STRUCTURE AND FUNCTION OF THE BC1 COMPLEX OF RHODOBACTER SPHAEROIDES


Antony Crofts, Beth Hacker, Blanca Barquera, Chang-Hyon Yun and Robert Gennis, Departments of Physiology and Biophysics, Biochemistry and Chemistry, University of Illinois, Urbana, IL 61801, USA.

INTRODUCTION

The ubiquinol:cytochrome c2 oxidoreductase (bc-complex) of Rb. sphaeroides has three main subunits which bear the prosthetic groups, and contribute to three catalytic sites and internal electron transfer pathways which define the modified Q-cycle mechanism. In this paper, we report on progress in modelling the structure of the bc-complex, and experiments using site directed mutagenesis and biophysical assay to probe the structural and function consequences of specific modifications to these subunits.

MODEL OF CYTOCHROME B.

Fig. 1 shows a preliminary model of the cytochrome b subunit. The structure shows six transmembrane helices, A-F (out of eight predicted), two amphipathic helices, ab and cd, the two heme groups of cytochromes (cyt) bh and bl, and two ubiquinone molecules identifying the two quinone reactive catalytic sites. Residues near the putative QR-site discussed below are high-lighted. In Fig. 2, the six transmembrane helices of the model are shown in a view from the cytoplasmic side, as helical wheels arranged about the hemes. The thickness of the lines reflects the degree of conservation (see below). The following constraints were used in building the model:

  • 1) The primary sequence as previously reported [1].
  • 2) Conserved hydrophobic spans which identify putative membrane spanning helices [2-4]. These spans are also identified on the basis of profiles using probability parameters from the distribution of residues in known or predicted membrane helices [5], or in buried helices identified in known structures in the Brookhaven Protein Data Bank [6].
  • 3) Spans showing conserved amphipathy at the helical repeat (hydrophobic moment) [4,7-9]. The identification of helix cd as an amphipathic helix was of importance in suggesting removal of this span from the membrane, to convert the original 9-helix model [2,3] to a structure containing 8 transmembrane helices, A-H (4,9). Amphipathy profiles identify spans at the N-terminal end (a helix), between helices A and B (ab helix), C and D (cd helix), and possibly between E and F (ef helix).
  • 4) A pattern of conservation of residues suggesting a helical motif (mutability moment) [9-12]. The mutability moment shows a vector such that in membrane helices at the lipid-protein interface, the unconserved face is the more hydrophobic face, and faces the lipid. In more amphipathic helices suggested to be at the protein water interface, the conserved face is the more hydrophobic face. This complementary pattern of mutability and hydrophobic moment provides a strong indication of the orientation of these helices at the protein-lipid or protein-water interface. Helices A and ab show strong mutabilty moments, and several of the membrane helices show strong moments for both hydrophobicity and mutability at their ends.
  • 5) The topological organization with eight transmembrane spans has been strongly supported by our phoA fusion experiments [13].
  • 6) Four conserved histidines, two each in helices B and D, which are spaced 14 residues apart so as to fall on the same side of the helix, have been suggested as likely ligands for the two heme groups [2,3]. Site-directed mutagenesis of these residues has demonstrated which histidines ligate which hemes [14].
  • 7) Four conserved glycines, 2 each in helices A and C, spaced, like the liganding histidines, 14 residues apart, are suggested to accommodate the packing needs of the hemes [15]. In support of this suggestion, in helix A, which shows a strong mutability moment, and in helix C, the glycines are on the conserved face of the helix (Fig. 1, 2).
  • 8) Inhibitor resistance mutations. An important feature leading to support of the 8-helix model was the mapping of lesions giving rise to resistance to inhibitors at the two quinone processing catalytic sites. In reaction centers from several bacteria, and photosystem II of green plants and algae, mutations giving rise to resistance to herbicides have all mapped to the QB-site (where quinone is reduced), which is blocked by these inhibitors. As we have previously pointed out [4], this pattern could be used to infer that similar mutations in cyt b would identify residues which contribute to the catalytic sites at which the affected inhibitors bind. In the 8-helix model, residue changes conferring resistance to inhibitors (diuron, antimycin, HQNO) acting at the QR- (quinone reducing) site all fall on the N- (protochemically Negative) side identified by the location of the heme of cytochrome (cyt) bh, while those giving resistance to inhibitors (stigmatellin, myxothiazol, mucidin) acting at the QO- (quinol oxidizing) site fall on the P-side (16-20), close to cyt bl.
  • 9) Specific mutagenesis.
  • Work on mutagenesis of residues thought to contribute to the QO-site in Rb. capsulatus by Daldal's group (20), and characterization of these mutant strains by Robertson, Dutton and colleagues (21) has been of importance in defining this site. Our own work exploring both QO- and QR-sites in Rb. sphaeroides is summarized in Table I and discussed below. We have reported similar work on mutagenesis of conserved residues elsewhere [14,22].

    Fig. 1. Cytochrome b subunit. The model is represented as a stereo pair for viewing by the crossed eye method. Transmembrane helices A-F, and amphipathic helices ab and cd are shown. Numbered residue are discussed in text. Top is cytoplasmic side.

    Fig. 2. Helical wheel representation of model, viewed from cytoplasmic side. Only transmem-brane helices A-F are shown.

    SITE DIRECTED MUTAGENESIS OF RESIDUES AT PREDICTED CATALYTIC SITES

    MUTATIONS AT THE PUTATIVE QR-SITE

    No stable spontaneous mutations giving rise to resistance at this site have been reported in the photosynthetic bacteria. In Table I the redox properties of strains with mutations constructed around the putative QR-site are summarized. The mutations of A52V, K251I and K251M mimic spontaneous mutations in yeast found to confer antimycin or diuron resistance. In the presence of saturating concentrations of antimycin, all the mutant strains showed reduction of cyt bh with kinetics and dependence on [QH2] similar to wild type, indicating a normal QO-site. In the absence of antimycin, all mutants showed a slower oxidation of cyt bh, as seen in the kinetics of cyt bh turn-over or the antimycin sensitive phase of the carotenoid change (Figs. 3, 4). The K251I and A52V mutations showed the least effect on both kinetics and antimycin titre. In strain K251M, the reduced rate of oxidation gave rise to a marked transient reduction of cyt bh. Strain H217A, and both mutations at D252 showed a much more marked block, and behaved as if constitutively inhibited by antimycin. In D252N, a slow antimycin sensitive turn-over could be detected from electrochromic measurements, but D252A and H217A were more strongly affected. In antimycin titrations measured from the kinetics of cyt bh reduction, strains K251M and A52V show on obvious resistance, with a higher titre, and a less sharp end-point, both indicative of a weaker binding. In the strains with a strongly inhibited cyt bh oxidation, antimycin had so small an effect as to make the titrations ambiguous. All strains except D252A and H217A could grow photosynthetically. In all the above six strains, redox titrations showed the higher potential form of this cytochrome (cyt b150,- Em,7 =150 mV in the absence of antimycin). Addition of antimycin induced oxidation of cyt b150, even in those strains with strongly inhibited turn-over at the QR-site. This provided an alternative method of assaying the antimycin binding (Fig. 5). Strains H217A and K251I showed end-points for the titration indistinguishable from the wild-type strain (BC17C); all other strains showed some degree of resistance. Antimycin induced oxidation showed the largest amplitude in strains H217A, K251M, D252A and D252N, and was also larger in K251I than in WT.

    Table I

    Mutations constructed at predicted quinol oxidizing (Qo) and quinone reducing (Qr) sites

    Mutations at the QR-site.

