Cryo-EM structure of haemoglobin at 3.2 Å determined with the Volta phase plate1
2
Maryam Khoshouei1
, Mazdak Radjainia2,3
, Wolfgang Baumeister1
& Radostin Danev1*
3
1 - Department of Molecular Structural Biology, Max Planck Institute of Biochemistry, 821524
Martinsried, Germany.5
2 - The Clive and Vera Ramaciotti Centre for Cryo-EM, Department of Biochemistry and Molecular6
Biology, Monash University, Victoria, 3800 Melbourne, Australia.7
3 – Present address: FEI, 5651 GG Eindhoven, The Netherlands.8
*Correspondence should be addressed to R.D. (danev@biochem.mpg.de).9
With the advent of direct electron detectors, the perspectives of cryo-electron microscopy10
(cryo-EM) have changed in a profound way1
. These cameras are superior to previous11
detectors in coping with the intrinsically low contrast of radiation-sensitive organic materials12
embedded in amorphous ice, and so they have enabled the structure determination of several13
macromolecular assemblies to atomic or near-atomic resolution. According to one14
theoretical estimation, a few thousand images should suffice for calculating the structure of15
proteins as small as 17 kDa at 3 Å resolution2
. In practice, however, we are still far away from16
this theoretical ideal. Thus far, protein complexes that have been successfully reconstructed17
to high-resolution by single particle analysis (SPA) have molecular weights of ~100 kDa or18
larger3
. Here, we report the use of Volta phase plate in determining the structure of human19
haemoglobin (64 kDa) at 3.2 Å. Our results demonstrate that this method can be applied to20
complexes that are significantly smaller than those previously studied by conventional21
defocus-based approaches. Cryo-EM is now close to becoming a fast and cost-effective22
alternative to crystallography for high-resolution protein structure determination.23
Given the radiation sensitivity of ice-embedded proteins, the low signal-to-noise ratio (SNR) of cryo-24
EM images is a limitation for SPA4
, restricting the size range of proteins that can be studied. In 1995,25
it was estimated, based solely on physical considerations, that the lower molecular weight limit of26
single particle cryo-EM would be 38 kDa5
. These considerations presumed the use of a perfect phase27
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
plate. It was suggested that the structure of 100 kDa proteins could be determined at 3 Å resolution28
from only 10,000 particles. Later, it was proposed that the theoretical limit might be even lower if29
perfect images could be taken2, 6
. With the technology at that time and cryo-EM images being far30
from perfect retaining only 10% of contrast, it seemed that obtaining a 3 Å reconstruction would be31
reserved for complexes with a molecular weight upwards of 4 MDa5
. Nowadays, obtaining ~3 Å32
resolution reconstructions has become almost routine and has been achieved with complexes that33
are much smaller in size1
. To date, the smallest protein solved to near-atomic resolution by single34
particle cryo-EM is the 3.8 Å resolution structure of the 93 kDa isocitrate dehydrogenase3
. Even so,35
single particle analysis reconstructions are still strongly biased towards larger symmetric complexes,36
indicating there is still a long way to go before the full potential of imaging proteins with electrons is37
reached.38
The difficulties in routinely obtaining high-resolution reconstructions of small molecular weight39
proteins are predominantly owed to poor representation of low spatial frequencies in electron40
micrographs obtained by conventional transmission electron microscopy (CTEM)4
. CTEM utilises41
phase contrast produced by spherical aberration (Cs) and the deliberate defocusing of the42
microscope’s objective lens. This approach creates oscillations in the contrast transfer function43
(CTF) of the microscope with some spatial frequencies of the object being transferred poorly, or not44
at all. One can compensate for this effect by varying the level of defocus from image to image, which45
is typically in the range of several hundreds to thousands of nanometres. By combining images that46
have different levels of contrast for given spatial frequencies an accurate representation of an object47
can be obtained. Nevertheless, the limitations due to reduced SNR resulting from contrast loss48
remain.49
