From Electron Crystallography to Single Particle Cryoem (Nobel Lecture)
Total Page:16
File Type:pdf, Size:1020Kb
See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/326270021 From Electron Crystallography to Single Particle CryoEM (Nobel Lecture) Article in Angewandte Chemie International Edition · July 2018 DOI: 10.1002/anie.201802731 CITATIONS READS 33 147 1 author: Richard Henderson Medical Research Council (UK) 193 PUBLICATIONS 26,896 CITATIONS SEE PROFILE Some of the authors of this publication are also working on these related projects: CryoEM View project All content following this page was uploaded by Richard Henderson on 11 November 2020. The user has requested enhancement of the downloaded file. Angewandte Nobel Lectures Chemie International Edition:DOI:10.1002/anie.201802731 Cryo-ElectronMicroscopy German Edition:DOI:10.1002/ange.201802731 From Electron Crystallography to Single Particle CryoEM (Nobel Lecture)** RichardHenderson* Keywords: biophysics ·cryoEM ·electron microscopy · Nobel lecture ·structural biology Introduction and Background in X-ray Crystallog- Bacteriorhodopsin at 7 ,then 3.5 ,Refinement raphy and Kinetics After completing an undergraduate physics degree at Following my return to the MRC-LMB,Igave atalk in the Edinburgh University in 1966, and deciding to pursue Ph.D. annual laboratory symposium in October 1973 about my ideas research in biophysics,Ihad the good fortune to consult for trying to solve the structure of bacteriorhodopsin. Since Professor Bill Cochran who suggested Iwrite to Max Perutz, bacteriorhodopsin had been shown[6] to consist of well- at that time head of the recently opened Medical Research ordered two-dimensional (2D) crystals in the membranes of Council Laboratory of Molecular Biology (MRC-LMB) in H. halobium,Ihad two ideas.One was to use X-ray powder Cambridge.Perutz offered me a3-year MRC Scholarship to diffraction of these native membranes with multiple heavy work with David Blow on the proteolytic enzyme chymo- atom derivatives to phase and resolve the problem of trypsin. Iarrived just as the chymotrypsin group was overlapping reflections.The other was to make 3D crystals calculating a3-dimensional (3D) Fourier map using two from detergent-solubilised monomeric bacteriorhodopsin. heavy-atom derivatives for phasing.Unfortunately,that first Neither of these ideas worked out, but in the same symposium map was only partly interpretable,with electron density for Iheard an impressive talk by Nigel Unwin about his work to only 10 of the 241 amino-acid residues recognisable,and since record high-quality electron microscope images of negatively Brian Matthews was just leaving for anew postdoctoral stained tobacco mosaic virus (TMV) using aphase plate that position at the NIH, Iwas invited to join the “chymotrypsin he had constructed from asingle thread of spider web silk team”, in which Paul Sigler was the only other scientist, to coated with gold. Afterwards,wediscussed the possibility of help determine the structure.After about 6months work recording images and electron diffraction patterns from 2D collecting data for athird heavy-atom derivative,the next 3D crystals of bacteriorhodopsin without using negative stain. A Fourier map proved to be fully interpretable,soIfound very productive 18-month collaboration ensued, culminating myself soon after my arrival in Cambridge transformed into in the determination of the 7 3D structure of bacteriorho- atrained X-ray crystallographer and co-author of apaper[1] dopsin,[7,8] shown in Figure 1and Figure 2, determined using describing the 3D structure of chymotrypsin. At the end of electron diffraction and electron microscopy of 2D crystals of that first year, Ithen embarked on my thesis research into bacteriorhodopsin at room temperature embedded in athin substrate and inhibitor binding to chymotrypsin, working film of glucose.Nigel and Iwondered why this electron initially alongside and then in collaboration with TomSteitz, crystallographic method had produced a3Ddensity map at who had arrived as apostdoctoral fellow that summer. By only 7 resolution, when there was nothing about the 1969 we had obtained anumber of informative 3D difference approach that intrinsically limited the resolution. We thought Fourier maps that allowed us to understand substrate and that the recording of images on film might be alimiting factor inhibitor binding to chymotrypsin and to explain the hydro- and spent time investigating different photographic emul- lytic mechanism.