
[48]. In L. japonicus, one FeSOD is localized to the chloroplast
whereas the other appears to be cytoplasmic [91]. FeSOD in the
chloroplast may associate with the plastid nucleoid and participate
in signalling or gene regulation [48].
Plants have MnSOD in their mitochondria, as expected. It was
also reported to be in chloroplasts of the unicellular alga
Chlamydomonas, but not in those of higher plants. Again, there
may be several genes, possibly to provide different SODs for differ-
ent sub-compartments of the mitochondrion. Plant MnSOD is also
found in peroxisomes, where its levels are regulated independently
of the levels of mitochondrial MnSOD [92]. Plant MnSODs share
some 70% homology, and are more distantly related to plant
FeSODs, suggesting separate evolutionary origins.
In general, animals possess MnSOD in their mitochondria10
(Crustaceans that employ Cu heavily for oxygen transport lack cyto-
plasmic Cu,ZnSOD and have a cytoplasmic MnSOD, in addition to the
expected mitochondrial MnSOD [93]). Mitochondrial MnSOD serves
complementary roles to those played by cytosolic and extracellular
Cu,ZnSOD, and accordingly, is regulated differently. In animals,
MnSOD is localized where it can metabolize superoxide produced
as a byproduct of respiratory electron transfer. This can be viewed
as a chronic problem, in contrast to bursts of reactive oxygen species
(ROS) produced by macrophages or elevated levels of ROS associated
9
A FeSOD has not yet been identified in a fungus, to my knowledge.
10
Animals seem to lack FeSOD.
with changes in peroxisome activity. However regulation of mito-
chondrial SOD activity has emerged as a very important aspect of
regulation of the levels of redox signalling molecules such as hydro-
gen peroxide (H2O2, a product of SOD activity) and nitric oxide (NO,
which is consumed by superoxide) [94,95]. For example, elevated
MnSOD levels have been related to the aggressiveness of some can-
cers [96,97]. This is an exciting and rapidly developing area, which is
unfortunately beyond the scope of this review.
8. Insights into evolution of eukaryotes, based on the
distribution of Fe and Mn SOD’s
Because Fe and/or MnSODs can be found throughout all king-
doms of life, they can tell us about even very early events in its
evolution. Indeed, several valuable analyses of SOD phylogenies
have appeared recently [47,59].
The FeSODs and MnSODs are understood to have arisen from
a common ancestor very early in the history of life, since they
are represented in both eubacteria and archaea. Mn/FeSODs
and FeSODs appear the more primitive based on their prevalence
in more primitive prokaryotes and primitive eukaryotes (Fig. 4),
as well as the lower extent of regulation associated with their
expression, compared to that of the MnSOD gene. Features spe-
cifically associated with Mn-based activity are common to bacte-
rial and archaeal MnSODs suggesting optimization of Mn use
prior to divergence of these kingdoms.11 However Fe use is asso-
ciated with quite different amino acid signatures in dimeric vs. tet-
rameric FeSODs suggesting more time for divergence within this
group, and/or possibly a return to use of Fe from among a Mn-uti-
lizing tetrameric lineage.
It is now accepted that all modern mitochondria stem from
endosymbiotic assimilation of a species of a-proteobacteria, which
contributed many genes for energy metabolism and heterotrophy
to the host genome [98]. Even amitochondriate eukaryotes
(protists) appear to be descended from an ancestor that once
possessed a mitochondrion. Nonetheless, the MnSOD of mitochon-
dria does not appear to be of bacterial origin.
The host organism that first acquired mitochondria appears to
have been of archaeal origin based on strong homologies between
eukaryotic and archaeal transcriptional and translational machin-
ery (reviewed by Martin [98]). It seems simplest to propose that
a cytoskeleton evolved first [98]. Once phagocytosis was possible,
the tremendous possibilities for accelerated acquisition of advan-
tageous traits via mergers and acquisitions would have immedi-
ately favoured the first organisms able to assimilate bacteria
[99]. Given the metabolic revolution precipitated by the transfer
of entire constellations of genes from the endosymbiont to the host
genome, it is not difficult to imagine that a burst of diversification
would follow, allowing rapid divergence and development of the
different branches of eukaryotes [100], and also eclipsing the
chances of any other branches of primitive life to compete and con-
fer similar status on their progeny to that acquired by eukaryotes.
In summary it seems reasonable to consider that acquisition of
mitochondria was very closely related to and even possibly insep-
arable from the evolution of eukaryotes, requiring only the possi-
bility of phagocytosis as a prerequisite [98,100].
