The [2Fe-2S] cluster E(m) as an indicator of the iron-sulfur subunit position in the ubihydroquinone oxidation site of the cytochrome bc1 complex.

Recent crystallographic and kinetic data have revealed the crucial role of the large scale domain movement of the iron-sulfur subunit [2Fe-2S] cluster domain during the ubihydroquinone oxidation reaction catalyzed by the cytochrome bc(1) complex. Previously, the electron paramagnetic resonance signature of the [2Fe-2S] cluster and its redox midpoint potential (E(m)) value have been used extensively to characterize the interactions of the [2Fe-2S] cluster with the occupants of the ubihydroquinone oxidation (Q(o)) catalytic site. In this work we analyze these interactions in various iron-sulfur subunit mutants that carry mutations in its flexible hinge region. We show that the E(m) increases of the iron-sulfur subunit [2Fe-2S] cluster induced either by these mutations or by the addition of stigmatellin do not act synergistically. Moreover, the E(m) increases disappear in the presence of class I inhibitors like myxothiazol. Because various inhibitors are known to affect the location of the iron-sulfur subunit cluster domain, the measured E(m) value of the [2Fe-2S] cluster therefore reflects its equilibrium position in the Q(o) site. We also demonstrate the existence in this site of a location where the E(m) of the cluster is increased by about 150 mV and discuss its possible implications in term of Q(o) site catalysis and energetics.

The ubihydroquinone:cytochrome (cyt) 1 c oxidoreductase, or bc 1 complex, is a key component of both respiratory and photosynthetic electron transfer chains (1)(2)(3). Using three redoxactive subunits (cyt b, cyt c 1 , and the iron-sulfur protein) and two active sites (Q o and ubiquinone reduction sites), it catalyzes the transfer of electrons from ubihydroquinone (QH 2 ) to a c-type cyt according to a mechanism known as the modified ubiquinone (Q) cycle (4,5). This electron transfer is coupled to a proton transport across the membrane and contributes to the generation of the electrochemical gradient subsequently used for ATP synthesis via the ATP synthase. The key of the bc 1 complex energetics relies on the bifurcation of electrons at the Q o site. Upon QH 2 oxidation at this catalytic site one electron is transferred to a high potential chain constituted of a [2Fe-2S] cluster carried by the iron-sulfur subunit and a c-type cyt borne by the cyt c 1 subunit. The other electron is transferred to a low potential chain constituted of two b-type hemes (high potential b-type heme and low potential b-type heme (b L )) and then to a Q or a semiubiquinone radical at the ubiquinone reduction site, all carried by the cyt b subunit.
Over the years, several hypotheses including the double occupancy model (6 -8), the proton-gated charge-transfer mechanism (9,10), the formation of a stable intermediate between the semiubiquinone at the Q o site and the reduced iron-sulfur subunit until the second electron is transferred to heme b L (11), the rolling over of this semiubiquinone from a [2Fe-2S] proximal to a heme b L proximal position during QH 2 oxidation (12)(13)(14), and the redox exchange between the two monomers of a dimeric enzyme complex (15,16) have been put forward to rationalize why the electrons emanating from QH 2 oxidation follow two thermodynamically different pathways. These models are based on kinetic data, analyses of the rate-limiting steps, energetics considerations, electron paramagnetic resonance (EPR) spectroscopy, and more recently crystallographic data. Indeed, during the past 4 years, several structures for mitochondrial bc 1 complexes obtained in the presence or absence of various inhibitors have been solved and revealed different conformations of the iron-sulfur subunit cluster domain in the enzyme complex (17)(18)(19)(20)(21). Similar data have also been obtained using electron microscopy with plastohydroquinone plastocyanin oxidoreductase, a homologue of the bc 1 complex in chloroplasts (22). These findings led collectively to the idea that the iron-sulfur subunit may move during QH 2 catalysis (17)(18)(19). More recently, this hypothesis has been confirmed by biochemical genetics (23)(24)(25)(26)(27)(28) and kinetic analyses (23,29) of various mutants located in the flexible region of the iron-sulfur subunit (hinge) linking its cluster domain and membrane anchor.
