Pulse Radiolysis Studies of Temperature ... - ACS Publications

Oct 28, 2010 - For J.A.F., Scripps: phone, 858-784-9235; fax, 858-784-2857; E-mail, [email protected]. For I.P., Weizmann: phone, +972-8-9344020; fax,...
6 downloads 4 Views 2MB Size
Biochemistry XXXX, XXX, 000–000 A DOI: 10.1021/bi100548n

Pulse Radiolysis Studies of Temperature Dependent Electron Transfers among Redox Centers in ba3-Cytochrome c Oxidase from Thermus thermophilus: Comparison of A- and B-Type Enzymes† )

Ole Farver,‡,§ Scot Wherland,§, William E. Antholine,^ Gregory J. Gemmen,z Ying Chen,z Israel Pecht,*,§ and James A. Fee*,z )

‡ Institute of Analytical Chemistry, Faculty of Pharmacy, University of Copenhagen, 2100 Copenhagen, Denmark, Department of Immunology, The Weizmann Institute of Science, 76100 Rehovot, Israel, Department of Chemistry, Washington State University, Pullman, Washington 99164, United States, ^National Biomedical ESR Center, Medical College of Wisconsin, Milwaukee, Wisconsin 53226 United States, and zDepartment of Molecular Biology, The Scripps Research Institute, La Jolla, California 92037, United States

§

Received April 12, 2010; Revised Manuscript Received October 27, 2010 ABSTRACT: The functioning of cytochrome c oxidases involves orchestration of long-range electron transfer (ET) events among the four redox active metal centers. We report the temperature dependence of electron transfer from the CuAr site to the low-spin heme-(a)bo site, i.e., CuAr þ heme-a(b)o f CuAo þ heme-a(b)r in three structurally characterized enzymes: A-type aa3 from Paracoccus denitrificans (PDB code 3HB3) and bovine heart tissue (PDB code 2ZXW), and the B-type ba3 from T. thermophilus (PDB codes 1EHK and 1XME). k,T data sets were obtained with the use of pulse radiolysis as described previously. Semiclassical Marcus theory revealed that λ varies from 0.74 to 1.1 eV, Hab, varies from ∼2  10-5 eV (0.16 cm-1) to ∼24  10-5 eV (1.9 cm-1), and βD varies from 9.3 to 13.9. These parameters are consistent with diabatic electron tunneling. The II-Asp111Asn CuA mutation in cytochrome ba3 had no effect on the rate of this reaction whereas the II-Met160Leu CuA-mutation was slower by an amount corresponding to a decreased driving force of ∼0.06 eV. The structures support the presence of a common, electron-conducting “wire” between CuA and heme-a(b). The transfer of an electron from the low-spin heme to the high-spin heme, i.e., heme-a(b)r þ heme-a3o f heme-a(b)o þ heme-a3r, was not observed with the A-type enzymes in our experiments but was observed with the Thermus ba3; its Marcus parameters are λ = 1.5 eV, Hab = 26.6  10-5 eV (2.14 cm-1), and βD = 9.35, consistent also with diabatic electron tunneling between the two hemes. The II-Glu15Ala mutation of the K-channel structure, ∼24 A˚ between its CA and Fe-a3, was found to completely block heme-br to hemea3o electron transfer. A structural mechanism is suggested to explain these observations.

Cytochrome c oxidases (CcO1) serve as terminal oxidants in oxygen respiring organisms (1, 2). They catalyze single electron transfer from cytochromes c to O2, reducing the latter to two H2O while translocating four protons from the inner space (mitochondrial matrix or bacterial cytoplasmic space) to their respective outer spaces (ref 3 and references therein). The A-, B-, and C-type dioxygen reductases dominate the larger superfamily of † This work was supported by the United States Public Health Service, GM35342 (J.A.F.), and the Edith and Bernard Shoor Foundation, USA (I.P.). *To whom correspondence should be addressed. For J.A.F., Scripps: phone, 858-784-9235; fax, 858-784-2857; E-mail, [email protected]. For I.P., Weizmann: phone, þ972-8-9344020; fax, þ972-8-9465264. E-mail, [email protected]. 1 Abbreviations: Tt, Thermus thermophilus; Pd, Paracoccus denitrificans; Bovine, Bos taurus; heme-bo, oxidized form of heme-b; heme-br, reduced form of heme-b; heme-a3o, oxidized form of heme-a3; heme-a3r, reduced form of heme-a3; heme-a(b), heme-a or heme-b, low-spin heme; kB, the Boltzmann constant, 8.617  10-5 eV/K; h, Planck’s constant, 4.136  10-15 eV 3 s; ET, electron transfer or electron tunneling; Ea, activation energy (eV); AA Arrhenius pre-exponential term (s-1); ΔH*, activation enthalpy (eV); ΔS* activation entropy (eV/K); κ, transmission coefficient; AM, Marcus pre-exponential term (K1/2/s); ΔEaM, Marcus free energy of activation (eV); ΔEo, relative redox potential (eV); Hab, electron coupling between donor and acceptor (eV or cm-1); λ (eV), reorganization energy (eV); β, attenuation factor (A˚-1); D, electron tunneling distance (A˚); WT, wild-type.

r XXXX American Chemical Society

heme-copper oxidases (4, 5). All have a high-spin heme in close proximity to the three-coordinate CuB locus, forming the core of the enzyme, the binuclear active site, into which O2, electrons, and protons enter and out of which product water and pumped protons exit via structurally suggestive paths and channels. The A- and B-family enzymes share highly similar structural and functional attributes, but they also differ in important ways. In particular, the ba3-oxidase from T. thermophilus (Tt), the first characterized member of the B-family (6), possesses a B-heme in the low-spin heme binding site, has an enlarged O2-uptake channel (7), lacks the so-called D- and Q-paths for proton uptake (8), and uses a structurally well-defined K-channel for uptake of both “pumped” and scalar protons (8). The reactions of these enzymes may be viewed as occurring within a relatively rigid environment (9) and have been the object of considerable experimental and theoretical attention, primarily on members of the A1-family (see refs (2, 10, 11) and (12-14) for overview) and to a lesser extent on the B-family ba3 oxidase (15-19). To varying degrees, the paths and channels that serve and/or connect the different centers have been charted within the available 3-dimensional structures (2, 7, 20, 21). Comparisons reveal significant variations between species, indicating that a single, conserved combination of amino acid sequence is not required for enzyme activity. However, the active site structure itself is pubs.acs.org/Biochemistry

B

Biochemistry, Vol. XXX, No. XX, XXXX

essentially constant in both sequence alignments and in the current 3-dimensional structures of A- and B-family enzymes (8). Pulse radiolysis has been widely employed in studies of electron transfer (ET) to and within a variety of multiredox center proteins (22-24), including the A1-family members, the bovine enzyme (25, 26), the bacterial one from P. denitrificans (27), and the B-family cytochrome ba3 oxidase (28). Here we extend this work to include mutants of the ba3 CuA site and to the temperature dependence of ET in the aa3 containing oxidases from bovine heart and P. denitrificans and the ba3-oxidase from T. thermophilus. MATERIALS AND METHODS Hepta-histidine tagged, recombinant cytochrome ba3, and the mutant forms used in this work were prepared, purified, and characterized as described by Chen et al. (29) and Keightley et al. (30); mutant forms of ba3 were constructed by standard techniques. The three mutants used in this study are II-Met160Leu and IIAsp111Asn in subunit II, and II-Glu4Gln/I-Lys258Arg/IIGlu15Ala, referred to as II-Glu15Ala. Relative to wild-type, recombinant ba3, the activities, as measured in ref 29, are 83% and 30% for mutant forms II-Met160Leu and II-Asp111Asn, respectively, with an error of ∼(10%. The II-Glu15Ala mutant is devoid of activity in this assay. The double mutation of the wild-type protein carrying the IIGlu4Gln/I-Lys258Arg was developed because of its crystallization properties and is in all measured functional aspects equivalent to the wild-type recombinant enzyme (31, 32). Mutations were also made in the water-soluble subunit II fragment, termed CuAT0 (33). The II-Asp111Asn mutant was not formed in the reconstitution mixture, as evidenced by the absence of the intense purple color that accompanies copper binding to the wild-type protein. In contrast, CuA formation in the II-Met160Leu mutant was indicated by this purple color, but it was not stable to subsequent purification steps. This is not to say that conditions cannot be found to produce these mutants in water-soluble fragments useful for more detailed study, in particular for measurements of their reduction potentials, as has been done by Ledesma et al. (34). Solutions of all proteins were characterized by their optical absorption spectra, as recorded in San Diego using an SLM/ AMINCO model DB3500, in Jerusalem using a Hewlett-Packard 8452A diode array spectrophotometer, and by multifrequency, low-temperature EPR at the National Biomedical ESR Center in Milwaukee (see below). The concentration of fully constituted protein was obtained using εred560-590 = 26 mM-1 cm-1 for determining enzyme concentrations (29) and a purity ratio, A413/A280, ranging from >0.7 to 0.9 as an indicator of purity. Solutions of ba3 contained 10 mM potassium phosphate, pH 7.5 buffer, and 1 mM dodecylmaltoside (DDM) were sealed in glass ampules and shipped from San Diego to Rehovot and/or Milwaukee on wet ice and subsequently maintained at 0 to 4 C. Enzyme samples so stored, or subjected to ∼75 C for several hours, followed by this type of storage, showed no decrease in ET rates or enzymatic activity. Protein structural Schemes 2, 3, and 4 were extracted from the protein data bank, PDB codes: 1XME and 1EHK. Additional PDB codes relevant to Scheme 3 are 3HB3 (Paracoccus aa3, 2GSM (Rhodobacter aa3), and 2OCC (bovine aa3). Multifrequency EPR spectra were recorded using loop-gap resonators and low-frequency microwave bridges designed and

Farver et al. Table 1: Molar Extinctions and Extinction Changes (M-1 cm-1) of Cytochrome ba3 at the Three Wavelengths Used in This Work wavelength (nm) extinction coefficient (M-1 cm-1) 790

560

445

1959 59 2094 1969

18325 15668 35739 19402

46962 44785 48395 113557

-1900 135 10 2035 -125 1910

-2657 17414 1077 20071 -16337 3734

-2177 1433 66595 3610 65162 68772

Oxidation State CuAoboa3o CuArboa3o CuAobra3o CuAoboa3r Process CuArboa3o-CuAoboa3o CuAobra3o-CuAoboa3o CuAoboa3r-CuAoboa3o CuAobra3o-CuArboa3o CuAoboa3r-CuAobra3o CuAoboa3r-CuArboa3o

built at the National Biomedical ESR Center, Medical College of Wisconsin (see ref 35 and references therein). X-band spectra were obtained on a Varian Century series spectrometer at Milwaukee. Time-resolved measurements of electron transfer were carried out on the Varian V-7715 linear accelerator at the Hebrew University in Jerusalem, Israel. Electrons accelerated to 5 MeV were employed. Protein solutions containing 5 mM 1-methyl-nicotinamide chloride (MNAþ Cl-), 10 mM potassium phosphate buffer at pH 7.5 and 1 mM DDM were deoxygenated by saturation with high purity Ar at a pressure slightly in excess of 1 atm. The solutions were irradiated with 0.05-1.5 μs pulses, and the yield of reducing MNA* radicals is ∼4 μM/μs pulse width. A 1.00 cm Suprasil cuvette (HELLMA) was employed, using either one or three light passes resulting in an optical path length of 1 or 3 cm, respectively. Reactions were performed under pseudo-first-order conditions, typically with a 20-fold excess of oxidized protein over reducing radicals. Under these conditions, less than one reducing equivalent was taken up by the enzyme during each pulse. Static spectra were acquired before, after, and at intervals during the experiments in order to determine levels of reduction and test for oxygen leaks. Quantitative features of the optical absorption spectra of cytochrome ba3 are important for both analysis and interpretation of the kinetic data. The current estimates of molar extinction coefficients used in this work are summarized in Table 1. Macro rate constants were extracted as described in ref 27. The error in the observed rate constants is on the order of 10-15%. The magnitude of the spectral changes are generally consistent with pulse duration and the extinction coefficients presented in Table 1, although low frequency noise associated with our optical system, on the order of several seconds, may affect amplitudes in some traces. The estimated error in the amplitudes is therefore ∼50%. Kinetic data used here for the Pd enzyme were obtained in the course of a collaboration with the laboratory of Dr. B. Ludwig (27). One set of k,T data for the bovine enzyme was obtained in the course of a collaboration with the laboratory of Dr. O. Einarsdottir (26). A second k,T set of data for this enzyme was obtained from Figure 6 of ref 25; these data are unique in being recorded at the 830 nm absorption band of CuA in the bovine enzyme. The chosen kinetic scheme was analyzed in terms of reaction rate theory, see ref 36, as presented in Appendix A of the Supporting Information.

Article

Biochemistry, Vol. XXX, No. XX, XXXX

Scheme 1: Kinetic Scheme for Reactions Studied in This Work (See Appendix A of Supporting Information for Justification of Assignments)

Table 2: Observed Rate Constants (s-1) and Equilibrium Constants for Reactions II and III of Scheme 1, Where kobs2 = kf2 þ kb2, K2 = kf2/kb2 and kobs3 = kf3 þ kb3, K3 = kf3/kb3 (WT = Wild-Type) Pd

Plots of k (s-1) vs T (K) were fitted using nonlinear regression as described in Appendix B of the Supporting Information. All k, T data sets are equally well-fitted by any of the exponential expressions for unimolecular processes eqs 1-4.   - Ea ð1Þ k ¼ AA exp Arrhenius kB T     kB T ΔS - ΔH K exp exp k ¼ transition state=Eyring h kB kB T ð2Þ ! 2 2 2 4π Hab - ðλ - ΔEo Þ ð3Þ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp k¼ Marcus h 4λkB T 4πλkB T

C

bovine

WT ba3

ba3 D111N

T

kobs2

K2

T

kobs2

K2

T

kobs2

K2

T

kobs2

K2

277 280 283 288 294 298 309

11650 15090 17130 18490 26610 30340 41390

2.10 2.09 2.07 2.05 2.03 2.02 2.00

275 277 278 281 286 289 290 294 299 300 307 309 312

10126 10880 11241 10611 12918 12965 13962 15887 15371 17270 18295 22353 19746

3.75 3.70 3.64 3.61 3.59 3.53 3.50 3.48 3.43 3.40 3.36 3.34 3.30

278 287 298 308 319

8380 8500 11200 21600 22400

0.63 0.42 0.53 0.34 0.44

293 296 303 314 322 332

7146 8600 13559 18926 19006 22014

0.66 0.64 0.54 0.42 0.33 0.27

WT ba3

D111N ba3

T

kobs3

K3

T

kobs3

M160L ba3 K3

T

kobs3

K3

278 287 298 308 319 348

333 722 770 1600 2500 6700

0.19 0.29 0.37 0.36 0.48 0.5

296

1100

0.32

285

610

0.32

k ¼ 1:0  1013 expð - βDÞ expð - E a M =kB TÞ distance dependence:

ð4Þ Commonly used symbols are defined in the list of abbreviations used. RESULTS General. Reaction I of Scheme 1 is diffusion controlled (28), specifically and irreversibly placing an electron on CuAo followed by slower relaxation via reactions II and III in Scheme 1. The latter reactions do not go to completion. By determining the optical density amplitudes due to CuAo reduction and CuAr reoxidation, monitored at 790 nm, or those of heme-bo reduction and heme-br reoxidation, monitored at 560 nm, and heme-a3o reduction, monitored at 445 nm, the equilibrium constants could be calculated for processes II and III in Scheme 1. The relations between the observed, macro rate constants, kobs2 and kobs3, and the elementary, micro rate constants, kf2, kb2, kf3, and kb3, are critically dependent on the relative magnitudes of kobs2 and kobs3 and the equilibrium constants, K2 = kf2/kb2 and K3 = kf3/kb3 (see ref 36). That these are kobs2 = kf2 þ kb2 and kobs3 = kf3 þ kb3 is justified in Appendix A of the Supporting Information. The observed rate and equilibrium constants for the three enzymes are presented in Table 2, and values of the elementary rate constants, calculated using these expressions, are shown in Tables S2 and S3 of the Supporting Information. The temperature dependence of kobs2 and kobs3 of reactions II and III in Scheme 1 has been investigated in the range 278-320 K for the wild-type Tt ba3 at pH 7.5. The results for reaction II are compared with those previously obtained for the cytochrome aa3 oxidases from P. denitrificans (27) and from bovine heart (25, 26). Reaction III of Scheme 1 has been observed in the Tt ba3 enzyme (28), but not in either the Pd or bovine cytochromes aa3. The temperature dependent data for Reactions II and III were

analyzed in depth using transition state theory (37) and the Marcus theory of electron tunneling (38, 39), as detailed in Appendix B of the Supporting Information. The results are presented in Tables 3 - 6 (see below) and considered further in the Discussion. Proteins Used. “CuA mutants” of ba3-oxidase were constructed, II-Met160Leu and II-Asp111Asn, the locations of which are indicated in Scheme 2. These mutations have significant effects on the spectral properties of the CuA center but none to small effects on electron transfer rates. The reduced minus oxidized optical absorption spectra of wild-type and the two CuA mutants of cytochrome ba3 are shown in Figure 1. In the visible and Soret regions, the three proteins are indistinguishable, indicating that both hemes are properly assembled; spectral features relevant to the kinetic studies are noted with arrows and described in the figure legends. We also prepared the II-Glu15Ala “K-channel mutation” of ba3 (8), which expressed in relatively small amounts but whose optical properties are shown in Figure S4 of the Supporting Information. The rationale for creating the CuA mutants was the expectation that the CuA center in each would have a different redox potential, and this would affect the rate of electron transfer between CuA and heme-b in a predictable manner. Such an effect has already been observed in the Pd enzyme as the II-Met227Ile mutation (40) and in the II-Met163Leu mutation of the Rhodobacter enzyme (35, 41). As previously noted by Zhen et al. (35) for the enzyme from Rhodobacter sphaeroides and by Zickermann et al. (40) for the P. denitrificans enzyme, replacing the weakly coordinated (42) methionine ligand with noncoordinating leucine or isoleucine causes significant effects on the spectral properties of the oxidized CuA center. These are expressed in shifts of the characteristic, near-infrared band at ∼800 nm, the formally 1þ charged

D

Biochemistry, Vol. XXX, No. XX, XXXX

Farver et al.

Table 3: Extraction of Extrathermodynamic Parameters from kf2 Using eqs 1 and 2 (CuAr þ heme a(b)o f CuAo þ heme a(b)r) parameter

Paracoccus kf2

bovine kf2 (OE)a

bovine kf2 (KK)a

bovine kf2 (combined)a

Thermus kf2

AA, s-1 Ea, eV ΔH*, eV ΔS*c, eV/K ΔG* (298 eV)

5.6 ( 0.12b  108 0.26 ( 0.00 0.25 ( 0.02 -83.8 ( 7.0  10-5 0.50

3.02 ( 0.03  106 0.14 ( 0.00 0.12 ( 0.00 -132.9 ( 8.7  10-5 0.51

2.10 ( 0.04  108 0.25 ( 0.00 0.22 ( 0.02 -99.2 ( 6.3  10-5 0.51

5.58 ( 0.09  106 0.16 ( 0.01 0.14 ( 0.01 -126.4 ( 0.14  10-5 0.51

3.36 ( 0.07  107 0.23 ( 0.02 0.23 ( 0.01 -108.4 ( 10.0  10-5 0.54

a OE and KK indicate that the data used here was obtained using enzyme from the laboratory of Drs. O. Einarsdottir (recorded at 607 nm) and K. Kobayashi (recorded at 830 nm), respectively (see text), while “combined” indicates the two data sets were combined for the analysis. bErrors reported reported correspond to 1σ from the fitting program (see footnotes to Table 4). Parameters derived therefrom, e.g. ΔG*, are expected to be greater than the estimated error in each of the k,T data sets (see text). cThe Eyring transmission coefficient, κ, is contained within ΔS* and is assumed equal to 1.

Table 4: Extraction of Parameters from kf2 Using Marcus’ Electron Tunneling Theory (eqs 4B-9B, Appendix B of Supporting Information) (CuAr þ heme a(b)o f CuAo þ heme a(b)r) parameter

Paracoccus kf2

bovine kf2 (combined)a

Thermus kf2

AM, (K1/2/s)b ΔEaM, eV ΔEo, eV λ, eV -1 |H ) P ab|, eV (cm -1 d * (ln s ) M ln kf2(298) (exptl) ln kf2(298) (calcd) βDe

1.61 ( 0.04c  1010 0.28 ( 0.02 0.018 1.1 24.1  10-5 (1.94) 27.45 9.92 10.19 9.6

1.60 ( 0.03  108 0.17 ( 0.01 0.031 0.74 2.16  10-5 (0.17) 28.07 9.53 9.41 13.9

9.63 ( 0.20  108 0.25 ( 0.000 -0.016 0.98 6.67  10-5 (0.54) 29.54 8.26 8.24 11.8

Rhodobacter kf2 (projected) 0.25 0 1 30.6  10-5 (2.46) 25.9 11.44

a Indicated errors in the primary fitting parameters, e.g., AM and ΔEMa, are output from the fitting program when one parameter is allowed to float while P the other is fixed; they correspond to one standard deviation rounded to the last decimal. bUnits are discussed in the text. cSee footnotes to Table 3. d M* = ||EaM/kBT| þ |2ln(Hab)exp|| (from equation 3) .

Table 5: Extraction of Extrathermodynamic parameters from kf3 using eqs 1 and 2 (See Table 3 for Details) (heme br þ heme a3o f heme bo þ heme a3r) parameter AA, s-1 Ea, eV ΔH*, eV ΔS*, eV/K ΔG* (298, eV)

Thermus kf3

Table 6: Extraction of Parameters from kf3 Using Marcus’ Electron Tunneling Theory (eqs 4B-9B, Appendix B of the Supporting Information) (See Table 4 for Other Relevant Footnotes) (heme br þ heme a3o f heme bo þ heme a3r) parameter

5.7 ( 0.1  108 0.37 ( 0.02 0.40 ( 0.04 -71.4 ( 11.0  10-5 0.62

[Cu(1)þ1.5-Cu(2)þ1.5] oxidized form of the cluster. The nearinfrared spectra shown in Figure 2 confirm the binuclear character of the CuA center in the mutant proteins used here. The first-derivative, X-band EPR spectra of our wild-type and mutant proteins are compared in Figure 3, and the second derivative spectra are shown in Figure S1 of the Supporting Information. Relevant EPR parameters are assembled in Table S1 of the Supporting Information. In all available structures of the cytochrome c oxidases, the NE-atom of the lower-numbered histidine ligand to CuA (His114 in ba3 numbering) faces the aqueous outside compartment and is in hydrogen bonding distance (e3 A˚) of the OD2 atom of an aspartate residue (Asp111 in ba3 numbering, see Scheme 2). In this arrangement, the His114-NE2 atom should be largely protonated at pH 7.5 while the Asp-OD2 atom should be largely deprotonated at pH 7.5. The mutation thus removes a negative charge from the second coordination shell of the CuA complex, and this is the most likely cause of the blue-shifted, near-infrared

AM, K1/2/s ΔEaM, eV ΔEo, eV λ, eV |H , eV (cm-1) P ab|exp a M* ln kf3(298) (exptl) ln kf3(298) (calcd) βDA aP a M* = ||E M/kBT| þ |2ln(Hab)exp||

Thermus kf3 1.69 ( 0.04  1010 0.39 ( 0.02 -0.026 1.5 26.6  10-5 (2.14) 31.5 5.34 5.68 9.35

band (Figure 2) and the altered EPR spectrum as shown in Figures 2 and 3 and Figure S1 of the Supporting Information. Kinetic Results. The changes in absorbance at each wavelength are followed over two time frames as illustrated for the recombinant, wild-type protein in Figure 4. Shown from top to bottom are the reduction of the CuA center by MNA* (reaction I of Scheme 1), reoxidation of CuAr by heme-bo (reaction II), and reoxidation of heme-br by heme-a3o (reaction III) as described in detail in ref 28. It is important to keep in mind that these changes result from events in singly reduced molecules. Similar traces for the ba3 II-Met160Leu and II-Asp111Asn mutants are presented in Figures S2 and S3 of the Supporting Information. Table S2 of the Supporting Information shows the primary kf2

Article and kb2 rate constants for reaction II of Scheme 1 for cytochromes ba3 from T. thermophilus, for aa3 from bovine heart and that extracted from Figure 6 of Kobayashi et al. (25), and for aa3 from P. denitrificans. These k,T data are plotted in Figure 5. Additional data are shown in Figures 5 and 6, as described in the legends. Table S3 of the Supporting Information shows similar data for kf3 and kb3 of Tt cytochrome ba3 and its IIAsp111Asn mutation (reaction III). Values of kf3 vs T for the heme-br to heme-a3o for ba3 are plotted in Figure 6. The equilibrium constants K2 and K3 (see Scheme 1) are largely independent of temperature as shown in Table 2. Moreover, single-temperature estimates of the rate of the CuAr þ a(b)o Scheme 2: Environment of the CuA Site in Cytochrome ba3 Showing Sites of Mutations (the Small Black Sphere Represents a Hydrogen Atom)

Biochemistry, Vol. XXX, No. XX, XXXX

E

f CuAo þ a(b)r reaction in the bovine enzyme (22, 43-45) report values of kobs near 2  104 s-1, consistent with the data shown in Table 2. The ba3 K-path mutant II-Glu15Ala (8) enzyme is inactive and appears unable to transfer electrons from heme-br to heme-a3o as illustrated in Figure S4 of the Supporting Information, which compares the normal spectrum of the fully reduced, wild-type enzyme with the partially reduced spectrum of the II-Glu15Ala mutant that occurs on addition of sodium dithionite. After several hours under anaerobic conditions, however, the mutant enzyme will become fully reduced. The purposes of examining this form of the enzyme in the pulse radiolysis experiment were to query whether something unexpected might occur at short times and to determine if CuAr to heme-bo ET is affected. Figure 7 shows clearly that electron transfer from CuAr to heme bo is normal, while that from heme-br to heme-a3o does not occur during the time of our measurement. Activation Parameters. The kf2,T and kb2,T data sets of Table S2 of the Supporting Information and the kf3,T and kb3,T data sets of Table S3 of the Supporting Information are unique in providing detailed information about the energetics of ET in cytochrome c oxidases when analyzed by the methods described in Appendix B of the Supporting Information. We have focused on forward reactions II and III of Scheme 1 and have not analyzed the corresponding back reactions, although the equilibrium constants kf/kb have been determined. Starting with the reaction, CuAr þ a(b)o f CuAo þ a(b)r, the rate constant for this reaction in the three different proteins differs by some 5-fold at 298 K, ∼4000 s-1 for Tt ba3 to ∼21000 s-1 for Pd aa3. A fourth enzyme,

FIGURE 1: Oxidized minus reduced optical absorption spectra of wild-type cytochrome ba3 from Thermus thermophilus, II-Met160Leu, and II-Asp111Asn showing arrows at the wavelengths at which kinetic data were collected. Subsequent to radiolysis (see Materials and Methods), an increase in absorbance at 560 nm indicates the appearance of heme-br, while an increase at 445 nm indicates appearance of heme-a3r. Estimates of extinction coefficients are given in Table 1.

F

Biochemistry, Vol. XXX, No. XX, XXXX

Farver et al.

FIGURE 3: Continuous wave X-band EPR spectra of native and the

FIGURE 2: Near-IR optical absorption spectra of native and mutant forms of fully oxidized cytochrome ba3 oxidase from T. thermophilus. (bottom) Recombinant, wild-type enzyme; (middle) the II-Met160Leu mutant; (top) the II-Asp111Asn mutant. Spectra are normalized at the 535 nm spectral maximum of the oxidized form of each protein and offset to emphasize differences in the shape and position of the long wavelength absorption band. Redox changes in the CuA site were routinely observed at 750 nm because of the greater sensitivity of the photomultiplier tube. At this wavelength, the extinction coefficients are approximately the same as at 790 nm (Table 1).

for which temperature dependent kinetic information is unfortunately not available, that can be considered part of this group, is the cytochrome aa3 from R. sphaeroides, having a forward rate constant of 93000 s-1 at pH 7.5 and 298 K (41). Transition State Theory. Laidler (46) suggests that if Ea g 0.2 eV, the reaction under consideration “...almost certainly [involves] the breaking of primary chemical bonds... .” The data in Table 3 are ambiguous in this regard because they straddle this border, ranging from 0.14 to 0.26 eV. However, assuming that the electron transfer mechanism is essentially the same for each enzyme, it is reasonable to suggest that neither bond making nor bond breaking is occurring. Values of ΔS* (with κ set to unity) are very small and fall in a narrow range (∼-80  10-5 to -130  10-5 eV/K; -18 cal/(mol K) to -30 cal/(mol K)). The pre-exponential term, AA, however, varies from ∼106 to 108 s-1. None of these parameters correlate with the differences in kf2 found with the different enzymes, but ΔG* at T correlates nicely with ln kf2, as shown in Figure S5, panel A in the Supporting Information, and indicates that the full range in the kf2 values is found in a ΔΔG* range of ∼0.04 eV (0.92 kcal/mol) at 298 K. Extrapolating the plot of ln kf2 vs ΔG*(298) to an intersection of the parallel line at ln(9.3  104 s-1) yields an estimate for ΔG* of ∼0.37 eV for the reaction, CuAr þ a(b)o f CuAo þ a(b)r in the Rs enzyme (41). Compensation between ΔH* and ΔS*. As already noted, there is a divergence of ΔH* and TΔS* values (Table 3), such that neither alone shows a trend that would explain the differences between the rate constants (at T) for the three enzymes. The good

mutant forms of fully oxidized cytochrome ba3 oxidase from T. thermophilus. Conditions of recording were: microwave frequency was 9.6 GHz at 5 mW power, modulation field was 5 gauss at 100 kHz, temperature of recording was 10 K. See Supporting Information for a compilation of relevant EPR parameters.

correlation with ΔG*, however, points to an underlying compensation between ΔH* and TΔS*, such that when ΔH* becomes smaller, for whatever reason, the negative ΔS* becomes larger, as is most clear in the case of the Bov(OE) sample (Table 3). Indeed, isokinetic plots of ΔH* vs ΔS* appear linear (not shown), and similar behavior has previously been observed in studies of electron transfer with azurins (47). Similar correlation is seen in the Marcus parameters (see Discussion), and the possible origin of these effects is discussed in Appendix C of the Supporting Information. Marcus Theory. Experimentally, Dutton and co-workers suggest that λexptl can only be extracted from experimental rate data when both T and ΔEo are systematically varied (48), which may be true for the electron transfer reactions of photosynthesis but cannot be generally true, as per eq 3. Certainly, one cannot use the Marcus equation if the driving force varies with temperature. Features of the Marcus interpretation of the pre-exponential term and of the exponential term differ from those of transition state theory. AM is numerically different from AA due to the inclusion of 1/T-2 in the pre-exponential term and, for the same reason, its units are K1/2 s-1. As expected, however, (Table 4), AM, like AA, varies over a range of ∼102. The Marcus activation energy, EaM, which is actually a free-energy in Marcus theory (39), differ very little, ∼0.01 eV, from EaA of transition state theory and show the same trend among the different enzymes. The extracted values of λ are ∼1 eV, except for the combined data for the bovine enzyme, which is somewhat smaller (∼0.7 eV). The values of Hab vary within the range of ∼2  10-5 to ∼24  10-5 eV. Clearly, neither λ nor Hab correlate with ln kf2 (see below) Earlier projections of the reorganization energy of this reaction in cytochrome c oxidases were made by Ramierez et al. (49), who suggested a value for this pathway of between 0.15 to 0.5 eV being lower than typical values of between ∼0.7 to 1.3 eV for internal electron transfer (see ref (49). and references therein. Farver et al. (26) suggest a value of 0.4 eV, and Brzezinski (44) also

Article

Biochemistry, Vol. XXX, No. XX, XXXX

G

FIGURE 4: Typical kinetic traces recorded from solutions of cytochrome ba3 of T. thermophilus subsequent to pulse radiolysis. Traces were

recorded at 750, 560, and 445 nm each over two time domains, left to right, for the native enzyme. (top panels) 750 nm: 107 μM protein; 5 mM 1-MNA; 10 mM phosphate pH 7.5; 0.05% DDM; Ar saturated; 1.5 μs pulse width; 24.9 C. (center panels) 560 nm and (bottom panels) 445 nm: 8.6 μM protein and 0.05 μs pulse width; 3.0 cm optical pass. See Materials and Methods for experimental details.

suggests a value of ∼0.3 eV, which would correspond to an EaM of ∼0.08 eV, a rate nearly independent of temperature. The results shown in Figure 5 and Table 4 are thus consistent with a significantly greater value of λ than earlier reports. Given the closeness of the four structures (Scheme 3) and more reasonable error limits, λ is likely to fall in the range of 1 ( 0.15 eV for each of the enzymes studied here. The reaction br þ a3o f bo þ a3r, reaction III of Scheme 1, has been observed by us only in the B-type enzyme (28) and not in the previously investigated A-type enzymes (26, 27). The Arrhenius and Eyring parameters of this reaction are presented in Table 5 (with κ = 1), while the Marcus parameters are shown in Table 6. As expected for this slower rate of reaction, the activation energy of kf3 is significantly larger: 0.37 eV vs 0.14 to 0.26 eV for reaction II. The activation entropy of kf3, ΔS* ∼ -70  10-5 eV/K, is, however, somewhat less negative than ΔS* reported for

kf2 ∼-108  10-5 eV/K (Table 5). The reorganization parameter is, however, significantly larger: ∼1.5 vs ∼1.0 for kf2, and the coupling energy, Hab ∼27  10-5 eV, falls on the high side of ∼24  10-5 eV, for kf2 of the Pd enzyme (Table 6). Interpretations of the extrathermodynamic parameters will be presented in the Discussion. DISCUSSION The elementary rate constants for reactions II and III report on the velocity of a unique ET event, and the effect of temperature on these rates reveals information about their energetics. As may be inferred from Schemes 3 and 4, an electron leaves a donor moeity (D), passes through a bridging structure (B), and ends up on the acceptor portion (A), designated D-B-A (see ref 50). Analysis of the kf2, T and kf3, T data in terms of reaction rate

H

Biochemistry, Vol. XXX, No. XX, XXXX

Farver et al.

FIGURE 6: Plots of kf3 vs T for reaction III of Scheme 1 (heme-br þ FIGURE 5: Plots of kf2 vs T for reaction II of Scheme 1 (CuAr þ heme-

b f CuA þ heme-a(b) ). From the top of the figure down as indicated by the annotation: open squares, Pd cytochrome aa3; open circles, bovine cytochrome aa3 obtained from the laboratory of Dr. O. Einarsdottir; narrow, open diamonds were extracted from Figure 6 of K. Kobayashi et al. assuming K2 = 3.4 at all temperatures (25), and open diamonds, Tt cytochrome ba3. The crosses are kf2,T pairs for the Tt ba3 II-Asp111Asn mutant, the single solid circle at 285 K represents the II-Met160Leu mutant, and the single solid inverted triangle represents the IIGlu15Ala mutant. The solid lines are fits to the Arrhenius equation k = AA exp(-Ea/kBT) with the values of AA and Ea presented in Table 3. In the case of the data from the bovine enzyme, all the points shown were used in the fitting. In the case of the Thermus enzyme, only the open diamond points were fitted; not all data points of Table S2 of the Supporting Information were used in the fitting. Conditions of recording are as described in the legend to Figure 4 and in ref 25. o

o

r

theories and available structures can provide significant detail about how the electron is actually transferred in these reactions. Indeed, there are numerous ramifications of the data and energetic analyses that merit discussion: From transition state theory (Eyring), one may find evidence for or against bond making and breaking. From Marcus theory, questions arise about gating, quantum mechanical coupling and reoganization energies, and ET distance. It is important to consider the compensation between the pre-exponential and exponential terms of the equations used to obtain transition state parameters. Data from the new mutants in the CuA cluster and in the proton mediating K-channel are important in understanding reactions II and III of Scheme 1. The more complicated reaction III of Scheme 1 merits comment in terms of transition state theories and in terms of recent evidence for superfast (nanosecond) ET between low-spin and high-spin

heme-a3o f heme-bo þ heme-a3r) for T. thermophilus cytochrome ba3. Conditions of recording are as described in the legend to Figure 4. Open diamond symbols correspond to wild-type ba3, the cross at 267 s-1, 296 K corresponds to this reaction in the II-Asp111Asn mutant, and the closed circle at 148 s-1, 285 K corresponds to the II-Met160Leu mutant at 285 K. The annotation indicates line fit parameters to the full Marcus expression, our eq 3.

hemes (see below). The implications of our observations are discussed along these lines. Electron Transfer from the CuA Center to Low-Spin Heme (kf2). “Gated” or “Pure” Electron Tunneling? The possible occurrence of “gated” electron transfer processes in proteins has been of concern to both theoreticians and experimentalists studying long-range electron transfer (51-57). A priori unknown, and, in some cases, possibly unknowable, is whether an electron transfer event occurs by a rapid tunneling event that follows a relatively slow, thermally induced structural change, e.g., D - B - AhðD - B - AÞ0 f Dþ - B - A in which the transition state parameters reflect the structural change D - B - AhðD - B - AÞ0 rather than electron tunneling. Hoffman and Ratner (55) state “...in many (most?) instances, the measured time course of a gated ET reaction is predicted to be indistinguishable from a reaction without gating.” Davidson (54), while also recognizing the possible ambiguity of the extrathermodynamic parameters, suggests that they may be of diagnostic use. For example, when analyzing an electron transfer reaction, if Hab is found to be

Article Scheme 3: Proposed Electron Tunneling Pathway for A- and B-Type Cytochrome c Oxidases (Small Black Spheres Identify the Two N-H---O Entities in the Structure; For Clarity, the D-Ring Propionate Has Been Omitted)

g990  10-5 eV (or 80 cm-1), electron transfer is likely preceded by a bond making or breaking process that is more properly characterized by the ΔH* and ΔS* of transition state theory. Our data of Tables 3-6 indicate a much weaker coupling, pointing toward “pure” electron tunneling. Commenting on the possible ambiguity in interpretations of k, T data, Devault (51) suggested that one should examine the intervening protein structure for evidence of interactions that might facilitate D-B-A electron tunneling. This now includes application of search programs that ferret out possible D-B-A pathways, compute relative efficiencies and interactions, and predict observed rate constants (58-61). In the case of cytochrome c oxidases, such work from Stuchebrukhov’s and Onuchic’s laboratories focused on the structure in Scheme 3 as the D-B-A arrangement for reaction II of Scheme 1. The four enzymes studied here possess a structure in which the covalent parts of donor CuA and of acceptor heme a(b)o are bridged by two N-H---O hydrogen bonds and three covalent bonds along the protein’s backbone, as already noted (49). Small differences between protein/ water structural fluctuations, electronic coupling, and effective distance modulation may account for the differences in kf2 observed for the four cytochrome c oxidases, see ref 62. It is appropriate to consider if any of our derived transition state parameters are inconsistent with pure electron tunneling. Reorganization Energy, λ, and Electronic Coupling, Hab. Hab and λ are linked in Marcus eq 3 because the unknown portion of the pre-exponential factor consists of Hab2/(λ)-2. These terms can be resolved only when ΔEo is known or it is known that ΔEo , λ (see Appendix B of the Supporting Information). It is the numerical value of Hab2/(λ)-2 that, in conjunction with the exponential term, determines the numerical rate constant; Hab and λ are not independent of each other. For reaction II, the values of λ range from ∼0.7 to 1.1 eV (Tables 4 and 6), the values of Hab range from ∼2.2  10-5 eV to ∼24  10-5 eV (Tables 4 and 6),

