The Role of the Cross-link His-Tyr in the Functional Properties of the Binuclear Center in Cytochrome c Oxidase*

Resonance Raman and Fourier transform infrared spectroscopies have been used to study the aa 3 -type cytochrome c oxidase and the Y280H mutant from Paracoccus denitrificans . The stability of the binuclear center in the absence of the Tyr 280 -His 276 cross-link is not compromised since heme a 3 retains the same proximal environment, spin, and coordination state as in the wild type enzyme in both the oxidized and reduced states. We observe two C-O modes in the Y280H mutant at 1966 and 1975 cm (cid:1) 1 . The 1975 cm (cid:1) 1 mode is assigned to a (cid:2) -form and represents a structure of the active site in which Cu B exerts a steric effect on the heme a 3 -bound CO. Therefore, the role of the cross-link is to fix Cu B in a certain configuration and distance from heme a 3 , and not to allow histidine ligands to coordinate to Cu B rather than to heme a 3 , rendering the enzyme inactive, as proposed recently (Das, T. K., Pecoraro, C., Tomson, F. L., Gennis, R. B., and Rousseau, D. L. (1998) Biochemistry 37, 14471–14476). The results provide solid evidence that in the

Cytochrome c oxidase (CcO) 1 couples the one-electron oxidation of cytochrome c to the four-electron reduction of molecular oxygen and links these electron transfers to proton translocation across the inner mitochondrial membrane, or the bacterial cytoplasmic membrane, respectively (1)(2)(3)(4)(5)(6). The enzyme contains four redox-centers; two hemes a and three associated copper atoms. Electron injection from cytochrome c to the homo-dinuclear copper center, Cu A , is followed by intramolecular electron transfer, via the low-spin heme a, to the binuclear center which contains a high spin heme a 3 and a Cu B atom. The latter two species serve as the catalytic site where O 2 is reduced to H 2 O. The free energy released in the electron-transfer reactions is conserved as an electrochemical proton gradient across the inner mitochondrial membrane and is used ultimately for ATP synthesis.
The crystal structures of mammalian CcO (7)(8)(9) and of the soil bacterium Paracoccus denitrificans (10,11) have been determined providing deep insight in the structural properties of the enzyme. The properties of the binuclear center are of particular importance, since the heme a 3 /Cu B center is the site where dioxygen reduction takes place and is the most probable site of the proton-electron coupling (2). One of the unique properties of the binuclear center that were determined by the crystal structure is the covalent link of Tyr 280 with one of the three histidine ligands of Cu B , namely His 276 . (If not stated otherwise, we adopt the residue numbering of P. denitrificans.) This specific tyrosine which is located at the end of the proton K-channel is highly conserved among the heme-copper oxidases and since its discovery it has been proposed to posses an important structural as well as functional role. On the basis of the properties of the Y 3 F mutant of Rhodobacter sphaeroides, analyzed by resonance Raman spectroscopy (RR), Rousseau and co-workers (12) proposed that the tyrosine-histidine crosslinking stabilizes the binuclear center. They suggested that in the absence of the His-Tyr cross-link one of the histidines normally bound to Cu B coordinates to the heme a 3 , leaving the binuclear center severely disrupted, and rendering the enzyme inactive. Based on the crystal structure of bovine CcO, Yoshikawa and co-workers (9) proposed a proton transfer mechanism from this tyrosine to ferric peroxide to generate a hydroperoxo adduct, and subsequently the electron transfer from Cu B 1ϩ via the cross-link would cleave the O-O bond of the ferric hydroperoxide. Moreover, other groups have suggested that the tyrosine can serve as a hydrogen atom donor during the cytochrome oxidase/O 2 reaction (3-4, 6, 11, 13). Recently, Babcock and co-workers (14) proposed that Tyr 280 is the source for both the proton and the electron required in the O-O bond cleavage, and Michel and co-workers (15) proposed that the EPR signal (g iso ϳ 2.0055) they observed in the cytochrome aa 3 /H 2 O 2 (P. denitrificans) reaction originated from the cross-linked tyrosine.
Structural information of the heme-Cu B center have been determined from studies of the CO-bound adducts (16 -28). In addition to revealing insights concerning the electronic and steric nature of the heme pocket, CO photodissociation studies provided a powerful tool for studying the dynamics and coordination chemistry in the heme-Cu B pocket after CO photolysis (29 -34). RR spectroscopy is a well adapted technique in the study of terminal oxidases, as it enables us to selectively enhance the vibrational modes of the hemes without interference from the protein matrix and thus identify their oxidation, spin, and ligation state by using established marker bands (35)(36)(37)(38)(39)(40)(41). Both RR and FTIR spectroscopy have been employed in the study of the carbonmonoxy derivatives of cytochrome aa 3 . The vibrational frequencies of the FeCO unit obtained by the two spectroscopic techniques have been identified in different types of heme-copper oxidases, revealing different conformations of the active site (16 -23). The two major conformers found are termed as ␣and ␤-forms and although their functional significance and the origin for the splitting have not been established, it has been demonstrated that in the ␣-form the frequencies of the (Fe-CO) and (C-O) deviate from the inverse linear curve that exists between the frequencies of these two modes in histidine-coordinated heme proteins, while in the ␤-form those frequencies are placed on the curve (17,(22)(23)(24)(25)(26)(27)(28). Recently, from the observed pH-dependent conformational changes in the binuclear site it was postulated that the different structures result from a change in the position of the Cu B atom with respect to the CO due to the presence of one or more ionizable groups (23). Furthermore, it was suggested that the possible candidates are the cross-linked, conserved tyrosine that is adjacent to the oxygen-binding pocket or one of the histidines that coordinate Cu B .
In an effort to gain additional information on the role of Tyr 280 in the catalytic function of CcO we have characterized the wild type and histidine mutant (Y280H) of CcO from P. denitrificans in the oxidized, reduced, and CO-bound forms by optical absorption, RR, and FTIR spectroscopies. Our studies show that the Cu B modification by the Y280H mutant results in only slight perturbation of the formyl group of heme a 3 . Therefore, the role of the cross-link is not to allow His 276 (His 240 in bovine, His 284 in Rb. sphaeroides) to coordinate to Cu B instead of the heme iron atom, as previously suggested (12), but to hold Cu B in a certain configuration and distance from heme a 3 . Without the cross-linking of Tyr 280 and His 276 , the heme pocket retains its active configuration that allows O 2 binding to heme a 3 . Upon direct mixing of O 2 to the CO-bound mixed-valence wild-type and Y280H enzymes, oxygenated species with similar Soret maxima at 438 nm are formed, which decay to the resting form of the enzymes.

EXPERIMENTAL PROCEDURES
Wild-type and mutant CcO was purified from P. denitrificans according to published procedures (42,43). The activity of wild-type and mutant CcO has been reported (43). Mammalian CcO was isolated from beef hearts (44). The samples were concentrated to 100 -150 M in 50 mM Hepes, pH 7.4, containing 0.1% dodecyl ␤-D-maltoside and stored in liquid nitrogen until use. The fully reduced CO derivative was prepared by flushing CO gas anaerobically to dithionite-reduced enzyme. The mixed-valence CO-bound enzyme was prepared by exposing an anaerobic solution of the resting enzyme to CO for 10 h. RR spectra were obtained from 30 to 40 M samples in an anaerobic cylindrical quartz spinning cell. The RR spectra were acquired by using a SPEX 1877 triplemate with an EG&G (model 1530-CUV-1024S) CCD detector. A Coherent Innova K-90 Krypton ion laser was used to provide the excitation wavelength of 413.1 nm. A Coherent 590 dye laser connected with a Coherent Innova 200 argon laser was used to provide the excitation wavelength of 431 nm. The power incident on the CcO sample was typically 4 -6 mW. FTIR spectra were obtained from 200 to 300 M samples with a Bruker Equinox 55 FTIR spectrometer equipped with liquid nitrogen-cooled MCT detector. The samples were loaded anaerobically into a cell with CaF 2 windows and a 0.025-mm spacer. The spectra were obtained as difference, using the buffer as background, and each spectrum is the average of 1000 scans. The spectral resolution used for the FTIR measurements was 2 cm Ϫ1 for the wild-type CO spectrum and 4 cm Ϫ1 for the Y280H mutant, respectively. Optical absorbance spectra were recorded before and after FTIR and Raman measurements to assess sample stability with a PerkinElmer Life Science Lamda 20 UV-visible spectrometer.

RESULTS AND DISCUSSION
The optical absorption spectra of resting (as isolated) and fully reduced aa 3 from P. denitrificans display maxima at 423 and 598 nm in resting form (Fig. 1A, trace a), and at 444 and 605 nm in the reduced form (Fig. 1A, trace b). The difference CO-bound-reduced spectrum displays a positive band at 430 nm and a shoulder at 416 nm with a trough at 449 nm (Fig. 1A, trace c). The optical spectrum of the resting Y280H mutant shown in Fig. 1B (trace a) shows in addition to the 423-nm band a shoulder at 441 nm in the Soret region and an increased absorption of the 604-nm band. This indicates that a sizeable percentage of heme a (ϳ30%) is reduced in the mutant. Addition of dithionite to Y280H (Fig. 1B, trace b) shifts the maxima at 441 and 604 nm, consistent with the maxima of the fully reduced wild-type enzyme. The difference spectrum of the reduced CO-bound minus reduced form is characteristic of CO binding to heme a 3 , as denoted by the peaks at 432 and 592 nm (Fig. 1B, trace c).
The high frequency region of the RR spectrum of the resting and reduced wild-type, which are shown in Fig. 2A (traces a and b), are in agreement with those previously reported for heme-copper oxidases (35-37, 39, 40). In the spectrum of the resting enzyme, the oxidation state marker v 4 is at 1371, establishing that both hemes are in the ferric (Fe 3ϩ ) state. The modes at 1477 and 1498 cm Ϫ1 originating from 3 of high-spin heme a 3 and low spin heme a, respectively, indicate that both hemes are six-coordinate. This is also consistent with structural data which indicate that in this form of the enzyme heme a is coordinated by two histidine ligands and that there is a bridging ligand between heme a 3 and Cu B (10,11). The core expansion region shows two vibrations at 1572 cm Ϫ1 (high-spin heme a 3 3ϩ ) and 1584 cm Ϫ1 (low-spin heme a 3ϩ ). The 1612 and 1635 cm Ϫ1 modes arise from 10 of heme a 3 3ϩ and heme a 3ϩ , respectively. The 1646 and 1671 cm Ϫ1 modes have been assigned as the CϭO stretching vibrations of the formyl groups (ϪCHO) of heme a 3ϩ and heme a 3 3ϩ , respectively (39,40). The spectrum from the fully reduced enzyme (Fig. 1A, trace b) has the oxidation state marker, 4 , at 1359 cm Ϫ1 indicating that both hemes are in the ferrous state. The modes at 1464 and 1490 cm Ϫ1 originate from the ferrous, five-coordinate high spin heme a 3 2ϩ and the ferrous six-coordinate low-spin heme a, respectively. The mode at 1519 originates from 19 of heme a 2ϩ . The modes at 1569 and 1584 cm Ϫ1 originate from 2 of heme a 3 2ϩ and heme a, respectively. The 1612 and 1662 have been assigned as the CϭO stretching vibrations of the formyl group (ϪCHO) of heme a 2ϩ and heme a 3 2ϩ , respectively (35,40). The 1623 cm Ϫ1 mode arises from the CϭC stretching vibration of heme a 2ϩ and heme a 3 2ϩ . The spectral perturbations caused by the Y280H mutation are very limited except for the decrease in intensities of the heme a 3ϩ modes and the concomitant appearance of modes originating from a fraction of heme a 2ϩ (Fig. 2B,  trace a). The reduction of heme a is apparent from the increase in intensities of 4 at 1356, 11 at 1519, 2 at 1584, 10 at 1612 cm Ϫ1 , as well the decrease in intensity of the formyl stretching vibration at 1646 cm Ϫ1 . All the modes associated with the high spin heme a 3 3ϩ are very similar to those of the wild-type enzyme. The only exception is the reduced intensity of the formyl stretching vibration at 1671 cm Ϫ1 . The RR spectrum of the fully reduced Y280H enzyme (Fig. 2B, trace b) is very similar to that of the fully reduced wild-type, with the exception of the formyl vibration at 1662 cm Ϫ1 which is weaker in the mutant.
