H. .,.. ., ..V .‘ H ,..... ‘ .,. ‘. ..... . . ..‘.’-...... . .. ...... .V. .....H. ..,...,... , 1‘: 9"y~~v( ~~~nun~ ......---_.... .qflqub IIIIIIIIIIIIIIIIIIIIIIIIIIII Thisistocettifythatthe dissertationentitled TtMEsRESOLVED RESONANCE QAHAN DETECTtoN INTERMEDIATE) Ml THE REDchcovJ or (\(‘rocHRoHE 0110A; E presentedby OF btoxf GEN BY CONSTANT! :4 05 A. VARoTsu has been accepted towards fulfillment of the requirements for P r degreein CHEMchL PHYSKS MM Major professor ,/ Date \v/W/X/ @?0 MS U is an Affirmative Action/Equal Opportunity Institution ‘fi LXBRARY Michigan State University A .__l‘ PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. E:J l—TCJ (—7 L j __J a ll __ IL___I[ u [- l—jT—l MSU Is An Affirmative Action/Equal Opportunity Institution chflJ TIME-RESOLVED RESONANCE RAMAN DETECTION OF INTERMEDIATES IN THE REDUCTION OF DIOXYGEN BY CYTOCHROME OXIDASE BY Constantinos A. Varotsis A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Chemical Physics Program 1990 ABSTRACT TIME-RESOLVED RESONANCE RANAN DETECTION OF INTERNEDIATES IN THE REDUCTION OF DIOXYGEN BY CYTOCNRONE OXIDABE BY Constantino: A. Varoteio Cytochrome oxidase, also known as cytochrome aaa, the terminal enzyme of the mitochondrial respiratory chain, catalyzes the reduction of molecular oxygen to water. In the work reported here, we used a combination of rapid mixing, laser photolysis, and a novel jet apparatus that produces a continuous stream of sample in air to detect the transient intermediates formed in the reactions of fully- reduced and mixed-valence cytochrome oxidase with dioxygen. In the spectrum recorded at 10 ps subsequent to carbon- monoxide photolysis of the fully-reduced enzyme in the presence of 02, a mode is observed at 571 cm'1 that shifts to 546 cm‘1 when the experiment is repeated with 1802. The appearance of this mode is dependent upon the laser intensity used and disappears at higher-incident energies. The high-frequency data, in conjunction with the mid- frequency data, allow us to assign the 571 cm'1 mode to the Fe-o stretching vibration of the low-spin 02 adduct that forms in the fully-reduced cytochrome oxidase/dioxygen reaction. The 571 cm”1 u(Fe-02) frequency in the fully- Constantino: A. Varotsis reduced enzyme/02 adduct is essentially identical to the 572 cm“1 frequency ‘we measured for ‘this mode. during thereduction of 02 by the mixed-valence enzyme, which indicates that the 02-hound cytochrome a3 is independent of the redox state of the cytochrome a/CuA pair. In the spectrum recorded at 800 ps in the reaction of the fully- reduced enzyme with 02, a mode is observed at 790 cm"1 that shifts to 755 cm‘1 when the experiment is repeated with 1"01,. The frequency of this vibration and the magnitude of the 1802 isotopic frequency shift allow us to assign the 790 cm"1 mode to the FeIV=0 stretching vibration of the a’*a3Iv=0/Cu31+ adduct that forms in the fully-reduced cytochrome oxidase. The appearance of this mode is not affected when 020 was used as a solvent. This suggests that the ferryl-oxo intermediate is not hydrogen-bonded. The high-frequency (1000-1700 cm") Raman data during the oxidase/02 reaction show that the oxidation of cytochrome a"’+ is biphasic. The faster phase is completed in 100 ps and is followed by a plateau region in which no further oxidation of cytochrome a occurs. These results are consistent. with the ibranched. pathway for ‘the oxidase/O2 reaction proposed by Hill and Greenwood (1984). Within the context of this scheme, the ferryl-oxo intermediate we have observed arises as the fourth and final electron enters the dioxygen reduction site. To the memory of my Father, my Mother, and Zoe ii ACKNOWLEDGMENTS I would like to thank Dr. Gerald T. Babcock for his friendship, encouragement, and guidance during the course of this work. I am also grateful to Drs. Katharine Hunt, Julius Kovacs, and Paul Parker for serving on my committee. Thanks also goes to Vada O'Donnell and Bill Draper for their efforts in the preparation of this manuscript. I also owe special thanks to Ron Haas and Deak Watters for technical services. Finally, I would like to thank all my fellow lab members for their general assistance. ifi TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 0 Vi LIST OF FIGURES O O O O O I O C O O O O O O O O O O O I O O O O O O O O O O I O O O I O O O O Vii CHAPTER 1 - GENERAL INTRODUCTION ...................... 1 The Individual Metal .............................. 6 Heme Absorption ................................... 12 Optical Absorption of Cytochrome Oxidase .......... 12 Raman Theory ...................................... 16 Oxygen Intermediates .............................. 22 References ........................................ 31 CHAPTER 2 - A SIMPLE MIXER/JET CELL FOR RAMAN SPECTROSCOPIC STUDIES Summary ........................................... 34 Introduction ...................................... 35 Cell Description and Performance .................. 37 References ........................................ 48 CHAPTER 3 - TIME-RESOLVED RAMAN DETECTION OF v(Fe-O) IN AN EARLY INTERMEDIATE IN THE REDUCTION OF 02 BY CYTOCHROME OXIDASE summary OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 49 References O..0.0.I...0.0.0.0...0.00.00.00.00...OO. 60 CHAPTER 4 - DIRECT DETECTION OF A DIOXYGEN ADDUCT OF CYTOCHROME A3 IN THE MIXED VALENCE- CYTOCHROME OXIDASE/DIOXYGEN REACTION Summary ........................................... 62 Introduction ...................................... 63 Experimental Procedures ........................... 67 Results ........................................... 68 Discussion ........................................ 81 References ........................................ 90 iv Page CHAPTER 5 - LATE APPEARANCE OF THE u(FeIV=O) VIBRATION FROM A FERRYL-OXO INTERMEDIATE IN THE CYTOCHROME OXIDASE/DIOXYGEN REACTION Summary ........................................... 94 Introduction ...................................... 95 Experimental Procedures ........................... 99 Results ........................................... 100 Discussion ..... ...... ............................. 111 Future Work ... ........ ............................ 117 References . .................... ..... ........ ...... 119 LIST OF TABLES Low-Frequency Raman Modes (cm") in Cytochrome Oxidase ............. .............. Vibrational Frequencies for Dioxygen-Bound Complexes ................................. ...... Vibrational Frequencies for Dioxygen-Bound Complexes .................................. ..... Vibrational Frequencies for Ferryl-Oxo Complexes Page 46 58 87 116 LIST OF FIGURES Sequence of electron carriers in the respiratory chain ............. ..... . ............ The location of cytochrome oxidase in eucaryotic organisms .......... ..... .. ........... Geometry and coordination properties for cytochrome a, cytochrome a3, CuA, and Gus in cytochrome c oxidase. Adapted from Ref. 12c .... A summary of current models for the coordination geometries of the iron and copper centers in reduced cytochrome oxidase. Adapted from Ref. 23 Optical absorption spectra of cytochrome oxidase: oxidized-—and fully reduced.... Adapted from Ref. 15c ...... ................... ... ............ Diagram illustrating the resonance Raman effect Reaction mechanism of fully-reduced cytochrome oxidase obtained by flash photolysis at room temperature. Adapted from Ref. 18 ........ . ..... Postulated intermediates in the reduction of 02 by cytochrome oxidase. Only the aa/CuB site is shown. L represents a bridging ligand in the oxidized form of the enzyme. Adapted from Ref. 23 ..... .... ..... ...... ......... . ........... Apparatus for room-temperature resonance Raman spectra of rapidly-mixed/flowing samples ........ High-frequency resonance Raman spectra of oxidized cytochrome oxidase. Spectrum A was recorded with a quartz capillary. Spectrum B was obtained with the mixer/jet cell. The energy of the 416 nm excitation wavelength was 0.8 mJ. The accumula- tion time was 2.5 min for both spectra and the flow rate was 0.2 ml/min. The sample concentra- tion was 80 pM. ........ ...... .. ........... . ..... vfi Page 11 14 18 25 29 39 42 Page Low-frequency resonance Raman spectra of oxidized cytochrome oxidase. Spectrum A was recorded with a quartz capillary. Spectrum B was obtained with the mixer/jet cell. The energy of the 416 nm excitation wavelength was 0.8 mJ. The accumulation time was 2.5 min for both spectra and the flow rate was 0.2 ml/min. The sample concentration was 80 pM. ............. .................... ..... 45 Instrumental configuration used for pulsed, time-resolved resonance Raman measurements of flowing cytochrome oxidase samples prepared by rapid mixing . ...... ............... ....... .... 54 Time—resolved resonance Raman spectra of cyto- chrome oxidase following initiation of the reaction with oxygen at room temperature. The energy of the 532 nm photolysis pump pulse was 1.3 mJ, sufficient to photolyze the enzyme-CO complex and initiate the 02 reduction reaction. The energy of the probe beam was 0.3 mJ for spectra A and B“ and 1.0 mJ for spectra C-E. The repetition rate for both the pump and probe pulses (10 ps duration) was 10 Hz. The pump- probe delay was 10 ms for the transient spectrum E. The accumulation time was 110 min for spectrum A, 70 min for spectrum 8, 5 min for spectra C and D, and 15 min for spectrum E. ................... 56 Time-resolved resonance Raman spectra of mixed valence cytochrome oxidase at the indicated times following initiation of the reaction with oxygen at room temperature. The energy of the 532 nm photolysis pump pulse was 1.3 mJ, sufficient to photolyze the enzyme -CO complex and initiate the 02 reduction reaction. The energy of the probe beam was 1.0 mJ for all spectra. The spectrum of the resting enzyme was obtained by using the probe beam only. The repetition rate for both the pump and probe pulses (10 ns duration) was 10 Hz. The accumulation time was 15 min for each spectrum. 70 Time-resolved resonance Raman spectra of mixed valence cytochrome oxidase following initiation of the reaction with oxygen at room temperature. The pump-probe delay was 10 us for spectra A and C and 2 us for spectrum B. The energy of the probe beam was 0.3 mJ for spectra 8 and C and 1.0 mJ for spectrum A. The accumulation time was 45 min for spectra B and C and 15 min for spectrum A. ...... 73 vfii Page Time-resolved resonance Raman spectra of mixed valence cytochrome oxidase following initiation of the reaction with oxygen at room temperature. The energy of the 532 nm photolysis pump pulse was 1.3 mJ. The energy of the probe beam was 0.3 mJ for spectra A and B and 1.0 mJ for spectra C, D, and E. The repetition rate for both the pump and probe pulses (10 ns duration) was 10 Hz. The pump-probe delay was 10 us for the spectra A-D and 10 ns for the transient spectrum E. The accumulation time was 120 min for spectrum A, 65 min for spectrum B, 5 min for spectra C and D, and 15 min for spectrum E. ............... 76 Difference spectra of the initial intermediate in the reaction of mixed valence cytochrome oxidase with isotopes of oxygen observed at 10 us into the reaction. Spectrum A was obtained by sub- tracting the low power 1802 spectrum (Figure 38) from the low power 1602 spectrum (Figure 3A). Spectrum B was obtained by subtracting the high power 1602 (Figure 3C) spectrum from the low power 1602 spectrum (Figure 3A). ................ 79 Time-resolved resonance Raman spectra of fully-reduced cytochrome oxidase at the indicated times. The energy of the 532-nm photolysis pump/pulse was 1.3 mJ, sufficient to photolyze the enzyme-CO complex and initiate the Oz-reduction reaction. The energy of the 427-nm probe beam was 0.8 mJ for spectra A and F and 0.3 mJ for spectra B-E. The repetition rate for both the pump and probe pulses (10-ns duration) was 10 Hz. The accumulation time was 15 min for spectra A and F and 50 min for spectra 8-8. The enzyme concentration was 50 pM after mixing, pH 7.4. ......................................... 102 Time-resolved resonance Raman spectra of fully-reduced cytochrome oxidase at the indicated times. The energy of the 532-nm photolysis pump/pulse was 1.3 mJ. The energy of the 441-nm probe beam was 0.8 mJ for spectrum A and 0.3 for spectrum B. The repetition rate for both the pump and probe pulses was 10 Hz. The accumulation time was 15 min for spectrum A and 65 min for spectrum B. The enzyme concentration was 50 pM after mixing, pH 7.4. ......................................... 106 Page Time-resolved resonance Raman spectra of fully-reduced cytochrome oxidase following initiation of the reaction with oxygen at room temperature. The energy of the 532-nm photolysis pump/pulse was 1.3 mJ. The energy of the 427-nm probe beam was 0.8 mJ for all spectra. The repetition rate for both the pump and probe pulses was 10 Hz. The accumulation time was 15 min for spectrum A and 20 min for spectra B-E. The enzyme concentration was 50 pM after mixing, pH 7.4. ........... ...... 110 CHAPTER 1 GENERAL INTRODUCTION The production of energy in the cells of aerobic organisms is based on the transfer of electrons through a series of redox proteins in the respiratory chain leading to the final reduction of oxygen to water (Figure 1.1). An essential element of this mechanism is cytochrome oxidase, the terminal enzyme in cellular respiration (Figure 1.2). This enzyme transfers electrons from the protein cytochrome c to oxygen and carries out the reduction of oxygen according to the overall reaction: 4 cytochrome c2+ + 02 + 4 H" .. 4 cytochrome c3+ + 2 H20. In eucaryotes, the components of the respiratory chain are located in the inner membrane of mitochondria, and the free energy released in redox reactions is used to generate a proton gradient across the membrane. The ATP synthase complex, present in the same membrane, catalyzes the synthesis of energy-rich ATP from ADP and phosphate with this proton gradient as the driving force. Thus, the mitochondrial respiratory chain and the Figure 1.1 Sequence of electron carriers in the respiratory chain; approximate midpoint potentials of the components at pH 7 are indicated on the right side of the Figure. NADH-Q reductase 1 Q l Cytochrome c re uctase Cytochrome c Cytochrome c oxidase O2 02 + 4H 4e 0.030 0.254 0.385 0.816 ‘ 2320 Figure 1.2 The location of cytochrome oxidase in eucaryotic organisms. Adapted from Ref. 2a. o 0:32.033 3033 2:253: ATP synthase complex supply the energy for the eucaryotic cell. Cytochrome oxidase has at least eight subunits and a M.W of 140,000 Daltons.1 Probably all of the redox metal centers are located in the two largest subunits (I and II). Subunit III has been implicated in the proton-translocating activity. Cytochrome oxidase contains four metal atoms per functional unit: two hemes, cytochrome a and a3; and two associated copper atoms, CuA and CuB. The four redox active metals can be divided into two pairs. Cytochrome a and CuA function together in the sense that they accumulate electrons from cytochrome c and then act as a two-electron donor to the second pair metal ions, namely cytochrome a8 and Gun. This second pair is present as a binuclear center that constitutes the catalytic site where oxygen binds and is reduced to H20. 1. The Individual Metals Cytochrome a In both oxidation states (II and III), the iron of cytochrome a remains low spin.2b IILis characterized in its oxidized form by an EPR spectrum with g-values of 3.03, 2.213, and 1.5.3 Comparison of these signals with those of model compounds suggested that, in the enzyme, this heme iron. is coordinated. to two .neutral histidine residues.4 Comparison of the resonance Raman spectra of cytochrome a 5 and model compounds points to the same conclusions, as does the comparison of the MCD spectrum of bis-imidazole heme a and cytochrome a.6 Gun The environment of this copper atom has been partially explained by the application of EPR3'8 and ENDORg'10 spectroscopy together with the incorporation of I‘M into the enzyme isolated from yeast.11a These methods allowed the identification of one histidine and one cysteine as ligands, with the possibility of a second cysteine. Cytochrome a3 and CuB The iron atom of cytochrome a3 and its associated CuB form a coupled binuclear center. The most conspicuous spectroscopic feature of this interaction is the lack of EPR signals from the individual metals, even when they occur in their oxidized, paramagnetic states. This unusual situation arises from antiferromagnetic coupling between the high-spin ferric ion and the CuB(II), forming a S = 2 EPR silent pair.3'8'12 In the oxidized, resting enzyme, this coupling may be facilitated by a bridging ligand, possible identified by' EXAFS as a chloride atom.11b .Although EPR silent, cytochrome a3 has been shown, from magnetic susceptibility13 and MCD measurements,14 to be high spin in the oxidized state and to remain so when reduced. Mossbauer spectroscopy15a confirms the high-spin nature of this ion: and, since no magnetic features are seen in the oxidized state at 4.2 K, supports the idea of spin coupling to a cupric ion. Figure 1.3 Geometry and coordination properties for cytochrome a, cytochrome a3, CuA, and CuB in cytochrome c oxidase. Adapted from Ref. 11C. N\_um nu mom 02 a o 53 83:3 neococoxovEc mm@ .36 \ Eu :3 N .o .558 $878 .56 A 3.5 2/ \z 3...: m: \u o/ 352 s. b. .L o I: N 0 63:71:53.. ~ I/ \II Ommll\ /I 33 ..s0 new. . o. 8: Io” u B exovmdm oz / Ecmow 0\oO_nO z Amtbv Ecnvv exoOm .nm .3 § 23m . m /M:o\ .. A25 2\ /z 35 -2 252 o v. % I .m— \uA\ NIUWU\IB m_n_l I 352 V I U. \ a 1.8; 7.3 so. .253 ...S 08. . 0;: N\_nw.nn0 \ :3: e: now Sow A .36 38.833. $8 « 30v 10 Figure 1.4 A summary of current models for the coordination geometries of the iron and copper centers in reduced cytochrome oxidase. Adapted from Ref. 23. H OH ’OOC C..H27 N N \ / Cu \ (1?) N 'OOC . 2+ I+ cyf g3 CUB O2 REDUCTION N O(?) CYTOCHROME g OXIDATION 12 II. Home Absorption The characteristics of the RR spectrum of a heme protein depend on the absorption band chosen in generating the RR effect. It is, therefore, necessary to outline the general characteristics of heme absorption spectra. The visible and near-ultraviolet absorption spectra of heme and metalloporphyrins are dominated by two 1r -+ «1* electronic transitions. Both transitions are polarized in the plane of the heme (x,y) and are of the same symmetry, Eu. The n + «* transitions are subject to strong configuration interaction, with the result that the transition dipoles are additive for the higher energy transition and largely cancel for the lower-energy one. The higher-energy transition is assigned to the intense (c .. 105 M“ cm") absorption band, called the Soret or 1 band, near 400 nm. The lower-energy transition is assigned to the a band, and has approximately tenfold less oscillator strength than the Soret transition. The lower-energy transition can "borrow" some of the intensity of the higher-energy transition through vibrational interactions. This produces a vibronic side band, called the )9 band, which occurs at an energy ~ 1300 cm'1 higher than the a band. III. Optical Absorption of Cytochrome Oxidase The spectrum of the fully-oxidized enzyme, as prepared, is characterized by broad absorption bands with peaks around 420 nm and 600 nm.1513 The position of the Soret band is 13 Figure 1.5 Optical absorption spectra of cytochrome oxidase: oxidized-—and. fully reduced---- Adapted from Ref. 15c. E. mM" cm" 14 200 - "\ .. ,’ \ O ) ISO - ’ ‘\\ :20— \ so— \ \ 40 O ’- 4 1 l 1 1 1 1 l 1 \T-C-l" 400 450 500 “--— OILIIIIIIIIJIIIJIIIL 500 550 600 650 700 omemm ‘ JilinlnLJIlInlnlIJIIJIL 650 700 750 800 850 Wavelength, nm 15 variable, depending on the method of preparation. Other distinctive features of the oxidized enzyme are the broad band centered at 830 nm (CuA'”) and a shoulder at 655 nm. On complete reduction of the enzyme, the Soret band shifts to 444 nm and the alpha band to 604 nm. The spectrum of cytochrome a is largely independent of the redox and ligation state of cytochrome a3 and vice versa. Therefore, it has been possible to deconvolute the optical spectra and assign spectral properties to the individual centers. By using half-reduced states (i.e., enzyme species that have been prepared so that the hemes have different oxidation levels) and a variety of ligands for either ferrous or ferric cytochrome a3, it has been shown that the cytochrome a makes the major contribution (80%) to the spectrum of the reduced enzyme at 605 nm, while the two cytochromes make approximately equal contributions at 444 nm. Carbon monoxide reacts with cytochrome oxidase only when both metals at the binuclear center are reduced, regardless of the oxidation state of cytochrome a and CuA. Carbon monoxide is bound only to ferrocytochrome a3 , whose absorption spectrum is characterized by a small decrease of the absorbance at 605 nm and new absorption bands at 592 and 430 nm.15 16 1?. Raman Theory The Raman effect derives from the inelastic scattering of electromagnetic radiation by matter. The molecule is promoted to higher vibrational quantum levels of the ground electronic state during this process. Molecular spectroscopic information is conveyed by the vibrational frequency shift, by the intensity of the Raman scattering process, and by the polarization of the Raman scattered light relative to the incident radiation. The intensity of the scattered radiation Imnr due to the transition from |m> to |n>, is given by equation 1, where (apa)mn is the transition polarization tensor with incident and scattered polarizations indicated by p and a, respectively. The expression for (“palmn is given by equation 2. The polarizability expression has a contribution from all of the rovibronic molecular excited states, as indicated by the summation over the excited states |eu> in equation 2. The weighing of an individual |eu> excited state contribution is determined by the values of the transition moment matrix elements in the numerator and by the values of the energy denominators. The energy denominator contains information on rev: the homogeneous linewidth for the transition between the ground state m and the excited state |eu>, and the detuning of excitation from resonance (uev-vo). For normal Raman scattering in which the excitation frequency is far 17 Figure 1.6 Diagram illustrating the resonance Raman effect 1 two hrs ____.Y_ _128 n5 18 (V0+ - vInn) IO Z [(aoo)rnnlz = Z:-;([ v (”l-Vo— Ir" (glue) v "—vo—il‘" [— aQ. (Elnlg) 0 V a—Vo-ira Mainly) 0Q. Ulv)(v|i) }(le.lv)(v|i) l) (f Iv) (vIQ.li) “m”? [MD UlQ.lv)(v|Q.Ii). 