- Email: [email protected]
Analysis of nano-crystals: Evaluation of heavy metal-embedded biological specimen by high voltage electron microscopy
Accepted Manuscript
Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy Hyeongseop Jeong , Seung Jo Yoo , Jonghan Won , Hyun-Ju Lee , Jeong Min Chung , Han-ul Kim , Gwang Joong Kim , Jin-Gyu Kim , Hyun Suk Jung , Jaekyung Hyun PII: DOI: Reference:
S0304-3991(18)30141-4 10.1016/j.ultramic.2018.07.003 ULTRAM 12609
To appear in:
Ultramicroscopy
Received date: Revised date: Accepted date:
19 April 2018 3 July 2018 7 July 2018
Please cite this article as: Hyeongseop Jeong , Seung Jo Yoo , Jonghan Won , Hyun-Ju Lee , Jeong Min Chung , Han-ul Kim , Gwang Joong Kim , Jin-Gyu Kim , Hyun Suk Jung , Jaekyung Hyun , Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy, Ultramicroscopy (2018), doi: 10.1016/j.ultramic.2018.07.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Highlights
High voltage electron microscopy (HVEM) visualized nano-crystals surrounding of heavy metal-embedded biological specimen
CR IP T
Electron crystallography determined chemical composition of the stain molecules
The grain size of nano-crystals of stained molecules from uranyl acetate
AN US
and uranyl formate were nearly identical, and consistent with that of uranium dioxide (UO2)
It identified that UO2 is the main contributor of image contrast of heavy
AC
CE
PT
ED
M
metal-embedded biological specimen
1
ACCEPTED MANUSCRIPT
Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy
Hyeongseop Jeong1*, Seung Jo Yoo1*, Jonghan Won2, Hyun-Ju Lee1, Jeong Min Chung4,
CR IP T
Han-ul Kim4, Gwang Joong Kim4, Jin-Gyu Kim1, Hyun Suk Jung4† and Jaekyung Hyun1,3†
1
Electron Microscopy Research Center, Korea Basic Science Institute, Chungcheongbukdo 28119,
Republic of Korea 2
Advanced Nano-Surface Research Team, Korea Basic Science Institute, Daejeon 34113, Republic of
AN US
Korea 3
Department of Bio-analytical Science, Korea University of Science and Technology, Daejeon 34113,
Republic of Korea 4
Department of Biochemistry, College of Natural Sciences, Kangwon National University, Gangwon-do,
M
24341, Republic of Korea
ED
These authors equally contributed to this work
PT
*
†
Corresponding author: Hyun Suk Jung
CE
Department of Biochemistry, College of Natural Sciences, Kangwon National University, 1 Kangwondaehak-gil, Chuncheon-si, Gangwon-do, 24341, Republic of Korea
AC
Phone: +82 33 250 8513; Fax: +82 33 259 5664 E-mail: [email protected]
†
Corresponding author: Jaekyung Hyun
Electron Microscopy Research Center, Korea Basic Science Institute, 161 Yeongudanji-ro, Ochang-eup, Cheongwon-gu, Cheongju-si, Chungcheongbuk-do 28119, Republic of Korea Phone: +82 42 865 3681; Fax: +82 42 865 3939 E-mail: [email protected] 2
ACCEPTED MANUSCRIPT ABSTRACT Heavy metal compounds are adsorbed onto biological specimen in order to enhance the contrast as well as to preserve the structural features of the specimen against electron beam-induced radiation damage. In particular, in combination with computational image
CR IP T
processing, negative staining is widely used for structural analysis of protein complexes to moderate resolutions. Image analysis of negatively stained biological specimen is known to suffer from limited achievable resolution due to dehydration and large grain size of staining
AN US
molecules although the extent of such effect remains somewhat dubious.
Stain molecules exist as grains under electron beam. However, clear observation of the crystalline nature of the grains and their association with biological specimen has not been thoroughly demonstrated. In this study, we attempted high-resolution TEM (HRTEM) using
M
high voltage electron microscopy and electron crystallography analysis for the detailed
ED
characterization of negatively stained biological specimen, focusing on physical state and chemical composition of the stain molecules. The electron crystallography analysis allowed
PT
for the identification of the crystal constituents of widely used stains, hence revealing the
CE
chemical nature and the morphology of the stain molecules at specimen level. This study
AC
re-evaluated generally accepted notions on negative staining, and may help correctly interpreting the structural analysis of stained biological specimen.
Keywords: High Voltage Electron Microscopy, Electron Crystallography, Negative Staining, Nano-crystal, Transmission Electron Microscopy, Uranium Dioxide
3
ACCEPTED MANUSCRIPT
Abbreviations: Transmission Electron Microscopy (TEM), High-resolution Transmission Electron Microscopy (HRTEM), Cryo-electron Microscopy (Cryo-EM), High-voltage Electron Microscope (HVEM), Selected Area diffraction (SAD), Inorganic Crystal Structure Database (ICSD), Uranyl Acetate (UA), Uranyl Formate (UF), Ammonium Molybdate (AM),
AC
CE
PT
ED
M
AN US
CR IP T
Sodium Phosphotungstate (SPT), Uranium Dioxide (UO2),
4
ACCEPTED MANUSCRIPT
1. INTRODUCTION Negative staining is one of the most widely used sample preparation methods for the visualization of biological specimen using transmission electron microscopy (TEM) [1, 2]. While the image contrast of biological specimen is poor due to weak electron scattering and
CR IP T
the use of minimal electron dose for the reduction of accumulated radiation damage, negative staining greatly improves image contrast due to enhanced electron scattering from heavy metal compounds adsorbed onto the specimen. In addition, the layer of stain molecules that
AN US
surround the biological specimen preserves the molecular features of the object under electron beam that otherwise completely destroys the structure. The technique is quick and easy to apply for most biological sample, and in particular, routinely employed for the structural
M
analysis of protein macromolecules to moderate resolutions as a preliminary experiment before high resolution structure determination using cryo-electron microscopy, or in cases
PT
heterogeneity [3-5].
