It makes better biological sense to interrupt oxidative damage at the level of O2than further downstream, because scavenging O2
depresses levels of all the reactive progeny simultan- eously. Superoxide disproportionates rapidly at pHs approaching its pK of 4.8, but at more common physiological pHs, spontaneous disproportion is slower (k2105s1 at neutral pH). Therefore, in order to suppress O2to less than the lethal concentration of1010M, in the face of the estimated rate of O2 formation of 5mM s154, superoxide dismutase (SOD) and/or superoxide reductase (SOR) is probably present at50mM in cells.55This may reflect the relatively high KMs of SOR and SOD (100mM, see below), which would decrease their efficiencies despite their very high second order rate constants ofk109M1s1. A number of parasites and pathogens export SOD. This is believed to contribute to virulence, presumably because the exported SOD defends the pathogen against the burst of O2
produced by phagocytic leukocytes.56
Superoxide can, in principle, be eliminated by oxidation, reduction, or disproportionation. The latter has the advantage of consuming two molecules of the offending substrate per reaction, but produces H2O2as well as O2, consistent with its dominance in aerobes (summarized inTable 1).
O2
oxidation to O2does not appear to be a prevalent strategy, possibly because the intracellular environment tends to be fairly reducing (ambient potentials 230 mV for oxygen tension corresponding to 10% air-saturated57) so that most centers with anEmsufficiently high to oxidize O2 are normally reducedin vivo. Superoxide reduction is employed by strict anaerobes, either because the O2resulting from dismutation is toxic, or because the cell interior is so reducing that any enzyme oxidized by O2 is rapidly re-reduced and would be subject to evolutionary optimization only for its ability to reduce O2. SODs constitute robust stand-alone defenses, since they do not rely on any additional electron donor or acceptor, but utilize O2
as both and consume two molecules of O2per turnover. The H2O2they produce is normally consumed by catalases and peroxidases. Indeed, whereas moderate over-expression of SOD tends to be bene- ficial,58there are instances in which high levels of overexpression of SOD alone were found to be deleterious.59This may be because overexpressed SOD suppresses the O2
concentration to the point that O2fails to activate the SoxR/S regulon and elicit overexpression of the several other enzymes besides SOD involved in minimizing oxidative damage.59,60
SODs have evolved on several independent occasions with the result that Cu,Zn–SOD is a completely different enzyme from Ni–SOD, and both are completely different from the Fe– and Mn–SODs (compareFigures 3and5). In contrast, Fe–SODs and Mn–SODs are strongly homo- logous to one another.61The different SODs are found in different organisms and associated with different cellular compartments in which O2
is generated (Table 1). Notwithstanding its final destination, mitochondrial Mn–SOD is encoded by a nuclear gene.
Table1Comparisonofthedifferenttypesofsuperoxidedismutasesandsuperoxidereductase. TypeofSOD orSORMetalcontent andligationLocationOligomerization* monomersizekcata (s1 )KM
Secondorder rateconstant (M1 s1 )Inhibition Cu,Zn–SODCuI/II (His4), ZnII (His3,Asp )Eukaryoticcytosol,peroxisomes, mitochondrialintermembrane space.PeriplasmofGramve bacteria,plantplastids, extracellularinmammals.
(1,2,4) *16kD0.910698 3.5mMat pH9.3 and5.5o C98
2109 atpH7 and25C80
pH<5,highpH pK10.7,CN N3 ,NCO ,H Fe–SODFeII/III (His3Asp , H2O/OH )
Prokaryotes,occasionallyin cytoplasmofsomeprotists, chloroplasts.