Biochemistry, Vol. XXX, No. XX, XXXX

I

Scheme 4: Structure Intervening between the Heme-b and Heme-a3 Sites

from which values of Hab2/(λ)-2 range roughly a hundred-fold from ∼4.8  10-10 to ∼576  10-10 (eV3/2) when λ is taken at 1 eV. It is the latter numbers that are important in comparing experimental rate constants to computed rate constants. Hab values for the structure in Scheme 3 have been computed from theoretical considerations. For this structure, Stuchebrukhov and co-workers (20) computed Hab ∼ 2  10-7 eV for the His-204 pathway in the Pd enzyme, and Onuchic and co-workers (21) computed values of Hab ranging from ∼3  10-7 eV to ∼10-6 eV, some hundred-fold smaller than those obtained from experiment. Nevertheless, both calculations were carried out in the spirit of proofs-of-principle, and theorists continue to develop more powerful theories to understand electron transfer reactions (20, 62-67). An important contribution of the present work is to provide experimental benchmarks for computational studies. Diabatic or Adiabatic? Marcus’ theory for long-range electron transfer applies to diabatic processes (38, 51, 68, 69). For readers who may be confused by the application of these words of Greek origin (διR-βRινεν, through-pass, or R-διRβRινεν, not-through-pass) to reaction rate theories, K. J. Laidler has written a historical perspective (70). The word adiabatic in electron transfer theory was initially described in terms of eq 5 k ¼ KðkB T=hÞ expð - ½ðΔEo þλÞ2 =4λkB T

ð5Þ

where κ is a transmission coefficient, ((see pp 50-55 of ref 71), not to be confused with the κ of transition state theory (see ref 72). Roughly speaking, the Marcus κ reflects the “allowedness” for electron transfer at the intersection of two potential energy surfaces, e.g. of D and A. When κ = 1, the electrons are following the slowly moving nuclei, electron transfer occurs each time the system achieves the intersection, there is a gradual, continuous change in the electronic quantum state, this being the definition of adiabatic ET (ref 69; see also p6ff of ref 73). Diagnostic of the adiabatic condition, therefore, is a strong coupling between D and A, i.e., Hab . kBT (71); see ref 74 for a deeper analysis. The word diabatic (often written with the Latin prefix nonmodifying the Greek word, adiabatic, i.e., nonadiabatic) applies when κ , 1, D and A are very weakly coupled, i.e., Hab , kBT; electron transfer then occurs by a “sudden” transition to a different potential energy surface, i.e., tunneling. For a diabatic process, the Marcus equation is written as our eq 3. On the basis of the data of Tables 4 and 6, where the largest value of Hab is ∼24  10-5 eV , (kB  298 = 0.026 eV), both reactions II and III of Scheme 1 occur in the diabatic regime. This is also in accord with Devault’s (51) first criterion in favor of electron tunneling, namely that from eq 1, AA , 1013 s-1, as is shown in Tables 3 and 5. By these criteria, our data support the conclusion that both reactions II and III of Scheme 1 are likely to be pure electron tunneling events.

J

Biochemistry, Vol. XXX, No. XX, XXXX

Farver et al.

FIGURE 7: Comparative behavior of wild-type cytochrome ba3 and the K-path mutant II-Glu15Ala in reactions II and III of Scheme 1. Presentation is similar to that in Figure 4. Top panels (red-colored traces) were recorded at 560 nm and indicate changes in heme-b. The time ranges from 350 μs in the upper left panel to 4 ms in the upper right panel. Bottom panels (black-colored traces) were recorded at 445 nm and indicate changes in heme-a3. The time ranges from 400 μs in the lower left panel to 4 ms in the lower right panel. Traces are scaled to maximum absorbance change at 0.3 ms and 560 nm. Experimental conditions were as described in the legend to Figure 4. The enzyme concentration was 8.6 μM, the pulse width was 0.05 μs, and the 3.0 cm pass optical arrangement was used. Estimates of extinction coefficients are given in Table 1.

Because the CuAr þ a(b)o f CuAo þ a(b)r reaction depends on electron-tunneling, the structure of Scheme 3 may represent the “bridge” of D-B-A. Strong support for this idea comes from disruption of this structure in the Rs enzyme (41) by introducing the CuA mutation, II-His260Asn (corresponding to His157 in ba3, see Scheme 2). This mutation causes kf2 to drop ∼2000-fold, from 90000 s-1 to 45 s-1. Along with the high conservation of structure (Scheme 3), these observations point to an essential role of this structure in the function of the enzyme; see, however, refs (75, 76). Compensation between Pre-Exponential and Exponential Marcus Terms. As might be expected from the underlying similarity of Arrhenius and Marcus equations (eqs 2 and 3), one should also find compensation between the pre-exponential and the exponential terms of the Marcus equation. Indeed, P this can be ,298) vs. the values of visualized by plotting ln(k f2 M* that are P given in Table 4. M* is equal to (|EaM/kBT| þ |2 ln(AM)|), and such plots are linear as shown in Figure S5 panel B (see also Appendix C of the Supporting Information). Understanding

the deeper meaning of structural differences between the four enzymes that cause this behavior may be very useful in identifying factors that influence long-range electron transfer in proteins (47). The cytochrome c oxidases are good candidates for such probing. Donor-Acceptor Separation Distance. Where donor and acceptor begin and end (?) is a general problem in long-range electron tunneling studies (refs (75, 77, 78). and references therein). This is particularly true in the case of metal centered DA systems, where there is a natural tendency to localize the electronic wave functions on the metal ions, whereas for the organic cofactors, distance between the nearest van der Waals’ surfaces is the more logical choice (75). Examination of the structures in Scheme 3 does not reveal obvious structural features, other than Cu and Fe, where the electron might reside prior to or subsequent to ET. We consider D (A˚) as an arbitrary distance and forego the widely used definition D = r - ro, r being the distance between the van der Waals’ surfaces of D and A, and ro corresponding to a van der Waals’ approach of ∼3.6 A˚.

Article Fitting the kf2,T data to eq 4 (equal to eq 10B of Appendix B of the Supporting Information) yields the dimensionless βD parameter, which varies from 9.32 to 13.9 with an average value of 11.8 for the three enzymes studied (Table 4). Neither β nor D is defined, but D is often assigned as “the” distance between the donor and acceptor moieties, making β an attenuation coefficient that reflects the nature of the intervening material. For a vacuum, β ∼ 2.8 (A˚-1), while in a random polypeptide matrix, β ∼ 1.5 (A˚-1), but β is sensitive to the specific chemical and structural nature of the intervening medium (see ref 79 and references therein), which can have large effects on its value. The distance between Cu1 of CuA and Fe of the low-spin heme is ∼19.5 A˚, from which β ranges from ∼0.5 to 0.7 (A˚-1), suggesting unusually efficient ET over a great distance. Taking a different approach, we set β = 1.5 A˚-1, typical of random polypeptide matrices (75, 80); then D (= βD/1.5) ranges from 6.4 to 9.2 A˚ with average D = 7.9 A˚. This distance corresponds to the 8.0 A˚ between NE2 of II-His157 (ba3 numbering) and OA2 of the C-ring propionate and encompasses the two N-H---O hydrogen bonds shown in Scheme 3, making it possible that this region of the proposed electron “wire” plays a special role in determining the overall rate of ET in reaction II. Mutations at the CuA Center. What do CuA mutations reveal about the role of CuA in this reaction? Previous studies (40, 41) and the work reported here with CuA-mutants show that the turnover numbers of these enzymes, including the wildtype forms, are small compared to the kf2/kb2 mediated equilibration rates observed for the reversible CuAr þ a(b)o f CuAo þ a(b)r reaction (43, 81-83). That the forward reaction cannot be the rate-limiting step in catalysis has important implications for understanding electron and possibly proton transfer. The primary results reported here for the two CuA mutants in ba3, II-Met160Leu and II-Asp111Asn, indicate significant changes in the optical and EPR spectra. Neither mutant shows large deviations from the native values of kf2, kb2, or K2 over a wide temperature range (see Figure 5). For the II-Met160Leu mutation, the activity is lowered to ∼83% of wild-type, while in the IIAsp111Asn mutation, the activity is lowered to ∼30% of wildtype. Detailed studies by Millet, Durham, Ferguson-Miller, and co-workers (41) with the Rs enzyme parallel ours with comporable results. They used a covalently linked Ru(II)-trisbypyridinecytochrome c complex to effect rapid and specific deposition of an electron on the CuA center of the Rs enzyme, demonstrating unequivocally the electron transfer path cr þ CuAo f co þ CuAr in both wild-type enzyme and two interesting CuA mutations; temperature dependencies were not measured. Nevertheless, kf2 was determined at 298 K = 93000 s-1 with a driving force of 0.046 eV. In the II-Met264Leu (Rs numbering) mutation, kf2 is lowered to 4000 s-1 and ΔEo is decreased to -0.072 eV for a total change in the redox potential of (0.046 þ 0.072 =) 0.118 eV, which accounts for the observed decrease of their kf2 at 298 K. For the II-Met160Leu ba3 mutant, Ledesma et al. (34) report 0 Eo values for the water-soluble portion of the ba3 subunit II of 0.29 eV for the wild-type CuA and 0.35 eV for the II-Met160Leu mutant, an increase of ∼0.06 eV. Taking the λ and AM from Table 4 and changing ΔEo by this amount yields values of kf2 from eq 3 such that kf2calcd (mutant) = 0.8kf2calcd (wild-type), while from experiment, kf2exptl (mutant) =0 0.6kf2exptl (wildtype) at 285 K. Although the change in Eo is small by comparison with that observed for the Rs enzyme, the results are consistent with eq 3 being operative in reaction II.

