Outcome of the First Electron Microscopy Validation Task Force Meeting

• DOI: 10.1016/j.str.2011.12.014
• Publication Date: 2012-02-07T14:43:45Z
• Authors: Richard Henderson; Andrej Sali; Matthew L. Baker; Bridget Carragher; Batsal Devkota; Kenneth H. Downing; Edward H. Egelman; Zukang Feng; Joachim Frank; Nikolaus Grigorieff; Wen Jiang; Steven J. Ludtke; Ohad Medalia; Pawel A. Penczek; Peter B. Rosenthal; Michael G. Rossmann; Michael F. Schmid; Gunnar F. Schröder; Alasdair C. Steven; David L. Stokes; John D. Westbrook; Willy Wriggers; Huanwang Yang; Jasmine Young; Helen M. Berman; Wah Chiu; Gerard J. Kleywegt; Catherine L. Lawson
• Keywords: Task force

This Meeting Review describes the proceedings and conclusions from inaugural meeting of Electron Microscopy Validation Task Force organized by Unified Data Resource for 3DEM (http://www.emdatabank.org) held at Rutgers University in New Brunswick, NJ on September 28 29, 2010. At workshop, a group scientists involved collecting electron microscopy data, using data to determine three-dimensional (3DEM) density maps, building molecular models into maps explored how assess models, other that are deposited Bank Protein public archives. The specific recommendations resulting workshop aim increase impact biology medicine. Structure analysis macromolecular complexes has become an essential tool structural research. is uniquely able organization not amenable methods (Frank, 2006Frank J. Three-dimensional assemblies: visualization biological molecules their native state.Second Edition. Oxford Press, York2006Crossref Google Scholar, Glaeser et al., 2007Glaeser R.M. Downing K.H. DeRosier D. Chiu W. Frank crystallography macromolecules. Oxford, York2007Google Scholar). More than thirty years ago, low-dose imaging computational averaging images two-dimensional (2D) crystals bacteriorhodopsin produced map revealed protein α helices spanning lipid bilayer (Henderson Unwin, 1975Henderson R. Unwin P.N. Nature. 1975; 257: 28-32Crossref PubMed Scopus (854) Subsequent advances unstained specimens embedded vitreous ice (cryo-EM) increasingly yielding wide variety near-atomic resolution. Applications icosahedral viruses chaperonins already demonstrate can be good enough trace Cα backbones de novo visualize some side-chain densities without aid X-ray (Chen 2011Chen D.H. Baker M.L. Hryc C.F. DiMaio F. Jakana Wu Dougherty M. Haase-Pettingell C. Schmid M.F. Jiang al.Proc. Natl. Acad. Sci. USA. 2011; 108: 1355-1360Crossref (73) Liu 2010Liu H. Jin L. Koh S.B. Atanasov I. Schein S. Zhou Z.H. Science. 2010; 329: 1038-1043Crossref (138) Zhang 2010aZhang Schröder G.F. Douglas N.R. Reissmann Fu C.J. Levitt Ludtke S.J. al.Nature. 463: 379-383Crossref (113) 2010bZhang X. Fang Q. Hui W.H. Cell. 141: 472-482Abstract Full Text PDF (146) unusually versatile used investigate structures under conditions close those cell. Specimens range highly purified, homogeneous heterogeneous conformations may assume different forms with or symmetry. Subnanometer resolution cryo-EM found useful providing illustrative snapshots machines such as ribosome, chaperonins, bound various cellular effectors ligands (Becker 2009Becker T. Bhushan Jarasch A. Armache J.P. Funes Jossinet Gumbart Mielke Berninghausen O. Schulten K. al.Science. 2009; 326: 1369-1373Crossref 1995Frank Zhu Penczek P. Li Y. Srivastava Verschoor Radermacher Grassucci Lata R.K. Agrawal 1995; 376: 441-444Crossref Miyazawa 2003Miyazawa Fujiyoshi N. 2003; 423: 949-955Crossref (849) Finally, tomography, which series collected region specimen tilted viewing angles, obtain 3D individual particles, including pleiomorphic ones whole-particle inadmissible (Grünewald 2003Grünewald Desai Winkler D.C. Heymann J.B. Belnap D.M. Baumeister Steven A.C. 302: 1396-1398Crossref (264) Scholar), well sections, even whole cells, provided they thicker approximately 0.7 μm (Al-Amoudi 2004Al-Amoudi Norlen L.P. Dubochet