Proc. Natl. Acad. Sci. USAVol. 82, pp. 3063-3067, May 1985Chemistry

Low-temperature magnetic circular dichroism studies of nativelaccase: Spectroscopic evidence for exogenous ligandbridging at a trinuclear copper active site

[coupled binuclear (type 3) copper/normal (type 2) copper/charge transfer transitions]

MARK D. ALLENDORF, DARLENE J. SPIRA, AND EDWARD I. SOLOMON*Department of Chemistry, Stanford University, Stanford, CA 94305

Communicated by John I. Brauman, December 21, 1984

ABSTRACT The detailed nature of N- binding at themulti-copper active site in native laccase is investigatedthrough a combination of low-temperature magnetic circulardichroism (LTMCD) and absorption spectroscopies. This com-bination of techniques allows charge-transfer spectral featuresassociated with N- binding to the paramagnetic type 2 Cu(II)to be differentiated from those associated with binding to theantiferromagnetically coupled, and therefore diamagnetic, bi-nuclear type 3 Cu(II) site. Earlier absorption titration studieshave indicated that N- binds with two different binding con-stants, yielding a high-affinity and a low-affinity form. Thestudies presented here are interpreted as strong evidence thatlow-affinity N- bridges the paramagnetic type 2 and diamag-netic type 3 binuclear Cu(ll) sites in fully oxidized laccase.This assignment is further supported by features in the MCDspectrum whose intensity correlates with an EPR signal associ-ated with uncoupled type 3 Cu(ll) sites. In these sites, N- hasdisplaced the endogenous bridge, thereby rendering the siteparamagnetic and detectable by both LTMCD and EPR spec-troscopy. High-affinity N- is found to bind to the paramagnet-ic type 2 Cu(ll) site in a limiited fraction of the protein mole-cules that contains reduced type 3 sites. Finally, the possiblerole of this trinuclear (type 2-type 3) Cu(ll) active site in en-abling the irreversible reduction of dioxygen to water is con-sidered.

The active site in Rhus vernicifera laccase contains four cop-per ions that together catalyze the four-electron reduction ofoxygen to water with concomitant oxidation of substrate (1).Based on their EPR properties, these Cu(II) centers havebeen classified as type 1 or blue (All < 90 x 10-4 cm-1), type2 or normal (All > 140 x 10-4 cm-'), and type 3 or coupledbinuclear (nondetectable by EPR). Laccase, the simplest ofthe multicopper oxidases (2, 3), contains only one of each ofthese Cu(II) types and, hence, provides the most appropriatesystem for determining how dioxygen bonding and reactivityat the oxidase active site differs from that in the hemocya-nins (4) and tyrosinase (5), which contain only a coupled bi-nuclear copper site. In the latter proteins, dioxygen bindsreversibly as peroxide and bridges the binuclear Cu(II) sitein a u-1,2 coordination geometry (6).An important simplification for study of the type 3 site has

been the preparation of a type 2 Cu(II)-depleted (T2D) lac-case form (7). As in hemocyanin, the coupled binuclearCu(II) site in fully oxidized (8) T2D laccase contains two te-tragonal Cu(II) ions that are antiferromagnetically coupledby superexchange through an endogenous protein bridge(OR-). However, whereas exogenous anions bind equator-ially and bridge the binuclear Cu(II) site in the hemocyanins

CU 2 CU +2I--, N~l- N"N

R

HEMOCYANINand

TYROSINASE

L

CU+2 CU/2

RL

LACCASE

FIG. 1. Comparison of exogenous ligand binding at coupled bi-nuclear Cu(II) sites.

and tyrosinase, exogenous ligands are found not to bridgethe type 3 site in T2D laccase (9, 10) (Fig. 1).