Although we now know that oxidation of QH 2 at the Q o site is followed by an apparently concerted transfer of the two electrons, the movement of the iron-sulfur subunit, and the release of two protons, little is known about the details of these reactions. In particular, the order of the events of electron transfer, deprotonation, and proton release and how the movement is controlled and coordinated in respect to other redox events remain unknown. Thus, information concerning the position of the cluster domain in the Q o site and its interactions with the Q o site occupants or with different subdomains of the Q o site is important. In recent years, the EPR signature of the reduced [2Fe-2S] cluster and its redox midpoint potential (E m ) value and the orientation of its g tensors have been used extensively for this purpose (6, 7, 30 -35). More recently, this approach was further complemented by the use of ␥-ray irra-diation at low temperature (36) to define the position of the oxidized [2Fe-2S] cluster of the iron-sulfur subunit. In this work we show that the E m value of the [2Fe-2S] cluster reflects the equilibrium position of the iron-sulfur subunit cluster domain in the Q o site. We also demonstrate the existence in this site of a location where the E m of this cluster is greatly increased and discuss the possible implications of these findings in terms of the energetics of Q o site catalysis.

EXPERIMENTAL PROCEDURES
Rhodobacter capsulatus strains used in this work were grown under semiaerobic conditions at 35°C in the dark in enriched medium supplemented with 2 mM MgCl 2 , 2 mM CaCl 2 , and 10 g/ml kanamycin (6), and their characterizations are described in detail in Refs. 23 and 24.
All biochemical and biophysical techniques were also described in Refs. 23 and 24 or in Ref. 37 in the case of the proteolytic assay using thermolysin. Oxidative titrations of the [2Fe-2S] cluster in chromatophore membranes were conducted potentiometrically according to Dutton (38), in the presence of the redox mediators listed in Ref. 24. The EPR titrations of the Q pool were performed as in Ref. 6 by monitoring the amplitude of the g x peak and in the presence of 50 M tetrachlorohydroquinone, 2,3,5,6-tetramethyl-1,4-phenylenediamine, 1,2-naphtoquinone-4-sulfonic acid, 1,2-naphtoquinone, N-methyl-dibenzopyrazine methyl sulfate, N-ethyl-dibenzopyrazine ethyl sulfate, duroquinone, and 2-hydroxy-1,4-naphtoquinone. Potassium hexachloroiridate was used instead of potassium ferricyanide to increase the ambient potential (E h ) above 500 mV. All chemicals were as described earlier (39) except that potassium hexachloroiridate was from Strem Chemicals and UHDBT was from B. L. Trumpower (Darmouth Medical School, Hanover, NH). Tridecylstigmatellin and MOA-stilbene were generous gifts from T. E. Wiggins (ICI Agrochemicals, Bracknell, UK), and P. R. Rich (University College, London, UK).

Mutations Located More than 20 Å Away from the [2Fe-2S]
Cluster of the Iron-Sulfur Subunit Drastically Affect Its E m Value-Previously, we had isolated several mutations in the R. capsulatus iron-sulfur subunit hinge region (between the residues 43 and 49) to investigate its role in the large scale domain movement of this subunit during Q o site catalysis (23,24). Detailed analyses of mutant bc 1 complexes indicated that indeed many of these mutations perturbed at varying degrees the Q o site events. In particular, several of these mutations, such as the ϩ1Ala and ϩ2Ala insertions (insertion of one or two alanine residues between Ala 46 and Met 47 ) (23), Delta3 deletion (deletion of the amino acid residues 45, 46, and 47) (24), and the V44L or A46T single substitution (40) or the 6Pro multiple substitution (replacement of the amino acid residues 44 -49 by six Pro residues) (24), led to the striking observation that the redox midpoint potential at pH 7 (E m7 ) values of their [2Fe-2S] cluster were increased by 30 -140 mV above that of the wild type [2Fe-2S] cluster ( Fig. 1A and Table I). Because of the large distance (over 20 Å) between the location of these mutations and the [2Fe-2S] cluster of the iron-sulfur subunit, these effects were considered unlikely to be direct. Rather, they could be mediated by a change in the position of the cluster domain in the Q o site leading to a change of the environment of this cofactor. However, for the ϩ2Ala and 6Pro (Fig. 1B), the V44L and A46T (40), the deletion Delta3, and the insertion ϩ1Ala (not shown) mutants, the EPR signature and in particular the g x peak of the [2Fe-2S] cluster remain unchanged even though the shape and position of this resonance is known to be very sensitive to the changes in the interactions of this metal cluster with its environment (7,30,33,40).

Class I Inhibitors Like Myxothiazol Decrease the High E m of the [2Fe-2S] Cluster in Various
Mutants-EPR titrations were next performed in the presence of myxothiazol, which is a representative of class I Q o site inhibitors (41). Earlier studies have shown that this inhibitor does not drastically change the E m value of the [2Fe-2S] cluster of the iron-sulfur subunit in the native bc 1 complex (30,42). However, in the case of the mutants described above, we have found that myxothiazol had a drastic effect and decreased the E m7 value of the 6Pro mutant by about 150 mV. It also brought down the E m7 values of the ϩ1Ala and ϩ2Ala insertion mutants to the range of 290 -320 mV ( Fig. 2A and Table I). A slight decrease in the E m7 value of the wild type bc 1 complex could also be observed, but this change was certainly within the reliability (approximately Ϯ 25 mV) of these measurements using chromatophore membranes. Remarkably, E m7 values ranging from 270 to 320 mV for the iron-sulfur subunit of the wild type bc 1 complex from R. capsulatus can be found in the literature (43)(44)(45). On the other hand, in contrast to the unmodified EPR signals seen with all of the mutants in the absence of inhibitor (i.e. in the presence of native Q), alterations of the g x peak shape could be readily detected in the presence of myxothiazol. This effect was clearly seen with the ϩ2Ala and 6Pro mutants (Fig. 2B), where the position of the g x resonance peak became narrower and shifted downfield.
The effect of MOA-stilbene, which is another class I inhibitor, on the E m value of the ϩ1Ala mutant [2Fe-2S] cluster was also tested and found to be similar to that seen with myxothiazol (not shown). Indeed, the E m7 value of 370 mV observed in the absence of inhibitor in this mutant decreased to about 315 mV in the presence of MOA-stilbene. This finding is in agreement with the earlier report that this inhibitor decreases the E m value of the [2Fe-2S] cluster by about 30 mV in the case of the native enzyme (45).
Class I Inhibitors Like Myxothiazol Displace the Iron-Sulfur Subunit Away from the Position That It Occupies in the Presence of Stigmatellin-Crystallographic data indicated that in FIG. 1. Potentiometric dark titrations and EPR spectra of the [2Fe-2S] cluster of various iron-sulfur subunit hinge mutants in the absence of inhibitor. A, the data points represent the normalized amplitudes of the EPR g y signal of the [2Fe-2S] cluster of the wild type (q), ϩ1Ala (ϫ), ϩ2Ala (‚), and 6Pro (OE) mutants recorded between 150 and 500 mV. In each case, the data were fitted to a n ϭ 1 Nernst equation to deduce the E m7 values indicated in Table I, and the curve fits are shown as black lines. B, the EPR spectra of the wild type (WT), ϩ2Ala, and 6Pro mutants are shown. The vertical line indicates the wild type g x value, which is to be compared with those observed with the mutants. the presence of a class I inhibitor, less electron density attributable to the [2Fe-2S] cluster of the iron-sulfur subunit was located in the b position, suggesting that the cluster domain moved toward a more "released" position (20,46). To probe whether in our mutants the iron-sulfur subunit is also located in a more "released" position in the presence of myxothiazol, we used a proteolysis assay that we had developed previously (37). The protease thermolysin cleaves the iron-sulfur subunit within its hinge domain to release an 18-kDa truncated form comprised of its extrinsic domain. This cleavage is inhibited when the native bc 1 complex is treated with stigmatellin or other inhibitors that lock the extrinsic domain of the ironsulfur protein in the stigmatellin position or in the 6Pro or ϩ2Ala mutants, which block it in a stigmatellin-like position even in the absence of these inhibitors (23)(24). Whereas in the native enzyme myxothiazol had a very limited effect on this proteolysis (37), it drastically increased the cleavage of the ϩ2Ala and the 6Pro mutants ( Fig. 3 and not shown). This finding therefore suggested that upon binding of myxothiazol, the extrinsic domain of the iron-sulfur subunit was displaced away from its initial stigmatellin-like position.
The  (30). Considering that this E m value is also high in the ϩ2Ala and 6Pro mutants, the effect of stigmatellin on these mutants was examined. The addition of stigmatellin increased the E m7 of the mutants and wild type  Table I). Only in the case of the 6Pro mutant this value reached 550 mV but was short of the 650 mV that would have been expected if the E m7 increase induced by the binding of stigmatellin was additive with that caused by the ϩ2Ala or 6Pro mutations. Yet, the EPR spectra obtained in the presence of this inhibitor revealed that all of the mutants responded properly to it as indicated by their g x signal ( Fig. 4B and not shown).
The effects of class II inhibitors weaker than stigmatellin such as tridecyl-stigmatellin and UHDBT were also tested with the 6Pro mutant. In the case of the wild type R. capsulatus bc 1 complex the E m7 increase induced by tridecyl-stigmatellin was higher than that seen with the bovine enzyme (47). This value was similar to that of the 6Pro mutant in the absence of any inhibitor. In the case of the 6Pro mutant, the [2Fe-2S] cluster E m7 only increased by about 50 mV. With UHDBT (48), the E m7 value of the wild type [2Fe-2S] cluster reached 350 -360 mV, whereas that of the 6Pro mutant remained unchanged (Table  I). Like with stigmatellin, the g x bands of the [2Fe-2S] cluster of the 6Pro mutant in the presence of tridecyl-stigmatellin or UHDBT were also similar to those observed with the native bc 1 complex, confirming the expected interactions between these inhibitors and the metal cluster (not shown).
The   ) refer to the potentiometric dark titrations experiments done at pH 7 in the presence of the natural quinone, 300 M myxothiazol, 100 M stigmatellin, 50 M E-␤-methoxyacrylate-stilbene (3-methoxy-2 (2-styryl phenyl) propenic acidmethylester), 100 M tridecyl-stigmatellin, 80 M 5-n-undecyl-6-hydroxy-4,7-dioxobenzothiazole or in the presence of the natural quinone but at pH 9, respectively. The E m values in mV were obtained by fitting the amplitude of the g y signal to a n ϭ 1 Nernst equation, as illustrated in Figs.  1, 2, and 4. b QH 2 and QH 2 (pH 9) refer to the potentiometric dark titrations experiments done at pH 7 or 9, respectively, in the presence of the natural ubihydroquinone at an E h between Ϫ100 and 200 mV. The E m values in mV were obtained by fitting the amplitude of the g x signal to a n ϭ 2 Nernst equation (Fig. 5) and are attributed to the midpoint potential of the Q/QH 2 couple at the Q o site (and not to that of the [2Fe-2S] cluster). c ND, not determined. For the strains used see the text.

FIG. 2. Potentiometric dark titrations and EPR spectra of the [2Fe-2S] cluster of various iron-sulfur subunit hinge mutants in the presence of myxothiazol.
A, the data points represent the normalized amplitudes of the EPR g y signal of the [2Fe-2S] cluster of the wild type (q), ϩ1Ala (ϫ), ϩ2Ala (‚), and 6Pro (OE) mutants recorded between 150 and 400 mV in the presence of 300 M myxothiazol. In each case, the data were fitted to a n ϭ 1 Nernst equation to deduce the E m7 values indicated in Table I, and the curve fits are shown as black lines. B, the EPR spectra of the wild type (WT), ϩ2Ala, and 6Pro mutants are shown. The vertical line indicates the wild type g x value, which is to be compared with those observed with the mutants.
the E m of this interaction follows the same pH dependence as the Q pool in chromatophore membranes (49,50). Thus, to further probe whether the interactions between the [2Fe-2S] cluster and the Q/QH 2 occupants of the Q o site had been modified in mutants exhibiting increased E m values for the [2Fe-2S] cluster, their response to the Q pool redox state were determined. In the case of the 6Pro mutant its g x ϭ 1.800 signal also titrated with an E m7 value around 90 mV. However, at low E h values, the g x signal was shifted slightly downfield and became narrower compared with the wild type ( Fig. 5 and Table I). Moreover, when a similar titration was performed with the ϩ1Ala mutant at pH 9 instead of pH 7, this mutant exhibited a pH-dependent E m change comparable with that observed with a native bc 1 complex (E m9 around Ϫ30 to Ϫ40 mV) ( Table  I). The overall findings therefore indicated that the mutations in the hinge region of the iron-sulfur subunit that increase significantly the E m values of the [2Fe-2S] cluster did not drastically change the relative affinity for QH 2 /Q at the Q o site, at least in respect to the reduced form of this cluster.

When the Iron-Sulfur Subunit Occupies a Specific Subdomain of the Q o Site, the E m of its [2Fe-2S] Increases
Drastically-In contrast to the mutants that modify the E m value of the [2Fe-2S] cluster by either affecting the hydrogen bond network around it (51-53) or changing the amino acid residues in its vicinity (40,44,54), the mutations in the hinge domain are located too far away (more than 20 Å) to have a direct effect. Thus, the high E m values observed in the mutant enzymes ( Fig.  1 and Table I) are probably due to a change in the environment of the [2Fe-2S] cluster mediated by a change in the position of the iron-sulfur subunit cluster domain, including a possibly stronger interaction between the cluster and Q/QH 2 at a subdomain of the Q o site. Multiple components such as the nature of the surrounding amino acid residues, the solvent accessibility, or the occupants of the Q o site define the properties of the immediate environment around the cluster, which in turn affects its E m . Small structural changes have already been observed between the different structures obtained by Iwata et al. (19), where the iron-sulfur subunit cluster domain occupies different locations in the Q o site. However, higher resolution structures are needed to precisely define these changes and to deduce the basis for the increase of the [2Fe-2S] cluster E m values.
As suggested by the proteolysis experiments (23,24,37), the alanine insertion mutants in general and the 6Pro mutant in particular are locked in a position similar to that adopted by the wild type iron-sulfur subunit in the presence of stigmatellin. The observation that the E m of the [2Fe-2S] cluster could be increased by either the presence of stigmatellin or the mutations bringing the cluster domain to a specific location in the Q o site led us to probe any correlation between the E m of the cluster and its location. The finding that in the mutants the effect of stigmatellin is not synergistic with that of the mutations indicated that in the native enzyme the E m increase is not solely a direct effect of the inhibitor per se but is also mediated by a shift in the equilibrium position of the iron-sulfur subunit cluster domain in the Q o site. In the presence of tridecylstigmatellin, a slight E m difference possibly reflecting a direct effect of the inhibitor was also seen between the native enzyme and the 6Pro mutant. This effect is expected considering that stigmatellin binds preferentially to the reduced protonated form of the cluster. Finally, it could be thought that the effect of stigmatellin in our mutants is incomplete because of increased resistance to stigmatellin, as seen with similar yeast mutants (28). However, the concentration of inhibitors used in our experiments was in large excess, and in the EPR spectra no g x ϭ 1.800 signal that could reflect any residual interaction with Q at the Q o site was detected.