(1,2) *22kD2.6104120 80Mat pH8.4 and25o C120
5108 at pH8,25C119pH5,highpH pK9,H2O2, N3 ,F120 Mn–SODMnII/III (His3Asp, H2O/OH )
Mitochondria,peroxisomesof eukaryotes.(2,4) *22kD1.25–4104147,148 40–50M atpHẳ9.3 and2C147
5.6108 atpH8.9, 25C189
highpH pK9.8,N3,F Ni–SODNiII ,NiII/III (Cys4,His2)Fungi.4*13kD1.3109 at pH7167CN ,H2O2165 SORFeII/III (His4,Cys, (Glu))Strictanaerobes,archae.4*14kD1.5109 atpH7.8, 25C186
pH<5.6,high pHpK9.5, N3 ,CN akcatvaluesarereportedforthesameconditionsasthosepertainingtoKM.
8.19.6 Cu,Zn-SOD CONTAINING SUPEROXIDE DISMUTASE
Cu,Zn-containing SODs (Cu,Zn–SODs) are known as the cytosolic and extracellular SODs of animals; however, they are also found in plants, animals, fungi and bacteria, and are encoded in several viral genomes.25,62 In animals, Cu,Zn–SOD is also found in the nucleus, the mitochondrial intermembrane space, and lysosomes.63 In plants, distinct Cu,Zn–SODs are found in the chloroplast and cytoplasm.62The prokaryotic Cu,Zn–SODs as a group are distinct from the eukaryotic Cu,Zn–SODs.63Recent reviews of Cu,Zn–SOD include those of Bordo and Bertini.63,64
8.19.6.1 Structure
As their name implies, Cu,Zn–SODs contain a CuI/II and a ZnII ion in each active site. The enzyme is most commonly a dimer of 16 kD monomers, although the human extracellular version is a homotetramer65and a few prokaryotic enzymes are monomers.63Monomeric versions of the human cytosolic enzyme have also been engineered.66 In all structurally characterized cases, the monomers are single antiparallel-barrel domains (Figure 3).63,67The active site is located on the exterior of the barrel at the base of a channel formed predominantly by loops. The Zn is completely occluded from solvent but the Cu has slight (10 A˚2) solvent exposure.
In the oxidized state, the active site CuII ion is coordinated by four histidines: His44 (N1), His46, His61, and His118 (N"2, numbering of bovine Cu,Zn–SOD is used throughout) in a geometry intermediate between square–planar and trigonal–bipyramidal (Figure 4).68–70 His61 also coordinates the ZnIIion via N1. The Zn ion is roughly tetrahedral and also coordinated by His69 and His78 (N1) and Asp81.63,68–70 Additional crucial residues include a buried Asp122 that hydrogen bonds to the His44 ligand of Cu and the His69 ligand of Zn, stabilizing the binuclear cluster, an Arg which is partially exposed to solvent near Cu, and a Cys whose participation in an SS bond is held to be essential for correct active site structure and metal center maturation.63 One molecule of solvent is coordinated to CuII based on NMR71 and is visible in many crystal structures in the axial position, adjacent to Arg141.68 However, some of the crystal structures purporting to describe the oxidized state may reflect varying contributions from reduced sites generated by X-ray photoreduction.72
The metal ions can be removed by dialysis against EDTA and reconstituted or replaced site- specifically.73,74 Cu, E–SOD, Cu,Zn–SOD lacking the two Zn ions but with the two Cu ions in their correct sites, retains 80% of native activity at pH 6, indicating that neither the Zn ion nor the bridging His are essential at low pH.75Similarly, Cu,Co–SOD retains full activity76and native structure,77and thus affords very valuable spectroscopic probes of the Zn site. NMR studies of the CuII,CoII–SOD active site have been used to elucidate the effects of ligand binding and
Figure 3 Ribbon of the Cu,Zn–SOD of bovine erythrocytes, based on the coordinates of Taineret al.69 (2SOD.pdb) All molecular structures are depicted using Molscript.191
mutations (reviewed by Bertiniet al.78). Optical studies of CuI,CoII–SOD were the first to indicate that the His61 bridge was broken upon Cu reduction (below).79
8.19.6.2 Mechanism and Insights from Spectroscopy
Superoxide dismutation appears to occur via two sequential reactions which are each first-order in superoxide, based on pulsed radiolysis.80 Thus, the metal ion alternately accepts and donates an electron in conjunction with proton transfer.
O2 ỵ HỵỵLCuII;ZnII!O2ỵLHCuI;ZnII ð1aị O2ỵHỵỵLHCuI;ZnII!H2O2ỵLCuII;ZnII ð1bị where L signifies the protein ligands to Cu and Zn.
The catalytic rate is limited by the rate of substrate diffusion into the active site under most conditions.81 A constellation of conserved positively charged residues is credited with produ- cing a positive electrostatic potential that ‘‘steers’’ O2 into the active site, in accordance with
His61 Asp81
His46
H2O
Cu Cu His78
Zn
His44 Zn
His61
His118
Asp122
Arg141
His69
H2O
Figure 4 Active site of Cu,Zn–SOD, based on the oxidized-state coordinates of Taineret al.69(2SOD.pdb, bovine erythrocytes) and the reduced state coordinates of Ogiharaet al.192(1JCV.pdb, from yeast). The Cu ions are shown in green and Zn in black, other atoms are colored according to cpk. The reduced state structure is depicted using heavy bonds and darkened colors including C atoms in gray, the oxidized state
structure is depicted with light bonds and light-colored atoms, including white C atoms.
inhibition of catalytic activity by high ionic strength or modification of positively charged residues,82,83 as well as calculations of Brownian dynamics.84 Mutations decreasing the nega- tive charge near the active site have produced enzymes with higher-than-native activity, including a P. leiognathi mutant Cu,Zn–SOD with kcat/KMẳ11010M1s1, the highest of
N CuII N ZnII His-N
N
N-His O-Asp–
N-His Arg
N CuI HN
ZnII N~ ~N
~N O~
N~
N CuII N ZnII N~ ~N
~N O~
N~
N CuI HN ZnII N~ ~N
~N O~
N~
O2–
O2
N N
CuII ZnII
~N
~N
~N O~
N~
O–2 N-His
~N
~N
~N His
~N HO2– H2O
H2O
O2•–
O2•–
H+, H2O His
NH3+
NH3+
H+
NH3+ NH3+
NH3+ H2O2
Scheme 1
Figure 5 Ribbon structure of Fe–SOD, based on the coordinates 1ISB.pdb of Lahet al.111The FeIIIions are depicted by black spheres.
any SOD.85 However, the effects of such mutations on residues surrounding the active site are large only at low ionic strength, whereas at physiological ionic strength, Arg141 and the metal ions themselves are thought to explain most of the enhanced substrate diffusion into the active site.86
The green color of CuII,ZnII–SOD reflects its d–d band at 14,700 cm1 (680 nm).87,88 The appearance of a UV ligand-to-metal charge-transfer band upon binding N3 or SCN,88 changes in the EPR signal induced by these and CN,89 and X-ray crystal structures show that anions coordinate to CuII in a weak axial site, displacing water.90 CN and N3
coordinate with quite different orientations, possibly due to steric interference between (larger) N3 and surrounding residues, as well as CN’s stronger ligand strength which gives it pre- eminence over the other ligands.91 However, the fact that both bind in the position at which coordinated solvent is occasionally observed suggests that this is where O2 binds too.72,92 O2 coordinated to CuII is proposed to be oxidized by inner sphere electron transfer to produce CuIþO2, with the latter escaping from the active site since its neutrality robs it of electrostatic stabilization (Scheme 1, showing proposed detailed mechanism of CuZn–SOD, based on Bordo et al.63 and Hart et al.72 Proposed hydrogen bonds are shown as dashed lines.). CuI releases His61 to become trigonal planar79,93,94 and His61 becomes protonated with a pK of 10.8 (Figure 4).95
Thus, Cu,Zn–SOD, takes up a proton.94,96 Because proton uptake accompanies reduction of the metal ion, the total charge of the active site is conserved, and the second substrate to bind in the cycle benefits from the same overall electrostatic steering as the first. The second O2 is proposed to hydrogen bond with Arg141 but not to coordinate directly to CuIbased on infrared spectroscopy.97Outer sphere electron transfer is therefore inferred, coupled to proton transfer to nascent product from His61.63Re-binding of deprotonated His61 to the now-CuIIsite would aid in product displacement and/or discourage O2H binding and restore the starting state of the active site (but see Fee and Bull98). Cu,Zn–SOD specific activity decreases at high pH with a pK of 10.7.99 This could reflect in part release of Cu, but the decrease also coincides with the pK of 10.8 of His61 in reduced Cu,Zn–SOD.95 Thus, low activity at high pH can be explained by competitive inhibition by OH binding to the oxidized state,100 in conjunction with loss of a proton required for product formation by the reduced state. The solvent isotope effect of 3.6 also indicates that a proton transfer step contributes to the rate.98
Between pH 5 and 9.5, Cu,Zn–SOD activity is only very weakly pH dependent. This has been attributed to the His61 ligand to Cu and its bridge to ZnII.101 The pK of His61 is depressed from14 to 10.8 due its coordination to ZnII, enabling it to coordinate to CuII but not CuI.102 Thus, it contributes to production of a favorableEmfor disproportionation of O2
, and couples proton transfer to electron transfer by taking up a proton upon Cu reduction. The ZnIIsite also forces His61 to bind CuIIin an equatorial position, and thus constrains most exogenous ligands including substrate to the weaker axial position. Product HO2is presumably also only relatively weakly bound and readily displaced.101
Besides its role in forestalling aging by preventing oxidative damage, Cu,Zn–SOD has wider significance to human health. Some 10% of familial amyotrophic lateral sclerosis (FALS) cases are associated with mutation in the SOD1 gene which encodes the cytoplasmic Cu,Zn–
SOD.103,104 Many different mutations have been identified, most of which are not believed to affect the active site directly.105Nonetheless, FALS victims are generally found to have lower- than-normal levels of total SOD activity.106 Since introduction of a FALS-associated mutant Cu,Zn–SOD results in onset of FALS symptoms even for mice bearing the normal Cu,Zn–
SOD gene, the mutant genes appear to have gained a new, dominant, and toxic pheno- type.107,108Many FALS-associated mutant Cu,Zn–SODs were found to be less well metallated than similarly expressed WT Cu,Zn–SOD and significantly less stable.105 Indeed, Fridovich has proposed that mutant Cu,Zn–SODs cause FALS by failing to enter the mitochondrial intermembrane space, embroiling heat shock proteins needed for other functions, and forming aggregates.109
8.19.7 Fe-CONTAINING SUPEROXIDE DISMUTASE
Fe-containing superoxide dismutase (Fe–SOD) is believed to be a relatively primitive SOD. It is constitutively expressed in many bacteria, and a few eukaryotes.25 Fe–SOD has been the subject of a few relatively recent reviews.110
8.19.7.1 Structure
Fe–SODs are dimers of 22 kDa monomers (Figure 5) with a few monomeric exceptions.25 Each subunit contains a single mononuclear Fe site in which Fe is coordinated by three His and one Asp residue as well as a molecule of solvent (Figure 6).111The solvent molecule is believed to be OHin (oxidized) FeIII–SOD and H2O in (reduced) FeII–SOD.112 The Fe is trigonal–bipyramidal with coordinated solvent at one apex and His26 at the other. The ligating amino acids derive from both of the two domains of a given subunit, so that the metal center is effectively bound between the two (Figure 5). A His ligand in one active site (His160) donates a hydrogen bond to Glu159 from the other subunit, so that the two active sites are linked, and the electrostatic charge of the metal ion is neutralized over all by the coordinated solvent, the Asp ligand and this second sphere Glu (numbering ofE. coliFe–SOD is used throughout).112,113The active site also contains a universally conserved Tyr residue (Tyr34) which hydrogen bonds to a conserved Gln (Gln69) which in turn hydrogen bonds to the coordinated solvent (Figure 6). Although the coordinated solvent also hydrogen bonds to the ligand Asp, Gln69 is its only link to the protein outside the active site.
All of the active site residues named above are enclosed in a shell of conserved aromatic side chains which have been proposed to insulate the active site from solvent and the surrounding protein.114 Note, however, that they may also provide important insurance against irreversible inactivation of the enzyme by its toxic substrate and toxic product, both of which can be activated by Fe. These and the one-electron nature of the chemistry catalyzed by Fe–SOD put the active site amino acids at risk of oxidation and covalent modification by catalytic intermediates and by-products. However the radicals of the His, Tyr, Trp, and Phe residues that characterize SOD are generally more table than those of the non-polar aliphatic hydrophobic amino acids found in the interiors of typical proteins, and therefore may persist long enough for the next molecule of substrate to enter the active site and re-reduce them, rescuing the active site. Indeed, Trp and His side chains get modified in the course of inactivation by H2O2,115and mutation of Tyr34 to the higher-Emresidue Phe increases Fe-substituted Mn–SOD’s susceptibility to inactivation.118
His73
Asn72
Asp156
His26
Trp122
Tyr34 Gln69 Gln69
Fe HO–
His160
Figure 6 Active site of FeIII–SOD, based on the coordinates of 1ISB.pdb Lahet al.111The FeIIIis in black and other atoms are colored according to cpk. Dashed lines indicate proposed hydrogen bonds.
8.19.7.2 Mechanism and Insights from Spectroscopy
Pulsed radiolysis studies indicate that the reaction proceeds via two sequential steps, both first order in O2
, in which O2
is alternately oxidized to O2or reduced to H2O2by the Fe ion.119 Thus, the Fe cycles between the III and II states. FeIII–SOD takes up a proton upon reduction throughout the pH range of activity,120 and chemical precedent112,121 as well as NMR studies (Milleret al.)193indicate that the proton is taken up by coordinated OHin the oxidized state to yield coordinated H2O in the reduced state. Thus the reactions may be written:
O2ỵHỵỵLFeIIIðOHị ! O2ỵLFeIIðOH2ị ð2aị O2ỵHỵỵLFeIIðOH2ị ! H2O2ỵLFeIIIðOHị ð2bị where L stands for the SOD protein, and the coordinated solvent molecule that couples proton transfer to electron transfer is in parentheses (Scheme 2, showing proposed detailed mechanism of Fe–SOD.).
The reaction proceeds with saturation kinetics withkcatẳ2.6104s1and aKMof 80mM for O2 at pH 8.4C and 25C,120 consistent with the observed second order rate constant for consumption of O2ofkẳ3108M1s1.122Activity decreases at high pH with a pK of9120
FeIII HO– His-N
His-N O-Asp– N-His
H H N Tyr34-O
H O
O2
FeIII HO–
~N
~N O-C~
N~
H H N
~O
H O
O2–
FeII
~NHOH
~N O-C~
N~
H H N
~O
H O
O2
FeII
~N HOH
~N O-C~
N~
NH H
~O
H O
FeIII –OH
~N
~N O-C~
N~
N H
~O
HO2– O H
O2
O2– H
H+ H+
Gln69
H2O2
•–
•–
–
Scheme 2
that affectsKMbut notkcat.122The rate-limiting step is nonetheless indicated to involve a proton based on a solvent isotope effect of 3 forkcatand because general acids accelerate the reaction by increasing kcat.120
Fe is high-spin in both oxidation states, with and without substrate analogs bound.123,124 Changes in the EPR signal and the appearance of a strong ligand-to-metal charge transfer band demonstrate that the competitive inhibitor N3 coordinates directly to FeIII, as do F and OH.120,123(Other substrate analogs such as SCN and ClO4
inhibit via an outer sphere anion binding site.120) Superoxide is believed to coordinate directly to FeIII in the equatorial plane, trans to the Asp ligand, to generate octahedral geometry111 and oxidation of O2
is
believed to occur via an inner sphere mechanism (Scheme 2). Considering that the coordination of FeIIIis oriented with the dominant (z)-axis defined by the axial ligand OH, the valence orbital is largely dxz, which would overlap favorably with a *-orbital of incoming O2,125 providing a plausible mechanism for inner sphere electron transfer from an end-on bound O2
(Scheme 3, showing proposed FeIII–SODdzzcomponent of the redox-active orbital and its interaction with incoming O2, based on crystal structures and DFT calculations.111,125). This is the reverse of the more familiar effective reduction of O2to O2
upon binding to FeIIin Mb and Hb, and suggests that O2does not engage in any H bonding in FeIII–SOD, since that would raise the O2/O2Em. Thus, we propose that O2
binds to FeIIIonly, and due to FeIII’s Lewis acidity, loses an electron.
