i like this post (click again to cancel)
i dont like this post (click again to cancel) remove favorite mark from this question (click again to restore mark)

I have collected over 370 NOE restraints for a 38 AA peptide. Because I have less than 10 restraints per residue, I am hesitant to eliminate too many because this will lower the eventual RMSD. I am using XPLOR-NIH to run SA on the peptide. However, I keep coming up with very high energy structures (sort energy ~ 1000). Even though the program generates violations, eliminating these often does not result in a lower energy or more ordered structure. Any suggestions would be much appreciated, thanks.

asked May 11 '10 at 11:03

E%20V%20Clarke's gravatar image

E V Clarke

updated May 11 '10 at 13:21

Evgeny%20Fadeev's gravatar image

Evgeny Fadeev

Hi, there are some similar questions with some information on the subject - take a look at the "Related questions" box on the right. Cheers. - Evgeny Fadeev (May 11 '10 at 13:24)

One Answer:
i like this answer (click again to cancel)
i dont like this answer (click again to cancel)

The number of restraints per residue is not as important as the distribution of restraints. The overall quality of a structure, in a case like this, depends critically on the distribution of NOe's among the following catergories: long range, medium range and short range (in terms of residues). If you have a sufficient number of long range NOe's, then a reasonably good model can be built from the long range NOe's only, using restrained energy minimization (or restrained molecular dynamics/SA) with a good force field. Remember that, in such a case a complete force field such as AMBER, CHARMM, etc has to be used, rather than the minimal force field that we use normally for NMR structure determination.

You could try to remove ALL intra-residue and sequential NOe's from your data and see if there are any violations when you generate a structure. If there are no violations using only long range data then you MAY be able to get away with a recalibration. However, if you observe violations, even with the long range restraints data set, then you have no option other than the use of ensemble based optimization methods.

Violations may be due to errors in assignment, due to calibration or due to multiple conformations sampled by the peptide. If you have multiple global conformations, then ensemble optimization is unavoidable. However, if a single major conformation exists, but local motion is significant then the problem is with your calibration. This is because, in the presence of motion, each class of NOe's has to be calibrated separately - in some dynamical models the distance dependence of the NOe's is no longer inversely proportional to r^6. Models ranging from 1/(r^3) ... to 1/(r^6) have been proposed for different types of models.

To assess the extent of conformational flexibility, it would be highly advisable to obtain additional data regarding J-coupling constants, H/D exchange rates, relaxation rates, etc.

If there is insufficient evidence to guarantee that you are dealing with a single dominant global conformation, then an ensemble based optimization method such as the approach described below should be used:

Maximum entropy approach to the determination of solution conformation of flexible polypeptides by global conformational analysis and NMR spectroscopy – Application to DNS1-c-[d-A2bu2, Trp4,Leu5]- enkephalin and DNS1-c-[d-A2bu2, Trp4, d-Leu5]enkephalin. M. Groth, J. Malicka, C. Czaplewski, S. Olstrokdziej, L. Lstrokankiewicz, W. Wiczk and A. Liwo. Journal of Biomolecular NMR. Volume 15, Number 4 / December, 1999


answered May 23 '10 at 05:51

sekhar%20Talluri's gravatar image

sekhar Talluri

updated May 23 '10 at 12:18

Thank you, both responses are helpful. I have considerable variation in the global conformations, so I'll need to keep at the ensemble optimization. I'll take a look at the reference you suggested. - E V Clarke (May 26 '10 at 11:03)

Your answer
Please start posting your answer anonymously - your answer will be saved within the current session and published after you log in or create a new account. Please try to give a good answer, for discussions, please use comments and please do remember to vote (login to vote)
toggle preview

powered by CNPROG