Homology among SOD sequences adds to this picture that the
MnSODs of modern eukaryotes’ mitochondria appears to be de-
rived from an archaeal MnSOD. Specifically, BLASTp homology
searches based on MnSOD from halobacterium identify mitochon-
11
It is difficult to make firm conclusions because many SODs are identified as Fe- or
Mn- specific based on their amino acid sequences rather than measurements of the
activity of Fe and Mn forms, creating a risk of circular logic. In addition, specificity is
often not absolute, with many prokaryotic SODs showing some amount of activity
with each metal ion, even if one metal ion performs better than the other.
 A.-F. Miller / FEBS Letters 586 (2012) 585–595
drial MnSODs as homologues, as well as SODs from other archaea.
However they are more distantly related to bacterial SODs. In con-
trast, MnSODs from a-proteobacteria display much higher homol-
ogy with SODs of other bacteria and not with eukaryotic
mitochondrial SODs or SODs from archaea. Thus the sequence
homologies of MnSODs suggest that -1- the host of the endosym-
biosis was of achaeal lineage, -2- the host’s MnSOD gene was re-
tained and augmented with a presequence to produce a
mitochondrial targeting peptide, yielding modern mitochondrial
MnSOD whereas -3- endosymbiont MnSOD and FeSOD genes ap-
pear to have been lost from animals and fungi. Many other genes
were lost from the endosymbiont’s genome, so loss of the SOD
genes would not be remarkable. Proposal 1 reiterates conclusions
from much more extensive works [99]. However proposal 2 sug-
gests two-way traffic in that mitochondria acquired a protein of
host origin, though not the gene.
Protists appear to be an exception to the above in that they
lack a MnSOD but possess a FeSOD instead. Phylogenetic analysis
does not support its acquisition from an a-proteobacterial endo-
symbiont that became the mitochondria. The protist mitochon-
drial FeSODA and FeSODC seem more closely related to
e-proteobacteria instead, and appear to stem from a different
gene acquisition event than the one in which the cytoplasmic
FeSODBs were obtained [101]. More information is needed to set-
tle this question.
Regarding chloroplasts, plant FeSODs were found to be more
closely related to cyanobacterial FeSODs than to those of Archaea
[47]. Thus, modern plant chloroplasts are believed to have retained
a SOD of endosymbiotic origin12 [47,48,59]. It may be beneficial for
chloroplasts to employ a SOD that does not require Mn, since photo-
system II requires Mn for function. Thus, use of a FeSOD may benefit
chloroplasts by avoiding competition for the same resource between
two valuable enzymes.
In summary, the path by which we inherited Cu,ZnSOD is still a
mystery, but MnSODs appear to have been inherited by eukaryotes
from an archaeal ancestor whereas the FeSODs of protists may be
derived from lateral gene transfer and the FeSODs of chloroplasts
are likely of cyanobacterial origin.
Archaeal SODs frequently display activity with either Mn or Fe
and could represent intermediates between FeSODs that preceded
them and the MnSODs of modern mitochondria. It is interesting to
note therefore that archaeal SODs such as that of Pyrobaculum lack
the Gln at position 69 that is typical of bacterial FeSODs but pro-
vide an H-bonding partner for coordinated solvent from position
146 (based on multiple sequence alignment). Our work indicates
that this would produce a much lower Em [14,47]. However the
side chain provided by position 146 is a His instead of a Gln, and
our work suggests that this would result in a raised Em which
would therefore mitigate the effect of the placement at position
146 and produce an intermediate Em [20]. This would be consistent
with this archaeal SOD’s ability to utilize either metal ion as a basis
for SOD activity [21], potentially providing an evolutionary bridge
to Mn use. Mutation of the His to Gln would have produced the
fully depressed Em and favoured Mn use to the point that Fe was
no longer active.
Thus the different kinds of SOD provide complementary win-
dows on the evolution of metalloenzymes. Old as they are, the
superoxide dismutases are sure to yield more surprises, answers
and questions.
12
The FeSOD gene has apparently been lost from the chloroplasts of some types of
algae [59]Fink, R.C. and Scandalios, J.G. (2002). Molecular evolution and structure-
function relationships of the superoxide dismutase gene families in angiosperms and
their relationship to other eukaryotic and prokaryotic superoxide dismutases. Arch.
Biochem. Biophys. 399, 19-36.
