Effect of Famoxadone on Photoinduced Electron Transfer between the Iron-Sulfur Center and Cytochrome c 1 in the Cytochrome bc 1 Complex*

Famoxadone is a new cytochromebc 1 Qo site inhibitor that immobilizes the iron-sulfur protein (ISP) in the bconformation. The effects of famoxadone on electron transfer between the iron-sulfur center (2Fe-2S) and cyt c 1 were studied using a ruthenium dimer to photoinitiate the reaction. The rate constant for electron transfer in the forward direction from 2Fe-2S to cyt c 1 was found to be 16,000 s−1in bovine cyt bc 1. Binding famoxadone decreased this rate constant to 1,480 s−1, consistent with a decrease in mobility of the ISP. Reverse electron transfer from cytc 1 to 2Fe-2S was found to be biphasic in bovine cyt bc 1 with rate constants of 90,000 and 7,300 s−1. In the presence of famoxadone, reverse electron transfer was monophasic with a rate constant of 1,420 s−1. It appears that the rate constants for the release of the oxidized and reduced ISP from the b conformation are the same in the presence of famoxadone. The effects of famoxadone binding on electron transfer were also studied in a series of Rhodobacter sphaeroides cyt bc 1 mutants involving residues at the interface between the Rieske protein and cytc 1 and/or cyt b.

The cytochrome (cyt) 1 bc 1 complex (ubiquinol:cytochrome c oxidoreductase) is an integral membrane protein in the electron transport chains of mitochondria and many respiratory and photosynthetic prokaryotes (1). The complex translocates four protons to the positive side of the membrane per two electrons transferred from ubiquinol to cyt c in a Q-cycle mechanism (2). A key-bifurcated reaction occurs at the Q o site in which the first electron is transferred from ubiquinol to the Rieske iron-sulfur center (2Fe-2S) and then to cyt c 1 and cyt c (1)(2)(3). The second electron is transferred from semiquinone in the Q o site to cyt b L and then to cyt b H and ubiquinone in the Q i site. Q o site inhibitors can be divided into three classes. Class Ia inhibitors such as myxothiazol and (E)-␤-methoxyac-rylate-stilbene alter the heme spectrum of cyt b L , class Ib inhibitors such as UHDBT alter the EPR spectrum of 2Fe-2S, and class Ic inhibitors such as stigmatellin alter both (1-3). X-ray crystallographic studies have revealed that these different classes of inhibitors occupy different subsites in the Q o pocket and have different effects on the mobility of the extramembrane domain of the Rieske iron-sulfur protein (ISP) (4 -8). Stigmatellin binding increases the midpoint redox potential (E m ) of 2Fe-2S by ϳ200 -250 mV (9) and immobilizes the ISP in the b conformation with the His-161 ligand on reduced 2Fe-2S in hydrogen-bonding contact with stigmatellin in the Q o site (4 -8). In contrast, (E)-␤-methoxyacrylate-stilbene binding to the Q o site completely eliminates the anomalous scattering signal for 2Fe-2S close to cyt b L , indicating the release of the ISP from the b conformation to a mobile state (7). These studies led to the proposal that the ISP functions as a mobile shuttle as it transfers an electron from Q-H 2 in the Q o site to cyt c 1 (4 -8). Experimental support for the mobile shuttle mechanism has been obtained from cross-linking and mutational studies that immobilize the ISP or alter the conformation of the neck region (10 -19).
