NMR of Membrane Proteins - Recent Developments in Biomolecular NMR - page 303

light-activated state of the GPCR rhodopsin. 292 Weak alignment in the portion
of the peptide that bound rhodopsin resulted from the magnetic susceptibility
of the rod membranes. The RDC restraints were used to support previously
reported trNOE data, 293 which was convoluted due to intermolecular cross-
relaxation between the peptide (ligand) and the protein. The study revealed
formation of an a-helical element in the C-terminus of G t a, a region which was
not ordered in the crystal structure, 294 demonstrating the complementary
information available through the use of the different techniques.
In cases where receptor resonances have been assigned, or HSQC-type
spectra can be recorded, it is possible to investigate interactions through
chemical shift perturbations. In the case of an assigned protein, movementof
known HSQC peaks allows location of the binding epitope. This forms the
basis of the SAR by NMR technique, whereby binding of small chemical
groups to adjacent sites on a target protein are identified through chemical
shift changes followed by lead optimisation 286,295 and used in other drug-
development studies. 296 In many cases assignment of the target protein may be
limiting, and other techniques such as Inter-ligand NOE (ILOE) 297 and
INPHARMA 298 (ILOE for Pharmacore mapping) require no prior protein
assignments. ILOE looks at ligands that bind at adjacent sites enabling lead
optimisation as for the SAR technique, whilst INPHARMA detects ligands
which compete for the same site. 286
Interesting structural information can also be provided by selective-labelling
schemes. For example, Bokoch et al. 299 use reductive methylation to label
lysine side-chains in the b 2 -adrenergic receptor with 13 C-labelled methyl
groups, enabling studies of the environment around the Lys305-Asp192 salt
bridge between ECL3-TM7 and ECL2 and the associated rearrangements on
activation. HSQC and STD-filtered-HMQC spectra were recorded demon-
strating an altered conformation of Lys305 on inverse-agonist (carazol)
binding, but no change on binding of a neutral agonist (alprenolol), whilst
agonists (formaterol) are seen to induce a different conformational change to
inverse agonists. Data with full agonists demonstrate attenuation of the
Lys305 resonance suggesting weakening of the Lys305-Asp192 salt bridge in
the active conformation of the b 2 AR. Combined with computational
modelling, the authors propose that the extracellular ends of TM6 and TM7
move on activation, whilst inverse agonists are expected to block TM6
motion. 299 This study demonstrates the valuable structural information that
can be gained from selective-labelling studies without the need for full
assignment, in particular in the region of the important extracellular loop
regions of GPCRs which may be more difficult to observe in crystal structures.
12.8 Conclusions
Recent developments have demonstrated that solution NMR is a viable
technique for the structural study of membrane proteins. A range of structures
has been solved, which indicate that many limitations have been overcome;