RE: AMBER: generating parameters for the NADH analogue H4NAD

From: Ross Walker <>
Date: Thu, 23 Feb 2006 11:00:30 -0800

Dear Simon,

> 1) How do i modify the pdb to add in the additional hydrogens
> and alter the hybridisation states of the carbons in the
> nicotinamide ring?

I would start from the files in $AMBERHOME/dat/contrib/NADH_and_NAD+

Here you will find NADH.lib which defines an NAD resiude. You can then load
this into xleap and if your changes are minor, extra protons etc you may be
able to use xleap's rudimentary editing window to draw in the extra protons.
You can then write this out as a pdb file and improve the structure by doing
a quick QM optimisation in something like Gaussian. HF 6-31G* should
suffice. You can then load that pdb back into leap to create your new unit.
Draw in the bonds and then choose the atoms types. As a first try I would
try to make suitable guesses based on the frcmod.NADH parameters and types.
You can then see how this works and if necessary try to refit some of the
parameters. This may be more reliable than using antechamber to
'automatically' assign types and parameters.

With regards to the charges you can do a simple resp fit on the modified
residue. This is what I did to generate the charges in the first place. The
70 or so atoms do not pose a problem with regards to computational overhead.

> 3) Some of the carbon-carbon bond lengths differ from NADH so
> would i need to say then minimise this structure in order to
> correct the geometry or would this simply occur when i
> perform my minimisations on the entire system prior to
> running dynamics?

Minimisation prior to running MD will fix a lot of the problems but as with
all hand build residues you should proceed slowly. E.g. when you start MD do
it with some weak restraints on everything but new residue, then heat the
system very slowly with a small timestep (0.5fs or so). Once it has moved
out of the bad starting geometry you can remove all the restraints and
switch to a more reasonable time step and temperature coupling.

Note, the coordinates used to make the inpcrd files will be those specified
in the full protein pdb file that you feed into leap. Leap will not use any
of the coordinates from the pdb file you used to create the residue unit
'unless' you have atoms missing in your protein pdb file. (e.g. missing
hydrogens will be placed based on their geometry in the residue template.)

> 4) Which charge method would be the most suitable considering
> the fact that the molecule has about 70 atoms and also are
> there any specific variables i should be aware of?

See above. I would definately use a RESP fit here, you should be able to do
it with R.E.D. Be careful using AM1-BCC with the NADH analogue, it may have
problems with the charged phosphate groups. It won't take long to do a full
Abinitio Resp fit on the residue.

> I appreciate i am asking quite a lot here but, I have no
> experience of generating my own parameters but would like to
> generate a realistic model because i`m aiming to study the
> motions of the substrate and interacting active site
> residues, compared to the physiological substrate, in some detail.

I would start by getting to grips with creating the new residue correctly.
You can then look to improving the parameters if necessary. One quick test
would be to take your newly created residue and run a normal mode analysis
on it and compare this to an ab initio frequency calculation.

I hope this helps...
All the best

|\oss Walker

| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel: +1 858 784 8889 | EMail:- |
| | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Sun Feb 26 2006 - 06:10:17 PST
Custom Search