A glycyl radical site in the crystal structure of a class III ribonucleotide reductase

Science

Volumn 283, Issue 5407, 1999, Pages 1499-1504

  Logan, Derek T ^a   Andersson, Jessica ^a   Sjöberg, Britt Marie ^a   Nordlund, Pär ^a  

^a STOCKHOLM UNIVERSITY   (Sweden)

Author keywords

[No Author keywords available]

Indexed keywords

DEOXYRIBONUCLEOTIDE; RIBONUCLEOTIDE REDUCTASE;

AMINO ACID SEQUENCE; ANAEROBIC METABOLISM; ARTICLE; CATALYSIS; CRYSTAL STRUCTURE; ENZYME ACTIVE SITE; ENZYME ANALYSIS; EVOLUTION; PRIORITY JOURNAL; PROTEIN DOMAIN; PROTEIN FOLDING; RNA ANALYSIS; SEQUENCE HOMOLOGY; STRUCTURE ANALYSIS;

ACETYLTRANSFERASES; AMINO ACID SEQUENCE; ANAEROBIOSIS; BINDING SITES; CRYSTALLOGRAPHY, X-RAY; DIMERIZATION; EVOLUTION, MOLECULAR; GLYCINE; HYDROGEN BONDING; MODELS, MOLECULAR; MOLECULAR SEQUENCE DATA; MUTATION; PROTEIN CONFORMATION; PROTEIN FOLDING; PROTEIN STRUCTURE, SECONDARY; RIBONUCLEOTIDE REDUCTASES; VIRAL PROTEINS;

EID: 0033525599     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.283.5407.1499     Document Type: Article

References (61)

1. P. Reichard, Science 260, 1773 (1993).
2. P. Nordlund, B.-M. Sjöberg, H. Eklund, Nature 345, 593 (1990).
3. B.-M. Sjöberg, Struct. Bonding 88, 139 (1997).
4. S. Licht, C. J. Gerfen, J. Stubbe, Science 271, 477 (1996).
5. A. Jordan and P. Reichard, Annu. Rev. Biochem. 67, 71 (1998).
6. X. Sun et al., J. Biol. Chem. 270, 2443 (1995).
7. P. Young, J. Andersson, M. Sahlin, B. M. Sjöberg, ibid. 271, 20770 (1996).
8. X. Sun et al., ibid. , p. 6827.
9. S. Ollagnier et al., ibid. 272, 24216 (1997).
10. D. S. King and P. Reichard, Biochem. Biophys. Res. Commun. 206, 731 (1995).
11. A. L. Persson, M. Sahlin, B.-M. Sjöberg, J. Biol. Chem. 273, 31016 (1998).
12. A. Aberg et al., ibid. 264, 12249 (1989).
13. E. Mulliez, S. Ollagnier, M. Fontecave, R. Eliasson, P. Reichard, Proc. Natl. Acad. Sci. U.S.A 92, 8759 (1995).
14. S. Booker, S. Licht, J. Broderick, J. Stubbe, Biochemistry 33, 12676 (1994).
15. U. Uhlin and H. Eklund, Nature 370, 533 (1994).
16. J. Riera, F. T. Robb, R. Weiss, M. Fontecave, Proc. Natl. Acad. Sci. U.S.A. 94, 475 (1997); A. Tauer and S. Benner, ibid., p. 53; A. Jordan et al., ibid., p. 13487.
17. J. Riera, F. T. Robb, R. Weiss, M. Fontecave, Proc. Natl. Acad. Sci. U.S.A. 94, 475 (1997); A. Tauer and S. Benner, ibid., p. 53; A. Jordan et al., ibid., p. 13487.
18. J. Riera, F. T. Robb, R. Weiss, M. Fontecave, Proc. Natl. Acad. Sci. U.S.A. 94, 475 (1997); A. Tauer and S. Benner, ibid., p. 53; A. Jordan et al., ibid., p. 13487.
19. 3 per dimer.
20. Superposition of structures was carried out with the SHP program (D. I. Stuart, unpublished data).
21. 2-terminal domain present in some other NrdD sequences, which corresponds to the overall activity allosteric site domain in R1. However, conservation of essential nucleotide recognition residues in NrdD sequences where the domain is present suggests that it will have a very similar structure to that seen in R1.
22. M. Eriksson et al., Structure 5, 1077 (1997).
23. The altered dimer interface makes it impossible for dNTPs to bind as they do in R1. dATP binds in a cleft between αA of one monomer and αB of monomer 2, and makes contacts to the hairpin loop between residues 90 to 103 of monomer 1 (Fig. 2A).
24. J. Stubbe, Adv. Enzymol. Relat. Areas Mol. Biol. 63, 349 (1990).
25. _ and W. A. van der Donk, Chem. Rev. 98, 705 (1998).
26. S. S. Mao et al., Biochemistry 31, 9733 (1992).
27. J. Stubbe, M. Ator, T. Krenitsky. J. Biol. Chem. 258, 1625 (1983).
28. M. Olcott, J. Andersson, B.-M. Sjöberg, ibid. 273, 24853 (1998).
29. 3H is found almost exclusively in water molecules, suggesting that H atoms pass through a protein-based intermediate [(13); R. Eliasson et al., Biochem. Biophys. Res. Commun. 214, 28 (1995)].
30. A. L. Persson et al., J. Biol. Chem. 272, 31533 (1997).
31. R. Lenz and B. Giese, J. Am. Chem. Soc 119, 2784 (1997).
32. P. E. M. Siegbahn, ibid. 120, 8417 (1998).
33. 441 is within hydrogen bonding distance of the 3′-OH of the modeled substrate and could act as a proton donor in the elimination reaction; however it is not completely conserved, being replaced by Thr and Leu in two archaebacterial sequences.
34. 441, the lowest energy transition state for the water elimination step was found when formate bridged the 2′-OH and 3′-OH positions (30).
35. 581 is directed toward the 2′-C of the modeled substrate. However, the role of this tyrosine in radical transfer is currently unclear, because a mutant of E. coli NrdD where it is altered to Phe retains 7% activity (8).
36. M. Frey, M. Rothe, A. F. Wagner, J. Knappe, J. Biol. Chem. 269, 12432 (1904).
37. Silver-stained gels of washed and redissolved crystals indicate the presence of NrdG in the crystals. The lack of electron density is probably due to partial occupancy. The affinity of NrdG for NrdD(C580A) may be affected by the integrity of the iron-sulfur-cluster, which we made no special effort to ensure was intact throughout purification and crystallization of the complex.
38. 580.
39. J. Knappe and A. F. Wagner, Methods Enzymol. 258, 343 (1995).
40. A. F. Wagner, M. Frey, F. A. Neugebauer, W. Schafer, J. Knappe, Proc. Natl. Acad. Sci. U.S.A. 89, 996 (1992).
41. J. B. Broderick et al., J. Am. Chem. Soc. 119, 7396 (1997).
42. R. Külzer, T. Pils, R. Kappl, J. Huttermann, J. Knappe, J. Biol. Chem. 273, 4897 (1998).
43. X. Sun et al., Proc. Natl. Acad. Sci. U.S.A. 90, 577 (1993).
44. The 10 residues between amino acids 275 and 284 that are missing in PFL and in the Methanococcus jannaschii class III RNR correspond to helix αE and a short loop, both on the side of the barrel distal from the active site. These could presumably be lost during evolution without affecting the structural integrity.
45. C. V. Parast et al., Biochemistry 34, 2393 (1995); S. G. Reddy et al., ibid. 37, 558 (1998).
46. C. V. Parast et al., Biochemistry 34, 2393 (1995); S. G. Reddy et al., ibid. 37, 558 (1998).
47. E. Mulliez, M. Fontecave, J. Gaillard, P. Reichard, J. Biol. Chem. 268, 2296 (1993).