HO– His-N
His-N
O-Asp–
N-His x
y z O
FeIII
Scheme 3
In the second half of the catalytic cycle, FeIIis reoxidized by O2in conjunction with proton transfer. This reaction is believed to proceed via outer-sphere electron transfer, since Fbinds to FeII–SOD but not in the inner sphere (Sorkin and Miller, unpublished),124but see ref. 194.
8.19.7.3 Protonation Steps Associated with Catalytic Activity
Reduction of O2 is unfavorable on electrostatic and electronic grounds unless it is coupled to protonation. While the protons required for production of H2O2derive ultimately from solvent,120 active site residues are most likely to mediate proton transfer to nascent product as a means of accelerating this step. The most important candidates for the roles of proton donors are the coordinated H2O (which will release a proton upon FeIIoxidation) and Tyr34 (below).
Nevertheless, the pH dependence of activity does not appear to reflect a rate-limiting proton transfer since KM but not kcat is pH dependent.120 The increase in KM at high pH can be accounted for by two active site pKs. In the oxidized state, OHbinds to FeIIIin competition with other anions and likely O2, with a pK of 8.5,120,126and the reduced state displays a pK of 8.5 attributed to deprotonation of Tyr34.127,128 Since F binding to FeII–SOD is inhibited by deprotonation of Tyr34,129it is possible that Fand O2bind via Tyr34’s OH proton, and that O2
becomes partially protonated as part of binding. The pK of Tyr34 in FeII–SOD is quite low for a Tyr, at 8.5. To the extent that Tyr34 transfers its proton to O2 as part of binding, O2
’s Em will rise and favor its acquisition of an electron from FeII. Since O2
is not coordinated to FeII, a hydrogen bond from Tyr34 to the distal O would not favor OO bond breakage and production of FeIVẳO. In contrast, in catechol dioxygenases such as 2,3- dihydroxybiphenyl 1,2-dioxygenase and cytochrome P450 which mediate OO bond cleavage, O2is believed to bind to Fe at one end and form a hydrogen bond and/or accept a proton at the other.20,130The cost of substrate’s not binding directly to the metal ion in FeII–SOD is decreased binding affinity, consistent with SOD’s relatively highKM. However, Tyr34’s phenolic O is only 5.5 A˚ from Fe. In-line geometry, with a 1.35 A˚ OObond2and a 2.7 A˚ hydrogen bond to O2
places the proximal O 1.1 A˚ from Fe. Thus, O2is not expected to point directly towards FeII, and even at a 90angle, electron transfer to the nearest O would not be rate limiting.
The rate-limiting step has been proposed to be displacement of the product and to involve a proton,120possibly transferred from coordinated solvent. It is unlikely that the proton acquired by departing H2O2is one of the two that had been part of the coordinated H2O molecule, but more likely that a proton released by coordinated solvent upon oxidation of FeII displaces another proton from the amide side chain of Gln69 towards Tyr34, which, being effectively deprotonated by O2, would draw such a relay in just this direction (Scheme 2). Thus, it is