Abstract

Introduction

Although the modified Q-cycle mechanism is well supported, there remain several problems in understanding the mechanism of catalysis at the molecular level. A more detailed understanding of the structure-function relationship at the catalytic sites requires a structural context, and this has recently been provided by X-ray crystallographic studies of the structure of the complexes from mitochondria (15-23). In this paper, we review the information on mechanism obtained from studies of mutations that effect the Q_o-site, and consider this in the context of the detailed structural information from crystallographic studies of the chicken heart mitochondrial bc₁-complex (20-22).

Methods

Results and Discussion

Structure of the catalytic subunits

In contrast, the ISP changes its position dramatically on binding stigmatellin, showing a rotational displacement in which the 2Fe2S center moves by ~16 Å from a position close to cyt c₁ in the absence, to a docking interface on cyt b in the presence of inhibitor. Theoretical considerations suggest that the structure would not be competent in all electron transfer steps in either of the positions found, and we have suggested that the ISP must move between these positions during turn over. Figure 1 summarizes the mechanism we have proposed, in which the movement of the catalytic domain of the ISP between reaction interfaces on cytochrome b and cytochrome c₁ subunits is required for electron transfer from quinol to cyt c₁ (31).

An extensive literature on mutations that affect inhibitor binding or kinetic function at the Q_o-site has been summarized recently by Brasseur et al. (27); our own recent work on site-directed mutations in the ef-loop, including the -PEWY- span, will be presented in detail elsewhere (28, and Barquera et al., in preparation). This work has identified spans in the sequence likely to contribute structurally; the C-terminal end of the transmembrane helix C, the N-terminal section of the cd loop, including the amphipathic cd-helix, the ef loop, including a coil from the end of the helix E, the -PEWY- turn, the small ef helix connecting the -PEWY- turn to transmembrane helix, and the N-terminal end of helix F. In the structure, these spans define two features: i) a quinol binding pocket, occupying a volume in which a difference electron density is seen in crystals containing myxothiazol or stigmatellin, and ii) a concave surface, which is a docking interface for the ISP. The transmembrane helices contributing to the site flare out to provide the volume, and to form the sides of the pocket, the -PEWY- loop acts as a spacer, and the amphipathic helices provide the bottom. They also, on their other faces, provide most of the ISP binding surface, with the ef-loop providing additional residues. From examination of the placing of these domains in the structure it is apparent that mutations in either can interfere with the catalytic mechanism.

Differential effects on binding of the two classes of inhibitor acting at the site have been noted, with strains resistant to stigmatellin, but not myxothiazol or mucidin, and vice versa, and strains resistant to both inhibitors (1, 27, 28). Binding studies have shown that occupation is mutually exclusive (3, 10, 12, 13, 32-35), so that only one occupant (quinone, stigmatellin, myxothiazol, or UHDBT) can be in the site, and the crystallographic studies confirm these earlier results. Mutations have several different effects on the changes in EPR spectrum of the Rieske center at g_x=1.800, which have been attributed to interaction with quinone at the Q_o-site. However, for any particular mutation, the effects on signals attributed to Q_ow and Q_os have been the same.

The Q_o-site is shown in greater detail in Fig. 2, in which residues known to effect inhibitor binding or catalysis are shown as stick models. The locations of stigmatellin and myxothiazol are modeled on the basis of the electron density differences observed in separate crystals containing these inhibitors, compared to the native structure.

A. Inhibitor resistance sites. Note that in many cases, several mutations have been generated at a site, with differential effects on inhibitor binding, depending on the nature of the change. The following color-coding is therefore somewhat simplistic. Myxothiazol resistant sites are shown in yellow, stigmatellin resistance sites are shown in red. Residues at which modification leads to resistance to both inhibitors are shown in green. Stigmatellin and myxothiazol are shown as ball and stick models, colored cyan and white respectively, occupying the volumes in which they are found in separate crystals. The heme of cyt b_L is shown as a magenta wireframe model. The iron sulfur protein is shown in the stigmatellin-induced configuration, as a blue backbone model.

Kinetics in the high potential chain

Myxothiazol binds at the Q_o-site of cyt b, and displaces quinol, preventing the delivery of electrons to the 2Fe2S center of the ISP. This simplifies the electron transfer chain so that the reactions of the high potential components can be examined separately. In Rb. sphaeroides, oxidizing equivalents from the primary donor (P⁺/P) of photochemical reaction center (RC) are delivered to cytochrome c₁ by cytochrome c₂, with a t_½; of about 150 ms. The electron transfer from the Rieske center to cyt c₁ cannot be resolved, because it is not rate determining. However, kinetic arguments suggest that the rate constant must be in the range 10⁵ s^-1 (5, 7, 9). In the absence of myxothiazol, the bifurcated reaction occurs, resulting in cytochrome b reduction. A ~200 ms lag observed before onset of reduction of cyt b_H is contributed by several processes: the delivery of oxidizing equivalents to cyt c₁ (~150 ms), the oxidation of the 2Fe2S center (<10 ms), and the movement of the ISP to its reaction interface with QH₂. The reactions of the ISP with its partners (QH₂ bound at the Q_o-site of cyt b, and the heme of cyt c₁) (Fig. 1) require a movement of ~ 16 Å between two reaction domains on these separate subunits.

 Kinetics of quinol oxidation

In the oxidation of quinol, the kinetics of electron transfer to the high potential chain and the cytochrome b chain are closely matched. The system behaves as if the oxidation involved a concerted, two-electron reaction with simultaneous delivery of electrons to two separate chains. However, it is generally accepted that this behavior reflects a sequential reaction in which the intermediate semiquinone is consumed rapidly, and never attains a significant concentration.
 

 Crofts and Wang (9) developed a model that accounted well for the kinetic features of quinol oxidation; although this provided an excellent fit to the kinetics over a wide range of conditions, one aspect of the mechanism of quinol oxidation was not explicitly address. It has been a long standing paradox that the bifurcated reaction at the Q_o-site delivered electrons stoichiometrically to the two separate chains, despite the strong thermodynamic potential favoring delivery of both electrons to the high potential chain (see Brandt (37) for a recent discussion).

 Double-occupancy model

Earlier work had shown an effect of the redox state of the quinone pool on the EPR spectrum of the 2Fe2S center of the complex (33, 38, 39). More recently, Ding et al. (34, 35, 40), using wild-type and mutant strains from Rb. capsulatus, have investigated the effects of extraction of ubiquinone, and concluded that different spectral changes can be detected at different local concentrations. They suggested that these reflect two different bound quinones at the Q_o-site, called Q_os and Q_ow for strongly and weakly binding species, occupying separate domains, at both of which Q and QH₂ bind with equal affinity (i.e. with no change in E_m with respect to the pool). They have extended these observations to mutant strains, and kinetic studies of turnover and inhibitor binding, and have discussed possible mechanisms and atomic details of ligation. Brandt (43) has recently expanded on the double occupancy model to suggest an explanation for the observation that electrons do not "leak" rapidly from the strongly reducing semiquinone at the Q_o-site into the high potential chain. The extensive work on the g_x signals of the 2Fe2S center has contributed a useful set of data on interactions of quinone and inhibitor species with the ISP, and on occupancy of the Q_o-site. Nevertheless, there are some problems with the interpretation offered (28).

Effects of Q_o-site occupancy on EPR spectra of the Iron Sulfur Protein.

 In summary, with the Q-pool oxidized and ISP reduced, a sharp band at g_x=1.80 is seen. When the pool is reduced, or Q extracted, or myxothiazol bound at the Q_o-site, the g_x signal shifts from 1.80 to ~1.765, and becomes smaller (34, 35, 39, 40). When the pool is extracted so as to leave ~2 Q/RC (1 as Q_A, plus ~2Q/bc₁-complex in the preparations used) an intermediate signal at g_x=1.783 is seen. Several possible explanations for these effects on the EPR spectrum can be offered:

1. Ding et al. (35, 40) suggested that the g_x=1.783 signal was due to the Q_os species, and that this would likely be the one forming a strong interaction with the ISP. From the structure, this would translate to Q_os occupying the distal end of the pocket (where stigmatellin binds). Q_ow was suggested to be an exchangeable species, through which coupling to the quinone pool could occur. In the context of the structure, Q_ow would be expected to bind at the proximal (myxothiazol) end, nearer heme b_L. Since the g_x=1.800 signal is ascribed to Q_ow, the interaction with the 2Fe2S center indicated by the signal would have to be indirect, and mediated by Q_os.
2. The g_x=1.800 signal is due to ISP_red interacting directly with Q at the Q_o-site, and is lost when the site is either empty, occupied by inhibitor, or by QH₂. We believe this explanation requires the fewest ad hoc hypotheses. However, if this is the case, then under conditions in which the quinone pool is oxidized before flash-activation, the reduced ISP would not be in position for its reaction with cytochrome c₁. It would be necessary to postulate that the diffusion time between cyt b and cyt c₁ docking interfaces is rapid compared to the <10 ms oxidation time of 2Fe2S observed, and that the binding at the cyt b interface under these conditions (Q oxidized, ISP reduced) must not be so great as to prevent the ISP leaving within this time. We discuss these parameters in greater detail elsewhere. (31).

1. None of the structures (22, 23) show a density in the uninhibited site expected for a tightly bound quinone species. In the Berry structures, a weak density in the unoccupied site is observed, which may represent a quinone species, either weakly bound with low occupancy in a fixed position, or mobile in the site. In the Xia et al. structure, the authors reported that no loss of density from the pocket was observed on binding inhibitors at the Q_o-site, although a loss of density was observed at the Q_i-site on binding antimycin, and this was attributed to displacement of quinone (23). Since their crystals contained sub-stoichiometric levels of quinone (~0.6 Q/complex), these results would suggest that the Q_i-site binds the quinone more tightly than the Q_o-site.
2. In the structures from both groups, the electron densities of inhibitors in the stigmatellin and myxothiazol crystals occupy overlapping volumes, suggesting, in line with competition studies, that these two inhibitors are mutually exclusive occupants of the site (see Fig. 2). Although at the current structural resolution, we cannot exclude the possibility that two ubiquinones could occupy the same site, it seems probable that only one quinone can occupy this volume.
3. The intermediate line-shape at g_x=1.783, seen when ~2 Q / bc₁-complex are present, is attributed to Q_os, a species that binds 30-fold more tightly than Q_ow (34, 35). Paradoxically, the other properties of these two species are similar. This paradox could be resolved if the line shape change reflected only one species. In attributing the 1.783 signal, Dutton and colleagues assumed a tight binding species. A difficulty with this assumption comes from the heterogeneity in distribution of redox components likely in a population of chromatophores (cf. 42). Extraction of chromatophores to the point giving rise to the Q_os signal would leave the distribution of quinone random, so that some chromatophores would have an excess of quinone over the 2 per bc₁-complex, and we would expect to see a sharp peak at 1.800 due to Q_ow. In another fraction, the quinone would be below this ratio, leaving some bc₁-complex without a quinone, with a shallow peak at 1.765. At the very least, we would expect in the overall population a peak representing the convolution of these two spectra, at a position intermediate between them, as previously seen in pentane extracted mitochondrial complex III by de Vries et al. (39).
4. The environment of the 2Fe2S center, and the protein interactions of the Rieske subunit, will change substantially as the ISP moves. In addition, the interaction between the spins of the reduced 2Fe2S center, and the oxidized cyt b_L (52) might also be expected to change during catalysis, since the distance between them would change by ~8 Å. Some perturbation of the spectrum of the 2Fe2S center might be attributed to these changes.

1. The g_x=1.800 signal is associated with a complex formed between quinone bound in the Q_o-site at the distal end, and the reduced ISP docked firmly at the interface with cytochrome b.
2. The signal is lost whenever this complex cannot form: - on reduction of Q, on binding of inhibitors, on extraction of quinone, in strains where mutation prevents binding at the distal end of the pocket, and in strains where mutation prevents docking of the ISP.
3. Only one quinone species occupies the site. The spectroscopic effects (g_x=1.783) attributed to Q_os arise from the statistical distribution of quinone in the heterogeneous chromatophore population, with additional contributions from changes in environment of the ISP due to changes in its interactions accompanying extraction of the last occupant, to leave a vacant site.

The paradox of the bifurcated reaction

In the presence of antimycin, the Q_o-site is free to turn over, and under steady state conditions in which quinol is available in the pool, the b-cytochrome chain becomes reduced, and the ISP is oxidized by the high potential acceptor. Under these circumstances, de Vries was able to detect a small amount of semiquinone at the Q_o-site. Since antimycin is an effective inhibitor, and net electron transfer is blocked under these conditions, the semiquinone must not pass its second electron to the ISP. How is this thermodynamically favored reaction prevented? Several factors might be involved:

1. a set of forces in the binding domain which disfavor the binding of the oxidized ISP, or
2. displacement of the semiquinone by ligand exchange to a position deeper in the pocket, and closer to its reaction partner, the cyt b_L heme, or
3. a conformational change in which movement of one or more residues "insulates" the quinone binding pocket from the ISP domain.

We favor the second of these possibilities as the main factor. The possibility of such a movement is supported by the deeper position in the pocket of myxothiazol compared to stigmatellin (Fig. 2). The recent evidence for a substantial movement of Q_B in its binding pocket when it is reduced to semiquinone in the bacterial reaction center (41), also shows the possibility of such movement. Stowell et al. suggest that the activation barrier which gives the high temperature dependence of electron transfer from Q_A^-Q_B to Q_AQ_B^-, might reflect this movement, and such considerations could obviously be applied to the reaction at the Q_o-site. In support of the third possibilitiy, the position of the head of stigmatellin (as currently modeled) overlaps the volume occupied by the ring of a tyrosine side chain (Tyr-279) in the uninhibited complex, and preliminary analysis suggests that the sidechain rotates from this position on binding of inhibitor (Berry, E.A., unpublished). In the absence of inhibitor, the tyrosine sits at the opening between the Q-binding pocket and the ISP interface. It seems likely that any occupant of the pocket that interacts with the ISP can only do so if the tyrosine is displaced. We suggest that the tyrosine might act as a trap-door controlling access to the occupant of the binding pocket. Stigmatellin also occupies some common volume with the position of the ring of Pro-271 in the uninhibited structure, which also appears to be involved in a small conformational change on occupation of the distal end of the pocket.

 The distribution of residues at which mutation results in modified function: distinct domains and modes of action

The extensive set of data provided by Ding et al. (34, 35, 40) on occupancy of the site in different mutant strains, as reflected in the g_x=1.800 signal of the 2Fe2S center, and our own work in collaboration with these authors (28), provide important clues about the mechanism. At the present resolution of the structure, we can distinguish several different classes of mutational change, which are color-coded in Fig. 2A (inhibitor resistance sites) and B (residues where change leads to a modified g_x=1.800 signal).

1. Residue changes at the interface with the ISP

Changes at Lys-288, Leu-282, Thr-145, Gly-143, Ile-269, which can be seen to project into the ISP interfacial surface in the Fig. 2B, eliminate the g_x=1.800 signal, which we have attributed to interaction between quinone at the Q_o-site and the reduced ISP. In some cases this is associated with a loss of activity, but some of these strains show substantial turnover. It seems possible that the mutations that eliminate or reduce the g_x signal interfere with the ISP binding, and prevent formation of the complex between Q and ISP_red that gives rise to this signal. The steric effects that give rise to loss of the EPR band might also be expected to interfere with the access of the ISP_ox to QH₂ at the site. The variable affects on turnover number would then be explained by the decrease in the residence time of the ISP at the docking interface, and therefore the probability of formation of the reaction complex.
 

2. Residue changes at the proximal end of the Q_o-binding pocket

Residue changes (yellow in Fig. 2B above), which lead to loss of activity without loss of the g_x=1.800 signal, cluster at the proximal end of the pocket, closer to the heme of cyt b_L, where myxothiazol binds. Many of the myxothiazol resistance strains also have residue changes at this end of the pocket (yellow and green in Fig. 2A). It seems reasonable to conclude that these mutations do not interfere with the binding of either Q at the distal end, or the ISP. However, they do prevent oxidation of QH₂. The inhibition of oxidation in these mutants must reflect some effect relating to the occupancy of the proximal end of the pocket. This could either be a second quinone species involved in electron transfer to the b_L heme (in a modified double-occupancy model), or a need for movement of the semiquinone to the proximal position to bring it closer to cyt b_L to allow rapid electron transfer.
 

