AMBER Archive (2009)

Subject: [AMBER] Re: [q4md-fft] Charge calculation or non standard residue

From: FyD (
Date: Sun Jul 05 2009 - 03:30:50 CDT

Dear Bagheri,

> I worked on metallo enzyme by AMBER package, for making library I need the
> charge of the atoms near the metal ions.
> my problems is ,
> can I use RESP server for calculating essential atom's charge?
> if yes what conformation I should upload?hole protein or the special
> fragment of protein that include this microenviroment contain metal ion?

You can indeed use R.E.D. Server (not "RESP Server") for that since
you might need to execute R.E.D.-IV. That being said you can work with
R.E.D.-III.x to prepare/understand the strategy, and you will need
Ante_R.E.D. as well which is only provided in the R.E.D.-III.x tools.

You could follow the following steps:
   (i) set up the topology & select the atoms to be involved in your
metal complex,
  (ii) get a correct QM geometry optimization output for your complex
(iii) run R.E.D.: you might need to adapt some keywords in the source
code (or R.E.D. Server: we will release a new charge model for metal
  (iv) adapt your new FF library with other existing ones.

Your questions are related to the first step (i). When you will use
R.E.D./R.E.D. Server, I strongly suggest you to use these tools in
their "mode 2":
This means you need to provide a QM geometry optimization output fully
characterized for each molecule used in the charge derivation as
input. More generally for each P2N file, you will need a QM geometry
optimization output.

For defining your metal complex, you need to extract the atoms you
believe important in this complex (you could start form a
characteristics PDB file for that), and run QM geometry optimization
for these selected atoms. You need to select a QM package for that,
pick up a correct theory level in this geometry optimization step, &
test various spin multiplicities: Time consuming work here.

For the selection of the atoms (your complex) involved in QM
computation, larger will be the sphere around your metal more accurate
should be the representation of your complex, but longer will be your
QM jobs... You could try piking up only the aminoacids (AA) connected
to the metal in their dipeptide form (CH3CO-AA-NHCH3) for instance.
You might have your own ideas here as well.

About the conformation to be selected, you could start from that
available in your PDB structure and see how the conformation of the
selected atoms in the absence of the protein (your selected complex)
is going to be affected by the QM geometry optimization job.

Then, when you got this optimized structure you can think to load it
in R.E.D. or R.E.D. Server ;-) Before running the job, you will have
to prepare the P2N file with the correct atom connectivities between
the metal the other atoms. Ante_R.E.D. should do the job if your
initial Cartesian coordinates are correct. If your initial Cartesian
coordinate set is bad, you will have to complete the atom
connectivities by hand. At the end, you get a FF library in the Tripos
mol2 file format ready to be loaded in LEaP.

regards, Francois

AMBER mailing list