Structural basis of transcription: H9251-Amanitin–RNA polymerase II cocrystal at 2.8 Å resolution David A. Bushnell, Patrick Cramer*, and Roger D. Kornberg † Department of Structural Biology, Stanford University School of Medicine, Stanford, CA 94305-5126 Contributed by Roger D. Kornberg, December 12, 2001 The structure of RNA polymerase II in a complex with the inhibitor H9251-amanitin has been determined by x-ray crystallography. The structure of the complex indicates the likely basis of inhibition and gives unexpected insight into the transcription mechanism. T he structure of 10-subunit 0.5-MDa yeast RNA polymerase II (pol II), recently determined at 2.8 Å resolution, reveals the architecture and key functional elements of the enzyme (1). The two largest subunits, Rpb1 and Rpb2, lie at the center, on either side of a nucleic acid-binding cleft, with the many smaller subunits arrayed around the outside. Rpb1 and Rpb2 interact extensively in the region of the active site and also through a domain of Rpb1 that lies on the Rpb2 side of the cleft, connected to the body of Rpb1 by an H9251-helix that bridges across the cleft. Proof that nucleic acids bind in the channel comes from the molecular replacement solution of a transcribing pol II complex at 3.3 Å resolution (2). This structure shows the template DNA unwinding some three residues before the active site, followed by nine base pairs of DNA–RNA hybrid. Adjacent regions of Rpb1 and Rpb2 form a highly complementary surface, resulting in extensive DNA–RNA hybrid–protein interaction. The ‘‘bridge’’ helix seems to play an important role, binding to both the second and third unpaired DNA bases and also to the coding base, paired with the first residue of the RNA. Comparison of the pol II structure in different crystal forms shows a division of the enzyme in several mobile elements that my facilitate DNA and RNA movement during transcription. Comparison of the pol II structure with that of the related bacterial RNA polymerase (3) suggests mobility of the bridge helix as well (2). The pol II structures open the way to many lines of investi- gation. Structures of cocrystals of pol II with interacting mole- cules can be solved, the full power of site-directed mutagenesis can be brought to bear on the transcription mechanism, and so forth. Here we report the structure of a cocrystal of pol II with the most potent and specific known inhibitor of the enzyme, H9251-amanitin. The active principle of the ‘‘death cap’’ mushroom, H9251-amanitin blocks both transcription initiation and elongation (4–6). The structure of the cocrystal suggests that H9251-amanitin interferes with a protein conformational change underlying the transcription mechanism. Materials and Methods Crystals of yeast pol II were grown as described and were soaked in cryoprotectant solution containing 50 H9262gH20862ml H9251-amanitin and 1 mM MgSO 4 for 1 week before freezing and x-ray data collection to 2.8 Å resolution (Table 1; ref. 7). Data collection was carried out at 100 K by using 0.5° oscillations with an Area Detector Systems Quantum 4 charge-coupled device (CCD) detector at Stanford Synchrotron Radiation Laboratory beam- line 11-1. Diffraction data were processed with DENZO and reduced with SCALEPACK (8). The previous 2.8-Å pol II structure was subjected to rigid body refinement against the cocrystal data. The R-free test set from the native form 2 pol II data was used for the pol II H9251-amanitin refinement (1). Refinement of the cocrystal structure was preformed by using CNS (9). A H9268A- weighted difference electron density map was consistent with the known structure of amanitin toxins (Fig. 1A). After positional and B-factor refinement of the pol II model and minor adjust- ments to