In-focus single particle cryo-EM enabled by the Volta phase plate (VPP) holds the promise of yielding50
up to a two-fold boost in SNR and therefore enhancing our ability to observe weak phase objects7
.51
The SNR of VPP images is high because transfer of contrast of low spatial frequencies is optimal52
and constant for images taken in focus. Unlike previous phase plate designs, VPP images also retain53
the high spatial frequencies of the specimen enabling structure determination at near-atomic54
resolution8, 9
. However, in-focus imaging with VPP requires very precise focusing8
and the typically55
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
strong Cs present in cryo-electron microscopes appears to be a limiting factor in attaining resolutions56
better than 3 Å by in-focus phase plate-imaging8
.57
Enabled by the ability to estimate and correct the phase shift of the VPP in CTFFIND410
and58
RELION211
, we therefore used a hybrid approach combining the strengths of CTEM and VPP12
(Fig.59
1). This involves applying a defocus of ~500 nm and correcting for the effects of CTF. We opted to60
apply this strategy to tetrameric Hgb, which mediates oxygen transport in blood and has a molecular61
weight of 64 kDa and C2 symmetry. We chose Hgb for its iconic status as the first protein structure62
alongside myoglobin that was solved using X-ray crystallography by Max Perutz in 1960, coincidently63
by overcoming the phase problem of X-ray crystallography13
.64
Commercially sourced human Hgb is in the non-functional ferric (Fe3+
) state referred to as metHgb.65
After vitrification of the metHgb, the sample was subjected to VPP-enabled imaging with multiframe66
movies taken at low defocus, as described above. The movies were corrected for motion and67
radiation damage using MotionCor214
. Hgb particles were readily discernible in VPP images (Fig.68
1a) and could be accurately picked because of their high contrast. 2D classification of automatically69
picked particles resulted in class averages with recognisable features and striking resemblance to70
the structure of Hgb (Fig. 1c). Class averages were selected for initial model building in EMAN using71
the common-line technique and taking advantage of the C2 symmetry15
. RELION 3D classification72
and refinement using half-split datasets of particles yielded the final map16
. The obtained 3D73
reconstruction had a resolution of 3.2 Å, as determined by the so-called “gold-standard” FSC=0.14374
criterion.75
At this level of resolution, side-chain densities and prosthetic haem groups are clearly resolved in76
our reconstruction (Fig. 2a, Fig. S1). We used an MD-based approach for model building and77
compared our atomic model with three conformers of ferrous (Fe2+
) Hgb present in a single crystal78
(PDB 4N7O) adopting the tight (T) and two relaxed states (R1/R2)17
. Rigid-body fitting was used to79
dock the α1 subunits yielding a good visual fit with a cross-correlation value of ~69%.80
Superimposition of docked α1 subunits and corresponding tetramers yields cross-correlation values81
of 43%, 47% and 62% for T, R1 and R2 states, respectively (Fig. 2b). This observation is in line with82
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
the fact that metHgb adopts an R-like state suggesting that conformational states can be determined83
for small proteins at high-resolution without crystallisation.84
Our results showcase how cryo-EM can be used to determine which conformational states are85
present in solution. It has become increasingly clear that simple allosteric models based on discrete86
states, which are arrested by tight crystal contacts potentially fail to provide a complete87
structure/function portrait and may be divergent from solution studies17
. Single particle analysis is88
inherently better suited than crystallography for visualising the full spectrum of conformational states89
that proteins adopt18
. Obtaining high-resolution structures of solution states may indeed be one of90
the main applications of structure determination by VPP as a technique complementary to X-ray91
crystallography and nuclear magnetic resonance spectroscopy.92
Given the ease of the data acquisition, it can be expected that near-atomic resolution maps will93
become routine for large parts of the proteome including membrane proteins. In conjunction with94
improved automation, and next generation direct electron detectors, cryo-EM is likely to become a95
major player in structure-based small molecule drug discovery for almost any drug target.96
97
Methods98
Sample preparation99
Human Hgb was commercially sourced (Sigma-Aldrich, St. Louis, MO). Frozen-hydrated specimens100
were prepared on plasma-cleaned Quantifoil R1.2/1.3 holey carbon EM grids (Quantifoil,101
Großlöbichau, Germany) using a Vitrobot Mark III (FEI, Hillsboro, OR) 5 s blotting time, 85% humidity102
and -5 mm blotting offset.103
Data acquisition104
Automated data collection was performed on a Titan Krios electron microscope (FEI, Hillsboro, OR)105
operated at 300 kV and equipped with a K2 Summit direct detector, a Quantum energy filter (Gatan,106
Pleasanton, CA) and an FEI Volta phase plate (FEI, Hillsboro, OR) using SerialEM software. Movies107
comprising 40 frames and a total dose of 40 eper
Å2
were recorded on a K2 Summit direct detection108
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
camera (Gatan) at a calibrated magnification of 95,200 corresponding to a magnified pixel size of109
0.525 Å.110
Data processing111
The recorded movies were subjected to motion correction with MotionCor214
. Particles were picked112
with Gautomatch (developed by Zhang K, MRC Laboratory of Molecular113
Biology, Cambridge, UK, http://www.mrc-lmb.cam.ac.uk/kzhang/Gautomatch/). Subsequently, the114
283,600 picked particles were extracted in Relion2 using a box size of 100 pixels11
. After performing115
2D classification in Relion2, the best-looking 2D class averages were selected to build an initial116
model in EMAN using the common-lines approach15
. After two rounds of 3D classification 164,300117
particles from the best looking class were subjected to 3D auto-refinement in Relion2. The final map118
was sharpened using a negative B-factor of 200. Local resolution was calculated with blocres from119
the Bsoft package19
. Flexible fitting of the Hgb crystal structure was performed using the NAMD120
routine in MDFF20
followed by real-space refinement in PHENIX21
. The data collection, refinement121
parameters and model statistics are summarised in Table 1.122
123
References124
125
1. Nogales, E. & Scheres, S.H. Cryo-EM: A Unique Tool for the Visualization of126
Macromolecular Complexity. Mol Cell 58, 677-689 (2015).127
2. Glaeser, R.M. Review: electron crystallography: present excitement, a nod to the past,128
anticipating the future. J Struct Biol 128, 3-14 (1999).129
3. Merk, A. et al. Breaking Cryo-EM Resolution Barriers to Facilitate Drug Discovery. Cell 165,130
1698-1707 (2016).131
4. Sigworth, F.J. Principles of cryo-EM single-particle image processing. Microscopy (Oxf) 65,132
57-67 (2016).133
5. Henderson, R. The potential and limitations of neutrons, electrons and X-rays for atomic134
resolution microscopy of unstained biological molecules. Q Rev Biophys 28, 171-193 (1995).135
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
6. Rosenthal, P.B. & Henderson, R. Optimal determination of particle orientation, absolute hand,136
and contrast loss in single-particle electron cryomicroscopy. J Mol Biol 333, 721-745 (2003).137
7. Danev, R., Buijsse, B., Khoshouei, M., Plitzko, J.M. & Baumeister, W. Volta potential phase138
plate for in-focus phase contrast transmission electron microscopy. Proc Natl Acad Sci U S A139
111, 15635-15640 (2014).140
8. Danev, R. & Baumeister, W. Cryo-EM single particle analysis with the Volta phase plate.141
Elife 5 (2016).142
9. Khoshouei, M. et al. Volta phase plate cryo-EM of the small protein complex Prx3. Nat143
Commun 7, 10534 (2016).144
10. Rohou, A. & Grigorieff, N. CTFFIND4: Fast and accurate defocus estimation from electron145
micrographs. bioRxiv (2015).146
11. Kimanius, D., Forsberg, B.O., Scheres, S. & Lindahl, E. Accelerated cryo-EM structure147
determination with parallelisation using GPUs in RELION-2. bioRxiv (2016).148
12. Danev, R., Tegunov, D. & Baumeister, W. Using the Volta phase plate with defocus for cryo-149
EM single particle analysis. bioRxiv (2016).150
13. Perutz, M.F. et al. Structure of haemoglobin: a three-dimensional Fourier synthesis at 5.5-A.151
resolution, obtained by X-ray analysis. Nature 185, 416-422 (1960).152
14. Zheng, S., Palovcak, E., Armache, J.-P., Cheng, Y. & Agard, D. Anisotropic Correction of153
Beam-induced Motion for Improved Single-particle Electron Cryo-microscopy. bioRxiv,154
