AMBER Archive (2009)

Subject: Re: [AMBER] looking for parameters for DHFR substrates and inhibitors... [one more try]

From: FyD (fyd_at_q4md-forcefieldtools.org)
Date: Wed Dec 02 2009 - 02:19:00 CST

Dear Sidney,

May-be you should think developing your own parameters ;-)

After looking in Wikipedia:
http://en.wikipedia.org/wiki/Methotrexate
http://en.wikipedia.org/wiki/Dihydrofolate_reductase
http://en.wikipedia.org/wiki/Folic_acid
it looks like these molecules derive all from L-glutamic acid. Right ?
You could easily imagine a building block approach based on a common
fragment L-glutamic acid and with various aromatic counterpart(s).

You could use R.E.D.-III.3 for that (or may-be R.E.D. Server/R.E.D.-IV
if you want to develop the FF libraries of all these molecules +
analogs in a single approach).

You will need to build a C-terminal fragment for L-glutamic acid
See http://q4md-forcefieldtools.org/Tutorial/Tutorial-3.php#17
  & a central fragment in same time
See http://q4md-forcefieldtools.org/Tutorial/Tutorial-3.php#15
or you can use the fragment CGLU (if charged) from the Amber force
field topology database

Then, either you split your aromatic counterparts in two using
additional constraints (inter-molecular charge constraint) during the
charge fit or you keep all in a single block. Once again, if a common
group is found in the aromatic part I would split the system in twos:
common & variable. You can find examples of such approaches in
R.E.DD.B. I think you can find answers of such charge constraints in
the last R.E.DD.B. projects:
See http://q4md-forcefieldtools.org/REDDB/projects/F-84/
See http://q4md-forcefieldtools.org/REDDB/projects/F-85/
In these two projects, you will find "connections" between various
building blocks of different natures.

Following a similar approach, my feeling is that you can easily
generate a project with various common & variable parts/blocks for all
your molecules & many analogs you might be interested in the future.

Advantage of the building block approach:
- you fully control the conformation(s) of the building blocks used &
- you can build FF library for the molecules you are interested and
for many analogs (variable parts).

In your case as you have an amino-acid, I would use the ff99SB force
field; similarly as what is done in the "F-85" R.E.DD.B. project.

If you are interested in starting such a work (quite quick since the
different building blocks are small), do not hesitate to ask questions
in the amber or q4md-fft mailing list.

regards, Francois

> There have been quite a few computational studies of DHFR, and I was
> surprised that I couldn't find any published parameters for the substrates
> and inhibitors readily available. Anyone have any insights into where I
> could find them? Using standard parameters for these ligands would be
> preferable to reinventing the wheel and doing my own parameterization.
> Thank you in advance for your help.
>
> Sid
>
> On Sat, Nov 21, 2009 at 8:47 PM, Sidney Elmer <paulymer_at_gmail.com> wrote:
>
>> Hi all,
>>
>> Does anyone know where I could find parameters for methotrexate,
>> dihydrofolate, and tetrahydrofolate? I was unable to find anything on
>> google. I also checked on REDDB but couldn't find any there, either. Any
>> leads would be very helpful. Thanks.
>>
>> Sid

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber