review proton pumping in the bc complex: a new gating ... proton pumping in the bc 1 complex: a new...

Biochimica et Biophysica Acta 1757 (2006) 1019–1034www.elsevier.com/locate/bbabio

Review

Proton pumping in the bc1 complex: A new gating mechanismthat prevents short circuits

Antony R. Crofts a,b,⁎, Sangmoon Lhee b, Stephanie B. Crofts b, Jerry Cheng b, Stuart Rose a

a Department of Biochemistry, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USAb Center for Biophysics and Computational Biology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA

Received 6 December 2005; received in revised form 3 February 2006; accepted 16 February 2006Available online 15 March 2006

Abstract

The Q-cycle mechanism of the bc1 complex explains how the electron transfer from ubihydroquinone (quinol, QH2) to cytochrome(cyt) c (or c2 in bacteria) is coupled to the pumping of protons across the membrane. The efficiency of proton pumping depends on theeffectiveness of the bifurcated reaction at the Qo-site of the complex. This directs the two electrons from QH2 down two differentpathways, one to the high potential chain for delivery to an electron acceptor, and the other across the membrane through a chaincontaining heme bL and bH to the Qi-site, to provide the vectorial charge transfer contributing to the proton gradient. In this review, wediscuss problems associated with the turnover of the bc1 complex that center around rates calculated for the normal forward and reversereactions, and for bypass (or short-circuit) reactions. Based on rate constants given by distances between redox centers in knownstructures, these appeared to preclude conventional electron transfer mechanisms involving an intermediate semiquinone (SQ) in the Qo-sitereaction. However, previous research has strongly suggested that SQ is the reductant for O2 in generation of superoxide at the Qo-site,introducing an apparent paradox. A simple gating mechanism, in which an intermediate SQ mobile in the volume of the Qo-site is a necessarycomponent, can readily account for the observed data through a coulombic interaction that prevents SQ anion from close approach to heme bL whenthe latter is reduced. This allows rapid and reversible QH2 oxidation, but prevents rapid bypass reactions. The mechanism is quite natural, and is wellsupported by experiments in which the role of a key residue, Glu-295, which facilitates proton transfer from the site through a rotational displacement,has been tested by mutation.© 2006 Elsevier B.V. All rights reserved.

Keywords: Q-cycle; bc1 complex; Proton circuit; Rhodobacter sphaeroides; Qo-site; Superoxide; –PEWY–span

Abbreviations: bc1 complex, ubiquinol:cytochrome c oxidoreductase (EC1.10.2.2); bL and bH, low- and high-potential hemes of cytochrome b,respectively; cyt, cytochrome; Em,(pH), midpoint redox potential at pH indicated(pH 7 assumed if not indicated); ESEEM, electron spin echo envelopemodulation spectroscopy; ISP, Rieske iron–sulfur protein; ISPH, reduced ISP;ISPox, oxidized, dissociated ISP; MOA-inhibitors, inhibitors with E-β-methoxyacrylate or similar group as pharmacophore; P-phase, N-phase, aqueousphases in which the sign of the transmembrane proton gradient is positive ornegative, respectively; PDB, Protein Data Bank; Q, oxidized form ofubiquinone; QH2, reduced form (hydroquinone, quinol) of ubiquinone; QHU,QU−, protonated and deprotonated forms of semiquinone; Qi-site (Qo-site),quinone reducing (quinol oxidizing) site of bc1 complex; Rb., Rhodobacter; RC,photosynthetic reaction center; SQ, semiquinone (with protonation stateunspecified); UHDBT, 5-undecyl-6-hydroxy-4,7-dioxobenzothiazol; UHNQ,2-undecyl-3-hydroxy-1,4-naphthoquinone⁎ Corresponding author. Department of Biochemistry, 419 Roger Adams Lab,

600 S. Mathews Ave., Urbana, IL 61801, USA. Tel.: +1 217 333 2043; fax: +1217 244 6615.

E-mail address: [email protected] (A.R. Crofts).

0005-2728/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.bbabio.2006.02.009

1. Introduction

The bc1 complex (Complex III of the mitochondrial respi-ratory chain) operates through a Q-cycle mechanism (reviewedin [1–4]). The key feature of this mechanism is the reaction at theQo-site in which ubihydroquinone (quinol, QH2) is oxidized.The two electrons from QH2 are passed to two different electrontransport chains — the first to a high potential chain of theRieske iron sulfur protein (ISP), cyt c1, cyt c (or c2 in bacteria),and a terminal oxidant. The second electron is transferred to alow potential chain of hemes bL, bH and the Qi-site, and crossesthe membrane, to contribute to both the electrical and chemicalwork of the proton gradient. It has generally been consi-dered that the reaction proceeds through separate 1-electronsteps, with generation of an intermediate SQ. However, asdiscussed further below, the viability of schemes involving

mailto:[email protected]

http://dx.doi.org/10.1016/j.bbabio.2006.02.009

Fig. 1. The modified Q-cycle. Schematic model showing the redox centers(hemes bH and bL in cyt b, heme c1 in cyt c1, and the 2Fe–2S cluster in ISP), theelectron transfer pathways connecting them (solid blue arrows), the binding andunbinding of substrates and products (QH2, Q (filled broad arrows) at the Qo-and Qi-sites, and cyt c at the interface on cyt c1),and the release and uptake ofprotons (solid red arrows). The membrane is indicated by purple dashed lines,the cyt b subunit by a dashed cyan box, the ISP subunit by a pale-blue dashedcircle, the cyt c1 subunit by a dashed pink circle, and cyt c by a dashed red circle.The Qo- and Qi-sites are indicated by the semiquinone intermediates. Inhibitorsdiscussed in the text are indicated by numbers at arrows pointing to the affectedsite as follows: at the Qo-site, 1, 2, 3,-stigmatellin, alkyl-HDBTs, and UHNQ,which H-bond to ISPH; 4, 5, 6, myxothiazol, other MOA-type inhibitors, andfamoxadone, which H-bond to the backbone NH of the –PEWY– glutamate; atthe Qi-site, 7, 8, antimycin and NQNO.

1020 A.R. Crofts et al. / Biochimica et Biophysica Acta 1757 (2006) 1019–1034

significant occupancy of the SQ state has recently beenchallenged.

Since the bifurcation is essential for proton pumping, aperennial question has been how the site operates to prevent bothelectrons passing down the more favored pathway, to the highpotential chain, to decouple this process. A second questionrelates to a design failure of the bc1 complex. When electronsback-up in the low potential chain, the Qo-site functions toreduce O2 to superoxide, the precursor of reactive oxygen spe-cies that lead to DNA and protein damage and cause cellularaging and death [5]. How does the mechanism minimize thisdeleterious reaction and other potential bypass reactions?

In this brief review, we will first provide an overview of theQ-cycle and its operation as a proton-pumping machinery, andaddress recent advances in our understanding of the critical Qo-site reaction. This will entail a brief survey of the kinetic, spec-troscopic, structural, and theoretical tools available, and the useof these to set constraints on viable mechanisms, especially withrespect to transfer of the first electron, which seems to be the ratedetermining process. We will then discuss the properties of thesecond electron transfer, the role of the SQ intermediate, and thedifferent strategies that have been suggested to explain how thebypass reactions are minimized. Finally, we will discuss a me-chanism incorporating coulombic gating that accounts for thepeculiarities of the Qo-site reaction in the context of a modelpreviously proposed [6].

2. The modified Q-cycle

The bifurcation of QH2 oxidation, the reaction at the heart ofthe Q-cycle, was first recognized by Wikstron and Berden [7] asan explanation for the phenomenon of “oxidant-induced reduc-tion of cytochrome b”[8]. It was Mitchell's genius to incorporatethis reaction into the proton-motive Q-cycle [9,10] by directingone electron across the membrane to reduce ubiquinone(quinone, Q) on the other side. He named the two quinone-processing sites Qo and Qi, to differentiate whether the H

+ cameout on oxidation of QH2, or in on reduction of Q. Garland [11]recognized that a “stand-alone” enzyme would need to delivertwo electrons to the Qi-site, requiring two turnovers of the Qo-site for net oxidation of 1 QH2, and called this a modified Q-cycle. Mitchell [10] discussed many variants of the Q-cycle, andlaid down some of the thermodynamic constraints, but the expe-rimental evidence available did not allow discrimination bet-ween these until the early 1980s. Detailed kinetic studies inchromatophores (the re-sealed invaginations of the inner bac-terial membranes of photosynthetic bacteria, produced by me-chanical disruption) showed that the modified Q-cycle couldaccount for the paradoxical behavior observed in this system[12,13] (reviewed in [4]). The version shown in Fig. 1 is aconsensus Q-cycle adopted by most workers in the field, and issupported by an extensive set of kinetic and thermodynamic datafor the partial processes [1,14]. The mechanism is also nowstrongly supported by structures [15,16] (reviewed in [17]), allof which show the redox centers, and their organization withrespect to membrane topology, predicted from these earlierexperiments.

Proton pumping is achieved indirectly by movement of twonegative charges across the membrane, carried by the electronspassing through the b-heme chain from Qo to Qi sites, and byrelease or uptake of H+, respectively, on oxidation or reductionof quinone, to give an overall yield of 2H+ pumped for eachQH2 oxidized:

QH2 þ 2cyt cþ þ 2HþN ⇆ Qþ 2cyt cþ 2Hþ

P þ 2HþP ðscalarÞ

In chromatophores, the electrogenic processes can be fol-lowed through the electrochromic carotenoid changes, whichprovide a “membrane voltmeter” function [18], and this con-venience has allowed a detailed matching of electrogenic eventsto partial processes [12,14,19,20]. By careful correction of ab-sorbance changes in the cytochrome α-band region for con-tributions from carotenoid changes [21], this approach wasextended to measurement of driving forces from electron trans-fer during development of the proton gradient in the coupledsteady-state [22–24]. Under static head conditions, the poiseof the electron transfer chain was close to that expected fromthe modified Q-cycle in equilibrium with the proton gradient,

Fig. 2. Model of the ES-complex, showing QH2 with H-bonds to His-161 of ISPand Glu-272 of cyt b. Model based in the structure containing stigmatellin, inwhich ubihydroquinone has been modeled in place of the inhibitor. The electrontransfer distance is shown as ∼7 Å (the equivalent bond in the stigmatellincomplex from PDB 1pp9 [41] is 6.91 Å).

1021A.R. Crofts et al. / Biochimica et Biophysica Acta 1757 (2006) 1019–1034

indicating a tight control. These results were in line with resultsfrom mitochondrial studies [25] in which the proton gradientwas varied through poising of the ATPase reaction. In thechromatophore experiments, the development of both electrontransfer poise and proton gradient could be resolved kinetically.

In chromatophores, where superior kinetic resolution andredox poising make it possible to separate out the partial pro-cesses contributing to the Q-cycle, activation energies can bemeasured or estimated for each of these within well-definedconstraints [26]. From such experiments, it was established thatthe reaction at the Qo-site is rate limiting under conditions ofsubstrate saturation, and that it is responsible for the highactivation energy of the overall reaction. However, this reactionis itself complex, and under non-saturating conditions, rates arealso in part determined by reactions involved in formation of theES-complex, and dissociation of products. A full understandingof mechanistic detail therefore requires analysis of the partialprocesses contributing to the Qo-site reaction [1,13,27], and theirstructural context [26].