    Residue cyt bl cyt bh Photo. 	Characteristics 
    G48A 	-90 	40 	+++ 	Turn-over indistinguishable from WT.
    G48V 	-90 	23 	 - 	Cyt bh oxidation blocked, Qo-site normal.
    G48D 	none 	none	 - 	Complex not assembled, but trace of FeS center.
    A52V 	-90 	32 	++ 	Qi-site slowed by 5, antimycin resistance.
    A52D 	none 	none 	 - 	Complex not assembled
    H217A 	-40 	50 	 - 	Complexed assembled, cyt bh oxidation blocked
    F244L 	none 	none 	 - 	Complex not assembled
    K251M 	-59 	47 	++ 	Qi-site slowed by 3, antimycin res., normal Qo-site.
    K251I 	-71 	42 	++ 	Qi-site slowed by 2, normal Qo-site.
    D252N 	-59 	68 	++ 	Cyt bh oxidation inhibited, normal at Qo-site.
    D252A 	-50 	48 	 - 	Cyt bh oxidation blocked, normal Q0-site.
    Mutations at the QO-site.
    Residue cyt bl cyt bh Photo. Characteristics 
    F144S 	-90 	50 	+++ 	Qo-site inhibited by 3, myxothiazol resistant
    N279Y 	n.d 	n.d 	++ 	Qo-site inhibited by 10
    E296D 	-89 	75 	+++ 	Qo-site inhibited by 2
    E295G 	-62 	44 	+++ 	Qo-site inhibited by 9
    E295Q 	-30 	64 	++ 	Qo-site inhibited by 50
    W296F 	-20 	60 	+++ 	Qo-site weakly inhibited
    W296L 	-19 	58 	++ 	Qo-site inhibited by 2
    Y297F 	-92 	60 	+++ 	Qo-site weakly inhibited
    Y297S 	-66 	46 	++ 	Qo-site inhibited by 25
    

    Interpretation of these results depends on an understanding of the mechanism of quinone reduction at the QR-site, and this at present is very incomplete. We have proposed that the antimycin-induced oxidation of cytochrom b150 involves a displacement of equilibria at the QR-site summarized below, where E.bH represents the QR-site with associated cytochrome bh:

    				QH2 Q
    E.b150-.Q.- + 2H+ === E.bH.QH2 == E.bH === E.bH.Q
    				AA ||
    				E.bH.AA
    

    The reactions indicated by the horizontal equilibria represent partial reactions in the turn-over of the two-electron gate thought to operate at the QR-site. The two-electron reduced state of the low potential chain of the complex (E.b150-.Q.-) can be reached either by reduction of the oxidized complex following binding of quinol at the QR-site (reactions proceeding to the left in the above scheme), or by electron transfer into the b-cytochrome chain following oxidation of 2QH2 at the QO-site:

    				  E.bL.bH.Q.-
    				    ||
    E.bL.bH.Q   ==  E.bL-.bH.Q   ==   E.bL.bH-.Q  ==  E.bL-.bH-.Q  ==  E.b150-.Q.-
    QH2 + 2Fe.2S+   Q + 2Fe.2S 	QH2 + 2Fe.2S+     Q + 2Fe.2S
    

    In the context of this mechanism, antimycin induced oxidation of cytochrome b150 occurs when antimycin is added to the complex in the two-electron reduced state, and reflects that fraction of complexes in which an electron resides initially on both cytochrome b150- and Q.­. The greater amplitude of the change in some strains would then indicate a greater stability of this state, due either to a stronger binding of QH2 or a change in the equilibrium constant, favoring a more stable semiquinone, for the electron transfer reaction from bound QH2. However, the mechanism also implies that electron transfer from cytochrome bh to the QR-site in the reaction: E.b150-.Q.- + 2H+ == E.bH.QH2 still occurs even in those strains with a severely inhibited cytochrome bh oxidation.

    Fig. 3. Kinetics of cytochrome bh reduction without and with antimycin. Chromatophores (approx. 30 µM BChl/ml) were suspended in 100 mM KCl, 50 mM MOPS at pH 7, 2 µM valinomycin and nigericin, a cocktail of redox mediators, and poised at Eh 100 mV. Kinetics were followed at 561-569 nm after a single flash. Left: No antimycin; right: antimycin added to saturation.

    Fig. 4. Antimycin-sensitive electrogenic reactions measured through the carotenoid change at 503 nm. Conditions as for Fig. 3, except that valinomycin and nigericin were omitted. Traces are subtractions of the change measured with no addition, and the change after antimycin was added to saturation.

    An additional effect also seen in the strains showing a marked antimycin induced oxidation was the stable reduction of cyt bh following flash activation in the absence of antimycin when chromatophores were poised at higher Eh (200 mV) where the pool was oxidized before flash activation. In order to explain this effect in terms of a more strongly stabilized semiquinone, it would be necessary to postulate that the two-electron reduced state of the low potential chain was reached either by a double turn over at the QO-site, or by a preferential oxidation of QH2 at the QR-site.

