Lecture 3 revision Flashcards
Why are important things to distinguish about protein structure?
- Spatial localisation of atoms
- Location of co-factors, metal ions
- Substrates and inhibitors
- Proteins dynamics/flexibility
- Quaternary structure
Where are protein structures contained
Protein data bank
What unit is used in atomic scale
Angstrom
1 Angstrom = 0.1nm
Carbon atom = 0.77A
C-C single bond = 1.54A
X-ray diffraction overview
- Crystal irradiated with X-ray
- Diffraction pattern is collected
- Fourier transform produces electron density map
- Protein structure modelled
How does X-ray diffraction work?
- Diffraction of wave through a small slit creates interference pattern
- Diffraction occurs when wavelength and slit are equivalent in size (0.7A between them)
- Positional info about source is contained in the diffraction pattern
Second X-ray diffraction theory
- Crystals have regular arrays of proteins atoms, where atoms act as ‘slits’ with 1A spacing
- The equivalent wavelength is in the X-ray region
What does crystallisation require in X-ray diffraction
- Large amount of pure, homogeneous protein
- Additives used to promote crystallisation
Limiting factors of crystallisation and X-ray diffraction
- Can’t predict exact conditions - try many
- Crystal growth can take weeks
- Quality is important over quantity
Hanging drop
Drop protein and low precipitant concentration (salt/PEG)
Reservoir is a high concentration of precipitant
Vapour diffusion from drop to reservoir
Collecting data from X-ray diffraction
- Crystal mounted in a metal loop
- Cooled in liquid nitrogen, and soaked in cryoprotectant
- Intense, focused X-ray beam 0.6-2.3A
- 3D diffraction pattern collected by rotating sample or detector
- Synchatron is an example of machinery to collect data
What is a problem with data analysis using X-ray diffraction
Phase problem:
- Fourier transform can construct an electron density map from diffraction pattern
- Phases of pattern required but only amplitude recorded
Solving phase problem
75% use molecular replacement - model of similar template - causes bias
25% use experimental phasing - heavy-atom containing isomorphous crystals and look for anomalous scattering
Ad initio phasing (rare) - only when resolution is less than 1.2A can phases be calculated directly
Model building in protein crystallography
- Protein model fitted to electron density map
- Iterative refinement improves fit of model
Features:
- Protons aren’t seen
- C/O/N are indistinguishable
- Resolution varies in structure
- ‘B-factors’ provide estimates on variation in certainty
Overview of nuclear magnetic resonance spectrometry (NMR spectrometry)
- Protein solution placed in magnet
- Radiofrequency pulse applied, decay recorded
- Fourier transformed to 1D or 2D spectra
4.Structural constraint used to model protein ensemble
Theory around NMR spectrometry
- Certain nuclei posses quantum mechanical property of spin
- Non-zero spin manifests as permanent magnetic dipole
- Outside of a magnetic field spin states are degenerate
- Inside a magnetic field spin states have different energies
- Exact magnetic field depends on local environment; both bonded and non-bonded neighbours
- Info used to determine protein structure
Which nuclei possess spin in NMR spectrometry?
Depends on atomic mass and number
1H - 1/2 spin
12C - 0
14N - 0
16O - 0
13C - 1/2 spin
15N - 1/2 spin
Stable isotopes used to introduce observable nuclei
What are the three important features of NMR spectra in data collection?
- Integration - signal proportional to number of nuclei
- Chemical shift - A measure of the local magnetic field relative to internal standard, given in ppm
- Chemical splitting - spin state of coupled nuclei divide the peak; pattern of splitting gives information about coupled nuclei.
What type of techniques are used in data collection using NMR spectrometry
- Multidimensional techniques.
- Proteins have too many atoms for 1D spectra.
- Important concept is resonance transfer, where transfer of energy by spin-spin coupling through chemical bonds.
What did combining heteronuclear 2D and 3D experiments with NMR spectrometry reveal?
Resonance transfer reveals bonded connectivity and allows peak assignment
NOE spectrometry in data collection using NMR spectrometry
Short range, non-bonded interactions can be observed using NOE spectrometry.
Nuclear Overhauser effect: transfer of energy by spin-spin coupling through space
Overview of data analysis using NMR spectrometry
- Assign chemical shifts to residues using sequential contacts
- Identify NOE interactions
- Derive torsional angles
- Build protein model using experimental restraints
- Ensemble of structures produced
Overview of electron microscope
- Accelerated electrons provide near atomic resolution
- Electrons focused by electromagnetic imaging lenses
- High vacuum prevents collisions
- TEM and SEM
Transmission EM
- Penetrates <80nm
- Sample fixed, embedded in matrix, and thinly sliced
- Electrons pass through sample
- Electron-dense material appears as dark areas
- High magnification and resolution
Can provide quaternary structural info
Scanning electron microscope
- Sample coated with 50nm of inert material
- Electron beam seeps over sample
- Secondary electrons are ejected and collected
- Variation in intensity creates image with depth and contrast
Tomography using TEM
- Sample moves up
- Knife cuts top off sample
- Knife retracts
- Block face imaged
Modelling using TEM
EM modelling can be combined with crystal structures:
- EM provides quaternary organisation
- X-ray diffraction/NMR provide atomic resolution
Cryo-electron microscopy
- Samples flash-frozen with liquid ethane (-88.5 degrees)
- Improved electron detectors
- Improved software to transform 2D to 3D
Process of ‘single particle’ Cyro EM
Sample ->
Grid ->
Electron microscope after freezing ->
2D projections after data collected ->
Particle alignment and averaging after particles picked ->
3D map ->
3D model
Pros and cons of EMs
Pros:
- True imaging technique
- provides quaternary structural info
- Excellent for large/flexible molecules
- Sample prep simple
Cons:
- Expensive
- Few images achieve atomic resolution
Alphafold
Protein structure prediction software:
1. Database searches for similar proteins
- Sequence alignment - primary and structural
- Neural network analysis
- Model generation, three iterations