593
Acknowledgements
This paper is dedicated to the memory and inspiring contribu-
tions to science of Prof. A.V. Xavier. The author is grateful to the
N.I.H. for funding under 1R01GM085302-01A.
References
[1] Blankenship, R.E. (2010) Early evolution of photosynthesis. Plant Physiol. 154,
434–438.
[2] Herbst, R.W., Guce, A., Bryngelson, P.A., Higgins, K.A., Ryan, K.C., Cabelli,
D.E., Garman, S.C. and Maroney, M.J. (2009) Role of conserved tyrosine
residues in NiSOD catalysis: a case of convergent evolution. Biochem. 48,
3354–3369.
[3] Wuerges, J., Lee, J.-W., Yim, Y.-I., Yim, H.-S., Kang, S.-O. and Carugo, K.D.
(2004) Crystal structure of nickel-containing superoxide dismutase reveals
another type of active site. Proc. Natl. Acad. Sci. USA 101, 8569–8574.
[4] Potter, S.Z. and Valentine, J.S. (2003) The perplexing role of copper-zinc
superoxide dismutase in amyotrophic lateral sclerosis (Lou Gehrig’s disease).
J. Biol. Inorg. Chem. 8, 373–380.
[5] Miller, A.-F. (2003) Superoxide processing in: Coordination Chemistry in the
Biosphere and Geosphere (Que, L. Jr.Jr. and Tolman, W., Eds.), pp. 479–506,
Pergamon, Oxford.
[6] Hart, P.J., Balbirnie, M.M., Ogihara, N.L., Nersissian, A.M., Weiss, M.S.,
Valentine, J.S. and Eisenberg, D. (1999) A structure-based mechanism for
copper–zinc superoxide dismutase. Biochem. 38, 2167–2178.
[7] Barondeau, D.P., Kassmann, C.J., Bruns, C.K., Tainer, J.A. and Getzoff, E.D.
(2004) Nickel superoxide dismutase structure and mechanism. Biochemistry
43, 8038–8047.
[8] DeLano, W.L. (2002). The PyMOL Molecular Graphics System. http://
www.pymol.org DeLano Scientific.
[9] Strange, R.W. et al. (2003) The structure of holo and metal-deficient wild-
type human Cu, Zn superoxide dismutase and its relevance to familial
amyotrophic lateral sclerosis. J. Mol. Biol. 328, 877–891.
[10] Lah, M.S., Dixon, M.M., Pattridge, K.A., Stallings, W.C., Fee, J.A. and Ludwig,
M.L. (1995) Structure-Function in E. coli Iron Superoxide Dismutase:
Comparisons with the Manganese Enzyme from T. thermophilus.
Biochemistry 34, 1646–1660.
[11] Ramirez, D.C., Gomez-Mejiba, S.E., Corbett, J.T., Deterding, L.J., Tomer, K.B. and
Mason, R.P. (2009) Cu, Zn-superoxide dismutase-driven free radical
modifications: copper- and carbonate radical anion-initiated protein radical
chemistry. Biochem. J. 417, 341–353.
[12] Edwards, R.A., Whittaker, M.M., Whittaker, J.W., Jameson, G.B. and Baker, E.N.
(1998) Distinct metal environment in Fe-substituted manganese superoxide
dismutase provides a structural basis of metal specificity. J. Am. Chem. Soc.
120, 9684–9685.
[13] Jackson, T.A. and Brunold, T.C. (2004) Combined spectroscopic/computational
studies on Fe- and Mn-dependent superoxide dismutases: insights into
second-sphere tuning of active site properties. Acc. Chem. Res. 37, 461–470.
[14] Vance, C.K. and Miller, A.-F. (2001) Novel insights into the basis for E. coli
SODs metal ion specificity, from Mn-substituted Fe–SOD and its very high Em.
Biochemistry 40, 13079–13087.
[15] Vance, C.K. and Miller, A.-F. (1998) Spectroscopic comparisons of the pH
dependencies of Fe-substituted-(Mn) superoxide dismutase and Fe-
superoxide dismutase. Biochemistry 37, 5518–5527.
[16] Bull, C., McClune, G.J. and Fee, J.A. (1983) The mechanism of iron EDTA
catalyzed superoxide dismutation. J. Am. Chem. Soc. 105, 5290–5300.
[17] Gupta, R. and Borovik, A.S. (2003) Monomeric MnIII/II and FeIII/II complexes
with terminal hydroxo and oxo lignads: probing reactivity via O–H bond
dissociation energies. J. Am. Chem. Soc. 125, 13234–13242.
[18] Barrette, J., Sawyer, W.C., Fee, D.T. and Asada, K. (1983) Potentiometric
titrations and oxidation–reduction potentials of several iron superoxide
dismutases. Biochemistry 22, 624–627.
[19] Miller, A.-F. and Sorkin, D.L. (1997) Superoxide dismutases: a molecular
perspective. Comments in Molecular and Cellular Biophysics 9, 1–48.
[20] Vance, C.K. and Miller, A.-F. (1998) A simple proposal that can explain the
inactivity of metal-substituted superoxide dismutases. J. Am. Chem. Soc. 120,
461–467.
[21] Yikilmaz, E., Rodgers, D.W. and Miller, A.-F. (2006) The crucial importance of
chemistry in the structure-function link: manipulating hydrogen bonding in
iron-containing superoxide dismutase. Biochemistry 45, 1151–1161.
[22] Miller, A.-F. (2008) Redox tuning over almost 1 V in a structurally-conserved
active site: lessons from Fe-containing superoxide dismutase. Acc. Chem. Res.
41, 501–510.
[23] Yamakura, F., Kobayashi, K., Ue, H. and Konno, M. (1995) The pH-
dependent changes of the enzymic activity and spectroscopic properties
of iron-substituted manganese superoxide dismutase. Eur. J. Biochem.
227, 700–706.
[24] Whittaker, M.M. and Whittaker, J.W. (1997) Mutagenesis of a proton linkage
pathway in Escherichia coli manganese superoxide dismutase. Biochemistry
36, 8923–8931.
[25] Stallings, W.C., Metzger, A.L., Pattridge, K.A., Fee, J.A. and Ludwig, M.L. (1991)
Structure–function relationships in iron and manganese superoxide
dismutases. Free Rad. Res. Comm. 12–13, 259–268.
594 A.-F. Miller / FEBS Letters 586 (2012) 585–595
[26] Jackson, T.A., Xie, J., Yikilmaz, E., Miller, A.-F. and Brunold, T.C. (2002)
Spectroscopic and computational studies on iron and manganese superoxide
dismutases: nature of the chemical events associated with active site pKs. J.
Am. Chem. Soc. 124, 10833–10845.
[27] Miller, A.-F., Sorkin, D.L. and Padmakumar, K. (2005) Anion-binding
properties of reduced and oxidized iron-containing superoxide dismutase
reveal no requirement for Tyrosine 34. Biochemistry 44, 5969–5981.
[28] Fee, J.A., McClune, G.J., Lees, A.C., Zidovetzki, R. and Pecht, I. (1981) The pH
dependence of the spectral and anion binding properties of iron containing
superoxide dismutase from E. coli B: an explanation for the azide inhibition of
dismutase activity. Israel J. Chem. 21, 54–58.
[29] Miller, A.-F., Padmakumar, K., Sorkin, D.L., Karapetian, A. and Vance, C.K.
(2003) Proton-coupled electron transfer in Fe-superoxide dismutase and Mn-
superoxide dismutase. J. Inorg. Biochem. 93, 71–83.
[30] Jackson, T.A., Karapetian, A., Miller, A.-F. and Brunold, T.C. (2004)
Spectroscopic and computational studies of the azide-adduct of
manganese superoxide dismutase: definitive assignment of the ligand
responsible for the low-temperature thermochromism. J. Am. Chem. Soc.
126, 12477–12491.
[31] Jackson, T.A., Karapetian, A., Miller, A.-F. and Brunold, T.C. (2005) Probing the
geometric and electronic structures of the low temperature azide adduct and
the product-inhibited form of oxidized manganese superoxide dismutase.
Biochemistry 44, 1504–1520.
[32] Whittaker, J.W. and Whittaker, M.M. (1991) Active site spectral studies on
manganese superoxide dismutase. J. Am. Chem. Soc. 113, 5528–5540.
[33] Tabares, L.C., Cortez, N. and Un, S. (2007) Role of tyrosine 34 in the anion
binding equilibria in manganese(II) superoxide dismutases. Biochemistry 46,
9320–9327.
[34] Schwartz, A.L., Yikilmaz, E., Vance, C.K., Vathyam, S., Koder Jr., R.L. and Miller,
A.-F. (2000) Mutational and spectroscopic studies of the significance of the
active site Gln to metal ion specificity in superoxide dismutase. J. Inorg.
Biochem. 80, 247–256.