061960 (2016).155
15. Tang, G. et al. EMAN2: an extensible image processing suite for electron microscopy. J Struct156
Biol 157, 38-46 (2007).157
16. Scheres, S.H. RELION: implementation of a Bayesian approach to cryo-EM structure158
determination. J Struct Biol 180, 519-530 (2012).159
17. Shibayama, N., Sugiyama, K., Tame, J.R. & Park, S.Y. Capturing the hemoglobin allosteric160
transition in a single crystal form. J Am Chem Soc 136, 5097-5105 (2014).161
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
18. Bai, X.C., Rajendra, E., Yang, G., Shi, Y. & Scheres, S.H. Sampling the conformational space162
of the catalytic subunit of human gamma-secretase. Elife 4 (2015).163
19. Heymann, J.B. & Belnap, D.M. Bsoft: image processing and molecular modeling for electron164
microscopy. Journal of structural biology 157, 3-18 (2007).165
20. Chan, K.Y., Trabuco, L.G., Schreiner, E. & Schulten, K. Cryo-electron microscopy modeling166
by the molecular dynamics flexible fitting method. Biopolymers 97, 678-686 (2012).167
21. Adams, P.D. et al. PHENIX: a comprehensive Python-based system for macromolecular168
structure solution. Acta Crystallogr D Biol Crystallogr 66, 213-221 (2010).169
170
Accession codes171
The cryo-EM map and atomic coordinates of Hgb were deposited to the Electron Microscopy Data172
Bank (EMDB) and Protein Data Bank (PDB) under accession codes EMD-3488 and PDB-5ME2,173
respectively. Raw data was made available at the Electron Microscopy Pilot Image Archive174
(EMPIAR) with accession code EMPIAR-10105.175
Acknowledgments176
We thank Prof. Jürgen Plitzko for his technical support. We also acknowledge Dr. Alexis Rohou177
and Sjors Scheres for support with CTFFIND4 and Relion2, respectively. We thank Dr. Matthew178
Belousoff for model building of the Hgb atomic model. We also thank Dr. Mike Strauss and Dr.179
Shelley Robison for critical reading of the manuscript. This work was supported by the Multi-modal180
Australian ScienceS Imaging and Visualisation Environment (www.massive.org.au).181
Author contributions182
MK, MR and RD were responsible for the conception, design, data analysis and interpretation of183
experiments. MK performed sample preparation and imaging. All authors wrote the manuscript.184
Competing financial interests185
RD is a co-inventor in US patent US9129774 B2 “Method of using a phase plate in a transmission186
electron microscope”. WB is on the Scientific Advisory Board and MR an employee of FEI Company.187
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
188
Figure 1. Phase plate-imaging of 64 kDa Hgb. (a) Electron micrograph of metHgb recorded at ~500189
nm underfocus with the Volta phase plate (VPP) (scale bar = 30nm). (b) Power spectrum of the190
image in (a), featuring contrast transfer function (CTF) Thon rings permitting defocus and phase shift191
estimation. (c) 2D class averages of Hgb showing secondary structure elements in projection. (d)192
Reconstructed 3D map and model of Hgb. (e) (Gold Standard?) Fourier shell correlation (FSC) plot193
indicating a resolution of 3.2 Å according to the FSC=0.143 criterion.194
195
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
196
Figure 2. Hgb at 3.2 Å resolution. (a) The iron atom of the prosthetic haem group is coordinated by197
the proximal histidine residue, as evidenced by a strong density connecting them. (b) VPP198
reconstruction fitted with 3 conformers of Hgb present in crystal structure PDB 4N7O. The199
reconstructed 3D map agrees best with chains A-D of PDB 4N7O representing the R2 state of Hgb.200
201
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
202
Figure S1. (a) Local resolution estimation. (b) Magnified details of the 3D map with refined model.203
204
.CC-BY 4.0 International licensenot peer-reviewed) is the author/funder. It is made available under a
The copyright holder for this preprint (which was.http://dx.doi.org/10.1101/087841doi:bioRxiv preprint first posted online Nov. 15, 2016;
Table 1
Data Collection
Particles used in final 3D refinement 175,374
Pixel size (Å) 0.525
Defocus (µm) -0.5
Voltage (kV) 300
Electron dose (e-
Å-2
) 40
Rms deviations
Bonds (Å) 0.009
Angles (°) 0.808
Validation
Clashcore, all atoms 4.5
Good outliers (%) 0.0
Ramachandran plot
Favored (%) 97.13
Allowed (%) 2.87
Outliers (%) 0.0
Refinement
Resolution (Å) 3.2
Map sharpening B-factor (Å2
) -200
Fourier Shell Correlation 0.143