AMBER Archive (2006)

Subject: Re: AMBER: AMBER minimization vs. Gaussian optimization

From: Thomas Cheatham (tec3_at_utah.edu)
Date: Wed Oct 04 2006 - 15:45:59 CDT

> I have a question regarding the sugar pucker of the nucleic acid residues.
> For watson-crick bases (nucleosides), the gaussian optimized structures
> are in 3'-endo conformation while the AMBER minimized structures are in
> 2'-endo conformation. Is the AMBER force field parameters choosen such
> that the nucleic acid residues are forced to have 2'-endo character?

I think that minima should be present at both C3'-endo and C2'-endo
(although they may be shifted a bit) for both AMBER and Gaussian and will
be different for ff94, ff98 and ff99 in AMBER.

What local minima is found will depend on the starting structure (and
environment). In practice, with ff94 in DNA, C3'-endo is only a transient
state (not clear if it is a minima during solvated dynamics) whereas with
ff99, distinct C3'-endo and C2'-endo populations are found. With RNA,
C3'-endo and C2'-endo puckers are minima.

So, I do not understand that gaussian finds 3'-endo and "AMBER" 2'-endo
except that minimization from a given structure can go anyway. One way to
force the pucker state is to restrain delta during the initial
minimization (as per what I did a long time ago in "How sensitive are
the results ..." paper in the proceedings of the 10th converstation paper at:
http://www.chpc.utah.edu/~cheatham/protected/publications/how_sensitive_jbsd.pdf
(username: steroid, password: steroid)

There I restrained C1'-C2'-C3'-C4' to between 30-40 deg for C3'-endo and
to between -38 to -40 for C2'-endo.

> For my modified nucleic acid residues, when I minimize them using AMBER
> force field, I am not getting the 2'-endo character. I am getting

I think you will have to force it to repucker during the minimization; try
using the restaints as outlined above initially...

--tom
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu