low energy spin state

1
-1/2
E
1/2
low energy spin state oriented with (parallel to)
magnetic field
2
Two Dimensional NMR
90o x
1D pulse and acquire
FT (t1)
t1
relaxation delay
90o x 90o x
t2
t1 delay
FT (t1,t2)
In a 2D experiment FIDs are acquired for (perhaps
512) incremented values of the t1 delay (to
generate a second frequency dimension). The data
are Fourier transformed with respect to both t1
and t2.
General rule all 2D NMR spectra contain the 1D
spectrum along the diagonal. Off-diagonal signals
indicate a correlation (coupling or NOE depending
on the type of spectrum) between 2 protons.
3
For example The COSY spectrum below illustrates
sequential spin coupling connectivities between
protons within an amino acid. In the TOCSY
spectrum each NH group is labelled with the
frequencies of all the other protons in its amino
acid side chain. COSY and TOCSY are useful for
spectral assignment. In a NOESY spectrum, off
diagonal signals correlate positions of protons
that are within 5 angstroms. Very useful for
structure determination. Note that COSY is the
2-dimensional analogue of spin decoupling
NOESY is the 2-dimensional analogue of
the difference NOE experiment
COSY TOCSY
4
NOESY off
diagonal crosspeaks correlate the chemical
shifts of pairs of protons that are close in
space (lt 5 angstroms). semi-quantitative
constraints on separation weak 1.5 - 5
Å medium 1.5 - 3.5 Å strong 1.5 - 2.5
Å
0
ppm
2
4
6
8
red lines highlight a weak NOE between protons at
8.3 and 3.7 ppm. Wherever these protons are in
the protein sequence, they must be close in space
(lt 5 Å) in the folded protein.
10
10 8
6 4
2 ppm
backbone NH aromatic
CHa CH b,g,d,e
methyl
5
The fingerprint region of a 2D 1H,1H spectrum
contains coupling (COSY) or spatial (NOESY)
correlations between backbone NH signals and
backbone CHa signals. The fingerprint region of
COSY and NOESY are used together for sequental
assignment of the NMR spectrum.
CHa
NH
In a COSY spectrum the fingerprint region
contains one cross peak for each amino acid (I.e.
intramolecular NH-CHa signal), except for proline
(no NH) and glycine (2 CHa i.e. 2 cross peaks
for gly).
NH signals here CHa signals
6
sequential assignment
COSY
CHa a
CHa b
through space correlation (NOESY) through bond
correlation (COSY)
NOESY
CHa a
CHa b
for any backbone torsion angles found in a
protein (f, ?), either the distance dNiNi-1 or
the distance dNiai-1 is short enough (lt3.5 Å) to
give an NOE effect.
NHa NHb
7
Ala1-Gly2-Ser3-Leu4-Gln5-Asp6-
8
must by Gly (only amino acid with 2 CHa)
COSY
NOESY
Gly CHas here
Ala CHa here
not in COSY-probably Gly NH-Ala CHa NOE
Gly NH here
also in COSY (therefore intraresidue NHCHa NOE)
i.e. Ala CHa - Ala NH
Ala1-Gly2-Ser3-Leu4-Gln5-Asp6-
L4
S3
Q5
G2
D6
A1 CHa
G2 L4 A1 Q5 S3
D6 NH
9
Structure calculation
Use computer energy calculation (commonly
simulated annealing using Xplor) to determine
structures that satisfy the NMR
constraints NMR-derived structural
constraints NOEs (by far the most important)
these are semi-quantitated e.g. weak NOE -
constraint 1.5 - 5 Å medium -
constraint 1.5 - 3.5 Å strong -
constraint 1.5 - 2.5 Å backbone dihedral angles
(phi) from NHCHa coupling constant hydrogen
bonds provide distance constraints on N-H--OC)
from hydrogen exchange measurements (novel
constraints orientational information from
residual dipolar couplings in partially aligned
samples) Nature of NMR structure Since no
single structure is defined by the full
constraint set, a family of structures is
calculated. These are usually presented as
overlays of backbone traces to give a visual
impression of the precision of the structure.
Would generally like 10 NOE constraints per
residue. NMR structures are usually well-defined
in the backbone (rmsd around 0.2-0.4 angstroms)
and less well defined in the side chains (0.5-0.8
angstrom rmsd). Investigation of ill-defined
regions can be made using 15N relaxation
measurements to distinguish between floppy
structure and those lacking sufficient
constraints. The more constraints, the better
(more precise and hopefully more accurate) the
structure.
10
Three (and more)-dimensional NMR
15N- and 15N,13C-labelling allows structure
determination of larger proteins (by simplifying
increasing complex spectra, by alleviating
inherent limitations due to slower tumbling of
proteins in solution, and by providing improved
assignment strategies). Structure determination
still relies on identification of NOEs.
15N NOESY HSQC 3D spectra in which H-H NOESY
information is resolved into separate planes in
the 3d dimension according to the frequency of
the amide 15N. Likewise for 15N TOCSY HSQC.
CH
CH
2
2
O
O
N C C N C
C N
H
H
H
H
H
HNCA 3D triple resonance spectrum in which cross
peaks in 3D correlate the frequency of 15N, its
directly attached 1H and the 13Ca of both its own
and preceding amino acid. Backbone can be
assigned directly by walking along the peptide
backbone. And there are many other examples (e.g.
see Dr. Schirras web site)
11
NMR in the Post Genomic Era (Structural
Proteomics) 30,000 genes in human genome (35
- 40 membrane proteins) perhaps 3,000-10,000
different protein folds (identification of all
folds plus homology modelling should allow
fitting a structure to all gene products even
without structure determination requires
increasing efficiency/throughput of structure
determination (even low resolution determination
of protein fold is valuable) improved
structural determination of high MW proteins and
membrane proteins actual/potential sources of
improvement 15N/13C labelling
expression of isolated domains (most protein
domains 50-150 amino acids) see Campbell
Downing paper on background reading list)
total or partial deuteration of proteins
(reduce line broadening from 1H-1H dipolar
interactions) dipolar couplings (partial
orientation of protein in solution determine
15N-1H and 13C-1H bond orientations)
accurate alignment of 2iary structure elements in
poorly constrained structures TROSY novel
pulse sequences that dramatically reduce line
broadening at very high fields
(see Riek et al. paper on reading list)
12
M. thermoautotrophicum Escherichia
coli Sacchromyces cerevisiae Thermotoga
maritima myxoma virus proteins from 513 ORF
having predicted MW lt 23kD expressed as
His-tagged fusions in E. coli in 15N media.
good promising mostly unfolded poor
loss of protein upon concentration
13
(No Transcript)