[35] Bull, C. and Fee, J.A. (1985) Steady-state kinetic studies of superoxide
dismutases: properties of the iron containing protein from Escherichia coli. J.
Am. Chem. Soc. 107, 3295–3304.
[36] Abreu, I.A., Rodriguez, J.A. and Cabelli, D.E. (2005) Theoretical studies of
manganese and iron superoxide dismutases: superoxide binding and
superoxide oxidation. J. Phys. Chem. B 109, 24502–24509.
[37] Miller, A.-F., Yikilmaz, E. and Vathyam, S. (2010) 15N-NMR characterization
of His residues in and around the active site of FeSOD. Biochim. Biophys. Acta
1804, 275–284.
[38] Maliekal, J. et al. (2002) Comparison and contrasts between the active site
pKs of Mn-superoxide dismutase and those of Fe-superoxide dismutase. J.
Am. Chem. Soc. 124, 15064–15075.
[39] Parker, M.W., Blake, C.C.F., Barra, D., Bossa, F., Schinina, M.E., Bannister, W.H.
and Bannister, J.V. (1987) Structural identity between the iron- and
manganese-containing superoxide dismutases. Protein Eng. 1, 393–400.
[40] Wintjens, R., Gilis, D. and Rooman, M. (2008) Mn/Fe superoxide dismutase
interaction fingerprints and prediction of oligomerization and metal cofactor
from sequence. Proteins, Str. Func. Bioinformat. 70, 1564–1577.
[41] Carlioz, A., Ludwig, M.L., Stallings, W.C., Fee, J.A., Steinman, H.M. and Touati,
D. (1988) Iron superoxide dismutase. nucleotide sequence of the gene from
Escherichia coli K12 and correlations with crystal structures. J. Biol. Chem.
263, 1555–1562.
[42] Stoddard, B.L., Howell, P.L., Ringe, D. and Petsko, G.A. (1990) The 2.1-Å
resolution structure of iron superoxide dismutase from Pseudomonas ovalis.
Biochemistry 29, 8885–8893.
[43] Yamakura, F., Sugio, S., Hiraoka, B.Y., Ohmori, D. and Yokota, T. (2003)
Pronounced conversion of the metal-specific activity of superoxide
dismutase from Porphyromonas gingivalis by the Mutation of a Single
Amino Acid (Gly155Thr) Located Apart from the Active Site. Biochemistry
42, 10790–10799.
[44] Yikilmaz, E., Porta, J., Grove, L.E., Vahedi-Faridi, A., Bronshteyn, Y., Brunold,
T.C., Borgstahl, G.E.O. and Miller, A.-F. (2007) How can a single second sphere
amino acid substitution cause reduction midpoint potential changes of
hundreds of millivolts? J. Am. Chem. Soc. 129, 9927–9940.
[45] Yikilmaz, E., Xie, J., Miller, A.-F. and Brunold, T.C. (2002) Hydrogen-bond-
mediated tuning of the redox potential of the non-heme Fe site of superoxide
dismutase. J. Am. Chem. Soc. 124, 3482–3483.
[46] Bafana, A., Dutt, S., Kumar, A., Kumar, S. and Ahuja, P.S. (2011) The basic and
applied aspects for superoxide dismutase. J. Mol. Cat. B: Enzymatic 68, 129–
138.
[47] Wolfe-Simon, F., Grzebyk, D., Schofield, O. and Flakowski, P.G. (2005) The role
and evolution of superoxide dismutases in algae. J. Phycol. 41, 453–465.
[48] Pilon, M., Ravet, K. and Tapken, W. (2011) The biogenesis and physiological
function of chloroplast superoxide dismutases. Biochim. Biophys. Acta 1807,
989–998.
[49] Youn, H.-D., Kim, E.-J., Roe, J.-H., Hah, Y.C. and Kang, S.-O. (1996) A novel
nickel-containing superoxide dismutase from Streptomyces spp. Biochem. J.
318, 889–896.
[50] Schmidt, A., Gube, M., Schmidt, A. and Kothe, E. (2009) In silico analysis of
nickel containing superoxide dismutase evolution and regulation. J. Basic
Microbiol. 49, 109–118.
[51] Eitinger, T. (2004) In vivo production of active nickel superoxide dismutase
from Prochlorococcus marinus MIT9313 is dependent on its cognate
peptidase. J. Bateriol. 186, 7821–7825.
