Role of Tyr-288 at the Dioxygen Reduction Site of Cytochrome bo Studied by Stable Isotope Labeling and Resonance Raman Spectroscopy*

To explore the role of a cross-link between side chains of Tyr-288 and His-284 at the heme-copper binuclear center, we prepared cytochrome bo where d4-Tyr, 1-[13C]Tyr, or 4-[13C]Tyr has been biosynthetically incorporated. Unexpectedly, the d4-Tyr-labeled enzyme showed a large decrease in the ubiquinol-1 oxidase and CO binding activities. Optical absorption and resonance Raman spectra identified the defect in the distal side of the heme-copper binuclear center. In the CO-bound d4-Tyr-labeled enzyme, a large fraction of the ν(Fe-C) mode was shifted from the normal 520-cm-1 band to a broad band centered around 491 cm-1, as found for the Y288F mutant. Our results suggested that the substitution of ring hydrogens of Tyr-288 with deuteriums slows down the formation of the His-Tyr cross-link essential for dioxygen reduction at the binuclear center.

reduction mechanism catalyzed by these oxidases have attracted considerable interests. Time-resolved spectroscopic studies have identified a number of intermediates in the reaction of the O 2 reduction by bovine CcO (8 -10). The reaction begins with binding of O 2 to the reduced heme a 3 . The resulting "oxy" (O) intermediate is rapidly transformed to the so-called "peroxy" (P) species, which has an oxidation state of the hemecopper moiety higher than the Fe 3ϩ /Cu 2ϩ state by two oxidative equivalent. Although the peroxy intermediate had been expected to have an intact O-O bond (11,12), our previous studies have clearly shown the definite FeϭO stretching Raman band at ϳ800 cm Ϫ1 for the P intermediate, suggesting that the O-O bond is already cleaved at this step (13)(14)(15)(16). Other chemical experiments also support this assignment (17,18). The subsequent one-electron reduction converts the P species to the next intermediate, "ferryl" (F) species with (FeϭO) at ϳ785 cm Ϫ1 (15,19,20). In cytochrome bo, two types of the F species were recently identified by time-resolved freeze-quench electron paramagnetic resonance spectroscopy (21).
The O-O bond scission in the O-to-P transition raised the question about the source of an additional oxidative equivalent. Cleavage of the iron-bound O-O bond requires four electrons. Three of them are provided by the binuclear center [Fe II /Cu I 3 Fe IV /Cu II ], but the source of the fourth electron remains to be solved. Because of the formation of a porphyrin -cation radical, oxidation of the heme macrocycle can provide one electron, as in horseradish peroxidase compound I. However, resonance Raman and optical absorption spectra eliminate this possibility (14,22,23). Another possibility is oxidation of an amino acid in the vicinity of the active center to form a so-called compound ES observed in cytochrome c peroxidase and prostaglandin H synthase (24 -26). Tyr-288 (if not stated otherwise, we adopt here the residue numbering of cytochrome bo) is highly conserved in the SoxM-type oxidases and is a potential candidate for such protein-based electron donors at the distal side of heme a 3 (or o). Tyr-288 is also involved in the uptake of protons through the K channel during the initial reduction of the binuclear center.
Crystallographic (27,28) and protein sequencing (29) studies on CcOs revealed the presence of a peculiar C ⑀ -N ⑀ covalent bond between Tyr-288 and His-284, one of three histidine ligands of Cu B (Fig. 1). The OH group of Tyr-288 could form a hydrogen bond with a diatomic ligand like O 2 (28), favoring the cross-linking to His-284 to participate in O 2 reduction (30 -35). Studies with model compounds showed that the His-Tyr linkage lowers the pK a of the phenol moiety by 1.1 to 1.8 (36 -38).
