AMBER Archive (2008)

Subject: Re: AMBER: Leap atoms CL and OXT do not have a type

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Thu Oct 02 2008 - 14:15:54 CDT

On Thu, Oct 2, 2008 at 7:36 PM, Bill Ross <ross_at_cgl.ucsf.edu> wrote:
>> I am planning to get the radial distribution vs the distance between
>> chloride ion and protein-atom which I know experimentally to be
>> closest. This for all protein atoms in this situation. Using the vdw
>> provided by Amber for chloride ion.
>
> You might want to do it on a per-atom-type basis to disentangle
> the effects.
>
>> A point of my concern is that the protein atoms at issue are on amino
>> acid side chains, in a couple of cases they belong to carboxylates
>> that, as they come from leap, are not protonated. From the X-ray
>> electron density map it is not entirely clear if, in such a case, it
>> was CO2- or CO2H, and thus the interaction with Cl- is even less
>> clear. Whether CO2- or CO2H depends on the pH the protein has sensed.
>
> CO2H would be the more sensitive test I think - the more attraction
> between Cl- and the molecule, the more sensitive it is to the vdw term.
>
>>
>> I am also organizing to measure pKa of the amino acids with software
>> that, via monte carlo, explores the various conformational situations.
>>
>> Thus, at which stage should the radial distribution procedure be best
>> carried out?
>
> I would run simple dynamics (that's what I know :-). You might even consider
> doing it with a simpler system if you can find an anion complex analogous to
> the crown ether for cations.

That is interesting, also because my protein is quite large. On the
fly, I remember of large azacrowns. In protonated form they complex
chloride ions, if I remember a couple of anions, one above and one
below the mean plane of the azacrown. X-ray diffraction data perhaps
exist. However, this would involve the gaff force field alone. Could
you imagine how to set up a periodic system for that, and would it be
representative? Actually, with my protein complex there is also amino
acid N (LYS) close to chloride.

francesco

>
> On the other hand, your simulation may not be very sensitive to the
> difference in Cl- radius involved here, especially since your solute
> is not positively charged, although you would need to consider the
> charges on the specific atoms in contact with the Cl-. Note that in
> the paper I cited, the whole calculation turned on the contact between
> the cation and the solute O's. (Also it happened that using uncorrected
> ion vdw reproduced experimental results, so there was some cancellation
> of errors that we could only hypothesize about.)
>
>> Unless this question would require the Aladino's lamp.
>
> :-)
>
> Bill
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu