How accurately can interproton distances in macromolecules really be determined by full relaxation matrix analysis of nuclear overhauser enhancement data

Content Provider	Semantic Scholar
Author	Clore, G. Marius Gronenborn, Angela M.
Copyright Year	1989
Abstract	In recent years considerable success has been achieved in determining three-dimensional structures of macromolecules in solution on the basis of approximate interproton distance restraints derived from nuclear Overhauser enhancement measurements (l-4). Improvements in the precision of such structure determinations can potentially be obtained by increasing either the number or the accuracy ofthe experimental restraints. To this end, several groups have been seeking to obtain more accurate interproton distances using full relaxation matrix analysis of the NOE data in order to account fully for mu ltispin effects and overcome the errors arising from the application of a simple two-spin approximation (5-14). The strategies generally proposed involve an iterative approach. A trial structure is first computed on the basis of a set of initial approximate interproton distance restraints using one of the available structure determination methods such as metric matrix distance geometry ( 15-I 7)) m inimization in torsion angle space ( 18, 19), restrained mo lecular dynamics (2022)) or dynamical simulated anneal ing (23-25). Based on this trial structure, a theoretical two-dimensional NOE (NOESY) spectrum is computed and compared to the experimental one. Adjustment of the interproton distance restraints and further structure refinement yield a new trial structure. The entire process is repeated until the experimental and theoretical NOESY spectra match up (26). Alternatively the matrix of theoretical NOESY cross-peak and diagonal-peak intensities derived from the trial structure can be merged with the experimental one to calculate a new set of interproton distances with which to refine the structure and produce a new trial structure; the process is again repeated in an iterative manner until no change in the interproton distances is observed from one cycle to the next ( 9-Z I ) . Because structure refinement is computationally intensive, it has generally been considered that one only needs to carry this process out for a single starting structure. The interproton distances derived in this manner are then assumed to accurately represent the true interproton distances and to provide a reliable data set with which to compute an ensemble of structures in order to obtain a measure of the conformational space consistent with the experimental NOE data (8) _ The underlying assumption in such an approach is that the interproton distances, or more specifically the cross-relaxation
Starting Page	398
Ending Page	409
Page Count	12
File Format	PDF HTM / HTML
DOI	10.1016/0022-2364(89)90388-0
Volume Number	84
Alternate Webpage(s)	https://spin.niddk.nih.gov/clore/Pub/pdf/147.pdf
Alternate Webpage(s)	https://doi.org/10.1016/0022-2364%2889%2990388-0
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article