Biochemistry, Vol. XXX, No. XX, XXXX

K

Recent Results with Photolytic Reduction of ba3. While this paper was in preparation, a communication from the Wikstr€om group appeared (19) making an elegant case for “oxidative reactivation (84)” of cytochrome ba3. However, that portion of their study dealing with the “as-isolated” form of the enzyme showed rather unusual behavior when compared to what is reported here. Thus, it is claimed that when the Ru(II)-tris-bipyridyl/analine photoreduction system was used to inject single electrons into ba3, the formation of CuAr is stated to be complete in a submicrosecond time range and hemes-b and a3 were maximally reduced in ∼10 μs compared, respectively, to the ∼140 μs and ∼1 ms phases reported here. This was followed by a slower phase, 230 μs, which involved partial oxidation of both hemes, perhaps due to reduction of CuBo. At this time, we are unable to rationalize our differences with this report. However, it is important to make direct measurements of CuA reduction by the Ru(II)/aniline system as the positively charged Ru(II) salts react quite slowly with the electrostatically neutral cytochrome c552 binding site of ba3 (85, 86). Electron Transfer from the Low-Spin Heme to the Binuclear Center (kf3). The results raise important questions about this reaction: What type of ET is occurring, gated or “pure” tunneling? Is the ET process diabatic or adiabatic? Why is ET not observed in the K-path mutation, II-Glu15Ala? Why is it not observed in pulse radiolysis of the A-type enzymes? Is there a rapid back reaction from a3r to heme-bo? The extrathermodynamic results presented in Tables 5 and 6 meet the criteria discussed under reaction II of Scheme 1 for diabatic electron transfer, notably that AA is ,1013 s-1 and Hab , kBT, and there exists an intervening medium that may provide electronic overlap between heme-br and heme-a3o, which is shown in Scheme 4. In this structure, the through-space distance between the two Fe-atoms is 13.9 A˚. One possible ET pathway consists of two Fe-N coordinate covalent bonds and 15 covalent bonds (21). The experimental βD parameter is 9.35 (Table 6). Using 13.9 A˚ as “the” distance between the Fe of heme-b and the Fe of heme-a3 yields β = 0.67 (A˚-1), suggesting an unusually efficient pathway, whatever that pathway may be. Assuming the protein structure between the two Fe centers is representative of random polypeptide, setting β = 1.5 and dividing into βD = 9.35 yields an effective distance of ∼6.2 A˚. Unfortunately, there are no obvious structural features separated by this distance. An alternative for heme-to-heme tunneling through covalent bonds is the possibility of through-space or direct-contact tunneling from methyl group CMD on heme-b to methyl group CMA on heme-a3 (5.2 A˚ in ba3) as shown in Scheme 4. This path is considered by the Stuchebrukhov (20) and Onuchic (21) groups to be the more likely tunneling pathway, and in the recently reported structure of cytochrome cbb3, the low-spin heme-b and the high-spin heme-b3 are in van der Waals’ contact (87), suggesting a direct heme-to-heme ET in that enzyme. Differences in the Protonation State of the Dinuclear Center. In the experiments described here (indeed for all experiments of this type), the electron from heme-br is moving to a CuB-Fea3 active site of unknown composition, charge, and homogeneity. During turnover of the enzyme, the demand for electrons and protons will vary with the chemistry occurring in the binuclear center, which must be regulated so that the scalar chemistry remains coupled to the proton pumping steps (88, 89). Because it is reasonable that inward fluxes of electrons and protons are chemically controlled (see refs (8, 17) for specific suggestions in

L

Biochemistry, Vol. XXX, No. XX, XXXX

this vein), it is likely that we observe electron tunneling from heme-br to heme-a3o only when the latter is chemically “ready” to receive the electron. The rate and order of heme-to-heme electron transfer during catalysis is probably determined by the relative affinities of each reaction intermediate for electrons and protons. Giuffre et al. (90) have measured the rate of reoxidation of reduced ba3 by O2 finding apparent first-order kobs ranging from ∼100 s-1 at 4.5 C to ∼300 s-1 at 25 C at 10 μM O2. These rates reflect the slow process in the oxidative cycle, being a factor of 2-3 less than k3obs at T (Table 2), suggesting that kf3 may be a rate limiting process. The Proton Uptake Mutant. The K-channel mutation, IIGlu15Ala, in ba3 blocks reaction III from a distance of ∼24 A˚. How can this be rationalized? Both wild-type and mutant forms of the enzyme are purified in the fully oxidized form, as indicated by optical absorption spectra (Figure S4 of the Supporting Information). The wild-type enzyme is very rapidly reduced by dithionite and oxidized by O2, but the mutant enzyme is very slowly reduced by dithionite and very slowly oxidized by O2 (8); both of the latter processes are unlikely related to the normal reaction of the enzyme and therefore are probably caused by the mutation. Moreover, as shown in Figure 7, the heme-a3o of the mutant enzyme does not rapidly receive and subsequently discharge an electron in the few milliseconds in which the wild-type enzyme receives an electron from heme-br. A simple reason for this behavior may be that the active site of the spuriously reoxidized II-Glu15Ala mutant differs from the normally oxidized enzyme by possessing one fewer protons in the binuclear center; this lowers its electron affinity such that it is unable to accept the electron from heme-br. The wild-type enzyme, having a functioning K-path, would be proton-sufficient, with a higher electron affinity, thus ready to accept an electron from heme-br, which is consistent with the suggested need for electric neutrality in the binuclear center (91). A simple explanation for the observation that the singly reduced, A-type enzymes, in our experimental conditions, do not undergo ET from heme-ar to heme-a3o is the absence of this enabling proton at the binuclear center in the resting enzyme. Super Fast (ns) Electron Transfer. From the present results, a marked difference emerges with respect to the heme-a3r species discussed in the work of Morgan et al. (43) and Pilet et al. (92) on the bovine protein and of Jasaitis et al. (93) on Escherichia coli cytochrome bo3. In the current study with ba3, the singly reduced binuclear center Fea3r/CuBo must be protonated at some position and have a redox potential only slightly different from that of heme-b (94), whereas in the binuclear center of E. coli cytochrome bo3, both metals are reduced (Feo3r/CuBr, possibly with CO bound to CuBr), which realistically may be considered to be strongly reducing. In the case of bo3, photolytically produced o3rCuBr-CO does indeed pass an electron to heme-bo in the nanosecond time range, but this cannot be a physiologically significant process. In wild-type ba3, we observe a rather slow equilibrium electron transfer between heme-a3 and heme-b. If such a rapid ET from reduced heme-a3r to heme-bo (as observed in the COphotolyzed cytochromes aa3 and bo3) was occurring in our studies, the equilibrium heme-a3r to heme-bo would be shifted far to the left, and there would be no evidence for electron transfer from heme-br to heme-a3o in cytochrome ba3. A fair comparison of aa3, ba3, and bo3 in this context would be to initiate the reverse reaction from the same electronic/protonation state of the binuclear center. Similar considerations also apply to the elegant studies of Adelroth et al. (95) and Ching et al. (96).

Farver et al. In summary, our data show that pulse radiolysis is a remarkably clean and powerful tool with which to explore mechanisms of electron-, and potentially proton transfer, in cytochrome c oxidases. Marcus parameters were obtained from the temperature dependence of reaction (II), CuAr þ a(b)o f CuAo þ a(b)r, which support the conclusions that electron transfer between these centers occurs along a well-defined structure by a diabatic process. The most likely pathway for this process includes the atoms extending from NE2 of the CuA(1) ligand, His157, to the terminal OA2 atom of the carboxy terminus of the C-ring of heme -(a)b. Select mutations near the CuA center in ba3 have either no effect on or change the rate of reaction II by small changes in the driving force. This is consistent with previous work showing that mutations which disrupt the structure in Scheme 3 drastically affect ET (41). Marcus parameters were also obtained for reaction (III), (a)br þ a3o f (a)bo þ a3r, which are consistent with a diabatic reaction within the structure of Scheme 4 and supportive of previous hypotheses and theoretical analyses (20),(21). A mutation at the head of the proton conducting K-path (8) blocks reaction III, suggesting that an “enabling proton” must be present on the dinuclear center prior to electron tunneling; such a proton may not be present in the resting state of the A-type enzymes, which would account for our failure to observe reaction III in these enzymes under our conditions. ACKNOWLEDGMENT We thank Eran Gilad from the Hebrew University of Jerusalem for his excellent support in running the accelerator, Dr. Olof Einarsdottir for providing bovine cytochrome aa3, and Dr. Bernd Ludwig for providing cytochrome aa3 from Paracoccus denitrificans. We thank Professor Daniella Goldfarb for generously confirming the EPR properties of one of our samples. J.A.F. thanks Drs. Frank Millet, Alexei Stuchebrukhov, Jay Winkler, James McCusker, Louis Noodleman, Leslie Dutton, Ilya Balabin, and Bruce Palfey for valuable discussions and critical commentary on the manuscript. SUPPORTING INFORMATION AVAILABLE Appendix A justifying the kinetic assignments in Scheme 1, Appendix B describing how transition state parameters were obtained, Appendix C discusses the origin of compensation between pre-exponential and exponential terms; five figures showing EPR, kinetic traces, optical P spectra of mutant proteins, and plots of ln(k) vs ΔG* and *; and three tables containing, EPR parameters of native and mutant CuA proteins, primary kf2 values, and primary kf3 values. This material is available free of charge via the Internet at http://pubs.acs.org. REFERENCES 1. Keilen, D. (1966) The History of Cell Respiration and Cytochrome, University Press, Cambridge, UK. 2. Hosler, J. P., Ferguson-Miller, S., and Mills, D. A. (2006) Energy transduction: proton transfer through the respiratory complexes. Annu. Rev. Biochem. 75, 165–187. 3. Richter, O.-M. H., and Ludwig, B. (2003) Cytochrome c oxidase: structure, function, and physiology of a redox driven molecular machine. Rev. Physiol. Biochem. Pharmacol. 147, 47–74. 4. Pereira, M. M., Santana, M., and Teixeira, M. (2001) A novel scenario for the evolution of haem-copper oxygen reductases. Biochim. Biophys. Acta 1505, 185–208. 5. Hemp, J., Han, S., Roh, J. H., Kaplan, S., Martinez, T. J., and Gennis, R. B. (2007) Comparative genomics and site-directed mutagenesis support the existence of only one input channel for protons in the

Article

6.

7.

8.

9.

10. 11. 12. 13. 14. 15. 16.

17.

18.

19.

20.

21. 22.

23.

24.

25.

26.

27.

28.

29.

