Yes, you are right. I think it was overlooked.
The atom types were incorrectly setup for LYN in parm94 and parm03 and
parm10. It is correctly setup in parm02 and parm02EP, and parm11 (not
released yet).
The atom types should be: -CT(H1,H1)-NT-(H,H) in LYN
and CT(HP,HP)-N3(H,H,H) for LYS.
Piotr
> Hi all,
>
> Lately I've been doing some small molecule work (including several
> side chain analogues) and noticed something odd about the atom types
> of the LYN (neutral lysine) side chain. In both AMBER ff9x/ff03 and
> GAFF, there are different atom types for:
>
> (1) protons bonded to a sp3 carbon bonded to a positively charged
> group (i.e., HP or hx) vs. protons bonded to a sp3 carbon bonded to a
> neutral, but electron-withdrawing group (i.e., H1 or h1)
> (2) a nitrogen atom in a positively charged amine group (i.e., N3 or
> n4) vs. a nitrogen atom in a neutral amine group (i.e., N or n)
>
> And indeed if one uses antechamber to assign atom types to methylamine
> and methylammonium, for example, the assigned GAFF atom types are
> different. For methylamine, the atom types are c3, n3, h1, and hn;
> for methylammonium, the atom types are c3, n4, hx, hn. (The same also
> occurs for n-butylamine and n-butylammonium, which are more directly
> analogous to the neutral and positively charged lysine side chains,
> respectively.)
>
> If one examines all_amino94.lib or all_amino03.lib, however, the atom
> types for the last two "groups" (i.e., CE+HE* and NZ+HZ*) of both the
> LYN and LYS side chains are CT, HP, N3, and H. In other words, no
> distinction is made between the atom types of the neutral vs.
> positively charged terminal groups.
>
> If one were to use N (the notionally "correct" atom type) instead of
> N3 for LYN, I don't think this would change things much as the vdW
> parameters are identical and the bonded parameters are generally very
> similar (and quite often identical). If one were to use H1 instead of
> HP for LYN, however, I'm not sure that the story would be the same as
> the vdW parameters differ significantly, thereby altering solute-
> solute and solute-solvent interactions. Admittedly the difference in
> net charge is probably far more significant in the resulting behavior
> of the neutral vs. charged side chain, but nonetheless it appears that
> Veenstra et al. found different optimal solutions for the HP and H1
> parameters when examining dimers of water with methylamine vs.
> methylammonium.
>
> What I am wondering is how this choice of atom types for the LYN side
> chain was rationalized given that there were/are existing alternative
> atom types (as demonstrated when using antechamber for small molecule
> analogues) or if this was possibly an oversight in defining the
> residue way back when.
>
> Best,
>
> Paul
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 27 2011 - 19:00:02 PDT
