[gmx-users] Amber missing psi dihedral at cationic N-terminus of protein/peptide?

[Christopher Neale]

Dear Users:

does anybody know why the amber force field does not apply a psi dihedral to the N-terminal residue when the N-terminus is charged (standard zwitterionic protein termini)? Looking at the amber software parm files (e.g., in file dat/leap/parm/parm99.dat ), the issue appears to originate in the amber force field itself, rather than stemming from the gromacs port of the amber force field.

I've posted to the amber mailing list: http://archive.ambermd.org/201412/0255.html but I have not yet received any response there and wonder if anybody on this list knows if this missing psi at a cationic N-terminus is intentional or an oversight? 

Here are the details on the missing psi dihedral in the gromacs topology files:

In gromacs, amber99.ff/ffbonded.itp has non-N-terminal psi dihedrals:
 N   CT  C   N     9     180.0      7.11280     1  ; 
 N   CT  C   N     9     180.0      8.36800     2  ; 

but nothing corresponding to the cationic N-terminus (N3 in place of N), so the following dihedral likely gets picked up:
 X   C   CT  X     9       0.0      0.00000     0  ; JCC,7,(1986),230

The same thing is true for all amber force fields in the gromacs 4.6.7 distribution and all amber force fields in the amber14 software suite. 

I have also verified the lack of a psi dihedral term by gmxdump on a gromacs .tpr file.

Note that opls and charmm forcefields under gromacs do have a non-zero psi value for charged N-termini:

OPLS:
In gromacs [ dihedraltypes ], oplsaa.ff/ffbonded.itp has N-terminal dihedrals:
  N      C      CT_2   N3      3     10.36376  -6.60654 -10.49347   6.73624   0.00000   0.00000 ; Psi  peptides AA N-CA-C(O)-N
  N3     CT_2   C      O       3      0.00000   0.00000   0.00000   0.00000   0.00000   0.00000 ; peptides  X-CT_2-C(O)-O

that are identical to the non-N-terminal dihedrals:
  N      C      CT_2   N       3     10.36376  -6.60654 -10.49347   6.73624   0.00000   0.00000 ; Psi  peptides AA N-CA-C(O)-N
  N      CT_2   C      O       3      0.00000   0.00000   0.00000   0.00000   0.00000   0.00000 ; peptides  N-CA-C(O)-O

CHARMM:
In gromacs, the charmm27 force field also accounts for the different name of the N-terminal nitrogen in the charged form (NH3 vs. NH1) for the psi dihedral:

NH1    C    CT1    NH1    9    0.00    2.5104    1
NH3    CT1    C    NH1    9    0.00    2.5104    1

and also for the N-CA-C-O dihedral:
O    C    CT1    NH1    9    0.00    0.0    1
O    C    CT1    NH3    9    0.00    0.0    1

Thank you,
Chris.