C-family (cbb3 oxidase) of heme-copper oxygen reductases. Biochemistry 46, 9963–9972. Zimmermann, B. H., Nitsche, C. I., Fee, J. A., Rusnak, F., and Munck, E. (1988) Properties of a copper-containing cytochrome ba3: a second terminal oxidase from the extreme thermophile, Thermus thermophilus. Proc. Natl. Acad. Sci. U.S.A. 85, 5779–5783. Luna, V. M. M., Chen, Y., Fee, J. A., and Stout, C. D. (2008) Crystallographic studies of Xe and Kr binding within the large internal cavity of cytochrome ba3 from Thermus thermophilus: structural analysis and role of oxygen transport channels in the heme-Cu oxidases. Biochemistry 47, 4657–4665. Chang, H.-Y., Hemp, J., Chen, Y., Fee, J. A., and Gennis, R. B. (2009) The cytochrome ba3 oxygen reductase from Thermus thermophilus uses a single input channel for proton delivery to the active site and for proton pumping. Proc. Natl. Acad. Sci. U.S.A. 1090, 16169–16173. White, G. F., Field, S., Oganesyan, V. S., Gennis, R. B., Yap, L. L., Katsourni, A., and Thomson, A. J. (2007) An EPR spin label study of the quinol oxidase, E. coli cytochrome bo3: a search for redox induced conformational changes. Biochemistry 46, 2355–2363. Babcock, G. T., and Wikstr€ om, M. (1992) Oxygen activation and the conservation of energy in cell respiration. Nature 356, 301–308. Brzezinski, P., and Gennis, R. B. (2008) Cytochrome c oxidase: exciting progress and remaining mysteries. J. Bioenerg. Biomembr. 40, 521–531. Siegbahn, P. E. M., and Blomberg, M. R. A. (2007) Energy diagrams and mechanism for proton pumping in cytochrome c oxidase. Biochim. Biophys. Acta 1767, 1143–1156. Kaukonen, M. (2007) Calculated reaction cycle of cytochrome c oxidase. J. Phys. Chem. B 111, 12543–12550. Popovic, D. M., Quenneville, J., and Stuchebrukhov, A. A. (2005) DFT/Electrostatic calculations of pKa values in cytochrome c oxidase. J. Phys. Chem. B 109, 3616–3626. Mattar, S., and Engelhard, M. (1997) Cytochrome ba3 from Natronobacterium pharaonis;an archaeal four-subunit cytochrome-c-type oxidase. Eur. J. Biochem. 250, 332–341. Siletsky, S. A., Belevich, I., Jasaitis, A., Konstantinov, A. A., Wikstr€ om, M., Soulimane, T., and Verkhovsky, M. I. (2007) Timeresolved single-turnover of ba3 oxidase from Thermus thermophilus. Biochim. Biophys. Acta 1767, 1383–1392. Fee, J. A., Case, D. A., and Noodleman, L. (2008) Toward a chemical mechanism of proton pumping by the B-type cytochrome c oxidases: application of density functional theory to cytochrome ba3 of Thermus thermophilus. J. Am. Chem. Soc. 130, 15002–15021. Smirnova, I. A., Zaslavsky, D., Fee, J. A., Gennis, R. B., and Brzezinski, P. (2008) Electron and proton transfer in the ba3 oxidase from Thermus thermophilus. J. Bioenerg. Biomembr. 40, 281–287. Siletsky, S., Belevich, I., Wikstr€ om, M., Soulimane, T., and Verkhovsky, M. I. (2009) Time-resolved OH f EH transition of the ba3 oxidase from Thermus thermophilus. Biochim. Biophys. Acta 1787, 201–205. Medvedev, D. M., Daizadeh, I., and Stuchebrukhov, A. A. (2000) Electron transfer tunneling pathways in bovine heart cytochrome c oxidase. J. Am. Chem. Soc. 122, 6571–6582. Tan, M.-L., Balabin, I., and Onuchic, J. O. (2004) Dynamics of electron transfer pathways in cytochrome c oxidase. Biophys. J. 86, 1813–1819. Winkler, J. R., Malmstrom, B. G., and Gray, H. B. (1995) Rapid electron injection into multisite metalloproteins: intramolecular electron transfer in cytochrome oxidase. Biophys. Chem. 54, 199–209. Farver, O., Pecht, I. (2007) Elucidation of Electron-Transfer Pathways in Copper and Iron Proteins by Pulse Radiolysis Experiments, in Progress in Inorganic Chemistry (Karlin, K., Ed.) pp 1-78, John Wiley & Sons, New York. Farver, O., Tepper, A. W. J. W., Wherland, S., Canters, G. W., and Pecht, I. (2009) Site-site interactions enhances intramolecular electron transfer in Streptomyces coelicolor laccase. J. Am. Chem. Soc. 131, 18226–18227. Kobayashi, K., Une, H., and Hayashi, K. (1989) Electron transfer process in cytochrome oxidase after pulse radiolysis. J. Biol. Chem. 264, 7976–7980. Farver, O., Einarsdottir, O., and Pecht, I. (2000) Electron transfer rates and equilibrium within cytochrome c oxidase. Eur. J. Biochem. 267, 950–954. Farver, O., Grell, E., Ludwig, B., Michel, H., and Pecht, I. (2006) Rates and equilibrium of CuA to heme a electron transfer in Paracoccus denitrificans cytochrome c oxidase. Biophys. J. 90, 2131–2137. Farver, O., Chen, Y., Fee, J. A., and Pecht, I. (2006) Electron transfer among the CuA-, heme b- and heme a3-centers of Thermus thermophilus cytochrome ba3. FEBS Lett. 580, 3417–3421. Chen, Y., Hunsicker-Wang, L. M., Pacoma, R. L., Luna, E., and Fee, J. A. (2005) A homologous expression system for obtaining

Biochemistry, Vol. XXX, No. XX, XXXX

30.

31.

32.

33.

34.

35.

36. 37. 38. 39. 40.

41.

42.

43.

44. 45.

46. 47.

48.

49.

50.

51.

M

engineered cytochrome ba3 from Thermus thermophilus HB8. Protein Expression Purif. 40, 299–318. Keightley, J. A., Zimmermann, B. H., Mather, M. W., Springer, P., Pastuszyn, A., and Fee, J. A. (1995) Molecular genetic and protein chemical characterization of cytochrome ba3 from Thermus thermophilus HB8. J. Biol. Chem. 270, 20345–20358. Liu, B., Luna, V. M., Chen, Y., Stout, C. D., and Fee, J. A. (2007) An unexpected outcome of surface engineering an integral membrane protein: improved crystallization of cytochrome ba3 from Thermus thermophilus. Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun. 63, 1029–1034. Liu, B., Chen, Y., Stout, C. D., Soltis, S. M., Doukov, T., and Fee, J. A. (2009) Combined microspectrophotometric and crystallographic examination of chemically-reduced and X-ray radiation reduced forms of cytochrome ba3 from Thermus thermophilus: structure of the reduced form of the enzyme. Biochemistry 48, 820–826. Slutter, C. E., Sanders, D., Wittung, P., Malmstrom, B. G., Aasa, R., Richards, J. H., Gray, H. B., and Fee, J. A. (1996) Water-soluble, recombinant CuA-domain of the cytochrome ba3 subunit II from Thermus thermophilus. Biochemistry 35, 3387–3395. Ledesma, G. N., Murgida, D. H., Ly, H. K., Wackerbarth, H., Ulstrup, U., Costa-Filho, A. J., and Vila, A. J. (2007) The Met axial ligand determine the redox potential in CuA sites. J. Am. Chem. Soc. 129, 11884–11885. Zhen, Y., Schmidt, B., Kang, U. G., Antholine, W. E., and FergusonMiller, S. (2002) Mutants of the CuA site in cytochrome c oxidase of Rhodobacter sphaeroides: I. Spectral and functional properties. Biochemistry 41, 2288–2297. Fleck, G. M. (1971) Chemical Reaction Mechanisms, Holt, Rinehart and Winston, New York. Leffler, J. E., Grunwald, E. (1963) Rates and Equilibria of Organic Reactions: As Treated by Statistical, Thermodynamic, And Extrathermodynamic Methods, John Wiley & Sons, New York. Marcus, R. A., and Sutin, N. (1985) Electron transfers in chemistry and biology. Biochim. Biophys. Acta 811, 265–322. Marcus, R. A., and Sutin, N. (1986) The Relation Between the Barriers for Thermal and Optical Electron Transfer Reactions in Solution. Comments Inorg. Chem. 5, 119–133. Zickermann, V., Verkhovsky, M., Morgan, J., Wikstr€ om, M., Anemuller, S., Bill, E., Steffens, G. C., and Ludwig, B. (1995) Perturbation of the CuA site in cytochrome-c oxidase of Paracoccus denitrificans by replacement of Met227 with isoleucine. Eur. J. Biochem. 234, 686–693. Wang, K., Geren, L., Zhen, Y., Ma, L., Ferguson-Miller, S., Durham, B., and Millett, F. (2002) Mutants of the CuA site in cytochrome c oxidase of Rhodobacter sphaeroides: II Rapid kinetic analysis of electron transfer. Biochemistry 41, 2298–2304. Blackburn, N. J., Ralle, M., Gomez, E., Hill, M. G., Pastuszyn, A., Sanders, D., and Fee, J. A. (1999) Selenomethionine-substituted Thermus thermophilus cytochrome ba3: Characterization of the CuA site using Se and Cu K-EXAFS. Biochemistry 38, 7075–7084. Morgan, J. E., Li, P. M., Jang, D.-J., El-Sayed, M. A., and Chan, S. I. (1989) Electron transfer between cytochrome a and Copper A in cytochrome c oxidase: a perturbed equilibrium study. Biochemistry 28, 6975–6983. Brzezinski, P. (1996) Internal electron-transfer reactions in cytochrome c oxidase. Biochemistry 35, 5611–5615. Zaslavsky, D., Sadoski, R. C., Wang, K., Durham, B., Gennis, R. B., and Millet, F. (1998) Single electron reduction of cytochrome c oxidase compound F: resolution of partial steps by transient spectroscopy. Biochemistry 37, 14910–14916. Laidler, K. J. (1972) Unconventional applications of the Arrhenius law. J. Chem. Educ. 49, 343–344. Farver, O., Jeuken, L. J. C., Canters, G. W., and Pecht, I. (2000) Role of ligand substitution on long-range electron transfer in azurins. Eur. J. Biochem. 267, 3123–3129. Gunner, M. R., Robertson, D. E., and Dutton, P. L. (1986) Kinetic studies on the reaction center protein from Rhodopseudomonas sphaeroides: the temperature and free energy dependence of electron transfer between various quinones in the QA site and the oxidized bacteriochlorophyll dimer. J. Phys. Chem. 90, 3783–3795. Ramirez, B. E., Malmstrom, B. G., Winkler, J. R., and Gray, H. B. (1995) The currents of life: the terminal electron-transfer complex of respiration [comment]. Proc. Natl. Acad. Sci. U.S.A. Collected 11949–11951. Newton, M. D. (1999) Control of electron transfer kinetics: models for medium reorganization and donor-acceptor coupling, in Elecron Transfer;from Isolated Molecules to Biomolecules (Jortner, J., Bixon, M., Eds.) pp 303-375, John Wiley & Sons, New York. Devault, D. (1980) Quantum mechanical tunneling in biological systems. Q. Rev. Biophys. 14, 387–564.