0 “ 5Q. 19 from resonance with an electronic transition in the molecule, the denominator only weakly depends upon excitation frequency, and all molecular rovibronic transitions contribute essentially in proportion to their transition moments. As resonance with an electronic transition. is approached, the first ‘term. in. equation 2 begins to dominate the sum over states of the Raman polarizability expression. If we display the dependence of the transition moment matrix elements upon vibrational motion, we can show that equation 2 predicts three distinct scattering mechanisms, A, B, and C, as discussed in the following paragraphs. The transition moment matrix elements have been factored into separate electronic and vibrational integrals. This is achieved within the Born-Oppenheimer approximation by writing the total wave-functions as products of electronic and vibrational parts. The change of the electronic wave-function, due to the origin shift, is incorporated by using the Herzberg-Teller expansion. The transition moment integrals between the ground and excited electronic states are denoted by angular brackets. Curved brackets denote the integrals between vibrational level, where |i) labels the initial vibrational level of the normal mode a in the ground electronic state, and |v) labels the vibrational level of mode a in the excited electronic state. If) labels the final vibrational level of mode a in the ground electronic state (f = 1+1 for Stokes Raman scattering 20 from node a). The Franck-Condon overlap factors (f |v) (v| i) will differ from zero only if the excited state equilibrium geometry is displaced along a symmetric normal mode coordinate relative to the ground state. This assumes identical orthonormal vibrational wavefunctions in the ground and excited states. A-term, Condon enhancement may also occur due to strong Franck-Condon overlaps when the excited state vibrational level are solutions to a different Hamiltonian than that of the ground state. For the 'more common case in large molecules of similar vibrational mode compositions in the ground and excited states, where the excited state geometry either expands or contracts relative to the ground state, the enhancement of totally symmetric vibrations scales roughly as (A)2/2, where A is the magnitude of displacement of ‘the excited. state potential surface along' the Raman active normal coordinate. For small displacements, only the fundamental shows significant enhancement. Larger displacements result in lengthy Franck-Condon progressions. Enhancement of symmetric vibrations by the A-term can also derive from vibrational force constant changes in the excited state as well as by alterations in the composition of the excited state normal coordinates (Duschinsky effect). Enhancement via the B-term derives from the non-Condon dependence of the electronic transition moment upon the vibrational coordinate. One example of a non-Condon enhancement mechanism is Herzberg-Teller vibronic coupling 21 of different electronic transitions. Both symmetric and nonsymmetric fundamentals can be enhanced by a B-term mechanism for a strongly-allowed transition, the magnitude of B-term enhancement of symmetric vibrations is significantly below that for A-term enhancement. B-term enhancement will dominate only for the nonsymmetric vibrations. If the transition is forbidden at the equilibrium geometry, enhancement of fundamentals cannot occur via either the A or B-terms; however, C-term enhancement of overtones and combinations can occur and will involve two quanta of vibrational mode a or the combination of one quantum of mode a and one quantum of mode b. The C—term only dominates enhancement for resonance excitation within forbidden electronic transitions. Resonance Raman studies of porphyrins provide an important correlation between the position of given vibrations and the oxidation and spin states of the central metal. The resonance Raman results for cytochrome oxidase discussed in this chapter are primarily those obtained with Soret excitation. Under this condition, the Franck-Condon mechanism dominates RR scattering, and all observed RR modes are polarized (p ~ 0.33). The vibrations in the 1000-1700 cm‘1 region represent primarily C-C and C-N stretching motions of the hemes. The following discussion concentrates on two primary RR indicator bands of cytochrome oxidase. The first is the v‘ symmetric vibration in the 1355-1375 22 cm"1 region, the frequency of which is sensitive to electron density in porphyrin. x* orbitals. In ‘practice, u‘ is sensitive to two effects: the oxidation state of the metal and the presence of axial ligands that have x-acid character and, thus, may withdraw porphyrin «it-electron density via the metal. The second RR indicator band is the V2 vibration which behaves as a reliable core-size indicator. Therefore, the lowering of the heme frequencies reflects expansion of the heme. A detailed analysis of several high- and low- frequency vibrations in cytochrome oxidase will be discussed in Chapters 3, 4, and 5. V. Oxygen Intermediatee In 1958, Okunukil7 reported that when 02 is added to the reduced-cytochrome oxidase, the Soret shifts to 428 nm. This newly-observed species was termed "oxygenated" and was believed to be an oxygen compound of cytochrome oxidase similar to oxyhemoglobin. It is now known that this simple view is incorrect and that reduced cytochrome oxidase reacts with 02 in approximately 1 ms at room temperature to regenerate the fully-oxidized enzyme.18 Because the reaction is extremely fast, manual stopped flow is not reliable: instead the 02 site can be blocked with CO. The CO dissociates slowly in the absence of light (k=2.5x10”s“) so that an anaerobic sample treated with CO can be mixed with 02 under conditions for which no significant reaction with oxygen occurs (30 sec) until a short light pulse is used to photodissociate the CO. Gibson and Greenwood18 studied the reaction of reduced cytochrome oxidase with 02 by combining rapid mixing and flash. photolysis, which. provided resolution on. the microsecond timescale. This "flow-flash” technique involves mixing the CO-bound reduced enzyme with O2 and then subjecting the mixture to an intense flash of light. The flash photolyzes the heme-CO bond to produce the unbound reduced enzyme free to react with 0,. Gibson and Greenwood proposed a model (Figure 7) suggesting that the initial bimolecular step is a combination of the enzyme with O2 in which 02 is not bound, as an inner-sphere ligand, to cytochrome as, but instead is trapped in the protein pocket. The second step is an intramolecular step occurring at a first-order rate of 6x10‘s'1 resulting in the transfer of O2 to its metal binding site at cytochrome a3 , which corresponds, in Chance’s notation, to compound A formation. The kinetic difference spectra generated in the O2 reaction may be reconciled with those from the static and reducive kinetic experiments by having about 40% of the cytochrome a oxidized with t‘/. ~ 35 pS.18 Such a suggestion requires a branch in the pathway of the reaction of the fully-reduced enzyme with 02. The scheme of Figure 7 suggests that the first event 24 Figure 1.7 Reaction mechanism of fully-reduced cytochrome oxidase obtained by flash photolysis at room temperature. Adapted from Ref. 18 25 FTully reduced 92’ 932’ CUA. CUB. .02 1x 10" M" s" 92’ (gaz‘oz) CUAO CUB. l 3x104 5'1 o ' 932202 Compound A Cqu CUB. /3x10° \ 91331022- gz’ 9332022- CUB‘, CUA. CUBZ‘ 7x 103 s'1 3. 20 30 93 -023- 9. 93 -023- Cu8 .2- CUAZO CUBZo 0 AH. 700 s-1 513’ 933' CuA2o CUBZo . 2 H20 26 after 02 binding is the transfer of two electrons to form a peroxo intermediate. It is at this stage that the branch is introduced. The two electrons transferred in this step may originate from either cytochrome a3 and CuB or from cytochrome a3 and cytochrome a. This heterogeneity may simply be a manifestation of an intrinsic difference in the rates of electron transfer from different metal centers in the enzyme to bound 02. The second electron-transfer step involves the oxidation of CuA, as evidenced by the change at 830 nm. There is evidence, from spectroscopic potentiometric-titration experiments, that the 830 nm band is nearly entirely due to CuA.19'2° The observed ratio of the fast/slow phases at 830 nm is 60:40, as predicted by the proportion in each. branch necessary ‘to account for the cytochrome a and a3 absorption contributions in the Soret and visible regions. Therefore, the branching model is able to account for the complex kinetics at 830 nm and retain CuA as the sole chromophore contributing at this wavelength. This result also suggests that interconversion between the two branches must be slow, relative to the rates of the individual steps in the branches. The second and third electron-transfer steps in each branch are shown to occur at identical rates. This implies that the rate limit to electron transfer within the complex is imposed by a shared event such as the binding of protons. Chance et al.21a developed techniques to study the reaction on a second-to-minute time scale at low 27 temperature. With this approach, Chance et al.,21a Clore et al.,21b 1.21C and Chan et a studied intermediates involved in dioxygen reduction and characterized them by using optical and EPR spectroscopies. Chan et al.21c confirmed the earlier findings of Clore et al.21b about the initial phases of the reaction involving the binding of O2 to form compound A and subsequent heterogeneous reaction of both CuA and Cyt a. They also show that the kinetics of the appearance of the EPR signal assigned to CuB are correlated to electron transfer from either CuA or Cyt a. These workers suggest that the decay of the CuB signal may be correlated to the breaking of the dioxygen bond. Although Gibson and Greenwood studied the absorption changes during the O2 reaction of reduced cytochrome oxidase by the rapid-reaction technique of flow-flash spectrophotometry in the Soret, visible and near i.r. spectral region, their data do not provide direct spectroscopic evidence to support the proposed structures for the various intermediates shown in Figure 7. Time- resolved resonance Raman spectroscopy provides a variety of advantages over time-resolved absorption spectroscopy. It offers not only the possibility of detecting the out-of- plane ligand vibrational modes and, thus, the structure of the bound 02 species through isotopic labeling of the 02, but information concerning heme geometric and electronic properties as well. Changes in these molecular parameters at the protein-active site can then be used to test whether Figure 1.8 Postulated intermediates in the reduction of Oz by cytochrome oxidase. Only the as/CuB site is shown. L represents a bridging ligand in the oxidized form of the enzyme. Adapted from Ref. 23. 3+ Fe°3 (oxidized) 1) 2e" e", H“ K 20H- 2+ Cu,3 (ferry!) 29 H- CUB Fe2+ 3 . (reduced ) 2+ Cua \ - O / HO Fe2+ °3 eZH+ (ferrous peroxy) Fe3+ “3 (ferric peroxy) 30 the postulated structures in Figures 7 and 8 do, in fact, occur. Babcock et. al.22'23 used two-color, time-resolved resonance Raman spectroscopy on flowing samples prepared by rapid mixing to study the reaction of fully- and partially- reduced cytochrome c oxidase with dioxygen. Although instrumental limitations restricted data acquisition to the high-frequency (>1000 cm") range, they concluded that the dioxygen reduction proceeds via a photolabile oxycytochrome a3 complex which has characteristics similar to oxyhemoglobin and oxymyoglobin. Moreover, they postulated that these intermediate species were replaced by nonphotolabile dioxygen adducts in which cytochrome a3 was oxidized with t%~60ps and t%~200ps in the fully-reduced and mixed-valence complexes. The aim of this work is to characterize the vibrational and structural properties of the intermediates formed in the reactions of dioxygen with fully-reduced and mixed-valence cytochrome oxidase by using time-resolved resonance Raman spectroscopy in conjunction with rap id-mixing/ f low techniques applied to the enzyme. 31 REFERENCES M. Wikstrom, K. Krab, and M. Saraste, "Cytochrome Oxidase, A Synthesis", Acad. Press, New York, 1981. a) P. E. Thornstrom (1988), Ph.D. Thesis, University of Goteborg and Chalmers University Of Technology. b) K. E. Falk, T. Vanngard, and J. Angstrom, FEBS Lett. 15, 23 (1977). C. R. Hartzell and H. Beinert, Biochim. Biophys. Acta, ggg, 318 (1974). W. E. Blumberg and J. Peisach in "Cytochrome Oxidase", T. E. King, Y. Oni, B. Chance, and K. Okunuki, Eds., Elsevier, Amsterdam, 1979, p. 153. G. T. Babcock, P. M. Callahan, M. R. Ondrias, and I. Salmeen, Biochemistry 20, 959 (1981). D. G. Eglinton, B. D. Hill, C. Greenwood, and A. J. Thomson, J. Inorg. Biochem. 21, 1 (1984). M. T. Wilson, C. Greenwood, M. Brunori, and E. Antonini, Biochem. J. 14_, 145 (1975). R. Aasa, S. P. J. Albracht, K. E. Falk, B. Lanne, and T. Vanngard, Biochim. Biophys. Acta 54;, 260 (1976). B. M. Hoffman, J. K. Roberts, M. Swanson, S. H. Speck, 10. 11. 12. 13. 14. 15. 16. 32 and E. Margoliash, Proc. Natl. Acad. Sci. USA 11, 1452 (1980). H. L. VanCamp, J. H. Wei, C. P. Scholes, and T. E. King, Biochim. Biophys. Acta 511, 238 (1978). a) T. H. Stevens, C. T. Martin, H. Wang, G. H. Brudvig, C. P. Scholes, and S. I. Chan, J. Biol. Chem. 251, 12106 (1982); b) R. Scott, M. Li, and S.—I. Chan (1988), Annals of the New York Academy of Science 559, 53-58; c) J. Centeno (1988), Ph.D. Thesis, Michigan State University. G. Palmer, G. Babcock, L. Vickery, Proc. Natl. Acad. Sci. U. S. A. 1;, 2206-2210 (1976). M. Tweedle, L. J. Wilson, L. Garcia-Iniquez, G. T. Babcock, and G. Palmer, J. Biol. Chem. 25;, 8065 (1978). a) G. T. Babcock, L. E. Vickery, G. Palmer, J. Biol. Chem. 251, 7907 (1976); b) A. J. Thomson, T. Brittain, C. Greenwood, and J. P. Springill, Biochem., J. 165, 327 (1977). a) T. A. Kent, L. J. Young, G. Palmer, J. A. Fee, and E. Munck, J. Biol. Chem. 258, 8543 (1983). b) Vanneste, W. H., Biochemistry 5, 838-847 (1966); c) F. G. Halaka (1981), Ph.D. Thesis, Michigan State University. J. Tang and. A. C. Albrecht in. "Raman Spectroscopy Theory and Practice" (ed. H. A. Szymanski), Vol. 2, pp. 33-68; D. L. Rousseau and P. C. Williams in "Topics in 17. 18. 19. 20. 21. 22. 23. 33 Current Physics", Vol. 2, "Raman Spectroscopy of Gases and Liquids" (Ed. A. Weber): R. J. H. Clark and B. Stewart, Struct. and Bonding 16, 1 (1979). K. Okunuki, B. Hagihara, I. Sekuzu, T. Horio, Proc. Int. Symp. Enz. Chem. 264 (1958). O. H. Gibson, C. Greenwood, Biochem. I. 86, 541 (1963); O. H. Gibson, C. Greenwood, Biol. Chem. (240, 2694 (1965): C. Greenwood, B. Hill, Biochem. J. 2.1.8.. 913 (1984): for review see: B. Hill, C. Greenwood, P. Nichols, Biochim. Biophys. Acta 21, 853 (1986) and A, Nafui and B. Chance, Ann. Rev. Biochem. 55, 137 (1986). G. T. Babcock, L. E. Vickery, and G. Palmer (1978), J. Biol. Chem. 25;, 2400-2411. H. Beinert, R. W. Shaw, R. E. Hansen, and C. R. Hartzell (1980), Biochim. Biophys. Acta 521, 458-470. a) B. Chance, C. Saronio, and J. S. Leigh, J. Biol. Chem. 2_$_Q, 9226-9237 (1975); 13) G. M. Clore, L. E. Andreasson, B. Karlsson, R. Aasa, and B. G. Malmstrom, Biochem. J. 185, 139-154 (1980); c) S. I. Chan, S. N. Witt, and D. F. Blair, Chemica Scripta ‘285, 51-56 (1988). G. T. Babcock, J. M. Jean, L. N. Johnston, G. Palmer, and W. H. Woodruff, J. Am. Chem. Soc. m, 8305-8306 (1984). G. T. Babcock, J. M. Jean, L. N. Johnston, W. H. Woodruff, and G. Palmer, J. Inorg. Biochem. 2;, 243-251 (1985). (JHAPTERJl A SIMPLE MIXER/JET CELL FOR RAMAN SPECTROSCOPIC STUDIES* SUMMARY A unique rapid mixing/jet apparatus that allows light scattering from a sample jet in air was constructed and applied successfully to observe low’ and. high frequency resonance Raman spectra of the oxygen metabolizing heme protein, cytochrome oxidase. The cell is designed to minimize sample consumption and is well suited to pulsed laser excitation and to multichannel detection. To illustrate some of the features of the cell design resonance Raman spectra obtained with the mixer/quartz capillary and those obtained with the mixer/jet cell are compared. *Varotsis, C.; Oertling W. A.: and Babcock, G. T., 1990, App. Spectroscopy, in press. 34 35 INTRODUCTION In biological applications of resonance Raman spectroscopy it is frequently desirable to reduce the instantaneous and long term power density of the focused laser beam in order to preserve the sample. Both flowing sample methods and laser beam defocusing techniques have been used successfully by various groups to minimize damage to photolabile samples.1"4 The use of flowing sample cells under conditions in which sample recycling is practical has the additional advantage that long acquisition times can be achieved with fairly minimal scattering consumption. Nonetheless, the flowing cells described to date are most useful when both fairly large volumes of sample are available and recycling is feasible. A further consideration with Raman flow cells involves the sample containment technique in the sample volume. Quartz capillaries are often used, but with these, quartz scattering is severe in the low frequency region, where it overlaps vibrational modes of the sample and makes their detection difficult. Several groups have avoided this problem by arranging their flowing cells so that the sample forms a free jet in air in the scattering volume (e.g. ref. 1, 2). A notable example of this is the microdroplet mixing technique developed by Kincaid and coworkers2 in which both rapid mixing and Raman scattering take place in air and is compatible with continuous wave laser excitation. 36 The necessity of developing the Raman cell described in this report arose when we tried to use pulsed laser excitation to observe resonance Raman scattering of photolabile intermediates formed in the irreversible, reduction of 02 by cytochrome oxidase. Because the reaction is irreversible each protein aliquot can be sampled only once, which necessarily precludes recycling. Thus, a major concern in the construction of the cell was to maximize the information we could Obtain per sample aliquot. An additional Objective was to be able to collect scattered light in the low frequency region efficiently as the most useful vibrations of the bound dioxygen substrate occur in this region. In the design that resulted we are able to achieve the following: (1) efficient mixing at low flow rates by using a modified eight jet Gibson type mixer3 and (2) a continuous flow of sample in air in the scattering volume at flow rates as low as 0.2 ml/min. At this flow rate and with a laser pulse repetition rate of 10 Hz, 1.2 scattering volumes pass between pulses, which insures that each laser pulse is incident on a fresh sample aliquot. With these flow and laser pulse frequency parameters, the total sample volume required per laser shot is 0.33 p1. Moreover, we are able to obtain Raman spectra with good resolution at laser pulse energies as low as 0.3 mJ. With this approach we have avoided photolysis of the transient intermediates and have detected Raman spectra of the first of these species.4 Finally, because the sample 37 is not contained by a quartz capillary in the scattering region, we are able to use multichannel detection to observe low frequency, as well as high frequency, spectra. CELL DESCRIPTION AND PERFORMANCE The design of the mixer/jet cell is shown in Figure 2.1. The exit port of the mixer is constructed so as to allow the flowing sample solution to form a continuous flow in air in the scattering volume, which eliminates quartz scattering. The continuous flow of sample is formed by two 0.66 mm i.d. glass micropipets. The mixer with the upper micropipet and the lower micropipet are each mounted on X- Y-z translators, which allows us to optimize the stability of the jet. Generally, we find that a gap of ~2 mm is appropriate. The iris diaphragm is also mounted on an X-Y- Z positioner and serves as the laser beam waist controller. The translation stage on which the cell is mounted allows X, Y and Z movements relative to the entrance slit of the spectrometer. To illustrate the performance of the cell, and particularly the advantages of the jet in overcoming the quartz scattering problem we show in Figure 2.2 and Figure 2.3 the resonance Raman spectra of the heme protein, cytochrome oxidase, in its oxidized state. This enzyme exhibits vibrational modes in the 200-1700 cm"1 region that arise from the heme macrocycle and, under certain 38 Figure 2.1 Apparatus for room-temperature resonance Raman spectra of rapidly-mixed/flowing samples 39 40 conditions, from iron-axial ligand motions. Resonance Raman excitation at 416 nm is produced by Stokes Raman shifting the third harmonic of a Nd: YAG laser in H2. Typical incident average powers were on the order of a few milliwatts (10 Hz) at this wavelength. The scattered radiation was collected with a SPEX 1877 Triplemate and detected by an EG & G PAR 1420 diode array detector. The resonance Raman spectra of oxidized cytochrome oxidase shown in Figure 2.2 and Figure 2.3 were recorded under identical conditions (laser power, sample concentration, and flow rate) with the exception that for spectra A a quartz capillary was used while spectra B were recorded with the mixer/jet cell described above. In the high frequency region (Figure 2.2) cytochrome oxidase displays Raman vibrations that are strongly enhanced with the Soret electronic transition. These arise from in plane ring modes that are coupled to u-- «* electronic transitions. The performance differences for the two scattering arrangements are apparent in the spectra. The vibrational modes at 1372, 1478, 1573, 1590, 1645, 1651, and 1675 cm”, which are clear in (B), lose resolution and apparent intensity in the spectrum obtained with the quartz capillary (A). The most direct way of probing the bonding to the iron atom in a Iheme protein is to 'monitor the iron-ligand vibrations, which occur below 700 cm"1 . Unfortunately, these out-of-plane vibrations are not strongly enhanced by Figure 2.2 41 High-frequency resonance Raman spectra of oxidized cytochrome oxidase. Spectrum A was recorded with a quartz capillary. Spectrum B was obtained with the mixer/jet cell. The energy of the 416 nm excitation wavelength was 0.8 mJ. The accumulation time was 2.5 min for both spectra and the flow rate was 0.2 ml/min. The sample concentration was 80 pM. RAMAN INTENSITY 42 41 6 nm EXCITATION 1372 B. JET ('3 30 PO) IO F A. QUARTZ CAPILLARY WAVENUMBERS 43 the dominant in-plane n - «* electronic transitions. Nonetheless, the low-frequency resonance Raman spectrum holds considerable promise for quantifying bond strain and changes in length for the axial ligand linkages and is a. major focus in a number of resonance Raman investigations of heme proteins. The 200 - 700 cm‘“1 region of the resonance Raman spectrum of resting cytochrome oxidase is shown in Figure 2.3. Table 2.1 compares and summarizes the frequencies Observed with the mixer/jet and with the mixer/quartz capillary. There are significant differences in intensity between the resonance Raman spectra Observed with the two devices, with a clear improvement in the signal-to-noise ratio of all the Raman modes in the spectrum taken with the mixer/jet arrangement. Of particular importance is the fact that the lower frequency region (<300 cm“) shown in Figure 2.3A sits atop a broad baseline that is due to quartz scattering, which makes the detection of vibrational modes with frequencies between 150 and 300 cm‘1 especially difficult. This is unfortunate because the ligand binding and dissociation pathways of heme proteins involve changes in the heme core size. This, in turn, depends on iron motion out of plane and hence on motion in the Fe-axial histidine bond. Thus, time-resolved resonance Raman studies of the Fe-his bond provide a direct means by which to monitor heme relaxation dynamics. However, the Fe-his stretching vibration occurs in the 200-240 cm‘1 region for most heme proteins, Figure 2.3 44 Low-frequency resonance Raman spectra of oxidized cytochrome oxidase. Spectrum A was recorded with a quartz capillary. Spectrum B was obtained with the mixer/jet cell. The energy of the 416 nm excitation wavelength was 0.8 mJ. The accumulation time was 2.5 min for both spectra and the flow rate was 0.2 ml/min. The sample concentration was 80 pM. INTENSITY RAMAN 45 LOW FREQUENCY 416nm EXCITATION 0.8m) 222 337 cc 1‘38 8 m :2 0” no N co v c Q fl 8. JET A. QUARTZ CAPILLAHY WAVENUMBEFIS Low Frequency Raman modes (cm‘l) in Cytochrome Oxidase 46 Table 2.1' Mixer/Jet Quartz Capillary 222/w — 250/w _ 265/w — 280/w — 337/s 337/w 373/5 373/w 404/s 404/w 685/vs 685/s Abbreviations: w - weak 5 - strong vs - very strong 47 including cytochrome oxidase.5'6 Thus the quartz capillary design is unlikely to be useful in monitoring this important. motion. The mixer/jet arrangement clearly circumvents the quartz scattering problem as shown in Figure 2.3B and thus opens to us the possibility of studying u(Fe-his) as a function of time. The data shown in Figure 2.2 and Figure 2.3 illustrate the unique opportunity of using multichannel techniques to observe Raman spectra of biological molecules in both the high and low frequency region without requiring high laser power' levels or long sampling’ times. IMoreover, small quantities of material are readily sampled. 48 REFERENCES Terner, J.: Spiro, T.G.; Nagumo, M.; Nicol, M.F.; El- Sayed, M.A. J. Am. Chem. Soc., 1_;, 3238 (1980). Caswell, D.S.: Spiro, T.G. J. Am. Chem. Soc., 192, 2796 (1987). Teraoka, J.; Harmon, P.A.: Asher, S.A. App. Spectroscopy 1988, submitted. Simpson, S.F.; Kincaid, J.R.: Holler, J.F. J. Am. Chem., 58. 3136 (1986). Oertling, W.A. and Babcock, G.T. J. Am. Chem. Soc., fl, 6406 (1985). Varotsis, C.; Woodruff, W.H.: Babcock, G.T. J. Am. Chem. Soc., 111, 6439 (1989); 112, 1297 (1990). Ogura, T.; Hon-nami, K.: Oshima, T.: Yoshikawa, 8.; Kitagawa, T. J. Am. Chem. Soc., 1_§, 7781 (1983). Salmeen, I.: Rimai, L.: Babcock, G.T. Biochemistry, 17, 800 (1978). CHAPTER 3 TIME-RESOLVED RAMAN DETECTION OF v(FE-O) IN AN EARLY INTERMEDIATE IN THE REDUCTION OF 02 BY CYTOCHROME OXIDASE“ SUMMARY Flash photolysis of carbon monoxy cytochrome oxidase subsequent to its rapid mixing with oxygenated buffer has been used to initiate the reduction of 02 by the enzyme. By delaying a second laser pulse relative to the photolysis pulse, time resolved resonance Raman spectra. have been recorded during the reaction. In the spectrum recorded at 10 ps after CO photolysis, a mode is observed at 571 cm'1 that shifts to 546 cm'1 when the experiment is repeated with 1802. The appearance of this mode is dependent upon the laser intensity used and it disappears at higher inci- dent energies. We assign this mode to the Fe-O stretching vibration of an early 02 adduct in the cytochrome oxidase/dioxygen reaction. Consideration of recent data on *Varotsis, C.: Woodruff, W. H.: and Babcock, G. T., J. Am. Chem. Soc., 1 1, 6439-6440 (1989): 112, 1297 (1990). 49 50 dioxygen adducts of other heme proteins and model hemes indicates that this early intermediate is most likely the cytochrome a3’*-O2 complex. Cytochrome oxidase contains four redox active centers per functional unit: cytochromes a and a3 and the copper atoms, CuA and CuB. Cytochrome c, the physiological substrate of cytochrome oxidase, transfers electrons to the cyta and CuA sites. These reducing equivalents are transferred to the binuclear cytaa...CuB center, which binds 02 and reduces it to H20. Although the reaction between 02 and cytochrome oxidase occurs too quickly to be studied by conventional stopped-flow techniques, Gibson and Greenwood1 showed that photolysis of the cytochrome a32*-CO complex of the enzyme in the presence of 02 could be used to circumvent this limitation. Babcock et al.2'3 adopted this approach and used time-resolved resonance Raman to study the reaction of fully- and partially-reduced cytochrome oxidase with 02. Although instrumental limitations restricted data acquisition to the high frequency (>1000 cm") range, they concluded that the reoxidation of cytochrome oxidase proceeds via a cytochrome a3’*-O2 complex that resembles oxymyoglobin and oxyhemoglobin. Direct detection of iron/bound oxygen ligand vibrations is necessary to test these conclusions as well as to provide detailed information on subsequent intermediates in the dioxygen reduction reaction. To this end, we have developed techniques that allow us to carry 51 out low-frequency Raman detection under flow/flash conditions: here we report u(Fe-O) for an early intermediate in the cytochrome oxidase/dioxygen reaction. Consideration of data on dioxygen adducts of other heme proteins and model hemes indicates that this early intermediate is most likely the cytochrome a3-02 complex. Cytochrome oxidase was prepared from beef hearts according to a modified Hartzell and Beinert procedure and dissolved in 50 mM HEPES, 0.5% lauryl maltoside, pH 7.4. The fully-reduced, carbon monoxide-bound enzyme is prepared by anaerobic reduction with 4mM sodium ascorbase and 1 pM cytochrome c under CO. The absorption spectrum of the enzyme-C0 complex (inset Figure 3.2) shows a Soret maximum at 430 nm, as expected.4 The enzyme solution and the O2 saturated buffer solution are placed in separated syringes and driven through two eight-jet mixers in series at flow rates of 0.4 ml/min by using a DSAGE 355 syringe pump. The syringe pump drive is set so that 2.5 changes of sample occurred in the scattering volume for each pump-probe pair. The exit port of the mixer is designed to allow the oxidase solution to form a free jet in air in the scattering volume, which eliminates quartz scattering. The reaction starts when the mixed solution comes into the laser beam. Photodissociation of carbon monoxide from a2*a’3*. CO is followed by the reaction of a2*a23* with 02. The spectrum arises from intermediates ‘which are jproduced. while the photodissociated enzyme stays in the laser beam. 52 The time-resolved resonance Raman experiment employs two Quanta Ray, pulsed lasers with pulse widths of 10 ns and repetition rates of 10 Hz and a digital delay generator which provides programmable triggering for the flash lamps and Q-switches (Figure 3.1). This delay is continuously monitored with a photodiode to collect scattering laser light from a glass slide mounted in front of the sample and a Tektronix oscilloscope to observe the delay. The pump pulse from the first laser (532 nm, 1.3 mJ) is sufficient to dissociate the CO and initiate the oxidase/oxygen reaction. The probe wavelength (427 nm) is provided by pumping stilbene 420 with the third harmonic output (355 nm) of the second laser; The scattered radiation is collected with a SPEX 1459 Illuminator, dispersed in a SPEX 1877 Triplemate, and detected by an EG&G PARC 1420 diode array detector. Time-resolved resonance Raman spectra of cytochrome oxidase at 10 us subsequent to carbon monoxide photolysis in the presence of 02 are shown in Figure 3.2A-D. Spectrum E is that of the photodissociation product of the reduced carbonmonoxy enzyme (pump-probe delay = 10 ns.). Spectrum A, obtained with a low energy, defocused beam (0.3 mJ), is similar to the 10 ns spectrum with the exception that a new mode appears at 571 cm“. Figure 3.28 shows that the 571 cm"1 mode in the 16O2 spectrum is downshifted to 546 cm’1 mode when the experiment is repeated with I 802 . This allows us to assign it as an Fe-O stretching motion in the 53 Figure 3.1 Instrumental configuration used for pulsed, time-resolved resonance IRaman. measurements of flowing cytochrome oxidase samples prepared by rapid mixing. 54» mea .7 . 3 6261.590 in i 0% 665...: O :28 2033.88 - \ T1 I (J mmm<4 m>o 1/ \.