In R. capsulatus, the effect of stigmatellin is less pronounced than that in the bovine enzyme (⌬E m ϭ 200 mV versus 250 mV) (30). The polar head group of stigmatellin probably interacts similarly in both cases with a histidine ligand of the [2Fe-2S] cluster (18), but the inhibitor may lock the iron-sulfur cluster domain in a slightly different environment in various species, leading to the observed differences in the E m values. The tridecyl-stigmatellin has the same polar head group as stigmatellin but a less hydrophobic tail, resulting in a weaker binding to the cyt b subunit. Consequently, it cannot lock as tightly as stigmatellin does the iron-sulfur cluster domain in the subdomain of the Q o site responsible for the increased stabilization of the reduced form. The other weak inhibitor UHDBT also increased the E m7 in the case of the wild type but not that of the 6Pro mutant. Because the EPR g x signal revealed a normal interaction with UHDBT in the case of the 6Pro mutant and because the E m of its [2Fe-2S] cluster did not decrease, apparently UHDBT does not move the cluster domain to a different location. The lower E m7 shift induced by UHDBT rather indicates that its ability to lock the native cluster domain is less efficient than the 6Pro substitution in the hinge domain. These findings are consistent with the binding of class II inhibitors per se not being the only factor responsible for the increased E m . Moreover, if the high E m in the 6Pro mutant was mediated only by a favored interaction between the cluster and Q and the protein environment had no influence at all, then one would have to conclude that UHDBT has a similar stabilizing effect on the reduced [2Fe-2S] cluster as Q does in this mutant. However, this effect is much more pronounced with the native enzyme. Finally, if the increased E m was mainly a consequence of a more favored interaction with Q and considering that this cannot be the case for the reduced cluster with QH 2 , one would expect in the mutants a significant change in the E m associated with the Q pool redox state, which is not observed (Fig. 5).
The Decrease of E m in the Presence of Class I Inhibitors Also Reflects a Change of the Position of the Iron-Sulfur Subunit at the Q o Site-Myxothiazol and MOA-stilbene lowered the E m value of the [2Fe-2S] cluster of the iron-sulfur subunit in the mutants analyzed. Although the effect of myxothiazol on the E m of the native enzyme was rather small, that on the 6Pro mutant was very significant (⌬E m of about 150 mV). Once again, direct interactions of the cluster with the inhibitor could not account solely for such a large decrease, indicating that changes in its environment must be important. The change in the location of the cluster domain out of the stigmatellin subdomain also correlated with an increased sensitivity to cleavage by thermolysin (Fig. 3). Movement of the cluster domain toward a more "released" position in the presence of myxothiazol is also supported by the crystallographic data of Kim et al. (20).
The The increased E m in the presence of inhibitors like stigmatellin or UHDBT therefore results at least partially from the binding of the inhibitor limiting the free diffusion of the cluster domain in the Q o site. Similarly, the ϩ2Ala, 6Pro, or Delta3 mutants in the hinge region increase the E m by displacing the equilibrium position of the cluster domain in the Q o site toward this high E m subdomain, and the more the cluster domain is trapped, the higher is the E m . In addition, to what extent the presence of Q at the Q o site also affects this equilibrium in various mutants remains to be determined.
Clearly, the presence in the Q o site of class I inhibitor like myxothiazol or MOA-stilbene excludes the cluster domain from the high E m subdomain, probably by inducing some conformational changes like the rotation of the side chain of glutamic acid 272 of cyt b as discussed in Ref. 14. The effect on the E m is barely detectable in the case of the native enzyme, unlike the ϩ2Ala, 6Pro, or even Delta3 mutants. Upon addition of a class I inhibitor, at least in the case of the ϩ2Ala and 6Pro mutants, the iron-sulfur cluster domain is released from the stigmatellin-like position, even though it cannot diffuse to cyt c 1 position as indicated by the absence of electron transfer from an initially reduced [2Fe-2S] cluster to cyt c 1 heme (23)-24). In these mutants the cluster domain is probably still constrained in the Q o site because of a steric hindrance at the level of the ef loop that needs to be crossed during the movement toward cyt c 1 (Fig. 6). Finally, it is noteworthy that in the ϩ2Ala and 6Pro FIG. 4. Potentiometric dark titrations and EPR spectra of the [2Fe-2S] cluster of various iron-sulfur subunit hinge mutants in the presence of stigmatellin. A, the data points represent the normalized amplitudes of the EPR g y signal of the [2Fe-2S] cluster of the wild type (q), ϩ1Ala (ϫ), ϩ2Ala (‚), and 6Pro (OE) mutants recorded between 300 and 600 mV in the presence of 100 M stigmatellin. In each case, the data were fitted to a n ϭ 1 Nernst equation to deduce the E m7 values indicated in Table I, and the curve fits are shown as black lines. B, the EPR spectra of the wild type (WT), ϩ2Ala, and 6Pro mutants are shown. The vertical line indicates the wild type g x value, which is to be compared with those observed with the mutants.

FIG. 5. Potentiometric dark titration of the Q pool and EPR spectra of the [2Fe-2S] cluster with the Q pool reduced in the case of the 6Pro mutant.
A, the data points represent the normalized amplitudes of the EPR g x signal of the [2Fe-2S] cluster of the wild type (q) and 6Pro (OE) mutant recorded between Ϫ100 and 200 mV. In each case, the data were fitted to a n ϭ 2 Nernst equation to deduce the E m7 values indicated in Table I, and the curve fits are shown as black lines. B, the EPR spectra of the wild type (WT) and 6Pro mutants obtained at an E h value of around 0 mV are shown. The vertical line indicates the wild type g x value, which is to be compared with those observed with the 6Pro mutant. mutants the EPR g x signal in the presence of Q or stigmatellin is native, but in the presence of myxothiazol this signal is shifted up-fields and becomes more narrow, indicating that the cluster/Q interactions are changed. Although the physical parameters influencing the shape and g value of the EPR g x signal are not well understood, nonetheless this observation still indicates the existence of several positions for the [2Fe-2S] cluster in the Q o site of the cyt bc 1 complex.
Energetic and Mechanistic Consequences-Mechanistic details of QH 2 oxidation reactions at the Q o site, including binding of the substrate QH 2 , liganding of QH 2 to the [2Fe-2S] cluster at the distal (i.e. stigmatellin binding), subdomain of the Q o site, deprotonation of the [2Fe-2S] cluster histidine ligand (His 161 in bovine numbering) to form probably a Hbonded complex with the bound QH 2 , transfer of two protons and two electrons (one to the [2Fe-2S] cluster and one to heme b L ), and release of the newly formed Q, are currently not well defined. During the past 20 years, different steps of these reactions have been considered to be rate-limiting (9 -10, 55, 56). A large number of factors, including the E h (5,55,57), the pH (9 -10, 53, 56), the pK a of the iron-sulfur subunit (attributed to the His ligand of the cluster) (53), the E m of the [2Fe-2S] cluster (40,44,(51)(52)(53), the temperature (53), and the viscosity (58) of the medium, are known to affect the rate of QH 2 oxidation. As to the activation energy, depending on the activity measured (steady-state or single turnover) and mutant strains used to address this question, it appears to be dependent only on the pH, the pK a of the iron-sulfur subunit, and the E m of the [2Fe-2S] cluster. However, in this case it is extremely difficult to change only one parameter at a time, and several key steps might have similar rates and contribute to the observed activation energy.