ED
where cryo-EM is precluded due to insufficient molecular mass and inherent molecular
The structural analysis of protein macromolecules in three dimensions often involves
CE
computational image processing such as single particle analysis [6, 7], electron
AC
crystallography [8, 9] and helical reconstruction [10, 11]. The achievable resolution obtained from the images of negatively stained specimen is typically limited to approximately 10 to 20 Å [3, 4, 12, 13]. The major factor that contributes to the limited resolution is known to be associated with the deformation of fine structural features within the protein molecules that results from the staining and dehydration. By contrast, cryo-electron microscopy (cryo-EM) is devoid of both staining and dehydration, and hence is the method of choice for near-atomic
5
ACCEPTED MANUSCRIPT
resolution structure determination [14-16]. Cryo-negative staining may be used as an alternative to retain the advantage of negative staining while the sample deformation is minimized by vitrification, resulting in 3D reconstruction comparable to that of cryo-electron microscopy data [17]. However, the reason behind limited achievable resolution of typical
CR IP T
negative staining remains somewhat dubious. For instance, an electron diffraction study has demonstrated possibility of data information up to 4 Å from a negatively stained catalase 2D crystal [18], but such resolution has never been realized for structural analysis of proteins by
AN US
2D image analysis or 3D structure reconstruction. Another misleading factor is the definition of “grain” that denotes the fineness of staining molecules that is often used in analogy to the level of details that can be visualized. However, the nature of the grains, or its association
M
with biological molecules that could affect the resolution and fidelity of structural analysis from negatively stained specimen, are still poorly characterized.
ED
In this study we attempted high resolution TEM imaging and electron crystallography of a
PT
biological specimen stained with uranium compounds in order to characterize the behavior of staining molecules on a specimen grid. The results provide visual evidence that the heavy
CE
metal molecules exist as nano-crystals surrounding protein particles under high electron beam,
AC
and the chemical constituents of the crystals were unambiguously determined.
6
ACCEPTED MANUSCRIPT
2. MATERIALS AND METHODS 2.1. Protein purification and sample preparation for electron microscopy Vaccinia virus D13 was used as a test specimen for the analysis of the behavior of stain molecules adsorbed onto the biological sample. Homogeneous preparation of D13 trimer was
CR IP T
purified by immobilized metal affinity chromatography and size exclusion chromatography as described previously [8]. For electron microscopy, 5 μL of D13 solution was applied onto a glow discharged EM grid coated with amorphous carbon film which is typically used for the
AN US
negative staining of protein samples. After 1 minute of sample adsorption, the grid was washed with three droplets of staining solution to replace the buffer with the staining solution. After 1 minute, excess stain solution was blotted away using a piece of filter paper and then
M
the grid was dried in air. The specimen grid for the analysis of nano-crystals that results from staining was prepared in the same way, but without the protein sample loading, due to
ED
possible crystalline contamination that may originate from protein sample buffer. The staining
PT
compounds used in the analysis include 1% (w/v) uranyl acetate, 0.75% (w/v) uranyl formate, 2% (w/v) ammonium molybdate and 2% (w/v) sodium phosphotungstate. The samples were
CE
examined using a transmission electron microscope operating at 120kV equipped with LaB6
AC
beam source (Tecnai G2 Spirit TWIN, FEI Company).
2.2. HRTEM imaging and analysis of crystals A high-voltage electron microscope (HVEM, JEOL Ltd., JEM ARM 1300S) operating at 1250kV was used for the detailed observation of the negatively-stained grids, focusing on the direct visualization and the selected area diffraction of nano-crystal. HRTEM and the electron
7
ACCEPTED MANUSCRIPT
diffraction images of the negatively-stained protein sample and stain nano-crystals were analyzed using the Digital Micrograph software (Gatan Inc.). Fast Fourier transform of the individual nano-crystals in 256 x 256 pixel boxes were used to estimate the distance from the center to each diffraction spots and the lattice angles. The lattice parameters were compared to
CR IP T
simulated electron diffraction and the atomic model that correspond to the stain nano-crystals, uranium dioxide, that were generated using Crystal Maker 2.7 software (CrystalMaker Software Ltd.). Selected area diffraction (SAD) images of the stain crystals were used to
AN US
measure d-spacing of the crystal, and this experimental d-spacing profile was compared to the diffraction profiles of uranium derivatives obtained from FIZ Karlruhe inorganic crystal
M
structure database (ICSD).
2.3. Imaging of dose-dependent stain crystal formation
ED
Dose-dependent crystal formation of stain nano-crystal formation was examined using Titan
PT
Krios TEM (FEI Company) equipped with Falcon III direct electron detector (FEI Company), operating at 300kV and at room temperature. Negatively stained D13 was imaged at nominal
CE
magnification of x250,000 using dose rate of approximately 100 e-/Å2/s, over 300 seconds of
AC
exposure time. The acquisition was performed using movie frame imaging, of which 1204 frames were recorded hence each movie frame was exposed to approximately 25 e-/Å2.
8
ACCEPTED MANUSCRIPT
3. RESULTS AND DISCUSSIONS 3.1. High Resolution Imaging of Negatively Stained Biological Specimen The existence of stain crystal has been demonstrated elegantly at the very early stage of structural investigation of protein complexes by Unwin [19]. The study demonstrated that the
CR IP T
formation of stain crystal is electron dose-dependent, as evidenced by a series of electron diffraction images showing that the crystal formation is promoted by accumulating electron dose. We first attempted to visually inspect the physical state and the behavior of staining
AN US
molecules using a typical negative staining procedure, a test protein sample Vaccinia virus D13 was stained using 1% uranyl acetate, followed by air-drying. The sample was first examined using a TEM operating at 120kV using various magnifications and defocus levels.
M
However, no crystalline feature was identifiable, most likely to be due to limited resolving power. We then performed HRTEM imaging using a HVEM, from which stain nano-crystals
ED
were clearly visible. The crystals were heavily associated around protein molecules with no
PT
apparent pattern in their distribution (Figure 1, A and B). At approximately 1 μm under-focus, the features of crystalline lattice were no longer visible (Figure 1, C and D). The images of
CE
nano-crystals and respective fast Fourier transforms indicate that the crystalline features are
AC
easy to be missed under normal imaging conditions used in the analysis of protein macromolecules (i.e. low dose, low magnification, and relatively high under-focus), especially when HRTEM is not intentionally attempted. It was also found that the crystalline patches are most noticeable on the region where the stain accumulation was heavy, usually surrounding the edge of protein molecules or grooves within the structure, suggesting that the stain molecules bound to the protein molecules vary in the extent of crystallization.
9
ACCEPTED MANUSCRIPT
In this study, for experimental simplicity and image acquisition using HRTEM, applied electron dose far exceeded the typical accumulated dose recommended for structural analysis of protein particles using cryo-EM (20~30 electrons per Å2). Accordingly, the crystals seen in our study is most likely formed during the strong electron beam exposure onto the specimen
CR IP T
as demonstrated before [17]. To elaborate approximate electron dose required for the formation of pronounced nano-crystals, a series of beam exposure onto the specimen was performed using movie frame image acquisition where each frame was exposed to
AN US
approximately 25 e-/Å2. Gradual growth of the crystal was discernable from the images when the accumulated electron dose reached up to tens of thousands e-/Å2 (Figure S1). The observation suggests that the effect of nano-crystal formation on the structural fidelity of the
M
negatively stained particles could be marginal if care is taken not to over-expose the sample
ED
with the beam or, most certainly, if low dose imaging is employed.
