7. Structural methods Flashcards
Describe the nuclear magnetic resonance phenomenon using the simplest example of just water
Now place ethanol (H3C-H2C-OH) in this field. What do you notice?
Which isotopes can be measured in NMR?
Place a sample of water in a strong magnetic field, measure the absorbance of different radiofrequencies Intensity and resonance frequency is proportional to magnetic field (eg 100Hz at 2.3T, 200Hz at 4.6T, etc)
With EtOH, you see 3 distinct peaks with relative intensities of 1, 2 and 3 (representative of the 3 different hydrogen environments with 1 2 and 3 Hs in each)
Lots of complicated math involving nuclear spins, but for this course just know that isotopes with an odd number of nucleons are most useful for NMR (1H, 19F, etc)
What is shielding and what is a result of this? What atomic property influences this?
How do aromatics/pi electrons affect this?
A probe (TMS usually) is used as a reference frequency as it does not vary. TMS creates characteristic shifts based on chemical environment, which is independent of magnetiv field
Shielding is the effect of electrons countering the reference frequency that results in different proton enfironments having different chemical shifts (positions on the X axis)
The more electronegative an atom, the less shielded (the more deshielded) it is, and the farther left (or “downstream”)
Downfield = deshield = e-ve
(Take away electrons, decrease shielding)
the pi electrons in an aromatic ring creates deshielded regions in the plane of the ring and shielded regions above/below the ring
How many different proton environments are there for Zwitterionic L-phenylalanine?
7-9 depending on how freely the Cb-Cg bond can rotate
Define the phenomenon of exhange broadening
If a molecule exists in two states (eg cis/trans, bound/unbound, etc) that create different chemical environments, it will exchange between those states at a rate kex that is defined by k1 + k2 (rates going from A to B and B to A)*
* If a ligand is binding an enzyme the kex is defined by kon[Lfree] + koff
If the exchange rate is very slow, the NMR spectrum will show two distinct peaks with intensities reflective of the ratio of the two states
If the exchange rate is very fast, they will converge into one peak at a frequency that is the average of the two states
Exchange broadening occurs in the case of an intermediate exchange rate
How does a 2D NMR work and what are its benefits? (in the simplified context of this course)
15N shows a crosspeak when it is bound to at least 1 Hydrogen, allowing
How does saturation transfer difference NMR work to detect ligand binding?
If done with varying concentrations of ligand, what can be determined via STD-NMR?
Selectively irradiating the receptor in solution with potential ligand 1 and 2
Irradiation causes ‘saturation’ and disappearance of NMR peak representing receptor
If eg ligand 1 binds and not 2, the ‘relaxation’ of the receptor will saturate ligand 2
Take the difference between spectra of before and after irradiation + saturation: Peaks that disappear are unaffected and don’t bind (ligand 2) peaks that show a difference became saturated (ligand 1)
Can fit binding isotherm equation to determine Kd with enough concentrations/data points
Picture: note that (D) has been scaled up to emphasise difference
How can you grow a protein crystal for X-ray diffraction, in general?
Describe the vapor diffusion method for growing a crystal
No one recipe, every protein is different
Need supersaturaion (salting out), which can be determined by a solubility phase diagram
Nucleation must occur (at supersaturation levels) before growth can begin
Vapor diffusion method:
Dilute drop of protein dissolved in volatile component (H2O) in a resevoir with a concentrated “mother liquor” containing precipitant. Eventually the volatile component will evaporate to equilibriate both solutions, leading to an increase of protein concentration in the drop, until nucleation occurs. As the protein grows, its concnetration in the drop will slowly decrease until the ‘saturation boundary’
What are trends that can influence protein crystalization?
- Flexible loops prevent crystalization
- Acidic proteins are more likely to form crystals than basic proteins
- Smaller proteins are more likely to crystallize than large proteins
- Ligands that bind and stabilize proteins can favor crystalization
- Monodisperse solutions of ptroteins are more likely to yield crystals
- Higher purity increases likelihood of crystalization
- Lysines are unvavored for crystal contacts (mutate or derivatize to reduce entropy)
- Try different orthologs for a given protein
What is the difference between crystalographic and non-crystalographic symmetry?
Describe unit cell and asymetric unit
Crystallographic symmetry: When two identical copies are related by a rotational symmetry that can be applied to the entire crystal
Unit cell: The repeated unit that makes a crystal by translation
Asymmetric Unit: The smallest unit that cannot be further decomposed by cyrstallographic symmetry
What is the most important variable in the complicated fourier transform equation for the determination of a protein crystal structure? What helps determine this variable?
PHASE
Molecular replacement: Requires a model with homologous structure, use its knwon phase to approximate phase of your crystal (have to be ~40% homologous)
Anomalous scattering: Crystalize protein with “anomalous scatterers” (heavy atoms) to create characteristic patterns in diffraction which can be traced to find location of atoms, which can then be traced to estimate phases
What is an electron density map?
How can one go from an EDM to a model of a protein?
What measures can be taken to improve accuracy of modelled protein?
How can ligands be crysallized with proteins and how can we make sure of this during diffraction?
Map of probable protein location. High enough resolution can show characteristic amino acid shapes and water molecules
Start by tracing backbone and secondary structures, add side chains later
The R-factor can increase accuracy:
Fobserved - Fcalculated = R (simplified)
Lower R = more accurate
Can be overfitted, keep Rfree (small subset of reflections not used for refinement) and compare to Rwork after
Can either cocrystallize or soak the crystals in ligand
Fo-Fc
If observed but not calculated (positive), it should be there
If calculated but not observed (negative) it shouldn’t be there