Final Material

Question 1

What’s is the data collection and process pathway of Cyro-EM?

Answer 1

1. Motion correction
2. CTF Correction
3. Particle picking
4. 2D classification
5. Ab Initio Reconstruction
6. Refinement

Question 2

Why do we need to glow discharge the grids for cyro EM?

Answer 2

Stores grids become hydrophobic so when we place the sample it will not want to spread so we must discharge to make it hydrophilic so proteins will want to spread out

Question 3

What are differences in contrast and resolution between negative stain EM and cyro-EM?

Answer 3

Negative staining EM is at a lower resolution but has a higher contrast so it is easier to pick out particles, while cyro-EM is at a lower contrast but a higher resolution so it is hard to pick out particles

Question 4

What is the goal when plunge freezing samples? What are the two methods?

Answer 4

To do it as fast as quickly so that water molecules stop moving before they crystallize.

1. Blotting and plunging
2. Spraying sample on as it is plunging

Question 5

Why do we need to clip our grid?

Answer 5

Allows the grid to lie flat when inserting into the microscope

Question 6

Why must we perform motion correction in cyro-EM? How is this done?

Answer 6

We need to do motion correction because as the electrons hit the ice, it melts and the samples move, we correct for motion by aligning each frame and averaging together to reduce blurring in images

Question 7

What is CTF? How does this relate to defocus?

Answer 7

CTF is contrast transfer function and it is similar to band pass filter where we lose information in the final image that is collected, so only limited amount of resolution is let through.

If we defocus ( move slightly away from focal point) at different values we can get and information at different resolution and contrast.

Question 8

Explain the relationship between the defocus value, resolution, power spectrum, and the contrast.

Answer 8

Higher defocus value = higher contrast and low resolution so the power spectrum (thon rings) will be very close together and less spaced out

Lower defocus value = lower contrast and high resolution so the power spectrum will show thon rings very spaced out

Question 9

What are the methods in particle picking?

Answer 9

1. Manual picking
2. Cross-correlation: automate particle picking by picking it based on high contrast areas
3. Blob picker
4. Template-based picking: using a homologous model as a template and create reprojections in 2D view and tell software to look for these type of images
5. Train a neural net: training AI

Question 10

What is the goal in 2D classification? What classifies this?

Answer 10

The goal is to throw away junk particles. Junk particles include overlapping particles, breaks in ice, contamination and uncentered particles.

Question 11

What is the projection theorem?

Answer 11

The 2D projection of a 3D object in real space is equivalent to taking a a central 2D slice out of the 3D Fourier transform

Question 12

What are the steps from going 2D to 3D?

Answer 12

1. Record projections
2. Calculate 2D Fourier transform of each image
3. Populate a 3D Fourier transform of an object with the slices
4. Inverse Fourier transform in 3D to recover the object

Question 13

What is Fourier shell correlation?

Answer 13

Gives an estimate of the cyro EM map resolution by collecting data and splitting it into 2 halves, calculate map, compare and at resolution where they don’t match is where the cut off. Correlation between the 2 3D maps, each calculation in Fourier space from and independent half of the data’s as a function of resolution. A measure of the signal to noise ratio in the map which decreases with increasing resolution.

Question 14

Explain model building in cyro? Why is difficult to build one ab initio?

Answer 14

Fit known structures into the map, start with a best guess and make changes. When building a model ab initio: figure out the path of the backbone, build secondary structure and match that to bits of the sequence that are predicted to have that structure, fit side chains to figure out where you are in the sequence.

It is difficult because there is no refinement of the map based on the model, THE MAP DOES NOT CHANGE AS YOU IMPROVE THE MODEL

Question 15

Explain some of the hurdles in cryo-EM?

Answer 15

1. Radiation damage: resolution worsens as it is more damaged
2. Sample preferred orientation: particles only show one orientation which limits the angles we see making it difficult to get a 3D projection
3. Conformational heterogeneity: proteins having different confirmations that don’t align

Question 16

What are some solutions to the preferred orientation limitation and sample heterogeneity in cyro?

Answer 16

In preferred orientation, we can tilt the sample to get different angles or we can have faster plunging times to give the sample less time to get to the water-air interface

For conformational heterogeneity, there is a computational method to pull out specific protein structures

Question 17

What are the advantages of cryo EM compared to crystallography? Limitations?

