Structural Biology - Cryo EM

Question 1

What are the advantages of cryoEM?

Answer 1

Allows molecules to be studied in their native environment
In a lipid/membrane environment
Other techniques have non-physiological buffers
Functionally relevant conformations (No crystal packing artifacts)
Can understand dynamics of protein
Relies on only a few mico litres of material
Allows imaging of a large scale of complexes
Macromolecular structures –> atomic resolution
High energy electrons are waves
Lambda smaller than interatomic distances
To achieve very high resolution information
Diffraction limit for EM not a problem like for light microscopy
Wavelength of visible light used for light microscopy
High energy electrons used for EM
Electrons can be focused by magnetic lenses
Images more powerful than diffraction patterns
No need to solve phase problem (image contains amplitude+phase info)

Question 2

What are the challenges of using cryo-EM?

Answer 2

EM operates in high vacuum to avoid unwanted scattering of electrons
Ex. scattering due to air so water must get sucked out of the sample before imaging
Ionizing radiation
Damaging to sample and gets destroyed when you image it
Forced to limit exposure of sample to electron beam to avoid damage
Collect very noisy samples (so must optimize noise:signal ratio)
CryoEM samples move when irradiated (beam induced motion) leading to blurry pictures

Question 3

Give a brief overview of preparing cryoEM (vitrification)

Answer 3

Vitrification method (to prepare sample)
Sample is transferred to a metal mesh
Sample is plunged into frozen liquid ethane
Sample forms a thin film across holes
Water vitrifies around the sample
Allows protein to be loaded into a vacuum

Question 4

Give a brief overview of how image analysis for 3D structures works

Answer 4

Proteins in random orientations are hit by an electron beam leaving a trace on the image
Computer places similar images in the same group
Computer generates high resolution 2D image from all the thousands of images
Computer calculates 3D reconstruction

Question 5

How is the Fourier transform applied in Cryo-EM (Broad answer)?

Answer 5

Electron microscope has an electron gun that is firing electrons at the sample
Sample diffracts electrons with Fourier transform
Electron lens (grey) refocuses electrons and inverse Fourier transform is carried out
In X-ray crystallography: Same concept but no electron lens and can’t do the inverse Fourier transform as missing information for it

Question 6

What are the properties of waves and how is frequency related to wavelength?

Answer 6

3 parameters in all waves: Frequency, Amplitude, Phase shift
In a multi-dimensional wave there will be multidimensional frequency
Frequency = 1/wavelength
Units for frequency: units^-1
Example:
Wavelength = 4 units
Frequency = 1/4 units^-1

Question 7

Explain starting with 2 piano keys how the Fourier transform works

Answer 7

Play 2 keys on piano with different sounds
Oscillations in air pressure
Composite wave of multiple frequencies combined
Into ear, neurons triggered by two frequencies
Fourier transform: Seperation of different sine wave frequencies
Get 2 different frequencies (loud and quiet)

Question 8

What is the Fourier transform?

Answer 8

Fourier transform calculates the component sine waves of a composite wave
Ex. composite wave of 4 sine waves with different amplitudes and phase shifts
Then plot individual waves in reciprocal space (represents wave in terms of its frequency instead of wavelength, frequency = 1/wavelength)
Frequency/reciprocal space units vs amplitude AND Frequency/reciprocal space units vs phase
look at image in notes
No two waves will have the same FT (it is a unique solution)

Question 9

What is the inverse Fourier transform?

Answer 9

Inverse Fourier transform puts the individual wave components back together
Frequency vs amplitude or phase shift graph –> sine waves of different frequencies –> multiple waves are added together to form a composite wave
Highest frequency wave also has the lowest amplitude

Question 10

What is the DC component?

