The world leader in serving science
2020/05/28 (Vybrané kapitoly z elektronové mikroskopie, MUNI.)
Cryo electron microscopy (for LS)
2
Content
• Introduction: bio-samples, scale, methods, etc.
• Sample preparation, navigation
• Fixation / vitrification,
• Plunge freezing (PF),
• High pressure freezing (HPF),
• Correlative microscopy (CLEM),
• Lamella preparation with ion beam (for SDB Tomo Workflow).
• Methods
• Cryo-Electron Tomography (cryo-ET) + STA,
• Single Particle Analysis (SPA),
• MicroED.
3
Bio-Samples: Wide Scale Across the Space
Adapted from: Alberts, Základy buněčné biologie, 2004.
4
Different Methods See Different Details
Adapted from: N/A; www.npt.co.uk/biotechnology/research/super-resolution-imaging.
5
From Image to Function…To reveal a structure and understand its function.
Adapted from Wiki, https://www.scientificpsychic.com/health/virus.html.
example of antiviral drug
TEM image
6
From Image to Function…Drug Design
• PROTEINS: “tiny molecular
machines. Perform most of the tasks
needed to keep cells alive. Drugs
can be used to turn proteins on or
off, to affect their action.”
• DRUGS: “small molecules that bind
to one specific protein and modify its
action. Antibiotics or anticancer
drugs are used to completely
disable a critical molecular machine.
These drugs can kill a bacterial or
cancer cell. Other molecules, such
as aspirin, gently block less-critical
proteins for a few hours.”
Adapted from https://cdn.rcsb.org/pdb101/learn/resources/flyers/how-do-drugs-work-flyer.pdf.
7
Selected Life Science Workflows to Study Objects of Interest
Remark. Micro-ED: structure of protein crystals. Room Temperature Tomography: 3D imaging on RT TEM.
Structure of
folded Protein
Function of Protein
in Cells
CryoEM: Single
Particle Analysis
CryoEM: Tomography
Cellular Organisation in
Tissue
Large Volume Analysis
Structure determination of
proteins and protein complexes
in their native state at “nearatomic”
resolution.
High resolution
reconstruction of large
molecular complexes in their
functional environment in
cells and tissues.
Modelling of cell-cell
interactions and organ
function based on volume
imaging at nm resolution.
SPA Cryo-ET LVA
8
Advantages of Cryo-EM
• Observation of biological systems in their native hydrated state.
• No artifacts that occur in chemical fixation and negative staining (direct
visualization of biological macromolecules instead of their contours in
background of negative stain).
• No need for the macromolecule crystallization.
• Suitable for larger protein complexes and pleomorphic structures.
• Possible 3D structure determination up to “near-atomic” resolution.
• Time-freezing of dynamic processes allows determination of molecules in
multiple different functional conformations. Cryo-ET enables to study these
conformations in the context of their cell environment.
• Cryo-ET bridges the gap between imaging of isolated macromolecules at
near-atomic resolution and large volume analysis at the cell-cell/tissue level;
“opens the window into the cell”.
20S proteasome
Remark. Images on the right: 20S Proteasome particles; near-atomic resolution model by Cryo-EM.
9
Structural Symphony: Role of Cryo-Electron Tomography / SPA within Imaging Technologies
FEI Company
AtomsProteins
100 µm 10 µm 10 nm 1 nm 1ÅStructure size 100 nm1 µm1 mm
VirusesCells
Single Particle cryo-TEM
Cryo-electron tomography (+ STA)
NMR
Light Microscopy
X-ray crystallography (XRD)/MicroED
Cryo Light Microscopy
Super Res
RESOLUTION
CONTEXT
Conventional SEM / TEM (+LVA)
Brno Philharmonic Orchestra image: the-wagnerian.com.
11
Deposited Structures in EMDB Solved by Cryo-EM
https://www.ebi.ac.uk/pdbe/emdb/index.html; Shashi Bushan: EMBO course: SPA presentation.
12
Sample Preparation is the Key…Fixation
FEI Company
Brno Philharmonic Orchestra image: the-wagnerian.com.