3. Residue changes at the distal end of the Q_o-binding pocket

Conclusions

1. Oxidizing equivalents are transferred to the site through cytochrome c₁ and the ISP. The mechanism involves a movement of the ISP between reaction domains on cytochrome c₁ and cytochrome b (Fig.1).
2. The reaction proceeds from a complex between the oxidized ISP and quinol. The reaction complex involves QH₂ bound at the distal end of the pocket, and the oxidized ISP docked tightly at the interface on cytochrome b. Binding of QH₂ and formation of the complex requires a displacement of Tyr-279, to allow access of the ISP to the quinol. This binding likely also leads to release of 1 H⁺ through an interaction at the site that changes the effective pK of the QH₂/QH^-.H⁺ dissociation reaction.
3. Electron transfer from QH₂ to the ISP leads to formation of semiquinone.
4. The semiquinone moves in the pocket to the proximal end, near heme b_L.
5. The Tyr-279 trap-door flips back to its "closed" position, insulating the semiquinone from further reaction with the ISP.
6. The semiquinone passes its electron to the cyt b_L heme, and the quinone exits the site.
7. At some point between c) and f), the second proton is released. This might be required for formation of the semiquinone, or its movement in the pocket, or the transfer of the second electron. We suggest that the pH dependence of the activation barrier (37) reflects this second deprotonation, and that the activation barrier is in one of these steps.
8. Meanwhile, formation of the semiquinone liberates the reduced ISP so that it can deliver an electron to cytochrome c₁ by the tethered diffusion mechanism we have previously suggested.