[52] Pesce, A., Capasso, C., Battistoni, A., Folcarelli, S., Rotilio, G., Desideri, A. and
Bolognesi, M. (1997) Unique structural features of the menomeric Cu, Zn
superoxide dismutase from Escherichia coli, revealed by X-ray
crystallography. J. Mol. Biol. 274, 408–420.
[53] Cannio, R., Fiorentino, G., Morana, A., Rossi, M. and Bartolucci, S. (2000)
Oxygen: friend or foe? Archaeal superoxide dismutases in the protection of
intra- and extracellular oxidative stress. Frontiers in Biosci. 5, 768–779.
[54] Ammendola, S. et al. (2008) Regulatory and structural differences in the Cu,
Zn-superoxide dismutases of Salmonella enterica and their significance for
virulence. J. Biol. Chem. 283, 13688–13699.
[55] Banci, L. et al. (2005) A prokaryotic superoxide dismutase paralog lacking two
Cu ligands: from largely unstructured in solution to ordered in the crystal.
Proc. Natl. Acad. Sci. USA 102, 7541–7546.
[56] Wilkinson, S.R., Prathalingam, S.R., Taylor, M.C., Ahmed, A., Horn, D. and
Kelly, J.M. (2006) Functional characterization of the iron superoxide
dismutase gene repertoire in Trypanosoma brucei. Free Radic. Biol. Med. 40,
193–195.
[57] Sturtz, L.A., Diekert, K., Jensen, L.T., Lill, R. and Culotta, V.C. (2001) A fraction
of yeast Cu, Zn-superoxide dismutase and its metallochaperone, CCS, localize
to the intermembrane space of mitochondria. J. Biol. Chem. 276, 38084–
38089.
[58] Fréalle, E., Noël, C., Viscogliosi, E., Camus, D., Dei-Cas, E. and Delhaes, L.q.
(2005) Manganese superoxide dismutase in pathogenic fungi: an issue with
pathophysiological and phylogenetic involvements. FEMS Immun. Med.
Micro. 45, 411–422.
[59] Fink, R.C. and Scandalios, J.G. (2002) Molecular evolution and structure-
function relationships of the superoxide dismutase gene families in
angiosperms and their relationship to other eukaryotic and prokaryotic
superoxide dismutases. Arch. Biochem. Biophys. 399, 19–36.
[60] Bueno, P., Varela, J., Giménez-Gallego, G. and del Río, L.A. (1995) Peroxisomal
copper, zing superoxide dismutase - characterization of the isoenzyme from
watermelon cotyledons. Plant. Physiol. 108, 1151–1160.
[61] McCord, J.M. and Fridovich, I. (1969) Superoxide dismutase an enzymic
function for erythrocuprein (hemocuprein). J. Biol. Chem. 244, 6049–
6055.
[62] Antonyuk, S.V., Strange, R.W., Marklund, S.L. and Hasnain, S.S. (2009) The
structure of human extracellular copper-zinc superoxide dismutase at 1.7 Å
resolution: insights into heparin and collagen binding. J. Mol. Biol. 388, 310–
326.
[63] Hong, Z., Lo Verde, P.T., Thakur, A., Hammarskjold, M.L. and Rekosh, D. (1993)
Schistosoma mansoni: a Cu/Zn superoxide dismutase is glycosylated when
expressed in mammalian cells and localizes to a subtegumental region in
adult schistosomes. Exp. Parasitol. 76, 101–114.
[64] Chang, L.Y., Slot, J.W., Geuze, H.J. and Crapo, J.D. (1988) Molecular
immunocytochemistry of the CuZn superoxide dismutase in rat
hepatocytes. J. Cell. Biol. 107, 2169–2179.
[65] Weisinger, R.A. and Fridovich, I. (1973) Mitochondrial superoxide dismutase.
Site of synthesis and intramitochondiral localization. J. Biol. Chem. 248,
4793–4796.
[66] Kira, Y., Sato, E.F. and Inoue, M. (2002) Association of Cu, Zn-type superoxide
dismutase with mitochondria and peroxisomes. Arch. Biochem. Biophys. 399,
96–102.
[67] Keller, G.-A., Warner, T.G., Steimer, K.S. and Hallewell, R.A. (1991) Cu, Zn
superoxide dismutase is a peroxisomal enzyme in human firboblasts and
hepatoma cells. Proc. Natl. Acad. Sci. USA 88, 7381–7385.
[68] Geller, B.L. and Winge, D.R. (1982) Rat liver Cu, Zn-superoxide dismutase. J.
Biol. Chem. 257, 8945–8952.
