Structure of the bc1-complex from X-ray crystallography:


An atomic structure for the mitochonrial bc1-complex has recently been announced (1, 2, 3, and Chang-An Yu, personal communication). At present, the information available relates to the general shape of the dimeric complex, with specific information about distances between prosthetic groups. No detailed information about the protein structure has yet been released. The current data confirm some of the main features of our structural model of cytochrome b, and are in agreement with a structure from electron microscopy previously suggested by Weiss and colleages (4, 5, and see below), but the X-ray structure has some surprises.

General features

The disposition of the protein is as expected from biochemical and previous structural studies, with a large fraction corresponding to the "core" proteins (subunits I and II) on the N-side, and cyt c1 and the FeS protein on the P-side.
The complex is dimeric, showing a two-fold symetry about an axis vertical to the membrane plane. The dimers show a tight association, making it unlikely that monomers would have an unperturbed function.
The dimensions of the dimer are about 130 Å in diameter and 151 Å in height, with the inter-membrane space region, the transmembrane region, and the matrix region 41 Å, 35 Å and 75 Å respectively (2).
These properties confirm the model suggested from electron microscopy studies by Weiss and colleagues (4, 5).

Prosthetic groups

The prosthetic group composition is as expected from biochemical studies: cyt bL, cyt bH, cyt c1, 2Fe.2S center.
Cyt bL --> cyt bH distance is 20 Å (13 Å edge to edge), perpendicular to membrane. Assuming that cyt bH is close to the antimycin binding site (29, and see below), the arrangement in the protein is as expected from our model.
The cyt bL hemes of the two monomers are 21 Å apart in the dimer, the cyt bH hemes are 33 Å apart, the FeS centers are 63 Å apart, and the cyt c1 hemes are 53 Å apart.
Cyt L --> 2Fe.2S distance is 26 Å.
2Fe.2S --> cyt c1 distance is 31 Å

These distances are summarized schematically here. The diagram is approximately to scale, but the rotations between the several different planes provided by the current data, cannot be represented in such a 2-D scheme.

Spatial considerations

The cyt bL hemes are nearest to the interface between monomers, and are quite close together; the 21 Å distance would indicate a minimal amount of protein (with ~10 Å required between helix centers) between the hemes. Sequence alignment and probability analysis suggest a well conserved face along the exterior sides of helices C and D in our model, and a somewhat conserved face along the exterior sides of helices B and F. Either of these could possibly indicate packing along the interface, with the surface provided by helices C and D providing the most plausible interface. Alternatively, the small distances could indicate that the structure of our model is wrong in the region corresponding to the Qo-site, and that the structure is more open to allow some specific interaction at the inter-heme site.
Because the cyt bH hemes are further apart then the cyt bL hemes, it seems likely that the the monomers lie at an angle to each other, with rotation about an axis along the line between the two bL hemes. Such a rotation is apparent in the Weiss structure (4, 5). A cartoon representation shows a possible distribution of prosthetic groups from the Deisenhofer data superimposed on the Weiss model.

The biggests surprise is the distance between the 2Fe.2S center and cyt c1. The kinetic data suggest that this reaction is rapid, with t½ < 10 µs. The distance of 31 Å would not permit this rate; further resolution of this paradox must await more detailed information about the protein structure.

There are 13 transmembrane helices apparent in the X-ray structure (2). Assuming that 8 of these belong to cyt b (SUIII)(as in our model), and that one belongs to cyt c1 (SUIV), the remaining four helices must be assigned to the remaining subunits (I, II, FeS, plus the small nuclear-encoded subunits). Neither of the "core" proteins (SUI, SUII) show obvious hydrophobic spans of sufficient length, though SUI has span between 99 and 118 which shows hydrophobic character, and a strong amphipathy; the many short hydrophobic spans look likely candidates for buried sheet. SUX and SUXI have plausible hydrophobic spans, and probably account for two of the four; SUVII has a strong hydrophobic span of length appropriate for a transsmembrane helix, but with -PP- in the center, which would not favor a helical configuration. SUV (the FeS protein) has a moderately hydrophobic span flanked by spans showing strong amphipathy at the helical repeat, which may indicate a transmembrane span with ends projecting from the membrane. SUIX has a hydrophobic span, but not long enough or hydrophobic enough to be a strong candidate.