• Bio-samples are full of water (body water content around 60%, brain 73%, cell 70%) => implications
for the sample preparation and observation in EM. (Most/less abundant elements: H, C, N, O/Na,
Mg, P, S.)
• Fixation = to stop the biological activity and to preserve the tissue structure for subsequent
treatments.
• “The objective is to process tissues and cells
without significant change in size, shape, positional
relationship of the cellular components and
to preserve as much of the biological activity and
chemical nature of cellular components…”*
*Dykstra M.J. et al., Biological EM microscopy (2003).
13
I love
Cryo!
Sample Preparation is the Key…Fixation
Chemical Fixation
Heat Preservation
Cryo Preservation
Freeze Dried
Normal Strawberry
14
Bio-Sample Preparation
SDB Cryo-ET
SPA
Adapted image by Andreas Kaech, University of Zürich. See as well: Handbook of Cryo-preparation methods for EM, edited by Cavalier A.,
Spehner D., Humbel B.M. (2009).
15
Preparation of Biological Samples for Cryo Electron Microscopy
Cryo-FIB
milling
Cryo-ultramicrotomy
CryoEM: SPA Cryo-electron tomography of thin specimen
3D RECONSTRUCTION & INTERPRETATION
IMAGING
Plunge-freezing in liquid ethane High-pressure freezing
VITRIFICATION
Purified particles Bacteria & yeasts / Eukaryotic cells
BIOLOGICAL SAMPLE
16
• Why, advantages
• The best method of preservation. Rapid freezing in milliseconds = near perfect preservation
(minimal chemical and physical changes if done well).
• Offers a SnapShot at a particular time, very important when studying function.
• Sample vitrification
• Cool the specimen so rapidly that there is not time for ICE (crystalline water) to form! → ICE is
what does the damage as it rips structures apart.
• Increase cooling speed by the reducing size of the specimen.
• Methods
• High pressure freezing (HPF, commonly up to ~200 µm), plunge freezing (PF), slam (metalmirror)
freezing, double jet propane freezing, spray freezing (all up to units or tens of µm).
Cryofixation (cryo/cryos means "icy cold" (from crystallos))
Adapted from Cheng et al., Current Micros. Contrib. to Advan. in Scien.and Tech. (2012); Szczesny et al.: doi.org/10.3109/02713689609017621.
17
Plunge Freezing
• sample thickness: up to ~5-10 µm
(cells, proteins, virus particles),
• 3 mm TEM grids,
• cryogen: liquid ethane/propane, atm pressure.
Plunge Freezing vs High Pressure Freezing (vitrification of native bio-specimens)
High Pressure Freezing
• sample thickness: commonly ~50-200 µm
(small organisms, tissue, cell cultures),
• 3 – 6 mm carriers,
• cryogen: “open system”: pressurized LN2.
HPF: Images adapted from leica.com; Andres Kaech: High pressure freezing / presentation.
18
Plunge Freezing
• Needs high rate of cooling (~105 K/s) to
temperatures below -140
• Fast plunge of blotted specimens (m/s).
• Liquid nitrogen cannot be used
because of low heat capacity.
• Liquid ethane or propane have good
properties and Tm close to Tb of LN2.
Tweezers
LN2 bath for cooling of secondary cryogen
°C.
19
Sample and TEM Grid Support
• Support to fragile TEM grids and possibility of robotic
handling and automated TEM loading.
20
Plunge Freezing Process
21
Time-Resolved Cryo-EM (TR Cryo-EM)
Adapted from: *Frank J., J. Struct. Biol. 200(3) (2017); Shaikh T.R. et al., PNAS 111 (27) (2014).
• *“Combination of the structural study
with kinetics, by capturing kinetic
intermediate states in a biological
reaction.”
• “Movies of a biological complex
functioning in real time.”
• “Fast reactions require a means of
mixing, reacting and depositing the
product on the grid in a fast,
controlled way.”
• Different approaches:
mixing-spraying, spraying-freezing,
flash-photolysis, etc.
Experimental setup of the mixingspraying
method and the design of the
mixing-spraying chip.
Cryo-EM maps of the
70S ribosomes
formed within 9.4 ms.
22
High Pressure Freezing
• Simultaneous high rate of cooling
(~104 K/s) to temperatures below
-140 at high pressure
(2100 bars).
• Pressurized LN2 flows over the
sample inside the carrier (“open
system”).
• Rapid cooling at high pressure allows
for the transition to amorphous
(glassy-like) state even for “thicker”
samples.
Image of HPF (*) adapted from leica.com.
*
°C
23
HPF Freezer
Depicted HPF Freezer: High Pressure Freezer Leica EM ICE.
24
The pressure-temperature dependent phase diagram of water
Images Fig.1, Fig. 6 adapted from: Dahl R. et al., J. of El. Micr. Tech. 13 (1989) + Bäuerlein F.J.B., PhD. thesis, 2018
(TUM/MPI); Kim Ch.U. et al., Acta Cryst. D61 (2005).
25
HPF: Cooling Rates
• Water has a very poor heat conductivity.
thicker specimen=lower cooling rate ice crystal formation=Segregation
26
HPF: FS and observation of resin embedded samples at RT
In 20% Dextran before freezing. After Freeze Substitution (FS) in Epon resin.
2 um
Drosophilla embryo
Images at the bottom adapted from Titze B. et al., Biol. Cell 108 (2016); Miranda K. et al., Molecular Rep. & Dev. 82 (2015).
Schematic representation of FS process.
27
Doing Cryo Electron Tomography on Vitrified Cells Requires a Workflow
ELECTRON
TOMOGRAPHY
FOCUSED ION BEAM
MILLING
VITRIFICATION
SUBTOMOGRAM
AVERAGING
CORRELATIVE
MICROSCOPY
½ DAY 1 TO 2 DAYS UP TO MULTIPLE DAYS DAYS- MONTHS
VISUALIZATION & ANALYSIS
28
Many Cells are Too Thick for Electron Cryo-Tomography
PHAGOSOME
CELL MEMBRANE
1 µm
RANGE
TILTAXIS
ACCESSIBLENON-ACCESSIBLE
29
A need for correlative microscopy / CLEM
Correlative microscopy: use of two or more microscopy techniques – preferably with different spatial
and/or temporal resolutions – to characterize the same region of interest in a sample.
Light microscope
FIB/SEM
30
Cryo-Fluorescence Microscopy: Fluorescence Mediated Targeting
Transfer device
Catridge
The eukaryotic cytoskeleton.
Actin filaments are shown in
red,and microtubules compo
sed of beta tubulin are in
green.
Blue: nucleus stained with
DAPI; Green: Tubulin
(microtubules); Red: F-Actin
stained with Texas Red X-
Phalloidin.
Cryo-Light Microscope (cryo-LM, Leica)
Images adapted from: thermofisher.com; http://rsb.info.nih.gov/ij/images/; leica.com.
31
Cryo-CLEM: Relocation of Region of Interest in the Cryo-EM
FIB/SEM
TEM TEM
+ LM data
FIB/SEM
LM
Right image adapted from: Kuba J. et al., Journal of Microscopy (in review) (2020).
32
Cryo-Tomography: high resolution imaging of structure and context for cell biology
Data Courtesy Max Planck Institute of Biochemistry
| Mahamid et al., Science 2016
Molecular Identification
Dynamics
Large Volumes
Only see what is tagged
No structure
No context
FLUORESCENCE IMAGING
33
AQUILOS
CRYO-FIB
GUIDED USER
SOFTWARE
LN2
DEWAR
SAMPLE LOADING
STATION
Cryo-FIB: Cryo-Sample Preparation…opening the window into the cell
34
Operating Principle and Function of Cryo-FIB/SEM in SDB Cryo-Tomo Workflow
Left: adapted from Kuba J. et al., J. of Micr. 2020 (in review). Right: Shaffer M. et al., J. Struct. Biol., Vol. 197(2) (2017).