3. Movement of the ISP extrinsic domain—a dancer in theQo-site ballet

A surprise from the structures was the large-scale movementof the extrinsic domain of the ISP [16], which is now recognizedas essential for delivery of electrons from the Qo-site to cyt c1[28–31]. The role of this movement in control of the Qo-sitereaction has been the topic of much speculation. It seems clearfrom molecular dynamics simulation [34] and from kineticmeasurements [14,35] that the movement itself is stochastic, andso rapid as not to be rate limiting for the overall forward reaction[32,33]. The critical issues for discussion therefore center on theinvolvement of the ISP mobile domain in formation of transientcomplexes with its reaction partners—with cyt b and theoccupants of the Qo-site (inhibitors, Q and QH2), or with cyt c1and its heme [6,36–38]. Structures showing the reduced ISP(ISPH) forming H-bonded complexes in situ with several Qo-siteinhibitors (stigmatellin, NQNO, HHDBT and UHDBT) occu-pying the domain of the Qo-site distal from heme bL (the distaldomain), have been solved crystallographically [15,16,38–41].The existence of such complexes was known from earlier bio-physical evidence, and EPR spectroscopy [42,43]. In addition, ithad been early speculated that the EPR line at gx=1.80 reflectedan interaction of ISPH with quinone (Q) bound at the Qo-site,and that QH2 also likely might bind in close association with ISPin formation of the ES-complex [44]. Of these complexes, thereactions involved in formation of the ES-complex at the Qo-siteare the most important in understanding mechanism [36,45–47],but because these occur only under the metastable condition inwhich oxidized ISP and ubihydroquinone are both present, theseare also the least accessible to direct observation and ther-modynamic characterization.

4. Configuration of the ES-complex

Because the ISP extrinsic domain is mobile, it acts as adiffusible second substrate (albeit, a tethered one) in the reaction

at the b-interface, so that the binding of two substrates, ISPoxand QH2, is needed for formation of the ES-complex. Since anelectron is transferred from QH2 to ISPox, the ES-complex isexpected to form with QH2 close to the ISP docking interface.The PDB file 1ntz has coordinates (with high B-factors) for aquinone occupant, but the experimental basis for these has notbeen discussed by the authors [48]. None of the other structurescurrently available shows any quinone species bound at the Qo-site. Modeling of quinone or quinol occupancy has thereforebeen based on occupancy of inhibitors. Because the rate-limit-ing reaction involves reduction of the oxidized ISP (ISPox),requiring a relatively short electron transfer path, the most ob-vious choice has been the stigmatellin structure, which shows adirect H-bond between Nε of His-161 of the reduced ISP and acarbonyl group of the occupant. A quinol modeled with H-bonds to the same ligands as stigmatellin can replace the inhi-bitor in the structure without strain, and fits within the electrondensity of the inhibitor [41,49]; the quinone species modeledin 1ntz is in a similar configuration. Models of this sort havebeen the starting point for most discussions of the ES-complex(Fig. 2) [36,49,50].

If a H-bond similar to that seen with stigmatellin, but from thequinol –OH to His-161 of ISPox, is involved in stabilization ofthe ES-complex, it would likely be to the imidazolate form[51,52]. The relative pK values for quinol and ISPox would favoran H-bond with the Nε of His-161 of ISPox in the dissociatedform as H-bond acceptor, and the quinol –OH as donor. Thisconfiguration was suggested on the basis of the thermodynamicproperties, but required the assumption that the pK on theoxidized form at 7.6 observed through redox titration of ISP wasassociated with the histidine involved in H-bonding [51]. Re-cently, Iwaki et al. [53] have shown that well-characterizedbands in the FTIR difference spectrum represent “the loss of theimidazolate form of one or both of the Nδ-ligated imidazolehistidine ligands to the [2Fe–2S] center”, which are “in theimidazolate form in the oxidized protein above pH 7.6, and …


(receive)… a redox-linked proton when the ISP becomesreduced”. This is the first direct confirmation of our earlierinterpretation.

The suggestion of such a H-bonded configuration opened thepossibility of an investigation of the binding energy involvedthrough the magnitude of thermodynamic effects assayed kine-tically through formation of the ES-complex [46,47]. Sinceformation of the ES-complex would, according to the abovehypothesis, pull both QH2 and ISPox out of solution, measure-ment of these effects should provide two of the values needed tocomplete the thermodynamic square describing formation of theES-complex (Fig. 3). The two effects assayed were thedisplacement of the Em value of the Q/QH2 couple involved information of the ES-complex [54–56], measured from theapparent Em for QH2 oxidation, and the pH dependence of QH2

oxidation [57,58], interpreted as a displacement of the pK ofISPox [46,47]. These had been reported in an extensive literature,but had not previously been linked. Values reported for theapparent Em of QH2 involved in formation of the ES-complexare ∼40 mV more positive than that measured for the quinonepool, while the apparent pK of ISPox assayed kinetically is ∼1.2pH units lower than the value measured in redox titrations.Within the range of measured values, these represent similarfree-energy contributions. A strong correlation between thechanges in apparent pK, measured from the pH dependence ofelectron transfer, and changes in the measured pK of ISPox inmutant strains, has provided firm support for the above hypo-thesis [59]. Taken together with the FTIR evidence [53], the datastrongly support the suggestion that the dissociated form of His-161 of ISPox is involved as H-bond acceptor in the ES-complexwith QH2.

Binding of QH2 in the absence of ISPox would involve otherbinding forces, including that due to the H-bond in which thesecond –OH of quinol participates. Using the binding of stig-matellin as a model, this –OHwould mimic the second ligand ofstigmatellin to Glu-272 of cyt b (a member of the highly con-served –PEWY– loop, numbered as in the chicken sequence),with the carboxylate group as acceptor. The free-energy contri-bution to binding could be estimated from the change in apparentaffinity of the QH2 substrate in mutant strains in which theglutamate was changed to other residues, assayed through thedependence of rate on [QH2]. This effect was quite small,suggesting an approximately 2-fold change in Km [49]. A

Fig. 3. Binding square involved in formation of the ES-complex. Formation ofthe ES-complex requires binding of two substrates, QH2 and ISPox. See text andlegend to Table 2 for explanation.

binding constant for ISPox in the absence of quinol could beestimated from crystallographic data [33]. The relatively weakoccupancy of ISP, as estimated from the electron densities rela-tive to the rest of the protein in structures of the native complexfrom different groups, was used to suggest that the affinity wasrelatively weak for either the b-position, or the c-position [33].These estimates provide constraints on the fourth value neededfor the thermodynamic binding square. The four values providethe same overall ΔGo′, in the range ∼−8 kJ/mol (or 83 mV inelectrical units) from the sum of separate ΔGo′ terms, forformation of the ES-complex by either pathway from vacant site(left) to ES complex (right) of the thermodynamic square (seeFig. 3).

5. The EP-complex

The EP-complex in which quinone interacts with ISPH hasbeen studied through several different approaches. As notedabove, the line at gx=1.80 in X-band CW-EPR depends on thepresence of both Q and ISPH. With the recognition fromENDOR and crystallographic data that the histidine ligands tothe [2Fe–2S] cluster were exposed [60,61], several groups notedthe possibility, based on the involvement of histidines in ligationof Q in bacterial reaction centers, that one of the histidines mightalso serve as a ligand to Q in the Qo-site [62–64]. From moredetailed ESEEM studies, the hyperfine interactions of the Nδ of ahistidine (likely H-161 in bovine numbering) were shown to besimilar in the gx=1.80 and stigmatellin complexes, and differentfrom those seen when the complex was dissociated on additionof myxothiazol [65]. Addition of myxothiazol or MOA-typeinhibitors, eliminated this band, presumably by displacing Q,and also led to a change in Em of ∼−30 mVof ISP [45,66–69],suggested to reflect the loss of the H-bond with quinone.Interestingly, when similar experiments were performed usingmutant strains in which the linking sequence through which theextrinsic head is attached to the anchoring N-terminal tail wasmodified [37], the myxothiazol-induced change in Emwas muchgreater, suggesting that additional forces had come into play[69]. The simplest interpretation was a change in strength of theH-bond arising from the closer proximity allowed by thelengthened link [45,68], but orientation-selective EPR measure-ments have suggested additional complications [67].

6. Forward electron transfer — the oxidation of QH2

Much discussion of the efficiency of the bifurcated reactionhas centered on “gating mechanisms” to prevent passage of bothelectrons to the high potential chain. Two natural gating pro-cesses are intrinsic to all Q-cycle mechanisms.

(i) The formation of the ES-complex will occur only underthe metastable conditions in which the QH2 substrate isavailable, and the high potential chain is oxidized. Nor-mally, this non-equilibrium state is generated only underconditions appropriate for net forward electron transfer.

(ii) Under normal operation (“state-3” conditions in mitochon-dria), the quinone pool is maintained at least partly


oxidized. This is necessarily the case in chromatophoressince otherwise, the photochemistry would fail; similarly,the dehydrogenases would be turned off if no Q wasavailable in respiratory chains. As a consequence, anacceptor is always available at the Qi-site, so that the lowpotential chain is partly oxidized, and hence can rapidlyremove the intermediate SQ product. Even under statichead (“state 4”) conditions, when the backpressure fromΔp is maximal, heme bL is about 50% oxidized [22–24,72,82].It is not widely appreciated that, because net forwardelectron transfer can occur only when both conditions (ISPand heme bL both oxidized) are satisfied, and net reverseelectron transfer when both ISP and bL are reduced, theseconstraints represent an intrinsic double-gating in bothdirections built into the Q-cycle. Under normal forwardconditions, the efficiency of bifurcation is high, andbypass or short-circuit reactions [1,70–72,74] are negligi-ble. Under static head conditions, heme bL becomes partlyreduced because of “back-pressure” from the protongradient, and a physiologically important bypass reactionoccurs, leading, in an aerobic system, to reduction of O2 toform superoxide [72,82,83]. Bypass reactions of this sortare also intrinsic to all Q-cycle mechanisms that involveformation of a SQ intermediate. However, the bypass rateis much greater when electron transfer to the Qi-site isblocked, and these inhibited conditions are normally usedto study the reactions. It is estimated that the rate ofsuperoxide generation in antimycin-inhibited mitochondriais ∼80% of the total cellular production of reactive oxygenspecies [70–72,82], and such rates might apply in mutantstrains, or myopathies with mutations in the Qi-site. Thedebate over how the complex has evolved to minimizethese bypass reactions can be framed in terms of thefollowing questions.

(iii) Are the rates of bypass reactions explicable in terms of theintrinsic electron transfer mechanisms, or

(iv) are there are additional special gating mechanisms thatinvolve, for example, allosteric control, conformationalcoupling, etc.

In order to explore these questions, it is necessary to establisha minimal model for the mechanism of the site at the molecularlevel.

7. Kinetic constraints on the reaction mechanism at theQo-site

Distances between redox centers involved in catalysis can bemeasured from the structures. Empirical evidence and theoret-ical considerations have suggested that distances determine the“intrinsic” rate constant, ko, in the Arrhenius–Randall–Wilkinstreatment, kcat=koe

−ΔG#/RT. When expressed in log10 form, andexpanded to include a Marcus term for the activation energy,and a Moser–Dutton [73] treatment of ko, this gives:

log10kcat ¼ 13� 0:6ðR� 3:6Þ � gðDGo þ kÞ2=k ð1Þ

where R is the electron transfer distance (the closest distancebetween conjugate rings for cytochromes, chlorophylls, etc.), λis the reorganization energy. The factor γ is 4.23 (F/(4×2.303 RT) at 298 K in the classical Marcus treatment, or3.1 in the Moser–Dutton equation. This latter value reflects thecontribution of quantum mechanical corrections that wereapplied to generate fits to experimental data. This useful equa-tion can be applied to discussion of the mechanism of the Qo-sitein the context of the structure, since it provides constraints de-termined by distances. For the Qo-site reaction, applicationdepends on the answers to three questions:

(i) What is the structure from which the reaction proceeds?(ii) What is the nature of the reaction?(iii) If the reaction is sequential, which of the two 1-electron

transfer steps is rate determining?

In the discussion of rates that follows, we will use theMoser–Dutton equation (with γ=3.1), which gives increasingly higherrate constants than the classical treatment as λ and –ΔGo

diverge. However, in most cases, plausible rate constants can becalculated using either approach. In the discussion of bypassrates below, the calculated SQ occupancy would vary dependingon choice of γ. Since subsequent calculations are “normalized”to this occupancy, relative rates and occupancies will for themost part show the same pattern, and lead to similar conclusions,with either choice. Where anomalies occur they will be pointedout.

From the considerations in the section above, the location ofthe ES-complex with QH2 in the distal domain seems firmlybased, and this will provide the starting point for the remainingdiscussion. In our view, and not withstanding recent suggestionsto the contrary [74,75], the evidence in favor of a sequentialelectron transfer with an intermediate semiquinone product isalso overwhelming. Perhaps, most compelling is the demon-stration that the overall rate is determined by the driving force forthe first electron transfer, and is relatively independent of thedriving force for the second electron transfer. The experimentalbasis for the first conclusion comes from several different groupswho have reported mutations of ISP in which the redox potentialof the [2Fe–2S] cluster has been modified [26,52,59,76,77]. Thedriving force is determined by the redox difference:

DGo V¼ �zFðEmðISPÞ � EmðSQ=QH2ÞÞ:

It seems unlikely that mutations in the ISP cluster-bindingdomain would affect the Em of the semiquinone couple. IfEm(SQ/QH2) is unaffected, changes in Em(ISP) would be directlyreflected in changes in the driving force. After normalization tothe wild-type rate, the results from separate studies crossing thebacterial/mitochondrial divide, and different genera of bacteria,have shown the same behavior, all data falling on the samecurve of log10ko v. ΔGo′, which can be well fitted by a classicalMarcus curve [47]. In contrast, no such relationship has beenreported for mutations changing Em(heme bL), although mutantshave been reported from several labs (cf. [26,36,78]). Similarly,the ∼60 mV change in Em of heme bL on reduction of heme bH


does not appear to have a significant effect on the rate oractivation energy.

Hong et al. [26] suggested that determination of kcat fromdistances in the structures could be used to put constraints onreactions at the Qo-site, based on the simple kinetic relationship:

v ¼ kcat½occupancy� ð2ÞThey evaluated a number of different mechanisms, and

concluded that, in the context of an ES-complex formed at thedistal end of the pocket, viable schemes could be limited to threetypes:

(i) The intermediate SQ product was constrained to the distaldomain, for example by a relatively stable ISPH.SQintermediate product [41,79]. In this case, a relatively highoccupancy was required for overall forward chemistry atthe observed rate because of the limitation on the rateconstant for the second electron transfer arising from thelong distance to the acceptor, heme bL.

(ii) The SQ product was liberated by dissociation of the ISPHafter the first electron transfer, and could move within thevolume of the Qo-site [36,49]. Hong et al. [26] favoredthis mechanism because of the need to minimize harmfulshort-circuit reactions by keeping [SQ] occupancy to aminimum. Because the rate constant is higher at shorterdistances to the acceptor heme bL, the observed rate couldbe attained at much lower SQ occupancy.

(iii) Double occupancy of the Qo-site by two quinones [44,80],arranged so as to bridge the distance between ISP andheme bL, would also provide a mechanism allowing ahigher rate constant for transfer of the second electron, andhence allow a lower occupancy. However, as discussed atlength elsewhere, the failure to find in the kinetic behavioror the structures the properties expected was seen asproblematical [26,36].

Among mechanisms rejected were all those in which the twoelectron transfers occurred from a common activated state, sinceonly one of the two steps, the first electron transfer, showedcontrol by driving force.

8. The first electron transfer from QH2 to ISPox

Identification of the first electron transfer as the limitingprocess raised the question of why this reaction was so slow. Asshown in Fig. 2, the electron transfer distance is <7.0 Å, which,with conventional parameters for driving force and reorganiza-tion energy (λ) in the Moser–Dutton equation, would beconsistent with a rate in the sub μs range rather than the 1.3×103

s−1 observed. The “theoretical” rate could be “slowed” byassuming both a strongly endergonic first electron transfer, and alarge value for λ, consistent with the high activation barrier [26].However, the reaction interface provided no obvious explana-tion for the high activation barrier. From the nature of theproposed ES-complex, and the redox properties of the reactionpartners, it was apparent that a proton transfer must be coupled to

the electron transfer reaction, and this insight leads to a moresatisfactory explanation. The pK values for QH2 and ISPoxdetermine the probability for proton transfer along the H-bondjoining them, through a Brønsted term, ΔGo′=2.303 RT(pKdonor−pKacceptor). Use of the measured pK values, togetherwith the overall free-energy, allowed construction of athermodynamic cycle from which the probability of differentpathways could be assessed. The likely sequence in the reactionis the pathway involving the transfer of a proton first, then theelectron; the electron transfer occurs from a proton configurationthat is kinetically favorable, but thermodynamically improbable[46,81]. The proton transfer step is strongly uphill (ΔGo′=21–27 kJ/mol), and contributes about half the activation barrier. Theremaining barrier is that for the electron transfer itself. This has adriving force closer to the isopotential range, and can occur withconventional reorganization energy and rate constant. The ob-served rate is slow because transfer occurs from a very lowoccupancy of the kinetically favorable proton configuration[47].

Rates of proton transfer along H-bonds can be extremelyrapid (in the range 1011 to 1013 s−1). Even allowing for thestrongly unfavorable equilibrium constant, rate constants inboth directions will be much more rapid than the overallelectron transfer rate, so the contribution of the protonbarrier can be treated using a simple kinetic expression,kapp=kETKproton. Expansion of the Moser–Dutton equation toinclude this term gives the following [47].

log10k ¼ 13� b2:303

R� 3:6ð Þ � gðDGo

ET þ kETÞ2kET

� pKQH2� pKISPox

� �:

9. Role of the pK of ISPox in determining rate

An interesting consequence of this treatment is that the pK ofISPox has two roles to play in determining the rate of reaction.One is through the Brønsted term in the above equation. Anadditional contribution, as noted above, arises from the role ofthe dissociated form of ISPox as the substrate involved information of the ES-complex, and dependence of the concen-tration of this form on pH and pK. The probability of formingES is dependent on pK. These different roles can be exploredthrough use of mutant strains in which the pK is modified.Raising the pK, as in strain Y156W [47,59], with pK changedfrom 7.6 to 8.5, leads to a change in the pH dependence ofelectron transfer to the higher value expected from the role information of ES. However, the higher pK would be expected tolower the Brønsted barrier, and therefore lead to a higher rateconstant. This effect is also observed, though convoluted withthe change linked to the change in driving force arising from thechanged Em. Because the two changes have opposing effects onrate, the net result is that at any fixed pH below the pK, the rateappears to depend only on the Em value, as observed [59].Further exploration of this relationship with new mutant strainsin which Ser-154 is modified support these conclusions(SangMoon Lhee and ARC, unpublished).


10. The second electron transfer

The second electron transfer involves oxidation of the SQ byheme bL. Examination of the available structures showedseveral features that provided clues to the mechanism [36,49].

(i) The volume of the site was larger than necessary to acco-mmodate the QH2 substrate.

(ii) Inhibitors were bound in different but overlapping do-mains; the H-bonding of stigmatellin by ISP constrained itto the domain distal from heme bL, but myxothiazol andMOA-type inhibitors were bound more proximal to hemebL.

(iii) Glu-272 was in different positions in different structures,either pointed towards the distal domain to act as a secondligand to stigmatellin, or rotated by >130°, to point to-wards heme bL in the native structure, or those with pro-ximal domain inhibitors.

(iv) Residues changed in strains resistant to myxothiazol butnot to stigmatellin lined the myxothiazol binding domainas expected, but changes at these positions correlated withtwo additional properties of mechanistic interest; (a)electron transfer was inhibited, but (b) the binding of Q toISPH indicated by EPR signal at gx=1.80 was not. Theseeffects were interpreted as showing that in these strains,binding of quinone species (including QH2) at the distaldomain was unimpeded, but that occupancy of the pro-ximal domain was blocked, and that such an occupancyby some intermediate was necessary during turnover.

On the basis of these properties, and of changes on muta-genesis of Glu-295 (the equivalent of Glu-272 in Rb. sphae-roides), Crofts et al. suggested a mechanism in which the SQcould move in the site closer to heme bL (Fig. 4) [34,36,49].Molecular dynamics simulations had revealed a buried waterchain to which the glutamate connected in the rotated position.

Fig. 4. Configurations of the Qo-site involved in electron transfer. In A, the configuratthrough the 7 Å distance shown between QH2 and the 2Fe–2S cluster of ISP, withintermediate QHU− would be in a similar configuration, but with the reduced 2Fe–2SIn B, the mobilized SQ has passed a proton to the –PEWY– glutamate (Glu-272 in cgroup contacts a water chain leading to the P-side aqueous phase. The QU− is now closSQ suggested (see text).

The rotation of Glu-272 therefore provided a mechanism for exitof the second H+ from the site. An additional point was that, witha shorter distance to heme bL, the necessary rate could still beachieved at a lower occupancy of SQ.

