Copper electron-nuclear double resonance of cytochrome c oxidase

Proc. Nati. Acad. Sci. USA Vol. 77, No. 3, pp. 1452-1456, March 1980

Biophysics

Copper electron-nuclear double resonance of cytochrome c oxidase (electron paramagnetic resonance)

BRIAN M. HOFFMAN*t, JAMES E. ROBERTS*, MAURICE SWANSONt, SAMUEL H. SPECKt, AND E. MARGOLIASHt Departments of *Chemistry and tBiochemistry and Molecular Biology, Northwestern University, Evanston, Illinois 60201

Contributed by Emanuel Margoliash, December 5,1979

ABSTRACT Electron-nuclear double resonance of copper was observed while monitoring the "intrinsic copper" electron paramagnetic resonance signal of cytochrome c oxidase (ferrocytochrome c:oxygen. oxi oreductase, EC 1.9.3.1) near g = 2. This unambiguously establishes the presence of the metal (Cu.) in the 'redox center responsible for this signal. The hyperfine couplings to copper are largely isotropic and the maximum value is about half that seen in type I blue copper proteins. The magnetic properties of this oxidized copper center are not consistent with those of a thiyl radical (R-S.) coordinated to Cu(I), and thus favor the identification of this redox center as a Cu(II) ion in a unique environment. Cytochrome c oxidase (ferrocytochrome c:oxygen oxidoreductase, EC 1.9.3.1), the terminal enzyme complex of the eukaryotic electron transport chain, is a transmembrane oligomeric protein consisting of six to ten polypeptides containing two a-type hemes and two copper atoms, as well as lipid (1, 2). Electron paramagnetic resonance (EPR) spectra of cytochrome oxidase exhibit signals from several paramagnetic centers (3-5). The most intense of these signals, near g = 2, was initially attributed to copper (6) and is often referred to as the "intrinsic copper" spectrum. Although this signal displays the characteristic form of a cupric ion center (gII > go ge), it nevertheless shows an unusually low value of gII, exhibits no hyperfine splittings (hfs) from copper at X or Q band, has one principalaxis g value of less than 2.00, and has an anomalous temperature-dependent spin-lattice relaxation rate (7-9). Simulations of X-band EPR spectra are consistent with the presence of a Cu(II) ion (5), and Froncisz et al. (9) have recently reported the resolution in.2- to 4-GHz EPR of hyperfine structure that may be-from copper. However, it has also been noted that the spectra features listed above are characteristic of thiyl radicals (R-S.) (10), and, in a recent electron-nuclear double resonance (ENDOR) investigation of cytochrome oxidase, proton and nitrogen ENDOR was observed, but copper could not be de-

tected (11). To determine whether the metal ion is indeed part of this oxidation-reduction center, we have reinvestigated the ENDOR of cytochrome c oxidase. Although the EPR of cupric complexes is. almost invariably straightforward, the observation of ENDOR from the copper nucleus has been unaccountably difficult. We have recently studied the copper ENDOR of a number of blue copper proteins and found the proper conditions for its detection (unpublished, and see refs. 12 and 13). The present communication reports the copper ENDOR from the intrinsic copper signal of cytochrome oxidase and presents a discussion of the electronic and geometric properties of this metal center.

EXPERIMENTAL PROCEDURES Cytochrome c oxidase was prepared from beef heart mitochondria according to Hartzell and Beinert (14) and was stored in liquid nitrogen in 0.02 M Tris.HCI, pH 7.4/0.25% Tween 20. It had an enzymic turnover of 350 electrons per sec as determined by the method of Nicholls et al. (15). Heme a concentration was determined spectrophotometrically by using a Ac605-6,0 = 13.1 mM-1 cm-1 (16). All EPR and ENDOR spectra were recorded on enzyme preparations that were diluted by 50% with glycerol and contained a final concentration of 0.25 mM cytochrome aa3. ENDOR experiments were performed at 2.0K, using the spectrometer described elsewhere (13), to which has been added a WAVETEK model 2001 sweep generator and a Fabritek model 1074 signal averager. ENDOR spectra were obtained with a 100-kHz filed modulation amplitude of -4 G (1 G = 10-4 tesla) and radiofrequency (rf) field strength of t1 G in the rotating frame. Microwave power and rf sweep rate varied, as discussed below. EPR spectra simulations were calculated by using program SIM 14 (17). RESULTS EPR. In Fig. 1 the X-band EPR spectrum of the g = 2 region of a frozen solution of cytochrome oxidase is shown. The signal from the "intrinsic Cu" site is free from features associated with adventitiously bound copper (7) and, except for possibly higher resolution brought about by the glycerol in the medium, is in complete accord with the spectrum published by van Camp et al. (11). Both these spectra (see Fig. i and ref. 11) are noticeably better resolved than the spectrum of Greenaway et al. (5), perhaps because of unfavorable conditions used by these authors, such as high microwave power, high modulation amplitude, and relatively high temperature. The superficial appearance is that of an axial g tensor (gII > g1), but Q-band spectra show a rhombic splitting of the g1 region and give the g values listed in Table 1 (3). As noted (10), the spectrum,. and in particular the g tensor, is at least as similar to that of a thiyl radical (18-20) as it is to the type I copper centers with most similar g tensors (21). Representative g values for the blue copper proteins and for a thiyl radical are also listed in Table 1.

ENDOR. ENDOR is performed by inducing nuclear transitions with a rf field while observing the EPR signal intensity with the external magnetic field, Ho, set a fixed value (22). Field positions at the extreme edges of the EPR spectrum, near either g. or gx (positions A and C in Fig. 1) will give single crystal-like

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

Abbreviations: ENDOR, electron-nuclear double resonance; EPR, electron paramagnetic resonance; hfs, hyperfine splittings; rf, radiofrequency.

this fact. 1452

Biophysics:

Proc. Natl. Acad. Sci. USA 77 (1980)

Hoffman et al.

1453

angle-dependent g value is equal to gy. Thus, patterns obtained at these fields are typically more poorly resolved and less susceptible to interpretation. With Ho set near any of the positions indicated in Fig. 1, when the frequency of the rf field is swept slowly (dv/dt < 10 MHz/sec) and the microwave power is relatively low (