PT
3.2. High Resolution Imaging of Staining Molecules on EM Grid We prepared EM grid without protein sample in order to avoid possible crystallization of the
CE
components in buffer solution that may interfere with clear examination of stain molecules.
AC
Some of the most commonly used stain solutions were subjected to HRTEM. These included uranyl acetate (UA, UO2(CH3COO)2), uranyl formate (UF, UO2(CHO2)2), ammonium molybdate (AM, (NH4)6Mo7O24) and sodium phosphotungstate (SPT, Na3O40PW12), dissolved in distilled water at concentrations used in routine protein sample preparation [2]. In all cases, small crystals were found all over the EM grid with varying sizes and crystalline lattices, from which the HRTEM images were collected (Figure 2). EM grids stained with UA and UF
10
ACCEPTED MANUSCRIPT
showed smallest crystals, typically smaller than 20nm in the longest dimension, in random orientations. The grids stained with other compounds also contained nano-crystals but micrometer-scale crystals with continuous lattice were often found. Considering typical achievable resolution of protein 3D reconstruction from negatively stained specimen is
CR IP T
approximately 10 to 20 Å, the size range of the crystals seen in the sample does may not be directly relevant to the resolution. Rather, crystallization of otherwise amorphous stain molecules will involve significant rearrangement of associated molecular structure of the
AN US
protein which, in consequence, causes additional deformation of the object. This conclusion was also drawn from initial evaluation of the effect of stain crystallization on the
M
reconstruction of stacked disk aggregates of Tobacco Mosaic virus protein [17].
3.3. Electron Crystallographic Analysis of Stain Nano-Crystals
ED
Crystals found in the EM grids stained with UA and UF were subjected to a further TEM
PT
analysis in order to analyze possible differences in their behaviors. The two compounds were chosen because they are derivatives from the same elements, but UF is known for finer
CE
granularity and better permeation that suits for high resolution analysis [20, 21]. In order to
AC
visually evaluate their physical state and chemical composition on the specimen, electron crystallography analysis was carried out. Small regions with best crystalline order, as judged by the sharpness and extension of diffraction spots in fast Fourier transform, from the field of a randomly chosen area within a grid were selected for electron crystallography analysis (Figure 3A and C). In the Fourier transforms, distance from the center of the pattern to each diffraction spot and the angle
11
ACCEPTED MANUSCRIPT
between lattice lines allowed us to obtain d-spacing profile and determine zone axis (Figure 3B and 3D), from which the identical diffraction profile was searched in the inorganic crystal structure database (ICSD). The diffraction patterns and d-spacing profiles of nano-crystals found on the grids stained with UA and UF were nearly identical, and consistent with that of
CR IP T
uranium dioxide (UO2) (figure 3E and 3F). In addition, the experimental diffraction profile agreed well with electron diffraction analysis of uranyl dioxide thin film reported elsewhere [22]. Because only a small number of crystalline patches were subjected to crystallographic
AN US
examination, the possibility of other types of crystals in the overall area still remained. Therefore, selected area diffraction (SAD) from larger field of view was recorded in order to make sure that the powder diffraction profile coincides with a single species (i.e. UO2)
M
(Figure 4). The resulting d-spacing profiles of UA and UF obtained from distinctive rings of intensity were identical with subtle differences, and it was clear that the majority of crystal
ED
constituents are indeed UO2.
PT
Taken together with physical association of stain crystals with protein molecules observed by HRTEM, the crystallographic analysis suggests that the degree of details identifiable from the
CE
use of both uranium compounds not only depends on stain granularity that is referred by the
AC
manufacturers and publications, but also by UO2 nano-crystals and consequential structure deformation of the specimen.
12
ACCEPTED MANUSCRIPT
4. CONCLUSIONS In this study, we addressed a fundamental question in the analysis of negatively stained biological sample; what are we really seeing in the image? We used HRTEM and electron crystallography, techniques that are not routinely employed for biological TEM, in order to
CR IP T
clarify the nature of stain molecules under electron beam. Our result demonstrate that “grain”, of which the size is commonly referred as an indicative of the degree of structural fineness retained in the image, is no longer valid when high enough electron dose promotes the
AN US
formation nano-crystals. In particular, we unambiguously identified that UO2 is the constituent of stain crystals in case of both UA and UF. Examination of chemical identity of stain molecules in low electron dose, that is before apparent formation of nano-crystals, as
M
well as direct comparison of resulting 3D reconstructions of negatively stained protein complexes at both low and high electron dose would provide further aid clarifying the relation
ED
between achievable resolution and negative staining. In the era of near-atomic cryo-EM
PT
structure determination, a careful revision of routine negative staining is still a valuable exercise that could improve the fidelity, and even to maximize resolution, of initial 3D
AC
CE
structure of biological macromolecules at moderate details.
13
ACCEPTED MANUSCRIPT
Acknowledgement This work was supported by Korea Basic Science Institute grant T38210 to J. Hyun, and in part by 2017 Research Grant from Kangwon National University (No. 520170496), Next-Generation BioGreen Program(SSAC, PJ013273042018) and Basic Science Research
CR IP T
Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (NRF-2015R1C1A1A01053611, 2018R1D1A1B07045580)
AN US
to H. S Jung.
Conflict of interest
AC
CE
PT
ED
M
The authors declare no conflict of interest arising from this work.
14
ACCEPTED MANUSCRIPT
References 1.
Rames, M., Y. Yu, and G. Ren, Optimized negative staining: a high-throughput protocol for examining small and asymmetric protein structure by electron microscopy. J Vis Exp, 2014(90): p. e51087. Harris, J.R., Negative Staining and Cryoelectron Microscopy 1997, New York: Taylor
CR IP T
2.
& Francis. 3.
Ohi, M., et al., Negative Staining and Image Classification - Powerful Tools in
4.
AN US
Modern Electron Microscopy. Biol Proced Online, 2004. 6: p. 23-34.
Bremer, A., et al., Has negative staining still a place in biomacromolecular electron microscopy? Ultramicroscopy, 1992. 46(1-4): p. 85-111.
Zhang, L., et al., Optimized negative-staining electron microscopy for lipoprotein
M
5.
studies. Biochim Biophys Acta, 2013. 1830(1): p. 2150-9. Zhou, Z.H., Towards atomic resolution structural determination by single-particle
ED
6.
7.
PT
cryo-electron microscopy. Curr Opin Struct Biol, 2008. 18(2): p. 218-28. Frank, J., Single-particle reconstruction of biological macromolecules in electron
Hyun, J.K., et al., Membrane remodeling by the double-barrel scaffolding protein of
AC
8.
CE
microscopy--30 years. Q Rev Biophys, 2009. 42(3): p. 139-58.
poxvirus. PLoS Pathog, 2011. 7(9): p. e1002239.
9.
Wisedchaisri, G., S.L. Reichow, and T. Gonen, Advances in structural and functional analysis of membrane proteins by electron crystallography. Structure, 2011. 19(10): p. 1381-93.
10.
Egelman, E.H., Reconstruction of helical filaments and tubes. Methods Enzymol,
15
ACCEPTED MANUSCRIPT
2010. 482: p. 167-83. 11.
Ge, P. and Z.H. Zhou, Hydrogen-bonding networks and RNA bases revealed by cryo electron microscopy suggest a triggering mechanism for calcium switches. Proc Natl Acad Sci U S A, 2011. 108(23): p. 9637-42. Kiselev, N.A., M.B. Sherman, and V.L. Tsuprun, Negative staining of proteins. Electron Microsc Rev, 1990. 3(1): p. 43-72.
13.
Durand, E., et al., Biogenesis and structure of a type VI secretion membrane core
AN US
complex. Nature, 2015. 523(7562): p. 555-60. 14.
Adrian, M., et al., Cryo-electron microscopy of viruses. Nature, 1984. 308(5954): p. 32-6.
Frank, J., Three-dimensional electron microscopy of macromolecular assemblies:
M
15.
CR IP T
12.
visualization of biological molecules in their native state. 2nd ed. Three-dimensional
ED
electron microscopy of macromolecular assemblies: visualization of biological
16.
Kuhlbrandt, W., Biochemistry. The resolution revolution. Science, 2014. 343(6178): p.
CE
1443-4.
De Carlo, S. and J.R. Harris, Negative staining and cryo-negative staining of
AC
17.
PT
molecules in their native state. 2006, New York: Oxford University Press.
macromolecules and viruses for TEM. Micron, 2011. 42(2): p. 117-31.
18.
Massover, W.H., P.F. Lai, and P. Marsh, Negative staining permits 4.0 A resolution with low-dose electron diffraction of catalase crystals. Ultramicroscopy, 2001. 90(1): p. 7-12.
19.
Unwin, P.N., Electron microscopy of the stacked disk aggregate of tobacco mosaic
16
ACCEPTED MANUSCRIPT
virus protein. II. The influence of electron irradiation of the stain distribution. J Mol Biol, 1974. 87(4): p. 657-70. 20.
Leberman, R., Use of uranyl formate as a negative stain. J Mol Biol, 1965. 13(2): p. 606. Knight, D.P., Negative staining of rat tail tendon collagen fibrils with uranyl formate.
CR IP T
21.
Tissue Cell, 1975. 7(4): p. 651-4. 22.
Lin, J., et al., Structure and properties of uranium oxide thin films deposited by pulsed
AC
CE
PT
ED
M
AN US
cd magnetron sputtering. Applied Surface Science, 2014. 301: p. 475-480.
17
ACCEPTED MANUSCRIPT
Table 1. Comparison between the experimental d-spacings of stain nano-crystals and the
Experimental d-spacing (UA)
Experimental d-spacing (UF)
(111)
0.315 nm
0.312 nm
0.320 nm
(002)
0.193 nm
0.192 nm
(113)
0.165 nm
0.164 nm
(024)
0.122 nm
0.122 nm
(224)
0.111 nm
AC
CE
PT
ED
M
0.112 nm
18
CR IP T
Plane
Reference d-spacing (UO2)
AN US
reference (UO2)
0.195 nm 0.164 nm 0.123 nm 0.112 nm
ACCEPTED MANUSCRIPT
Figures and figure legends
AN US
CR IP T
Figure 1.
M
Fig. 1. HRTEM images and the fast Fourier transforms of negatively-stained protein sample.
ED
(A) HRTEM image of negatively-stained D13, recorded at just below at-focus and (B) a higher magnification image. (C) Micrographs of same area, recorded at under-focus typically
PT
used for structural analysis of protein complexes (~1μm under-focus), and (D) a higher
CE
magnification image. Nano-scale crystals are found around the protein particles at close-to-focus imaging condition, whereas such features diminished at high defocus level.
AC
Fourier transforms are shown in the insets. Scale bars are 20 nm in (A) and (C), 5 nm in (B) and (D), 5 1/nm in the insets, respectively.
19
ACCEPTED MANUSCRIPT
AN US
CR IP T
Figure 2.
Fig. 2. HRTEM image and the fast Fourier transform of stain molecules on EM grid. HRTEM micrographs of EM specimen grids treated with (A) UA, (B) UF, (C) AM and (D) SPT in the
M
absence of protein sample. Crystals with varying sizes and lattice dimensions are clearly seen
ED
in the micrographs and respective FFTs. The Fourier transforms are shown in the insets. Scale
AC
CE
PT
bars are 5 nm in the micrographs, 10 1/nm in the insets, respectively.
20
ACCEPTED MANUSCRIPT
M
AN US
CR IP T
Figure 3.
ED
Fig. 3. HRTEM and electron crystallographic analysis of UA and UF nano-crystals. (A, C)
PT
HRTEM micrographs of UA and UF nano-crystals, and (B, D) respective Fourier transforms. (E) A simulated model of UO2 with a molecular unit cell demarked by a box, and (F) the
CE
corresponding Fourier transform. All zone axes of the selected nano-crystals in this figure are
AC
[1-10]. The HRTEM and crystallographic analysis indicates that UO2 is the main constituent of the stain nano-crystals. Unit cell dimensions are indicated in (B), (D) and (F). Scale bars are 2 nm in (A) and (C), 5 1/nm in (B) and (D), respectively.
21
ACCEPTED MANUSCRIPT
CR IP T
Figure 4.
Fig. 4. Selected area diffraction (SAD) analysis of stain nano-crsytals. (A) SAED images of
AN US
UA nano-crystals and (B) UF nano-crystals. The diffraction analysis unambiguously demonstrates that the majority of stain nano-crystals are composed of UO2. Scale bars are 5
AC
CE
PT
ED
M
1/nm.
22
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Graphical abstract
23
Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy Hyeongseop Jeong , Seung Jo Yoo , Jonghan Won , Hyun-Ju Lee , Jeong Min Chung , Han-ul Kim , Gwang Joong Kim , Jin-Gyu Kim , Hyun Suk Jung , Jaekyung Hyun PII: DOI: Reference:
S0304-3991(18)30141-4 10.1016/j.ultramic.2018.07.003 ULTRAM 12609
To appear in:
Ultramicroscopy
Received date: Revised date: Accepted date:
19 April 2018 3 July 2018 7 July 2018
Please cite this article as: Hyeongseop Jeong , Seung Jo Yoo , Jonghan Won , Hyun-Ju Lee , Jeong Min Chung , Han-ul Kim , Gwang Joong Kim , Jin-Gyu Kim , Hyun Suk Jung , Jaekyung Hyun , Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy, Ultramicroscopy (2018), doi: 10.1016/j.ultramic.2018.07.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Highlights
High voltage electron microscopy (HVEM) visualized nano-crystals surrounding of heavy metal-embedded biological specimen
CR IP T
Electron crystallography determined chemical composition of the stain molecules
The grain size of nano-crystals of stained molecules from uranyl acetate
AN US
and uranyl formate were nearly identical, and consistent with that of uranium dioxide (UO2)
It identified that UO2 is the main contributor of image contrast of heavy
AC
CE
PT
ED
M
metal-embedded biological specimen
1
ACCEPTED MANUSCRIPT
Analysis of Nano-crystals: Evaluation of Heavy Metal-Embedded Biological Specimen by High Voltage Electron Microscopy
Hyeongseop Jeong1*, Seung Jo Yoo1*, Jonghan Won2, Hyun-Ju Lee1, Jeong Min Chung4,
CR IP T
Han-ul Kim4, Gwang Joong Kim4, Jin-Gyu Kim1, Hyun Suk Jung4† and Jaekyung Hyun1,3†
1
Electron Microscopy Research Center, Korea Basic Science Institute, Chungcheongbukdo 28119,
Republic of Korea 2
Advanced Nano-Surface Research Team, Korea Basic Science Institute, Daejeon 34113, Republic of
AN US
Korea 3
Department of Bio-analytical Science, Korea University of Science and Technology, Daejeon 34113,
Republic of Korea 4
Department of Biochemistry, College of Natural Sciences, Kangwon National University, Gangwon-do,
M
24341, Republic of Korea
ED
These authors equally contributed to this work
PT
*
†
Corresponding author: Hyun Suk Jung
CE
Department of Biochemistry, College of Natural Sciences, Kangwon National University, 1 Kangwondaehak-gil, Chuncheon-si, Gangwon-do, 24341, Republic of Korea
AC
Phone: +82 33 250 8513; Fax: +82 33 259 5664 E-mail: [email protected]
†
Corresponding author: Jaekyung Hyun
Electron Microscopy Research Center, Korea Basic Science Institute, 161 Yeongudanji-ro, Ochang-eup, Cheongwon-gu, Cheongju-si, Chungcheongbuk-do 28119, Republic of Korea Phone: +82 42 865 3681; Fax: +82 42 865 3939 E-mail: [email protected] 2
ACCEPTED MANUSCRIPT ABSTRACT Heavy metal compounds are adsorbed onto biological specimen in order to enhance the contrast as well as to preserve the structural features of the specimen against electron beam-induced radiation damage. In particular, in combination with computational image
CR IP T
processing, negative staining is widely used for structural analysis of protein complexes to moderate resolutions. Image analysis of negatively stained biological specimen is known to suffer from limited achievable resolution due to dehydration and large grain size of staining
AN US
molecules although the extent of such effect remains somewhat dubious.
Stain molecules exist as grains under electron beam. However, clear observation of the crystalline nature of the grains and their association with biological specimen has not been thoroughly demonstrated. In this study, we attempted high-resolution TEM (HRTEM) using
M
high voltage electron microscopy and electron crystallography analysis for the detailed
ED
characterization of negatively stained biological specimen, focusing on physical state and chemical composition of the stain molecules. The electron crystallography analysis allowed
PT
for the identification of the crystal constituents of widely used stains, hence revealing the
CE
chemical nature and the morphology of the stain molecules at specimen level. This study
AC
re-evaluated generally accepted notions on negative staining, and may help correctly interpreting the structural analysis of stained biological specimen.
Keywords: High Voltage Electron Microscopy, Electron Crystallography, Negative Staining, Nano-crystal, Transmission Electron Microscopy, Uranium Dioxide
3
ACCEPTED MANUSCRIPT
Abbreviations: Transmission Electron Microscopy (TEM), High-resolution Transmission Electron Microscopy (HRTEM), Cryo-electron Microscopy (Cryo-EM), High-voltage Electron Microscope (HVEM), Selected Area diffraction (SAD), Inorganic Crystal Structure Database (ICSD), Uranyl Acetate (UA), Uranyl Formate (UF), Ammonium Molybdate (AM),
AC
CE
PT
ED
M
AN US
CR IP T
Sodium Phosphotungstate (SPT), Uranium Dioxide (UO2),
4
ACCEPTED MANUSCRIPT
1. INTRODUCTION Negative staining is one of the most widely used sample preparation methods for the visualization of biological specimen using transmission electron microscopy (TEM) [1, 2]. While the image contrast of biological specimen is poor due to weak electron scattering and
CR IP T
the use of minimal electron dose for the reduction of accumulated radiation damage, negative staining greatly improves image contrast due to enhanced electron scattering from heavy metal compounds adsorbed onto the specimen. In addition, the layer of stain molecules that
AN US
surround the biological specimen preserves the molecular features of the object under electron beam that otherwise completely destroys the structure. The technique is quick and easy to apply for most biological sample, and in particular, routinely employed for the structural
M
analysis of protein macromolecules to moderate resolutions as a preliminary experiment before high resolution structure determination using cryo-electron microscopy, or in cases
PT
heterogeneity [3-5].