35
Cryo-FIB: Cryo-Sample Preparation Flow
36
Cryo-FIB: Cryo Infrastructure
AQUILOS
CRYO-FIB
37
Prep Station Controller
(pumping, venting, heating)
A
B
Aquilos: Sample Loading, Transfer, Preparation Station
Cryo-FIB
Autogrid shuttle
(closed shield)
Heating Plate for Tools
Stand for
Transfer Rod
Transfer Pot
(+ docked
Transfer Rod)
38
Aquilos: Load Lock, Cryo Transfer Rod, Stage
LOAD LOCK
(continously
pumped)
CRYO
TRANSFER ROD
Cryo Shuttle
Shutter in
open
position
Two Slots for Autogrids
39
Contamination: specles, large particles, homogeneous layer; devitrification
“cell”
“protective
layer”
contamination
cell
40
Lamella Preparation
1
2
3
SEM
FIB
41
Lamella Preparation Automation: Cryo AutoTEM live
42
Lamella Milling Issues
Electrons Ions
44
Protective Coating: Cold Deposition (using GIS = gas injection system)
GIS scheme on the right adapted from Rigort A. et al., Archives of Biochem. and Biophys. 581 (2015).
Remark. Hayles M. F. et al., A technique for improved FIB milling of cryo-prepared LS specimens, JOM 226 (Pt 3) (2007).
Do not forget to close
the GIS valve…
Gas
molecules
Deposited
material
Volatile
products
FIB
GIS
RT
Cryo
45
without Pt GIS coatingwith Pt GIS coating
Pt layer
Protective Coating: Preventing Beam Erosion
Milling direction
Adapted from Shaffer M. et al., J. Struct. Biol., Vol. 197(2) (2017); Rigort A. et al., Archives of Biochem. and Biophys. 581
(2015).
46
Cryo-FIB Lamella: Milling Strategy
SEM (+ contrast
difference)
TEM
Fig. a,b adapted from Rigort A. et al., Archives of Biochem. and Biophys. 581 (2015). Bottom image: Schaffer et al., Journal of Structural Biology
2016. Image on the right: Electron tomography (2010), edited by J. Frank.
Lamella thickness and
uniformity, tilting during
milling.
Distribution of scattered electrons
for vitreous ice.
47
uncoated Pt sputter
coated
Conductive Coating: SEM, TEM
Adapted from (image on the left) Schaffer et al., Journal of Structural Biology (2016), doi:10.1016/j.jsb.2016.07.010.
just polished sputtered
cells on the TEM grid
lamella
48
Conductive Coating: Example of Retractable In-Chamber Magnetron Sputter
Principle of DC Magnetron Sputtering.
Image on the right adapted from: http://www.semicore.com/news/94-what-is-dc-sputtering.
49
Lamella Preparation Approaches
Adapted from Kuba J. et al., Journal of Microscopy 2020 (in review).
50
CLO Flow
51
CLO live
52
SDB Cryo-Tomo Workflow: SEM Cryo-Imaging
SDB cryo-tomo workflow for structure determination of cells.
Cryo Auto Slice and View or end-pointing
during lamella preparation.
cryo-lamella (mouse brain)
3D display, Chlamydomonas
vitrified cell*
*
53
Cryo-ASV: Site Preparation
54
SEM Cryo Imaging: Visualization of subcellular features.
Chlamydomonas
Cell (schematic)
GOLGI
CHLOROPLAST
(THYLAKOIDS)
PYRENOID
(RUBISCO)
STARCH
NUCLEUS
ER
GOLGI
NUCLEAR
PORES
MITOCHONDRIA
Aquilos Cryo-FIB: Cryo ASV of Chlamydomonas
55
Aquilos: Live Demo
56
Data collection: meet the big guy.
Image on the left: https://blogs.ubc.ca/corinnedy/history-2/. TEM scheme: Bäuerlein F.J.B., PhD. thesis, 2018 (TUM/MPI).
57
Cryo electron tomography (Cryo-ET) principle
compustage
e-beam
e-dose
busy microscropist
detector
Adapted from: left - Electron tomography (2010), edited by J. Frank; right: Weber M.S. et al., Cells 8 (57) (2019).
holder
sample on TEM grid
2D projections
3D reconstruction
(tomogram)
tilt series
58
Cryo-ET principle
59
Data collection geometry: ET vs SPA
Adapted from Electron tomography (2010), edited by J. Frank.