Answer 10

The DC component is a wave with infinitely long wavelength
Frequency = 1/infinity = 0 units^-1
Has highest amplitude
DC = direct current
Used to adjust baseline of a wave

Question 11

Explain a 2D Fourier transform for a wave of frequency 2 along A and 0 along B

Answer 11

Waves have no directionality: 2,0 wave AND -2,0 wave (called Friedel pairs)
Central point (0,0) is the DC component (has frequency 0 so infinite wavelength along a and b)
DC component: Used to raise or lower background greyscale
Here amplitude of all waves is between 0 and 1
All other waves have zero amplitude
Look at image in notes

Question 12

Explain how a 2D Fourier transform works for a composite wave. Wave 1: frequency 2 along a and frequency 0 along b. Wave 2: frequency 3 along a and b

Answer 12

Combine 2 waves to form a composite wave then Fourier transform
5 points: 2 points for every wave and a DC component
(-2,0) (2,0) (3,3) (-3,-3) (0,0)

Question 13

How is Fourier transform and inverse Fourier transform used for more complex 2D images?

Answer 13

Start with a 2D image and perform Fourier transform
Extract individual Fourier components
Get many dots with different brightness
Brightness corresponds to amplitude
Highest amplitude is the DC component
Next highest amplitude is 1,0 and next is 3,0
Use inverse Fourier transform for the highest 3 amplitudes
Get an image of the 3 composite waves

Question 14

How can filters be applied to the Fourier transform and what would the resulting image look like?

Answer 14

Filter out highest spacial frequency information: get blurry image
Filter our low frequency info (only have high), remove DC component that adds a level of grey = only sharp features and dark image
Only medium frequency and no DC = blurry and not sharp, dark image

Question 15

How are phases important in the Fourier transform and to make an image?

Answer 15

Phases are more important to describe the image than amplitude
Need amplitude AND phase information to make an image

Question 16

Compare electron and light microscopes

Answer 16

Light microscope: source is visible light, lens focuses light
Electron microscope: source is electrons, lens magnet focus electron waves, operates under vacuum conditions (don’t want electrons to be scattered by air)
Both have:Same architecture
Condenser lens and objective lens
Detection mechanisms
Apertures

Question 17

What is resolution and why does electron microscopy have higher resolution than light microscopes?

Answer 17

Resolution: the ability to resolve 2 points
Resolution is directly proportional to the wavelength of light used for imaging
Visible light = 0.61 x 400nm wavelength = 200nm resolution limit of instrument
Electrons = 0.61 x 0.0025 wavelength = 0.0015nm resolution limit
Electrons have a much smaller wavelength so have a smaller resolution limit
Will not get this resolution for a structure: due to radiation damage and flexibility of sample

Question 18

What occurs when an electron encounters an atom - what different types of electrons are there?

Answer 18

Transmitted electrons: Most electrons pass through the atom (80%) and are not diffracted
Elastically scattered electrons: contribute to image
Inelastically scattered electrons: energy is not conversed –> energy released as ionizing radiation –> contributes to noise in image and damage to atom
3x as many in elastically scattered than elastically scattered
Back scattering: can’t produce and information for image

Question 19

What is the structure of an electron microscope?

Answer 19

Electrons extracted from source
Condenser system: series of lenses focus the beam and controls intensity
Sample: scatters electrons
Objective lens: generates contrast
Objective aperture: removes highly scattered electrons
Projector system: magnifies the signal
Imaging: detection methods
High vacuum to prevent unwanted scattering

Question 20

How are electrons emitted from the source at the top of the microscope?

Answer 20

Tungsten filament:
Diffuse source of electrons
Accelerating electrons heats up filament
Extraction of electrons from filament tip under high pressure/vacuum
Thermionic emission
Less coherent
Like tip of light bulb

Field emission gun:
More coherent electron beam
Like laser pointer

Question 21

Why are magnets required in an electron microscope?

Answer 21

Electron waves are focused by magnets
Copper coils are wound tightly into a disk
Electron beam passes though the middle hole
Change current of copper wires to focus/bend path of electrons

Question 22

What are the 3 lenses of an EM?

Answer 22

Condenser lens: control intensity + convergence of electron beam, focus initial electrons from source
Objective lens: generates contrast
Projector lens: magnifies image before detector

Question 23

What are apertures?