    MUTATIONS IN THE -PEWY- REGION

    Although much work has been described on inhibitor resistance of spontaneous mutants, more detailed characterization of the effects on kinetic parameters has only been reported for mutants in photosynthetic bacteria (1,14,20-22). We have extended our own work in this area to cover mutations in the loop connecting E and F helices which contains the highly conserved -PEWY- span (Table I). Preliminary characterization of these strains shows that non-conservative changes in the PEWY region lead to modifications of quinol oxidation, but do not markedly effect the binding of inhibitors specific to the QO-site. The kinetics of cytochrome bh reduction in the presence of antimycin, which reflects the rate of QH2 oxidation at the QO-site, are shown for several mutant strains in Fig.6. The conservative mutations E295D, W296F and Y297F showed kinetics most similar to the wild type (BC17C), but the non-conservative

    mutations showed a marked reduction in rate of quiol oxidation. Similar experiments at different

    ambient redox potentials to modulate the degree of reduction of the pool were performed to determine the apparent Km for QH2, and the results are shown in Fig. 7. All strains showed an increase in rate with reduction of the pool, suggesting that the underlying mechanism was unchanged. In strains E295D and W296F, the main effect seemed to be on Vmax, and Km was little effected. With Y297F, both Vmax and Km were effected. However, for the serverely disfunctional strains, the limited range over which the [QH2] could be varied by redox titration precluded more detailed kinetic analysis. The effects on quinol oxidation suggest a role for these residues in catalysis or structure. Surprisingly, none of the mutants was completely devoid of activity, and none failed to assemble.

    Fig. 5. Titration of antimycin induced oxidation of cytochrome b150. Spectra were taken of chromatophores (~150 µM BChl/ml) in the medium of Fig. 3. Difference spectra (inset, Eh 100 - 200 mV) show change on adding antimycin. In the Fig., the amplitude of the change is plotted against concentration of antimycin. Amplitude was normalized to total cyt bh measured from difference at Eh -20 - 200 mV.

    Fig. 6. Kinetics of cyt bh reduction in strains mutated in the -PEWY- span. Conditions as for Fig. 3, with 10 µM antimycin added. Eh was poised at 100 mV.

    Fig. 7. Rate of cyt bh reduction as a function of [QH2]. Rates were measured from experiments similar to those of Fig. 6, except that Eh was varied over the range 200 - 90 mV. [QH2] was calculated assuming a value of 60 (Q + QH2) / bc1 complex [23], using the Nernst-Peters equation.

    DISCUSSION

    The constraints on the model discussed above provide severe limitations on the general topology of the cyt b subunit. For six of the transmembrane helices, A - F, choice among alternative arrangements for packing the structure around the liganding helices B and D is strongly dependent on the weight attached to the suggested role of the conserved glycines ([15] and item 7 above). Tertiary models of the helices showing distribution of the inhibitor resistance lesions define the catalytic sites, but with reference to the packing of helices A and C against the hemes. Support for the orientation shown comes from our characterization of strains with specific mutations of G48 (G33 in yeast) (see Table I), which show that only residues of small volume are tolerated, and from the L198F mutation in yeast which confers funiculosin resistance. This residue is next to H197, which is one of the ligands to cyt bh, and its position is therefore define with respect to the heme planes. We are presently exploring the roles of the other conserved glycines, I213 (equivalent to L198 in the Rb. sphaeroides sequence), and other residues shown as projecting into the volume of the QO- and QR-sites in the model.

    All the mutations constructed around the putative QR-site showed effects on cyt bh oxidation and/or antimycin binding. This provides some reassurance that our modelling of the structure has correctly identified these residues as close to the binding pocket. None of the mutants showed any marked change in affinity for antimycin (cf. the 3000 fold change in affinity for atrazine in the S264G mutant of Amaranthus hybridus), and it seems unlikely that any of these residue is an essential ligand. It is interesting to note that on the model of Fig. 2, H217 and D252 are neighbors; possibly they represent a counter-ion pair, and this could be readily tested by further mutations.

    The mutations constructed in the -PEWY- region likewise provide some reassurance about the model. Although none of the residues modified has been previously identified in inhibitor resistance strains, several resistance lesions have been found on either side of this span, suggesting that it must be close to the QO-site. The finding that the main effects of mutation in this conserved span are on the kinetics of QH2 oxidation is consistent with this picture.

    While it is premature to speculate about details of function or mechanism in the context of this preliminary structure, there are some possibly interesting motifs. At both sites, the quinone binding pocket is formed between membrane spanning helices, close to the interface between lipid and aqueous phases. At the QO-site, a well defined amphipathic helix (cd) contains residues which when modified confer resistance to inhibitors at this site, and this is shown in the model as "capping" the binding pocket. At the QR-site, a conserved span showing helical amphipathy (helix a) is found to the N-terminal side of a residue associated with diuron resistance in yeast. This could well "cap" the other binding domain. In both cases, the overall structure would then be very similar to that of the quinone binding pockets of the bacterial reaction centers.

    REFERENCES

    1. Yun, C.-H., Beci, R., Crofts, A. R., Kaplan, S. and Gennis, R. B. (1990) Eur. J. Biochem. 194, 399-411.
    2. Widger, W.R., Cramer, W.A., Herrmann, R.G. and Trebst, A. (1984) Proc. Natl. Acad. Sci. 81, 674-678.
    3. Saraste, M. (1984) FEBS Lett. 166, 367-372.
    4. Crofts, A.R., Robinson, H.H., Andrews, K., Van Doren, S. and Berry, E. (1987) In Cytochrome Systems: Molecular Biology and Bioenergetics (Papa, S., Chance, B. and Ernster, L., eds.) pp. 617-624, Plenum Publ., New York.
    5. Rao, J.K. and Argos, P. (1986) Biochim. Biophys. Acta, 869, 197-214.
    6. Walsh, L.L., Bobak, M. and Crofts, A.R. (1990) 4th. Symp. Protein Soc. Abstract.
    7. Eisenberg, D. (1984) Ann. Rev. Biochem. 53, 595-523
    8. Cornette, J.L., Cease, K.B., Margalit, H., Spouge, J.L., Berzofsky, J.A. and DeLisi, C. (1987) J. Mol. Biol. 195, 659-685
    9. Crofts, A.R., Wang, Z., Chen, Y., Mahalingham, S., Yun, C.-H. and Gennis, R.B. (1990) in Highlights in Ubiquinone Research (Lenaz, G., Barnabei, O, Rabbi, A. and Battino, M., eds.) pp. 98-103, Taylor & Francis, London, New York, Philadelphia.
    10. Komiya, H., Yeates, T.O., Rees, D.C., Allen, J.P. and Feher, G. (1988) Proc. Natl. Acad. Sci. USA, 85, 9012-9016
    11. Rees, D.C., DeAntonio, L. and Eisenberg, D. (1989) Science, 245, 510-513
    12. Crofts, A.R., Yun, C.-H., Gennis, R.B. and Mahalingham, S. (1989) in Proceedings of the VIIIth. Internatl. Cong. Photosynth., Stockholm, in Press.
    13. Yun, C.-H., Van Doren, S.R., Crofts, A.R., and Gennis, R.B. (1991) J. Biol. Chem. 266, 10967-10973.
    14. Yun, C.-H., Crofts, A.R. and Gennis, R.B. (1991) Biochemistry. In press.
    15. Tron, T., Crimi, M., Colson, A.-M. and Degli Esposti, M. (1991) Eur. J. Biochem., in press
    16. di Rago, J.-P., Perea, X. and Colson, A.-M. (1986) FEBS Lett., 208, 208-210.
    17. Colson, A.-M., Meunier, B. and di Rago, J.-P. (1987) in Cytochrome Systems (Papa, S., Chance, B. and Ernster, L., Eds.), pp. 135-136, Plenum Publishing Corp.
    18. Howell, N. & Gilbert, K. (1988) J. Mol. Biol. 203, 607-618.
    19. di Rago, J. P. & Colson, A.-M. (1988) J. Biol. Chem. 263, 12564-12570.
    20. Daldal, F., Tokito, M.K. and Davidson, E. (1989) The EMBO J. 8, 3951-3964
    21. Robertson, D.E., Daldal, F. and Dutton, PL. (1990) Biophys. J. 58, 11249-11260
    22. Yun, C.-H., Wang, Z., Crofts, A.R. and Gennis, R.B. (1991) J. Biol. Chem. In press.
    23. Crofts, A. R. (1985). In: The Enzymes of Biological Membranes, (Martonosi, A.N., ed.), Vol. 4, pp. 347-382, Plenum Publ. Corp., New York.