Struct. Biol. 2004; 148: 131-135Crossref (125) Al-Amoudi 2007Al-Amoudi Díez Betts M.J. Frangakis A.S. 2007; 450: 832-837Crossref (156) Frank, McIntosh, 2007McIntosh J.R. Cellular Microscopy. Volume 79. Elsevier Science, San Diego2007Google Medalia 2002Medalia Weber Nicastro Gerisch G. 2002; 298: 1209-1213Crossref (497) For extensive review procedures, see Henderson, 2012Baker T.S. Henderson Cryomicroscopy Biological Molecules.in: Arnold E. Himmel Rossmann M.G. International Tables Crystallography F, Macromolecules. West Sussex, UK: John Wiley & Sons, Ltd, 2012: 593-614Crossref Scholar. Interpretation frequently involves model. Models consist atoms "coarse-grained" objects representing multiple atoms, residues, secondary structure segments, shape-based features. A model given complex often computed assembling experimentally determined atomic homology subunits. subunit either rigid (Chapman, 1995Chapman M.S. Acta Crystallogr. 51: 69-80Crossref 2001Jiang Mol. 2001; 308: 1033-1044Crossref (189) Lasker 2009Lasker Topf Sali Wolfson H.J. 388: 180-194Crossref (53) Roseman, 2000Roseman A.M. D 2000; 56: 1332-1340Crossref (106) Rossmann, 2000Rossmann 1341-1349Crossref (127) Volkmann Hanein, 1999Volkmann Hanein 1999; 125: 176-184Crossref (132) Wriggers 1999Wriggers Milligan R.A. McCammon J.A. 185-195Crossref (349) Chacón, 2001Wriggers Chacón Structure. 9: 779-788Abstract (87) Scholar) allowed flex (Fabiola Chapman, 2005Fabiola Chapman 2005; 13: 389-400Abstract (77) Rusu 2008Rusu Birmanns Bioinformatics. 2008; 24: 2460-2466Crossref (18) 2007Schröder Brunger A.T. 15: 1630-1641Abstract Tama 2004Tama Miyashita Brooks 3rd, C.L. 337: 985-999Crossref (129) 2005Topf B. 149: 191-203Crossref (57) 2008Topf Webb 16: 295-307Abstract Trabuco 2008Trabuco L.G. Villa Mitra 673-683Abstract (235) 2011Trabuco Schreiner Hsin 173: 420-427Crossref (21) 2000Wriggers Drew D.L. Biophys. 79: 1670-1678Abstract 2011Zhang Ma Joachimiak L.A. Frydman 19: 633-639Abstract while being fit map; precautions need taken avoid over-fitting introducing too many refinable parameters relative available. higher resolutions (better 6 Å mostly α-helical 4 β-stranded structure), it possible recognize known folds subunits (Jiang Khayat 2010Khayat Lander G.C. Johnson J.E. 170: 513-521Crossref (7) Saha 2010Saha 26: i301-i309Crossref In addition features stereochemistry, modeling also utilize types information, symmetry, proximities proteomics experiments, residue chemical cross-linking, related homologous structures, SAXS profiles (Alber 2008Alber Förster Korkin Annu. Rev. Biochem. 77: 443-477Crossref Increasingly, described literature archives, where retrieved independent assessment, use, development new tools visualization, fitting, validation. EMDataBank, (http://emdatabank.org; Lawson 2011Lawson Best Bi Feng van Ginkel Devkota Lagerstedt al.Nucleic Acids Res. 39: D456-D464Crossref (72) Scholar; Figure 1), provides joint deposition retrieval (EMDB) archive coordinates fitted volumes (PDB) archive. Currently, more 1,000 EM 400 map-derived available (Figure 2).Figure 23DEM Entries EMDB PDB, Cumulative YearShow full captionStatistics December 31, 2011: 1322 entries, 427 entries.View Large Image ViewerDownload Hi-res image Download (PPT) Statistics entries. Every uncertainty. Therefore, assessment errors essential, especially when techniques practitioners. addition, all rapidly developing fields, our enthusiasm go further faster, there risk avoidable mistakes, both large small, made production interpretation maps. Such mistakes have adverse effect undermining credibility general. It therefore important develop checking validating goal establishing set best practices field. Historically, field any notorious blunders but, scientific disciplines, handful papers reported erroneous results. While rarity these incidents heartening, do provide justification cautious. early days crystallography, published projection was sometimes overly optimistic (Hayward Stroud, 1981Hayward Stroud 1981; 151: 491-517Crossref (36) before importance correcting beam tilt realized 1986Henderson Baldwin J.M. Lepault Zemlin Ultramicroscopy. 