The present set provides diffraction data to 3.2 A; a native data set, and sets with inhibitors (antimycin or UHDBT) bound have been solved.

Inhibitor binding

The antimycin binding site is close to a b-heme, assumed to be cyt bH. The difference electron density map ± antimycin shows a strongly defined density for the inhibitor, and a loss of density which probably corresponds to the displaced ubiquinone.

The UHDBT binding site is close to the 2Fe.2S center, on the inner side between the center and the dimer interface, and therefore between the center and cyt bL.

The present crystals diffract to 2.8 - 2.9 Å, an improved set has been collected to this resolution after a visit to Grenoble synchrotronn, and is currently under refinement.
In the Deisenhofer crystals, the definition of structure in the "core" subunits is excellent, good in the membrane spanning region, and somewhat less ordered in the volume occupied by cyt c1 and FeS proteins (J. Deisenhofer, personal communication, mid May 1996).

Other crystallographic studies

The structure from Xia et al. is the only full structure of the bc1 complex currently reported. Similar studies are in progress in several other labs, and are expected to yield structures of similar or better resolution in the near future (6-10). Smith, Cramer and colleagues have previously reported a high resolution structure for a solubilized cyt f of the b6f-complex (11). Michel, Link and colleagues have recently reported on a high resolution structure for the solubilized FeS protein from the beef heart complex (12-14).

Note added after b6f-complex meeting in Regensburg

The Xia/Yu/Deisenhofer structure was still "in progress", with much of the chain tracing still to be completed. They have a 2.6 A data set now, and expect to have a complete structure from that during the next 1-2 months. There were no protein details, and incomplete CA-only backbones for the main helical segments. The distances are as in the previously available abstracts, but a new big surprise is the orientation of cyt bL, which they show as nearly flat in the membrane plane. This is a surprise because the spectroscopic data (16-19) seemed to indicated unambiguously that the two hemes of cyt b were more or less vertical. The other new information comes from the myxothiazol binding sites (more or less between, and equidistant from, the centers of bL and FeS), and some more distances. The antimycin binding sites are 33 A apart (a little more than the bH Fes); the bL to c1 distance was 33-35 A; the angle between c1, FeS and bL is 66 degrees. Of the 13 transmembrane helices, 8 are ascribed to cyt b, 1 to cyt c1, and 1 (at least) to Fes (as predicted by modeling, and the biochemical properties of the native subunit, and that from a mutant strain with the putative helical anchor removed); core I (SUI) contributes 1, and the other 2 come from the small subunits. One of the small subunits (I think XI) is missing from the structure.
An approximate model of the prosthetic groups based on the data available is shown in this Fig.
Because the Xia/Yu/Deisenhofer structure is incomplete, it wasn't possible to fit the Iwata/Link/Michel structure into the complete complex.

Recent news

Ed Berry has solved the phasing problem in his crystals of the chicken-heart complex (see Ed's WWW-pages), and has made significant progress towards a structure. Data has been resolved sufficiently to show -helices, and prosthetic groups. He has similar iron-iron distances to the Xia /Yu /Deisenhofer structure except for FeS-cyt c1 and FeS-cyt bL distances, where the former are closer, and the latter more distant.

Colors are:
Monomer 1: cyt bH, light green; cyt bL, dark blue; FeS, orange; cyt c1, blue-green
Monomer 2: cyt bH, pale blue; cyt bL,blue; FeS, red; cyt c1, yellow-green

View of a Model Structure (Xia/Yu/Deisenhofer information)

You will need Netscape 2.0 or higher, and Chemscape Chime to see the model within your viewer


Colors are:
Monomer 1: cyt bH, light green; cyt bL, pale blue; FeS, red; cyt c1, yellow; Qi, light green; Qo, light blue
Monomer 2: cyt bH, green; cyt bL,blue; FeS, orange; cyt c1, green; Qi, yellow green; Qo, blue

Note that the coordinate positions of the Fe atoms are based on data available from published abstracts, information from symposium presentations, not on crystallographic data (see also 15). Although the positions shown are approximately in agreement with the data, the true positions and orientations may differ significantly. The coordinates of non-Fe atoms, and orientation of rings, etc., is speculative.