[69] Bordo, D., Djinovic, K. and Bolognesi, M. (1994) Conserved patterns in the Cu,
Zn superoxide dismutase family. J. Mol. Biol. 238, 366–386.
[70] Tomalski, M.D., Eldreidge, R. and Miller, L.K. (1991) A baculovirus homolog of
a Cu/Zn superoxide dismutase gene. Virology 184, 149–161.
[71] Dos Santos, W.G., Pacheco, I., Liu, M.Y., Teixeira, M., Xavier, A.V. and LeGall, J.
(2000) Purification and characterization of an iron superoxide dismutase and
a catalase from the sulfate-reducing bacterium Desulfovibrio gigas. J.
Bacteriol. 182, 769–804.
[72] Regelsberger, G., Atzenhofer, W., Ruker, F., Peschek, G.A., Jakopitsch, C.,
Paumann, M., Furtmuller, P.G. and Obinger, C. (2002) Biochemical
characterization of a membrane-bound manganese-containing superoxide
dismutase from the cyanobacterium Anabaena PCC 7120. J. Biol. Chem. 277,
43615–43622.
[73] Li, T., Huang, X., Zhou, R.B., Liu, Y.F., Li, B., Nomura, C. and Zhao, J.D. (2002)
Differential expression, localization of Mn, Fe superoxide dismutases in the
heterocystous cyanobacterium Anabaena sp. strain PCC 7120. J. Bact. 184,
5096–5103.
[74] Esposito, L., Seydel, A., Aiello, R., Sorrentino, G., Cendron, L., Zanotti, G. and
Zagari, A. (2008) The crystal structure of the superoxide dismutase from
Helicobacter pylori reveals a structured C-terminal extension. Biochim.
Biophys. Acta 1784, 1601–1606.
[75] Amo, T., Atomi, H. and Imanaka, T. (2003) Biochemical properties and
regulated gene expression of the superoxide dismutase from the facultatively
aerobic hyperthermophile Pyrobaculum calidifontis. J. Bacteriol. 185, 6340–
6347.
[76] Atzenhofer, W., Regelsberger, G., Jacob, U., Peschek, G.A., Furtmuller, P.,
Huber, R. and Obinger, C. (2002) The 2.0 Å resolution structure of the
catalytic portion of a cyanobacterial membrane-bound manganese
superoxide dismutase. J. Mol. Biol. 321, 479–489.
 A.-F. Miller / FEBS Letters 586 (2012) 585–595
[77] Nakamura, T. and Uegaki, K. (2011) Crystal structure of the cambialistic
superoxide dismutase from Aeropyrum pernix K1 - insights into the enzyme
mechanism and stability. FEBS J. 278, 598–609.
[78] Yamano, S. and Maruyama, T. (1999) An azide-insensitive superoxide
dismutase from a hyperthermophilic archaeon, Sulfolobus solfataricus. J.
Biochem. 125, 186–193.
[79] Kardinahl, S., Anemüller, S. and Schäfer, G. (2000) The hyperthermostable Fe-
superoxide dismutase from the archaeon Acidianus ambivalens:
characterization, recombinant expression, crystallization and effects of
metal exchange. J. Biol. Chem. 381, 1089–1101.
[80] Kirby, T.W., Lancaster, J.R. and Fridovich, I. (1981) Isolation and
characterization of the iron-containing superoxide dismutase of
Methanobacterium bryantii. Arch. Biochem. Biophys. 210, 140–148.
[81] Takao, M., Yasui, A. and Oikawa, A. (1991) Unique characteristics of
superoxide dismutase of a strictly anaerobic archaebacterium
Meinanobacterium thermoautotrophicum. J. Biol. Chem. 266, 14151–14154.
[82] May, B.P. and Dennis, P.P. (1989) Evolution and regulation of the gene
encoding superoxide dismutase from the archaebacterium Halobacterium
cutirubrum. J. Biol. Chem. 264, 12253–12258.
[83] Loftus, B. et al. (2005) The genome of the protist parasite Entamoeba
histolytica. Nature 433, 865–868.
[84] Bécuwe, P. et al. (1996) Characterization of iron-dependent endogenous
superoxide dismutase of Plasmodium falciparum. Mol. Biochem. Parasitol.
76, 125–134.
[85] Asojo, O.A., Schott, E.J., Vasta, G.R. and Silva, A.M. (2006) Structures of
PmSOD1 and PmSOD2, two superoxide dismutases from the protozoan
parasite Perkinsus marinus. Acta Crystallogr. Sect. F 62, 1072–1075.