1986; 147-178Crossref (496) proved remarkably easy get absolute hand wrong subnanometer (Böttcher 1997Böttcher Wynne S.A. Crowther 1997; 386: 88-91Crossref (540) Kühlbrandt Wang, 1991Kühlbrandt Wang D.N. 1991; 350: 130-134Crossref (257) 1997Li Lee Jap B.K. Nat. 4: 263-265Crossref (96) 2000Zhou He Rixon F.J. 288: 877-880Crossref (212) Images untilted together information needed correctly this property (Belnap 1997Belnap Olson N.H. 120: 44-51Crossref (34) Cheng 2002Cheng Trus B.L. Newcomb W.W. Brown J.C. Virol. 76: 7855-7859Crossref (30) Conway 1997Conway J.F. Zlotnick Wingfield P.T. Stahl 91-94Crossref (301) single particle field, five between 2002 2005 independently same receptor complex, 1.3 MDa inositol phosphate receptor, tetramer responsible calcium release endoplasmic reticulum. Two were negative stain three amorphous (da Fonseca 2003da P.C. Morris Nerou E.P. Taylor C.W. Proc. 100: 3936-3941Crossref (65) Hamada 2003Hamada Terauchi Mikoshiba Chem. 278: 52881-52889Crossref 2002Jiang Q.X. Thrower E.C. Chester D.W. Ehrlich B.E. Sigworth EMBO 21: 3575-3581Crossref (81) Sato 2004Sato Ogura Iwasaki Hiroaki Tani 336: 155-164Crossref (62) Serysheva 2005Serysheva I.I. Hamilton S.L. 345: 427-431Crossref (54) Although differences partly explained biochemical preparation, likely due determination. recent study ∼10 (Ludtke 2011bLudtke Tran T.P. Ngo Q.T. Moiseenkova-Bell V.Y. 1192-1199Abstract substantially earlier agrees qualitatively one (Hamada so evidence convergence. each studies time, absence appropriate validation meant prove correct show incorrect. Additionally, common practice crystal discrepancy model-based 2.5 ryanodine remains unresolved illustrates challenge fitting fragments low-resolution (Serysheva 2008Serysheva Cong Eramian 105: 9610-9615Crossref (60) Tung 2010Tung C.C. Lobo P.A. Kimlicka Van Petegem 468: 585-588Crossref (66) study, docked indicated antibody labeling, global fragment performed. These protocol sufficient completely alter final illustrate only identify best-fitting location confidence interval, but consider conformational variability docked. Given current rapid size, productivity, community, timely suggest guidelines validating, annotating, depositing models. There opportunity synergize experimental efforts bringing respective communities. establish standards share software databases. Similar (Read 2011Read R.J. Adams P.D. Arendall W.B. Emsley Joosten R.P. Kleywegt G.J. Krissinel E.B. Lütteke Otwinowski Z. al.Structure. 1395-1412Abstract (74) nuclear magnetic resonance (NMR) spectroscopy (http://www.wwpdb.org/workshop/2010/nmr_validation.html), communities (Schwede 2009Schwede Honig Berman H.M. Jones Brenner S.E. Burley S.K. Das Dokholyan N.V. 17: 151-159Abstract (48) serve constructive examples. Twenty eight participants 19 academic institutions worldwide attended 2010 (http://vtf.emdatabank.org). discussed issues computation ways strengthen collaboration participants' consensus formulated recommendations, aimed On first day, 14 presentations focused computing raw given. second "map" "model" discussion groups asked address questions respectively, report findings, make future. two referred Map Group Model respectively throughout rest review. We now summarize reached among meeting. concerned derivation, annotation, archiving, distribution, publication based intended starting point refinement broader community interested EM. Group's began enumerating attributes map, most method prepare sample symmetry object examined. prepared ways, here we describe four techniques. First, negatively stained samples usually adsorption onto carbon film, followed washing few drops stain, 1% uranyl acetate, then dried, interest seen low-density regions been excluded. Second, ice-embedded normally blotting thin film solution containing interest, plunge-freezing liquid ethane temperature just above its freezing (Dobro 2010Dobro Melanson Jensen McDowall A.W. Methods Enzymol. 