4649
Inorg. Chem. 1992, 31, 4649-4655
Redox Chemistry of Superoxide Dismutase. Cyclic Voltammetry of Wild-Type Enzymes and Mutants on Functionally Relevant Residues H. A. Azab,t*ll L. Banci,s M. Borsari,l C. Luchinat,*v+M. Sola,$ and M. S. Viezzolit Department of Chemistry, University of Florence, Via G. Capponi 7, 50121 Florence, Italy, Institute of Agricultural Chemistry, University of Bologna, viale Berti Pichat 10, 401 27 Bologna, Italy, and Department of Chemistry, University of Modena, Via Campi 183, 41 100 Modena, Italy Received March 2, 1992
The reduction potential of human cuprozinc superoxide dismutase and of several of its functionally relevant mutants have been measured through cyclic voltammetry. The reduction potential of the bovine enzyme has been also measured and compared with literature values. The human enzyme has a slightly higher redox potential than the bovine isoenzyme ( E o = 0.36 f 0.01 and 0.32 f 0.01 V vs NHE, a t pH 7.4, respectively). The redox properties of the bovine copper+obalt derivative are very similar to those of the native protein. The pH dependence of the Eo value in the wild-type enzyme and its pH independence in the Asn-124 mutant, which has an empty zinc binding site over the entire pH range, is ascribed to the uptake of a proton by His-63 upon reduction. A pK, value of 10.8 for this group is obtained from IH N M R titrations. It is proposed that also in the zinc-deprived derivative the copper-His-63 bond is broken upon reduction. Sizably negative reduction potentials were estimated for CN- and N3--inhibited enzymes. The values are below the reduction potential of dioxygen to superoxide.
Introduction Copper(I1)-zinc(I1) superoxide dismutases (SOD)are metalloenzymes formed by two identical subunits, each containing one copper(I1) and one zinc(I1) i0n.l-3 Copper and zinc are connected by a histidine imidazolate bridge. The chemical, structural, and catalytical properties of SOD have been extensively characterized through a variety of spectroscopic techniques4-10 and X-ray studies3,Il as well as through substitution of relevant amino acid residues by site-directed mutagenesis.I2-l8 The role of SOD seems to be that of catalyzing the dismutation of superoxide anion, strongly cytotoxic, to dioxygen and hydrogen University of Bologna. University of Florence. University of Modena. 11 Permanent address: Assiut University, Assiut, Egypt. (1) (a) Fridovich, I. Adu. Enzymol. 1974,41,35-97. (b) Fridovich, I. Ado. t 8
Enzymol. Relat. Areas Mol. Biol. 1986, 58, 61-97. (2) Fee, J. A. In Metal Ions in Biological Systems; Sigel, H., Ed.; Marcel Dekker: New York, 1981; Vol. 13, pp 259-298. (3) Tainer, J. A.;Getzoff, E. D.; Beem, K. M.; Richardson, J. S.; Richardson, D. C. J. Mol. Biol. 1982, 160, 181-217. (4) Valentine, J. S.; Pantoliano, M. W. In Copper Proteins; Spiro, T. G., Ed.; Wiley: New York, 1981; Vol. 3, pp 291-358. ( 5 ) Pantoliano, M. W.; McDonnell, P. J.; Valentine, J. S.J. A m . Chem. Soc. 1979, 101, 6454-6456. (6) Lippard,S. J.; Burger, A. R.; Ugurbil, K.;Pantoliano, M. W.; Valentine, J. S.Biochemisfry 1977, 16, 1136-1141. (7) Ming, L. J.; Valentine, J. S. J . A m . Chem. SOC.1987, 109,2426-2428. (8) Ming, L. J.; Banci, L.; Luchinat, C.; Bertini, I.; Valentine, J. S. Inorg. Chem. 1988, 27, 4458-4463. (9) Pantoliano, M. W.; Valentine, J. S.; Nafie, L. A. J. Am. Chem. SOC. 1982. 104. 63104317. (10) Berzni, 1.i Banci, L.; Piccioli, M.; Luchinat, C. Coord. Chem. Reu. 1990, 100. 67-103. (1 I ) Tainer, J.A.;Getzoff, E. D.;Richardson, J.S.; Richardson,D. C. Nature (London) 1983, 306, 284-287. (12) Beyer, W. F.; Fridovich, I.; Mullenbach, G. T.; Hallewell, R. A. J. Biol. Chem. 1987, 23, 1 1 182-1 1187. (13) Banci, L.; Bertini, I.; Luchinat, C.; Hallewell, R. A. J . A m . Chem. SOC. 1988, 110, 3629-3633. (14) Bertini, I.; Banci, L.; Luchinat, C.; Bielski, B. H. J.; Cabelli, D. E.; Mullenbach. G. T.: Hallewell. R. A. J. A m . Chem. SOC.1989, 111, 714719. (15) Banci. L.; Bertini, I.; Luchinat, C.; Hallewell, R. A. Ann. N . Y . Acad. Sei. 1988. 542, 37-52. (16) Banci. L.; Bertini, 1.; Cabelli, D. E.; Hallewell, R. A,; Luchinat, C.; Viezzoli, M. S . Free Radical Res. Commun. 1991, 12-13, 239-251. (17) Banci,L.;Bertini,I.;Cabelli, D. E.;Hallewell,R. A.;Tung, J. V.;Viezzoli, M. S.Eur. J. Biochem. 1991, 196, 123-128.