As Hong et al. [26] pointed out, several critical parametersneeded to assess this mechanismwere unknown, in particular theSQ occupancy, and the distance over which electron transferoccurs. Osyczka et al. [75] introduced the idea that the occu-pancy of SQ could be estimated from the rate of short-circuitreactions, since all of them involve SQ as an intermediate. Thisapproach, and the earlier arguments of Hong et al., now make itpossible to determine important constraints on mechanism thatallow a choice between the options previously discussed [26].The problems can be summarized through the terms in Eq. (2) –occupancy, rate, and rate constant – the latter provided by thedistances through which electron transfer reactions involvingSQ might occur (Fig. 5). Starting with a SQ in the distal position(Fig. 5A), where it would be formed from an ES-complex of thesort discussed above, evolution of the state of the system under“bypass-conditions” (forward electron transfer inhibited becauseheme bL is reduced), can occur through several pathways thatlead to loss of the SQ [5,22,75]. Muller et al. [5] showed that thedominant pathway (∼70%) was by oxidation, either by O2 underaerobic conditions, or through the high potential chain (ISP,heme c1, cyt c) under anaerobic conditions. Because cyt c canrapidly oxidize the superoxide generated under aerobic condi-tions, both pathways could be assayed through the reduction ofcyt c in the presence of antimycin. Oxidation under anaerobicconditions necessarily requires the dissociation of the interme-diate product (ISPH.SQ) after the first electron transfer, becausemovement of the ISPH is necessary for reduction of heme c1 [5].The immediate oxidant for SQ is ISPox on its return fromdelivery of an electron to cyt c1. It is assumed that ISPox acceptselectrons from the same position as in the ES-complex, through adistance of ∼7.0 Å. It is likely that O2 reacts through a similardistance, since the docking site for the ISP would provide an

ion before transfer of the first electron is shown. The first electron transfer occursthe constraints discussed in the text. Immediately after the electron transfer, thecluster dissociated so as to liberate both ISPH and SQ for subsequent movement.hicken numbering), the side chain of which has rotate so that the carboxylic acide enough to heme bL to deliver an electron rapidly, even at the low occupancy of

Fig. 5. Parameters determining electron transfer rates under bypass conditions.The ubiquinone and protein configurations in panels A and B are the same as forFig. 4A and B. In panel C, the SQ is shown in the position expected from theconstraints of coulombic gating suggested in the text. The 12.4 Å distance fromSQ in the distal domain to heme bL is based on the crystalographically-determined position of UHDBT (1sqv), while the 6.3 Å distance from SQ in theproximal domain to heme bL is based on the position of myxothiazol (1sqp).These values are used for compatibility with the discussion of Osyczka et al.[74]. The notes inserted in the figure report the rate constants, occupancies andrates as follows: (a) The rate constant for oxidation of SQ by ISPox calculatedfrom the 7 Å distance and appropriate values for ΔGo and λ, and the occupancyof SQ needed to match the observed rate of ∼ 3 s− 1. (b) The rate constantcalculated using the 12.4 Å distance, and the rate of electron transfer for the 2ndelectron at the occupancy in (a). (c) The rate constant calculated using the 6.3 Ådistance, and the rate of electron transfer for the 2nd electron at the occupancy in(a). (d) The rate constant calculated using the 6.3 Å distance, and the rate ofreduction of SQ by heme bL

− at the occupancy in (a), if SQ were at the proximalposition. (e) The rate constant calculated using the 12.4 Å distance, and the rateof reduction of SQ by heme bL

− at the occupancy in (a), if SQ were restricted tothe distal position by coulombic gating.


access port when the domain was not there. These assumptionsprovide a value for the distance in the Moser–Dutton equation,allowing calculation of a plausible rate constant. From themeasured rate and Eq. (2), we should now be in a position tocalculate occupancy. Osyczka et al. [75] discuss the problem interms of the stability constant for SQ formation under equi-librium conditions,Kstab. This provides a measure of occupancy/bc1 complex through [SQ]=(0.25 Kstab)

1/2, but is otherwise notrelevant to the Qo-site mechanism. Using the ∼7.0 Å distance, arate for bypass of ∼0.3 s−1 measured from the decay of reducedheme bL, a driving force for electron transfer from SQ to ISPox of

−0.5 V, and a conventional value for λ ∼1 V, they calculated anoccupancy in the range 5×10−9 SQ/bc1 complex. The rate ofelectron transfer from SQ in the distal domain to heme bL at thisoccupancy, which would determine the net rate of forwardelectron transfer, would be ∼0.25 s−1, clearly inconsistent withthe measured forward rate of 1.3×103 s−1 [26]. The rate ofbypass measured by Osyczka et al. (∼0.3 s−1) was 10-foldslower than rates commonly reported in the literature [71]. Thelatter are measured under “oxidant-induced reduction” condi-tions, with the two b-hemes completely reduced, which might beexpected to maximize SQ occupancy. In estimation of theoperational SQ occupancy under conditions of normal forwardelectron transfer, we have to keep in mind the possibility ofgating effects from the heme. If we ignore such effects, the SQoccupancy would likely be lower than that calculated frommaximal bypass rates because the rate constant for SQ removal ishigher than for formation. However, using a bypass rate of3.0 s−1 [5], in the range commonly reported, giving an occu-pancy ∼4×10−8, the calculated rate for transfer of the secondelectron would still be 3 orders of magnitude slower than theobserved forward rate (Fig. 5A a, b). In order to achieve themeasured rate, an occupancy in the range >10−5 SQ/(bc1complex) would be needed. However, and here's the problem,this would be expected to allow rates for bypass reactions similarto those for forward chemistry, and would lead to massive in-efficiency in proton pumping. We can endorse the conclusionreached by Osyczka et al. “Thus, there is no single value of Kstab

that will permit rapid sub-millisecond forward and reverseelectron tunneling and simultaneously limit short-circuit tun-neling to the timescale of seconds” [75], and recognize the needfor some gating mechanism.

The suggestion that the SQ might move in the site [26,36,49]provides some escape from this paradox. Using the SQ occu-pancy of 4×10−8 calculated from the observed bypass rate, thechange in rate constant for the electron transfer from SQ to hemebL obtained by changing the distance from 12.4 to 6.3 Å, allowsfor a >1000-fold increase in rate, to a value greater than theobserved rate (Fig. 5B(c)). However, we are now forced toconsider possible consequences of allowing the SQ close toheme bL, and the potential for short-circuit by other bypassreactions, such as those involving direct reduction of SQ byreduced heme bL [1,5,22,74]. Given the Em values of heme bL(−90 mV) and the SQ/QH2 couple (likely >400 mV), thereaction would be strongly exergonic, and would occur over ashort distance, leading, at the occupancy proposed, to a calcu-lated bypass rate similar to the forward rate (Fig. 5(d)). Similarscenarios can be demonstrated for other positions because nomatter what particular position we choose, rates for forward,back, and bypass reactions are in the same range, leading toprediction of crippled proton pumping efficiencies.

11. Mechanisms that prevent accumulation of SQ

In view of these difficulties, it is not surprising that Osyczkaet al. [74,75] would suggest that mechanisms allowingaccumulation of SQ should be discarded. However, this maybe a case of throwing the (SQ) baby out with the bath-water— a


drastic step that ignores the very strong evidence that bypassreactions do involve SQ generated at the Qo-site [5,71,72,83].Before this approach can become acceptable, plausible hypo-theses to account for the anomalies in terms of mechanisms thatprevent SQ accumulation will be needed. The recent literaturehas spawned a plethora of options, including concerted mecha-nisms [27,84–86], and gated mechanisms involving indirecteffectors operating on several different players [58,79,87–90].Osyczka et al. [74] rejected all current schemes, and consideredtwo possibilities of their own; (a) a concerted process in whichboth electron transfers occurred from a common intermediate atthe activation barrier, and (b) a double-gated mechanism inwhich SQ was generated only under circumstances in which itwas immediately consumed. However, both these proposals areproblematical, as discussed more extensively below.

11.1. Concerted mechanisms

Hong et al. [26] had previously considered the evidenceagainst some concerted mechanisms, and the points they raisedare still valid. These and other difficulties can be summarized asfollows:

(i) In a concerted reaction of the sort proposed, both electrontransfers would have to occur at the same rate. If bothelectron transfers occur from a common activated state,then both would have to have the same intrinsic rateconstant, ko, and the same control by driving force. Fromthe lack of control in the second electron transfer, this isclearly not the case.

(ii) Similarly, the distance for transfer of both electrons totheir separate acceptors would have to be about the same,with similar structure intervening between the QH2 donorand the two different acceptors. Neither the location forthe reaction complex roughly midway between ISPox andheme bL, nor the structural similarity, has any justificationfrom crystallographic data.

(iii) Separation of two electron charges from QH2 in a phase ofrelatively low dielectric would entail obvious physicalproblems. It is therefore necessary to address the questionof the co-transfer of protons required by the overall re-action equation, and the synchronization of these pro-cesses with the electron transfers. No consideration wasgiven to these aspects.

(iv) The exponential term in the Arrhenius equation providesan estimate of the occupancy of the activated state. Underconditions of substrate saturation, kcat=koexp(–ΔG#/RT)≈ ko[ES

#]. This is because, in the standard treatment,–ΔG#/RT= lnK#, and K# = [ES#]/[ES], so that [ES#]becomes equal to the fractional occupancy of the activatedstate if [ES#]<< [Etot]. For an activation energy of∼600 mV, occupancy would be ∼10−10. Applying Eq.(2) to this occupancy shows that a rate constant∼1013 s−1

is required to give the observed rate. This is the valueexpected when the reactants are at van der Waals contact.No plausible configuration of the Qo-site would allow bothelectron transfers to occur from such a distance.

(v) It is clear in any scenario that if the activated state isgenerated in the distal domain, electron transfer to hemebL could not occur at a realistic rate at this occupancy[26].

(vi) Recognizing that the constraint suggested by this sim-plistic treatment might be too harsh, there are still prob-lems even with the most permissive scenario. For areaction complex midway between ISPox and heme bL, therate constant calculated using the Moser–Dutton equationwith λ=2.0, R=9.6 Å, andΔGo =0, is kcat ∼1.4×103 s−1.This is just adequate for concerted electron transfer froma fully occupied ES-complex in this position. However,such a scenario is difficult to justify. None of thestructural or kinetic evidence discussed above is compat-ible with such a position, and the plausibility of thecalculated rate constant depends on quantum mechanicalcorrections (implicit in the γ=3.1 of the Moser–Duttontreatment) to “lower” the barrier. With a classical Marcusbarrier, this configuration gives rate constants 100-foldtoo slow.

(vii) An earlier mechanism discussed by Crofts and Wang [27]involved a SQ intermediate at the occupancy of theactivated state, and is formally equivalent to a concertedreaction, since the second electron transfer occurred fromthis state. A similar mechanism was more recently dis-cussed by Kim et al. [90]. The slope of the Marcus curveexpected is much steeper than observed, and on this basis,mechanisms of this class were excluded [26].

11.2. Double-gated mechanisms

The discussion of how double-gated mechanisms mightfunction [74,75] has been framed in rather general terms so as toallow formation of SQ only if both chains were oxidized andQH2 was available as substrate for the forward reaction, or bothchains were reduced and Q available as substrate for the reversereaction (see above). In either case, the SQ would always berapidly removed on formation, so as to keep the occupancy low.There is no objection in principle to double-gating, since it isintrinsic to the Q-cycle mechanism. However, the mechanismsproposed by Osyczka et al. [74] were rather lacking in detail,and it was not clear what additional mechanistic componentswould be necessary to forestall bypass reactions, for instanceunder “oxidant-induced reduction” conditions. Rich [95] hasdiscussed some specific “logic-gated” mechanisms, includinggating functions linked to the redox state of the b-hemes, thatwould prevent binding of QH2 or formation of SQ [84,96], butOsyczka et al. [75] preferred the more general framework. Inneither case were particulars included that might be testedexperimentally. The structural context was weakly defined, andthe evidence discussed above for specific features (such as theconfiguration of the ES-complex with a direct H-bond from His-161 of ISP, and a direct H-bonding role for Glu-272) wasignored. Instead, participation of waters as H-bonded bridgesbetween the quinone species and unidentified ligands wasproposed [75]. No evidence for such a configuration is avail-able. The authors did not suggest experimental tests, except in

Table 1Electron transfer rates for forward and bypass reactions at the Qo-site in wildtype and in strains mutated at Glu-295 of cyt b in Rb. sphaeroides

Strain Bypass rate(e/bc1/s)

Qo-site rate(e/bc1/s)

Bypass as% rate

WT 1.35 237 0.6E295W 1.5 2.0 75E295L 2.8 7.18 39E295G 1.46 2.7 54E295Q 1.76 9.8 18E295D 1.56 14 11E295K 1.25 3.2 39


the general sense of accounting for the low rates of short-circuitreactions they measured. This argument was weakened by thefact that the rates they reported were 10-fold lower than thosemeasured elsewhere.

The types of mechanism proposed by Osyczka et al. [74,75]would be expected to prevent significant occupancy of SQ, soneither accounts for the bypass rates that are observed, or theproperties of these reactions that have led others to recognizethat a SQ is involved. A plausible scheme in either the concertedor the double-gated scenario would therefore need to be supple-mented by additional hypotheses.

12. Coulombic gating

An alternative approach is to look for explanations in thecontext of mechanisms in which SQ is a natural intermediate. Apromising starting point is the mechanism involving mobility ofthe SQ in the site [36,49]. If the SQ could be prevented fromgetting close to reduced heme bL, the troubling short-circuitreactions associated with that configuration would be avoided.The mechanism summarized in Figs. 4 and 5 shows a gatingprocess that is a natural consequence of the chemistry. It involvesthe other dancer in the Qo-site ballet – Glu-272 – previouslyproposed as a proton-carrying group involved in getting thesecond proton out of the Qo-site [36,49]. The pirouette of Glu-272 was apparent in early structures from Berry's group [16,32].In the stigmatellin structure PDB 2bcc, it was involved as aligand to the inhibitor. In contrast, in the native structure PDB1bcc, or in the presence of myxothiazol [36], the side chain wasrotate around by >130° to connect to a chain of H2O moleculesleading to the aqueous phase. This water chain was first iden-tified in amolecular dynamics simulation of ISPmovement [34],but has now been seen in higher resolution structures of bc1complex from yeast [50] and bovine mitochondria [41]. Theinitial configuration in the mechanism suggested started with theES-complex discussed above (Fig. 4A). After the first electrontransfer and dissociation of ISPH, a neutral SQ, QHU, was left atthe distal end of the site, H-bonded to the Glu-272 carboxylateside chain. After H+ transfer from QHU to glutamate, the sidechain rotated to release the H+ to the water chain, leaving a SQanion, QU−, free in the site. We assumed that with heme bLoxidized, QU− could diffuse to the proximal position, andtransfer the electron to heme bL (Fig. 4B). However, as noted in[36], if heme bL were reduced, the net change in local negativechange would have coulombic consequences—the QU− wouldbe repelled. This would have two effects, (a) the reaction leadingto formation of the anionic form (on transfer of the secondH+ outof the site) would by discouraged, and (b) coulombic repulsionwould prevented any QU− formed from getting close to heme bL

−.The SQ would then be restricted to the neutral QHU form,constrained to the distal domain (Fig. 5C)—just what is neededto prevent the rapid short-circuit reactions. These coulombiceffects therefore provide the gating necessary to minimize by-pass reactions while allowing rapid forward and reversechemistry. The coulombic gating we have proposed functionsin the context of a mechanism that requires formation of SQ,albeit at very low occupancy. This is in contrast to the suggestion

of Osyczka et al. [75], where the gating was proposed to preventformation of SQ.

13. Mutational experiments to test the role of the–PEWY– glutamate

In the mechanism proposed, Glu-272 obviously plays animportant role. We have previously reported experiments onstrains in which the equivalent residue (Glu-295) in Rb. sphae-roides was mutated to glutamine, glycine, or aspartate in strainsE295Q, E295G and E295D, respectively [36,49]. All strainsshowed substantially inhibited electron transfer through the Qo-site, resistance to stigmatellin, and a modest increase in Km forQH2. The E295Q, E295G strains (the E295D strain was nottested) also retained a normal gx=1.80 EPR signal, diagnostic ofan uninhibited reaction of Q with ISPH at the b-interface. Theseproperties were consistent with the mechanistic role for the ES-complex proposed.We have recently regenerated these and othermutations in a His-tagged background to facilitate spectroscopicstudies. We have compared the bypass reactions in these strainswith wild type under conditions designed to minimize artifacts,and we show preliminary results and example traces in Table 1 andFig. 6. Several features of the results are noteworthy, and providestrong support for the coulombic gating mechanism proposed.

(i) It is clear from Table 1 that all strains showed bypass ratessimilar to or greater than those seen in wild type.

(ii) It will also be obvious that all mutant strains showed astrongly inhibited forward electron transfer. However, justas important is the fact that none showed the completeinhibition seen with mutations at other essential residues.

Each of these features is interesting. The bypass rates showclearly that SQ was being produced in all cases, at occupanciessimilar to or greater than in wild type under bypass conditions.Yet the inhibited rates show that the SQ was unable to reduceheme bL at the wild type rate. It is not clear what extent ofinhibition the measured rates represent, because, as argued atlength by Hong et al. [26], the rates for the second electrontransfer in the absence of Qo-site inhibitor are not rate limiting. Inthe presence of antimycin, the rate of reduction of heme bL, seenwhen heme bH is pre-reduced is essentially the same as the rate ofreduction of heme bH [97], despite that fact that the driving forcediffers due to the coulombic effect, which lowers the Em of hemebL by ∼60 mV. In addition, the activation barrier for both

Fig. 6. Turnover of the Qo-site in wild type and E295Wmutant on flash activation of the bc1 complex. Kinetic traces showing the changes in cytochromes and reactioncenter on flash illumination of strains with wild type sequence (A and C) or mutation of Glu-295 to Trp (E295W) (B and D). The traces in A and B are followingillumination with a single flash, on a rapid time scale. Those in C and D are following a group of 6 flashes spaced at 10 ms. For cytochrome changes (heme bH in green,heme bL in blue, heme c1 in red), an upward deflection shows reduction. For reaction center (RC, in black), an upward deflection shows oxidation. All traces weremeasured in the presence of antimycin to inhibit oxidation of heme bH. In the presence of myxothiazol, the flux through the Qo-site was <0.3 electrons/bc1 complex/s(not shown). The dashed lines show slopes used to measure rates. The wild type used was a mutant strain (E295K) that had reverted to wild type, to control for allgenetic engineering manipulations, but the behavior was essentially the same as His-tagged wild type. Experimental protocols were as previously reported [1], andmutant strains were generated using conventional procedures [46,59]. Strain E295Wwas grown under anaerobic photosynthetic conditions in a medium supplementedwith DMSO so as to allow growth without limitation by the mutation.


reactions was the same [26]. Because no control by driving forcewas apparent from this step, the rate must be fast enough that anyslowing due to the change in driving force is not reflected in theobserved rate [26]. Under the conditions of the experiments inTable 1, the reaction was limited by the relatively oxidized stateof the quinone pool, but the intrinsic limitation was in the firstelectron transfer step, as discussed above. In wild type, with anunconstrained second electron transfer, a rate (determined byrate constant and occupancy as discussed above) in the range>10-fold higher than that for the first step seems reasonable, or>104 s−1. The observed rates in the mutant strains thereforelikely represent a much more substantial inhibition than is at firstapparent. Remarkably, the rates observed are in the rangeexpected from electron transfer from the distal domain at theoccupancy calculated from the bypass rate (see Fig. 5), asdiscussed more extensively below.

(iii) In Fig. 6, example traces of the kinetics of reaction center(RC) and cytochrome changes following flash excitationof chromatophores from a strain with wild type sequence,and in E295W, one of the more severely inhibited strains,are shown. The kinetics observed show clearly that, inboth mutant and wild type strains, the distribution ofelectrons in the high and low potential chains iscontrolled by the equilibrium constants, as previouslydemonstrated [12]. That for the Qo-site reaction controlsthe overall distribution with respect to the quinone pool.In all cases, the Qo-site reaction is observed through the“oxidant-induced reduction of cyt b” [8], which isinitiated when, with QH2 available in the pool, oxidizingequivalents from the RC, generated on flash illumination,reach the Qo-site via ISPox. In Fig. 6A and B, the stronglyinhibited rate in the mutant is apparent from the slow rate

Table 2

Reaction ΔGo/F(mV)′

kforward

First electron transferQH2+ISPox+E·bL⇆ ISPox·QH2·E·bL −80 a 1.4×105 M−1 s−1

ISPox·QH2·E·bL⇆ ISPH+·QH−·E·bL 300 b 105 s−1

ISPH+·QH−·E·bL⇆{ISPH+QH−}#·E·bL 360 c 106 s−1

{ISPH+QH−}#·E·bL⇆ ISPH+(QHU)·E·bL −300to −140 d

1011 s−1

QH2+ISPox+E·bL⇆ ISPH+(QHU)·E·bL 280 to 440d 1.3×103 s−1

Second electron transferISPHb·Qd

UH…−OOC-glu·E·bL⇆ISPHc+Q

U−·HOOC-glu·E·bL−30 e >104 s−1

ISPHc+heme c1⇆ ISPox+heme c1−+H+ 30 f 105 s−1

QU−·HOOC-glu·E·bL -{water chain}⇆QpU−·E·bL− −OOC-glu{water chain+H+}

−30e >104 s−1

QpU−·E·bL−Q+E·bL

− −280to −440d

>104 s−1

ISPH·QUH· E·bL+heme c1⇆ ISPox+heme c1

−+Q+E·bL−+2H+

−310to −470d g

>104 s−1

In all reaction equations, E stands for enzyme, and represents the Qo-site incytochrome b.A Visual Basic program, Marcus–Bronsted, which allows the user to explorehow the Marcus curves for the two electron transfer reactions at the Qo-site varywith different parameters for driving force, SQ occupancy, pK of ISPox, etc., isavailable for downloading from the URL below. Files available include theprogram in executable (.exe) form, and all source code files needed forcompilation, together with an illustrated user manual.http://www.life.uiuc.edu/crofts/Marcus_Bronsted/.a The value in electrical units is derived from apparent binding constants.