481: 63-82Crossref (19) 1988Dubochet Adrian Chang J.J. Homo Schultz 1988; 129-228Crossref case against background ice. Third, samples, 2D crystals, continuous medium glucose, trehalose, tannic acid (Unwin 1975Unwin 94: 425-440Crossref (501) treatment preserves high-resolution diffraction order, contrast matching low obscure envelope. sections tissue plastic embedding (Glauert Lewis, 1998Glauert Lewis P.R. Specimen Preparation Transmission Microscopy.in: Glauert Practical Princeton London1998Crossref high pressure cryo-sectioning (Ladinsky, 2010Ladinsky 165-194Crossref (9) tomographic collection Plastic requires care fixation artifacts, produce compression artifacts; nonetheless, widely valuable methods. nature depends principally Thus, obtained crystallographic single-particle approaches (Frank 1988Frank Degn 345-360Crossref (16) Scholar); helical Fourier-Bessel (Diaz 2010Diaz Rice W.J. Stokes 482: 131-165Crossref (13) (Egelman, 2010Egelman E.H. 167-183Crossref (26) symmetries no (Rochat Chiu, 2012Rochat R.H. Cryo-electron tomography virus particles.in: Egelman Comprehensive Biophysics. 2012Crossref (2) finally tomogram sub-tomogram average (Schmid Booth, 2008Schmid Booth C.R. 161: 243-248Crossref (46) clear needs assessing accuracy satisfactory does yet exist, open research problem. However, number necessary validity, detect whether incorrect certain circumstances. majority require least reconstruction post-processing, masking filtration. Most access portion metadata reconstruction. additional explicitly Examples below. cannot experiment resolve chiral directly map. Tilt experiments offer validate help place limits orientation accuracy. include random-conical (Radermacher, 1988Radermacher Microsc. Tech. 359-394Crossref orthogonal (Leschziner Nogales, 2006Leschziner A.E. Nogales 2006; 153: 284-299Crossref (51) (Baumeister 1999Baumeister Grimm Walz Trends Cell 81-85Abstract (206) tilt-pair parameter plots 2011Henderson Chen J.Z. Grigorieff Passmore Ciccarelli Rubinstein J.L. Stewart P.L. Rosenthal P.B. 413: 1028-1046Crossref 2003Rosenthal 333: 721-745Crossref (342) web-based service (https://cryoem.nimr.mrc.ac.uk/software/). established comparison component whose hands previously. availability subunits, domains within determination (Baker 2003Baker Bowman B.R. Quiocho F.A. 331: 447-456Crossref (23) Kanamaru 2002Kanamaru Leiman P.G. Kostyuchenko V.A. Chipman Mesyanzhinov V.V. Arisaka 415: 553-557Crossref (208) 2004Leiman 118: 419-429Abstract (135) capsid surface lattice (such T = 7l) easily distinguished arrangement hexameric capsomeres freeze-fracture, metal shadowed particles (Prasad 1993Prasad B.V. Prevelige P.E. Marietta R.O. Thomas King 1993; 231: 65-74Crossref Additional ensuring agreement projections (if generated) class-averages, reference-free class-averages fully represented projections, coverage orientations (Orlova 1996Orlova E.V. Heel 1996; 3: 547-552Crossref Tang 2007Tang Peng Mann D.S. Rees 157: 38-46Crossref (374) criteria represent reliable variance local determination, bootstrap-based (Penczek 2006Penczek Spahn C.M. 154: 184-194Crossref Fourier Shell Correlation (FSC) measurements (Ménétret 2007Ménétret Schaletzky Clemons Jr., W.M. Osborne A.R. Skånland S.S. Denison Gygi S.P. Kirkpatrick Park al.Mol. 28: 1083-1092Abstract measures interpret structures. better 20 0.5 FSC criterion, RMEASURE (Sousa Grigorieff, 2007Sousa 201-210Crossref (63) estimate signal-to-noise correlation neighboring Transform terms. Possible bias overfitting noise should estimated statistics possible. Experimentalists encouraged own according listed To whole, EMDataBank table existing description what required technique, circumstances und