Title: NMR and Proteins
1NMR and Proteins
-Lecture 3
2NMR is not a competing technique with X-ray
crystallography by a complementary one.
There are reasons one might use each technique.
3Comparison of NMR and X-ray crystallography (of
proteins)
- Similarities
- NMR and X-ray techniques require pure protein
- protein must be folded
- protein must be in the correct fold (i.e. natural
fold) - -purification techniques that unfold the protein
and then refold it may be undesirable
- Both require relatively minimal instrument
experimental time (1-2 days) - But extensive computational and analysis time
(1-6 months)
4Differences (experimental)1 NMR requires i)
soluble protein, ii) low salt buffer, iii) no
additives (especially organic compounds) iv)
concentration 5-10mg/ml, in 600 ul (i.e. 5 mg
of protein) v) need to use labelled protein,
(i.e. radio labelled -C13, N15) vi) due to v),
protein must express in the C13 / N15 labelled
minimal media vii) protein must be stable under
experimental conditions -usually 30ºC for 24
hours
5Differences (experimental)2 c.f. X-ray
requires i) A protein crystal, ii) Crystal
must diffract, iii) concentration used to
generate crystal 1-100mg/ml, in 2 ul (i.e. ug
of protein) iv) to find crystallization
conditions may require gt100mg v) For completely
new structure (with new fold), may need selenium
methionine vi) Due to v), protein must express
in the selenium methionine labelled minimal
media vii) Crystal must be stable under
experimental conditions (i.e. must have
freezing conditions)
6Example 1D NMR spectra
36 a.a protein
one protein side chain
-Average protein solved by NMR is 100-200 residues
7(No Transcript)
8What do we need from the NMR experiment
- Sequence data tells you what the primary
structure is. - To get secondary and tertiary structure need
information on distances between atoms - To get this information you need to perform 2D,
3D, 4D NMR - 1D NMR involves data from one atom e.g. an H1 or
C13 - 2D/3D/4D NMR requires data from more than one
atom at the same time - e.g. H1-C13 (2D), or H1-C13..H1 (3D),
H1-C13..H1-N15 (4D)
9Example 2D spectra HH
From this data we need to solve two problems. i)
Which peak corresponds to what atom in what
amino acid ii) How does each atom relate to the
other atoms
10Idiots guide to protein NMR terminology
COSY- Correlation spectroscopy Gives
experimental details of interaction between
hydrogens connected via a covalent
bond NOESY- Nuclear Overhauser effect
spectroscopy Gives peaks between pairs of
hydrogen atoms near in space (1.5-5 Ã…) (and not
necessarily sequence)
11So I have a NMR spectra, what do I do next The
COSY data tells you which peaks on the spectra
belong to which amino acid, the NOE spectra tells
you what other atoms that atom is near
Asn
Gly
NOE indicated the asparagine amino-hydrogen is
near a glutamate acidic hydrogen
Identified as an asparagine amino-hydrogen from
COSY spectra
Glu
12From the amino acid sequence of the protein you
can calculate the which amino acid you have
identified in the COSY and NOE spectra
e.g. your COSY spectra identifies an ALA, the NOE
data identifies it as being close to a GLN on one
side and a ASP on the other. That ASP intern has
the ALA on one side and a HIS on the
other. Therefore your sequence is
GLN-ALA-ASP-HIS or HIS-ASP-ALA-GLN You find
in your amino acid sequence that you have
HIS-ASP-ALA-GLN 34
35 36 37 Therefore the unknown ALA is
ALA 36. This procedure is repeated for the
whole protein
13What next?
- From NOE I know close atom-atom distances, but
that doesnt give a structure - The information you have up to this stage is a
list of distance constraints - The structure can be determined by inputting this
information to computer minimization software. - The computer program also contains information
about amino acids, bond lengths/angles and
standard information about atom-atom interactions
such as minimum distance (i.e. van der waals
radii) - With all this information you can generate a
model of the structure. - Important NMR gives you a number of possible
solutions - (all almost identical, rmsd lt1Ã…), This can range
from 5-20 models - X-ray crystallography give one average structure
14Example of an NMR structure, showing 10 models
15Over simplification
The previous method applies to small lt10kDa
proteins For larger proteins additional
techniques are required HSQC, TOSEY Details of
these techniques are outside the scope of this
course
16How well do X-ray/NMR agree
- Not worth the time to get a structure that is
already been solved - Most proteins that have been solved by both
methods give almost identical structures (e.g.
bovine pancreatic trypsin inhibitor, plastocyanin
and e.coli thioredoxin) - There are a few example of significant
differences between structures - (e.g. glucagon (a peptide hormone) and
metallothionein (binds seven cadmium ions)) - The reasons are usually due to conformational
variability in the different experimental
conditions the protein was exposed to, (e.g. pH,
temp, salt concentration) e.g. glucagon. - In the case of metallothionein, the crystal
structure had been wrongly determined (pre
modern computing techniques), the data was
re-examined it agreed with the NMR data - Important The fact that NMR and X-ray structures
agree is an important point in justifying the use
of crystals in structural determination
17Limitations of NMR
- Protein must be soluble at 5mg/ml in a simple low
salt, no additive buffer - -many proteins require complex buffers for
solubility - Protein must be stable at 30ºC for 24hrs
- -some proteins degrade at prolonged exposure to
physiological temperatures - Protein must be lt30kDa (current advances are
pushing the upper limit to 40kDa) - -most proteins are larger
- Reason for size limitation is physical not
technological- to deconvolute signals from larger
proteins need larger magnet. Current magnets
800-900MHz. - -larger, more powerful, machines gt1000MHz. This
is in the microwave radiation range, therefore
extreme heating of sample is a problem
18Size of NMR machine necessary to solve a 100 kDa
protein
19Advantages of NMR
- NMR gives dynamic structure not static
- -can observe movement in flexible regions or
substrate induced changes - Dont need crystals
- Generally quicker for small proteins lt10kDa
- NMR can give dynamic information on proteins
bigger than 40kDa
20NMR as a tool for dynamic analysis of a protein
Case study on Cytochrome P-450BM-3
- The protein is 120 kDa, the domain of interest is
45kDa - -cannot solve structure by NMR
- Protein structure solved by X-ray crystallography
- -no substrate bound structure available
Dilemma The substrate did not appear to fit into
the enzyme Solution Use NMR to examine
substrate-protein interaction
21How can NMR be used on a 45kDa protein
- P-450s contain a heme group in the catalytic
center - -heme contains iron (Fe (III))
- Fe(III) is paramagnetic,
- -therefore effects the magnetic properties of
the surrounding atoms - When the substrate enters the binding site it
changes the paramagnetic properties of the iron,
which changes the magnetic properties of the
surrounding amino acids - NMR measures magnetic properties, therefore the
changes induced by the substrate entry can be
measured by NMR - From these changes, distance constraints for the
surrounding atoms can be established
22What did the experiment show?
Phe
6Ã…
Substrate
Crystal Structure
Information from NMR
23- Experiment gave information not available by
X-ray crystallography - Example of how NMR and X-ray are complimentary
techniques - Shows the main advantage of NMR
- - can observe dynamics (movement) of protein in
solution