Several possible ways for normalization to concentration are discussed in [47].b From ΔGo=2.303(RT/F)(pKdonor–pKacceptor), using 11.3 for pKquinol, and

6.3 for pKISPox(bound). This latter value is appropriate for the pK of the formbound in the ES-complex [47].c Assuming ΔG#

overall=660 mV from the ES-complex [26].d The energy level of the intermediate product is uncertain. The main

constraint is the value used for rate constant, determined by the position fromwhich the SQ transfers its electron. From myxothiazol to the nearest conjugateatom of heme bL is ∼6.3 Å, and a quinone modeled to occupy this position hasa similar distance. A closer position for Q can be found (with a similar energy,after minimal side chain displacement) at a distance of 4.7 Å, giving a 10-foldfaster rate constant. The higher energy level shown here corresponds to theoccupancy of 4×10−8 discussed in the text. Choice of a value for ΔGo for thisintermediate state propagates to the other values indicated by a superscript d[26].e These values are somewhat arbitrary, but reflect relatively small equilibrium

constants, based on the difference in Km for O2 in production of superoxide forantimycin and myxothiazol inhibited yeast complex [5].f From the Em values for ISP and cyt c1 [12–14].g The sum of free energies for the first and second electron transfers is

constrained by the ΔGo for the overall reaction, given by the Em values, usingΔGo′/F=− (Em(ISP)+Em(hemebL)−2Em(Q/QH2))∼−30 mV [12].


of reduction of heme bH (green trace) observed in thepresence of antimycin following a single flash. In wildtype, the reduction is substantially complete at 40 ms butin the mutant, barely detectable. If six flashes are given tofully oxidize the reaction center and maximize thedriving force available in the high potential chain (Fig.6C and D), and the time of measurement is extended,additional features are apparent, and a more completereduction of heme bH can be observed in themutant. Notethat, in the rise kinetics for both strains, the reduction ofheme bL lags that of bH, as expected from the differenceinEm values. This lag is apparent in the wild type strain inFig. 6A, where no reduction of heme bL was seen afterthe first flash, but heme bH was rapidly reduced. Thereduction of heme bL on the second and subsequentflashes can be readily seen if the time scale in Fig. 6C isexpanded (cf. [97]). In the E295W mutant (Fig. 6D), theturnover of the Qo-site is so slow that the reaction has notyet come to equilibrium at the time of the last flash, andthe traces showing reduction of both heme bH and bLcontinue to rise, the former even out until 4 s. However,as the high potential chain relaxes, with RC and then cyt cgoing re-reduced, the controlling effect of the Qo-siteequilibrium constant can be seen, leading to the re-oxidation of heme bL before any relaxation of heme bH.Essentially, the same relaxation effects can be seen in thewild type (Fig. 6C), but they start from the higher level ofreduction of the b-heme chain achieved by the end of theflash sequence.

It is clear from these results that in both wild type andmutant, the equilibrium constants controlling the kinetic effectsare similar, but that both forward and reverse rate constants areinhibited in the mutant strain. Similar effects were observed forall mutant strains [1].

(iv) The bypass rates were measured from the flux ofelectrons into the terminal acceptor of the high potentialchain, the oxidized reaction center, immediately afterthe last flash of the sequence. At this point in the kinetictrace, heme bH and heme bL were essentially completelyreduced in the wild type; in the mutant, heme bH was∼80% oxidized and heme bL was almost completelyoxidized. Intermediate levels of reduction of heme bLwere seen in less severely inhibited strains [1]. Since theSQ was likely the source of electrons in all cases, wecan say that the redox state of heme bL determinesneither the binding of QH2, nor the formation of SQ.From this, it seems likely that neither the docking of,nor the electron transfer to, ISPox are modified by theredox state of either b-heme, and that we can excludedany specific dependence on redox state of the b-hemesin gating of these partial processes.

The behavior summarized above is readily explained interms of the hypothesis proposed here. The mutation of Glu-295 frustrates the proton exit mechanism for the secondproton. This constrains the SQ to the neutral form, QHU,

and to the distal domain, and likely also constrains Q to thedistal domain. The electron transfer rates from the distaldomain are limited by the greater distance to heme bL. Fromthe distances, we might expect the rate constants to changeby a factor of ∼ 1000, as discussed above, but both forwardand reverse rate constants would be changed by the sameratio, since the determining factor will be the samedifference in distance between the occupant in distal andproximal locations and heme bL.

http://www.life.uiuc.edu/crofts/Marcus_Bronsted/


14. Conclusions

In this review, we have discussed the evidence that supports asimple mechanism for the Qo-site of the bc1 complex,summarized in the partial reactions of Table 2, for which allnecessary thermodynamic and kinetic values are given. Themechanism accounts for the experimentally observed propertiesof the Qo-site as it functions in the bc1 complex of Rb.sphaeroides. All components of the mechanism are natural, andhave been derived from measured properties of the system. Inaddition to the reaction properties, the mechanism requirescertain structural features, which have all been based on knownstructures. The same mechanism accounts for most of the datafrom mitochondrial studies.

This simple mechanism is by no means generally accepted.Other groups have interpreted their data as showing mechanismsthat are more complex. These include interactions betweenredox groups over substantial distances through postulatedconformational changes. For many of these mechanisms, somefeatures have been shown to be untenable in the context of theRb. sphaeroides complex [26,75], and our experiments withGlu-295 mutants provide additional constraints. However,evidence for long-range interactions between Qo- and Qi-sitesin both mitochondrial and bacterial complexes has come fromstructures [39,91], and from a recent EPR investigation of theconfiguration of the ISP mobile domain at the Qo-site on inhi-bition or mutagenesis at the Qi-site [92]. Similarly, in our ownhands, mutations at the Qi-site designed to make the Rb. sphae-roides complex sensitive to funiculosin (V209A and V209A-I213L mutant strains), generated sensitivity, but at the Qo-site(Sangjin Hong and ARC, unpublished results). The mechanisticsignificance of these long-range effects is uncertain, since nokinetic control arising from such interactions has been demon-strated. On the other hand, it has been demonstrated that the rateof oxidation of QH2 on a single turnover is unaffected by bindingof antimycin [14], and that binding of antimycin has no effect onthe configuration of the Qo-site detectable in the highest reso-lution structure (2.1 Å) yet reported [93].

It is as well to recognize that the evolutionary divide betweenbacteria andmitochondria represents∼2 billion years of separateevolution, and is evidenced on the mitochondrial side by asubstantial increase in structural complexity following incorpo-ration into the eukaryote cell. This has presumably reflectedsome functional purpose. It is tempting, therefore, to ascribe thedifferences between mitochondrial and bacterial complexes tothe greater structural complexity of the mitochondrial com-plexes, and the fine-tuning of mechanism to fit the complex to itseukaryotic role. On the other hand, some of the apparentdifferences between bacterial and mitochondrial systems mightreflect difficulties in working with the latter. Most work reportedhas been using the solubilized mitochondrial complexes instopped-flow experiments using unnatural quinone substrates.The time resolution is not as rapid as that available throughphotoactivation in chromatophores, and in most work, noattempt has been made to measure the separate kineticcontributions of the two b-hemes. The use of unnatural qui-nones, especially apparent in earlier work, introduces different

driving forces, and may also lead to different contributions ofbypass reactions [94]. Even when a good analog such as decylubiquinone is used, limitations are introduced by the low criticalmicellar concentration (<20 μM, [98]). These deficiencies, andthe uncertainty of some of the values for redox potentials ofcenters in the yeast complex, mean that it has not been possibleto analyze the data with the precision needed to discriminatebetween hypotheses. It might be as well to defer argument untilconstraints from data are better established. On the other hand,there is an old adage, “The heat of the controversy is inverselyproportional to the hardness of the data”; despite the caveat, 45years for one of us (ARC) in the bioenergetics field has shownthat the controversy is half the fun.

Acknowledgement

This work was supported by grant GM 35438 from NIH (toA.R.C.).

References

[1] A.R. Crofts, The bc1 complex: what is there left to argue about? in: M.Wikstrom (Ed.), Biophysical and Structural Aspects of Bioenergetics,Royal Society of Chemistry, Cambridge, 2005, pp. 121–155.

[2] C. Hunte, H. Palsdottir, B.L. Trumpower, Protonmotive pathways andmechanisms in the cytochrome bc1 complex, FEBS Lett. 545 (2003)39–46.

[3] W.A. Cramer, H.M. Zhang, J.S. Yan, G. Kurisu, J.L. Smith, Evolution ofphotosynthesis: time-independent structure of the cytochrome b6fcomplex, Biochemistry 43 (2004) 5921–5929.

[4] A.R. Crofts, The Q-cycle—A personal perspective, Photosynth. Res. 80(2003) 223–243.

[5] F. Muller, A.R. Crofts, D.M. Kramer, Multiple Q-cycle bypass reactions atthe Qo-site of the cytochrome bc1 complex, Biochemistry 41 (2002)7866–7874.

[6] A.R. Crofts, S.J. Hong, N. Ugulava, B. Barquera, R. Gennis, M. Guergova-Kuras, E.A. Berry, Pathways for proton release during ubihydroquinoneoxidation by the bc1 complex, Proc. Natl. Acad. Sci. U. S. A. 96 (1999)10021–10026.

[7] M.K.F. Wikström, J. Berden, Oxidoreduction of cytochrome b in thepresence of antimycin, Biochim. Biophys. Acta 283 (1972) 403–420.

[8] E.C. Slater, The cytochrome b paradox, the BAL-labile factor and the Q-cycle, in: V.P. Skulachev, P.C. Hinkle (Eds.), Chemiosmotic ProtonCircuits in Biological Membranes, Addison-Wesley Publ. Co., Reading,Mass, 1981, pp. 69–104.

[9] P. Mitchell, Proton motive redox mechanism of the cytochrome b–c1complex in the respiratory chain: proton motive ubiquinone cycle, FEBSLett. 56 (1975) 1–6.

[10] P. Mitchell, Possible molecular mechanisms of the protonmotive functionof cytochrome systems, J. Theor. Biol. 62 (1976) 327–367.

[11] P.B. Garland, R.A. Clegg, D. Boxer, J.A. Downie, B.A. Haddock, Proton-translocating nitrate reductase of Escherichia coli, in: E. Quagliariello, S.Papa, F. Palmieri, E.C. Slater, N. Siliprandi (Eds.), Electron TransferChains and Oxidative Phosphorylation, North-Holland Publishing Co.,Amsterdam, The Netherlands, 1975, pp. 351–358.

[12] A.R. Crofts, S.W. Meinhardt, K.R. Jones, M. Snozzi, The role of thequinone pool in the cyclic electron-transfer chain of Rhodopseudomonassphaeroides A modified Q-cycle mechanism, Biochim. Biophys. Acta(BBA)-Bioenergetics 723 (1983) 202–218.

[13] A.R. Crofts, The mechanism of ubiquinol:cytochrome c oxidoreductasesof mitochondria and of Rhodopseudomonas sphaeroides, in: A.N.Martonosi (Ed.), The Enzymes of Biological Membranes, Plenum Publ.Corp., New York, 1985, pp. 347–382.

[14] A.R. Crofts, V.P. Shinkarev, D.R.J. Kolling, S. Hong, The modified


Q-cycle explains the apparent mismatch between the kinetics ofreduction of cytochromes c1 and bH in the bc1 complex, J. Biol.Chem. 278 (2003) 36191–36201.

[15] D. Xia, C.-A. Yu, H. Kim, J.-Z. Xia, A.M. Kachurin, L. Zhang, L. Yu, J.Deisenhofer, Crystal structure of the cytochrome bc1 complex from bovineheart mitochondria, Science 277 (1997) 60–66.

[16] Z. Zhang, L.-S. Huang, V.M. Shulmeister, Y.-I. Chi, K.-K. Kim, L.-W.Hung, A.R. Crofts, E.A. Berry, S.-H. Kim, Electron transfer by domainmovement in cytochrome bc1, Nature (London) 392 (1998) 677–684.

[17] A.R. Crofts, The cytochrome bc1 complex–function in the context ofstructure, Annu. Rev. Physiol. (2004) 689–733.

[18] J.B. Jackson, A.R. Crofts, The high energy state in chromatophores fromRhodopseudomonas spheroides, FEBS Lett. 4 (1969) 185–189.

[19] J.B. Jackson, A.R. Crofts, The kinetics of light induced carotenoid changesin Rhodopseudomonas spheroides and their relation to electrical fieldgeneration across the chromatophore membrane, Eur. J. Biochem. 18(1971) 120–130.

[20] E.G. Glaser, A.R. Crofts, A new electrogenic step in the ubiquinol:cytochrome c2 oxidoreductase complex of Rhodopseudomonas sphaer-oides, Biochim. Biophys. Acta (BBA)-Bioenergetics 766 (1984) 322–333.

[21] G. Venturoli, M. Virgili, B.A. Melandri, A.R. Crofts, Kinetic measure-ments of electron transfer in coupled chromatophores from photosyntheticbacteria: a method of correction for the electrochromic effects, FEBS Lett.219 (1987) 477–484.

[22] Y. Chen, Kinetic studies of mechanism of ubiquinol:cytochrome c2oxidoreductase of R. sphaeroides: a steady state approach. PhD thesisBiophysics, pp. 143, University of Illinois at Urbana-Champaign 1989.

[23] Y. Chen, A.R. Crofts, Operation of UQH2:cyt c2 oxidoreductase of Rb.sphaeroides in the steady state, in: M. Baltscheffsky (Ed.), CurrentResearch in Photosynthesis, vol. III, Kluwer Academic Publishers,Dordrecht, 1990, pp. 287–290.

[24] A.R. Crofts, Z. Wang, Y. Chen, S. Mahalingham, C.-H. Yun, R.B. Gennis,Function, steady-state turnover, and structure of the ubiquinol:cytochromec2 oxidoreductase of Rhodobacter sphaeroides, in: G. Lenaz, O. Barnabei,A. Rabbi, M. Battino (Eds.), Highlights in Ubiquinone Research, Taylorand Francis, London, 1990, pp. 98–103.

[25] J.J. Lemasters, W.H. Billica, Non-equilibrium thermodynamics ofoxidative phosphorylation by inverted inner membrane vesicles of ratliver mitochondria, J. Biol. Chem. 256 (1981) 12949–12957.

[26] S.J. Hong, N. Ugulava, M. Guergova-Kuras, A.R. Crofts, The energylandscape for ubihydroquinone oxidation at the Qo-site of the bc1 complexin Rhodobacter sphaeroides, J. Biol. Chem. 274 (1999) 33931–33944.

[27] A.R. Crofts, Z. Wang, How rapid are the internal reactions of theubiquinol:cytochrome c2 oxidoreductase? Photosynth. Res. 22 (1989)69–87.

[28] E. Darrouzet, M. Valkova-Valchanova, C.C. Moser, P.L. Dutton, F. Daldal,Uncovering the [2Fe2S] domain movement in cytochrome bc1 and itsimplications for energy conversion, Proc. Natl. Acad. Sci. U. S. A. 97(2000) 4567–4572.

[29] H. Tian, S. White, L. Yu, C.-A. Yu, Evidence for the head domainmovement of the Rieske iron–sulfur protein in electron transfer reaction ofthe cytochrome bc1 complex, J. Biol. Chem. 274 (1999) 7146–7152.

[30] H. Tian, L. Yu, M.W. Mather, C.-A. Yu, Flexibility of the neck region ofthe Rieske iron–sulfur protein is functionally important in the cytochromebc1 complex, J. Biol. Chem. 273 (1998) 27953–27959.

[31] E. Darrouzet, F. Daldal, Protein–protein interactions between cytochrome band the Fe–S protein subunits duringQH2 oxidation and large-scale domainmovement in the bc1 complex, Biochemistry 42 (2003) 1499–1507.

[32] A.R. Crofts, M. Guergova-Kuras, L.-S. Huang, R. Kuras, Z. Zhang, E.A.Berry, The mechanism of ubiquinol oxidation by the bc1 complex: the roleof the iron sulfur protein, and its mobility, Biochemistry 38 (1999)15791–15806.

[33] A.R. Crofts, S. Hong, Z. Zhang, E.A. Berry, Physicochemical aspects ofthe movement of the Rieske iron sulfur protein during quinol oxidation bythe bc1 complex, Biochemistry 38 (1999) 15827–15839.

[34] S. Izrailev, A.R. Crofts, E.A. Berry, K. Schulten, Steered moleculardynamics simulation of the Rieske subunit motion in the cytochrome bc1complex, Biophys. J. 77 (1999) 1753–1768.

[35] G. Engstrom, K. Xiao, C.-A. Yu, L. Yu, B. Durham, F. Millett,Photoinduced electron transfer between the Rieske iron–sulfur proteinand cytochrome c1 in the Rhodobacter sphaeroides cytochrome bc1complex: effects of pH, temperature, and driving force, J. Biol. Chem. 277(2002) 31072–31078.

[36] A.R. Crofts, B. Barquera, R.B. Gennis, R. Kuras, M. Guergova-Kuras,E.A. Berry, The mechanism of ubiquinol oxidation by the bc1complex: the different domains of the quinol binding pocket, and theirrole in mechanism, and the binding of inhibitors, Biochemistry 38(1999) 15807–15826.

[37] E. Darrouzet, M. Valkova-Valchanova, F. Daldal, The [2Fe–2S] cluster Em

as an indicator of the iron–sulfur subunit position in the ubihydroquinoneoxidation site of the cytochrome bc1 complex, J. Biol. Chem. 277 (2002)2464–3470.

[38] X. Gao, X. Wen, C.-A. Yu, L. Esser, S. Tsao, B. Quinn, L. Zhang, L. Yu, D.Xia, The crystal structure of mitochondrial cytochrome bc1 in complexwith famoxadone: the role of aromatic–aromatic interaction in inhibition,Biochemistry 41 (2002) 11692–11702.

[39] S. Iwata, J.W. Lee, K. Okada, J.K. Lee, M. Iwata, B. Rasmussen, T.A.Link, S. Ramaswamy, B.K. Jap, Complete structure of the 11-subunitbovine mitochondrial cytochrome bc1 complex, Science 281 (1998)64–71.

[40] H. Palsdottir, C.G. Lojero, B.L. Trumpower, C. Hunte, Structure of theyeast cytochrome bc1 complex with a hydroxyquinone anion Qo siteinhibitor bound, J. Biol. Chem. 278 (2003) 31303–31311.

[41] E.A. Berry, L.S. Huang, Observations concerning the quinol oxidation siteof the cytochrome bc1 complex, FEBS Lett. 555 (2003) 13–20.

[42] J.R. Bowyer, P.L. Dutton, R.C. Prince, A.R. Crofts, The role of the Rieskeiron–sulfur center as the electron donor to ferricytochrome c2 inRhodopseudomonas sphaeroides, Biochim. Biophys. Acta (BBA)-Bioen-ergetics 592 (1980) 445–460.

[43] T.A. Link, U. Haase, U. Brandt, G. von Jagow, What information doinhibitors provide about the structure of the hydroquinone oxidation site ofubiquinol:cytochrome c oxidoreductase, J. Bioenerg. Biomembranes 25(1993) 221–232.

[44] H. Ding, D.E. Robertson, F. Daldal, P.L. Dutton, Cytochrome bc1 complex[2Fe–2S] cluster and its interaction with ubiquinone and ubihydroquinoneat the Qo site: a double-occupancy Qo site model, Biochemistry 31 (1992)3144–3158.

[45] A.R. Crofts, V.P. Shinkarev, S.A. Dikanov, R.I. Samoilova, D. Kolling,Interactions of quinone with the iron–sulfur protein of the bc1 complex: isthe mechanism spring-loaded? Biochim. Biophys. Acta 1555 (2002)48–53.

[46] A.R. Crofts, M. Guergova-Kuras, R. Kuras, N. Ugulava, J. Li, S. Hong,Proton-coupled electron transfer at the Qo-site: what type of mechanismcan account for the high activation barrier? Biochim. Biophys. Acta 1459(2000) 456–466.

[47] A.R. Crofts, Proton-coupled electron transfer at the Qo-site of the bc1complex controls the rate of ubihydroquinone oxidation, Biochim.Biophys. Acta 1655 (2004) 77–92.

[48] X. Gao, X. Wen, L. Esser, B. Quinn, L. Yu, C.-A. Yu, D. Xia, Structuralbasis for the quinone reduction in the bc1 complex: a comparative analysisof crystal structures of mitochondrial cytochrome bc1 with bound substrateand inhibitors at the Qi site, Biochemistry 42 (2003) 9067–9080.

[49] A.R. Crofts, S.J. Hong, N. Ugulava, B. Barquera, R.B. Gennis, M.Guergova-Kuras, E. Berry, Pathways for proton release during ubihy-droquinone oxidation by the bc1 complex, Proc. Natl. Acad. Sci. U. S. A.96 (1999) 10021–10026.

[50] C. Hunte, J. Koepke, C. Lange, T. Roßmanith, H. Michel, Structure at2.3 Å resolution of the cytochrome bc1 complex from the yeastSaccharomyces cerevisiae co-crystallized with an antibody Fv fragment,Structure 8 (2000) 669–684.

[51] N.B. Ugulava, A.R. Crofts, CD-monitored redox titration of the RieskeFe–S protein of Rhodobacter sphaeroides: pH dependence of the mid-point potential in isolated bc1 complex and in membranes, FEBS Lett. 440(1998) 409–413.

[52] C. Snyder, B.L. Trumpower, Mechanism of ubiquinol oxidation by thecytochrome bc1 complex–pre-steady-state kinetics of cytochrome bc1


complexes containing site-directed mutants of the Rieske iron–sulfurprotein, Biochim. Biophys. Acta 1365 (1998) 125–134.

[53] M. Iwaki, G. Yakovlev, J. Hirst, A. Osyczka, P.L. Dutton, D. Marshall, P.R.Rich, Direct observation of redox-linked histidine protonation changes inthe iron–sulfur protein of the cytochrome bc1 complex by ATR-FTIRspectroscopy, Biochemistry 44 (2005) 4230–4237.

[54] E.H. Evans, A.R. Crofts, In situ characterization of photosynthetic electrontransport in Rhodopseudomonas capsulata, Biochim. Biophys. Acta 357(1974) 89–102.

[55] R.C. Prince, P.L. Dutton, Single and multiple turnover reactions in theubiquinone-cytochrome b–c2 oxidoreductase of Rhodopseudomonassphearoides: the physical chemistry of the major electron donor tocytochrome c2, and its coupled reactions, Biochim. Biophys. Acta 462(1977) 731–747.

[56] M. Snozzi, A.R. Crofts, Electron transport in chromatophores fromRhodopseudomonas sphaeroides GA fused with liposomes, Biochim.Biophys. Acta (BBA)-Bioenergetics 766 (1984) 451–463.

[57] T.A. Link, G. von Jagow, Zinc ions inhibit the QP center of bovine heartmitochondrial bc1 complex by blocking a protonatable group, J. Biol.Chem. 270 (1995) 25001–25006.

[58] U. Brandt, J.G. Okun, Role of deprotonation events in ubihydroquinone:cytochrome c oxidoreductase from bovine heart and yeast mitochondria,Biochemistry 36 (1997) 11234–11240.

[59] M. Guergova-Kuras, R. Kuras, N. Ugulava, I. Hadad, A.R. Crofts, Specificmutagenesis of the Rieske iron sulfur protein in Rhodobacter sphaeroidesshows that both thermodynamic gradient and the pK of the oxidized formdetermine the rate of quinol oxidation by the bc1 complex, Biochemistry39 (2000) 7436–7444.

[60] R.J. Gurbiel, T. Ohnishi, D. Robertson, F. Daldal, B.M. Hoffman, Q-BandENDOR spectra of the Rieske protein from Rhodobacter capsulatusubiquinol-cytochrome c oxidoreductase show two histidines coordinatedto the [2Fe–2S] cluster. Biochemistry 30 (1991) 11579–11584.

[61] S. Iwata, M. Saynovits, T.A. Link, H. Michel, Structure of a water solublefragment of the Rieske iron sulfur protein of the bovine heartmitochondrial cytochrome bc1 complex determined by MAD phasing at1.5 Å resolution, Structure 4 (1996) 567–579.

[62] P.R. Rich, in: A. Kotyk, et al. (Eds.), Highlights of Modern Biochemistry,V.S.P. Publishers, Prague, 1989, pp. 903–912.

[63] D.E. Robertson, F. Daldal, P.L. Dutton, Mutants of ubiquinol-cytochromec2 oxidoreductase resistant to Qo site inhibitors: consequences forubiquinone and ubiquinol affinity and catalysis, Biochemistry 29 (1990)11249–11260.

[64] U. Brandt, U. Haase, H. Schägger, G. von Jagow, Significance of the“Rieske” iron–sulfur protein for formation and function of the ubiquinol-oxidation pocket of mitochondrial cytochrome c reductase (bc1 complex),J. Biol. Chem. 266 (1991) 19958–19964.

[65] R.I. Samoilova, D. Kolling, T. Uzawa, T. Iwasaki, A.R. Crofts, S.A.Dikanov, The interaction of the Rieske iron sulfur protein with occupantsof the Qo-site of the bc1 complex, probed by 1D and 2D Electron SpinEcho Envelope Modulation, J. Biol. Chem. 277 (2002) 4605–4608.

[66] R.E. Sharp, B.R. Gibney, A. Palmitessa, J.L. White, J.A. Dixon, C.C.Moser, F. Daldal, P.L. Dutton, Effect of inhibitors on the ubiquinone bind-ing capacity of the primary energy conversion site in the Rhodobacter cap-sulatus cytochrome bc1 complex, Biochemistry 38 (1999) 14973–14980.

[67] J.W. Cooley, A.G. Roberts, M.K. Bowman, D.M. Kramer, F. Daldal, Theraised midpoint potential of the [2Fe2S] cluster of cytochrome bc1 ismediated by both the Q(o) site occupants and the head domain position ofthe Fe–S protein subunit, Biochemistry 43 (2004) 2217–2227.

[68] V.P. Shinkarev, D.R.J. Kolling, T.J. Miller, A.R. Crofts, Modulation of themidpoint potential of the [2Fe–2S] Rieske iron sulfur center by Qo

occupants in the bc1 complex, Biochemistry 41 (2002) 14372–14382.[69] E. Darrouzet, M. Valkova-Valchanova, F. Daldal, The [2Fe–2S] cluster Em

as an indicator of the iron–sulfur subunit position in the ubihydroquinoneoxidation site of the cytochrome bc1 complex, J. Biol. Chem. 277 (2002)3464–3470.

[70] A. Boveris, Determination of the production of superoxide radicals andhydrogen-peroxide in mitochondria, Methods Enzymol. 105 (1984)429–435.

[71] F. Muller, The nature and mechanism of superoxide production by theelectron transport chain: its relevance to aging, J. Am. Aging Assoc. 23(2000) 227–253.

[72] J.F. Turrens, A. Alexandre, A.L. Lehninger, Ubisemiquinone is theelectron donor for superoxide formation by complex III of heartmitochondria, Arch. Biochem. Biophys. 237 (1985) 408–414.

[73] C.C. Moser, P.L. Dutton, Engineering protein structure for electron transferfunction in photosynthetic reaction centers, Biochim. Biophys. Acta 1101(1992) 171–176.

[74] A. Osyczka, C.C. Moser, P.L. Dutton, Fixing the Q-cycle, TrendsBiochem. Sci. 30 (2005) 176–182.

[75] A. Osyczka, C.C. Moser, F. Daldal, P.L. Dutton, Reversible redox energycoupling in electron transfer chains, Nature 427 (2004) 607–612.

[76] E. Denke, T. Merbitzzahradnik, O.M. Hatzfeld, C.H. Snyder, T.A. Link,B.L. Trumpower, Alteration of the midpoint potential of the Rieskeiron–sulfur protein by changes of amino acids forming H-bonds to theiron–sulfur cluster, J. Biol. Chem. 273 (1998) 9085–9093.

[77] T. Schröter, O.M. Hatzfeld, S. Gemeinhardt, M. Korn, T. Friedrich, B.Ludwig, T. Link, Mutational analysis of residues forming hydrogen bondsin the Rieske [2Fe2S] cluster of the cytochrome bc1 complex ofParacoccus denitrificans, Eur. J. Biochem. 255 (1998) 100–106.

[78] G. Brasseur, A. Sami Saribas, F. Daldal, A compilation of mutationslocated in the cytochrome b subunit of the bacterial and mitochondrial bc1complex, Biochim. Biophys. Acta 1275 (1996) 61–69.

[79] T.A. Link, The role of the “Rieske” iron sulfur protein in the hydroquinoneoxidation (Qp

−) site of the cytochrome bc1 complex: the “proton-gatedaffinity change” mechanism, FEBS Lett. 412 (1997) 257–264.

[80] U. Brandt, Bifurcated ubihydroquinone oxidation in the cytochromebc1 complex by proton-gated charge transfer, FEBS Lett. 387 (1996)1–6.

[81] Y. Zu, M.-J. Manon, M.M.-J. Couture, D.R.J. Kolling, A.R. Crofts, L.D.Eltis, J.A. Fee, J. Hirst, The reduction potentials of Rieske clusters: theimportance of the coupling between oxidation state and histidineprotonation state, Biochemistry 42 (2003) 12400–12408.

[82] S.S. Korshunov, V.P. Skulachev, A.A. Starkov, High protonic potentialactuates a mechanism of production of reactive oxygen species inmitochondria, FEBS Lett. 416 (1997) 15–18.

[83] F.L. Muller, A.G. Roberts, M.K. Bowman, D.M. Kramer, Architecture ofthe Qo site of the cytochrome bc1 complex probed by superoxideproduction, Biochemistry 42 (2003) 6493–6499.

[84] B.L. Trumpower, A concerted, alternating sites mechanism of ubiquinoloxidation by the dimeric cytochrome bc1 complex, Biochim. Biophys.Acta 1555 (2002) 166–173.

[85] C.-A. Yu, X. Wen, K. Xiao, D. Xia, L. Yu, Inter- and intra-molecularelectron transfer in the cytochrome bc1 complex, Biochim. Biophys. Acta1555 (2002) 65–70.

[86] C.H. Snyder, E.B. Gutierrez-Cirlos, B.L. Trumpower, Evidence for aconcerted mechanism of ubiquinol oxidation by the cytochrome bc1complex, J. Biol. Chem. 275 (2000) 13535–13554.

[87] U. Brandt, The chemistry and mechanics of ubihydroquinone oxidation atcenter P (Qo) of the cytochrome bc1 complex, Biochim. Biophys. Acta1365 (1998) 261–268.

[88] S.S. Klishin, W. Junge, A.Y. Mulkidjanian, Flash-induced turnover of thecytochrome bc1 complex in chromatophores of Rhodobacter capsulatus:binding of Zn2+ decelerates likewise the oxidation of cytochrome b, thereduction of cytochrome c1 and the voltage generation. Biochim. Biophys.Acta 1553 (2002) 177–182.

[89] A.Y. Mulkidjanian, Ubiquinol oxidation in the cytochrome bc1 complex:reaction mechanism and prevention of short-circuiting, Biochim. Biophys.Acta 1709 (2005) 5–34.

[90] H. Kim, D. Xia, C.A. Yu, J.Z. Xia, A.M. Kachurin, L. Zhang, L. Yu, J.Deisenhofer, Inhibitor binding changes domain mobility in the iron–sulfurprotein of the mitochondrial bc1 complex from bovine heart, Proc. Natl.Acad. Sci. U. S. A. 95 (1998) 8026–8033.

[91] C. Lange, C. Hunte, Crystal structure of the yeast cytochrome bc1 complexwith its bound substrate cytochrome c, Proc. Natl. Acad. Sci. U. S. A. 99(2002) 2800–2805.

[92] J.W. Cooley, T. Ohnishi, F. Daldal, Binding dynamics at the quinone


reduction (Qi) site influence the equilibrium interactions of the iron sulfurprotein and hydroquinone oxidation (Qo) site of the cytochrome bc1complex, Biochemistry 44 (2005) 10520–10532.

[93] L.S. Huang, D. Cobessi, E.Y. Tung, E.A. Berry, Binding of the respiratorychain inhibitor antimycin to the mitochondrial bc1 complex: a new crystalstructure reveals an altered intramolecular hydrogen-bonding pattern, J.Mol. Biol. 351 (2005) 573–597.

[94] J.L. Cape, J.R. Strahan, M.J. Lenaeus, B.A. Yuknis, T.T. Le, J.N.Shepherd, M.K. Bowman, D.M. Kramer, The respiratory substraterhodoquinol induces Q-cycle bypass reactions in the yeast cytochromebc1 complex—Mechanistic and physiological implications, J. Biol. Chem.280 (2005) 34654–34660.

[95] P.R. Rich, The quinone chemistry of bc complexes, Biochim. Biophys.Acta 1658 (2004) 165–171.

[96] L. Zhang, Z. Li, B. Quinn, L. Yu, C.-A. Yu, Nonoxidizable ubiquinolderivatives that are suitable for the study of the ubiquinol oxidation site inthe cytochrome bc1 complex, Biochim. Biophys. Acta 1556 (2002)226–232.

[97] S.W. Meinhardt, A.R. Crofts, The role of cytochrome b-566 in the electrontransfer chain of Rhodopseudomonas sphaeroides, Biochim. Biophys.Acta 723 (1983) 219–230.

[98] G. Lenaz, M. Degli Esposti, Physical properties of ubiquinones in modelsystems and membranes. In G. Lenaz (Ed.), Coenzyme Q, Wiley and Sons,Ltd., Chichester, U.K, pp. 83–105.

review proton pumping in the bc complex: a new gating ... proton pumping in the bc 1 complex: a new...

Documents