Answer 17

• Does not require crystals
• Samples can be partially heterogenous
• closer to physiological conditions
• requires a small amount of samples
  Limitations
• sample movement
• radiation damage
• preferred orientation
• conformational heterogeneity

Question 18

Coarse grained energy functions:

Answer 18

Effectively captured hydrophobic burial, formation of secondary structure and atomic overlap

Question 19

High-resolution, atomically detailed energy functions:

Answer 19

More accurate but slower to evaluate and introduced many local minima into the landscape and makes them harder to navigate efficiently.

Question 20

What are the computational models of protein energetics?

Answer 20

Coarse grained energy functions
High-resolution, atomically detailed energy function

Question 21

What are the steps in template based modeling?

Answer 21

1. Select a suitable structural template
2. Align the target sequence to the template structure
3. Perform molecular modeling to account for mutations, insertions and deletions present
4. Side chain optimizations at mutated position and rebuild the backbone around insertions and deletions

Question 22

What are the steps in template free modeling? What are the requirements?

Answer 22

1. Construct a multiple sequence alignment with related proteins
2. Predict secondary structures and residue contacts
3. Assemble 3D models
4. Refine and rank models

Requirements:
- Conformational sampling strategy for generating candidate models
- ranking criteria for selection of native-like conformation

Question 23

How can we make contact predictions from residue covariation? What are correlation mutations?

Answer 23

Correlated mutations are when one residue is mutated, another residue also changes so they are most likely to interact. A MSA can be used to predict residue contacts based on correlated mutations. This covariation is attributed to the need to preserve favorable residue interactions. When measuring for coevolution we get contacts since they are evolutionarily conserved together.

Question 24

What is the limitation of contact predictions?

Answer 24

The dépendance of deep MSA (larger databases, number of sequences in the multiple sequence alignments)

Question 25

What is fragment assembly?

Answer 25

An approach to conformational sampling and predictions include torsion angles and secondary structures, models built from short contiguous backbone fragments (3-15 residues) based on protein of known structure. Libraries are selected for each fragment to pose a possible local backbone structure. Several thousand simulations conducted with fragment insertions until the final lowest energy model is found

Question 26

What is Monte Carlo Sampling?

Answer 26

A simulation to build 3D model from fragment assembly. Employs randomly selected conformational moves and occasional uphill steps to escape local minima. Favorable sampling moves represented by green arrows and rejected ones are in red arrows. Downhill moves that decrease energy are accepted w probability of 1 and those moving uphill with a low probability. In fragment replacement moves, we start from a random or fully extended confirmation and repeatedly select a random window of protein and insert the structure of a randomly selected fragment from the corresponding fragment library.

Question 27

How can we perform model refinement in computational structural biology?

Answer 27

- Monte Carlo simulations - molecular dynamics simulation

Question 28

Explain molecular dynamic simulations

Answer 28

Model is placed in a simulation box surrounded by water molecules and force field calculations for each atom is determined as time progresses in femtoseconds which provides broad sampling of the energy landscape

Question 29

What are the two methods in protein design? Explain them

Answer 29

1. Template based design: sequence and structure of naturally evolved proteins are modified to achieve new functions 2. De novo design (template free): novel protein backbones and sequences are generated from scratch guided by design requirements and physiochemical constraints, no restricted by evolutionary contraints, design proteins that you haven’t seen before and can be done with fragment assembly

Question 30

Explain de novo interface design. What are some applications

Answer 30

Docking: create a model in which the proteins are brought near each other so that their surfaces are adjacent, sequence optimization simulations are then used to search for amino acids to stabilize interaction, emphasize hydrogen binding and good hydrophobic packing Applications: re-engineering a signaling molecule to get a specific downstream effect, stabilizing proteins

Question 31

How can we validate our protein design?

Answer 31

- stability: DSF - oligomerization: SEC - functional assays - high resolution structural determination

Question 32

What is the idea to drug pathway?

Answer 32

- identify a disease and and unmet need to treat it - identify a target that is causing the disease - identify the compound that will interact with the target - drug development - clinical trials

Question 33

What is drug design?

Answer 33

A process of finding new medicine based on knowledge of the biological target. We design molecules that are complementary in shape and charge target they will bind.

Question 34

What is drug discovery?

Answer 34

A process that identifies synthetic molecules for comprehensive evaluation as a potential drug candidate and involves screening hits, medicinal chemistry, and optimizations of hits

Question 35

What are the screening strategies in drug design?

