AMBER Archive (2003)

Subject: force field problem

From: Bayly, Christopher (christopher_bayly_at_merck.com)
Date: Mon Feb 10 2003 - 14:01:37 CST

I think an all_amino*.in library needs to be generated for parm99.dat, and
the X-CA-C-X torsion parameters need to be carefully examined and probably
changed or removed in all force fields including and since parm94.dat

The X-CA-C-X torsion parameters are causing a big problem for me. Since
parm94, this has been set to a high barrier (14.5 kcal/mol 2-fold
"interpolation based on C6H6") because it described the phenolic C-OH in
tyrosine. This barrier is way too high for regular small-molecule contexts
like a benzamide fragment (phenyl-C(=O)-NH-R). I need to do MD on systems
where the Xray structure clearly shows non-planar torsions in this key
benzamide fragment; amber breaks two H-bonds to force it planar. If I just
change this torsion in parm94.dat, all tyrosines will presumably become too
floppy around their phenolic fragment.

Lo and behold, the parm99 force field has partly fixed the problem by
re-paramaterizing assuming the tyrosine CZ to be a CA atom type. However,
the leap libraries all_amino*.in only incorporate this change into tyrosine
(changing CZ from type "C" into type "CA") in all_amino02.lib, which also
has different charges etc. throughout (for polarization I believe).
However, it still retains the X-CA-C-X torsion at 14.5 kcal/mol, so I would
also have to modify it for benzamide-type fragments, or benzoic esters, or
benzoyl-anythings... that's a lot of small molecules. gaff.dat also retains
X-CA-C-X torsion at 14.5 kcal/mol, presumably as a legacy number.

Suggestions:
1) To correctly describe X-CA-C-X torsions in small molecules, the default
X-CA-C-X torsion at 14.5 kcal/mol needs to be changed in force fields dating
from parm94. In parm99 and later ffs, X-CA-C-X torsion at 14.5 kcal/mol
should be viewed as debris and removed, or better yet, parameterized well.
2) Once changed, pre-parm99 amino acid libraries will have compromised tyr
residues at CZ, the phenolic carbon.
3) Since parm99 is the only two-body additive amber ff which has properly
typed TYR CZ as atom type CA, leap lib files all_amino99.in,
all_aminoct99.in, and all_aminont99.in should be generated and made
available.

Having been out-of-touch for a while, is anyone(s) taking care of continuing
ff development for the AMBER force field?

Christopher Bayly

------------------------------------------------------------------------------
Notice: This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (Whitehouse Station, New Jersey, USA) that may be confidential, proprietary copyrighted and/or legally privileged, and is intended solely for the use of the individual or entity named on this message. If you are not the intended recipient, and have received this message in error, please immediately return this by e-mail and then delete it.

==============================================================================