Acknowledgments

Reference

1. Gennis, R.B., Barquera, B., Hacker, B., van Doren, S.R., Arnaud, S., Crofts, A.R., Davidson, E., Gray, K.A. and Daldal, F. (1993) J. Bioenerg. Biomembr. 25, 195-210.
2. Brandt, U. and Trumpower, B.L. (1994) CRC Crit. Rev. Biochem. 29, 165-197.
3. Schägger, H., Brandt, U., Gencic, S. and von Jagow, G. (1995) Methods Enzymol. 260, 82-96.
4. Mitchell, P. (1976) J. Theor. Biol. 62, 327-367
5. Crofts, A. R., Meinhardt, S. W., Jones, K. R. and Snozzi, M. (1983) Biochim. Biophys. Acta, 723, 202-218
6. Glaser, E.G. and Crofts, A.R. (1984). Biochim. Biophys. Acta, 766, 322-333.
7. Crofts,  A.  R.  (1985). In: The Enzymes of  Biological  Membranes, (Martonosi, A.N., ed.), Vol. 4, pp. 347-382, Plenum Publ. Corp., New York.
8. Robertson, D.E. and Dutton, P.L. (1988) Biochim. Biophys. Acta 935, 273-291.
9. Crofts,  A.R. and Wang, Z. (1989) Photosynth. Res. 22, 69-87
10. Bowyer, J. R., Dutton, P. L., Prince, R. C., & Crofts, A. R. (1980) Biochim. Biophys. Acta 592, 445-460.
11. Meinhardt,  S.W. and Crofts, A.R. (1982). FEBS Lett., 149, 217-222.
12. Link, T.A., Haase, U., Brandt, U. and von Jagow, G. (1993) J. Bioenerg. Biomemb. 25, 221-232
13. von Jagow, G. and Link, T.A. (1986) Methods Enzymol. 126, 253-271
14. Crofts, A. R. and Wraight, C. A. (1983). Biochim. Biophys. Acta, 726, 149-186.
15. Iwata, S., Saynovits, M., Link, T.A. and Michel, H. (1996) Structure 1996, 4: 567-579.
16. Xia, D., Yu, C.-A., Deisenhofer, J., Xia, J.-Z. and Yu, L. (1996) Abstracts. Biophys. Soc. Meeting, Baltimore, Feb. 1996.
17. Xia, D., Kim, H., Deisenhofer, J., Yu, C.-A., Xia, J.-Z. and Yu, L. (1996) Abstracts, #SO355. XVII Meeting, Intntl. Union Crystallography, Seattle.
18. Kim, H., Xia, D., Deisenhofer, J., Yu, C.-A., Kachurin, A. and Yu, L. (1996) Abstracts, #SO356. XVII Meeting, Intntl. Union Crystallography, Seattle.
19. Yu, C.-A., Xia, J.-Z., Kachurin, A.M., Yu, L., Xia, D., Kim, H. and Deisenhofer, J. (1996) Biochim. Biophys. Acta 1275, 47-53.
20. Berry, E.A., Zhang, Z., Huang, L.-S., Chi, Y.-I., and Kim, S.-H. (1997) Abstracts, Biophy. Soc. Ann. Meeting, New Orleans, March 1997, Tu-PM-G4
21. Berry, E.A., Shulmeister, V.M., Huang, L.S. and Kim, S. H. (1995) Acta Crystallographica 51, 235-239.
22. Zhang, Z., Huang, L., Chi, Y.-I., Kim, K.K., Crofts, A.R., Berry, E.A., and Kim, S.-H. (1997) Manuscript in preparation.
23. Xia, D., Yu, C.-A., Kim, H., Xia, J.-Z., Kachurin, A.M., Zhang, L., Yu, L., Deisenhofer, J. (1997) Science, 277, 60-66
24. Yun, C.-H., Beci, R., Crofts, A.R., Kaplan, S. and Gennis, R.B. (1990) Europ. J. Biochem. 194, 399-411
25. Yun, C.-H., Crofts, A.R. and Gennis, R.B. (1991) Biochemistry, 30, 6747-6754
26. Konishi, K., Van Doren, S.R., Kramer, D.M., Crofts, A.R., and Gennis, R.B. (1991)  J. Biol. Chem. 266, 14270-14276
27. Brasseur, G., Sami Saribas, A. and Daldal, F. (1996) Biochim. Biophys. Acta 1275, 61-69.
28. Crofts, A.R., Barquera, B., Bechmann, G., Guergova, M, Salcedo-Hernandez, R., Hacker, B., Hong, S. and Gennis, R.B. (1995) In Photosynthesis: from light to biosphere. (Mathis, P., ed.), Vol. II, pp. 493-500. Kluwer Academic Publ., Dordrecht.
29. Crofts, A.R., Hacker, B., Barquera, B., Yun, C.-H. and Gennis, R. (1992) Biochim. Biophys. Acta 1101, 162-165
30. Degli Esposti, M., De Vries, S., Crimi, M., Ghelli, A., Patarnello, T., and Meyer, A. (1993) Biochim. Biophys. Acta 1143, 24-271
31. Berry, E.A., Zhang, Z., Huang, L.-S., Kuras, R. , Guergova-Kuras, M. and Crofts, A.R. (1997) Abstract, Meeting on "Reaction centers of photosynthetic purple bacteria: structure, spectroscopy, dynamics", Cadarache, June 1997
32. Meinhardt,  S.W.  and Crofts, A.R. (1982). FEBS Lett. 149, 223-227.
33. Matsuura, K., Bowyer, J.R., Ohnishi, T., & Dutton, P.L. (1983) J. Biol. Chem. 258, 1571-1579.
34. Ding, H., Robertson, D.E., Daldal, F. and Dutton, P.L. (1992) Biochemistry 31, 3144-3158.
35. Ding, H., Moser, C.C., Robertson, D.E., Tokito, M.K., Daldal, F. and Dutton, P.L. (1995) Biochemistry 34, 15979-15996.
36. Glaser, E.G., Meinhardt, S.W. and Crofts, A.R. (1984) FEBS Lett. 178, 336-342
37. Brandt, U. (1996) Biochim. Biophys. Acta 1275, 41-46.
38. Siedow, J.N., Power, S., De La Rosa, F.F. and Palmer, G. (1978) J. Biol. Chem. 253, 2392-2399
39. De Vries, S., Albracht, S.P.J., Berden, J. and Slater, E.C. (1982) Biochim. Biophys. Acta 681, 41-53
40. Ding, H., Daldal, F. and Dutton, P.L. (1995) Biochemistry 34, 15997-16003
41. Stowell, M. H. B., McPhillips, T. M., Rees, D. C., Soltis, S. M., Abresch, E. and Feher, G. (1997) Science, 276, 812-815
42. Crofts, A.R., Guergova-Kuras, M., Hong, S. and Bechmann, G. (1997) Photosynth. Research, in press.