Answer 35

- high throughput screening: libraries or drug like compound evaluated against a protein target in a biochemical or cell-assay - virtual screening: a virtual library of existing compounds may be docked into a 3D protein structure to computationally predict their activity against a target

Question 36

What is structure-based drug design?

Answer 36

This relies on the knowledge of the 3D structure and computer modeling techniques. If the structure of the target is known but not bound, we need to do virtual screening to dock the compounds and perform molecular dynamics to virtually improve binding. If the structure is not known, we need to determine the structure of the target bound to compound and redesign to improve binding. Can reduce time and cost by narrowing down what is tested experimentally

Question 37

What is the remarkable cancer drug?

Answer 37

Gleevac

Question 38

Explain Gleevac and the disease it included.

Answer 38

Gleevac is the first successful targeted therapy to help CLM. It is a kinase inhibitor and it had a single target

Question 39

Explain the active and and inactive state of a kinase.

Answer 39

Active: DFG “in” conformation and the alpha C helix “in” Inactive: DFG “out” conformation opening a deep hydrophobic binding pocket

Question 40

How can the gatekeeper residue help with targeted drugs?

Answer 40

Smaller gatekeeper residues will allow the design of inhibitors that will be more targeted because it gives enough room to fit smaller inhibitors in that back pocket compared to the larger hydrophobic pocket

Question 41

How can gatekeeper residues lead to kinase drug resistance?

Answer 41

If there is a bulkier gatekeeper residue, this can sterically exclude inhibitors because drugs cannot fit into these pockets

Question 42

How can gatekeeper residue mutations help us?

Answer 42

Usually, Gly is not a gatekeeper residue but they will be in diseases so this allows us to make a drive that targets kinases with these Gly gatekeeper residues

Question 43

Explain a type 1 inhibitor

Answer 43

This inhibits kinases that are in the active state and binds into the ATP binding site

Question 44

Explain type 2 inhibitor

Answer 44

This inhibitor inhibits inactive state of kinases in the ATP binding pocket and this create a a large hydrophobic pockets so we can extend drive to fill this up

Question 45

Explain type 3 inhibitor

Answer 45

This is an allosteric inhibitor and it binds the Kinase in an active state at a site near the ATP Binding pocket

Question 46

Explain type 4 inhibitor

Answer 46

This is an allosteric inhibitor that binds at a remote location from the ATP binding site

Question 47

What is Bidentate binding?

Answer 47

This is when a compound simultaneously binds to the ATP binding pocket and also a unique pocket outside the ATP cleft to make the compound selective. This improves affinity and is more selective

Question 48

What is a bivalent inhibitor? Give an example

Answer 48

Inhibitors that target kinases at two different binding sites. This combines a type 1 inhibitor with a compound that targets a remote secondary binding site. Abl kinase and fused with BCR to drive cellular proliferation and leukemia (Gleevac)

Question 49

How do we solve the phase problem in X-ray Crystallography?

Answer 49

Fourier transform and determining the structure factor of each reflection to determine the electron density map . We can also perform multiple isomerous replacement and molecular replacement

Question 50

When performing crystal optimization in crystallography, how can we do this?

Answer 50

Sparse screening matrix by varying the PEG and salt

Question 51

What do we want to prevent when harvesting our crystals?

Answer 51

Crystalline ice and breaks

Question 52

What is the process for protein crystallization?

Answer 52

1. Start with pure high quality protein 2. Use précipitant solutions to force protein to crystallize 3. Optimize crystals using sparse matrix screening 4. Harvest crystal in loop 5.Transfer to cryoprotectant solution 6. Flash freeze in liquid nitrogen 7. Store for later data collection

Question 53

Explain the X-day Crystallography process.

Answer 53

1. Crystallize purified protein 2. Freeze crystals 3. Shoot crystals with high powered x-rays and collect diffraction images 4. Process images by measuring intensities and symmetry 5. Solve phase problem 6. Generate electron density map 7. Build model and refine data 8. Assess quality of structure

Question 54

What can the diffraction pattern in crystallography tell us?

Answer 54

The diffraction pattern is used to determine the space group which defines how the molecules are packed in the crystal lattice

Question 55

How is protein expression induced?

Answer 55

It is induced by adding IPTG

Question 56

What are inclusion bodies?

Answer 56

When proteins are misfolded and aggregate into cellular deposits