\ Il— Lonesome em: oucso~cmns Rke~ Ream soocrmsa-~ an: Ae-ogaroo 22 Louamsou aaqxo w-~ 5S ovum Z40mmm NEE. Figure 3.2 55 Time-resolved resonance Raman spectra of cytochrome oxidase following initiation of the reaction with oxygen at room temperature. The energy of the 532 nm photolysis pump pulse was 1.3 mJ, sufficient to photolyze the enzyme-CO complex and initiate the 0,2 reduction reaction. The energy of the probe beam was 0.3 M for spectra A and BH and 1.0 mJ for spectra C- E. The repetition rate for both the pump and probe pulses (10 ns duration) was 10 Hz. The pump-probe delay was 10 us for the spectra A-D and 10 ns for the transient spectrum E. The accumulation time was 110 min for spectrum A, 70 min for spectrum B, 5 min for spectra C and D, and 15 min for spectrum E. RAMAN INTENSITY 56 427 nm EXCITATION . 'K 'r = 293 K I “‘1 td :10 HS 749 AUbUHUIANk t 685 .3" of CO Aosoronon 749 1 1 A, l .m- \‘\/‘N" \|\. . 600 g so A. 1602' Lop. 1 O 1802' Lop. v on 1602- H.P. 180,- HP. 10 ns. WAVENUMBERS 57 cytochrome a3/02 complex, as the 25 cm"1 shift is in agreement with that expected from the two-body harmonic oscillator approximation for Fe-Oz. Spectra C and D were obtained with relatively high energies (1mJ) and the absence of any modes located at 571 cm”1 (Figure 3.2C, 1“02) and 546 cm’1 (Figure 3.20, 1"02) indicates photodissociation of the oxy ligand, as was observed in the high frequency experiments , 2 ' 5 and further supports our assignment of the 571 cm"1 mode in the cytochrome oxidase/O2 complex. The most reasonable assignment of the 571 cm‘1 mode is that it arises from a cytochrome a32*-O2 complex. Such an assignment is consistent with the photolability of this species,2b but more important, it is in reasonable agreement with u(Fe’*-02) frequencies observed in other heme Fe2+-02 complexes. Table 3.1 summarizes several of these frequencies; the 571 cm“1 mode for the oxidase intermediate is similar to u(Fe2*-02) for several dioxygen- bound heme species. Several further points can be made from the Table 3.1. First, the u(Fe-02) in the oxidase intermediate is 5 cm'1 lower than that of the imidazole- heme a Fe2+-02 complex, despite the fact that the model compound reproduces the immediate coordination sphere that is expected to occur around the iron in the protein environment. We do not regard this decrease as mechanistically significant, as discussed below. Second, the oxidase species has a frequency that is close to the 58 Table 3.1 Vibrational frequencies for dioxygen bound complexes. u(Fe-O) ref Cytochrome Oxidase 571 this work Kb 567 11 Mb 570 9 HRP III 562 9 Im (heme a) Fern-O2 576 7 (mp) -Fe-O-O-Fe (TMP) 574 6 (Pip) (TPP)Fe2+-02 575 12 (TPP)Fe2+-02 509 12 Abbreviations: Hb, hemoglobin Mb, myoglobin HRP, horseradish Peroxidase Im, imidazole Pip, piperazine TPP, meso-tetraphenyl porphyrin 59 574 cm'1 observed for the iron-oxygen stretching frequency for the five coordinate p-peroxo dimer reported by Nakamoto and. co-workers.6 Despite the similarity in ‘those two frequencies, we nevertheless favor a cyt a3“-02 structure for the intermediate we detect. The basis for this lies in our expectation that cytochrome a3 will retain its proximal histidine ligand during catalysis and thus that a peroxy a3 species will have u(Fe-O) at significantly higher frequencies than the five coordinate model compound.7 A similar frequency increase in the iron-oxygen stretching frequency is apparent in Table 3.1 when one compares the five-coordinate (TPP)Fe“-O2 complex (u(Fe-O)=509 cm") to the six-coordinate (Pip) (TPP)Fe'”-02 species (u(Fe-O)=575 cm“) and is also apparent in heme ferryl oxo species.7. The v(Fe2*-O) frequency at 571 cm'1 indicates that the cytochrome a3 -02 complex is unperturbed by distal effects in. the cytochrome as/CuB binding site. ‘Weakening and rupture of the =0 bondza's'13 occurs subsequent to formation of the initial dioxygen-a32+ adduct. 60 REFERENCES a) Gibson, Q; Greenwood, C. Biochem. J. 1963, 86, 541. b) Greenwood, C.; Gibson, Q. J. Biol. Chem. 1967, 252, 1782. c) Greenwood, C.; Hill, B. Biochim. J. 1984, 2:18, 913. d) Hill, 8.; Greenwood C.; Nichols, P. Biochim. Biophys. Acta 1986, 851, 91. e) This technique has been modified for low temperature application by Chance and co-workers and has been used to study the oxidase/O2 reaction in frozen samples by several groups. See: Chance, B.: Saronio, C.; Leigh, J. S. J. Biol. Chem. 1975, 259, 9226. Clore, G. M.: Andreasson, L. E.; Karlsson, B.; Aasa, R.: and Malmstrom, B. G. Biochem J. 1980, 185, 139. Chan, S. 1.; Witt, S. N.; Blair, D. F. Chemica Scripta 1988, 285, 51. For review, see Naqui A. and Chance B. Ann. Rev. Biochem. 1986, 55, 137. a) Babcock, G.T.; Jean, J. M.: Johnston, L. N.; Palmer, G.; Woodruff, W. H. J. Am. Chem. Soc. 1984, 196, 8305. b) Babcock, G. T.: Jean, J. M.; Johnston, L. N.: Woodruff, W. H.: Palmer, G. J. Inorg. Biochem., 1985, g;, 243. Babcock, G. T. "Biological Applications of Raman Scattering", Spiro, T.G., ed., 1988, Vol. 3, 295. Vanneste, W. H. Biochem. 1966, 5, 838. Varotsis, C. and Babcock, G.T., unpublished results. Paeng, I.R.: Shiwaku, H.; Nakamoto, K. J. Am. Chem. Soc. 1988 110, 1995. 10. 11. 12. 13. 14. 61 Oertling, W.A.: Kean, R. T.; Wever, R.; Babcock, G.T. Inorg. Chem., 1990, in press. Blair, D.F.; Witt, S.N.: Chan, 8.1. J. Am. Chem. Soc., 1985, _Q1, 7389. Van Wart, H.: Zimmer, J. J. Biol. Chem. 1985, 2;, 8372. M. Wikstrom, K. Krab and M. Saraste, "Cytochrome Oxidase, A Synthesis", Acad. Press, New York, 1981. Brunner, H. Naturwissenschaften 1974, 61, 129. Nakamoto, K.; Paeng, I. R.: Kuroi, T.; Iosbe, T.: Oshio, H. J. Mol. Structure 1988, 182, 293. Wikstrom, M. Proc. Nat. Acad. Sci. USA 1981, 18, 4051. A loose point focus at the sample was used to decrease the probe power density. We estimate that the probe power density used to record spectra A and B was 12- fold less than in spectra C-E. CHAPTER 4 DIRECT DETECTION OF A DIOXYGEN ADDUCT 0F CYTOCHROME A, IN THE MIXED VALENCE CYTOCHROME OXIDASE/DIOXYGEN REACTION* 80W! Time-resolved resonance Raman spectra have been recorded during the reaction of mixed valence (a3+a3’+) cytochrome oxidase with dioxygen at room temperature. In the spectrum recorded at 10 ps subsequent to carbon monoxide photolysis a mode is observed at 572 cm'1 that shifts to 548 cm‘1 when the experiment is repeated with 1”0,2. The appearance of this mode is dependent upon the laser intensity used and it disappears at higher incident energies. The high frequency data, in conjunction with the mid-frequency data, allow us to assign the 572 cm‘1 mode to the Fe-O stretching vibration of the low-spin 02 adduct that forms in the mixed valence cytochrome oxidase/dioxygen *Varotsis, C.; Woodruff, W. H.; and Babcock, G. T., 1990, J. Biol. Chem., in press. 62 63 reaction. The 572 cm’1 v(Fe-02) frequency in the mixed valence/02 adduct is essentially identical to the 571 cm”1 frequency we measured for this mode during the reduction of 02 by the fully reduced enzyme (Varotsis, C., Woodruff, W.H. and Babcock, G.T. J. Am. Chem. Soc., 111, 6439-6440 (1989), 112, 1297, (1990)), which indicates that the 02- bound cytochrome a3 site is independent of the redox state of the cytochrome a/CuA pair. The photolabile oxy intermediate is replaced by photostable low- or intermediate-spin cytochrome a33+ with tz : 200 us. INTRODUCTION Cytochrome oxidase, the terminal enzyme complex of the mitochondrial respiratory chain, catalyzes the transfer of electrons from reduced cytochrome c to molecular oxygen. Four electrons are funnelled into 02 in this reaction to reduce it to 2H20; concomitantly, the free energy made available in the electron transfer reactions that occur during 02 reduction is used to pump protons from the matrix to the cytosolic side of the inner mitochondrial membrane. The free energy stored in the proton gradient is used ultimately to drive adenosine triphosphate formation(1). The enzyme isolated from bovine heart contains two hemes, cytochrome a and as, two associated copper atoms, CuA and CuB, magnesium and zinc(2,3). The cytochromes and copper atoms are redox active, but the roles of the other metals, if any, remain uncertain. The complex, intramolecular 64 electron transfers that occur between cytochrome a/CuA and the binuclear center cytochrome as/CuB, where the four- electron reduction of molecular oxygen to water occurs, have been studied by various spectroscopic techniques(4- 23). The focus in this work has been on characterizing intermediates that occur in dioxygen reduction, which is essential for elucidating the chemical mechanism of the redox processes catalyzed by the enzyme. Although the reaction between cytochrome oxidase and 02 occurs too quickly (tx = 1 ms) to be studied by conventional stopped flow techniques, Gibson and Greenwood(4) used photolysis of the cytochrome a32+'C0 complex of the enzyme in the presence of 02 to circumvent this limitation. Hill and Greenwood(8) showed that by starting from the mixed valence state in which only cytochrome a3 and CuB are reduced, partially reduced oxygen species were generated as intermediates on a microsecond- to-millisecond time scale at room temperature. Chance et a1. (9) developed techniques to study the reaction on a second-to-minute time scale at low temperatures. With this approach, Chance and coworkers(9), Clore et al.( 10), Chan et al.(11,12) and Denis(13) studied intermediates involved in dioxygen reduction and characterized them by using optical and EPR spectroscopies. The intermediate species formed in ‘the reaction, of the enzyme 'with. 02 at room temperature are, however, not yet well defined, although a variety of structures have been postulated (e.g. 1, 9, 10— 65 12, 15, 16, 19, 21, 22). Time-resolved resonance Raman spectroscopy provides the unique possibility of detecting out-of-plane ligand vibrational modes and thus the structure of the bound dioxygen species through isotopic labeling of the 02. Moreover, information concerning heme geometric and electronic properties is accessible. Monitoring the time evolution of these species provides the opportunity to explore the dioxygen reduction mechanism in detail. Babcock. et al.(18,19) adapted the Gibson/Greenwood technique and used time-resolved resonance Raman to study the reaction of fully and partially reduced oxidase with 02. Although instrumental limitations restricted data acquisition to the high frequency (>1200 cm“) region, they concluded that photolabile oxy species were the initial intermediates in both reactions. They postulated that these dioxygen adducts were replaced by non-photolabile species in which cytochrome a3 was oxidized with t% = 60 us and 200 ps in the fully reduced and mixed valence complexes, respectively. Ogura et al.