As an example, one hypothesis is that the rate-limiting state is associated with the formation of the semiquinone radical. However, studies of mutants affecting the E m of the [2Fe-2S] cluster (51)(52)(53) show that this change affects the activation energy less than expected from a simplistic "Arrhenius" approach. The results obtained in this work further indicate that the problem is even more complex. Indeed the E m of the [2Fe-2S] cluster varies in function of its position in the Q o site, and yet this E m is a key component in the energy calculations. Even if the overall ⌬G for the reaction is unaffected by this finding, its implications on the distribution of the energy between the different steps of the mechanism need to be considered. An increase in the E m should help in the formation of the semiquinone at the stigmatellin subdomain, and a decrease in the E m when the cluster move toward cyt c 1 should favor electron transfer to this cofactor. This would make of the iron-sulfur subunit in the bc 1 complex the first enzyme that modulates its [2Fe-2S] cluster E m by a large scale movement of this cofactor, but how this is accomplished remains to be defined. In the case of the nitrite reductase, a large scale domain movement to exchange the heme c ligand of this enzyme and modify its potential during the catalysis has been described (59). If the interactions of the [2Fe-2S] cluster with Q/QH 2 at the Q o site are responsible for the increased E m when the protein is already in the right position, like in the 6Pro mutant, then this finding points out that in the native enzyme an energy-requiring step would be to bring the head domain into this specific position. Another point that needs to be considered in detail is the role of the movement in various other steps. Indeed, even if in the native enzyme the movement is not rate-limiting, it can become so by increasing the viscosity (58) by mutating the hinge region (23)(24)(25)(26) or the interface with the cyt b subunit. 2 Therefore, ultimate care is needed when associating the in-  (18), b (the coordinates of this structure have been provided by S. Iwata as a privileged communication), intermediate (19), myxothiazol (the coordinates of this structure have been provided by E. A. Berry as a privileged communication), and c 1 (19) positions are shown as space filling models with the sulfur atoms in yellow and the iron atoms in red, orange, cyan, green, and blue, respectively. B and C represent schematically the multiple positions that the iron-sulfur subunit cluster domain can occupy in the cyt bc 1 complex of the wild type and the ϩ2Ala or 6Pro mutants, respectively. Only the hemes, inhibitors, [2Fe-2S] clusters, and amino acid residues 253-270 of the ef loop are kept from A, and the iron-sulfur subunits in various positions are represented as blue ovals. Note that the equilibrium positions of the iron-sulfur cluster domain have changed in the mutants or in the presence of the inhibitor myxothiazol (Myx) or stigmatellin (Stig), with the intensity of its blue color reflecting various levels of occupancy of the different locations. crease in the activation energy with the lower E m in the ironsulfur subunit cluster domain mutants (51)(52)(53).
The finding that the Q o site of the bc 1 complex has multiple subdomains that can change the properties of the [2Fe-2S] cluster of the iron-sulfur subunit also raises additional questions. In the case of the ϩ2Ala mutant, we have found that one round of QH 2 oxidation can occur provided that the [2Fe-2S] cluster is initially oxidized (23,60), even though the iron-sulfur subunit cluster domain apparently cannot leave the Q o site. Thus, does QH 2 oxidation take place with the iron-sulfur subunit locked at the stigmatellin subdomain, or does this subunit still have enough freedom to move between the multiple positions in the Q o site? Moreover, calculations of the molecular surface for the histidine ligand (His 161 in bovine numbering) when the iron-sulfur subunit moves from cyt c 1 position to the stigmatellin position indicate a clear decrease in the water accessibility that should affect the pK a of this residue. Thus, not only the E m but also the protonation state of the iron-sulfur subunit probably changes between its multiple positions in the Q o site, and how does this affect QH 2 oxidation? Hopefully, future studies of the ϩ2Ala or 6Pro mutants will bring insightful answers to these issues, which are critical for our understanding of Q o site catalysis.