Famoxadone and azoxystrobin are new Q o site inhibitors that have significantly different effects on cyt bc 1 than other Q o site inhibitors (20 -22). X-ray crystallographic studies revealed that famoxadone binding in the Q o site is stabilized by a network of interactions among the three aromatic groups on famoxadone and aromatic residues in the binding pocket ( Fig. 1, top) (23). Famoxadone binding leads to extensive conformational changes on the surface of cyt b and triggers a long range conformational change in the ISP from the mobile state to a state with 2Fe-2S proximal to cyt b (23). In contrast to stigmatellin, famoxadone increases the E m of 2Fe-2S by only 26 mV and immobilizes both the oxidized and reduced ISP in the b conformation. 2 This is consistent with the finding that famoxadone is more deeply buried in the Q o site than stigmatellin and that it does not form a hydrogen bond with the His-161 ligand on reduced 2Fe-2S (5,8,23). Azoxystrobin also immobilizes the ISP in the b conformation but has only a minor effect on the E m of 2Fe-2S, decreasing it by 24 mV. 2 In this paper, the effects of famoxadone and azoxystrobin on electron transfer between the Rieske iron-sulfur center and cyt c 1 are studied using the binuclear ruthenium complex, Ru 2 D, to rapidly photooxidize or photoreduce cyt c 1 (10). Binding famoxadone to the Q o site of bovine cyt bc 1 decreased the rate constant for electron transfer from 2Fe-2S to cyt c 1 from 16,000 s Ϫ1 to 1,480 s Ϫ1 , consistent with a decrease in the mobility of ISP. Famoxadone binding also decreased the rate constant for reverse electron transfer from cyt c 1 to 2Fe-2S to 1,420 s Ϫ1 , indicating that the rate constants for the release of oxidized and reduced ISP from the b conformation are the same in the presence of famoxadone. Famoxadone binding was also found to decrease the rate of electron transfer from 2Fe-2S to cyt c 1 in R. sphaeroides cyt bc 1 as well as mutants involving residues at the interface between the Rieske protein and cyt c 1 and/or cyt b.
Generation and Expression of R. sphaeroides Cyt bc 1 Mutants-Mutations were constructed by site-directed mutagenesis using the Altered Sites system from Promega. The single-stranded pSELNB3503 (28)  A plate-mating procedure (28) was used to mobilize the pRKDfb-cFmBmCHQ plasmid in Escherichia coli S17-1 cells into R. sphaeroides BC17 cells as described previously (12). Growth of E. coli cells and plasmid-bearing R. sphaeroides cells were carried out as described previously (12). The identity of the mutations was confirmed by DNA sequencing before and after photosynthetic or semi-aerobic growth of the cells as described previously (12). Mutant cytochrome bc 1 was purified as described by Xiao et al. (12). Bottom, x-ray crystal structure of bovine cyt bc 1 P6 5 22 crystals (c 1 conformation) (6). ISP is blue, cyt c 1 is yellow, and the residues that were mutated are shown as sticks. R. sphaeroides numbering is used.
Determination of Enzyme Activity and Redox Potential of the 2Fe-2S Cluster in Mutant Cyt bc 1 -The cyt bc 1 activity was determined in an assay mixture containing 100 mM Na ϩ /K ϩ phosphate buffer, pH 7.4, 300 M EDTA, 100 M cyt c, and 25 M 2,3-dimethoxy-5-methyl-6-(10bromodecyl)-1,4-benzoquinol (Q 0 C 10 BrH 2 ) at 23°C using the method described by Xiao et al. (12). The redox potentials of 2Fe-2S in cyt bc 1 mutants were determined as described previously (12). The reduction of cyt c 1 was followed by measuring the increase of the ␣-absorption (553-545 nm) in a Shimadzu UV2101 PC spectrophotometer. The reduction of 2Fe-2S was followed by measuring the negative CD peak at 500 nm of partially reduced complex minus fully oxidized complex in a JASCO J-715 spectropolarimeter (29 -31). The same samples were used for the absorption and CD measurements. The redox potentials of 2Fe-2S were calculated from the redox states of heme c 1 and 2Fe-2S at pH 8.0 using 280 mV for the midpoint redox potential of heme c 1 (32).
Flash Photolysis Experiments-Transient absorbance measurements were carried out by flash photolysis of 300-l solutions contained in a 1-cm glass semi-microcuvette. The excitation light flash was provided by a Phase R model DL1400 flash lamp-pumped dye laser using coumarin LD 490 to produce a 480-nm light flash of Ͻ0.5-s duration. The detection system has been described by Heacock et al. (33). Samples typically contained 5 M cyt bc 1 in a buffer with 0.01% dodecylmaltoside. In photoreduction experiments, 10 mM aniline and 1 mM 3CP were used as sacrificial donors, and catalytic concentrations of horse cyt c and bovine cyt oxidase (20 nM) were present to maintain cyt bc 1 in the oxidized state. In photooxidation experiments, paraquat or [Co(NH 3 ) 5 Cl] 2ϩ was used as sacrificial acceptors, and Q 0 C 10 BrH 2 was used to reduce cyt bc 1 . To regenerate reduced quinol throughout the flash experiments, 1 mM succinate and 50 nM SCR were included. Redox mediators included p-benzoquinone (⑀ m ϭ ϩ280 mV) and TMPD (⑀ m ϭ ϩ275 mV) used in concentrations of 10 and 2 M, respectively. The experiments were carried out aerobically to rapidly reoxidize the highly absorbing reduced paraquat.

Effects of Famoxadone and Azoxystrobin on Electron Transfer between 2Fe-2S and Cyt c 1 in Bovine Cyt bc 1 -An important
goal toward understanding the mechanism of electron transfer in cyt bc 1 is to determine what factors control the conformation of the ISP in each state of the complex, the dynamics of the changes between the different conformations, and the rate of electron transfer in each of the conformations. X-ray crystallographic studies are providing valuable information on the conformations of the ISP including the b conformation ( Fig. 1, top), the c 1 conformation (Fig. 1, bottom), and intermediate and mobile conformations (4 -8). However, it has been difficult to determine the kinetics of electron transfer from 2Fe-2S to cyt c 1 (34, 35) as well as the dynamics of ISP conformational changes. The development of the ruthenium photoreduction method provides an opportunity to measure electron transfer between 2Fe-2S and cyt c 1 in both the forward and reverse directions and thus provides kinetic information on two different initial redox states of cyt bc 1 (10). Moreover, it is becoming clear that the measured rates of electron transfer are probably ratelimited by conformational changes in the ISP (36). The binuclear complex Ru 2 D contains the 2,2Ј:4Ј,4Љ:2Љ,2ٞ-quaterpyridine ligand, which bridges the two ruthenium atoms ( Fig.  2A, inset) (24). Ru 2 D has a charge of ϩ4, which allows it to bind with high affinity to the negatively charged domain on cyt c 1 (10). The electrochemical properties of Ru 2 D are similar to those of the widely used ruthenium tris-bipyridine complex ( Table I). The metal-to-ligand excited state of Ru 2 D with a lifetime of 0.5 s is both a strong oxidant and a strong reductant and can rapidly oxidize or reduce cyt c 1 in the presence of appropriate sacrificial electron acceptors or donors (10).
To study electron transfer in the forward direction, the ruthenium dimer Ru 2 D was used to rapidly photooxidize cyt c 1 in bovine cyt bc 1 with cyt c 1 and 2Fe-2S was initially reduced ( Fig.  2A). The excited state of Ru 2 D oxidizes cyt c 1 within 1 s according to Scheme 1. Only one of the two ruthenium centers in Ru 2 D is photoexcited in this experiment, and this one is represented in Scheme 1. The sacrificial electron acceptor paraquat was present in the solution to oxidize Ru(II*) and/or Ru(I).
The rapid photooxidation of cyt c 1 shown by the initial decrease in 552-nm absorbance was followed by biphasic reduction of cyt c 1 with rate constants of k 1 ϭ 16,000 Ϯ 3,000 s Ϫ1 and k 2 ϭ 250 Ϯ 50 s Ϫ1 (Fig. 2A). The rate constant k 1 has been assigned to electron transfer from reduced 2Fe-2S to photooxidized cyt c 1 , whereas k 2 is correlated with the oxidant-induced reduction of cyt b H (Scheme 2) (10). The rate constant k 2 thus represents electron transfer of the first electron from Q-H 2 to 2Fe-2S and cyt c 1 followed rapidly by the transfer of the second electron from the semiquinone to cyt b L and cyt b H (10). The experimental rate constant k 1 is much smaller than the theoretical value predicted for electron transfer between 2Fe-2S and cyt c 1 in the crystallographic c 1 state, 1 ϫ 10 6 -2 ϫ 10 7 s Ϫ1 (5, 10). Therefore, it appears that the measured rate constant for electron transfer is "gated" by changes in the conformation of the ISP (10). This interpretation is consistent with previous studies showing that k 1 has a large temperature dependence with an activation energy of 59 kJ/mol (10). X-ray diffraction studies have indicated that the ISP is largely in the b conformation when both cyt c 1 and 2Fe-2S are reduced (37). Therefore, the rate constant a Determined by cyclic voltammetry at a platinum disc-working electrode in acetonitrile with respect to a saturated calomel reference electrode (24).
b Calculated from the excited state energy and the ground state potential as described previously (39). The addition of 30 M famoxadone to bovine cyt bc 1 led to a single phase of reduction of photooxidized cyt c 1 with a rate constant of 1,480 Ϯ 250 s Ϫ1 , and no reduction of cyt b H was observed at 562 nm (Fig. 2B). The rate constant was independent of the concentration of famoxadone over the range of 10 -100 M, consistent with a large binding constant. These results indicate that famoxadone binds strongly to the Q o site and decreases the rate constant for electron transfer from 2Fe-2S to cyt c 1 from 16,000 to 1,480 s Ϫ1 . It appears that famoxadone binding significantly decreases the rate constant for the conformational change from the b state to the c 1 state. The amplitude of the 552-nm transient for cyt c 1 reduction in the presence of famoxadone is 31% of the amplitude of the initial photooxidation, approximately the same as in the absence of famoxadone (28%). Because 2Fe-2S and cyt c 1 have the same redox potentials at pH 8.0, the electron transfer from 2Fe-2S to cyt c 1 would reach equilibrium with cyt c 1 50% reduced and the amplitude of the cyt c 1 reduction would be expected to be 50% of the initial photooxidation. It is quite possible that the ISP is lost from a fraction of cyt bc 1 molecules during purification, accounting for the smaller amplitude of the observed cyt c 1 reduction. The similar reduction amplitudes in the presence and absence of famoxadone are consistent with the finding that famoxadone binding has only a small effect on the redox potential of 2Fe-2S, increasing it by 26 mV. The addition of stigmatellin completely eliminated the reduction of cyt c 1 , indicating that stigmatellin displaced famoxadone from the Q o site and locked reduced ISP strongly in the b conformation (Fig.  2C). This indicates that stigmatellin has a much higher affinity for the Q o site than famoxadone. The addition of 30 M azoxystrobin to bovine cytochrome bc 1 resulted in a single phase of cyt c 1 reduction with a rate constant of 3,400 Ϯ 600 s Ϫ1 (Fig.  2D). This result is consistent with x-ray crystallographic studies showing that azoxystrobin binding immobilizes the ISP in the b conformation. 2 The addition of 30 M stigmatellin to bovine cyt bc 1 treated with 30 M azoxystrobin led to the elimination of cyt c 1 reduction, indicating that stigmatellin displaced azoxystrobin from the Q o site.
To study reverse electron transfer from cyt c 1 to 2Fe-2S, Ru 2 D was used to photoreduce cyt c 1 in oxidized bovine cyt bc 1 according to Scheme 3 (Fig. 3). Electron transfer from excited state Ru(II*) to cyt c 1 was complete in 1 s, the lifetime of the excited state of Ru 2 D. The sacrificial electron donors aniline and 3CP were present in the solution to reduce Ru(III). The photoreduction of cyt c 1 was followed by biphasic oxidation with rate constants of 90,000 Ϯ 15,000 s Ϫ1 and 7,300 Ϯ 1,200 s Ϫ1 and relative amplitudes of 57 and 43%, respectively (Fig. 3A). The difference in kinetics compared with forward electron transfer is apparently the result of the initial redox states of the enzyme. X-ray diffraction studies have indicated that a smaller fraction of ISP is in the b conformation in the fully oxidized complex than in the complex with cyt c 1 and that 2Fe-2S is reduced (37). It is reasonable to assign the fast phase to the mobile conformation of the ISP and the slow phase to the conformation initially in the b state (10). With this assumption, the fast phase would be gated by fluctuations in conformation between the mobile state and the c 1 state, whereas the slow phase would be gated by the conformational change from the b state to the mobile state. The addition of famoxadone resulted in a single phase of cyt c 1 oxidation with a rate constant of 1,420 Ϯ 200 s Ϫ1 (Fig. 3B). The amplitude of the single phase in the presence of famoxadone was nearly the same as the sum of the two phases in the absence of inhibitor. This result is consistent with the finding that ISP is nearly all in the b conformation in the presence of famoxadone (23). The rate constant k d for the change in conformation from the b state to the mobile state must therefore be much smaller than the rate constant k f from the mobile state to the b state in the presence of famoxadone. The rate constant k d is 1420 s Ϫ1 in the presence of famoxadone, whereas the rate constant k f could be 90,000 s Ϫ1 or even higher. It is interesting that the rate constants for forward and reverse electron transfer are the same in the presence of famoxadone. This indicates that the rate constant k d does not depend on the redox state of 2Fe-2S, consistent with the finding that famoxadone binding changes the redox potential of 2Fe-2S by only 26 mV. The temperature dependence of k d gives activation parameters of ⌬H ‡ ϭ 19.1 Ϯ 1.8 kJ/mol and ⌬S ‡ ϭ Ϫ121 Ϯ 6 J/mol⅐K for the conformational change from the b state to the mobile state in the presence of famoxadone (Fig. 4).
The present studies indicate that famoxadone binding significantly decreases the rate constant for the release of the ISP from the b conformation in cyt bc 1 . X-ray crystallography studies have revealed that famoxadone binding leads to significant conformational changes in three domains on the surface of cyt b: residues 163-171 that contact the neck region of the ISP; residues 262-268 in the ef loop that are part of the ISP-docking crater; and residues 252-256 in the middle of the ef loop that connects the Q o site to the other two surface domains (Fig. 1,  top) (23). The ef loop may play a role in relaying famoxadoneinduced conformational changes in the Q o pocket to the ISP crater and the neck contact domain to decrease the rate of release of the ISP. The linkage between the conformation of the Q o site and the dynamics of the ISP movement could be a key to how the enzyme promotes the transfer of the first electron from Q-H 2 to 2Fe-2S but inhibits the transfer of the second electron from semiquinone to 2Fe-2S. The potential role of the ef loop in regulating the dynamics of ISP domain movement during the catalytic cycle is particularly intriguing.
Effects of Famoxadone on Electron Transfer between 2Fe-2S and Cyt c 1 in R. sphaeroides Cyt bc 1 -Electron transfer in the forward direction was studied using Ru 2 D to photooxidize cyt c 1 in R. sphaeroides cyt bc 1 according to Scheme 1 in the presence of the sacrificial electron acceptor [Co(NH 3 ) 5 Cl] 2ϩ . The reduction of photooxidized cyt c 1 was biphasic with rate constants of k 1 ϭ 80,000 Ϯ 15,000 s Ϫ1 and k 2 ϭ 1,500 Ϯ 300 s Ϫ1 (Fig. 5A). k 1 is attributed to electron transfer from 2Fe-2S to cyt c 1 , whereas k 2 is due to subsequent electron transfer from Q-H 2 to 2Fe-2S and cyt c 1 followed by electron transfer from the semiquinone to cyt b L and cyt b H (Scheme 2) (10). Previous studies of the effects of temperature, pH, and redox potential demonstrated that k 1 is not rate-limited by true electron transfer in the c 1 state but rather is gated by conformational changes from the b state and the mobile state to the active c 1 state (36). According to this analysis, the population of the c 1 state is small but the rate constant for electron transfer from 2Fe-2S to cyt c 1 in the c 1 state is much larger than the observed value of k 1 , 80,000 s Ϫ1 . The addition of famoxadone leads to monophasic reduction of cyt c 1 with a rate constant of 4,800 Ϯ 800 s Ϫ1 (Fig.  a Enzymatic activity is expressed as micromole of cyt c-reduced/min/nmol cyt b at 25°C. The error limits are Ϯ 15%. b ⌬E m is the difference in redox potential between 2Fe-2S and cyt c 1 at pH 8.0, 25°C. The error limits are Ϯ 5 mV. ND, not determined. c k 1 is the rate constant for electron transfer from 2Fe-2S to cyt c 1 measured from the 552-nm transient at pH 9.0, 25°C. The error limits are Ϯ 20%. d k 1 is the rate constant for electron transfer from 2Fe-2S to cyt c 1 in the presence of 30 M famoxadone at pH 9.0, 25°C. The error limits are Ϯ 20%. 5B). There is no reduction of cyt b H in the presence of famoxadone, consistent with displacement of Q-H 2 from the Q o site. The most reasonable explanation of these results is that famoxadone binding stabilizes the b state of the ISP and that the observed rate constant k 1 is gated by the rate constant k d for the release of the ISP from the b state to the mobile state and the c 1 state. The rate constant k f for the conformational change from the mobile state to the b state is expected to be much larger than k d in the presence of famoxadone and could be 80,000 s Ϫ1 or larger. Electron transfer was also studied in the reverse direction using Ru 2 D to photoreduce cyt c 1 in the oxidized complex according to Scheme 3 in the presence of the sacrificial electron donors aniline and 3CP. Oxidation of photoreduced cyt c 1 was biphasic with rate constants of 84,000 Ϯ 15,000 s and 4,800 Ϯ 800 s Ϫ1 and relative amplitudes of 60 and 40%, respectively. The addition of famoxadone led to monophasic oxidation of cyt c 1 with a rate constant of 6,800 Ϯ 1,200 s Ϫ1 . It appears that the Rieske protein is largely in the b conformation in the presence of famoxadone and the rate of release to the c 1 conformation is decreased. The fact that the rate constants for forward and reverse electron transfer are nearly the same in the presence of famoxadone indicates that the rate constant k d does not depend on the redox state of 2Fe-2S.