[86] Sienkiewicz, N. et al. (2004) Identification of a mitochondrial superoxide
dismutase with an unusual targeting sequence in Plasmodium falciparum.
Mol. Biochem. Parasitol. 137, 121–132.
[87] Kanematsu, S. and Asada, K. (1979) Arch. Biochem. Biophys. 195, 535–545.
[88] Narasipura, S.D., Chaturvedi, V. and Chaturvedi, S. (2005) Characterization of
Cryptococcus neoformans variety fattii SOD2 reveals distinct roles of the two
superoxide dismutases in fungal biology and virulence. Mol. Microbiol. 55,
1782–1800.
[89] Van Camp, W., Bowler, C., Villarroel, R., Tsang, E.W.T., Van Montagu, M. and
Inze, D. (1990) Characterization of iron superoxide dismutase cDNAs from
plants obtained by genetic complementation in Escherichia coli. Proc. Natl.
Acad. Sci. USA 87, 9903–9907.
[90] Moran, J.F., James, E.K., Rubio, M.C., Sarath, G., Klucas, R.V. and Becana, M.
(2003) Functional characterization and expression of a cytosolic iron-
595
superoxide dismutase from cowpeak root nodules. Plant. Physiol. 133, 773–
782.
[91] Rubio, M.C., Becana, M., Sato, S., James, E.K., Tabata, S. and Spaink, H.P. (2007)
Characterization and genomic clones and expression analysis of the three
types of superoxide dismutases during nodule development in Lotus
japonicus. Molecular Plant-Microbe Interact. 20, 262–275.
[92] del Río, L.A., Sandalio, L.M., Altomare, D.A. and Zilinskas, B.A. (2003)
Mitochondrial and peroxisomal manganese superoxide dismutase:
differential expression during leaf senescence. J. Exp. Bot. 54, 923–933.
[93] Brouwer, M., Brouwer, T.H., Grater, W. and Brown-Peterson, N. (2003)
Replacement of a cytosolic copper/zinc superoxide dismutase by a novel
cytosolic manganese superoxide dismutase in curstaceans that use copper
(haemocyanin) for oxygen transport. Biochem. J. 374, 219–228.
[94] Holley, A.K., Dhar, S.K. and St Clair, D.K. (2010) Manganese superoxide
dismutase vs P53 regulation of mitochondrial ROS. Mitochondrion 10, 649–
661.
[95] Lisanti, M.P., Martinez-Outschoorn, U.e., Lin, Z., Pavlides, S., Whitaker-
Menezes, D., Pestell, R.G., Howell, A. and Sotgia, F. (2011) Hydrogen
peroxide fuels aging, inflammation, cancer metabolism and metastasis: the
seed and teh soil also needs ‘fertilizer’. Cell Cycle 10, 2440–2449.
[96] Quiros-Gonzalez, I., Sainz, R.M., Hevia, D. and Mayo, J.C. (2011) MnSOD drives
neuroendocrine differentiation, androgen independence, and cell survival in
prostate cancer cells. Free Radic. Biol. Med. 50, 525–536.
[97] Mohr, A., Buneker, C., Gough, R.p. and Zwacka, R.M. (2008) MnSOD protects
colorectal cancer cells from TRAIL-induced apoptosis by inhibition of Smac/
DIABLO release. Oncogene 27, 763–774.
[98] Martin, W., Hoffmeisster, M., Rotte, C. and Henze, K. (2001) An overview of
endosymbiotic models for the origins of eukaryotes, their ATP-producing
organelles (mitochondria and hydrogenosomes), and their heterotrophic
lifestyle. Biol. Chem. 382, 1521–1539.
[99] Poole, A.M. and Neumann, N. (2011) Reconciling an archaeal origin of
eukaryotes with engulfment: a biologically plausible update of the Eocyte
hypothesis. Res. Microbiol. 162, 71–76.
[100] Gray, M.W., Burger, G. and Lang, B.F. (1999) Mitochondrial evolution. Science
283, 1476–1481.
[101] Dufernez, F., Yernaux, C., Gerbod, D., Noël, C., Chauvenet, M., Wintjens, R.,
Edgcomb, V.P., Capron, M., Opperdoes, F.R. and Viscogliosi, E. (2006) The
presence of four iron-containing superoxide dismutase isozymes in
Trypanosomatidae: Characterization, subcellular localization, and
phylogenetic origin in Trypanosoma brucei. Free Radic. Biol. Med. 40, 210–225.