I'm gearing up for a project in which we're looking at the structure of a short peptide that should fold upon binding divalent cations. What software should I use? Does CNS handle this well, and how does one incorporate metals into the picture?
asked Mar 16 '10 at 14:36
Hi Josh, your question can be split into two parts: which software can be used, and how to calculate structures of small(er) molecules that can aggregate.
Long answer, but the bottom line is that the choice is probably between NIH-XPLOR and CYANA, but then there may still be issues specific to this kind of a task.
1) software packages
I think the dominant core packages for the NMR structure calculations are NIH-XPLOR and CYANA (which I have not had a chance to use yet). In my understanding NIH-XPLOR and CNS had at some point split from the common ancestor and CNS is now more used by X-Ray crystallographers.
CYANA is a successor of DYANA, now authored by Peter Guentert and coworkers. Cyana is available commercially, NIH-XPLOR is free for academia.
There is also UNIO, created by Torsten Herrmann - UNIO is a software that enables NIH-XPLOR and I beleive DYANA to automatically assign NOE's and more. UNIO is a successor of Atnos/Candid. The particularly nice feature of UNIO (IMO) here is that it can analyze raw NMR data, not hand-picked peak lists. In the Clubb lab (when I've been a postdoc there) we've been successful at calculating a number of protein structures with Atnos/Candid. I highly recommend this tool for the solution of macromolecular structures by NMR.
UNIO claims to be able to automatically generate structures based on raw unassigned data, it probably does have high requirements on the quality of the data, though.
Even though UNIO and CYANA can automatically assign NOEs and generate structures, for a small peptide this will not be necessary - as manual assignments will probably be more reliable and unlike in a bigger protein correctness of assignments may be more important.
2) your particular application
The problem is that small peptides tend to have less defined structure in a way than say proteins. Chances are that all side-chains will not have a prevalent conformation because there is no packing of side-chains. It's very possible that the backbone will not be very ordered too.
The issue here is that NOE's that you will see may contradict one another and there will be no single structure that will satisfy all restraints at once. For that there must be a way to apply restraints to an ensemble instead of every single structure - and how to do it - I have no idea. Perhaps somebody else could suggest something...
Here is a great post by Sekhar Talluti in response to a related question .
Another challenge is that you might not know the degree of aggregation, but here you can use some hypothesis and build a model anyway.
A slight potential complication arises when the molecule has non-standard moyeties - in that case you'll have to dig deeper in order to construct the model (it's the easiest to construct the model from pre-defined standard pieces) - it is definitely possible with NIH-XPLOR, maybe with CYANA too.