In comparing exogenous ligand reactivity in T2D and na-tive laccase, major differences have been observed (11).Whereas the fully reduced native enzyme is readily oxidizedby 02, the binuclear Cu(I) site in T2D laccase is stable toaerobic oxidation (8, 12). In addition, N- binds to oxidizedT2D laccase to generate a single N- -- Cu(II) charge-trans-fer (CT) transition at 450 nm (K 200 M-1, As 800M-1lcm-1), while two N- molecules have been reported (11,13) to bind to native laccase (K1 60,000 M-1 at 500 nm,Ae500 = 500 M-'-cm-', Ae410 = 630 M-l cm-l; K2 60 M-1at 400 nm, AE - 1900 M-1lcm-1). In earlier studies (11), itwas proposed that the type 2 site indirectly stabilizes ligandbinding at the type 3 site. Through low-temperature magnet-ic circular dichroism (LTMCD) spectroscopy, the nature ofthe interaction of exogenous ligands with the type 3 and type2 centers can now be clearly defined.The new absorption features observed in N- reactions

with native laccase have been associated (11, 13) with equa-torial N- -+ Cu(II) CT. These transitions can only occur atthe oxidized type 2 and type 3 centers, as the type 1 site inlaccase is reasonably considered to contain no exchangeablepositions. In order for a CT transition to be observed in theLTMCD spectrum, it should (i) have significant absorptionintensity and (ii) be associated with a magnetically degener-ate ground state. Two structural units exist in laccase fromwhich N- -* Cu(II) CT bands may originate: an isolated,mononuclear type 2 Cu(II) and a binuclear type 3 Cu(II) pair.The first has a magnetically degenerate ground state with S= 1/2 , giving rise to EPR signals in the g 2.2 region. Aligand -- Cu(II) CT band originating from this ground statewill have an associated MCD band whose intensity isproportional to 1/T, for kT >> gf3 H (C term). The secondstructural unit contains a pair of antiferromagnetically cou-pled S = 1/2 Cu(II) ions, with total spin states ST = 0,1, whichare split by the exchange coupling 2J. Since for laccase 2J >

Abbreviations: MCD, magnetic circular dichroism; LTMCD, low-temperature MCD; T2D, type 2 copper depleted; CT, charge trans-fer.*To whom correspondence should be addressed.

3063

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. NatL Acad Sci USA 82 (1985)

550 cm-' (14, 15), the S=1 state is not thermally accessible. 8000An MCD band associated with a CT transition from the S=0 ground state will be temperature independent and weakerby a factor of 100-1000 at 10 K than bands associated withparamagnetic ground states (16). Thus, through LTMCD 6000studies of laccase, one can determine whether the groundstate of the Cu(II) associated with an N- -- Cu(II) CT transi-tion is paramagnetic and hence due to type 2 Cu(II) or isdiamagnetic and thus associated with the type 3 site. These 4000LTMCD studies strongly suggest that exogenous anions di-rectly coordinate to and bridge between the type 2 and type 3Cu(II) centers, thereby defining a new type of Cu(II) activesite, which is trinuclear. 2000

MATERIALS AND METHODS

Laccase was purified (17, 18) from the acetone powder(Saito and Company, Osaka, Japan) of the Japanese lacquertree, Rhus vernicifera. Routinely, the protein solutions wereconcentrated to =0.4 mM by using an Amicon YM-10 mem- ebrane. To obtain glasses of high optical quality, the protein -0.50was dialyzed into 50% (vol/vol) glycerol/0.1 M potassium .25x/cphosphate, pH 6.0, which further concentrated the sample to=1 mM (the protein concentration for all spectra presentedhere). For a given titration experiment, this protein was di- 300 400 500 600 700vided into 300- to 400-1ul aliquots, and 5 A.l of N- solution (of Waveengt,0nappropriate N- concentration) was added directly to each Wavelength, nmsample. These solutions were then equilibrated for -16 hr at FiG. 2. Native laccase spectra. The upper absorption spectra are40C. at 298 K (-) and 77 K (---). The lower spectra are variable-tem-

Absorption spectra were recorded by using two quartz perature MCD spectra at 50 kG: a, 59.8 K; b, 20.4 K; c, 11.8 K; d,disks spaced by a rubber gasket (1.4-mm path length) on a 4.75 K; and e, 1.75 K. Units of E are liter/mol-cmkG.Cary 17 spectrophotometer. The sample was mounted on anAir Products LT-3-110 Heli-tran liquid helium cryotip, whichwas fixed in the Cary 17 sample compartment. A LakeshoreCryotronics DTC-500 allowed spectra to be taken from 80-300 K. Equilibrium binding constants were calculated as de- 4.0 -.scribed (19).EPR spectra were recorded at 9.39 GHz on a Bruker ER

220 D-SRC spectrometer. An Air Products LTD-3-110 Heli-tran liquid helium transfer refrigerator and the Lakeshoretemperature controller maintained samples at 8 K.MCD spectra were recorded on a JASCO J500C spectro- 2.0 -

polarimeter with a modified sample compartment to accom-modate a superconducting magnet. Magnetic fields up to 60kG were produced by an Oxford SM4 superconducting mag-net/cryostat. Samples were mounted as described for thelow-temperature absorption spectra. Depolarization wasmeasured by recording the CD spectrum of a nickel tartratesolution placed before and after the sample (zero magnetic 0.30ffield) and was rarely observed. Sample temperatures were 0.3measured with a carbon glass resistor, calibrated from 1.5-300 K, by Cryogenic Calibrations (Pitchcott, Aylesbury,Buckinghamshire, U.K.). The temperature was maintained(±0.05 K) by an Oxford DTC-2 temperature controller con-nected to a Rh/Fe resistance thermometer.

RESULTS AND DISCUSSION

Fig. 2 shows the absorption at 298 and 77 K and the LTMCDspectra at 1.7-59.8 K of native laccase between 800 and 300nm. As the temperature was decreased, the 614-nm region ofthe absorption spectrum sharpened, similar to that previous- -0.60ly reported (20). In the LTMCD spectra of Fig. 2, the strong1/T dependence of each of the bands (610,55, 450, and 345nm) is evident and indicates that they are associated with the Wavelength, nmparamagnetic types 1 and 2 Cu(II) sites. FIG. 3. Titration of native laccase with N3 . (Upper) AbsorptionUpon N- titration of native laccase, several sets of over- spectra at 298 K. (Lower) LTMCD (4.9 K) spectra at 50 kG.

lapping spectral features appeared in the 550- to 300-nm Native laccase; ---, 0.25 and 0.50 protein equivalents of N3; -spectral region, which are assigned as N- -+Cu(II) CT tran- 2.5 and 9.0 protein equivalents of N;., 38.0 protein equivalentssitions (Fig. 3). Those features observed at N- concentra- of N-.

3064 Chemistry: Allendorf et aL

Proc. NatL Acad Sci. USA 82 (1985) 3065

tions s 0.5 protein equivalents (i.e., c 0.5 equivalents of N3per equivalent of laccase) are designated "high-affinity"based on their high binding constant (K > 10 M' at 298 and4.9 K). The second set appears at N- concentrations > 0.5protein equivalents and is designated "low-affinity" binding(K = 102-103 M-1 at 298 and 4.9 K). The low-affinity fea-tures further display two behaviors: for N- concentrationsup to 9.0 protein equivalents, the intensity of the LTMCDfeatures increases continuously, following that of the ab-sorption bands. However, at higher concentrations dramaticdecreases occur in some of the LTMCD features, while theabsorption bands continue to increase. Because the low-af-finity features dominate the spectra, these spectral changesare examined first.

Difference absorption and LTMCD spectra, relative to abaseline N- concentration of 0.5 protein equivalents (to cor-rect for the high-affinity N- features; see above) are shownin Fig. 4 for a titration of native laccase with NW. In the 550-to 370-nm region of the 298 K absorption spectrum, a strongpeak is observed at =400 nm with a broad shoulder centeredat =500 nm. In the LTMCD spectrum at 4.9 K (Fig. 4 Low-er), three sharp, negative bands appear at 485, 442, and 340nm with an equally intense positive feature at 385 nm. SinceMCD intensity is proportional to absorption intensity, thechanges observed in an MCD band with the addition of N3should reflect the changes occurring in the N - *Cu(II) CTabsorption feature corresponding to that MCD band. This

Wavelength, nm

333 400 500

:C

1500/

b

1000 A

5000

0.50-

3.0 2.5 2.0

Energy, 10 4xcm1

FIG. 4. Low-affinity N- spectral features. (Upper) Differencespectra at 298 K relative to an absorption baseline with N- concen-tration of 0.5 protein equivalents. Baseline variations at 77 K be-cause of stray light and cracking of the sample glass preclude thecalculation of accurate difference spectra at this temperature. (Low-er) LTMCD (4.9 K) difference spectra at N_ concentrations of 2.5(spectrum a), 9.0 (spectrum b), and 38.0 (spectrum c) protein equiva-lents.

parallel behavior is observed in Fig. 4 for the 485-nmLTMCD and 500-nm absorption bands; both increase contin-uously as the concentration of N- is increased. Since thereis a band in the LTMCD spectrum corresponding to the 500-nm absorption feature, the ground state of this electronictransition must be paramagnetic. Therefore, these two fea-tures can be assigned to a single CT transition from N- tothe type 2 Cu(II).The behavior of the remaining MCD bands in Fig. 4 is con-

siderably more complex. Continuous intensity increaseswith increasing N- concentration are observed for the 400-nm absorption band, but parallel changes in the MCD spec-trum do not occur. Initially, simultaneous (as evidenced bythe common crossover points at 351 and 422 nm) and contin-uous increases in the MCD intensity are seen at 340, 385, and442 nm as the N- concentration increases from 0.5 to 9.0protein equivalents. When the N- concentration is increasedfurther, however, the intensity of all three bands decreases,preventing a simple correlation between the absorption bandat 400 nm and the MCD features.

In the EPR spectrum (Fig. 5), however, a new signal isobserved between 3200 and 3800 G at <40 K whose intensitydependence with N- concentration can be directly correlat-ed with these MCD bands. This EPR signal intensity mea-sured at geff = 1.86 first increases and then decreases withincreasing N- concentration (Fig. 6). The spectral shape, pHdependence, and chemistry of this new signal are very simi-lar to a signal observed (unpublished data) for N- binding tofully oxidized T2D laccase and which has been associatedwith uncoupled type 3 sites. This uncoupling occurs in a lim-ited fraction of type 3 sites when N- and H+ competitivelydisplace and protonate the endogenous bridge. This elimi-nates the superexchange pathway resulting in a pair of para-magnetic, dipolar-coupled Cu(II) centers whose associatedN- -*Cu(II) CT transitions will now be observed in theLTMCD spectrum. Therefore, based on the quantitative cor-relation of the intensity of the LTMCD and EPR signals (Fig.6), the 340-, 385-, and 442-nm LTMCD bands are assigned asN- type 3 (uncoupled) Cu(II) CT transitions. The ob-served decrease of these bands and the EPR signal at N3concentrations > 9 protein equivalents suggests that an addi-tional N- binds, perhaps shifting the features to a differentspectral region.

a PI

b

0 3000 6000Gauss

FIG. 5. EPR spectrum at 8 K of native laccase (trace a) and na-tive laccase with 9 protein equivalents of N_ (trace b). Arrow pointsto g = 1.86.

Chemistry: Allendorf et aL

Proc. NatL Acad Sci. USA 82 (1985)

2.0

1.5-

o

0.5-~~

0 1 0 20 30 40

Protein Equivalents N3

FIG. 6. Changes of EPR, LTMCD, and absorption intensitieswith N_ concentration, normalized to the intensity of the spectrumwith N_ at 9 protein equivalents for each type of spectroscopy. ,LTMCD (4.9 K) at 485 nm; ---, A4w at 298 K; -.-, LTMCD (4.9 K)at 385 nm (intensity adjusted to maintain constant crossover pointsat 351 and 422 nm);. , g = 1.86 EPR signal at 8 K. All intensitieswere determined by using the baseline with N_ concentration at 0.5protein equivalents. Error bars with double hatch marks correspondto the EPR signal. The graph represents the observed behavior ofthese spectral features in four sets of experiments.

Alternatively, the changes in the absorption band at =400nm as a function of N- concentration do not reflect the in-tensity decrease in the LTMCD spectrum at 385 nm (Fig. 6).Thus, an additional CT spectral feature must be contributingto the absorption spectrum at 400 nm. The quantitative cor-relation of the LTMCD intensity changes with those of thetype 3 EPR signal (Fig. 6) requires that within ± 5%, all ofthe 385-nm LTMCD signal be associated with the uncoupledtype 3 sites. This indicates that the additional absorption fea-ture has no measurable LTMCD intensity associated with it.From the N- binding constant calculated from the 77 K ab-sorption at 400 nm combined with the integrated EPR inten-sity, it is estimated that .70% of the absorption intensity hasno LTMCD intensity; hence, it must arise from a transitionassociated with a diamagnetic site. Therefore, this additionalfeature, which dominates the absorption spectrum at 400nm, can be assigned to an N- -3 type 3 (coupled) Cu(II) CTtransition.The intensity changes for the 485-nm LTMCD band asso-

ciated with the paramagnetic type 2 sites are also included inFig. 6. This transition clearly shows a behavior with N3 con-centration that is qualitatively similar to that of the 400-nmabsorption band (i.e., both continuously increase), and bind-ing constants calculated for the two bands are estimated tobe the same within experimental error (K 102-103 M-1).Thus, low-affinity N- binding generates CT transitions fromboth a diamagnetic (400-nm absorption band) and a paramag-netic (385-nm LTMCD band) ground state with similar bind-ing constants. This demonstrates that either a single N3bridges the type 2 and coupled binuclear type 3 Cu(II) sites,or alternatively, that two N- molecules are binding withvery similar binding constants at the laccase active site, oneto the type 2 and one to the type 3 center.The assignment of both the 485-nm LTMCD and the 400-

nm absorption features to a single bridging N- is supported

by related studies at pH 7.0. Earlier work (11, 13) has shownthat the 400-nm absorption intensity is greatly reduced athigh pH due to a reduction of this N- binding constant; wefind a similar decrease of the 485-nm LTMCD intensity. It isunlikely that binding of N- to a binuclear type 3 vs. a mono-nuclear type 2 Cu(II) site would exhibit this same pH effect.Additionally, the binding constant of high-affinity N-3which corresponds to N- binding to an oxidized type 2 sitein the presence of a reduced type 3 site (see below), is notsignificantly affected by this pH change. These results indi-cate that the oxidation state of the type 3 site strongly affectsboth the N- binding constant of the type 2 site (low vs. highaffinity) and the pH perturbation of that binding. Finally, N3binding studies of fluoride-treated laccase indicate that fluo-ride binding at the type 2 Cu(II) directly competes with N3binding at the type 3 center (unpublished data). This recipro-cal interaction between the type 2 and type 3 sites togetherwith the correlated behavior of the 400-nm absorption and485-nm LTMCD features under a variety of conditionsstrongly indicates that a single N- bridges the type 2 andtype 3 sites.When low concentrations of N3 (<0.5 equivalents of N3

per equivalent of protein at a protein concentration of 1.0mM) are added to native laccase, a broad increase is ob-served in the N- -+Cu(II) CT region (Fig. 3 Upper). Thespectral changes are more clearly seen in the difference ab-sorption spectrum at 298 K shown in Fig. 7 Upper. In thecorresponding difference MCD spectrum, two bands of op-posite sign are clearly resolved at 510 and 445 nm (Fig. 7Lower). Substantial evidence indicates that these featuresare associated with N3 binding to a limited fraction of pro-tein molecules that contain Cu(I) sites. First, the LTMCDbands obtain their maximum intensity with less than stoi-

Wavelength, nm333 400 500

Energy, 104Xcm1

FIG. 7. Difference absorption spectrum at 298 K (Upper) anddifference LTMCD spectrum at 4.9 K (Lower) of N3 at a concentra-tion of 0.5 protein equivalents (high-affinity) relative to native lac-case. Difference spectra were not calculated for wavelengths <400nm because small changes of the large absorption intensity of nativelaccase in this region relative to the magnitude of the spectralchanges with N_ limits the accuracy of such subtractions.

3066 Chemistry: Allendorf et aL

Proc. NatL Acad Sci. USA 82 (1985) 3067

chiometric equivalents of N3 ('0.5 equivalents at 1.0 mMprotein). Second, x-ray absorption edge studies have shownthat '-'12% of the copper in native laccase is reduced (21) butcan be oxidized by peroxide. Finally, peroxide oxidationsuppresses essentially all of the absorption and LTMCD in-tensity normally associated with high-affinity N- binding tonative laccase, as has been similarly demonstrated in carefulEPR studies (22). The presence of bands in the LTMCD cor-responding to the new absorption intensity indicates that N3is binding to the type 2 Cu(II). Thus, the high-affinity formcontains N- bound to an oxidized type 2 Cu(II) site in thepresence of a reduced type 3 site. It should be noted that pH7.0 LTMCD/EPR/Absorption results suggest that there mayalso be a small fraction (<15%) of oxidized type 3/reducedtype 2 sites also present in the high-affinity form.