the model, an H9251-amanitin model was placed. The H9251-amanitin model was generated from 6H11032-O-methyl-H9251-amanitin (S)-sulfoxide methanol solvate monohydrate as obtained from the Cambridge Structure Database [accession code 3384082 (10)]. To conform to the known composition and stereochem- istry of H9251-amanitin, the 6H11032-O-methyl group was removed from the 6H11032-O-methyltryptophan residue (H9251-amanitin position 4) and the stereochemistry of the sulfoxide was modified to R. Topology and refinement parameter files for use in CNS for the H9251-amanitin structure were generated by using HIC-UP (11). Rigid body refinement was performed on the H9251-amanitin alone, followed by positional and B-factor refinement of the entire pol II-H9251- Abbreviation: pol II, polymerase II. Data deposition: The atomic coordinates have been deposited in the Protein Data Bank, www.rcsb.org (PDB ID code 1K83). *Present address: Institute of Biochemistry, Gene Center, University of Munich, 81377 Munich, Germany. † To whom reprint requests should be addressed. E-mail: firstname.lastname@example.org. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Table 1. Crystallographic data Space group I222 Unit cell, Å 122.5 by 222.5 by 374.2 Wavelength, Å 0.965 Mosaicity, ° 0.44 Resolution, Å 20–2.8 (2.9–2.8) Completeness, % 99.8 (99.4) Redundancy 3.9 (2.9) Unique reflections 124,441 (12,292) R sym , % 6.7 (21.6) Values in parentheses correspond to the highest-resolution shell. R sym H11005 H9018 i,h H20841I(i,h) H11002H20855I(h)H20856H20841H20862H9018 i,h H20841I(i,h)H20841, where H20855I(h)H20856 is the mean of the I observations of reflection h. R sym was calculated with anomalous pairs merged; no sigma cut-off was applied. Table 2. Refinement statistics Nonhydrogen atoms 27,906 Protein residues 3,490 Water molecules 69 Anisotropic scaling (B11, B22, B33) H110026.3, H110026.9, 13.1 rms deviation bonds 0.0083 rms deviation angles 1.4 Reflection test set 3,757 (3.0%) R cryst H20862R free 22.9H2086228.0 Average B factor overall 57 Average B factor pol 57 Average B factor amanitin 78 Average B factor water 35 R cryst/free H11005H20858 h H20648F obs (h)H20841 H11002 H20841F calc (h)H20648H20862H20858 h H20841F obs (h)H20841. R cryst and R free were calculated from the working and test reflection sets, respectively. 1218–1222 H20841 PNAS H20841 February 5, 2002 H20841 vol. 99 H20841 no. 3 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.251664698 Fig. 1. Stereo image of final H9251-amanitin structure. (A)H9268A-weighted F obs H11002F calc elec- tron density at 2.8 Å resolution (red) con- toured at 3 sigma calculated from the ini- tial pol II placement beforeH9251-amanitin was included in the model. The final H9251-aman- itin structure is shown (ball and stick mod- el). (B) H9268A-weighted 2F obs H11002 F calc electron density at 2.8 Å resolution (blue) con- toured at 1.2 sigma, superimposed on the final H9251-amanitin structure (ball and stick model). Only the electron density around H9251-amanitin is shown. This figure was gen- erated by using BOBSCRIPT and RASTER3D (21–23). Fig. 2. Location ofH9251-amanitin bound to pol II. (A) Cutaway view of a pol II-transcribing complex showing the location ofH9251-amanitin binding (red dot) in relation to the nucleic acids and functional elements of the enzyme. Adapted from ref. 24. (B) Ribbons representation of the pol II structure (top view in refs. 1 and 7). Eight zinc atoms are shown in light blue, the active site magnesium is magenta, the region of Rpb1 around H9251-amanitin is light green (funnel) and dark green (bridge helix), the region of Rpb2 near H9251-amanitin is dark blue, and H9251-amanitin is red. This figure was prepared by using RIBBONS (25). Bushnell et al. PNAS H20841 February 5, 2002 H20841 vol. 99 H20841 no. 3 H20841 1219 BIOCHEMISTRY amanitin complex and further minor adjustment of the model, giving a final free-R factor of 28% (Table 2). The refined H9268A-weighted 2F obs H11002 F calc map (Fig. 1B) clearly shows density for the main chain atoms. Some of the side chains, however, such as that of the 4,5-dihydroxyisoleucine residue, are only partially visible (ordered) in the map. The stereo chemistry of the 4,5-dihydroxyisoleucine H9253 hydroxyl is important in amanitin inhibition, suggestive of a role in hydrogen bonding (12). Poor ordering in our cocrystal indicates that at least in yeast, the proposed hydrogen bond is not formed. This may partially explain the lesser sensitivity of Saccharomyces cerevisiae to H9251-amanitin compared with other eukaryotes (4). Results and Discussion The H9251-amanitin binding site is beneath a ‘‘bridge helix’’ extending across the cleft between the two largest pol II subunits, Rpb1 and Rpb2, in a ‘‘funnel’’-shaped cavity in the pol II structure (Fig. 2 A and B). Most pol II mutations affecting H9251-amanitin inhibition map to this site (Table 3), showing that it is functionally relevant and not an artifact of crystallization (13–15). Pol II residues interacting with H9251-amanitin are located almost entirely in the bridge helix (in the previously defined ‘‘cleft’’ region of Rpb1) and in an adjacent part of Rpb1 on the Rpb2-side of the cleft [in the previously defined funnel region of Rpb1 (Fig. 3 A and B; Table 3)]. There is a strong hydrogen bond between hy- droxyproline 2 of H9251-amanitin and bridge helix residue Glu- A822. There is an indirect interaction involving the backbone carbonyl group of 4,5-dihydroxyisoleucine 3 of H9251-amanitin, hydrogen-bonded to residue Gln-A768, which is, in turn, hydrogen-bonded to bridge helix residue His-A816. Finally, there are several hydrogen bonds between H9251-amanitin and the region of Rpb1 adjacent to the bridge helix. Binding of H9251-amanitin therefore buttresses the bridge helix, constraining its position with respect to the Rpb2-side of the cleft. This mode of H9251-amanitin interaction can account for the biochemistry of inhibition. There is little if any influence of H9251-amanitin binding on the affinity of pol II for nucleoside triphosphates (5, 16). Moreover, after the addition of H9251-aman- itin to a transcribing pol II complex, a phosphodiester bond can Table 3. Hydrogen bonds, buried surface area, and known amanitin mutants Residue in yeast H9004 surface area, Å 2 H-bond Residue in human Mutations Val-A719 H1100232 Asn-A742 Leu-A722 0 Leu-A745 Mouse L745F (13) Asn-A723 H1100222 Asn-A746 Arg-A726 H1100263 NH1 to AMA pos. 4 O 3.0 Å Arg-A749 Mouse R749P (14) Drosophila melanogaster R741H(15) Asp-A727 H110027 Asp-A750 Phe-A755 H110028 Lys-A778 Ile-A756 H1100248 Ile-A779 Mouse I779F (14) Ala-A759 H110027 Ser-A782 Gln-A760 H1100233 Gln-A783 Cys-A764 0 Val-A787 Caenorhabditis elegans C777Y(15) Val-A765 H110022 Val-A788 Gly-A766 H110021 Gly-A789 Gln-A767 H1100234 N to AMA pos. 4 O 3.1 Å Gln-A790 O to AMA pos. 5 N 3.2 Å Gln-A768 H1100216 OE1 to AMA pos. 3 O 2.6 Å Gln-A791 Ser-A769 H1100237 N to AMA pos. 2 O 3.3 Å Asn-A792 Mouse N792D (14) Gly-A772 H1100224 Gly-A795 C. elegans G785E (15) Lys-A773 H110024 Lys-A796 Arg-A774 H110022 Arg-A797 Tyr-A804 H110022 Tyr-A827 His-A816 H1100213 His-A839 Gly-A819 H1100219 Gly-A842 Gly-A820 H110028 Gly-A843 Glu-A822 H1100215 OE2 to AMA pos. 2 OD2 2.6 Å Glu-A845 Gly-A823 H1100213 Gly-A846 Asp-A826 H110022 Asp-A849 Thr-A1080 H110021 Thr-A1103 Leu-A1081 H1100263 Leu-A1104 Lys-A1092 H1100237 Lys-A1115 Lys-A1093 H110021 Asn-A1116 Gln-B763 H1100216 Gln-B718 Pro-B765 H1100211 Pro-B720 Total H11002541 H9004 surface area (Å 2 ) is the change in