No unusual stereochemical influences on heme a 3 are apparent in the Y280H mutant that would modify the heme a 3 ligand-binding site. In addition, no evidence for any histidine coordination from Cu B to heme a 3 is apparent. Thus, our data do not support the conclusions of Das et al. (12) that, without the cross-linking of Tyr 288 and His 284 (R. sphaeroides residue numbering), the heme pocket becomes severely disrupted and one of the histidines bound to Cu B coordinates to heme a 3 , lowering its redox potential. Fig. 3 shows the low-frequency RR spectra of the fully reduced wild-type and Y280H mutant, and that of the wild-type CO-bound form. The RR spectrum of the fully reduced enzyme (trace a) is characterized by the Fe-His stretching mode at 220 cm Ϫ1 and porphyrin modes of both hemes a and a 3 . Very similar spectra were obtained for the Y280H mutant as shown in trace b. The Fe-His stretching mode is 6 cm Ϫ1 higher in P. denitrificans than it is in CcO (36). The binding of CO to heme a 3 2ϩ is accompanied by the disappearance of the Fe-His stretching vibration at 220 cm Ϫ1 and of the porphyrin mode at 369 cm Ϫ1 (trace c). In the 400 -600 cm Ϫ1 region of the RR spectrum of the CO complex, one frequency for (Fe-CO) is detected at 517 cm Ϫ1 . Assignment of this frequency is confirmed by isotope ( 13 CO) replacement experiment (trace d), where the corre-sponding line appears at 513 cm Ϫ1 , close to the value expected for a two-harmonic oscillator between iron and CO. The difference spectrum (trace e) confirms the above assignment. In the 13 CO-bound adduct we also detect a line at 559 cm Ϫ1 . Although the presence of a porphyrin mode at ϳ580 cm Ϫ1 partially obscures band assignment with 12 CO in this region, the difference spectrum shows that the 559 cm Ϫ1 band shifts to 573 cm Ϫ1 when the experiment is repeated with 12 CO. We assign the 573 cm Ϫ1 mode to the Fe-C-O bending mode ␦(Fe-C-O). The frequencies of the 517 and 573 cm Ϫ1 modes are similar to those that have been reported for the aa 3 -type oxidases from beef heart (16), R. sphaeroides (17), and aa 3 -600 from Bacillus subtilis (21).
The increased frequency of the Fe-His mode we observe in the aa 3 from P. denitrificans can be attributed to the strength of the H-bond of the proximal His 411 ligand to Gly 387 (41). From the model compound work, the complex with a weaker (or absent) hydrogen bond to the proximal His is expected to have the weaker Fe-His bond and the lower frequency vibration (36). The observation of the Fe-His at 220 cm Ϫ1 in conjunction with the high frequency data further supports our conclusion that heme a 3 2ϩ in the Y280H mutant is five-coordinated, highspin, and the proximal environment of heme a 3 2ϩ in the mutant behaves in a manner analogous to the wild type. Since The Fe-His 411 forms the only covalent linkage between the fivecoordinate heme a 3 and the protein, protein structural changes should be manifested through this bond. From the absence of such changes in the proximal environment of heme a 3 2ϩ , any protein structural change due to the mutation is not widespread but is more localized in the Cu B environment.
