NMR and SPR

Question 1

Give an overview of NMR?

Answer 1

NMR spectroscopy measures the absorption of radio-frequency energy by atomic nuclei with spin = 1/2 (e.g. 1H, 13C, 15N, 31P) when placed in a strong magnetic field
The electron cloud around the nucleus causes small changes in the resonance frequency – called the chemical shift (d)

Nuclei separated by up to 3 covalent bonds show splitting of their peaks – called J-coupling
1H nuclei that are within 5Å distance exchange magnetization according to distance – called the nuclear Overhauser effect (NOE)

Intensity - proportional to the concentration of the species

Question 2

Describe an NMR sample?

Answer 2

A 5mm diameter glass tube containing 0.5 ml of sample
NMR is insensitive - needs 50-500 mM = high protein concentration
Protein usually dissolved in buffer (Tris, phosphate, HEPES etc.)
NMR structures are in solution - no need to crystallise

Question 3

Describe the NMR spectrometer?

Answer 3

Higher field magnet = better sensitivity and better resolution
Current highest available = 1.0 GHz
Cryogenically cooled ‘probes’ significantly boost sensitivity (by 2-3-fold)

Question 4

What properties of proteins would we choose NMR to solve a protein structure?

Answer 4

NMR is the best choice for:
Dynamic / mobile or very flexible systems
Partly or fully unfolded systems
Weakly interacting complexes
Native environment e.g. in cells

Question 5

What properties of proteins would we not choose NMR to solve a protein structure?

Answer 5

NMR is not the best choice for:
High resolution structures of large proteins and complexes
Proteins that cannot be expressed in simple systems eg E. coli (isotopic labelling is required)

Question 6

Describe stable istopic labelling?

Answer 6

Need to enrich our sample in 15N (or 13C) – as these isotopes are present at very low abundance
Usually overexpress protein in E coli, as typically cheapest
Can use very simple substrates, e.g. 15NH4Cl, 13C-glucose

Question 7

What are some types of isotopic labelling?

Answer 7

Selective labelling/selective unlabelling
Deuteration
Uniform C/N labelling
Methyl selective labelling
Paramagnetic labelling
Fractional labelling
Segmental labelling

Question 8

Describe how NMR can be used to solve protein structures?

Answer 8

NMR can be used to completely solve a small protein or nucleic acid 3D structure (<25 kDa)
Complex process using multiple different types of NMR data
Uses computational calculations (e.g. ‘molecular dynamics’) to find structures that agree with the NMR data

Mechanism - sample preparation, spectroscopy, resonance assignments (which peak is what), gather information on - secondary structure/conformational constraints/global fold, calculate the structure which agrees with all the data, refinement and final ensemble

Question 9

Describe 1D proton spectrum of a small protein?

Answer 9

Each peak is a different hydrogen atom
Lots of information about structure and dynamics of the protein, but too much overlap to be able to extract it
So we move on to 2D

Question 10

Describe 2D heteronuclear NMR?

Answer 10

We can resolve overlap by increasing the dimensionality of the spectrum
2D spectrum spreads the 1D resonances out into two dimensions – allows interactions between them to be monitored
If the 2nd dimension is a nucleus other than 1H (e.g. 15N), it is called an ‘HSQC’
HSQC = heteronuclear single-quantum correlation

You can get one peak for each amide group for example - and therefore one peak per residue
2D 1H-15N HSQC gives a peak for every N-H group
Each amino acid residue (apart from Pro)
Each side chain of Asn, Gln and Trp (also sometimes His, Lys, Arg)
Provides a ‘fingerprint’ of the folded conformation of a polypeptide chain

Question 11

Describe 3D ‘ triple resonance’ NMR?

Answer 11

Separate into three independent frequency dimensions
E.g. HNCA experiment – correlates HN, 15N and 13C within each residue
Key types of experiment for resonance assignment
We can produce links between amino acids in the polypeptide chain

Question 12

Describe resonance assignment from 3D NMR?

Answer 12

Example: 3D HNCACB experiment
15N,13C-labelled protein
Here, we are matching ‘strips’ taken from the 3D spectrum that correspond to each 1H-15N pair in the protein
We’re looking for the same chemical shift in the same residues as it must be following on from the previous one
Software program assists matching of intra- & inter-residue correlations

Question 13

Give some NMR parameters and the structural information they give?

Answer 13

Chemical shifts - resonance assignments, secondary structure
J-couplings - dihedral angles, Karplus curves
New J-couplings - direct idnetification of H-bonds
NOE - interatomic distances
Solvent exchange - H-bonds
Relaxation/line widths - mobility, dynamics, conformational exchange, projection angles
Residual dipolar couplings - projection angles

Question 14

What are NOEs as structual restraints?

Answer 14

Nuclear Overhauser effect - measures interatomic distances
NMR can measure distances between protons < 5.5 Å apart in the folded protein by running NOESY (Nuclear Overhauser Effect SpectroscopY) experiments
This ‘web’ of 1H-1H distances can be used to work out the 3D structure

Question 15

What is significant about using NOEs for interatomic distance measuring?

Answer 15

In reality - due to spin diffusion - we can only classify NOE intensities into approximate distance ranges
Classify into ‘strong’, ‘medium’, ’weak’ by peak integral
Assign an ‘upper bound’ to the distance for each, e.g.
Strong = <2.5 Å
Medium = <3.5 Å
Weak = <5.5 Å

Question 16

Describe 2D and 3D NOE Spectroscopy?

Answer 16

A 3D NOESY separates the 1H NOEs into a 3rd dimension according to the chemical shift of the directly bonded 13C or 15N, massively reducing the overlap problem
Each strip contains NOEs from one NH group to all other hydrogen atoms close by
This can separate into aliphatic hydrogens and amide/aromatic hydrogens = reduce overlap

Question 17

Describe residual dipolar coupling as a structural restraint?

Answer 17

Residual Dipolar Couplings (RDCs) are long-range orientational restraints
They report on the orientation of a bond (e.g. H-N) relative to the orientation of the protein in the magnetic field
RDCs are measured in a partially aligning medium
They are complementary to NOEs (short-range distance restraints)

Effects average out in isotopic solution but if samples are partially aligned some of this effect is retained
Nuclei as little magnets affecting each other
Usually averages to zero but if you align it doesn’t

Question 18

Describe the workflow of NMR?

Answer 18

Optimize sample – construct length, buffer conditions etc. to get the best NMR data (weeks-months)
Prepare labelled samples – 15N, 15N/13C, AA-selective etc. (days-weeks)
Acquire NMR data – 2D, 3D, triple resonance, NOESY, RDC etc. (days-weeks)
Assign resonances (weeks)
Calculate structures (weeks-months)

Question 19

Describe the computer calculations of structures?

Answer 19

The structural data (NOEs, RDCs etc.) are used as restraints in computational molecular dynamics (MD) simulations
Standard bond lengths, angles and Van der Waals radii are assumed and enforced

Molecular dynamic simulations start with a bunch of random structures, then applies kinetic energy (i.e. ‘heat’) to the atoms so they can find their energy minima when ‘cooled down’ again (’simulated annealing’)
Structures that ‘disobey’ the NMR restraints are ‘penalized’ energetically
The ‘best’ structures (those that fit the experimental data best) have the lowest energy

Question 20

What is the outcome of these computer calculations?

Answer 20

NMR structures are ensembles of models that fit the experimental data
More variability where there are fewer restraints - these are often also more mobile regions

Question 21

Describe in cell NMR?

Answer 21

Cells are crowded environments containing many different proteins
In-cell NMR allows the study of proteins in their native environment
Allows post translational modifications and interactions with other proteins and ligands
Need isotopically labelled proteins in unlabelled cells

Question 22

What do we do for in cell NMR for bacteria/eukaryotic cells?

Answer 22

For bacteria - we grow in unlabelled media and then moved into labelled media just before we express our protein

For eukaryotic cells - we grow bacteria and isotopically label - it is then transferred into eukaryotic cells, the manipulated cells are then collected for in-cell NMR

Question 23

What are some in cell NMR limitations?

Answer 23

Limited sample viability time (cells die)
Can be difficult to get high enough concentrations of the protein of interest
Proteins that interact with multiple binding partners and have multiple conformations may be difficult to study

Question 24

Describe NMR studies of disordered proteins?

Answer 24

Some proteins are natively disordered and may or may not fold in response to binding a ligand.
These cannot be studied by techniques such as X-ray crystallography
NMR doesn’t need crystals so can be used to study disordered proteins and peptides
This also allows studies of protein folding

Question 25

Describe methyl NMR studies of large molecular machines?

Answer 25

We can’t normally study these large proteins but we can use methyl
Labelling only the methyl groups of Ile, Leu, Val, Met, Ala with 13C,1H (all else 12C and 2H) provides very sensitive probes of protein structure, function and dynamics
Applicable even to very large (MDa) complexes, e.g. 20S proteasome (1.1 MDa)

Question 26

Describe protein dyanamics from NMR?

Answer 26

Dynamics (protein movement) is essential for biology – protein structures are not static
NMR is one of the only tools available to understand dynamics at atomic resolution

Excited spins will lose their energy and ‘relax’ back to equilibrium with a characteristic time constant (T1 and T2 relaxation)
These processes are assisted by the fluctuating fields generated by motions of nearby spins (mostly 1H)
Measuring this relaxation (R1 and R2 rates) can tell us about dynamics (amplitudes and rates of motion) within the protein

Question 27

Give some examples of monitoring protein dynamics?

Answer 27

Catabolite activator protein CAP binding its cofactor cAMP
Fast (ps-ns) dynamics - conformational entropy & binding
Slow (μs-ms) dynamics - catalysis, folding and function

NMR elucidates mechanistic differences between anti-cancer drugs that target fibroblast growth factor receptor (FGFR) kinase

Question 28

What is NMR good at monitoring?

Answer 28

‘Gold standard’ technique for detecting ligand binding – extremely definitive
Structurally (atomically) resolved
Protein-observe or ligand-observe approaches
Kd can be determined for weak binding
Low sensitivity – high (labelled) protein consumption (protein-observed)

Question 29

What is the effect of binding on NMR spectra?

Answer 29

What we see depends on the exchange rate kex compared with the difference in chemical shift between free and bound states
We see an average peak between the two
The peak reduces in height (broad) as the height represents the concentration, but this represents the merging of the two peaks
Exchange rate is most influenced by the koff for binding

Fast exchange, intermediate exchange and slow exchange

Question 30

Describe fast exchange?

Answer 30

Kex > Δv
Observe single peak at the weighted average chemical shift
This peak moves position as we add more ligand
As kex is usually fast when binding is weak (fast koff), we normally see fast exchange with weak binding (Kd > 10-5 M)

Plot Kd (change in peak distance) v concentration  
Find the Kd between each of the 'peaks' shifts

Question 31

Describe slow exchange?

Answer 31

Kex < Δv
Observe separate peaks for bound and free
As we add ligand, free peak disappears while bound peak increases
As kex is usually slow when binding is tight (slow koff), we normally see slow exchange with tight binding (Kd < 10-7 M)

Plot height of bound peak v ligand concentration

Question 32

Describe intermediate exchange?

Answer 32

Kex = Δv
If the values of kex and Δv are similar (within 10-fold), we see a mixed picture
Peaks broaden and shift, then sharpen again as we add more ligand
Typically for Kd between 10-5 M and 10-7 M

Question 33

What is chemcial shift mapping of binding sites?

Answer 33

Chemical shift perturbations (CSPs) on ligand binding can be mapped onto a structure of the unbound protein to reveal the binding site
Need to assign NMR resonances first

Question 34

Describe protein-observed and ligand observed NMR?

Answer 34

The effects of binding can be seen in the NMR spectrum of both the ‘ligand’ and the ‘receptor’
So, in principle, we can use either to monitor binding

Protein observe: map to residues to identify binding site
Ligand observe: don’t need labelled proteins but no binding site info

Question 35

How are these chemical shift mapping of binding sites produced?

Answer 35

STD – saturation transfer difference (saturate signals from protein)

This transfers to any ligands that are in contact with the protein so do 1D and then again with saturation
Any peaks that reduce/disappear come from ligands that bind
Measure off – resonance one (no saturation and subtract STD one from 1D so only peaks that are left (difference) are from ones that aren’t in saturated spectrum and therefore bind
Can give info on what parts of ligand interact but not where on protein and also use a mixture of ligands at once

Question 36

How is this NMR technique applied in the pharmaceutical industry?