ED
where cryo-EM is precluded due to insufficient molecular mass and inherent molecular
The structural analysis of protein macromolecules in three dimensions often involves
CE
computational image processing such as single particle analysis [6, 7], electron
AC
crystallography [8, 9] and helical reconstruction [10, 11]. The achievable resolution obtained from the images of negatively stained specimen is typically limited to approximately 10 to 20 Å [3, 4, 12, 13]. The major factor that contributes to the limited resolution is known to be associated with the deformation of fine structural features within the protein molecules that results from the staining and dehydration. By contrast, cryo-electron microscopy (cryo-EM) is devoid of both staining and dehydration, and hence is the method of choice for near-atomic
5
ACCEPTED MANUSCRIPT
resolution structure determination [14-16]. Cryo-negative staining may be used as an alternative to retain the advantage of negative staining while the sample deformation is minimized by vitrification, resulting in 3D reconstruction comparable to that of cryo-electron microscopy data [17]. However, the reason behind limited achievable resolution of typical
CR IP T
negative staining remains somewhat dubious. For instance, an electron diffraction study has demonstrated possibility of data information up to 4 Å from a negatively stained catalase 2D crystal [18], but such resolution has never been realized for structural analysis of proteins by
AN US
2D image analysis or 3D structure reconstruction. Another misleading factor is the definition of “grain” that denotes the fineness of staining molecules that is often used in analogy to the level of details that can be visualized. However, the nature of the grains, or its association
M
with biological molecules that could affect the resolution and fidelity of structural analysis from negatively stained specimen, are still poorly characterized.
ED
In this study we attempted high resolution TEM imaging and electron crystallography of a
PT
biological specimen stained with uranium compounds in order to characterize the behavior of staining molecules on a specimen grid. The results provide visual evidence that the heavy
CE
metal molecules exist as nano-crystals surrounding protein particles under high electron beam,
AC
and the chemical constituents of the crystals were unambiguously determined.
6
ACCEPTED MANUSCRIPT
2. MATERIALS AND METHODS 2.1. Protein purification and sample preparation for electron microscopy Vaccinia virus D13 was used as a test specimen for the analysis of the behavior of stain molecules adsorbed onto the biological sample. Homogeneous preparation of D13 trimer was
CR IP T
purified by immobilized metal affinity chromatography and size exclusion chromatography as described previously [8]. For electron microscopy, 5 μL of D13 solution was applied onto a glow discharged EM grid coated with amorphous carbon film which is typically used for the
AN US
negative staining of protein samples. After 1 minute of sample adsorption, the grid was washed with three droplets of staining solution to replace the buffer with the staining solution. After 1 minute, excess stain solution was blotted away using a piece of filter paper and then
M
the grid was dried in air. The specimen grid for the analysis of nano-crystals that results from staining was prepared in the same way, but without the protein sample loading, due to
ED
possible crystalline contamination that may originate from protein sample buffer. The staining
PT
compounds used in the analysis include 1% (w/v) uranyl acetate, 0.75% (w/v) uranyl formate, 2% (w/v) ammonium molybdate and 2% (w/v) sodium phosphotungstate. The samples were
CE
examined using a transmission electron microscope operating at 120kV equipped with LaB6
AC
beam source (Tecnai G2 Spirit TWIN, FEI Company).
2.2. HRTEM imaging and analysis of crystals A high-voltage electron microscope (HVEM, JEOL Ltd., JEM ARM 1300S) operating at 1250kV was used for the detailed observation of the negatively-stained grids, focusing on the direct visualization and the selected area diffraction of nano-crystal. HRTEM and the electron
7
ACCEPTED MANUSCRIPT
diffraction images of the negatively-stained protein sample and stain nano-crystals were analyzed using the Digital Micrograph software (Gatan Inc.). Fast Fourier transform of the individual nano-crystals in 256 x 256 pixel boxes were used to estimate the distance from the center to each diffraction spots and the lattice angles. The lattice parameters were compared to
CR IP T
simulated electron diffraction and the atomic model that correspond to the stain nano-crystals, uranium dioxide, that were generated using Crystal Maker 2.7 software (CrystalMaker Software Ltd.). Selected area diffraction (SAD) images of the stain crystals were used to
AN US
measure d-spacing of the crystal, and this experimental d-spacing profile was compared to the diffraction profiles of uranium derivatives obtained from FIZ Karlruhe inorganic crystal
M
structure database (ICSD).
2.3. Imaging of dose-dependent stain crystal formation
ED
Dose-dependent crystal formation of stain nano-crystal formation was examined using Titan
PT
Krios TEM (FEI Company) equipped with Falcon III direct electron detector (FEI Company), operating at 300kV and at room temperature. Negatively stained D13 was imaged at nominal
CE
magnification of x250,000 using dose rate of approximately 100 e-/Å2/s, over 300 seconds of
AC
exposure time. The acquisition was performed using movie frame imaging, of which 1204 frames were recorded hence each movie frame was exposed to approximately 25 e-/Å2.
8
ACCEPTED MANUSCRIPT
3. RESULTS AND DISCUSSIONS 3.1. High Resolution Imaging of Negatively Stained Biological Specimen The existence of stain crystal has been demonstrated elegantly at the very early stage of structural investigation of protein complexes by Unwin [19]. The study demonstrated that the
CR IP T
formation of stain crystal is electron dose-dependent, as evidenced by a series of electron diffraction images showing that the crystal formation is promoted by accumulating electron dose. We first attempted to visually inspect the physical state and the behavior of staining
AN US
molecules using a typical negative staining procedure, a test protein sample Vaccinia virus D13 was stained using 1% uranyl acetate, followed by air-drying. The sample was first examined using a TEM operating at 120kV using various magnifications and defocus levels.
M
However, no crystalline feature was identifiable, most likely to be due to limited resolving power. We then performed HRTEM imaging using a HVEM, from which stain nano-crystals
ED
were clearly visible. The crystals were heavily associated around protein molecules with no
PT
apparent pattern in their distribution (Figure 1, A and B). At approximately 1 μm under-focus, the features of crystalline lattice were no longer visible (Figure 1, C and D). The images of
CE
nano-crystals and respective fast Fourier transforms indicate that the crystalline features are
AC
easy to be missed under normal imaging conditions used in the analysis of protein macromolecules (i.e. low dose, low magnification, and relatively high under-focus), especially when HRTEM is not intentionally attempted. It was also found that the crystalline patches are most noticeable on the region where the stain accumulation was heavy, usually surrounding the edge of protein molecules or grooves within the structure, suggesting that the stain molecules bound to the protein molecules vary in the extent of crystallization.
9
ACCEPTED MANUSCRIPT
In this study, for experimental simplicity and image acquisition using HRTEM, applied electron dose far exceeded the typical accumulated dose recommended for structural analysis of protein particles using cryo-EM (20~30 electrons per Å2). Accordingly, the crystals seen in our study is most likely formed during the strong electron beam exposure onto the specimen
CR IP T
as demonstrated before [17]. To elaborate approximate electron dose required for the formation of pronounced nano-crystals, a series of beam exposure onto the specimen was performed using movie frame image acquisition where each frame was exposed to
AN US
approximately 25 e-/Å2. Gradual growth of the crystal was discernable from the images when the accumulated electron dose reached up to tens of thousands e-/Å2 (Figure S1). The observation suggests that the effect of nano-crystal formation on the structural fidelity of the
M
negatively stained particles could be marginal if care is taken not to over-expose the sample
ED
with the beam or, most certainly, if low dose imaging is employed.