Structural information such as the geometry of the bound CO to the heme and its interactions with Cu B has been determined from the vibrational modes involving the CO. Although the (Fe-CO) and ␦(Fe-C-O) we have observed is similar to other aa 3 -type oxidases, the I ␦ /I ϳ 0.25 we observe is low compared with those of CcO and R. sphaeroides aa 3 -type oxidase (I ␦ /I ϭ 0.43) and significantly higher than that of cytochrome bo 3 (20) (I ␦ / I ϭ 0.1). It has been argued that a high I ␦ / I is an indication of a strong interaction between the CO and Cu B . Consequently, our data suggest that the Fe . . . Cu B distance is longer in our case. However, despite the increased Fe-Cu B distance we have not been able to detect a low-energy Fe-C bond (ϳ490 cm Ϫ1 ) corresponding to the high energy C-O bond (ϳ1950 cm Ϫ1 ) of the ␤-conformer. Recently, it was postulated that the two distinctly different Fe-CO modes observed in the RR of the Rb. sphaeroides spectra result from a change in the position of Cu B with respect to the CO due to the presence of one or more ionizable groups in the vicinity of the binuclear center (23). Additional experiments are in progress in our laboratory to address this possibility in the aa 3 oxidase from P. denitrificans. Fig. 4 shows the FTIR spectra of the CO-bound wild-type and Y280H mutant of aa 3 from P. denitrificans at room temperature. For comparison, the fully reduced aa 3 -CO complex of the mammalian enzyme is included (trace c). Trace a, shows that the C-O stretching modes from the P. denitrificans aa 3 enzyme split into three components just as was found in the lowtemperature experiments (27). The major component is centered at 1966 cm Ϫ1 (␣-form) and two minors are located at 1956 cm Ϫ1 (␤-form) and 1975 cm Ϫ1 . The frequency of the major CO stretching mode at 1966 cm Ϫ1 , which we detect in P. denitrificans is 3 cm Ϫ1 higher than the corresponding frequency of the CO-bound heme a 3 of CcO (trace c). Trace b shows that the Y280H mutant has two conformers that have (CO) at 1966 and 1975 cm Ϫ1 . Similar results have been obtained in the low-temperature experiments. 2 The structural basis for the splitting of the enzyme into the ␣and ␤-forms has not been determined and no information regarding the origin of the 1975 cm Ϫ1 has been reported. The FWHM is ϳ5 cm Ϫ1 in the C-O modes of the major conformer in both the mammalian and P. denitrificans enzymes indicating the absence of a wide distribution of allowed CO conformations. Thus, these data confirm the similarity in the properties of the active sites in these two terminal oxidases. The presence of the 1956 and 1975 cm Ϫ1 modes in the P. denitrificans do indicate, however, that the binuclear site of the bacterial enzyme, while similar, is not identical to its mammalian counterpart. Conversion between the ␣and ␤-forms is pH-dependent and has been attributed to changes in the iron-copper distance (23). It has also been demonstrated from the low-temperature FTIR data that the amplitudes of the bands attributed to the ␣and ␤-forms are temperature and pH-dependent (27). Thus the ␣-form represents a constricted pocket that will not allow CO to coordinate to heme iron without strong distal polar interactions between the CO and the copper atom, while in the ␤-form the Cu B atom is moved away from the bound CO. Based on the above interpretation regarding the origin of the ␣and ␤-forms we assign the 1975 cm Ϫ1 mode to the ␥-conformation in which Cu B is moved closer to the CO-bound heme a 3 , thereby the Fe-C-O moiety is further distorted from its preferred symmetry in the ␣-form. Unlike the wild-type enzyme which has a prominent ␣-form, the mutant enzyme exhibits the 1975 cm Ϫ1 (␥form)/1966 cm Ϫ1 (␣-form) in a 1.8 ratio indicating that the ␥-confomer is the major confomer in the mutant. Consequently, in the absence of the cross-link Tyr-His, the Cu B atom has moved further closer to the CO-bound heme a 3 . This is further supported by the absence of the ␤-form (1956 cm Ϫ1 ) in the mutant enzyme in which the CO is bound without anomalous polar or steric interactions. Both C-O stretches for the Y280H mutant are broad indicating a wide distribution of allowed CO conformations. In the absence of Tyr 280 , the heme a 3 -Cu B distance has changed and Cu B is not fixed in a certain position resulting in different Cu B conformations. The different conformations of Cu B in the mutant reflect significant differences in the heme environment, thereby alter the properties of the CO modes observed in the FTIR spectra. In the CO derivatives, the different conformations of Cu B could easily cause the change in the C-O frequency and bandwidth since it is well established that ligand frequencies and bandwidths in heme proteins are modulated by the properties of the distal environment.