Answer 36

NMR has a role to play in screening weak binders, getting info on where they bind, fragments
Screening to lead characterisation (fragments/weak binders often too weak for other detection methods/xtal)
NMR can be used to screen a library of ‘fragments’ – very low-mw compounds that can be starting-points for rational drug design

Test library of fragments for weak binders
Identify those that interact with neighbouring regions on the protein, try and link them together to produce a more drug-like molecule

Question 37

Describe kinetics?

Answer 37

The energy barriers and the collision rate govern reaction kinetics
Energy barriers: reaction through a high energy transition state/activated complex
Difference in energy ground-transition = energy barrier that reaction has to overcome (activation energy)
Without a barrier then rate is limited by diffusion – how likely they are to collide in correct orientation so faster at higher temperatures

Question 38

Describe association and dissociation rates?

Answer 38

These rates are related to the equilibrium constant
Binding – reversible so in dynamic equilibrium between bound and unbound
Rates of binding and release different and related to equilibrium constant (Kd)
At equilibrium, concentrations don’t change but still binding and dissociation going on
Rate of a chemical reaction in solution is limited by diffusion

Breaking a very strong non-covalent bond
If the equilibrium is very skewed towards the bond state in lab purposes it is irreversible

Question 39

How can we measure binding kinetics by SPR?

Answer 39

Surface plasmon resonance e.g. Biacore

1. Have a gold chip with one binding partner immobilised
2. Flow other over the surface of it
3. Shine polarised light at the chip at interface of chip and gold
4. Produces charge density waves that reduces angle of reflected light at angle known as resonance angle that is proportional to mass on the chip
5. As binding and release occur, get change in mass so change in angle of reflected light – can measure this to follow binding and release

Interaction profile= sensorgram (measure as function of time)
As flowing binder over surface, get rise in mass – measure association then flow buffer and measure dissociation
Then regenerate by flowing buffer than strips binder off

Question 40

Give an example of using SPR?

Answer 40

Immobilise antibody, wash with antigen - binding curve
Wash with antibody that binds a different way – more binding curve, if the second antibody doesn’t bind, or competes with immobilised antibody – no second curve
Different antibodies with similar affinities (Kd) can have very different kinetics (same max but slower/faster on/off rates Kd=koff/kon
Therefore we can compare these too

Question 41

What are some methods for measuring immobilisation of a sensor chip?

Answer 41

1. Amine coupling - directed or random
   Problem – multiple sites so could be multiple orientations – could inactivate or occlude binding site
2. His tag labelled protein binds to nickel NTA on the sensor chip = single orientation and may be able to dissociate again with nickel
   Can’t use EDTA
3. Streptavidin-biotin - binds to streptavidin on the chip

Question 42

Describe the elements of the sensorgram?

Answer 42

First part of the curve increasing – association
While binder is flowing Steeper = faster rate, so fit this part of the curve to get Ka

Plateu = steady state - still flowing ligands over the chip
When the rate of association equals the rate of dissociation so no net change in concentration bound
Plotting height of the plateau as a function of concentration of the ligand produces a saturation curve and Kd

Decreasing curve
When stop flowing binder – dissociation but not new molecules so not as much binding (we have switched to buffer)
Fit this curve to determine koff – steeper = faster

Question 43

What are the effects of mass transport on SPR?

Answer 43

Sometimes the measured rate depends on the flow rate
Slows the apparent kinetics (i.e. is artefact)
If flow too slow or too much immobilised ligand then shortage of ligand binding so measured rate is measuring step 1 instead
Can be compensated by increasing flow rate or decreasing immobilization density

Question 44

What is regeneration in SPR?

Answer 44

Washing the sensor to remove ligand while keeping the immobilized receptor intact
Sometimes can wash off what was bound so you can re-use chip - other times would need to immobilise again to re-use as wash that off too

We can use: salts, acids, bases, detergents, denaturants (allows the chip to be reused)
Eg use acid Ph buffer – proteins partially unfolded and +ve charged so repel each other
Must not damage the ligand (immobilised) one though
Or use high salt or specific chemicals to break interaction (salt for ionic interactions

Important to use the mildest possible conditions that would cause dissociation so as not to irreversibly damage the ligand