You can download Chime by clicking here. Chime plug-in

Late breaking news

Hans Deisenhofer/Tony Crofts, -by phone 8/1/96

Structure is progressing well for the "core" subunits and for cytochrome b, but resolution of structure for FeS and Cyt c1 is still problematic. The electron densities in the well resolved subunits have been fitted to the sequence.

Density around the bL hemes is better fitted by horizontal hemes, but a perpendicular arrangement now seems more likely. They can't fit amino acid side chains to the model with horizontal hemes, but can if the model has hemes perpendicular. Heme ligands are provided by 4 histidines for the perpendicular model; other aspects of the structure (sequence of helices around hemes) are looking more like our model.

Hans now thinks that the "horizontal" hemes might be an artifact of crystallography.

The distances for irons previously provided in abstracts and EBEC paper are probably O.K.,- signals from irons are high, so locations are probably correct for crystals. However, although crystals form an active complex when dissolved, the Rieske protein may have an inactive configuration in this crystal form. As the RISP is not involved in crystal contacts, crystal packing forces cannot be held responsible for this.

Probably the poor resolution of electron density in FeS and c1 indicates some disorder in their structure. This is also consistent with the weak signal from the FeS irons. Also consistent with some disorder is the fact that they can not yet identify the Rieske protein structure provided by Michel/Link coordinates in their electron density.

Hans believes that the form in the crystal might be an informative structure, and that the protein may adopt different configurations. It's possible that the crystals show one configuration selected from several that may be involved in the normal catalytic cycle. This would suggest a considerable conformational change in catalysis, with a rotation of the Rieske protein as an important feature.

Latest news

Abstract, Meeting on "Reaction centers of photosynthetic purple bacteria: structure, spectroscopy, dynamics", Cadarche, June 1997.

On the mechanism of ubiquinol oxidation by the bc1-complex: the role of the Rieske subunit, and its mobility.