It has been established by Rousseau et al. (12, 16 -18, 20) that the (Fe-CO) and (CO) frequencies of heme proteins and the correlation between them reflect: (a) the identity and properties of the proximal ligand because the bound CO competes with the proximal ligand for the same iron d z 2 orbital, and (b) indicate the polarity of the distal heme pocket. A highly polar environment favors -back donation, resulting in an increased (Fe-CO) and reduced (C-O) due to the increased density in the CO antibonding orbitals. Additional information concerning the properties of the proximal environment in heme-copper oxidases and how it influences the ligand properties on the distal site has been deduced from the unique inverse linear correlation that exists between the frequencies of Fe-His and the Fe-CO stretching modes. Recently, we have shown that in heme-Cu B oxidases the strength of the proximal histidine Hbonding interaction affects the strength of both the Fe-C and C-O bonds which are further influenced by the Cu B distal environment (21). We consider both proximal and distal effects on the origin of the (Fe-CO) and (CO) frequencies we have observed.
The ϳ4 cm Ϫ1 downshift in the (Fe-CO) of P. denitrificans, when compared with that found in CcO, is brought about by a stronger hydrogen bonding interaction of the proximal histidine, and by distal effects on the heme a 3 -bound CO exerted by Cu B . This argument is supported by the observed high frequency of the Fe-His stretching mode at 220 cm Ϫ1 , as compared with aa 3 -and bo 3 -type oxidases. A similar conclusion concerning the effect of the proximal ligand to the properties of the distal CO was reached recently by Wang and co-workers (45) for prostaglandin H synthase. On the other hand, Das et al. (23) has argued recently that the pH dependence of the Fe-C-O modes they observed in Rb. sphaeroides cannot be attributed to 2 P. Hellwig, submitted for publication. proximal effects due to the absence of any significant pH-dependent change of the proximal Fe-His stretching frequency. The proximal protein pocket in heme a 3 appears to be hydrophobic and inaccessible to solvent as indicated by Raman studies in D 2 O (41). Consequently, the invariance of the frequency of the Fe-His does not necessarily rule out a hydrogen bond interaction to the proximal histidine of heme a 3 3ϩ . This is further supported from the crystal structure of the P. denitrificans enzyme where it is shown that the proximal His to heme a 3 is indeed hydrogen-bonded to Gly 387 . The frequencies of the Fe-His and Fe-CO of the enzyme place it on the correlation curve shown in Fig. 5A, and posit that it has the same structure as that of the major ␣-form of the mammalian and Rb. sphaeroides enzymes. In the absence of a proximal effect as indicated by the strength of the Fe 2ϩ -His located at 220 cm Ϫ1 in the mutant, the anomalously high C-O stretching mode we observed in the CO-bound Y280H mutant is attributed solely to distal effects. Proximal effects should not be taken under consideration since the Fe-His stretching frequency remains unaffected by the mutation, evidence that a global conformational change is unlikely to have taken place.
The properties of the heme a 3 -Cu B binuclear center have been determined from the correlation of (Fe-CO) and (C-O) frequencies. The major component (␣-form) of the wild-type aa 3 , shown in Fig. 5B, deviates from the inverse linear correlation curve that exists between the frequencies of (Fe-CO) and (C-O) between histidine coordinates proteins but fits to the curve of the rest of the terminal oxidases. Although we were unable to detect the (Fe-CO) in the ␤-form, the frequency of the (C-O) in the ␤-form is identical to that reported for the ␤-form of aa 3 from Rb. sphaeroides suggesting that in P. denitrificans the frequencies of the ␤-form are placed on the curve of the histidine-coordinated proteins in which CO can bind to the iron without anomalous polar and steric interactions.
The overall similarity between the frequencies and relative enhancements for the vibrational resonances of oxidized heme a 3 in the wild-type and Y280H mutant indicates that the protein milieu surrounding the heme a 3 is the same in these two forms. A comparison of the high-frequency resonance Raman spectra of the mutant enzyme with that of the wild type, upon reduction, shows no significant differences. However, a conformational change is likely to occur in the binuclear center as indicated by the differences in the intensity and bandwidth of the formyl line of a 3 in both oxidation states of the heme a 3 in the Y280H mutant compared with the wild-type enzyme. This, nevertheless, cannot be attributed to a different oxidation, spin, or ligation change since we have established that no such changes occur as a result of the mutation. Therefore, other possibilities should be considered.
In the absence of a conformational change in heme a 3 the only possibility to account for the reduced intensity of the heme a 3 formyl group is the histidine ligands coordinated to Cu B . His 276 is the histidine that forms the cross-link with Tyr 280 in the wild-type and is not close to the formyl group, thus it is unlikely that it can interact with the formyl group. His 325 and His 326 are on the same helix and are close to the formyl group. The closest residue to the formyl oxygen atom of heme a 3 is His 325 (His 290 in bovine). It is noteworthy that the absence of Tyr 280 in the heme pocket results in the loss of the hydrogenbonding interaction between the hydroxy group on the farnesyl chain of heme a 3 and the hydroxy group on the tyrosine. If the observed changes were due to a change in the hydrogen bonding state of the formyl group then, a frequency shift would have been expected (46). We postulate that changes in the Fe-Cu B distance could modulate the position of His 325 with respect to the formyl group. Since no frequency shift is observed, we postulate that the intensity difference is due to a change in the geometry of the formyl group that does not allow the electronic coupling between the formyl group and the porphyrin core to be as effective as in heme a 3 of the wild type enzyme. The complete absence of the CϭO stretching vibration of the formyl group (ϪCHO), reported in the reduced form of the Y280F mutant from Rb. sphaeroides, was attributed to an altered heme a 3 conformation. The question arising is whether the change in the orientation of the formyl, which is more evident in the Y280F mutant from Rb. sphaeroides than the Y280H aa 3 from P. denitrificans, could have an effect in the catalytic function of the enzyme. Recently, Das et al. (46) reported a redox-link deprotonation event at the binuclear site of the quinol oxidase from Acidianus ambivalens, in which the changes in heme a 3 formyl C-O stretching mode upon heme reduction were attributed to a change in H-bonding to the formyl group. As a possible candidate for the pH-dependent changes they observed suggested His 290 (bovine sequence numbering) and discussed the implications for proton translocation. However, since this is a single observation and no similar observation has been made for any other aa 3 CcO we cannot consider that the small conformational change we detect in the formyl group of heme a 3 in the Y280H mutant will influence the functional properties of the binuclear center.
Oxygenated intermediate(s) that occur after the decay of the oxyintermediate (I m ) in the MV/O 2 reaction may be generated by direct mixing of oxygen with the CO-bound MV oxidase. These species have been observed by others to have relatively long lifetimes, and thus, they can be studied in a conventional absorption spectrometer. The Soret region optical absorption spectra of oxygenated species formed this way are shown in Fig. 6. The difference spectrum (MV-CO minus oxidized) of the wild-type enzyme (panel A, inset), with maxima and minima at 432 and 411 nm, agrees with that of the CO-bound MV Y280H enzyme (panel B, inset). In both the wild-type and Y280H MV enzymes, mixing with O 2 shifts the Soret maximum to 438 nm (1-18 min). This indicates that in both the wild-type and mutant enzymes O 2 spontaneously replaces CO. The progression of changes in the Soret indicates that the decay of the oxygenated 438-nm species to the pulsed and subsequently to the resting form occurs on a time scale of tens of seconds. These results suggest that Tyr 280 is not involved in the formation and decay of the 438-nm species because similar observations were obtained in the wild-type and mutant enzymes. Similar observations in the decay of II m to the resting form have been reported in the bovine MV CcO/O 2 reaction by Rousseau and co-workers (47). It was also demonstrated by the same authors that intermediates formed by mixing O 2 with CO-bound MV CcO and allowing O 2 to spontaneously replace CO are the same as those of II m formed in the flow-flash-probe experiments (47).
The data reported here have shown that the cross-link His-Tyr generates a unique environment around Cu B and holds it in a certain distance and position from heme a 3 . In the absence of the cross-linking, Cu B moves further toward the heme a 3 in the CO bound form of the enzyme, affecting the properties of the bound ligands to heme a 3 . Thus, in the Y280H mutant, heme a 3 retains its proximal His ligand and can be oxidized by oxygen. The oxidation of heme a 3 in the Y280H mutant and in the wild-type proceeds through oxygenated species with Soret maxima at 438 nm. The nature of the bound oxygen intermediates following the oxy species remains to be determined. The characterization of the functional/structural implications to the heme a 3 -Cu B center by the Y280H mutation reported here, and the determination of the initial electron transfer steps in the Y280H/O 2 reaction will lead to a better understanding of the oxidative phase of cytochrome c oxidase. These experiments are in progress in our laboratory.