[gmx-users] Double vs. Charged Single bonded Oxygen in GROMOS96.1

[Justin A. Lemkul]

Peyman Yamin wrote:
> On Monday 27 October 2008 18:08, Justin A. Lemkul wrote:
>> Peyman Yamin wrote:
>>> Hi again dears,
>>>
>>> I should ask for another comment. I have a Phorphorus to which two Oxygen
>>> atoms are bonded. One O is double bonded and the other has only a single
>>> bond while being negatively charged. In the topology file I get from
>>> prodrg both Os are denoted with OA (GROMOS96.1) atom type. both have the
>>> same charge (-0.421) and mass (15.9994). I guess the charge is reasonable
>>> for the double-bonded one, but not for the one with a single bond and a
>>> free electron. How could I correct this?
>> There is no true double or single bond.  Think resonance!
> 
> 
> Impressive answer! But, well, I'm not sure about that resonance thing?
> 
> ...C-O-P(+3)-O-C... So far it's OK! but P still has 3 fellow electrons to 
> share. it does share two of them at once with another O. I understand till 
> here. there is another O who wants to get 2 electrons shared, but only 
> receives one: it needs more => so we call it a "negatively charges fellow 
> oxygen", don't we?

I think there are going to be two distinct atom types.  For a few examples of 
how to deal with phosphates, see the .rtp entry for ATP, or look at the lipids 
that Peter Tieleman makes available online.  If we consider a traditional Lewis 
structure for the phosphodiester (hopefully this translates alright):

      O
      ||
R--O--P--O--R
       |
       O-

Really the =O and --O^- are equivalent in terms of atom type (OM in Gromos) and 
charge.  The O in the phosphodiester linkages are likely to bear similar 
charges, but have a different atom type (OE in Gromos).  Hence why proper 
parameterization takes a substantial amount of time.

Be careful if you're using the topology straight from PRODRG.  Its 
interpretation of charge groups and charges often requires a bit of refinement. 
  Also realize that Gromos96 is not the best for reproducing the properties of a 
lipid, although I don't know if isolated lipids have been considered.  The 
Berger lipid parameters (Gromos/OPLS combination) are widely used...but that's a 
whole other debate :)

> 
> Well I'm well aware of my chemistry knowledge fading away, but you'll tell me 
> an OA will do well for both??
>

Maybe for the phosphodiester O, but not for the =O and --O^-; they are more like 
a carbonyl O (atom type OM).

-Justin

> Thanks for the time, indeed,
> Peyman
> 
>> -Justin
>>
>>> Cheers,
>>> Peyman
> 

-- 
========================================

Justin A. Lemkul
Graduate Research Assistant
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================