Answer 23

Apertures: removes highly scattered electrons
Metal plate with holes
Condenser aperture: reduces spherical aberrations, reduces spot size
Objective aperture: increases contrast and removes scattered electrons that reduce contrast

Question 24

Why are direct electron detectors useful compared to non-direct electron detectors?

Answer 24

Before: no direct electron detector
Convert from electron to light signal and back to electron signal = loss of information
Now: direct electron detector
No intermediate step, can detect electron scattering without conversion to light
Increased sensitivity of detector (can detect lower electron doses)
Better signal to noise
Fast readout (can take movies)

Question 25

How are direct electron detectors useful when collecting images of samples in the ice

Answer 25

Sample is trapped in vitreous ice
Electron beam causes changes in potential energy of system
It releases stress/tension so particles that are trapped in the ice can move
If collecting only one snapshot (indirect detector) = blurry image because particles are moving
Direct electron detectors (takes many images) = corrects the movement of particles so clear image

Question 26

What is contrast?

Answer 26

Contrast: the difference in intensity between scattered and unscattered electron waves
Difference between light and dark
Amplitude contrast: corresponds to the particle properties of an electron
Phase contrast: corresponds to the wave properties of an electron - this is the most important aspect

Question 27

What is amplitude contrast?

Answer 27

Amplitude: difference from centre line to the top of the crest or trough of a wave function
Amplitude contrast: detecting difference in amplitude of the scattered and unscattered electrons
Differences in amplitude can occur due to:
Thickness of material - not important in cryo-EM because material is very thin
Density of material - if atoms have similar density, doesn’t contribute to amplitude contrast. If stained with heavy metal –> can increase contrast because they scatter strongly
Scattering from a lattice - imaging diffraction of electron beam when it interacts with a crystal lattice
In cryo-EM: amplitude contrast is small
It is enhanced by the objective aperture which removes highly scattered electrons

Question 28

What is phase contrast (how do electron waves interact with sample waves)?

Answer 28

Electrons are plane waves that travel down the column of the microscope (straight line)
Planar wave interacts with spherical wave (from sample) to form a complex wave (sum of 2 waves)
Complex wave propagates like ripples in a pond
When electron encounters different scattering centres, it will ripple in different ways
Waves can be constructive (in phase) or destructive (out of phase)
Spherical wave oscillates from 0-360 degrees

Question 29

How does the objective lens focus the electron beam with one scattered wave at a low scattering angle? what does the detector see? What is the vector diagram?

Answer 29

Incident plane wave + spherical wave –> complex wave + unscattered plane wave (80% are unscattered)
Scattered wave has a phase shift of 90 degrees
Lens: bends/focuses electron beam
Introduces an additional phase shift of lambda/4 to bring back wave to detector (lambda = wavelength of electron)
Scattered and unscattered beam have different path lengths to reach the detector
Detector sees only the sum of the scattered and unscattered waves
Records the the intensity (amplitude squared)
Vector diagram showing changes in amplitude of wave
Red = unscattered wave with high amplitude
Black = scattered wave phase shift 90 degrees and lower amplitude (because scattering centres don’t scatter wave that much)
Lens shifts scatters wave by lambda/4 = total phase shift is 180 degrees
Scattered wave contributes maximally but in opposite direction of unscattered wave = negative contrast
Look at image in notes

Question 30

What happens when there are multiple scattered waves at a high scattering angle and what is the vector diagram?

Answer 30

Many waves are being scattered at different angles
Higher scattering angle means wave is carrying higher resolution information
A wave is scattered at a higher angle
Phase shift of 90 degrees + lens phase shift of lambda/2 = 270 degrees
Vector diagram shows changes in amplitude of the waves
Minimal difference in amplitude of resulting wave (contributes minimally) - because wave is shifted “sideways” and not in the opposite direction which would reduce the contrast

Question 31

What happens when a wave is scattered at an even higher angle?

Answer 31

Even higher scattering angle
Phase shift of 90 degrees + lens phase shift 3lambda/4 = 360 degrees
Scattered wave contributes amplitude in same direction = contributes maximally to image = positive contrast

Question 32

How does the structure of the lens allow it to bend electrons?

Answer 32

Lens: electromagnetic coils that focus electron waves and causes an additional phase shift
Strength depends on current through the lens
Have spherical aberrations = magnetic field is not uniform
Electrons are bend more strongly at the periphery (for higher scattering angles)

Question 33

How is defocusing used to generate contrast?

Answer 33

Everything in focus: no contrast (very hard to see image)
To focus: bend electrons more strongly and shorten path length to reach image plane (by applying weaker current)
Defocusing: generates contrast
Objective lens is used for defocusing
Can control amount of defocus through the phase contrast applied to image
Want to be slightly under focus: beam will not point directly at one point
Will increase contrast: because there is more difference between scattered and unscattered wave
If hit it with focused beam would get no contrast because waves will be pointing at same point

Question 34

Explain the contrast transfer function

Answer 34

CTF: Detecting oscillations in positive and negative contrast due to scattering angle of the electron wave
lambda/4: negative contrast
lambda/2: minimal contrast
3lambda/4: positive contrast
When scattering angle is increased, there is higher resolution - because oscillations between positive and negative contrast become closer together
Defocus introduces a phase shift in the electron wave which affects the CTF
Oscillations can leave artefacts in the image
Defocus causes distortions and delocalisation of signal so need to account for/correct for it in the CTF otherwise get distorted image

Question 35

What are some other ways to generate contrast?

Answer 35

Induce 90 degree phase shift via carbon plate in microscope
Thin film with hole in the centre
Can image samples at focus and still see high contrast images
Used for imaging of small protein complexes or thicker samples for tomography

Question 36

What are the size requirements for cryoEM?

Answer 36

Before: EM only used for big molecules ex. ribosomes
Now: direct electron detectors + algorithms allow detection of smaller molecules (>150kDa)

Question 37

What are the quantity requirements for CryoEM?

Answer 37

Only need little material
3.5-4.5 micro litres
0.01-0.1mg/ml
Less than x-ray crystallography

Question 38

What are the environmental requirements for CryoEM?

Answer 38

Can visualize in lipid/membrane environment (more native)
Can’t do this with x-ray

Question 39

How can homogeneity be ensured for cryoEM? Why is it important to maintain it?

Answer 39

Sample purity is important (like X-ray)
Material needs to be homogenous (all the same) to get a high resolution structure
First purification with affinity/size exclusion chromatography
Test purity with: SDS-PAGE and activity assay
Then use negative stain: tool to optimize sample preparation

Question 40

explain compositional and conformational heterogeneity

Answer 40

Compositional heterogeneity: example only 2 of 3 complexes binding
Conformational heterogeneity: how to limit trapping of different conformations (ex. inhibitors)

Question 41

What is the main goal of sample preparation?

Answer 41

Improve signal to noise ratio
Preserve native structure of protein
Microscope operates under a vacuum so aqueous material is dehydrated OR frozen in vitreous ice

Question 42

How does negative stain work and why is it used?

Answer 42

Copper grid has thin carbon film on top that protein absorbs to
Carbon film is very hydrophobic –> need to put in a glow discharge machine that applies a negative charge on the carbon surface –> sample absorbs to grid and liquid distributes over surface
Apply sample and series of wash steps before applying heavy metal stain
Heavy metals are radioactive: needs to be done in a controlled environment
Heavy metals have high atomic numbers –> will scatter electrons more
Increases amplitude contrast
Example: Uranium, tungsten
Are only imaging stain around object, not the object itself so sample is not sensitive to radiation damage
Reveals shape of molecule

Question 43

What are the advantages and disadvantages of using a heavy metal stain?

Answer 43

Advantages:
Very fast speed of screening
High contrast: doesn’t suffer from low signal to noise ratio
High amplitude contrast because of heavy metals with large nucleus that highly scatter electrons
Not sensitive to radiation damage, can increase radiation dose
Biological material is not present, only the stain around it

Disadvantages:
Resolution limited by size of stain (20Angstroms)
So not good for chemical information
Protein distortion in depth/thickness

Question 44

What is vitrification and how is it used as a sample preparation technique?

