Chemistry Reference
In-Depth Information
optimization of sample conditions that can minimize exchange on this timescale a
particularly important consideration for membrane proteins [
155
].
Relaxation optimization strategies have since been extended to side chain
aliphatic groups with the important recognition by Kay and coworkers that the
coherence transfer pathways available to a rapidly rotating, isolated
13
C
1
H
3
spin
system in a slowly tumbling macromolecule undergoes different rates of relaxation
[
288
,
289
]. Most significantly, they established that the HMQC is a TROSY type of
experiment for isolated
13
C
1
H
3
systems since, unlike the HSQC experiment, there is
no interconversion of slowly- and rapidly-relaxing coherences. The benefits of this
effect are optimal for
2
H,
12
C-labeled proteins with specific incorporation of a
single
13
C
1
H
3
group in a subset of amino acid types (e.g., ILV [
290
]). Deuterium
spin relaxation measurements in
13
CHD
2
-labeled methyl groups can also take
advantage of this relaxation optimization strategy [
291
], opening the door to the
study of dynamic structural states for extremely large protein complexes, as
demonstrated for the 670-kDa 20S proteasome core particle [
271
].
Although unassigned methyl group resonances can be used as reporters of
structural transitions in large proteins [
257
,
292
], any requirement for site-specific
information requires that these methyl peaks be assigned. This is straightforward
when backbone assignments are available, since TROSY-type sequences can be
used to correlate methyl proton or carbon shifts to amide
1
H and
15
N chemical
shifts [
293
-
295
]. Alternatively, in some cases it has been possible to transfer
assignments made on smaller fragments or isolated subunits to these same
components in the intact complex [
257
,
266
,
295
]. Meanwhile, efforts are also
being made to automate the process of methyl shift assignment, with one protocol
using X-ray structure-based chemical shift predictions along with methyl-NOESY
data to correctly assign 99% of a 300-kDa ILV-labeled proteasome [
296
]. This
approach may become important for large single-chain membrane protein-
detergent complexes, particularly since these samples would not easily be adapted
to the fragment-based approach, and
1
H-
15
N spectral quality may not be sufficient
for through-bond correlations with methyl groups. However, in the event that
backbone assignments are not available or methyl correlation spectra do not
allow unambiguous assignment, methyl shift assignments can still be made by
using spectra of mutants [
266
], sometimes paired with NOEs, or paramagnetic
relaxation enhancement patterns [
297
].
5 Strategies for Membrane Protein Structure Determination
by Solution NMR
5.1 NOE-Based Methods of Structure Determination
The approach used for structure determination of a membrane protein varies
significantly depending on the size of the complex, the protein fold, and the quality