Effect of Mutations in the Iron-Sulfur Protein and cyt b on Electron Transfer in Cytochrome bc 1 -Mutations in R. sphaeroides cyt bc 1 were generated by site-directed mutagenesis to characterize the interaction of the ISP with cyt c 1 and with cyt b (Fig. 1). Flash photolysis studies were carried out to determine the effects of these mutations on the rate constants k 1 for electron transfer from 2Fe-2S to cyt c 1 and k 2 for electron transfer from Q-H 2 to 2Fe-2S as described above (Table II). The effects of famoxadone on the kinetics of selected mutants was also examined (Table III). The ISP mutant P166C is completely inactive in photoinduced electron transfer, suggesting that this mutation may have led to a critical alteration in the conformation of the ISP. The P150C mutation near the 2Fe-2S center dramatically decreases the rate constant k 2 for electron transfer from Q-H 2 to 2Fe-2S down to only 4 Ϯ 1 s Ϫ1 but does not significantly affect the rate constant k 1 for electron transfer from 2Fe-2S to cyt c 1 . It appears that the structural change caused by this mutation greatly affects the interaction of the ISP with the cyt b peptide but does not affect the dynamics of the interaction of the ISP with cyt c 1 . Famoxadone binding decreased the rate constant k 1 somewhat more than for wild-type cyt bc 1 , indicating that the ISP was held more tightly in the b conformation (Table III). The K70C mutation on the ISP decreases k 2 down to 250 Ϯ 50 s Ϫ1 but does not affect k 1 , indicating a significant effect on the interaction of ISP with the cyt b peptide but no effect on the interaction with cyt c 1 . In contrast, the D143C mutation decreases k 1 down to 20,000 Ϯ 4,000 s Ϫ1 but does not affect k 2 . This finding suggests that the acidic residue Asp-143 on the surface of the ISP may be involved in the interaction with cyt c 1 . Surprisingly, famoxadone binding only decreased k 1 to 10,700 Ϯ 2,000 s Ϫ1 , indicating that this mutant is not held as tightly in the b conformation as wild-type ISP. The G153C mutation affects both rate constants, decreasing k 1 to 13,000 Ϯ 2,600 s Ϫ1 and k 2 to 450 Ϯ 80 s Ϫ1 . This residue close to the 2Fe-2S center appears to be important for the interaction with both cyt b and cyt c 1 . Famoxadone binding to the G153C mutant decreased the rate of electron transfer from 2Fe-2S to cyt c 1 to 1,500 Ϯ 300 s Ϫ1 , indicating that this mutant is held more tightly in the b conformation than wild-type ISP. Among the cyt b mutations, only Y302C affected the photoinduced kinetics, decreasing k 1 to 12,000 Ϯ 2,000 s Ϫ1 but not greatly affecting k 2 . It appears that the ISP is held more tightly in the b conformation in this mutant than in wild-type enzyme. Earlier studies have shown that the rate constant k 1 for electron transfer from 2Fe-2S to cyt c 1 was not greatly affected by ISP mutations S154A and Y156W, which decrease the redox potential of 2Fe-2S significantly (36). These results provided evidence that k 1 is not rate-limited by true electron transfer in the c 1 state but is gated by conformational changes from the b state and from the mobile state to the c 1 state. The effects of famoxadone binding on k 1 in the S154A and Y156W mutants were comparable with the effect on wild-type cyt bc 1 . Xiao et al. (12) have previously shown that the formation of a disulfide cross-link between ISP and cyt b in the K70C/A185C mutant led to a complete loss of steady-state activity, providing experimental evidence for the mobile shuttle mechanism of the ISP. The K70C/A185C mutant was totally inactive in photoinduced electron transfer, providing further confirmation of the mobile shuttle mechanism. Xiao et al. (38) previously prepared the P33C/G89C and N36C/G89C mutants where each has a disulfide cross-link between the tail region of the ISP and cyt b. These mutants both have good steadystate enzyme activity, providing evidence for the intertwined dimer structure of cyt bc 1 . The photoinduced electron transfer kinetics is not greatly affected by cross-linking in these mutants, providing further evidence that the tail region of the ISP is not involved in the mobile shuttle mechanism.