48. Reviewed in G. Sawers and G. Watson, Mol. Microbiol. 29, 945 (1998).
49. Z. Otwinowski and W. Minor, Methods Enzymol. 276, 307 (1997).
50. Collaborative Computational Project No. 4, Acta Crystallogr. D50, 760 (1994).
51. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly- Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X- PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
52. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly- Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X- PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
53. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly- Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X- PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
54. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly-Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X- PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
55. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly- Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X-PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
56. 2 was removed from the mother liquor for the nucleotide soaks because it caused precipitation of the nucleotides in the presence of PEG. The presence of Os and Hg in the crystals was determined at the time of data collection by x-ray fluorescence spectroscopy. Sites in MMA and KOs were found independently by anomalous difference Pattersons, although they turned out to have four common sites; IrCl was solved by difference Fourier with phases from the first two, but was poorly substituted. All Hg and some of the Os sites are covalently linked to free Cys residues and were used as markers during model building. Heavy-atom parameter refinement and phasing was carried out with MLPHARE [Z. Otwinowski, Proceedings of the CCP4 Study Weekend (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1991), p. 80]. To minimize bias, only one set of isomorphous differences was refined for each derivative; for the other wavelengths, anomalous differences only were used [D. T. Logan, M.-H. Mazauric, D. Kern, D. Moras, EMBO J. 14, 4156 (1995)]. The hand of the space group and of the anomalous differences was determined by inspecting electron density maps solvent-flattened with DM [K. Cowtan, Joint CCP4 ESF-EACBM Newslett. Protein Crystallogr. 31, 34 (1994)] for interpretability. Solvent flattening improved the mean figure of merit from 0.61 to 0.74. Despite the weak anomalous signals and common sites, over 300 residues of poly- Ser model could be built into these maps, which was refined with the program TNT [D. E. Tronrud, L. F. ten Eyck, B. W. Matthews, Acta Crystallogr. D53, 240 (1987)]. at first by rigid body refinement, then very tightly restrained atomic refinement. Calculated phases were combined with experimental phases, and maps were subjected to further density modification. Many cycles of such model building and phase improvement were carried out. When all visible side chains had been built, one cycle of torsion angle dynamics [L. M. Rice and A. T. Brünger, Proteins 19, 277 (1994)] was carried out with the program X- PLOR. Further refinement against the second data set was performed with TNT and Refmac [G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Cystallogr. D53, 240 (1997)].
57. A. L. Morris, M. W. MacArthur, E. G. Hutchinson, J. M. Thornton, Proteins 12, 345 (1992).
58. All figures were made using the programs Molscript [P. J. Kraulis, J. Appl. Crystallogr. 24, 946 (1991)], Bobscript [R. M. Esnouf, J. Mol. Graphics Modelling 15, 132 (1997)], and Raster3D [E. A. Merritt and M. J. Murphy. Acta Crystallogr. D50, 869 (1994)].
59. All figures were made using the programs Molscript [P. J. Kraulis, J. Appl. Crystallogr. 24, 946 (1991)], Bobscript [R. M. Esnouf, J. Mol. Graphics Modelling 15, 132 (1997)], and Raster3D [E. A. Merritt and M. J. Murphy. Acta Crystallogr. D50, 869 (1994)].
60. All figures were made using the programs Molscript [P. J. Kraulis, J. Appl. Crystallogr. 24, 946 (1991)], Bobscript [R. M. Esnouf, J. Mol. Graphics Modelling 15, 132 (1997)], and Raster3D [E. A. Merritt and M. J. Murphy. Acta Crystallogr. D50, 869 (1994)].
61. We acknowledge the contributions of P. Young and M. Westman to protein purification, assistance with data collection from V. Stojanoff and A. Thompson (BM14), M. Capel (X12B), K. Knudsen (SNBL). and outstation staff at EMBL, Hamburg. We also thank M. Thunnissen for help with early map interpretation. This work was funded by grants from the Swedish Natural Sciences Research Council to D.L., P.N., and B.-M.S and from the Swedish Cancer Foundation to B.-M.S. It was carried out in part at the NSLS, which is supported by the U.S. Department of Energy, Division of Materials Sciences and Division of Chemical Sciences. The structure of the anaerobic ribonudeotide reductase NrdD subunit has been deposited with the Protein Data Bank (accession number 188B).