0020-166919211331-4649$03.00/0
p e r o ~ i d e . ~
400
E
I
W
350
300
250
200
5
6
9
B
7
10
11
pn
Figure 6. pH dependence of the redox potential for HSOD-AS Cu2E2Asn-124 (v),and Cu2E2 HSOD-AS (0). 6 6
.
.
.
l
,
.
~
.
.
,
,
.
,
,
,
.
.
.
.
,
l
.
,
.
,
.
.
.
,
,
,
,
.
.
,
l
(m),
8 7 (8 21
l
t
\
this order of magnitude between voltammetric and potentiometric values of the reduction potential of metalloproteins are largely documented40and arise primarily from the intrinsically different mechanism of the electron exchange which occurs in heterogeneous and homogeneous phases, respectively. Whichever comparison would be purely speculative. Different Eo values were also previously obtained from potentiometric titrations with MVZ+ (+0.33 and +0.26 V at pH 7.4) depending on the protein p r e p a r a t i ~ n .Again, ~ ~ values of +0.40 and +0.25 V wereobtained for BESOD a t pH 7.4 from titrations with KsFe(CN)b and K2IrC16, r e ~ p e c t i v e l y . ~It~is, ~apparent ~ that the potential value is affected by the nature of the mediator, which may possibly influence the stability of the protein. Moreover, evidence was obtained that the above anions act not only as simpleredox partners but also interact closely with the copper centers, though the perturbation of the copper site by ferricyanide has been recently e~claded.2~ Our data appear to be obtained on a protein neither perturbed by the promoter nor affected by adsorption phenomena onto the electrode. We have no elements for claiming our potential value more reliable than the others as the absolute value for the reduction potential of SOD. However, we stress again that, within the experimental error, the variations within our sets of data with pH, inhibitor binding to the protein, and mutations in the active site are real and show an inherent consistency with known properties of the enzyme. From the present data no evidence is obtained for a previously proposed inequivalence of the two protein subunits:29 more precisely, the shape of the voltammetric signal sets the upper value for the difference in the reduction potentials, if any, at about 0.03 V. While the reduction potential of all the species investigated is found pH-dependent due to the involvement of one proton in the redox equilibrium, with a low-pH Eo value of about 0.45-0.50 V and a high-pH limit value of 0.24-0.25 V, respectively, the enzymatic activity of SOD is known to be pH-independent up to about pH 10, while with a further increase in pH it decreases with an approximately unit slope in log units.15J6 Hence, the kinetics of SOD-catalyzed superoxide disproportionation is not affected at all by the relative thermodynamic stability of the redox states of the copper ion around neutral pH. Instead, it depends on the protonation state of a group whose pKa is close to that regulating the redox potential. The pKavalues of 8.9,8.8, and 8.2 determined here for HSOD, HSOD-AS, and BESOD, respectively, must be due to a group present in the reduced form and absent in the oxidized form. Most probably, this pKa is relative to the deprotonation of the Zn-bound imidazole group of His-63 of the reduced form,4J1q29-33 according to the following equilibria:
n
i o w p ~ :Z n - N g N - C u ( I 1 )
J
+ e- + H+
n
+ Cu(1)
Zn-NON-H
v
n
high pH: Zn-N"-cu(lI)
r 6
t
5
,
.
.
,
l
6
.
.
.
.
7
,
.
.
.
.
8
.
.
p6,
,. 1 . . . , , . . , , 1 . , . , , . . ,
9
io
!!
:2
v
+
e-
n
Zn-NGN
+
Cu(1)
13
DH
Figure 7. Plot of the IH NMR chemical shift (600 MHz, 27 "C)versus pH of the C-2 proton of the five metal binding histidines in CulZnllSOD: ( 0 )His-63; (A)His-80; (0)His-48; (V)His-71; ( 0 )His-120. The scale refers to His-63, while right scales are relative to His-80, to His-48 and His-120, and to His-71 (values in parentheses).
investigated, in line with the close similarity of their chemical behaviors, while the bovine enzyme invariably shows lower potential values. The value obtained for BESOD a t pH 7 (0.330 f 0.010 V) islower than that recently determined throughvisible spectroelectrochemistry (0.403 f 0.005 V).27 Discrepancies of
The case of the Asn- 124 mutant is thus particularly interesting as far as the observed pKa is concerned. In this species the zinc site is empty and, unlike the C U ~ derivative E~ of the wild-type protein, the Cuz+ ion does not migrate in this site at al!raline pH. The Eo values are pH independent and leveled at the wild-type high-pH value. Apparently, the absence of zinc abolishes the group undergoing deprotonation in the wild-type reduced protein. This is a strong evidence in favor of His-63 as the candidate group. In CuzE2 SOD His-63 is either coordinated to copper as the other three histidines or noncoordinated, given the preference of Cu(1) for low coordination numbers. In the first hypothesis
Redox Chemistry of Superoxide Dismutase His-63 would be a normal ligand and therefore expected to have a lower pKa whencoordinated toCu(I1) than toCu(1). Therefore, we would predict a pH-independent Eo value up to the pKa of the oxidized form, after which the Eo values should decrease with pH. This is obviouslynot the case. In the second hypothesis His-63 would be a free residue, with a pKa around 5-6 for the cationic to neutral form and then a pKa around 14 for the neutral to anionic form. Therefore a pH independence of Eo is expected over the whole range from 5-6 to 14. This behavior is consistent with the experimental data. Accordingly, we propose that the structure of Cu(1) in Cu2E2SOD is the same as that in CUIZn2SOD. The residual pH dependence of Eo observed for the wild-type Cu2Ez-derivative above pH 7 is most probably linked to the migration of the Cu2+ ion to the Zn2+ site that gives rise to exchanging species with a different site occupancy, each characterized by an individual reduction potential. Overall, the data on the C U ~ Ederivatives Z strongly support the assignment of the pKa around 8-9 observed in the reduction potential of the holoenzyme to the ionization of the zinc-coordinated His-63 in the reduced protein. NMR titration experiments, coupled with the full assignment of the active site protons recently perf~rmed,~' have allowed us to determine the actual value of the microconstant for the deprotonation of His-63 in the reduced form. The determined value of 10.8 is sizably higher than that measured electrochemically, but still within the possible range of deviations due to the electrode-dependent phenomena mentioned above. It is worth noting that the pKa of 10.8 measured here for the deprotonationof His-63 in the reduced form is in the same ballpark as the pKa measured for the loss of activity at high pH. Further experiments are, however, needed to establish whether this coincidence is just fortuitous or if indeed the enzyme requires a protonated His-63 to be active. The redox potential for all the mutants investigated, bearing substitutions on most of the relevant residues in the active cavity resulted to be, even if to a different extent, invariably lower than those of the wild-type protein over all the pH range investigated. Such a decrease of Eo values for mutants having a Gln residue instead of a Glu residue in position 132 and/or position 133, and an Ala or Gln residue instead of Lys in position 136, can hardly be interpreted on simple electrostatic bases, since the above replacements cause opposite variation of charge in the proximity ofthecopper ion. MutantsonThr-137 (Figure4) show Eo values very close to that of the wild-type enzyme: substitutions with Ala and Ser scarcely affect the potential (see Table I), especially in
Inorganic Chemistry, Vol. 31,
NG.22, 1992 4655
the latter case, where the functionally relevant OH group is maintained. It should be noted that substitution of the positively charged Lys- 136 residue with neutral residues maintain the pH dependenceofthereduction potential. SinceLys-136is theclosest acid-base group to copper (besides His-63) (Figure 4), this is a further support for His-63 being the group responsible for the observed pH dependence. The pattern of the redox potential for the mutants does not correlate with their activities. This finding is not surprising since the Eo values for the mutants remain well within the limits defined by the Eo values of the two catalyzed semireactions. The overall enzymatic catalysis is the result of many factors that can affect to a different extent the dismutation rate. The metal derivative of BESOD with Co2+ inserted in the Zn2+ site has been deeply investigated through *HNMR spectroscopy and proved to be extremely informative on the structural features of the metal sites and on the binding mode of a number of inhibitors. That this information can be confidently transferred to the native enzyme is further demonstrated by the small changes of the redox potential of SOD induced by such a metal substitution. The slightly higher Eo values for the cobalt derivative, if significant, can be tentatively attributed to a shift of the His-63 pK, toward higher values as the result of a lower binding strength toward the cobalt ion as compared to the zinc ion. The dramatic changes of Eo upon binding of azide and cyanide clearly indicate that in both cases the anion binds to the copper ion, stabilizing the oxidized state. The extent of the potential lowering is such that in both cases the bound SOD would be unable to disproportionatethe superoxideanion also from a purely thermodynamic point of view as its redox potential is no longer intermediate with those of the two pairs involved in the 02dismutation.
Acknowledgment. Thanks are expressed to Professor I. Bertini of the University of Florence for stimulating the present research, for the many subsequent discussions, and for critical reading of the manuscript. Thanks are also expressed to Prof. G. Battistuui Gavioli of the University of Modena for competent and valuable advice on the electrochemical aspects of the work. We gratefully acknowledge the long-standing collaboration with Dr. R. A. Hallewell and Chiron Corp., Emeryville,CA, which made possible the obtainment of the mutants investigated in this work. Finally, fellowship support for H.A.A. from the 'Third World Academy of Science in Italian Laboratories" (Trieste) is gratefully acknowledged.