(20,21) applied continuous wave, one laser Raman techniques and reported transient resonance Raman spectra of intermediates formed within 100 us after photolysis of the fully reduced C0 complex. Neither of these studies reported direct detection of the iron/bound oxygen ligand vibration. More recently, observation of oxygen isotope sensitive ligand vibrations have been reported. Rousseau et al. (22) 66 observed a mode at 477 cm'1 under intense laser illumination, which they attributed to Fe3+-0H‘ motion in a cytochrome aa/hydroxide adduct. Owing to uncertainties as to the physiological relevance of the experimental protocol, however, the catalytic significance of this species is unclear. In our own work (Varotsis et al., 1989), we have extended the flow/flash time-resolved approach so as to be able to observe the low frequency region of the Raman spectrum at early times in the 0.2/reduced cytochrome oxidase reaction. In our initial work, we observed an isotope sensitive mode at 571 cm‘1 at 10 us into the reduction reaction that we assigned to u(Fe2+—02) in the initial reduced cytochrome oxidase/dioxygen adduct. In the experiments reported here, we have applied our flow/flash, time—resolved Raman approach to study the reaction between mixed valence cytochrome oxidase and dioxygen. The use of this complex allows us to obtain further insight into the reaction of cytochrome oxidase with 02 at room temperature; the interaction of this species with 02 also serves as a benchmark for studies in which partially reduced oxygen species, such as hydrogen peroxide or superoxide, are added to the enzyme. Finally, it provides us with information as to whether the low temperature oxygenated species(9-14) are populated at room temperature. Our results indicate that the dioxygen adduct of cytochrome a32+ is formed in the mixed valence enzyme 67 and that the redox states of cytochrome a and of CuA do not influence the vibrational characteristics of this species. EXPERIMENTAL PROCEDURES Cytochrome oxidase was prepared from beef hearts by using a modified Hartzell-Beinert preparation(24) and was frozen under liquid N2 until ready for use. The enzyme was solubilized in 50 mM HEPES (4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid) at pH 7.4 with 0.5% dodecyl fi-D-maltoside. The mixed valence/C0 enzyme was prepared by exposing an anaerobic solution of the resting enzyme to carbon monoxide for five hours(8). The absorption spectrum of the mixed valence'CO complex (inset, Fig. 3) shows a Soret. maximum at 430 nm. as expected(25). The enzyme solution and an oxygen saturated buffer solution (50 mM HEPES, 0.5% dodecyl B-D-maltoside, pH 7.4, 1.2 mM [02]) were placed in separate syringes and driven through two eight-jet tangential Gibson type mixers in series at a flow rate of 0.4 ml/min with a SAGE 355 syringe pump. The syringe pump drive was set so that 2.5 changes of sample occurred in the scattering volume for each pump-probe pulse pair. The exit port of the mixer is designed to allow the oxidase solution (40 pH after mixing) to form a free jet in air in the scattering volume, which eliminates quartz scattering(26). The reaction is initiated when the mixed valence'C0/02 solution comes into the scattering volume. Photodissociation of carbon monoxide from a3+a32+'C0 is 68 followed by reaction of a3+a32+ with oxygen. In the time- resolved resonance Raman experiment, two Quanta Ray DCR 2A pulsed lasers with pulse widths of 10 ns and repetition rates of 10 Hz were used. A digital delay generator (Stanford Research Systems, Inc., Model DG 535) provided programmable triggers for the flash lamps and Q-switches. The pump pulse from the first laser (532 nm, 1.3 mJ) was sufficient to dissociate C0 and initiate the oxidase/oxygen reaction. The probe wavelength (427 nm) was provided by pumping stilbene 420 with the third harmonic output (355 nm) of the second laser. The scattered radiation was collected with a SPEX 1459 Illuminator, dispersed in a SPEX 1877 Triplemate, and detected by an EG&G PAR 1420 diode array detector. A linear sloping background was subtracted from the resonance Raman spectra in Figure 3 and Figure 4, but no smoothing was done. Optical absorption spectra were obtained with a Perkin-Elmer Lamda 5 uv-visible spectrophotometer. RESULTS In Figure 1 (B-F) we present high frequency time- resolved resonance Raman spectra of mixed valence (a3+a2+) cytochrome oxidase at various delay times (10 ps - 500 p5) subsequent to carbon monoxide photolysis in the presence of 02. These spectra, as well as those of the 10 ns photoproduct (Fig. 1A) and resting enzyme (Fig. 1G), were recorded with ~1 mJ/pulse. With 427 nm excitation, the Figure 4.1 69 Time-resolved resonance Raman spectra of mixed valence cytochrome oxidase at the indicated times following initiation of the reaction with oxygen at room temperature. The energy of the 532 nm photolysis pump pulse was 1.3 mJ, sufficient to photolyze the enzyme -C0 complex and initiate the 02 reduction reaction. The energy of the probe beam was 1.0 mJ for all spectra. The spectrum of the resting enzyme was obtained by using the probe beam onLy. The repetition rate for both the pump and probe pulses (10 ns duration) was 10 Hz. The accumulation time was 15 min. for each spectrum. 70 427 nm EXCITATION 1 mJ/pulse «2.: cl 9 .m m e R G. F. 500113 E. 20011: C. 50115 A1003 >._..mzm._.z_ Z._._mzm._.z_ Z<_2 Hb 567 (39) Mb 570 (40) HRP III 562 (40) Im (heme a) Fe2+ - 02 576 (45) (TMP) Fe - o — o — Fe (TMP) 574 (41) (Pip) (TPP) Fe2+ - 02 575 (42) (TPP) Fe2+ - 02 509 (42) aAbbreviations: a3+ a32+, mixed valence cytochrome oxidase; a2+ a32+ fully reduced cytochrome oxidase; Hb, hemoglobin, Mb, myoglobin; HRP, horseradish peroxidase: Im, imidazole; Pip, piperazine; Fe(TMP), tetramesitylporphyrinatoiron; TPP, meso- tetraphenylporphyrin. 88 oxidase that was substantially higher than that observed for the C0 complexes of other heme proteins. A clear difference between the mixed valence and fully reduced oxy cytochrome a3 species, however, is their lifetimes. Hill et al.(6,7) reported that the oxy intermediate in the fully reduced enzyme was undetectable owing to its short lifetime. We interpreted our earlier Raman data(18) to indicate that the oxy species was detectable and that it decayed with a halftime consistent with a rate constant of ~ 2 x 10‘ s’1 at 25°C. 0rii’s optical data(16) and our recent Raman data on u(Fe-0) in the a32+/02 adduct confirm the detectability of this species: there is some disagreement, however, as to its decay rate constant and this merits further study. It is clear, however, that the mixed valence oxy species reacts to form subsequent species at a substantially slower rate than the fully reduced enzyme. Hill et al.(8) and 0rii(16) detected this species optically and we estimated a decay rate constant of ~ 3.5 x 103 s"1 in our earlier Raman work(19), in reasonable agreement with the value of 6 x 103 s’1 reported by Hill et al.(8). The data in Figure 1 are consistent with these rate constants as photostable intermediates are formed in the 100-500 #5 time regime. In the 500 p8 spectrum the position of u at 1371 cm“1 and of 4 the formyl at 1674 cm”1 indicates that cytochrome a3 is oxidized. Moreover, its V2 vibration has shifted underneath the v2 vibration of low-spin cytochrome a3+ at 89 1589 cm“. This indicates the formation of a low-spin or intermediate-spin(30) cytochrome a33+ complex at 500 us in agreement with the optical data reported by Greenwood et al.(8). The relationship of this species to trapped intermediates in the mixed valence/02 reaction, to species that occur at longer times, and to forms of the enzyme that occur when the oxidized enzyme reacts with H202 is under investigation. 90 REFERENCES Wikstrom, M., Krab, K. and Saraste, M. (1981), "Cytochrome Oxidase, A Synthesis", Acad. Press, New York. Caughey, W.S., Smythe, G.A., O’Keeffe, D.H., Maskasky, J. and Smith, M.L. (1975) J. Biol. Chem. 250, 7602- 7622. Einasdottir, 0. and Caughey, W.S. (1984) Biochem. Biophys. Res. Commun. 124, 836-842. Gibson, Q. and Greenwood C. (1963) Biochem. J. 86, 541-554. Greenwood, C. and Gibson, Q. (1967) J. Biol. Chem. 242, 1782-1787. Hill, B. and Greenwood, C. (1984) Biochem. J. 218, 913-921. Hill, B., Greenwood C. and Nichols, P. (1986) Biochim. Biophys. Acta 853, 91-113. Hill, B.C. and Greenwood, C. (1983) Biochem. J. 215, 659-667. Chance, B., Saronio, C. and Leigh, J.S. (1975) J. Biol. Chem. 250, 9226-9237. lo. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 91 Clore, G.M., Andreasson, L.E., Karlsson, B., Aasa, R. and Malmstrom, B.G. (1980) Biochem. J. 185, 139-154. Chan, S.I., Witt, S.N. and Blair, D.F. (1988) Chemica Scripta 28A, 51-56. Blair, D.F., Witt, S.N. and Chan, 8.1. (1985) J. Am. Chem. Soc. 107, 7389-7399. Denis, M. (1981) Biochim. Biophys. Acta, 634, 30-40. Chance, B., Saronio, C. and Leigh, J.S. (1975) Proc. Natl. Acad. Sci. U.S.A. 72, 1635-1640. 0rii, Y. (1984) J. Biol. Chem. 259, 7187-7190. 0rii, Y. (1988) Annals of the New York Academy of Science 550, 105-117. Babcock, G.T. (1988) "Biological Applications of Raman Scattering" (Spiro, T.G., ed.), Vol. 3, 295-346. Babcock, G.T., Jean, J.M., Johnston, L.N., Palmer, G. and Woodruff, W.H. (1984) J. Am. Chem. Soc. 106, 8305- 8306. Babcock, G.T., Jean, J.M., Johnston, L.N., Woodruff, W.H. and Palmer, G. (1985) J. Inorg. Biochem. 23, 243- 251. Ogura, T., Yoshikawa, S. and Kitagawa, T. (1985) Biochim. Biophys. Acta 832, 220-223. Ogura, T., Yoshikawa, S. and Kitagawa, T. (1989) Biochemistry 28, 8022-8027. Han, S., Ching, Y.-C., and Rousseau, D.L. (1989) J. Biol. Chem. 264, 6604-6607. Varotsis, C., Woodruff, W.H. and Babcock, G.T. (1989) 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 92 J. Am. Chem. Soc. 111, 6439-6440: (1990) 112, 1297. Hartzell, R. and Beinert, H. (1974) Biochim. Biophys. Acta. 368, 318-338. Vanneste, W.H. (1966) Biochemistry 5, 838-847. Varotsis, C., Oertling, W.A. and Babcock, G.T. (1990) App. Spectroscopy, in press. Babcock, G.T., 'Callahan, P.M., Ondrias, M.R. and Salmeen, I. (1981) Biochemistry 20, 959-966. ‘Woodruff, W.H., Dallinger, R.F., Antalis, T.M. and Palmer, G. (1981) Biochemistry 20, 1332-1338. Argade, P.V., Ching, Y.-C. and Rousseau, D.L. (1986) Biophys. J. 50, 613-620. Carter, K.R,, Antalis, T.M., Palmer, G.M., Ferris, N.S. and Woodruff, W.H. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 1652-1655. Ching, Y.-C., Argade, P. and Rousseau, D. (1985) Biochemistry 24, 4938-4946. Rousseau, D.L., Singh, S., Ching, Y.-C. and Sassaroli, M. (1988) J. Biol. Chem. 263, 5681-5685. Babcock, G.T. and Chang, C.K. (1979) FEBS Lett. 97, 358-362. Wikstrom, M. (1981) Proc. Nat. Acad. Sci. U.S.A. 78, 4051-4053. Wikstrom, M. (1989) Nature 338, 776-778. Alben, J.0., Moh, P.P., Fiamingo, F.G. and Altschuld, R.A. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 234-237. Findsen, E.W., Centeno, J., Babcock, G.T. and Ondrias, 38. 39. 40. 41. 42. 43. 44. 45. 93 M.R. (1987) J. Am. Chem. Soc. 109, 5367-5372. Dyer, R.B., Einarsdottir, 0., Killough, P.M., Lopez- Garriga, J.J. and Woodruff, W.H. (1989) J. Am. Chem. Soc. 111, 7657-7659. Brunner, H. (1974) Naturwissenschaften 61, 129-130. Van Wart, H. and Zimmer, J. (1985) J. Biol. Chem. 250, 8372-8377. Paeng, I.R., Shiwaku, H. and Nakamoto, K. (1988) J. Am. Chem. Soc. 110, 1995-1997. Nakamoto, K., Paeng, I.R., Kuroi, T., Iosbe, T. and Oshio, H. (1988) J. Mol. Struct. 189, 293-300. Einarsdottir, Olof, Choc, M.G., Weldon, S. and Caughey, W. (1988) J. Biol. Chem. 263, 13641-13654. Argade, P.V., Ching, Y.-C. and Rousseau, D.L. (1984) Science, 225, 329-331. Oertling, W.A., Kean, R.T., Wever, R., and Babcock, G.T. (1990) Inorg. Chem., in press. (HIAPTFJKS LATE APPEARANCE OF THE u (Few: O) VIBRATION FROM A FERRYL-OXO INTERMEDIATE IN THE CYTOCHROME OXIDASE/DIOXYGEN REACTION" SUMMARY Time-resolved resonance Raman spectra have been recorded during the reaction of fully-reduced (a2*a32*) cytochrome oxidase with dioxygen at room temperature. In the spectrum recorded at 800 ps subsequent to carbon- monoxide photolysis a mode is observed at 790 cm'1 that shifts to 755 cm“1 when the experiment is repeated with 1802. The frequency of this vibration and the magnitude of the 1802 isotopic frequency shift lead us to assign the 790 cm'1 mode to the FeIV=0 stretching vibration of a ferryl-oxo cytochrome as intermediate that occurs in the reaction of fully-reduced cytochrome oxidase with dioxygen. The appearance and vibrational frequency of this mode were not affected when D20 was used as a solvent. This result suggests that the ferryl-oxo intermediate is not hydrogen- *Varotsis, C. and Babcock, G. T. (1990), submitted. 94 95 bonded. We have also recorded Raman spectra in the high- frequency (1000-1700 cm“) region during the oxidase/02 reaction that show that the oxidation of cytochrome a2+ is biphasic. The faster phase is complete within 100 ps and is followed by a pdateau region in which no further oxidation of cytochrome a occurs. The plateau persists to ~500 p8 and is followed by the second phase of oxidation. These results on the kinetics of the redox activity of cytochrome a are consistent with the branched pathway discussed by Hill, Greenwood, and Nichols (Biochim. Biophys. Acta 8_;, 91-113 (1986)) for the oxidation of reduced cytochrome oxidase by 02 at room temperature. Within the context of this scheme, the ferryl-oxo intermediate we have observed arises as the fourth. and final electron enters the dioxygen reduction site. INTRODUCTION In the mitochondrial electron transport chain of eucaryotic organisms, cytochrome oxidase functions as the oxygen-activating enzyme. The overall reaction catalyzed by the enzyme is the rapid reduction of dioxygen to water. The free energy released in the electron-transfer reactions that occur during 02 reduction is conserved as an electrochemical proton gradient across the inner mitochondrial membrane and is used ultimately for adenosine triphosphate synthesis(Wikstrom et al., 1981). Mitochondrial cytochrome c oxidase contains two hemes, cytochrome a and a3, and two 96 copper atoms, designated CuA and CuB. The low potential sites, cytochrome a and CuA, function together in the sense that they oxidize cytochrome c and subsequently transfer the reducing equivalents to the high-potential binuclear site, which contains cytochrome a3 and CuB, where oxygen binding and reduction to H20 take place. Although the cytochrome oxidase/dioxygen reaction has been extensively studied by various spectroscopic techniques at room and low temperatures(Gibson and Greenwood, 1963; Hill and Greenwood, 1983, 1984: Hill, et al., 1986; 0rii, 1984, 1988; Oliveberg et al., 1990; Chance et al., 1975a,b: Clore et al., 1980; Blair et al., 1985; Chan et al., 1988), the reaction mechanism is not yet fully understood. For example, the precise chemical identity of various intermediates and the jpathway(s) by’ which electrons are transferred to the binuclear site are important mechanistic questions for which further information is required. The reaction is rapid under physiological conditions (ty. = 1 ms)(Gibson and Greenwood, 1963; Greenwood and Gibson, 1967). Nonetheless, the room-temperature flow/flash technique developed by Gibson and Greenwood(1963), has provided a way to resolve the reaction kinetically. Hill and Greenwood(1983; 1984), 0rii(1984; 1988) and.<01iveberg' et al.(1990) used this technique to show that partially-reduced intermediates were generated at room temperature. Chance et al.(1975a,b), Clore et al.(1980), Blair et al.(1985), and Chan et al.(1988) studied intermediates involved in dioxygen 97 reduction. at low 'temperatures and. characterized them Iby using optical and EPR spectroscopies. Wikstrom(1981, 1989) was able to reverse electron flow through the enzyme to trap what was postulated to be a peroxy intermediate that precedes ferryl-oxo formation in the reaction scheme he proposed. Wikstrom(1989), Blair et al.(1985), and Chan et al.(1988) have postulated heterolytic cleavage of the 0=0 bond to form a cytochrome a3 ferryl-oxo (FeIV=0) intermediate as electrons from the cytochrome a/CuA sites enter the oxygen-bound binuclear site. The formation and decay rates of intermediate species in this process are not well determined at room temperature, although recently 0rii(1988) proposed that a ferryl-oxo species occurs at ~100 ps in the oxidase/dioxygen reaction, consistent with the rapid oxidation of a fraction of the a/CuA centers that has been inferred from optical measurements(Hill and Greenwood, 1984; Hill et al., 1986). Furthermore, the protonation steps involved, as well as the specific structure of the cytochrome a3/02 adducts, are not yet well defined, although a variety of structures have been postulated(Hill and Greenwood, 1983; 1984; Hill et al. 1986; Chance et al., 1975a,b; Clore et al., 1980; Blair et al., 1983; 1985; Chan et al., 1988; Oliveberg et al., 1990). Raman spectroscopy is a structure-specific vibrational technique and, relative to optical or EPR spectroscopies, has the potential to provide more detailed information on 98 the intermediate structures that occur during the oxidation of cytochrome oxidase by 02. Babcock et al.(1984; 1985), using pulsed excitation, and later Ogura et al.(1985; 1989), with. continuous-wave lasers, showed. that. a 'time-resolved Raman approach was feasible. Although a photolabile cytochrome a3“-02 species was postulated as the initial intermediate in 02 reduction in this work (Babcock et al., 1984), the measurements were confined to the high-frequency region. Recently, Varotsis et al.(1989; 1990a), Han et al.(1990a; 1990b), and Ogura et al.(1990) have confirmed the occurrence of this oxy species by monitoring the u(Fea32*- 02) vibration directly in the reaction of both the mixed valence and fully-reduced enzyme with 02. In the experiments reported here, we have continued the two-color, pulsed irradiation. Raman approach to intermediates that occur at later times in the fully-reduced cytochrome oxidase/dioxygen reaction. The pulsed technique is particularly useful in this application, as the time resolution is determined by the programmable time delay between the short (10 ns) pump and probe laser flashes. This contrasts with a c. w. laser approach in which the time resolution is determined by the residence time of the reacting sample in the beam. The latter approach becomes ambiguous kinetically as this residence time increases. Our results indicate that cytochrome a is oxidized in a biphasic manner. Partial oxidation occurs at early' times after mixing (t<100 #8), consistent with optical data reported by 99 several workers (Hill and Greenwood, 1984; Hill et al. 1986; 0rii, 1988; Brunori and Gibson, 1983; Oliveberg et al., 1990) and with branched schemes for the overall reaction (Hill and Greenwood, 1984; Hill et a1. 1986; Clore et al., 1980; Blair et al., 1983; 1985; Chan et al., 1988). Despite this agreement with the absorption experiments, there is no clear indication of a conventional ferryl-oxo species at this three-electron level of reduction. Instead, we detect u(FeIV=0) at later times in the reaction (t-BOO us), as the more slowly-reacting fraction of cytochrome a is oxidized. The frequency of ‘this ferryl-oxo stretching' motion, 790 cm“, is insensitive to H20/D20 exchange, suggesting that it is not hydrogen bonded. Within the context of the branched kinetic schemes that have been developed from the earlier spectroscopic work, the appearance of this ferryl accompanies the arrival of the fourth electron in the oxygen-bound binuclear site. EXPERIMENTAL PROCEDURES Cytochrome oxidase was prepared from beef hearts by using a modified Hartzell-Beinert(1974) preparation and was frozen under liquid N2 until ready for use. The enzyme was solubilized in 50 mM HEPES (4-(2-hydroxyethyl)-1- piperazineethanesultonic acid) at pH 7.4 with 0.5% dodecyl fi-D-maltoside. 'The absorption spectrum. of resting cytochrome oxidase shows a maximum at 421 nm, which is characteristic of rapidly-reacting enzyme. The fully- 100 reduced, carbon-monoxide-bound enzyme was prepared by anaerobic reduction with 4 mM sodium ascorbate and 1 pM cytochrome c under C0 and shows a Soret maximum at 430 nm, as expected (Vanneste, 1966). The pD of solutions prepared in D20 buffer was measured by using a pH meter and assuming pD=pH (observed) +0.4. The experimental techniques used for the measurements of time-resolved Raman spectra have already been reported(Varotsis, et al., 1989; 1990agb). The probe wavelength (441 nm) was provided by pumping coumarin 440 with the third harmonic output (355 nm) of a quanta-ray DCR 2A pulsed laser. RESULTS Figure 1 shows high-frequency resonance Raman spectra of fully-reduced (a’*a32*) cytochrome. oxidase Tat 'various delay times subsequent to carbon-monoxide photolysis in the presence of 02. With 427 nm excitation, cytochrome a2* and cytochrome a32+ contribute roughly equally to the resonance Raman spectrum of the fully-reduced enzyme through an 0-1 enhancement mechanism because of their coincident absorption maxima near 443 nm(Babcock et al., 1981; Woodruff et al., 1981; Argade et al., 1986; Babcock, 1988). Excitation at 427 nm also enhances vibrations of cytochrome a3+ and those of oxygenated cytochrome as, which have absorption maxima in the 427-nm range. In the 10-ns photoproduct spectrum (Figure 1A), the oxidation state marker is at 1355 cm“, establishing that both cytochromes are in the ferrous state. Figure 5.1 101 Time-resolved resonance Raman spectra of fully-reduced cytochrome oxidase at the indicated times. The energy of the 532-nm photolysis pump/pulse was 1.3 mJ, sufficient to photolyze the enzyme-C0 complex and initiate the 02-reduction reaction. The energy of the 427-nm probe beam was 0.8 mJ for spectra A and F and 0.3 mJ for Spectra B- E. The repetition rate for both the pump and probe pulses (lo-ns duration) was 10 Hz. The accumulation time was 15 min for Spectra A and F and 50 min for Spectra B-E. The enzyme concentration was 50 pM after mixing, pH 7.4. 102 427 nm EXCITATION T = 298 K run ... >._._mzm._.z_ Z