antagonist for Tie2 receptors expressed on endothelial cells in vitro and in vivo. How- ever, we also unexpectedly found that Ang2 acts as an agonist for Tie2 receptors ectopi- cally expressed on nonendothelial cells (fi- broblasts). One possible explanation for these disparate effects is that endothelial cells express an accessory component or components required for Ang2 to act as an antagonist. Tie1, which has no known li- gand and is expressed predominantly in en- dothelial cells, may be such a component, perhaps by undergoing Ang2-induced het- erodimerization with Tie2 and thus prevent- ing Tie2 activation; however, preliminary evidence argues against this possibility (17). The observation that Ang2 can activate Tie2 on certain cultured cells raises the question of whether this activation occurs in vivo, perhaps in a special subclass of endo- thelial cells, or in the rare nonendothelial cell types that express Tie2, such as the early cells of the hemopoietic lineage (11). REFERENCES AND NOTES ___________________________ 1. W. Risau, FASEB J. 9, 926 (1995); Nature 386, 671 (1997). 2. J. Folkman, Nature Med. 1, 27 (1995). 3. D. Hanahan and J. Folkman, Cell 86, 353 (1996). 4. T. Mustonen and K. Alitalo, J. Cell Biol. 129, 895 (1995). 5. G. H. Fong, J. Rossant, M. Gertenstein, M. L. Breit- man, Nature 376, 66 (1995); F. Shalaby et al., ibid., p. 62; P. Carmeliet et al., ibid. 380, 435 (1996); N. Ferrara et al., ibid., p. 439. 6. D. Jackson, O. V. Volpert, N. Bouck, D. I. H. Linzer, Science 266, 1581 (1994). 7. M. S. O’Reilly et al., Cell 79, 315 (1994). 8. M. S. O’Reilly et al., ibid. 88, 277 (1997). 9. S. Davis et al., ibid. 87, 1161 (1996). 10. D. J. Dumont, T. P. Yamaguchi, R. A. Conlon, J. Rossant, M. L. Breitman, Oncogene 7, 1471 (1992); P. C. Maisonpierre, M. Goldfarb, G. D. Yancopoulos, G. Gao, ibid. 8, 1631 (1993); T. N. Sato, Y. Qin, C. A. Kozak, K. L. Audus, Proc. Natl. Acad. Sci. U.S.A. 90, 9355 (1993); H. Schnurch and W. Risau, Develop- ment 119, 957 (1993); S. F. Ziegler, T. A. Bird, K. A. Schneringer, K. A. Schooley, P. R. Baum, Oncogene 8, 663 (1993). 11. A. Iwama et al., Biochem. Biophys. Res. Commun. 195, 301 (1993). 12. D. J. Dumont et al., Genes Dev. 8, 1897 (1994); T. N. Sato et al., Nature 376, 70 (1995). 13. C. Suri et al., Cell 87, 1171 (1996). 14. Cloning of Ang2 cDNAs resulted from screening a human fetal lung and a mouse adult uterus cDNA library (both from Clontech) with the full-length mouse 32 P-labeled Ang1 cDNA probe at 55° or 65°C (25). 15. Mutation of Cys 245 in Ang1, which is not shared between Ang1 and Ang2, does not alter its agonistic properties. For some experiments, Ang1* (a recom- binant version of Ang1 with a modified NH 2 -terminus and mutated Cys 245 that is easier to produce and purify) was used instead of Ang1; for details on ex- pression and purification of Ang1, Ang1*, and Ang2, see S. Davis et al., in preparation. 16. J. Partanen et al., Mol. Cell. Biol. 12, 1698 (1992). 17. P. C. Maisonpierre et al., data not shown. 18. K. K. Hirschi and P. A. D’Amore, Cardiovasc. Res. 32, 687 (1996). 19. J. K. Findlay, J. Endocrinol. 111, 357 (1986); H. S. Phillips, J. Hains, D. W. Leung, N. Ferrara, Endocri- nology 127, 965 (1990); D. Shweiki, A. Itin, G. Neufeld, H. Gitay-Goren, E. Keshet, J. Clin. Invest. 91, 2235 (1993); E. Bacharach, A. Itin, E. Keshet, Proc. Natl. Acad. Sci. U.S.A. 89, 10686 (1992). 20. T. M. Schlaeger et al., Proc. Natl. Acad. Sci. U.S.A. 94, 3058 (1997). 21. R. Schweitzer, R. Howes, R. Smith, B.-Z. Shilo, M. Freeman, Nature 376, 699 (1995). 22. V. Cioce et al., J. Biol. Chem. 271, 13110 (1996). 23. W. P. Arend, H. G. Welgus, R. C. Thompson, S. P. Eisenberg, J. Clin. Invest. 85, 1694 (1990). 24. M. Vikkula et al., Cell 87, 1181 (1996). 25. P. C. Maisonpierre et al., Science 247, 1446 (1990). 26. C.-J. S. Edgell, C. C. McDonald, J. B. Graham, Proc. Natl. Acad. Sci. U.S.A. 80, 3734 (1983). 27. P. C. Maisonpierre, P. F. Jones, S. Davis, G. D. Yancopoulos, in preparation. 