Consequently, Tyr-288 could provide hydrogen to the ironbound O-O, which leads to rapid cleavage of the bond with the concomitant formation of a tyrosine neutral radical. Electron paramagnetic resonance studies on the reaction of H 2 O 2 with the oxidized enzyme revealed an electron paramagnetic resonance signal with partially resolved hyperfine structure ascribable to a tyrosine radical (17,39,40). Using radioactive iodide labeling followed by peptide mapping, Proshlyakov et al. (34) demonstrated the presence of the tyrosine neutral radical at the cross-linked His-284-Tyr-288 during P formation. In model studies, a C-O stretching vibration of the tyrosine radical was assigned to 1516 (38) or 1530 (37) cm Ϫ1 , which is higher than the 1489 cm Ϫ1 found in the reaction of H 2 O 2 with the oxidized cytochrome bo (41). Recent attenuated total reflection-Fourier transform infrared study on bovine CcO revealed strong negative bands at 1313 and 1547 cm Ϫ1 and a positive band at ϳ1519 cm Ϫ1 in the P M intermediate (42). These results indicate that the cross-linked Tyr facilitates the cleavage of the O-O bond during P formation.
To explore the role of the Tyr-His covalent bond, we prepared three stable isotope-labeled enzymes. We constructed a Tyr auxotrophic strain harboring an overexpression vector for cytochrome bo and grew it in a synthetic medium supplemented with L-d 4 -Tyr, L-1-[ 13 C]Tyr, or L-4-[ 13 C]Tyr as a sole L-Tyr source. Unexpectedly, the d 4 -Tyr-labeled enzyme showed a reduction in the ubiquinol-1 oxidase and CO binding activities. Optical absorption and resonance Raman (RR) spectra identified the defect in the distal side of the heme-copper binuclear center. Our results suggest that the substitution of ring hydrogens of Tyr-288 with deuteriums slows down the formation of the His-Tyr cross-link essential for dioxygen reduction at the binuclear center. A possible self-catalyzed mechanism for the formation of the His-Tyr bond is discussed. Bacterial Strains-E. coli strains GO103/pHN3795-1 (cyo ϩ ⌬cyd/ cyo ϩ ), ST4533/pHN3795-H333A (⌬cyo cyd ϩ /cyo Ϫ ) (43), and ST4676/ pMFO9-Y288F (⌬cyo cyd ϩ /cyo Ϫ ) (44) were used for isolation of the wild type cytochrome bo and H333A and Y288F mutant oxidases, respectively. The tyrosine auxotroph, GO103Y, was constructed as follows. The ⌬tyrA16::Tn10 locus was transduced into GO103 by P1 phage grown on strain N3087 (CGSC 6662). A transductant was verified by Tet r and Tyr Ϫ phenotypes and transformed with pHN3795-1 to give GO103Y/pHN3795-1.
Purified enzymes were stored in 50 mM Tris-HCl (pH 7.4) containing 0.1% sucrose monolaurate at Ϫ80°C until use. The concentration of the enzymes was determined by the pyridine ferrohemochromogen method using a millimolar extinction coefficient of 20.7 for heme B, assuming that the enzyme has two hemes. Reduced forms were prepared by adding a slight excess of freshly prepared anaerobic dithionite solution under argon atmosphere into the degassed enzyme solution. CO-bound forms were prepared by introducing gaseous CO into the reduced samples.
Resonance Raman Measurements-RR spectra were obtained with a single polychromator (DG-1000; Ritsu Oyo Kogaku) equipped with a liquid N 2 -cooled CCD detector (CCD-1340/400-EB; Princeton Instruments). The excitation wavelength used was 413.1 nm from a Kr ion laser (BeamLok 2060; Spectra Physics) for the air-oxidized, reduced, and CO-bound forms. The laser power at the sample point was adjusted to 1.2 mW for the reduced form and to 0.05-0.1 mW for the air-oxidized and CO-bound forms to prevent photoreduction and photodissociation, respectively. Raman shifts were calibrated with indene, CCl 4 , and aqueous solution of ferrocyanide, and the accuracy of the peak positions of the well defined Raman bands was Ϯ 1 cm Ϫ1 . All measurements were performed at room temperature with a spinning cell. The enzyme concentration for RR experiments was 50 M heme in the 50 mM Tris-HCl (pH 7.4) with 0.1% sucrose monolaurate. To measure absorption and RR spectra of the fully oxidized Y288F, the air-oxidized form, where hemes are partly reduced (49,50), was treated with ferricyanide and passed through a Sephadex G-200 column to remove the excess ferricyanide.
Metal Content Analysis-The purified oxidase solutions (0.2-1 mM) were diluted to 0.5-1 M by Nanopure water (18.3 M⍀ purity; Barnstead), and copper and iron contents were determined by inductively coupled plasma atomic emission spectroscopy at Shimadzu Analysis Center. Each analysis was complemented by appropriate control experiments.
Miscellaneous-Optical absorption spectra of purified oxidases (ϳ5 M heme) were recorded on a Hitachi UV-3200 UV/Vis spectrophotometer at room temperature. The samples were sealed inside a 1-cm path length quartz cuvette. Heme contents of the oxidases were analyzed by reverse-phase HPLC after extraction with acid acetone as described previously (51). Ubiquinol-1 oxidase activity was measured spectrophotometrically with a millimolar extinction coefficient of 13.25 at 275 nm at 0.5 mM ubiquinol-1 (3).

Effects of Isotope Labeling of Tyrosines on Quinol Oxidase Activity and Metal
Contents-Three different isotope-labeled tyrosines were biosynthetically incorporated into the wild type cytochrome bo in the Tyr auxotroph: 1) L-1-[ 13 C]Tyr with a 13 C atom at the main chain carbonyl, 2) L-4-[ 13 C]Tyr with a 13 C atom bound to the phenolic oxygen, and 3) L-d 4 -Tyr where all ring hydrogens are deuterated ( Fig. 1). In the enzymes, all tyrosines are isotopically labeled.
Ubiquinol-1 oxidase activity and metal contents are summarized in Table I (Table I) as reported previously (43,44,(52)(53)(54). In Y288F, the loss of ubiquinol oxidase activity was accompanied by the replacement of heme o with heme b at the binuclear center and a partial loss of the CO binding activity (44) as reported for Y288L (54) and heme bb-type wild type oxidase, which has been isolated from the heme O synthase mutant (55). The CO binding activity and the copper content of H333A were higher than those reported previously (52), probably because of a difference in expression vectors (i.e. a multicopy vector pHN3795-H333A versus a single copy vector pMFO1-H333A) (43).
Unexpectedly, the quinol oxidase and CO binding activities of d 4 -Tyr-WT were reduced to 19 and 63%, respectively, of those of 12 (Table I), effects of the d 4 -Tyr labeling on the oxidase activity must be different from those of the Y288F and H333A mutations. Accordingly, effects of the d 4 -Tyr labeling on the binuclear center were examined by optical and resonance Raman spectroscopy.
Effects of Isotope Labeling of Tyrosines on Optical Absorption Spectra of Cytochrome bo-Optical absorption maxima of 1-[ 13 C] Tyr-WT and 4-[ 13 C]Tyr-WT in the air-oxidized, dithionite-reduced, and reduced CO-bound states are summarized in Table II. They are similar to those of 12 C WT with the Soret peak at 409, 428, and 418 nm, respectively. The Soret maxima of the CO-bound form of d 4 -Tyr-WT was shifted to 426 nm from 418 nm of 12 C WT (Table II) because of significant contribution from the unbound form (428 nm), which results from the low CO binding activity of heme o (Table I)  Effects of Isotope Labeling of Tyrosines on Resonance Raman Spectra in the High Frequency Region-To gain further insight into the molecular structure of the heme-copper binuclear center, we have employed RR spectroscopy. The 413.1-nm excited RR spectra of the air-oxidized, reduced, and CO-bound forms are shown in Fig. 3, A, B, and C, respectively. It is well known that bands in the high frequency region can be used as sensitive markers of the oxidation state ( 4 ) and spin and coordination states ( 2 and 3 ) of the heme iron (56). In the air-oxidized states, all the Tyr-labeled WT and H333A mutant enzymes showed almost similar RR spectra to that of the unlabeled 12 C WT enzyme (Fig. 3A). The 4 band of all the preparations was observed at 1371 cm Ϫ1 with a shoulder at 1361 cm Ϫ1 . The shoulder intensity becomes relatively stronger upon increase of laser power and, therefore, is attributed to the photoreduced species (data not shown). Typically, 5-coordinate high spin, 6-coordinate high spin, and 6-coordinate low spin hemes give the 3 band at 1490 -1500, 1475-1485, and 1500 -1510 cm Ϫ1 , respectively (57). Therefore, the 3 bands observed at 1477 and 1505 cm Ϫ1 for 12 C WT, Tyr-labeled WT, and H333A mutant are assignable to high spin heme o and low spin heme b, respectively. d 4 -Tyr-WT exhibited no specific features compared with 12 C WT. In the heme bb-type Y288F mutant, a high spin heme b at the binuclear center was partly photoreduced and gave the 4 and 3 bands at 1361 and 1492 cm Ϫ1 at the expense of the 1371-and 1477-cm Ϫ1 bands. For the dithionite-reduced 12 C WT, Tyr-labeled WT, and H333A mutant, the 3 bands observed at 1472 and 1492 cm Ϫ1 were assigned to 5-coordinate high spin heme o and 6-coordinate low spin heme b, respectively (Fig. 3B). Again, there are no specific features in d 4 -Tyr-WT. In the reduced Y288F mutant, the 1472-cm Ϫ1 band disappeared and only the 1492-cm Ϫ1 band was observed, which coincides with the conversion of heme bo-type to heme bb-type oxidase (Table I). When CO was added to the dithionite-reduced state, a new 4 band appeared at 1371 cm Ϫ1 for [ 12 C]Tyr-WT, 1-[ 13 C]Tyr-WT, and 4-[ 13 C]Tyr-WT enzymes, suggesting that CO binds to high spin heme o (Fig. 3C). The presence of the 3 band at 1504 cm Ϫ1 supports this assignment. For d 4 -Tyr-WT, on the other hand, 4 and 3 appeared at 1361 and 1493 cm Ϫ1 , respectively, which were distinct from 12 C WT but close to the H333A mutant. This would suggest the formation of an amino acid-bound 6-coordinate low spin state in heme o similar to the heme b moiety. In the case of the Y288F mutant, the whole spectrum was little affected by addition of CO, presumably because of photodissociation of the iron-bound CO even under the quite low laser power condition (ϳ0.05 mW). Fig. 4 shows the dithionite-reduced minus CO-bound RR difference spectra. The spectra for [ 12 C]Tyr-WT, 1-[ 13 C]Tyr-WT, and 4-[ 13 C]Tyr-WT were alike, yielding four negative bands at 1408, 1503, 1548, and 1579 cm Ϫ1 . In contrast, the four peaks were indistinct for d 4 -Tyr-WT, H333A, and Y288F. The 1579-cm Ϫ1 band seemed to be upshifted to 1588 cm Ϫ1 for the H333A mutant. The Fourier transform infrared study by Tomson et al. (58) on cytochrome bo revealed that four positive peaks at 1412, 1483, 1549, and 1579 cm Ϫ1 in the CO-photodissociated minus CO-bound difference spectra at 80 K shifted upon labeling with L-15 N ␦,⑀ -His or L-d 4 -Tyr. The 1549-and 1579-cm Ϫ1 bands were assigned to the coupled His-Tyr ring modes on the basis of isotope-labeling data and normal mode calculations. Accordingly, the unexpected behavior of the 1548and 1579-cm Ϫ1 bands of d 4 -Tyr-WT might be because of mode changes caused by ring deuterization of Tyr. Fig. 5 shows the RR spectra of all the dithionite-reduced forms of Tyr-labeled WT and mutants. They were similar to those of unlabeled 12 C WT. The Fe-His stretching mode, (Fe-His) , of cytochrome bo has been assigned to the 208-cm Ϫ1 band by isotopic substitution of the heme iron (52). All the preparations here provided the intense (Fe-His) mode at 208 -209 cm Ϫ1 , which disappeared upon addition of CO (data not shown). Thus, neither substitutions of Tyr-288 and His-333 nor the isotope labeling of Tyr-288 affect the (Fe-His) mode of heme o. It is noted that d 4 -Tyr-WT does not show any peculiar property. Fig. 6A shows low frequency RR spectra of the CO-bound enzymes. The Fe-CO stretching mode, (Fe-C) , appeared at 520 cm Ϫ1 for [ 12  zymes and exhibited low frequency shift by ϳ10 cm Ϫ1 upon C 18 O substitution, which is comparable with a theoretical value (12 cm Ϫ1 ) expected for an isolated diatomic oscillator. The (Fe-C) mode was shifted from 520 to 497 cm Ϫ1 in the Cu B mutant H333A (59) and to 491 cm Ϫ1 in the Y288F mutant. It is noted that the intensity of the 520-cm Ϫ1 band of d 4 -Tyr-WT is about 40% of the 12 C WT when the spectra are normalized with the 675-cm Ϫ1 band. Probably, the (Fe-C) band was substantially transformed to the broad 491-cm Ϫ1 band as found in Y288F. The presence of the 491-cm Ϫ1 band was confirmed by the difference spectrum (Fig. 6B) and deconvolution of the band (data not shown). The broadness of the Fe-C stretching band suggests a loss of steric constraints for the CO binding to heme o and thus the serious perturbation of the Fe o ⅐⅐⅐Cu B side of the binuclear center.

Effects of Isotope Labeling of Tyrosines on Cytochrome bo-
Isotope labeling of amino acid residues is a powerful tool for signal assignments for NMR, IR, and Raman spectroscopies and generally applied on the premise that it has little effect on the function of proteins. In contrast to L-1-[ 13 C]Tyr and L-4-[ 13 C]Tyr, specific stable isotope labeling of tyrosines in cytochrome bo with L-d 4 -Tyr resulted in a large decrease in ubiquinol-1 oxidase activity and CO binding activity, although it did not affect the contents of three redox metal centers, heme b, heme o, and Cu B (Table I). Optical absorption and RR spectra of d 4 -Tyr-WT showed the presence of a primary defect at the distal side of the heme o-Cu B binuclear center (Table II, Figs. 2-6). Among 20 tyrosines present in subunit I, only Tyr-288 is invariant and located in the proton channel. Two other relatively conserved tyrosines, Tyr-61 and Tyr-173, are not components of proton channels (7), and their mutations did not affect the catalytic activity (44,54). In addition, tyrosines are not expected to be in the high affinity quinone-binding site for the oxidation of quinols (7). The amount of the normal and active enzyme present in the d 4 -Tyr-labeled cells was only 20 -40% of the control cells (Table I, Fig. 6). If the d 4 -Tyr labeling perturbed tyrosines other than Tyr-288, the aerobic growth of the E. coli cells would be severely reduced under our growth conditions where cytochrome bo serves as a sole terminal oxidase in the cytochrome bd-deficient strain. It is unlikely, because a final yield of d 4 -Tyr-WT was ϳ50% of 1-[ 13 C]Tyr-WT. Thus, labeling of other tyrosines with L-d 4 -Tyr would not affect catalytic activity and proton pumping.
Our data indicate that d 4 -Tyr-WT consists of two populations; the major component gives the broad Fe-CO stretching band at 491 cm Ϫ1 similar to that of the Y288F mutant, and this is enzymatically inactive. The rest of the molecules are active and show the Fe-CO stretching band at 520 cm Ϫ1 (Fig.  6). Thus, it became evident that "hydrogens" bound to the phenolic ring of Tyr-288 in addition to its OH group are essential for the functional assembly of the heme o-Cu B binuclear center (44,54).
The Fe-CO stretching modes of the heme-copper terminal oxidases, which appear around 520 cm Ϫ1 for the ␣-conformer, are higher than those of other hemoproteins because of the presence of Cu B near the iron-bound CO. Except for the Cu B ligand mutants like H333A, the frequencies of (Fe-C) and (C-O) of the ␣-conformer do not fall on the well recognized inverse correlation plot for the proteins with histidine as a proximal ligand (30, 59 -61). Recently, RR studies on aa 3 -type CcO from  Rhodobacter sphaeroides revealed the pH-dependent interconversion between the ␣and ␤-conformers with the Fe-CO stretching modes at 517 and 494 cm Ϫ1 , respectively (49). It was postulated for the ␤-conformer that some different structures result from a change in the position of the Cu B atom with respect to the iron-bound CO because of the presence of the cross-linked Tyr-288 and/or one of the histidines that coordinate Cu B (49). Judging from the (Fe-C) frequency, it is interesting that d 4 -Tyr yields the binuclear center structure similar to that of the ␤-conformer. This frequency is lower than the 497 cm Ϫ1 of the H333A mutant. Therefore, we are tempted to infer that the formation of Tyr-288-His-284 covalent linkage determines the position of Cu B and also the Fe-C stretching frequency and that the failure in the formation of the covalent linkage due to deuteration of Tyr-288 changes the coordination environment of CO, yielding the broad (Fe-C) band around 491 cm Ϫ1 . In contrast, 4-[ 13 C]Tyr-WT and 12 C WT have similar properties to each other. A slightly reduced quinol oxidase activity may be ascribed to a mass effect of carbon at the C4 atom of the tyrosine ring, which would affect the role of a general acid-base catalyst of Tyr-288. However, additional studies are needed to make a definitive determination.
Role of Tyr-288 in Heme-Copper Terminal Oxidases-Previously, we proposed a new role of Tyr-288 as a general acid-base catalyst in the reaction of the oxidized cytochrome bo with hydrogen peroxide (41). Calculations of electrostatic energy on the Paracoccus denitrificans CcO indicate that Tyr-288 is expected to be protonated in the oxidized state when a negative ligand such as OH Ϫ is bound in the binuclear center (62). In our model, Tyr-288 in a tyrosinate form abstracts a proton from hydrogen peroxide and, in turn, provides the proton to cleave the O-O bond and produces a water molecule and a tyrosine neutral radial. For Tyr-288 to function as a general acid-base catalyst the pK a of its OH group must be close to neutral pH.
The C ⑀ -N ⑀ covalent linkage between Tyr and His could modulate the pK a value of Tyr-288 and the redox potential (37)(38)(39) and might optimize the geometry of Cu B relative to heme o, facilitating the O-O bond cleavage by electron-coupled hydrogen transfer from the cross-linked tyrosine (30 -35). Therefore, if the yield of the post-translational modification decreases, the activities of the heme-copper terminal oxidases would decrease.
Substitutions of Tyr-288 with Leu or Phe completely eliminated quinol oxidase activity because of the replacement of heme o with heme b at the binuclear center (44,54), as found for heme bb-type wild type enzyme isolated from heme O synthase mutants (51,55). These mutations also affect the Cu B content and the CO binding activities. Redox-induced protein structural changes were perturbed in the Y288F and H333A mutants (63). The misincorporation of heme B into the binuclear center, which would be caused by disruption of a hydrogen bond between the OH group of Tyr-288 and the OH group of the hydroxyethyl farnesyl chain of heme o (28) and the lack of the phenolic OH group at Tyr-288 as the acid-base catalyst (41) would result in a nonfunctional enzyme. Such changes could affect the configuration and the distance of the Cu B atom from heme o and thus initiate the dioxygen reduction ability.