PT
3.2. High Resolution Imaging of Staining Molecules on EM Grid We prepared EM grid without protein sample in order to avoid possible crystallization of the
CE
components in buffer solution that may interfere with clear examination of stain molecules.
AC
Some of the most commonly used stain solutions were subjected to HRTEM. These included uranyl acetate (UA, UO2(CH3COO)2), uranyl formate (UF, UO2(CHO2)2), ammonium molybdate (AM, (NH4)6Mo7O24) and sodium phosphotungstate (SPT, Na3O40PW12), dissolved in distilled water at concentrations used in routine protein sample preparation [2]. In all cases, small crystals were found all over the EM grid with varying sizes and crystalline lattices, from which the HRTEM images were collected (Figure 2). EM grids stained with UA and UF
10
ACCEPTED MANUSCRIPT
showed smallest crystals, typically smaller than 20nm in the longest dimension, in random orientations. The grids stained with other compounds also contained nano-crystals but micrometer-scale crystals with continuous lattice were often found. Considering typical achievable resolution of protein 3D reconstruction from negatively stained specimen is
CR IP T
approximately 10 to 20 Å, the size range of the crystals seen in the sample does may not be directly relevant to the resolution. Rather, crystallization of otherwise amorphous stain molecules will involve significant rearrangement of associated molecular structure of the
AN US
protein which, in consequence, causes additional deformation of the object. This conclusion was also drawn from initial evaluation of the effect of stain crystallization on the
M
reconstruction of stacked disk aggregates of Tobacco Mosaic virus protein [17].
3.3. Electron Crystallographic Analysis of Stain Nano-Crystals
ED
Crystals found in the EM grids stained with UA and UF were subjected to a further TEM
PT
analysis in order to analyze possible differences in their behaviors. The two compounds were chosen because they are derivatives from the same elements, but UF is known for finer
CE
granularity and better permeation that suits for high resolution analysis [20, 21]. In order to
AC
visually evaluate their physical state and chemical composition on the specimen, electron crystallography analysis was carried out. Small regions with best crystalline order, as judged by the sharpness and extension of diffraction spots in fast Fourier transform, from the field of a randomly chosen area within a grid were selected for electron crystallography analysis (Figure 3A and C). In the Fourier transforms, distance from the center of the pattern to each diffraction spot and the angle
11
ACCEPTED MANUSCRIPT
between lattice lines allowed us to obtain d-spacing profile and determine zone axis (Figure 3B and 3D), from which the identical diffraction profile was searched in the inorganic crystal structure database (ICSD). The diffraction patterns and d-spacing profiles of nano-crystals found on the grids stained with UA and UF were nearly identical, and consistent with that of
CR IP T
uranium dioxide (UO2) (figure 3E and 3F). In addition, the experimental diffraction profile agreed well with electron diffraction analysis of uranyl dioxide thin film reported elsewhere [22]. Because only a small number of crystalline patches were subjected to crystallographic
AN US
examination, the possibility of other types of crystals in the overall area still remained. Therefore, selected area diffraction (SAD) from larger field of view was recorded in order to make sure that the powder diffraction profile coincides with a single species (i.e. UO2)
M
(Figure 4). The resulting d-spacing profiles of UA and UF obtained from distinctive rings of intensity were identical with subtle differences, and it was clear that the majority of crystal
ED
constituents are indeed UO2.
PT
Taken together with physical association of stain crystals with protein molecules observed by HRTEM, the crystallographic analysis suggests that the degree of details identifiable from the
CE
use of both uranium compounds not only depends on stain granularity that is referred by the
AC
manufacturers and publications, but also by UO2 nano-crystals and consequential structure deformation of the specimen.
12
ACCEPTED MANUSCRIPT
4. CONCLUSIONS In this study, we addressed a fundamental question in the analysis of negatively stained biological sample; what are we really seeing in the image? We used HRTEM and electron crystallography, techniques that are not routinely employed for biological TEM, in order to
CR IP T
clarify the nature of stain molecules under electron beam. Our result demonstrate that “grain”, of which the size is commonly referred as an indicative of the degree of structural fineness retained in the image, is no longer valid when high enough electron dose promotes the
AN US
formation nano-crystals. In particular, we unambiguously identified that UO2 is the constituent of stain crystals in case of both UA and UF. Examination of chemical identity of stain molecules in low electron dose, that is before apparent formation of nano-crystals, as
M
well as direct comparison of resulting 3D reconstructions of negatively stained protein complexes at both low and high electron dose would provide further aid clarifying the relation
ED
between achievable resolution and negative staining. In the era of near-atomic cryo-EM
PT
structure determination, a careful revision of routine negative staining is still a valuable exercise that could improve the fidelity, and even to maximize resolution, of initial 3D
AC
CE
structure of biological macromolecules at moderate details.
13
ACCEPTED MANUSCRIPT
Acknowledgement This work was supported by Korea Basic Science Institute grant T38210 to J. Hyun, and in part by 2017 Research Grant from Kangwon National University (No. 520170496), Next-Generation BioGreen Program(SSAC, PJ013273042018) and Basic Science Research
CR IP T
Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT and Future Planning (NRF-2015R1C1A1A01053611, 2018R1D1A1B07045580)
AN US
to H. S Jung.
Conflict of interest
AC
CE
PT
ED
M
The authors declare no conflict of interest arising from this work.
14
ACCEPTED MANUSCRIPT
References 1.
Rames, M., Y. Yu, and G. Ren, Optimized negative staining: a high-throughput protocol for examining small and asymmetric protein structure by electron microscopy. J Vis Exp, 2014(90): p. e51087. Harris, J.R., Negative Staining and Cryoelectron Microscopy 1997, New York: Taylor
CR IP T
2.
& Francis. 3.
Ohi, M., et al., Negative Staining and Image Classification - Powerful Tools in
4.
AN US
Modern Electron Microscopy. Biol Proced Online, 2004. 6: p. 23-34.
Bremer, A., et al., Has negative staining still a place in biomacromolecular electron microscopy? Ultramicroscopy, 1992. 46(1-4): p. 85-111.