Answer 44

Vitrification: frozen water in a transparent glass like state so water does not scatter electrons (NOT crystalline state)
Obtain vitrified state depending on temperature and speed by which it is frozen
Sample is put in -190 degrees (temperature of liquid nitrogen)
Not all ice is the same: ice formation is dependent on the temperature/speed by which it is frozen
Ice contaminants or surface contaminants after freezing
Crystalline ice is bad
Everything has to be kept at liquid nitrogen temperatures: once sample is frozen it has to stay frozen

Question 45

What would be the ideal way of getting particles into the ice for vitrification?

Answer 45

Want to get particles into ice in a randomised distribution of orientations
Particles that are imaged need to be in all different orientations
Forces at air-water interface
Air is hydrophobic and liquid is hydrophilic
Impacts how molecules are embedded in the ice

Question 46

What is the reality of getting particles into the ice for vitrification?

Answer 46

It is a big challenge to get a random distribution of orientations
Introduce carbon support: pulls particles down so they are no longer influenced by the air-water interface
But molecules can have preferential orientations when interacting with the grid
Air water interface can cause proteins to denature

Question 47

How is protein adsorption to a support controlled? which type of grid is used?

Answer 47

Controlling protein adsorption to a support
Before: used grid with carbon film but this caused extra scattering = BAD
Now: use graphene which is transparent to electron beam (doesn’t scatter) = decreases background noise
Controlling charge applied onto grid (from glow discharge machine) used to control adsorption of protein

Question 48

How is the sample loaded into the EM after vitrification?

Answer 48

Use a temperature controller (similar to room-temp holder)
Once everything is frozen needs to stay frozen
Holding station is filled with liquid nitrogen

Question 49

What is radiation damage? How does it arise?

Answer 49

Radiation damage arises from inelastically scattered electrons
No conservation of energy so electrons are released as ionising radiation which destroys bonds in the sample
Increased time of exposure = increased damage
First the high resolution information is destroyed
Same as in x-ray crystallography

Question 50

How is low dose imaging used to minimise radiation damage?

Answer 50

Use low electron doses at low magnification to find area that we want to take a picture of
At high magnification find the focus point (area close to where exposure is occuring)
Defocus image
Apply parameters (defocus values, exposure settings, roundness of beam)
Then point beam at the sample
This minimizes radiation damage
Need to redefine focus every few points (as sheet is not perfectly flat and parameters will change)

Question 51

How do direct electron detectors help to minimise radiation damage?

Answer 51

Use of direct electron detectors = increased sensitivity = don’t need as much exposure
Can take very fast exposures to collect a movie with many frames
First frame will have little radiation damage –> give this frame a higher weighting when aligning images

Question 52

Give an overview of the sample preparation and imaging process

Answer 52

Sample purification
Preparation (negative stain, sample heterogeneity, quality control)
Plunge freezing - formation of vitreous/glassy ice that is inserted into microscope under vacuum
Form an initial model via tomography –> particle alignment –> final structure
OR subframe collection (collect fast exposures, all have diff. radiation damage) –> low dose sample navigation and imaging –> align and average frames, determine orientation parameters (eliminate noise) –> estimate defocus and CTF –> particle alignment and classification –> 3D reconstruction

Question 53

How are 2D projections of 3D images used to reconstruct the 3D object? How is the signal to noise ratio enhanced?

Answer 53

Taking 2D projections of 3D objects and then reconstructing original 3D object
Need to know how the original object was oriented within the ice
Collecting single particles: need 100,000 - 1,000,000 single particles to get a high resolution 3D image/structure
Collect images in many different orientations
Individual images are very noisy (due to non-interacting/inelastically scattered electrons) and have low contrast
Enhance signal : noise ratio by taking the average from individual single images

Question 54

What is the projection theorem?

Answer 54

Projection theorem: 2D projections from different angles can be used to reconstruct the 3D structure
2D projection of 3D object is a central slice through the 3D Fourier transform of the object
Collect 2D projections from different orientations of the object
Need to fill up all 2D Fourier transforms in the 3D box (otherwise get distortions)
Take inverse FT to get reconstruction of 3D structure

Question 55

How is the projection theorem applied to EM?

