Confirmation of the involvement of protein domain movement during the catalytic cycle of the cytochrome bc1 complex by the formation of an intersubunit disulfide bond between cytochrome b and the iron-sulfur protein.

To study the essentiality of head domain movement of the Rieske iron-sulfur protein (ISP) during bc(1) catalysis, Rhodobacter sphaeroides mutants expressing His-tagged cytochrome bc(1) complexes with three pairs of cysteines engineered (one cysteine each) on the interface between cytochrome b and ISP, A185C(cytb)/K70C(ISP), I326C(cytb)/G165C(ISP), and T386C(cytb)/K164C(ISP), were generated and characterized. Formation of an intersubunit disulfide bond between cytochrome b and ISP is detected in membrane (intracytoplasmic membrane and air-aged chromatophore), and purified bc(1) complex was prepared from the A185C(cytb)/K70C(ISP) mutant cells. Formation of the intersubunit disulfide bond in this cysteine pair mutant complex is concurrent with the loss of its bc(1) activity. Reduction of this disulfide bond by beta-mercaptoethanol restores activity, indicating that mobility of the head domain of ISP is functionally important in the cytochrome bc(1) complex. The rate of intramolecular electron transfer, between 2Fe2S and heme c(1), in the A185C(cytb)/K70C(ISP) mutant complex is much lower than that in the wild type or in their respective single cysteine mutant complexes, indicating that formation of an intersubunit disulfide bond between cytochrome b and ISP arrests the head domain of ISP in the "fixed state" position, which is too far for electron transfer to heme c(1).

movement of the head domain of the iron-sulfur protein (ISP) 1 during bc 1 catalysis. This suggestion arose from observation of a particularly low electron density area in the intermembrane space portion of the complex, where the extramembrane domains of ISP and cytochrome c 1 reside (2). This movement hypothesis was further supported by the observation of various positions for 2Fe2S in the different crystal forms (3,4) and in complexes loaded with different inhibitors (4,5).
In tetragonal I4 1 22 crystals of native oxidized bovine cytochrome bc 1 complex, the position of the 2Fe2S cluster is 27 Å from heme b L and 31 Å from heme c 1 (the "fixed state" position) (2,5). Binding of stigmatellin or 5-n-undecyl-6-hydroxy-4,7dioxobenzothiazole enhances the electron density of the anomalous scattering peak of 2Fe2S, suggesting that these inhibitors arrest the mobility of ISP in the fixed state position (5). Conversely, binding of (E)-␤-methoxyacrylate-stilbene or myxothiazol to the complex abolishes the electron density of the anomalous scattering peak of 2Fe2S, suggesting that these inhibitors increase the mobility of ISP in the crystal and that 2Fe2S has no predominant position (referred to as the "released" or "loose" position) in this inhibited state (5). In orthorhombic crystals (P2 1 2 1 2 1 ) of the chicken enzyme, binding of stigmatellin shifts 2Fe2S from the so-called "distal or c 1 position" to the "proximal or b position" (4). The b position in the P2 1 2 1 2 1 crystal is believed to be the same as the fixed state position observed in I4 I 22 crystals. In bovine P6 5 crystals 2Fe2S is located between the b state and c 1  If movement of the head domain of ISP is required for bc 1 catalysis, locking the head domain of ISP in a given position should abolish the bc 1 complex activity. One way to lock the 2Fe2S cluster of ISP in a fixed position is to form a disulfide bond (disulfide bridge) between a pair of genetically engineered cysteines on the interface between the head domain of ISP and cytochrome b. However, genetic manipulation of bovine heart mitochondria is not practical. Rhodobacter sphaeroides is an ideal system for studying the intersubunit disulfide bond formation by molecular genetics. The four-subunit bacterial complex is functionally analogous to the mitochondrial enzyme; the largest three subunits (cytochrome b, cytochrome c 1 , and ISP) are homologous to their mitochondrial counterparts, and this system is readily manipulated genetically. In addition, R. spha-eroides expressing His 6 -tagged cytochrome bc 1 complex has been prepared (6,7). This greatly speeds up the isolation of the bc 1 complex from wild type or mutant cells.
In fact, the study of the neck region of ISP using this system (6,8) provided the first functional evidence for movement of the head domain of ISP during bc 1 catalysis. The molecule of ISP can be divided into three domains: head, tail, and neck, with the 2Fe2S cluster located at the tip of the head (9,10). Because the three-dimensional structures of the head and tail domains are rigid and are the same in the fixed and released states, a bending of the neck is required for movement of the head domain. For the neck region to bend, some flexibility is imperative. Mutants with increased neck rigidity, generated by deletion or double-or triple-proline substitution, have greatly reduced electron transfer activity with an increased activation energy (6). Formation of a disulfide bond between two engineered cysteines, having only one amino acid residue between them, in the neck region near the transmembrane helix, also drastically reduces electron transfer activity (8), presumably because of increased neck rigidity. Cleavage of the disulfide bond by reduction or alkylation restores activity to that of the wild type enzyme (8). These results clearly demonstrate a need for neck flexibility in catalysis.
To further establish that movement of the head domain of ISP is essential for the bc 1 complex, we generated mutants expressing His 6 -tagged bc 1 complex with pairs of cysteine substituted (one cysteine each) at the interface between cytochrome b and the head domain of ISP. We predicted that formation of an intersubunit disulfide bond between the engineered cysteine pair would arrest the mobility of ISP to the fixed state and decrease electron transfer activity. Herein we report procedures for generating three cysteine pair mutants with one each on cytochrome b and ISP in close proximity (interface) to each other. Mutants with single cysteine substitutions at indicated positions were also generated and characterized to confirm that the generated cysteine pair mutants are not at critical positions in the bc 1 complex and, hence, are suitable for this study. The photosynthetic growth behavior, cytochrome bc 1 complex activity, SDS-PAGE patterns, and EPR characteristics of the 2Fe2S cluster in purified complexes from wild type and mutant strains were examined and compared as were the rates of pH induced intramolecular electron transfer between 2Fe2S and heme c 1 .
The single-stranded pSELNB3503 (12) was used as the template for mutagenesis. A plate mating procedure (12) was used to mobilize the pRKDfbcF m B m C H Q plasmid in Escherichia coli S17-1 cells into R. sphaeroides BC17 cells. The presence of engineered mutations was confirmed by DNA sequencing before and after photosynthetic or semiaerobic growth of the cells. Expression plasmid pRKDfbcF m B m C H Q was purified from an aliquot of a photosynthetic or semi-aerobic culture using the Qiagen Plasmid Mini Prep kit. Because R. sphaeroides cells contain four types of endogenous plasmids, the isolated plasmids lack the purity and concentration needed for direct sequencing. Therefore, a 2.5-kilobase pair DNA segment containing the mutation sequence was amplified from the isolated plasmids by the polymerase chain reaction and purified by 1% agarose gel electrophoresis. The 2.5-kilobase pair polymerase chain reaction product was recovered from the gel with an extraction kit from Qiagen. DNA sequencing and oligonucleotide syntheses were performed by the Recombinant DNA/Protein Core Facility at the Oklahoma State University.
Growth of Bacteria-E. coli cells were grown at 37°C in LB medium. For photosynthetic growth of the plasmid-bearing R. sphaeroides BC17 cells, an enriched Siström's medium containing 5 mM glutamate and 0.2% casamino acids was used. Photosynthetic growth conditions for R. sphaeroides were essentially as described previously (6). Cells harboring mutated fbc genes on the pRKDfbcFBC H Q plasmid were grown photosynthetically for one or two serial passages to minimize any pressure for reversion. For semi-aerobic growth of R. sphaeroides, an enriched Siström's medium supplemented with 20 amino acids and extra rich vitamins was used. These semi-aerobic cultures were grown in 500 ml of enriched Siström's medium in 2-liter Bellco flasks with vigorous shaking (220 rpm) for 26 h at 30°C. The inoculation volumes used for both photosynthetic and semi-aerobic cultures were at least 5% of the total volume. Antibiotics were added to the following concentrations: ampicillin (125 g/ml), kanamycin sulfate (30 g/ml), tetracycline (10 g/ml for E. coli and 1 g/ml for R. sphaeroides), and trimethoprim (100 g/ml for E. coli and 30 g/ml for R. sphaeroides).
Enzyme Preparations and Activity Assay-Chromatophore and intracytoplasmic membrane (ICM) were prepared as described previously (6) and stored at Ϫ80°C in the presence of 20% glycerol until use. The His 6 -tagged cytochrome bc 1 complexes were purified from frozen chromatophores by the method of Tian et al. (6). Purified cytochrome bc 1 complexes were stored at Ϫ80°C in the presence of 20% glycerol. To assay ubiquinol-cytochrome c reductase activity, chromatophores, ICM, or purified cytochrome bc 1 complexes were diluted with 50 mM Tris-Cl, pH 8.0, containing 200 mM NaCl to a final concentration of cytochrome b of 5 M. Five l of the diluted samples were added to 1 ml of assay mixture containing 100 mM of Na ϩ /K ϩ phosphate buffer, pH 7.4, 0.3 mM of EDTA, 100 M of cytochrome c, and 25 M of Q 2 H 2 . Activities were determined by measuring the reduction of cytochrome c (the increase of the absorbance at wavelength 550 nm) in a Shimadzu UV 2101 PC spectrophotometer at 23°C, using a millimolar extinction coefficient of 18.5 for calculation. The nonenzymatic oxidation of Q 2 H 2 , determined under the same conditions, in the absence of enzyme, was subtracted during specific activity calculations.
Determination of pH-induced Reduction and Oxidation of ISP and Cytochrome c 1 in the Partially Reduced Wild Type and Mutant bc 1 Complexes-The wild type or mutant bc 1 complex was diluted in 3 ml of 20 mM Tris-Cl buffer, pH 8.0, containing 200 mM NaCl and 0.01% dodecylmaltoside. The concentration of cytochrome c 1 was adjusted to about 10 M. Different amounts of NaOH or HCl were added to give the indicated pH levels. Fully oxidized or reduced cytochrome c 1 and ISP was obtained by addition of K 3 Fe(CN) 6 or sodium ascorbate. Reduction of cytochrome c 1 was followed by measuring the increase of the ␣-absorption (553-545 nm) in a Shimadzu UV 2101 PC spectrophotometer. Reduction of ISP was followed by measuring the negative CD peak, at 500 nm, of partially reduced ISP minus fully oxidized complex in a JASCO J-715 spectropolarimeter (13)(14)(15). The same samples were used for the absorption and CD measurements. Instrument settings for the spectropolarimeter were: scan speed, 100 nm/min; step resolution, 1 nm; accumulation, 10 traces for averaging; response, 1 s; bandwidth, 1.0 nm; sensitivity, 10 mdeg; and slit width, 500 m.
Determination of Redox Potentials of the 2Fe2S Cluster in Wild Type and Mutant bc 1 Complexes-The redox status of heme c 1 or the 2Fe2S cluster was determined as described above. The cytochrome c 1 partially reduced wild type and mutant bc 1 complexes were prepared and used for the absorption and CD measurements. The redox potentials of ISP were calculated from the redox statuses of heme c 1 and 2Fe2S, at pH 8.0, using 230 mV for the redox potential of heme c 1 (16). 1 Complexes-The method used is essentially the same as that previously reported (15). The His 6 -tagged, cytochrome c 1 half-reduced cytochrome bc 1 complex was prepared as described for the preparation of the oxidized complex (6), except the cytochrome c 1 in the dodecylmaltoside-solubilized chromatophore was reduced 50% with sodium ascorbate before being applied to a nickel-nitrilotriacetic acid column. The cytochrome c 1 half-reduced complex was diluted in 20 mM Tris-Cl buffer, pH 6.8 or 8.9, containing 200 mM NaCl and 0.01% dodecylmaltoside to a cytochrome c 1 concentration of around 10 M and mixed with buffers of various pHs at room temperature in an Olis stopped flow rapid scanning spectrophotometer. Oxidation or reduction of cytochrome c 1 was monitored by the decrease or increase of absorption at 553 nm minus 545 nm.

Determination of Electron Transfer Rates between the 2Fe2S Cluster and Heme c 1 in Wild Type and Mutant bc
Other Biochemical and Biophysical Techniques-Protein concentration was determined by the method of Lowry et al. (17). Cytochrome b (18) and cytochrome c 1 (16) contents were determined according to published methods. SDS-PAGE was performed according to Laemmli (19) using a Bio-Rad Mini-Protean dual slab vertical cell. Samples were digested with 10 mM Tris-Cl buffer, pH 6.8, containing 1% SDS, and 3% glycerol in the presence and absence of 0.4% ␤-ME for 2 h at 37°C before being subjected to electrophoresis. The intersubunit disulfide bond linked adduct protein was obtained by electrophoretic elution (20) of a protein band, with an apparent molecular mass of 64 kDa, from SDS-PAGE of the purified A185C(cytb)/K70C(ISP) mutant complex without ␤-ME treatment. Western blotting was performed with rabbit polyclonal antibodies against cytochrome b, cytochrome c 1 , and ISP of the R. sphaeroides bc 1 complex (6). The polypeptides separated by SDS-PAGE gel were transferred to polyvinylidene difluoride membrane for immunoblotting. Goat anti-rabbit IgG conjugated to alkaline phosphatase or protein A conjugated to horseradish peroxidase was used as the second antibody. EPR spectra were recorded in a Bruker ER 200D apparatus equipped with liquid Nitrogen Dewar, at 77 K. The instrument settings are detailed in the figure legends.

Photosynthetic Growth Behaviors of Mutants Carrying Cysteine Substitutions in the Interface between Cytochrome b and the Head Domain of ISP-Three pairs of amino acid residues:
Ala 185 (cytb)/Lys 70 (ISP), Ile 326 (cytb)/Gly 165 (ISP), and Thr 386 -(cytb)/Lys 164 (ISP) were selected for mutation to cysteines. These choices were based on the three-dimensional structural model of the four-subunit cytochrome bc 1 complex of R. sphaeroides (Fig. 1A) constructed with coordinates from bovine cytochromes b and c 1 , ISP, and subunit XII (21). The distances between these three cysteine pairs are 6.1, 6.1, and, 7.5 Å, respectively, in the bacterial complex (Fig. 1B). They are 6.5, 6.8, and 6.4 Å (Table I), respectively, when calculations are based on corresponding residues in the bovine enzyme. Mutants with a single cysteine substitution, at positions Ala 185 -(cytb), Ile 326 (cytb), Thr 386 (cytb), Lys 70 (ISP), Gly 165 (ISP), or Lys 164 (ISP), were also generated and used as controls.
For a cysteine pair mutant to be useful in this study, the engineered cysteine positions must not be critical for cytochrome bc 1 complex activity. Because the bc 1 complex is absolutely required for the photosynthetic growth of R. sphaeroides, whether or not the engineered cysteine positions are critical to the complex can be determined by assaying photosynthetic growth. Mutants with cysteine substitutions at critical positions in the complex will not grow photosynthetically, whereas mutants with substitutions at noncritical positions will grow.
Because Gly 165 of ISP is a critical position, mutant I326C-(cytb)/G165C(ISP) does not support photosynthetic growth and cannot be used to study the effect of disulfide bond formation on the bc 1 complex. The structural importance of Gly 165 was further investigated by substituting alanine or threonine at this position. The ISP:G165A or G165T substitution also results in cells that do not grow photosynthetically, indicating that the size of the amino acid side chain at position 165 of ISP is critical. A similar size-activity relationship was previously observed for Gly 158 of cytochrome b in Rhodobacter capsulatus (22) and Ser 155 of cytochrome b in R. sphaeroides (23).
On the other hand, mutants A185C(cytb)/K70C(ISP) and T386C(cytb)/K164C(ISP) support photosynthetic growth, indicating that the engineered cysteine positions are noncritical to the complex. Therefore, formation of an intersubunit disulfide bond between cytochrome b and the ISP head domain was examined with these two cysteine pair mutants.
Formation of a Disulfide Bond between Cytochrome b and  (Table I). When these chromatophore preparations were subjected to Western blot analysis using antibodies against R. sphaeroides cytochrome b and ISP, no protein band corresponding to the adduct of cytochrome b and ISP was observed, indicating that no disulfide bond is formed between the two engineered cysteines in mutants A185C(cytb)/ K70C(ISP) and T386C(cytb)/K164C(ISP) during anaerobic photosynthetic growth. The lack of disulfide bond formation is expected because photosynthetic growth is under strict anaerobic conditions, whereas disulfide bond formation is an oxidative process. Without oxygen no disulfide bond can be formed even if the two cysteines are in favorable positions.
When the His 6 -tagged bc 1 complexes were purified from these six freshly prepared cysteine mutant chromatophores, all but the A185C(cytb)/K70C(ISP) mutant complex have the same bc 1 activity found in their respective chromatophores (Table I), based on cytochrome b content. The A185C(cytb)/K70C(ISP) mutant complex, when assayed immediately after preparation, has about 23% of the bc 1 complex activity found in its freshly prepared chromatophores. Activity in this cysteine pair mutant complex decreased during storage at 0°C. About 7% of the activity remained after 24 h. Under identical conditions, no activity loss was observed for wild type and mutant complexes of A185C(cytb), K70C(ISP), T386C(cytb), K164C(ISP), and T386C(cytb)/K164C(ISP).
To see whether or not the loss of bc 1 complex activity observed for mutant A185C(cytb)/K70C(ISP) results from disulfide bond formation during purification, SDS-PAGE patterns of these purified mutant complexes, with and without ␤-ME treatment, were examined (Fig. 2). When purified complexes were treated with SDS at 37°C for 2 h and subjected to electrophoresis in the absence of ␤-ME (Fig. 2, lanes marked with  Ϫ), the A185C(cytb), K70C(ISP), T386C(cytb), and K164C(ISP) mutants have the same electrophoretic pattern as that of the wild type complex. They all contain five protein bands with apparent molecular masses of 41, 33, 31, 23, and 14 kDa. Western blot analysis with antibodies against R. sphaeroides cytochrome b, cytochrome c 1 , ISP, and subunit IV identified these five protein bands, with decreasing molecular masses, as band I, cytochrome b; bands II and IIa, cytochrome c 1 ; band III, ISP; and band IV, subunit IV. The lack of any protein band with an apparent molecular mass of 64 kDa, the size of an adduct band of cytochrome b and ISP, indicates that no intersubunit disulfide bond is formed between an engineered cysteine in cytochrome b (A185C or T386C) or ISP (K70C or K164C) and an endogenous cysteine in ISP or cytochrome b, respectively.
This lack of intersubunit disulfide bond formation between an engineered cysteine in cytochrome b or ISP and an endogenous cysteine in ISP or cytochrome b is as expected. The R. sphaeroides cytochrome bc 1 complex has nine endogenous cysteine residues: one in cytochrome b, four in cytochrome c 1 , and four in the ISP. It has been established that two cysteines (Cys 129 and Cys 149 ) in the ISP serve as ligands for the 2Fe2S cluster (24) and two cysteines (Cys 37 and Cys 40 ) in cytochrome c 1 are covalently bonded to heme c 1 (25). Thus, there are five endogenous free cysteines in the complex: two in ISP (Cys 134 and Cys 151 ), one in cytochrome b (Cys 54 ), and two in cytochrome c 1 (Cys 145 and Cys 169 ). In the three-dimensional structural model, Cys 134 is on the loop ␤4-␤5 and Cys 151 is on the loop ␤6-␤7 of ISP (21). They form an intrasubunit disulfide bond that brings the two loops together and stabilizes this region. Therefore, they are not available for intersubunit disulfide bond formation with the engineered cysteines on cytochrome b. Because the endogenous cysteine in cytochrome b (Cys 54 ) is, respectively, 41.6 and 43.2 Å from the engineered cysteines at positions 70 and 164 of ISP (Fig. 1B), formation of an intersubunit disulfide bond between this endogenous free cysteine and the engineered cysteines on ISP is unlikely.
The presence of two bands (II and IIa) for cytochrome c 1 in SDS-PAGE of the bc 1 complex, in the absence of ␤-ME, may result from the formation of an intramolecular disulfide bond between Cys 145 and Cys 169 in some of the cytochrome c 1 molecules. The cytochrome c 1 molecules with the intramolecular disulfide bond move faster (band IIa) than those without this bond. This explanation is consistent with the presence of a single protein band, with slower electrophoretic mobility, in SDS-PAGE patterns obtained after treating the complex with SDS and ␤-ME prior to electrophoresis (Fig. 2, lanes marked  with ϩ). No doubt ␤-ME reduces the intramolecular disulfide bond present in some cytochrome c 1 molecules, thus converting the fast moving band to the slower one. Probably intramolecular disulfide bond formation occurs only after the complex has a The distances were measured from C-␤ to C-␤ (except with glycine, which is to C-␣). b Ps, photosynthetic growth. ϩϩ, the cells growth rate is essentially the same as that of the wild type cells; ϩ, the cells can grow photosynthetically but at a rate slower than that of the wild type cells; Ϫ, no photosynthetic growth in 4 days.
c Enzymatic activity is expressed as mol of cytochrome c reduced/min/nmol cytochrome b at room temperature. been dissociated with SDS. The cytochrome bc 1 complex, in its native state, with or without ␤-ME treatment, has the same electron transfer activity. Moreover, the distance between Cys 169 and Cys 145 of cytochrome c 1 in the structural model of this bacterial complex is 23.5 Å (Fig. 1B), too large for disulfide bond formation. The electrophoretic pattern of the T386C(cytb)/K164C(ISP) mutant complex in the absence of ␤-ME (Fig. 2, lane 11) shows a faint band with an apparent molecular mass of 64 kDa that disappears in the presence of ␤-ME (Fig. 2, lane 10). The failure of this 64-kDa protein to react with antibodies against cytochrome b and ISP and the lack of ␤-ME effect on cytochrome bc 1 activity of this mutant complex lead us to assign this protein as a contaminant rather than an adduct of ISP and cytochrome b. Thus, no intersubunit disulfide bond is formed between cytochrome b and ISP in the T386C(cytb)/K164C(ISP) mutant complex.
The electrophoretic pattern of the A185C(cytb)/K70C(ISP) mutant complex (Fig. 2, lane 5), in the absence of ␤-ME, differs from those of wild type and mutant complexes of A185C(cytb), K70C(ISP), T386C(cytb), K164C(ISP), and T386C(cytb)/ K164C(ISP) in two aspects: the appearance of a protein band with an apparent molecular mass of 64 kDa and a decrease in the band intensities of cytochrome b and ISP. The 64-kDa protein band is established as an adduct of cytochrome b and ISP, resulting from intersubunit disulfide bond formation between these two subunits, by the following experimental results. First, the 64-kDa protein band reacts with antibodies against cytochrome b and ISP (data not shown); second, when the mutant complex is treated with SDS and ␤-ME and then subjected to SDS-PAGE in the presence of ␤-ME, the 64-kDa protein band disappears, and the band intensities of cytochrome b and ISP increase (Fig. 2, lane 4); and third, the protein isolated from the 64-kDa gel slice, after treatment with ␤-ME, can be resolved into cytochrome b and ISP (data not shown).
Although the 64-kDa protein band that reacts with antibodies against cytochrome b and ISP is absent from freshly prepared chromatophores from A185C(cytb)/K70C(ISP) mutant cells, it appears in ICM preparations obtained from cells grown semi-aerobically. The A185C(cytb)/K70C(ISP) mutant ICM has no cytochrome bc 1 complex activity (Table II). Also, when freshly prepared mutant chromatophores are incubated at 0°C under air, the 64-kDa protein band intensity increases as the bc 1 activity decreases. It usually takes more than a week for the loss in activity to reach 50%. These results confirm that formation of a disulfide bond from the two engineered cysteines on cytochrome b and ISP in the A185C(cytb)/K70C(ISP) mutant complex is promoted by oxygen and is directly related to the loss of bc 1 activity in this mutant complex.
Effect of ␤-ME on the Disulfide Bond Formation and bc 1 Activity-To further confirm that the formation of a disulfide bond between cytochrome b and ISP causes the A185C(cytb)/ K70C(ISP) mutant complex to lose bc 1 complex activity, the effect of ␤-ME on bc 1 complex activity and disulfide bond formation was examined. When purified A185C(cytb)/K70C(ISP) mutant complex, which has been incubated at 0°C for 24 h, was treated with ␤-ME, the activity was restored to the same level as that in freshly prepared chromatophores (Table II). No adduct of cytochrome b and ISP was detected in the ␤-ME treated A185C(cytb)/K70C(ISP) mutant complex. When this complex was purified from the freshly prepared chromatophores in the presence of 100 mM ␤-ME, it had the same activity as that found in the chromatophores, and no cytochrome b-ISP adduct was detected. A similar ␤-ME effect is observed for the cytochrome bc 1 complex in mutant ICM (Table  II). It should be emphasized that the observed activity restoration or preservation is not due to nonenzymatic reduction of cytochrome c by ␤-ME, because only antimycin-sensitive cytochrome c reduction is used for activity calculations. Under identical conditions, ␤-ME has no effect on the bc 1 activities in the A185C(cytb), K70C(ISP), T386C(cytb), K164C(ISP), and T386C(cytb)/K70C(ISP) mutant complexes.
Effect of Disulfide Bond Formation between Cytochrome b and ISP on EPR Characteristics and Redox Potential of the 2Fe2S Cluster-Although evidence presented above clearly Enzymatic activity is expressed as mol of cytochrome c reduced/ min/nmol cytochrome b at room temperature. b ϩ samples were prepared with buffers containing 100 mM ␤-ME. c Ϫ samples were prepared in the absence of ␤-ME. d The bc 1 activity was measured after the sample had been stored on ice for 1 day. demonstrates that the loss of bc 1 complex activity in the A185C(cytb)/K70C(ISP) mutant correlates with the formation of a disulfide bond between the two engineered cysteines in cytochrome b and ISP, it is unknown how disulfide bond formation causes the activity loss. Because this disulfide bond is formed in the interface between cytochrome b and ISP, it is possible that the microenvironments or the redox potential of the ISP cluster are altered, thus leading to activity loss. To test this possibility, EPR characteristics and redox potentials of the 2Fe2S cluster in the A185C(cytb)/K70C(ISP), with and without ␤-ME treatment, are compared.
As shown in Fig. 3, the 2Fe2S in the A185C(cytb)/K70C(ISP) mutant complex, with or without ␤-ME treatment, has the same EPR spectrum, with the g x signal at 1.775, g y at 1.900, and g z at 2.020. This result indicates that the microenvironments of the ISP cluster in the A185C(cytb)/K70C(ISP) mutant complex are not affected by disulfide bond formation between the two engineered cysteines in cytochrome b and ISP. Therefore, the loss of bc 1 activity is not due to a change of microenvironments in the ISP.
However, it should be noted that replacing the K70 of ISP with cysteine in the A185C(cytb)/K70C(ISP) mutant complex shifts the g x signal of the 2Fe2S cluster from 1.800 to 1.775. This is deduced from the observation that the 2Fe2S cluster in the A185C(cytb) mutant complex has an EPR spectrum identical to that observed in the wild type complex, with resonance at g x ϭ 1.800, g y ϭ 1.900, and g z ϭ 2.020 (Fig. 3), whereas the 2Fe2S in the K70C(ISP) mutant complex has a broadened g x signal that is shifted to 1.768. The g x signal of 2Fe2S in the A185C(cytb)/K70C(ISP) mutant complex is sharper than that in the K70C(ISP) bc 1 complex but broader than that in the wild type or A185C(cytb) mutant complex. As expected, the effect on the bc 1 complex is small because the K70C(ISP) and ␤-MEtreated A185C(cytb)/K70C(ISP) mutant complexes retain more than 50% of the bc 1 activity found in the wild type.
The line shape of the g x signal of the 2Fe2S cluster is thought to be mediated by the oxidation state of ubiquinone present in the Q o site (26 -30). The g x of bc 1 from R. sphaeroides is at g ϭ 1.800 when ubiquinone is present but shifts to 1.750 and broadens when ubiquinol is present. When ubiquinone is extracted from chromatophore membranes, the g x signal of the "depleted state" is at g ϭ 1.765 and is considerably broader than those seen in the presence of either ubiquinone or ubiquinol. The change in the g x signal because of oxidation-reduction state of Q in the Q o site of the bc 1 complex is similar to that observed for the substitution of Leu for Phe 144 (F144L) in the cytochrome b from R. capsulatus (29). The F144L bc 1 complex in R. capsulatus chromatophores was reported to have a very low turnover rate with a broadened, redox state-insensitive, g x value at 1.765. It was suggested that these properties of the F144L complex resulted from a reduced affinity for quinone and quinol exhibited by the Q o center of the mutated complex. Because the Lys 70 of ISP is in the vicinity of the putative Q o pocket of cytochrome b, perhaps substitution of Lys 70 with cysteine, as in the K70C and ␤-ME-treated A185C/K70C mutant complexes, reduces the affinity of the Q o site of cytochrome b for quinone and quinol and thus decreases activity.
The redox potentials of the 2Fe2S clusters in the complexes of wild type, A185C(cytb), K70C(ISP), and A185C(cytb)/ K70C(ISP) are 231, 228, 234, and 232 mV, respectively. Because the redox potentials are similar, the loss of activity in the cysteine pair mutant cannot be attributed to a change of the redox potential of the 2Fe2S cluster of ISP.
Effect of the Disulfide Bond Formation between Cytochrome b and ISP on the Rate of Intramolecular Electron Transfer between 2Fe2S and Heme c 1 -One way to unambiguously establish that formation of an intersubunit disulfide bond between cytochrome b and ISP arrests the movement of the head domain of ISP to the fixed state is to compare the rate of intramolecular electron transfer between 2Fe2S and heme c 1 in wild type and cysteine pair mutant complexes.
It has been reported that intramolecular electron transfer between heme c 1 and the 2Fe2S cluster in the bovine complex can be induced by changing the pH of the enzyme solution (15). This is based on the fact that the redox potential of heme c 1 is independent of pH, whereas the redox potential of 2Fe2S is pH-dependent (higher the pH lower the redox potential). At pH 8.0 heme c 1 and 2Fe2S have the same redox potentials. Thus, when the pH of the enzyme solution is raised above 8.0, the The complexes were incubated with 5 mM sodium ascorbate on ice for 30 min and frozen in liquid nitrogen. EPR spectra were recorded at 77 K with the following instruments settings: microwave frequency, 9.28 Hz; microwave power, 20 milliwatts; modulation amplitude, 20 G; modulation frequency, 100 kHz; time constants, 0.1 s; and scan rate, 20 G/s. 2Fe2S becomes less reduced than heme c 1 if the preparation is 50% reduced at pH 8.0. However, when the pH is adjusted to a lower value than 8.0, the 2Fe2S becomes more reduced than heme c 1 . Electron shuffling between 2Fe2S and heme c 1 , in the partially reduced complex, is pH-dependent.
To be sure that the electron transfer between heme c 1 and 2Fe2S in the R. sphaeroides bc 1 complex can also be induced by a change of pH, the redox status of heme c 1 and 2Fe2S in a cytochrome c 1 half-reduced wild type bc 1 complex was monitored at various pHs (Fig. 4). Similar to results obtained with the bovine complex, at pH 8.1, heme c 1 has the same redox potential as the 2Fe2S cluster in the R. sphaeroides bc 1 complex. When the pH is increased, 2Fe2S becomes less reduced, and when it is lowered 2Fe2S becomes more reduced; heme c 1 changes in opposite directions. Fig. 5 shows time trace of acidification and alkalizationinduced intramolecular electron transfer between 2Fe2S and heme c 1 in wild type and A185C(cytb)/K70C(ISP) mutant complexes. The rates of electron transfer from heme c 1 to 2Fe2S induced by acidification (reverse) and from 2Fe2S to heme c 1 (forward) induced by alkalization of wild type bacterial complex both have a reaction half-life (t1 ⁄2 ) of about 1-2 ms (Fig. 5, A and  C). The rates of pH-induced electron transfer between 2Fe2S and heme c 1 in the A185C(cytb)/K70C(ISP) mutant complex are very slow; the t1 ⁄2 for the forward and backward reactions are 100 and 10 s, respectively (Fig. 5, D and B). The rates of electron transfer between 2Fe2S and heme c 1 in mutant complexes of A185C(cytb) and K70C(ISP) are the same as those observed in the wild type complex (data not shown). These results confirm that formation of a disulfide bond between the two engineered cysteines on cytochrome b and ISP in the A185C(cytb)/K70C(ISP) mutant complex arrests the movement of the head domain of ISP, required for bc 1 catalysis, to a fixed state. When the 2Fe2S cluster of ISP is in the fixed state, it is about 31 Å away from heme c 1 , too far for electron transfer.
The observation that the rates of electron transfer between 2Fe2S and heme c 1 in mutant complexes of A185C(cytb) and K70C(ISP), which have, respectively, 100 and 53% of the ubiquinol-cytochrome c reductase activity found in the wild type complex, are the same as that observed in the wild type complex suggests that electron transfer between 2Fe2S and heme c 1 is not the rate-limiting step in the electron transfer reaction catalyzed by the bc 1 complex (from ubiquinol to cytochrome c). This suggestion is also consistent with the observation that the rate of electron transfer from 2Fe2S to heme c 1 in the wild type bacterial complex is comparable with that in the bovine complex (15), even though the rate of electron transfer from ubiquinol to cytochrome c catalyzed by the bacterial complex is only one-tenth that of the bovine complex.
The fact that formation of an intersubunit disulfide bond between the engineered cysteines at the position 185 of cytochrome b and 70 of ISP results in loss of bc 1 activity suggests that movement of the head domain of ISP is independent from cytochrome b, at least in the cd2 region where the A185C is located. However, some regions of cytochrome b may have a synchronous movement with the head domain of ISP and may even provide the driving force for the movement of the ISP head. Formation of an intersubunit disulfide bond between such a region and the head domain of ISP should not have an adverse effect on electron transfer activity. FIG. 5. Time trace of the redox change of cytochrome c 1 in the cytochrome bc 1 complex upon changing the pH from 6.9 to 8.9 or from 8.9 to 6.9. Purified cytochrome bc 1 complexes were diluted in 20 mM Tris-Cl buffer, pH 6.9 (for the alkalization experiment) or 8.9 (for the acidification experiment), containing 200 mM NaCl, 0.01% dodecylmultoside to a cytochrome c 1 concentration of about 10 M. The diluted complex was rapidly mixed with an equal volume of buffer containing enough NaOH or HCl to cause the desired pH changes, at room temperature, in an Olis stopped flow rapid scanning spectrophotometer. A and B, oxidation of cytochrome c 1 in wild type (WT) and A185C(cytb)/K70C(ISP) mutant complexes, respectively. C and D, reduction of cytochrome c 1 in wild type and A185C(cytb)/K70C(ISP) mutant complexes, respectively. The reaction was monitored at 553 nm Ϫ 545 nm. Complex diluted with an equal volume of the same buffer was used as a base line.