solvent-exposed surface as calculated with program AREAIMOL, using a standard probe radius of 1.4 Å. Potential hydrogen bonds with a donor-acceptor distance below 3.3 Å were included. Residues that are different between yeast and human are in bold. Mutations are changes in Rpb1 in eukaryotes that are known to affect H9251-amanitin inhibition. H9251-Amanitin also seems to make a contact with part of the disordered loop between A1081 and A1092. Unfortunately, only density for H110111 amino acid appears, preventing placement of this loop or even reliable determination of which amino acid in the disordered loop is responsible for this interaction. 1220 H20841 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.251664698 Bushnell et al. still be formed (17, 18). The rate of translocation of pol II on DNA is, however, reduced from several thousand to only a few nucleotides per minute (5, 6). These findings are consistent with binding of H9251-amanitin too far from the active site to interfere with nucleoside triphosphate entry or RNA synthesis (or its reversal) (Fig. 2A; ref. 5). They may be explained by a constraint on bridge helix movement. It was previously sug- gested that such movement is coupled to DNA translocation. The suggestion was based on two observations. First, in the structure of a pol II-transcribing complex, bridge helix residues directly contact the DNA base paired with the first base in the RNA strand. Second, although the sequence of the bridge helix is well conserved, the conformation is different in a bacterial RNA polymerase structure, with bridge helix residues in position to contact the second base in the DNA strand (1, 2, 19). Movement of bridge helix residue Glu-A822 by as little as 1 Å would extend the length of the donor-acceptor pair for the hydrogen bond to hydroxyproline 2 of H9251-amanitin beyond 3.3 Å, effectively breaking the bond. Structural derivatives of H9251-amanitin show the importance of bridge helix interaction for inhibitory activity. The derivative proamanullin, which lacks the hydroxyl group of hydroxyproline 2, involved in hydrogen bonding to bridge helix residue Glu- A822, and which also lacks both hydroxyl groups of 4,5- dihroxyisoleucine 3, is about 20,000-fold less inhibitory than H9251-amanitin. This effect is caused almost entirely by the alteration of hydroxyproline 2, because alteration of 4,5-dihydroxyisoleu- cine 3 alone, in the derivative amanullin, reduces inhibition only about 4-fold (4, 20). Other changes in H9251-amanitin structure may affect inhibition indirectly, by diminishing the overall affinity for pol II. For example, shortening the side chain of isoleucine-6 of H9251-amanitin reduces inhibition by about 1,000-fold. This side chain inserts in a hydrophobic pocket of pol II in the cocrystal structure. Thus three lines of evidence on H9251-amanitin inhibition, coming from biochemical studies of transcription, from structure- activity relationships, and from cocrystal structure determina- tion, converge on a simple picture. Binding of H9251-amanitin to pol II permits nucleotide entry to the active site and RNA synthesis but prevents the translocation of DNA and RNA needed to empty the site for the next round of synthesis. The inhibition of translocation is caused by interaction of H9251-amanitin with the pol II bridge helix, whose movement is required for translocation. We thank N. Thompson and R. Burgess (McArdle Laboratory for Cancer Research, Univ. of Wisconsin) for generously providing anti- body for protein purification. We thank B. Freedman (North American Mycological Association) for providing samples of Amanita phalloides. We thank COMPAQ (Houston) for providing a Unix workstation. This research is based in part on work done at