[2,3] sions.Wealso thought that the film scanners that were My transition from X-ray crystallographer to electron available in the 1970s for digitising the images might be crystallographer followed indirectly from my postdoctoral degrading the information and spent time building and experiences at Yale where Ihad decided to work on improving film scanners.This produced only fairly small membrane protein structure and had tried to tackle voltage- improvements. gated sodium channels (VGSCs) from garfish olfactory nerves.Ihad found that the VGSCs,assayed by atritiated- tetrodotoxin ligand-binding assay were unstable after solubi- [4] [*] R. Henderson lisation in detergent, so had switched to working on the MRC Laboratory of Molecular Biology small, stable and abundant membrane protein bacteriorho- Francis Crick Avenue, Cambridge CB2 0QH (UK) dopsin that had been discovered by Walther Stoeckenius and E-mail:[email protected] his collaborators in the purple membrane fraction from H. [**] Copyright The Nobel Foundation 2017. We thank the Nobel halobium.[5,6] Foundation, Stockholm,for permission to print this lecture. 10804 www.angewandte.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Angew.Chem. Int. Ed. 2018, 57,10804 –10825 Angewandte Nobel Lectures Chemie Figure 2. The structure of bacteriorhodopsin at 7 resolution in 3D from 18 images and 15 diffraction patterns.The collage shows (a) freeze-fracture picture from Walther Stoeckenius, (b) electron dif- fraction pattern obtained much later using aphosphor/fibre-optics/ CCD camera, (c) the 1975 balsawood model of asingle bacterio- rhodopsin molecule.[8] Figure 1. The first projection structure at 7 resolution of the purple rattlesnake venom crotoxin using an electron microscope in membrane calculated in October 1974 using 36 reflections obtained by Berlin with aliquid-helium superconducting objective lens, room-temperature electron diffraction and imaging of glucose-embed- ded 2D crystals of bacteriorhodopsin.[7] that Ibecame convinced electron cryomicroscopy could produce high quality images.Wetherefore embarked, as alast resort, on using electron cryomicroscopy for high- At that stage,having come into structural biology through resolution phase determination (see Figure 4). In earlier X-ray diffraction in which all the phases of the Fourier years,Bob Glaesersgroup had shown that freezing thin 3D components,asobserved through Bragg diffraction from the crystals of catalase could produce good electron diffraction crystal lattice,had to be determined indirectly,Ialso thought patterns and images[14,15] and that there was abenefit in terms that electron diffraction was intrinsically more promising than of reduced radiation damage,[16] but Ihad been unconvinced electron microscopy because the elegant simplicity of record- by earlier attempts to show that electron cryomicroscope ing electron diffraction patterns compared favourably with images of purple membrane contained high-resolution infor- the multiple difficulties of recording good images.Wethere- mation.[17] fore spent several years trying to extend the resolution of the Thechange of emphasis from diffraction to imaging bacteriorhodopsin structure using anumber of diffraction- proved to be very challenging.Ibegan with avisit to Jacques based approaches.Figure 3summarises the different ideas we Dubochetslaboratory at the European Molecular Biology tried. TomCeska tried to make heavy atom derivatives.[10] Laboratory (EMBL) in Heidelberg in 1984 working with Jean Joyce Baldwin and Michael Rossmann tried molecular Lepault to record images on their hybrid Zeiss/Siemens replacement.[11,12] David Agard tried to extend the phases microscope with the same design of superconducting liquid- using amulti-parameter model building approach (unpub- helium objective lens as on the Berlin microscope.Wespent lished). Although all of these approaches gave hints of success aweek with that home-constructed microscope,which turned that were encouraging at times,none of them were powerful out to be very unreliable.Fortunately,wemanaged to obtain enough to give phases that resulted in convincing maps that just one image that showed diffraction beyond 4 resolution, were interpretable much beyond the resolution obtained in although because of the difficulty of alignment and the short 1975. It was not until Tzyy-Wen Jeng and WahChiu mean time between failures,that image had over 5000 of demonstrated, in acollaboration with Fritz Zemlin,[13] that astigmatism. We did not pursue further imaging at EMBL. images showing clearly visible diffraction spots at 3.9 Nevertheless,that was the first image that allowed us to begin resolution could be obtained from thin 3D crystals of developing procedures for the computer-based processing of Angew.Chem. Int.Ed. 2018, 57,10804 –10825 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim 10805 Angewandte Nobel Lectures Chemie Figure 3. Overview of methods used in the early 1980s to try to solve the structure of bacteriorhodopsin at high resolution. (a) An optical diffraction pattern of ahigh-resolution projection image