28. Probes for mouse embryo sections were 560- and 680-nucleotide (nt) cRNAs extending from 5' leader sequence to, respectively, codon 165 in mouse Ang1 cDNA and codon 155 in mouse Ang2 cDNA. Probes for rat ovary sections were VEGF, a 141-nt cRNA spanning the last 47 rat VEGF codons (which are shared among most VEGF RNA splice-variants); Ang1, a 773-nt cRNA spanning the last 257 codons of mouse Ang1 cDNA; and Ang2, a 315-nt cRNA spanning codons 380 to 485 in rat Ang2 cDNA. 29. To obtain staged ovulating ovarian tissue, we injected immature (postnatal day 29) female Sprague-Dawley rats ( Zivic-Miller) with 5 U of pregnant mare serum gonadotropin (Calbiochem). Rats were killed and ova- ries were surgically removed beginning at 56 hours after injection and at 9- to 32-hour intervals thereafter. 30. We thank L. S. Schleifer and P. R. Vagelos for en- thusiastic support; M. Goldfarb, M. Wang, R. Ross- man, D. Datta, Y. Qing, T. Schlaeger, J. Lawitts, and J. Bruno for their contributions; J. Springhorn for HUCEC cells; and C. Murphy, E. Hubel, C. Ru- dowsky, and E. Burrows for graphics work. T.N.S. was partly supported by Hoffmann–LaRoche Inc. 7 March 1997; accepted 13 May 1997 Crystal Structure of the Cytochrome bc 1 Complex from Bovine Heart Mitochondria Di Xia, Chang-An Yu, Hoeon Kim, Jia-Zhi Xia, Anatoly M. Kachurin, Li Zhang, Linda Yu, Johann Deisenhofer* On the basis of x-ray diffraction data to a resolution of 2.9 angstroms, atomic models of most protein components of the bovine cytochrome bc 1 complex were built, including core 1, core 2, cytochrome b, subunit 6, subunit 7, a carboxyl-terminal fragment of cytochrome c 1 , and an amino-terminal fragment of the iron-sulfur protein. The positions of the four iron centers within the bc 1 complex and the binding sites of the two specific respiratory inhibitors antimycin A and myxothiazol were identified. The membrane- spanning region of each bc 1 complex monomer consists of 13 transmembrane helices, eight of which belong to cytochrome b. Closely interacting monomers are arranged as symmetric dimers and form cavities through which the inhibitor binding pockets can be accessed. The proteins core 1 and core 2 are structurally similar to each other and consist of two domains of roughly equal size and identical folding topology. Ubiquinol–cytochrome c oxidoreductase (bc 1 complex) is a component of the eu- karyotic or bacterial respiratory chain and of the photosynthetic apparatus in purple bacteria. In mitochondria, this enzyme cat- alyzes electron transfer from ubiquinol to cytochrome c, which is coupled to the translocation of protons across the mito- chondrial inner membrane from the matrix space (negative or N side) to the intermem- brane space (positive or P side). Thus, bc 1 contributes to the electrochemical proton gradient that drives adenosine triphosphate (ATP) synthesis (1). The purified mito- chondrial bc 1 complex contains 11 protein subunits (2, 3); it consists of 2165 amino acid residues and four prosthetic groups with a total molecular mass of 248 kD. The amino acid sequences of all subunits are known; some of them were determined by peptide sequencing (4) and others were de- duced from nucleotide sequences (5). The essential redox components of bc 1 are the two b-type hemes b L (also called b 565 ) and b H (b 562 ), one c-type heme (c 1 ), one high- potential iron-sulfur cluster (2Fe-2S Rieske center), and ubiquinone. On the basis of functional data (1), the proton-motive Q cycle has been favored as a model for bc 1 function (6). The key fea- ture of the model is that there are two separate ubiquinone or ubiquinol binding sites; ubiquinol is oxidized at site Q o , near the P side of the inner mitochondrial mem- brane, and ubiquinone is reduced at site Q i , near the N side of the membrane. Accord- ing to the Q cycle model, one electron is transferred from ubiquinol to the Rieske D. Xia, H. Kim, and J. Deisenhofer are in the Howard Hughes Medical Institute and Department of Biochemis- try, University of Texas Southwestern Medical Center, Dallas, TX 75235, USA. C.-A. Yu, J.-Z. Xia, A. M. Ka- churin, L. Zhang, and L. Yu are in the Department of Biochemistry and Molecular Biology, Oklahoma State University, Stillwater, OK 74078, USA. * To whom correspondence should be addressed. E-mail: jd@howie.swmed.edu SCIENCE  VOL. 277  4 JULY 1997  www.sciencemag.org 60