60
Cryo-ET (with STA) workflow*
• Data collection (automated by dedicated SW): collection
geometry (single-, dual-tilt axis etc. strategy), tilt scheme;
stage drift: ROI tracking, autofocus).
• Image pre-processing: frame alignment (DED); defocus
determination (gradient: tilt, sample thickness) > CTF
correction (3D).
• Tilt series alignment: correction for shifts, rotation,
magnification changes; fiducial or feature/patch tracking
based.
• Tomogram Reconstruction (dedicated SW packages:
Eman, Spider, Scipion, …): different algorithms (WBP,
ART, SIRT, DFM…) to process tilt series into tomogram.
• Particle picking for STA (template matching and/or manual;
starting reference vs biased structure determination).
• STA = sub-tomogram averaging (3D particle averaging
from reconstructed volume).
• Post-processing and visualization.
*Software tools for Molecular microscopy: https://en.wikibooks.org/wiki/Software_Tools_For_Molecular_Microscopy.
Scheme adapted from Briggs J.A.G. et al.: Methods in Enzymology 579 (2016).
61
Tilt scheme, dose distribution
sample holder, single tilt
increase of the sample effective thickness
Adapted from: left - Briggs J.A.G. et al.: Methods in Enzymology 579 (2016); right top: Galaz-Montoya J.G. et al., Biophys. Rep. 3 (1-3) 2017.
62
Projection theorem and missing wedge, WBP
Adapted from Electron tomography (2010), edited by J. Frank.
Projection theorem. “The 2D Fourier transform of a projection
of the object is identical to a central section of the object’s
3D Fourier transform.”
N ~ D/d (Crowther; cylindrical object, fully mapped); D object dia, d resolution.
real space Fourier space
object
N projections
WBP reconstruction
Effects of the missing
wedge on a 2D image.
63
Missing wedge
Adapted from: left – Koster A.J. et al., J. of Struct. Bio. 120 (1997); right – McEwen B.F. et al., Methods in Cell Bio. 49 (ch. 6).
64
How does tomography data look like?
TOMOGRAM VISUALIZATION
Data Courtesy Max Planck Institute of Biochemistry | Bykov et al., eLife 2017.
65
Cryo-ET at the Cytoplasm/Nucleoplasm Border
500 nm 200 nm
TEM projection Slice from 3D reconstruction 3D reconstruction (300 nm)
67
Cryo-ET: Visualizing the Molecular Sociology at the Hela Cell Nuclear Periphery
Mahamid et al., Science, 351(6276) 2016; cell scheme on the right: Alberts, Základy buněčné biologie, 2004.
68
Subtomogram averaging (STA/STP): structural biology in-situ
Adapted from Briggs Briggs J.A.G. et al.: Methods in Enzymology 579 (2016); Mahamid et al., Science. (2016).
Particle picking (template matching)
69
STA allows to map different functional states of the same protein
Adapted from Albert et al., PNAS 114 (2017).
Proteasome localization and activity (functional state and interaction partners)
by visualizing its macromolecular structure within the native cellular
environment (Chlamydomonas reinhardtii). Scale bar: 200 nm. Tomo: lam. thick.
< 200 nm, tilt incr. 2°, ±60°, px 3.4 Å, defocus form -4 to -5.5 µm, total dose < 100 e/Å.
Subtomogram averages of 26S assembly states
(res. ~ 21 Å).
(Remark. Particle picking: SPA template from EMDB: low-pass
filtered 20S core particle attached to one 19S regulatory particle.)
70
Volta Phase Plate (VPP)
Adapted from: Danev R. et al., PNAS, 111 (2014); Bäuerlein F.J.B., PhD. thesis, 2018 (TUM/MPI); Fukuda Y. et al., Journal of
Structural Biology 190 (2015).
CTEM (-8 um defocus), VPP (-1um defocus)
Cryo-ET
SPA
71
Direct Detectors & Dose Fractionation & Drift Correction
Image on the left: http://www.directelectron.com/products/de-series; DQE graph: Bäuerlein F.J.B., PhD. thesis, 2018 (TUM/MPI).