Zhang, L., et al., Optimized negative-staining electron microscopy for lipoprotein
M
5.
studies. Biochim Biophys Acta, 2013. 1830(1): p. 2150-9. Zhou, Z.H., Towards atomic resolution structural determination by single-particle
ED
6.
7.
PT
cryo-electron microscopy. Curr Opin Struct Biol, 2008. 18(2): p. 218-28. Frank, J., Single-particle reconstruction of biological macromolecules in electron
Hyun, J.K., et al., Membrane remodeling by the double-barrel scaffolding protein of
AC
8.
CE
microscopy--30 years. Q Rev Biophys, 2009. 42(3): p. 139-58.
poxvirus. PLoS Pathog, 2011. 7(9): p. e1002239.
9.
Wisedchaisri, G., S.L. Reichow, and T. Gonen, Advances in structural and functional analysis of membrane proteins by electron crystallography. Structure, 2011. 19(10): p. 1381-93.
10.
Egelman, E.H., Reconstruction of helical filaments and tubes. Methods Enzymol,
15
ACCEPTED MANUSCRIPT
2010. 482: p. 167-83. 11.
Ge, P. and Z.H. Zhou, Hydrogen-bonding networks and RNA bases revealed by cryo electron microscopy suggest a triggering mechanism for calcium switches. Proc Natl Acad Sci U S A, 2011. 108(23): p. 9637-42. Kiselev, N.A., M.B. Sherman, and V.L. Tsuprun, Negative staining of proteins. Electron Microsc Rev, 1990. 3(1): p. 43-72.
13.
Durand, E., et al., Biogenesis and structure of a type VI secretion membrane core
AN US
complex. Nature, 2015. 523(7562): p. 555-60. 14.
Adrian, M., et al., Cryo-electron microscopy of viruses. Nature, 1984. 308(5954): p. 32-6.
Frank, J., Three-dimensional electron microscopy of macromolecular assemblies:
M
15.
CR IP T
12.
visualization of biological molecules in their native state. 2nd ed. Three-dimensional
ED
electron microscopy of macromolecular assemblies: visualization of biological
16.
Kuhlbrandt, W., Biochemistry. The resolution revolution. Science, 2014. 343(6178): p.
CE
1443-4.
De Carlo, S. and J.R. Harris, Negative staining and cryo-negative staining of
AC
17.
PT
molecules in their native state. 2006, New York: Oxford University Press.
macromolecules and viruses for TEM. Micron, 2011. 42(2): p. 117-31.
18.
Massover, W.H., P.F. Lai, and P. Marsh, Negative staining permits 4.0 A resolution with low-dose electron diffraction of catalase crystals. Ultramicroscopy, 2001. 90(1): p. 7-12.
19.
Unwin, P.N., Electron microscopy of the stacked disk aggregate of tobacco mosaic
16
ACCEPTED MANUSCRIPT
virus protein. II. The influence of electron irradiation of the stain distribution. J Mol Biol, 1974. 87(4): p. 657-70. 20.
Leberman, R., Use of uranyl formate as a negative stain. J Mol Biol, 1965. 13(2): p. 606. Knight, D.P., Negative staining of rat tail tendon collagen fibrils with uranyl formate.
CR IP T
21.
Tissue Cell, 1975. 7(4): p. 651-4. 22.
Lin, J., et al., Structure and properties of uranium oxide thin films deposited by pulsed
AC
CE
PT
ED
M
AN US
cd magnetron sputtering. Applied Surface Science, 2014. 301: p. 475-480.
17
ACCEPTED MANUSCRIPT
Table 1. Comparison between the experimental d-spacings of stain nano-crystals and the
Experimental d-spacing (UA)
Experimental d-spacing (UF)
(111)
0.315 nm
0.312 nm
0.320 nm
(002)
0.193 nm
0.192 nm
(113)
0.165 nm
0.164 nm
(024)
0.122 nm
0.122 nm
(224)
0.111 nm
AC
CE
PT
ED
M
0.112 nm
18
CR IP T
Plane
Reference d-spacing (UO2)
AN US
reference (UO2)
0.195 nm 0.164 nm 0.123 nm 0.112 nm
ACCEPTED MANUSCRIPT
Figures and figure legends
AN US
CR IP T
Figure 1.
M
Fig. 1. HRTEM images and the fast Fourier transforms of negatively-stained protein sample.
ED
(A) HRTEM image of negatively-stained D13, recorded at just below at-focus and (B) a higher magnification image. (C) Micrographs of same area, recorded at under-focus typically
PT
used for structural analysis of protein complexes (~1μm under-focus), and (D) a higher
CE
magnification image. Nano-scale crystals are found around the protein particles at close-to-focus imaging condition, whereas such features diminished at high defocus level.
AC
Fourier transforms are shown in the insets. Scale bars are 20 nm in (A) and (C), 5 nm in (B) and (D), 5 1/nm in the insets, respectively.
19
ACCEPTED MANUSCRIPT
AN US
CR IP T
Figure 2.
Fig. 2. HRTEM image and the fast Fourier transform of stain molecules on EM grid. HRTEM micrographs of EM specimen grids treated with (A) UA, (B) UF, (C) AM and (D) SPT in the
M
absence of protein sample. Crystals with varying sizes and lattice dimensions are clearly seen
ED
in the micrographs and respective FFTs. The Fourier transforms are shown in the insets. Scale
AC
CE
PT
bars are 5 nm in the micrographs, 10 1/nm in the insets, respectively.
20
ACCEPTED MANUSCRIPT
M
AN US
CR IP T
Figure 3.
ED
Fig. 3. HRTEM and electron crystallographic analysis of UA and UF nano-crystals. (A, C)
PT
HRTEM micrographs of UA and UF nano-crystals, and (B, D) respective Fourier transforms. (E) A simulated model of UO2 with a molecular unit cell demarked by a box, and (F) the
CE
corresponding Fourier transform. All zone axes of the selected nano-crystals in this figure are
AC
[1-10]. The HRTEM and crystallographic analysis indicates that UO2 is the main constituent of the stain nano-crystals. Unit cell dimensions are indicated in (B), (D) and (F). Scale bars are 2 nm in (A) and (C), 5 1/nm in (B) and (D), respectively.
21
ACCEPTED MANUSCRIPT
CR IP T
Figure 4.
Fig. 4. Selected area diffraction (SAD) analysis of stain nano-crsytals. (A) SAED images of
AN US
UA nano-crystals and (B) UF nano-crystals. The diffraction analysis unambiguously demonstrates that the majority of stain nano-crystals are composed of UO2. Scale bars are 5
AC
CE
PT
ED
M
1/nm.
22
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN US
CR IP T
Graphical abstract
23