N

Biochemistry, Vol. XXX, No. XX, XXXX

52. Agmon, N., and Hopfield, J. J. (1983) Transient kinetics of chemical reactions with bounded diffusion perpendicular to the reaction coordinate: intramolecular processes with slow conformational changes. J. Chem. Phys. 78, 6947–6959. 53. Ogawa, M. Y., Wishart, J. F., Young, Z., Miller, J. R., and Isied, S. S. (1993) Distance dependence of intramolecular electron transfer across oligoprolines in [(bpy)2RuIIL•-(Pro)n-CoIII(NH3)5]3þ, n = 1-6: different effects for helical and nonhelical polyproline II structures. J. Phys. Chem. 97, 11456–11463. 54. Davidson, V. L. (1996) Unraveling the kinetic complexity of interprotein electron transfer reactions. Biochemistry 35, 14035–14039. 55. Hoffman, B. M., and Ratner, M. A. (1987) Gated electron transfer: When are observed rates controlled by conformational interconversion? J. Am. Chem. Soc. 109, 6237–6243. 56. Engstrom, G., Xiao, K., Yu, C.-A., Yu, L., Durham, B., and Millett, F. (2002) Photoinduced electron transfer between the Rieske ironsulfur protein and cytochrome c1 in the Rhodobacter sphaeroides cytochrome bc1 complex. J. Biol. Chem. 277, 31072–31078. 57. Davidson, V. L. (2008) Protein control of true, gated, and coupled electron transfer reactions. Acc. Chem. Res. 41, 730–738. 58. Beratan, D. N., Onuchic, J. O. (1996) The protein bridge between redox centers, in Protein Electron Transfer (Bendall, D. S., Ed.) pp 23-42, BIOS Scientific, Oxford. 59. Siddarth, P., and Marcus, R. A. (1993) Correlation between theory and experiment in electron-transfer reactions in proteins: electronic couplings in modified cytochrome c and myoglobin derivatives. J. Phys. Chem. 97, 13078–13082. 60. Beratan, D. N., Betts, J. N., and Onuchic, J. O. (1991) Protein electron transfer rates set by the bridging secondary and tertiary structure. Science 252, 1285–1288. 61. Beratan, D. N., and Balabin, I. A. (2008) Heme-copper oxidases use tunneling pathways. Proc. Natl. Acad. Sci. U.S.A. 105, 403–404. 62. Balabin, I., and Beratan, D. N. (2008) Peristence of structure over fluctuations in biological electron-transfer reactions. Phys. Rev. Chem. 101, 158102-1–158102-4. 63. Small, D. W., Matyushov, D. V., and Voth, G. A. (2003) The theory of electron transfer reactions: What may be missing? J. Am. Chem. Soc. 125, 7470–7478. 64. Matyushov, D. V. (2007) Energetics of electron-transfer reactions in soft condensed media. Acc. Chem. Res. 40, 294–301. 65. Milischuk, A. A., Matyushov, D. V., and Newton, M. D. (2006) Activation entropy of electron transfer reactions. Chem. Phys. 324, 172–194. 66. Larsson, S. (1998) Electron transfer in proteins. Biochim. Biophys. Acta 1365, 294–300. 67. Ungar, L. W., Newton, M. D., and Voth, G. A. (1999) Classical and quantum simulation of electron transfer through a polypeptide. J. Phys. Chem. B 103, 7367–7382. 68. Bolton, J. R., Archer, M. D. (1991) Basic Electron Transfer Theory, in Advances in Chemistry pp 7-23, American Chemical Society, Washington, DC. 69. Barbara, P. F., Meyer, T. J., and Ratner, M. A. (1996) Contemporary issues in electron transfer research. J. Phys. Chem. 100, 13148–13168. 70. Laidler, K. J. (1994) The meaning of “adiabatic”. Can. J. Chem. 72, 936–938. 71. Bixon, M., Jortner, J. (1999) Electron transfer;from isolated molecules to biomolecules, in Electron Transfer;From Isolated Molecules to Biomolecules (Jortner, J., Bixon, M., Eds.) pp 734, John Wiley & Sons, New York. 72. Eyring, H. (1962) The transmission coefficient in reaction rate theory. Rev. Mod. Phys. 34, 616–619. 73. Astruc, D. (1995) Electron Transfer and Radical Processes in Transition-Metal Chemistry, VCH Publishers, New York. 74. Voorhis, T. V., Kowalczyk, T., Kaduk, B., Wang, L.-P., Cheng, C.-L., and Wu, Q. (2009) The diabatic picture of electron transfer, reac tion barriers, and molecular dynamics. Annu. Rev. Phys. Chem. 61, 149–170. 75. Moser, C. C., Anerson, J. L. R., Dutton, P. L. (2010) Guidelines for tunneling in enzymes. Biochim. Biophys. Acta in press.

Farver et al. 76. Moser, C. C., Page, C. C., and Dutton, P. L. (2006) Darwin at the molecular scale: selection and variance in electron tunnelling proteins including cytochrome c oxidase. Phil. Trans. R. Soc. 361, 1295–1305. 77. Moser, C. C., Chobot, S. E., Page, C. C., and Dutton, P. L. (2008) Distance metrics for heme protein electron tunneling. Biochim. Biophys. Acta 1777, 1032–1037. 78. Gray, H. B., and Winkler, J. R. (2003) Electron tunneling through proteins. Q. Rev. Biophys. 36, 341–372. 79. Beratan, D. N., and Skourtis, S. S. (1998) Electron transfer mechanisms. Curr. Opin. Chem. Biol. 2, 235–243. 80. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S., and Dutton, P. L. (1992) Nature of biological electron transfer. Nature 355, 796–802. 81. Hill, B. C. (1993) The sequence of electron carriers in the reaction of cytochrome c oxidase with oxygen. J. Bioenerg. Biomembr. 25, 115–120. 82. Hosler, J. P., Fetter, J., Tecklenburg, M. M. J., Espe, M., Lerma, C., and Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a model for mitochondrial cytochrome c oxidase. Purification, kinetics, proton pumping, and spectral analysis. J. Biol. Chem. 267, 24264–24272. 83. Keightley, J. A., Sanders, D., Todaro, T. R., Pastuszyn, A., and Fee, J. A. (1998) Cloning and expression in Escherichia coli of the cytochrome c552 gene from Thermus thermophilus HB8. Evidence for genetic linkage to an ATP-binding cassette protein and initial characterization of the cycA gene products. J. Biol. Chem. 273, 12006–12016. 84. Baker, G. N., Noguchi, M., and Palmer, G. (1987) The reaction of cytochrome oxidase with cyanide: preparation of the rapidly reacting form and its conversion to the slowly reacting form. J. Biol. Chem. 262, 595–604. 85. Giuffre, A., Forte, E., Antonini, G., D’Itri, E., Brunori, M., Soulimane, T., and Buse, G. (1999) Kinetic properties of ba3 oxidase from Thermus thermophilus: effect of temperature. Biochemistry 38, 1057–1065. 86. Muresanu, L., Pristovsek, P., Lohr, F., Maneg, O., Mukrasch, M. D., Ruterjans, H., Ludwig, B., and Lucke, C. (2006) The electron transfer complex between cytochrome c552 and the CuA domain of the Thermus thermophilus ba3 oxidase: A combined and computational approach. J. Biol. Chem. 281, 14503–14513. 87. Buschmann, S., Warkentin, E., Xie, H., Langer, J. D., and Michel, H. (2010) The structure of cbb3 cytochrome oxidase provides insights into proton pumping. Science 329, 327–330. 88. Verkhovsky, M. I., Morgan, J. E., and Wikstr€ om, M. (1992) Intramolecular electron transfer in cytochrome c oxidase: a cascade of equilibria. Biochemistry 31, 11860–11863. 89. Verkhovsky, M. I., Morgan, J. E., and Wikstr€ om, M. (1995) Control of electron delivery to the oxygen reduction site of cytochrome c oxidase: a role for protons. Biochemistry 34, 7483–7491. 90. Giuffre, A., Stubauer, G., Sarti, P., Brunori, M., Zumft, W. G., Buse, G., and Soulimane, T. (1999) The heme-copper oxidases of Thermus thermophilus catalyze the reduction of nitric oxide: evolutionary implications. Proc. Natl. Acad. Sci. U.S.A. 96, 14718–14723. 91. Rich, P. R. (1995) Towards an understanding of the chemistry of oxygen reduction and proton translocation in the iron-copper respiratory oxidases. Aust. J. Physiol. 22, 479–486. 92. Pilet, E., Jasaitis, A., and Liebl, U. (2004) Electron transfer between hemes in mammalian cytochrome c oxidase. Proc. Natl. Acad. Sci. U.S.A. 101, 16198–16203. 93. Jasaitis, A., Hohansson, M. P., Wikstr€ om, M., Vos, M. H., and Verkhovsky, M. I. (2007) Nanosecond electron tunneling between the hemes in cytochrome bo3. Proc. Natl. Acad. Sci. U.S.A. 104, 20811–20814. 94. Sousa, F. L., Verissimo, A. F., Baptista, A. M., Soulimane, T., Teixeira, M., and Pereira, M. M. (2008) Redox properties of Thermus thermophilus ba3: different electron-proton coupling in oxygen reductases? Biophys. J. 94, 2434–2441. 95. Adelroth, P., Brzezinski, P., and Malmstr€ om, B. G. (1995) Internal electron transfer in cytochrome c oxidase from Rhodobacter sphaeroides. Biochemistry 34, 2844–2849. 96. Ching, E., Gennis, R. B., and Larsen, R. W. (2003) Kinetics of intramolecular electron transfer in cytochrome bo3 from Escherichia coli. Biophys. J. 84, 2728–2733.