Edward A. Berry#, Zhaolei Zhang#, LiShar Huang#, Richard Kuras*, Mariana Guergova-Kuras* and Antony R. Crofts*, Center for Biophysics and Computational Biology, U. of Illinois at Urbana-Champaign*, and Lawrence Berkeley National Laboratory, U. C. Berkeley#.
The positions of Fe-centers in the mitocondrial bc1-complex of beef heart have been made available by Xia, Deisenhofer, Yu and colleagues. We have compared these positions with those from our own work on the complex from chicken heart, beef heart (in hexagonal crystals) and from rabbit. We have noted that, although the heme Fe-atoms appear to occupy equivalent positions in all the structures, the positions of the 2Fe2S centers of the Rieske subunit (ISP, or iron sulfur protein) are different. There is a displacement of ~10 between the position in beef enzyme as solved by Xia et al., and that in the chicken, the rabbit, and the hexagonal form of the beef enzyme. We suggest that the different ISP conformations reflect a displacement of the subunit which is a necessary part of the catalytic mechanism of the Qo-site of the complex. We note that in none of the crystal forms is the complex in a form in which the static configuration could be competent in all partial reactions of quinol oxidation. However, in each of the two main positions, the configuration could function for a part of the reaction cycle. We suggest that the ISP must move between the positions seen in different crystal forms. In confirmation of this possibility, we have observed that in the chicken enzyme, stigmatellin induces a displacement of the ISP from its position close to cyt c1, to a position close to that inferred for the Xia et al. structure. The ISP binds at a concave interface on cyt b, which brings the 2Fe2S center close to a buried myxothiazol binding site (as observed from the electron density difference in myxothiazol-containing crystals), which is likely the quinol binding pocket. We discuss the mechanism of quinol oxidation in the context of such a movement. We note that movement of the ISP as a necessary part of turn-over throws new light on many anomalous points from the literature.
  1. Xia, D., Yu, C.-A., Deisenhofer, J., Xia, J.-Z. and Yu, L. (1996) Abstracts. Biophys. Soc. Meeting, Baltimore, Feb. 1996.
  2. Xia, D., Kim, H., Deisenhofer, J., Yu, C.-A., Xia, J.-Z. and Yu, L. (1996) Crystal structure of beef heart mitochondrial bc1-complex. Abstracts, #SO355. XVII Meeting, Intntl. Union Crystallography, Seattle.
  3. Kim, H., Xia, D., Deisenhofer, J., Yu, C.-A., Kachurin, A. and Yu, L. (1996) X-ray crystallographic studies on specific inhibitors of mitochondrial bc1 complex. Abstracts, #SO356. XVII Meeting, Intntl. Union Crystallography, Seattle.
  4. Weiss, H. (1987) Structure of mitochondrial ubiquinol-cytochrome c reductase (Complex III). Curr. Topics in Bioenergetics, 15, 67-90
  5. Karlsson, B., Hovmoller, S., Weiss, H. and Leonard, K. (1983) Structural studies of cytochrome reductase. Subunit topography determined by electron microscopy of membrane crystals of a subcomplex. J. Mol. Biol. 165, 287-302
  6. Kawamoto, T. Kubota, T. Matsunaga, K. Fukuyama, H. Matsubara, K. Shinzawa-Itoh & S. Yoshikawa (1994) New crystal forms and preliminary X-ray diffraction studies of mitochondrial cytochrome bc1 complex from bovine heart. J. Mol. Biol. 244: 238-41
  7. E. A. Berry, L. S. Huang, T. N. Earnest & B. K. Jap (1992) X-ray diffraction by crystals of beef heart ubiquinol: cytochrome c oxidoreductase. J Mol Biol 224: 1161-6
  8. Berry, E.A, Shulmeister, V.M, Huang, L.S, Kim, S.H. (1995) A new crystal form of bovine heart ubiquinol:cytochrome c oxidoreductase - determination of space group and unit-cell parameters. Acta Crystallographica section D-Biological Crystallography, 51, 235-239
  9. Lee, J.W., Chan, M., Law, T.V., Kwon, H.J. and Jap, B.K. (1995) Preliminary cryocrystallographic study of the mitochondrial cytochrome bc1 complex - improved crystallization and flash-cooling of a large membrane protein. J. Mol. Biol. 252, 15-19.
  10. Kubota, M. Kawamoto, K. Fukuyama, K. Shinzawa-Itoh, S. Yoshikawa & H. Matsubara (1991) Crystallization and preliminary X-ray crystallographic studies of bovine heart mitochondrial cytochrome bc1 complex. J Mol Biol 221: 379-82
  11. Martinez, S.E., Smith, J.L., Huang, D., Szczepaniak, A. and Cramer, W.A. (1993) Crystallographic studies of the lumen-side domain of turnip cytochrome f. In Research in Photosynthesis (Murata, N., ed.) Vol. II, pp. 495-498. Kluwer Academic Publ., Dordrecht
  12. Link, T.A., Saynovits, M., Assmann, C., Iwata, S., Ohnishi, T. and Von Jagow, G. (1996) Isolation, characterization, and crystallization of a water-soluble fragment of the Rieske iron-sulfur protein of bovine heart mitochondrial bc1 complex. Europ. J. Biochem. 237, 71-75.
  13. Iwata, S., Saynovits, M., Link, T.A. and Michel, H. (1996) Structure of a water soluble fragment of the 'Rieske' iron-sulfur protein of the bovine heart mitochondrial cytochrome bc1 complex determined by MAD phasing at 1.5 resolution. Structure 1996, 4: 567-579.
  14. Link, T.A and Iwata, S. (1996) Functional implications of the structure of the 'Rieske' iron-sulfur protein of bovine heart mitochondrial bc1 complex. Biochim. Biophys. Acta 1275, 54-60.
  15. Yu, C.-A., Xia, J.-Z., Kachurin, A.M., Yu, L., Xia, D., Kim, H. and Deisenhofer, J. (1996) Crystallization and preliminary structure of beef heart mitochondrial cytochrome-bc1 complex. Biochim. Biophys. Acta 1275, 47-53.
  16. Ericinska et al. Biochim. Biophys. Acta 501, 63
  17. Simpkin, D., Palmer, G., Devlin, F.J., McKenna, M.C., Jensen, G.M., and Stephens, P.J. (1989) The axial ligands of heme in cytochromes: A near-infrared magnetic circular dichroism study of yeast cytochromes c, c1, and b and spinach cytochrome f, Biochemistry 28, 8033-8039
  18. Riedel et al., (1991) J. Biol. Chem. 266, 17838
  19. Palmer, G. and Degli Esposti, M. (1994) Biochemistry 33, 176-185