[Dock-fans] Does prepare_amber.pl (i.e, LEaP understand capping?

[Francesco Pietra]

Hi:

For reasons that would be long to detail, I would like to work (DOCK6)
with a model where the first and last residues are capped with ACE and
NME respectively. That for more than one stretch of standard amino
acids. The stretches are separated from one another by a TER line

The way I have built the pdb file is not understood by
prepare_amber.pl, i.e., by leap, as it emerges from filename.log
(where filename is the prefix of the pdb file).

As an example with one stretch, the first amino acid:

ATOM      1  N   LEU     1    xyz
ATOM      2  CA  LEU     1
ATOM      3  CB  LEU     1
ATOM      4  CG  LEU     1
ATOM      5  CD1 LEU     1
ATOM      6  CD2 LEU     1
ATOM      7  C   LEU     1
ATOM      8  O   LEU     1
ATOM      9  HA  LEU     1
ATOM     10  HB2 LEU     1
ATOM     11  HB3 LEU     1
ATOM     12  HG  LEU     1
ATOM     13 HD11 LEU     1
ATOM     14 HD12 LEU     1
ATOM     15 HD13 LEU     1
ATOM     16 HD21 LEU     1
ATOM     17 HD22 LEU     1
ATOM     18 HD23 LEU     1
ATOM     19  H   LEU     1

The corresponding ACE:

HETATM 4018  C   ACE    46     xyz
HETATM 4019  O   ACE    46
HETATM 4020  CH3 ACE    46
HETATM 4021 HH31 ACE    46
HETATM 4022 HH32 ACE    46
HETATM 4023 HH33 ACE    46

The various capping residues are NOT separated from one another by a TER line.

LEaP does not understand that there is capping (while Chimera
understands that and reproduces perfectly the whole structure). LEaP
adds a new atom, named H, to each beginning terminal, and "missing"
OXT to each terminal last. Then it tries to join first ACE with last
NME. Perhaps I could avoid the last issue by placing a TER line
between the capping residues but before embarking in a perhaps
hopeless avenue, I ask for experience, or insight how to configure the
pdb file.

Without capping, the model is treated correctly by prepare_amber.pl,
though it has charges that the true protein has not, which is likely
to alter any docking.

Thanks
francesco pietra