SPECTROSCOPICALLY EFFECTIVE MODEL FORTHE TYPE 2-TYPE 3 ACTIVE SITE

LTMCD spectroscopy of the N- -*Cu(II) CT transitions hasbeen used to probe the ground states of the copper centersinvolved in N3 binding to native laccase. The simultaneouspresence of LTMCD and absorption features at 485 nm showthat N- binds to the paramagnetic type 2 site, while an ab-sorption feature at =400 nm with no corresponding LTMCDfeature indicates that N- also binds to the diamagnetic type3 site. This assignment is further supported by the appear-ance at high concentrations of N- and low pH of an LTMCDfeature at 385 nm that directly correlates with an EPR signalassociated with N- binding to an uncoupled type 3 site.

Additional data indicates that the type 2 and type 3 N3binding sites must be close together. Changing the oxidationstate of the type 3 site alters the binding constant of N- forthe type 2 Cu(II) by more than a factor of 10 and changes thepH dependence of this binding. Additionally, the binding ofN3 at the type 3 site is strongly inhibited by binding fluorideto the type 2 center. These results, coupled with the parallelbinding constant and pH behavior of N3 binding at the type2 and coupled type 3 sites leads to a spectroscopically effec-tive model for N3 binding to native laccase: a single N3binds to a trinuclear Cu(II) site, bridging the type 2 and cou-pled type 3 Cu(II) centers (Fig. 8). Hence, the type 2 and atleast one of the type 3 Cu(II) ions must be separated by <5.2A. While the unusually high binding constant for N3 boundto an oxidized type 2 Cu(II) site in the presence of a reducedtype 3 site (high-affinity N3) may also relate to such a bridg-ing interaction, this cannot be spectroscopically demonstrat-ed by LTMCD, as N3 binding to Cu(I) does not produce anabsorption feature in an accessible spectral region.

Finally, the presence of a trinuclear copper site in laccasethat is capable of binding and bridging small molecules sug-gests the possibility that oxygen reduction may involve athree-electron-reduced dioxygen intermediate. While therehave been a number of reports of two-electron-reduced, per-oxide-level intermediates, the two stable forms, peroxylac-case (23, 24) and peroxy-T2D (25) laccase, have been shown

I-, Cu "'2N

N

N

"I-, /

Cu/2 CU+2

R I

TYPE 2

TYPE 3

FIG. 8. Trinuclear Cu(II) active site model for N- binding bynative laccase.

(21) through x-ray absorption edge studies to involve oxida-tion but not binding by peroxide. Alternatively, in reactionsof anaerobically reduced laccase with 02, a paramagnetic in-termediate has been observed (26) and, although copper ionoxidation states are not clear (27, 28), the presence of an 0-radical has been proposed and would be consistent withthree-electron transfer.

Thus, it is clear that the interaction of exogenous ligandsat the coupled binuclear Cu(II) site in native laccase is verydifferent from that in the hemocyanins and tyrosinase. Fur-ther, the lack of 02 reactivity of the type 3 site in T2D lac-case would appear to be a consequence of the strong in-volvement of the type 2 Cu(II) in the oxygen reactivity ofthis trinuclear copper exogenous ligand binding site in nativelaccase. Similarly, the inability of exogenous ligands tobridge the coupled binuclear Cu(II) site in laccase is consist-ent with the type 3 site instead being structurally defined forasymmetric exogenous ligand bridging to the type 2 copperof the trinuclear copper active site.We thank Philip Stephens for very useful discussions concerning

the design of our MCD spectrometer. We also thank the NationalInstitutes of Health (AM31450) for support of this research.1. Reinhammar, B. (1972) Biochim. Biophys. Acta 275, 245-259.2. Fee, J. A. (1975) Struct. Bonding (Berlin) 23, 1-60.3. Malkin, R. & Malmstrom, B. G. (1970) Adv. Enzymol. Relat. Areas

Mol. Biol. 33, 177-244.4. Solomon, E. I. (1981) in Copper Proteins, ed. Spiro, T. G. (Wiley,

New York), pp. 42-108.5. Himmelwright, R. S., Eickman, N. C., LuBien, C. D., Lerch, K.

& Solomon, E. I. (1980) J. Am. Chem. Soc. 102, 7339-7344.6. Eickman, N. C., Himmelwright, R. S. & Solomon, E. I. (1979)

Proc. Natl. Acad. Sci. USA 76, 2094-2098.7. Graziani, M. T., Morpurgo, L., Rotilio, G. & Mondovi, B. (1976)

FEBS Lett. 70, 87-90.8. LuBien, C. D., Winkler, M. E., Thamann, T. J., Scott, R. A., Co,

M. S., Hodgson, K. 0. & Solomon, E. I. (1981) J. Am. Chem. Soc.103, 7014-7016.

9. Spira, D. J., Winkler, M. E. & Solomon, E. I. (1982) Biochem.Biophys. Res. Commun. 107, 721-726.

10. Spira, D. J. & Solomon, E. I. (1983) Biochem. Biophys. Res. Com-mun. 112, 729-736.

11. Winkler, M. E., Spira, D. J., LuBien, C. D., Thamann, T. J. &Solomon, E. I. (1982) Biochem. Biophys. Res. Commun. 107, 727-734.

12. Hahn, J. E., Co, M. S., Hodgson, K. O., Spira, D. J. & Solomon,E. I. (1983) Biochem. Biophys. Res. Commun. 112, 737-745.

13. Morpurgo, L., Rotilio, G., Finazzi-Agro, A. & Mondovi, B. (1974)Biochim. Biophys. Acta 336, 324-328.

14. Solomon, E. I., Dooley, D. M., Wang, R. H., Gray, H. B., Cer-donio, M., Mogno, F. & Romani, G. L. (1976) J. Am. Chem. Soc.98, 1029-1031.

15. Dooley, D. M., Scott, R. A., Ellinghaus, J., Solomon, E. I. &Gray, H. B. (1978) Proc. Natl. Acad. Sci. USA 75, 3019-3022.

16. Stephens, P. J. (1976) Adv. Chem. Phys. 35, 197-264.17. Reinhammar, B. (1970) Biochim. Biophys. Acta 205, 35-47.18. Reinhammar, B. & Oda, Y. (1979) J. Inorg. Biochem. 11, 115-127.19. Byers, W., Curzon, G., Garbett, K., Speyer, B. E., Young, S. N.

& Williams, R. J. P. (1973) Biochim. Biophys. Acta 310, 38-50.20. Dooley, D. M., Rawlings, J., Dawson, J. H., Stephens, P. J., An-

dreasson, L. E., Malmstrom, B. G. & Gray, H. B. (1979) J. Am.Chem. Soc. 101, 5038-5046.

21. Penner-Hahn, J. E., Hedman, B., Hodgson, K. O., Spira, D. J. &Solomon, E. I. (1984) Biochem. Biophys. Res. Commun. 119, 567-574.

22. Morpurgo, L., Desideri, A. & Rotilio, G. (1982) Biochem. J. 207,625-627.

23. Farver, O., Goldberg, M. & Pecht, I. (1980) Eur. J. Biochem. 104,71-77.

24. Farver, 0. & Pecht, I. (1981) in Copper Proteins, ed. Spiro, T. G.(Wiley, New York), pp. 152-218.

25. Farver, O., Frank, P. & Pecht, I. (1982) Biochem. Biophys. Res.Commun. 108, 273-278.

26. Andrdasson, L. E., Branddn, R. & Reinhammar, B. (1976) Bio-chim. Biophys. Acta 438, 370-378.

27. Aasa, R., Branddn, R., Deinum, J., Malmstrom, B. G., Reinham-mar, B. & Vanngard, T. (1976) FEBS Lett. 61, 115-119.

28. Andrdasson, L. E. & Reinhammar, B. (1979) Biochim. Biophys.Acta 568, 145-156.

Chemistry: Allendorf et aL