SSRL, which is funded by the U.S. Department of Energy Office of Basic Energy Sciences. The structural biology program is supported by the National Institutes of Health National Center for Research Resources Biomedical Tech- Fig. 3. Interaction of H9251-amanitin with pol II. (A) The chemical structure of H9251-amanitin, with residues of pol II that lie within 4 Å [determined by using CONTACT (26)] placed near the closest contact. The CH9251sofH9251-amanitin are labeled with blue numbers. Hydrogen bonds are shown as dashed lines with the distances indicated. (B) Stereoview of the H9251-amanitin binding pocket. Ball and stick models of H9251-amanitin (red bonds) and of pol II residues within 4 Å (gray bonds) are shown. Rpb1 from A700 to A809 (funnel region) is light green. Rpb1 from A810 to A825 (bridge helix) is dark green. Rpb2 from B760 to B769 is blue. This figure was generated by using BOBSCRIPT and RASTER3D (21–23). Bushnell et al. PNAS H20841 February 5, 2002 H20841 vol. 99 H20841 no. 3 H20841 1221 BIOCHEMISTRY nology Program and the Department of Energy Office of Biological and Environmental Research. P.C. was supported by a postdoctoral fellowship of the Deutsche Forschungsgemeinschaft. D.A.B. was supported by postdoctoral fellowship no. PF-00-014-01-GMC from the American Cancer Society. This research was supported by National Institutes of Health Grant GM49985 (to R.D.K.). 1. Cramer, P., Bushnell, D. A. & Kornberg, R. D. (2001) Science 292, 1863–1876. 2. Gnatt, A. L., Cramer, P., Fu, J., Bushnell, D. A. & Kornberg, R. D. (2001) Science 292, 1876–1882. 3. Zhang, G., Campbell, E. A., Minakhin, L., Richter, C., Severinov, K. & Darst, S. A. (1999) Cell 98, 811–824. 4. Wieland, T. & Faulstich, H. (1991) Experientia 47, 1186–1193. 5. Chafin, D. R., Guo, H. & Price, D. H. (1995) J. Biol. Chem. 270, 19114–19119. 6. Rudd, M. D. & Luse, D. S. (1996) J. Biol. Chem. 271, 21549–21558. 7. Cramer, P., Bushnell, D. A., Fu, J., Gnatt, A. L., Maier-Davis, B., Thompson, N. E., Burgess, R. R., Edwards, A. M., David, P. R. & Kornberg, R. D. (2000) Science 288, 640–649. 8. Otwinowski, Z. & Minor, W. (1997) Methods Enzymol. 276, 307–326. 9. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., et al. (1998) Acta Crystallogr. D 54, 905–921. 10. Wieland, T., Gotzendorfer, C., Dabrowski, J., Lipscomb, W. N. & Shoham, G. (1983) Biochemistry 22, 1264–1271. 11. Kleywegt, G. J. & Jones, T. A. (1998) Acta Crystallogr. D 54, 1119–1131. 12. Zanotti, G., Wieland, T., Benedetti, E., Di Blasio, B., Pavone, V. & Pedone, C. (1989) Int. J. Pept. Protein Res. 34, 222–228. 13. Bartolomei, M. S. & Corden, J. L. (1987) Mol. Cell. Biol. 7, 586–594. 14. Bartolomei, M. S. & Corden, J. L. (1995) Mol. Gen. Genet. 246, 778–782. 15. Chen, Y., Weeks, J., Mortin, M. A. & Greenleaf, A. L. (1993) Mol. Cell. Biol. 13, 4214–4222. 16. Cochet-Meilhac, M. & Chambon, P. (1974) Biochim. Biophys. Acta 353, 160–184. 17. Vaisius, A. C. & Wieland, T. (1982) Biochemistry 21, 3097–3101. 18. Gu, W., Powell, W., Mote, J., Jr., & Reines, D. (1993) J. Biol. Chem. 268, 25604–25616. 19. Darst, S. A. (2001) Curr. Opin. Struct. Biol. 11, 155–162. 20. Zanotti, G., Mohringer, C. & Wieland, T. (1987) Int. J. Pept. Protein Res. 30, 450–459. 21. Esnouf, R. M. (1997) J. Mol. Graph. Model. 15, 132–134. 22. Esnouf, R. M. (1997) J. Mol. Graph. Model. 15, 112–113. 23. Merrit, E. A. & Murphy, M. E. P. (1994) Acta Crystallogr. D 50, 869–873. 24. Klug, A. (2001) Science 292, 1844–1846. 25. Carson, M. (1997) in Methods in Enzymology, eds. Sweet, R. M. & Carter, C. W. (Academic, New York), Vol. 277, pp. 493–505. 26. Collaborative Computational Project, No. 4 (1994) Acta Crystallogr. D 50, 760–763. 1222 H20841 www.pnas.orgH20862cgiH20862doiH2086210.1073H20862pnas.251664698 Bushnell et al.