Das et al. (64) studied the Y288F mutant of CcO from R. sphaeroides with RR spectroscopy. They found that Y288F retained Cu B (ϳ70%) and that heme a 3 is of 6-coordinate configuration in both oxidized and reduced states, probably because of coordination of His-284 to heme a 3 . In contrast, the Y288H mutant of CcO from P. denitrificans retained 5-coordinate heme a 3 (50). The (Fe-C) mode of the Y288 mutant from E. coli was observed at 491 cm Ϫ1 , whereas the Y288 mutant of CcOs from R. sphaeroides (64) and P. denitrificans (50) gave an intense (Fe-C) band at 493 (Y288F) and 517 cm Ϫ1 (Y288H), respectively. Structural changes at the binuclear center caused by the Tyr-288 mutations may be different in CcO and quinol oxidase. Although there are some phenotypic differences, the Tyr-His cross-link is also essential in CcO for the proper configuration of the heme-Cu B binuclear center.
Unexpected phenotypes of d 4 -Tyr-labeled cytochrome bo could be explained by a loss of the C ⑀ -N ⑀ covalent linkage between Tyr-288 and His-284 as illustrated in Fig. 7. Because of the high pK a of tyrosine (10.1), the OH group of the conserved tyrosine (Tyr-288) in the vicinity of the heme-copper binuclear center is expected to be protonated. Electrochemical calculations on the ferric CcO suggest the presence of a hydroxide ion coordinating to Cu B (62). During the assembly and maturation of the Cu B center in the wild type, the conserved tyrosine (Tyr-288) in the vicinity of the heme-copper binuclear center could undergo oxidation by Cu B 2ϩ , and a resultant tyrosine neutral radical would attack the N ⑀ of the nearby Cu B ligand histidine (His-284) (Fig. 7B). Release of a proton from the C ⑀ of Tyr-288 by the N ⑀ of His-284 leads to the formation of a covalent bond between the C ⑀ of Tyr-288 and the N ⑀ of His-284 and the Tyr radical to a tyrosinate (Fig. 7C). In d 4 -Tyr-WT, substitution of protons on the phenol ring by deuteriums would slow down the release of the proton from Tyr-288, thereby generating some unknown products and, as a result, reducing the yield for the formation of the Tyr-His cross-link essential for the cleavage of the O-O bond at the O-to-P transition of the dioxygen reduction cycle. Accordingly, only a part of d 4 -Tyr-WT could exhibit the oxidase activity.
In bacterial and mammalian CcO, a limited acid hydrolysis at the conserved Asp-271-Pro-272 bond allowed the isolation and N-terminal sequence analysis of the C-terminal half of subunit I (29). Further, the Asp-313-Pro-314 bond in the Thermus thermophilus enzyme facilitated the isolation and mass spectroscopic analysis of a 43-residue fragment containing the His-284-Tyr-288 cross-link (29). We tried to demonstrate a lack of the His-284-Tyr-288 cross-link in d 4 -Tyr-WT and identified ϳ40% of subunit I fragments by proteolytic or chemical cleavage followed by mass spectroscopic analysis. 2 However, because of the hydrophobic nature of transmembrane helices and the absence of convenient cleavage sites around the active site, we were unable to confirm the absence of such a covalent bond in the d 4 -Tyr-labeled cytochrome bo.
Conclusion-The His-Tyr cross-link in the heme-Cu B binuclear center of terminal oxidases is essential for the proper geometry of the Cu B atom relative to heme o. Characterization of cytochrome bo that has been biosynthetically labeled with L-d 4 -Tyr, L-1-[ 13 C]Tyr, or L-4-[ 13 C]Tyr revealed that the d 4 -Tyrlabeled enzyme was unexpectedly defective in the ubiquinol-1 oxidation. Optical absorption and RR spectra identified the defect in the distal side of the heme-Cu B binuclear center but not in the proximal side and heme o itself. Our results suggest that the substitution of ring hydrogens of Tyr-288 with deuteriums practically decreases in the formation of the His-Tyr cross-link, thereby decreasing the amount of functional enzyme capable of the O-O bond scission.