Answer 55

Electron beam interacts with the sample in the electron microscope
Sample is embedded in ice in random orientations
Leads to scattering of electron beam
Detector measures intensity of electron beam to collect a 2D projection of the original sample
Collects series of 2D images of the sample at different angles
Use 2D Fourier transforms
Need to fill up the sphere with 2D FTs to make a 3D FT (in Fourier space)
Each 2D FT is related to another 2D FT depending on the rotation of the object
Inverse FT to reconstruct the 3D object (in real space)
Noisy, low contrast images make determining orientations problematic - inaccuracies in the alignment degrade the quality of the average

Question 56

How are orientation parameters determined? How are these used to correctly remodel the 3D structure?

Answer 56

Need to know the relationship between each of the 2D FT to correctly fill 3D FT
Need to know orientation and rotation of 2D object
Initial position, translation (in X and Y) and rotation defines orientation parameters
Decrease noisy 2D projections by averaging (increase signal to noise ratio)
Need to know orientational parameters as these are used to align randomly oriented projections
If misalign projections = get average that is not a true representation of the object (distortion)

Question 57

What 2 factors impact the average of 2D projections?

Answer 57

Alignment (using orientation parameters) and heterogeneity (averaging over molecules that are not the same)

Question 58

What are the sources of heterogeneity in cryo-EM? How do you avoid it?

Answer 58

Compositional heterogeneity: Complex of 3 subunits but only 2 are occupied (because of averaging see partial occupancy for that position), complexes disintegrating, contaminants
Conformational heterogeneity: Proteins are dynamic, due to averaging see partial occupancy of conformation

Single particle cryo-EM allows in silico (computer) purification processes

Avoid overlaps when taking image to avoid heterogeneity
Concentration should be low enough so there is enough space between objects to prevent overlaps
Ice thin enough to have one level of depth

Question 59

Why is it important to correct for motion?

Answer 59

Molecules freeze on ice = physical strain on ice = electron beam releases strain = molecule in ice can rotate
Blurry image if don’t correct for motion of particles

Question 60

Why is it important to estimate defocus and correct for it in the CTF?

Answer 60

Need to image at defocus to get contrast (if image at focus get no contrast)
Find focus and then take the image under focus
Get distortions in CTF (oscillations of positive and negative contrast)
Need to calculate CTF for each image and apply corrections for it to get positive contrast
If didn’t correct for CTF: will get distortions in the image

Question 61

How are particles selected and aligned?

Answer 61

Individual particles are selected using automated software/machine learning
Still contains noise and ice contaminants
Alignment occurs in 2D
Rotations of particles doesn’t matter at this stage
Particles are separated from ice contaminants and denatured particles (will only have shadow of object)

Question 62

What is projection matching?

Answer 62

Projection matching: determines rotational relationships
Estimate an initial 3D low resolution model by tomography to find the reference object
Reference object is rotated by certain amount to generate a portfolio of rotations in silico (in computer)
Find the best match of projection and known image to determine rotation
Iterative process: always improving model that feeds into the next cycle

Question 63

How are samples purified in silico and why is this important?

Answer 63

Refined model could be a mixture of conformations
Because only had crude purification steps (getting rid of contaminants) for 2D images
Need to remove things that are not your particle / separate 2 conformational states of a protein
Separate in-silico (computer) by 3D classification methods
Ex. When not able to isolate by chromatography

Question 64

What is a validation map? Why is signal to noise important in this?

Answer 64

How do I know who has got it right?
Signal to noise ratio is important
Detergent micelles contributes to noise in membrane proteins, can lead to missasignment of particles to orientations
CCD detectors (indirect) = source of noisy images due to conversion to light intermediate
Direct detectors improved signal to noise ratio
3 structures for HIV: is there secondary structure in the map that matches the model?

Question 65

What is model bias?

Answer 65

Aligning noise to match model –> will align noise into shape –> phantom shape
No particle in raw image: if align noise to model of a molecule get a low resolution image of the molecule that is not actually there