72
Single Particle Analysis (SPA) Workflow
Data acquisition
(2D images of particles)
Reconstruction
(3D model)
others…
Biochemistry
(protein purification)
Sample preparation
(vitrification)
automated data collection: e.g. EPU sw
73
SPA Workflow
Sample preparation (proteins, viral particles, …)
Data acquisition
(no tilt series)
Data processing
Adapted from: image on the right - http://cns.fas.harvard.edu/CryoEM.
74
SPA Principle - video
Dr. Wen-Ti Liu: http://www.novalix-pharma.com/.
75
SPA: Data Processing Chain
2D images /
projections
class
averages
3D model reprojection
refined class
average
• frames: motion correction,
• particle picking,
• CTF correction,
• alignments and classification
(particle shifts, rotation; classes)
• relative orientation of the particles /
classes in the space (euler angles),
• reconstruction algorithm (2D slices),
e.g.: FT > filling of 3D FT space > FT-1
• reproject in all directions (angular step),
• realign, reclassify (filters),
• filling of 3D FT space > FT-1
initial
model
• known structure,
• tomography,
• de novo / common lines
• random conical tilt, …
Adapted from: Jensen G.: Getting started in cryo-EM; Orlova E.V. et al., Chem. Rev. 111 (2011); Carroni M. et al., Methods 95 (2016).
76
357,022 Particles
2D classification
2392 Micrographs
Particle picking
3740 Micrographs
Micrograph selection
Example: Apoferritin sample reconstructed at 1.6 Å.
1.62 Å Reconstruction
1.9 Å Reconstruction
Ctf Refinement + Particle Polishing
265,955 Particles
3D refinement + Postprocessing
Final 3D Refinement +
Postprocessing
Camera Falcon 3 EC
Pixel size(Å) 0.52
Dose rate
(e/pix/sec)
0.5
Total dose (e/Å2) 52
Dose fractions 100
Exposure time
(sec)
29
Number of
images
3740
Defocus values
-1.2, -1, -0.8,
-0.6,-0.4
• Data Acquisition on Krios
TEM using EPU sw &
Falcon 3 EC camera.
• Sample: apoferritin
~3 mg/ml.
77
Example: Apoferritin reconstruction at 1.6 Å resolution.
3D reconstruction of Apoferritin Atomic Structure Docking
Ferritin
78
From Sample Screening to High Resolution Data Acquisition
TalosL120CGlaciosTitanKrios
Image on the left adopted from Passmore & Russo, Methods Enzymol, 579 (2016).
79
Example: Cryo-EM structure of the 2019-nCoV spike (S) glycoprotein.
Adapted from: left - https://www.scientificanimations.com; right - Wrapp et al., Science 367 (2020) and Sheeren M.A. et al., Journal of
Advanced Research 24 (2020). Remark. Build you own virus: CellPAINT-2D.
81
MicroED: Crystal Protein Lamella Workflow
PRODUCTION +
PURIFICATION
CRYSTALLIZATION DATA PROCESSING
REFINEMENT &
VISUALIZATION
1-2 DAYSFEW DAYS ½ DAY FEW DAYSMINUTES MINUTES TO HOURS
FREEZING CRYO-FIB MILLING
FEW HOURS
X-ray crystallography
(XRC)
>50 µm
MicroED (MED)
<0.5 µm
Too small for XRC
Too large for MED
0.5-50 µm
DATA COLLECTION
82
Principle of MicroED
3D structure
3D structureFT space
83
Micro-crystals of Lysozyme
SDB: Electrons
1
0
µ
m
10 µm
50nm
SDB: Ions
10 µm
TEM
Electron Density
Map
84
MicroED for Pharmaceutical Molecule Structure Workflow (RT)
• Drug (paracetamol, ibuprofen,…) is turned
into to smooth powder.
• Powder is applied on carbon film on TEM grid
and observed in TEM.
• No cryo temperature needed.
Image on the right: Jones C.G. et al. ACS Cent Sci. 4(11) 2018.
Product quality control.