AMBER Archive (2008)

Subject: Re: AMBER: Constant pH with new parameters (amber10)

From: Markus Kaukonen (
Date: Sun Jul 20 2008 - 03:08:35 CDT

Dear All/Adrian,

Thank's for the comments!

My trials were pretty preliminary here (my first ever), running 500ps constant pH
oxidised and reduced poplar plastocyaning. Did not do any convergence
checking. Was running at pH 7.0 and salt 0.1. Just calculated the
total charge after the run using the truncated (0 or 1) value of the
protonation fraction given by Charge differences were
caused by one lysine and one glutamic acid in this particular case.
Just wanted to point out that (otherwise than the force field, ff99 or ff03) identical input proceduce produced different results.

Motivation for my constant pH runs was that we were calculating redox potentials
and pKa values for various things using free energy QM/MM. I noticed
that using 'standard' protonation states in plastocyaning resulted much worse redox potential compared to experiment than in some other test systems.
Here we used ff03 force field, so it would have been nice to continue using ff03 also for the constant pH runs, but as it proved out, this is not possible at the moment.

It is true that the total charge state of a protein is an uncertain business. However, some references gave number -8e for poplar plastocyaning (without giving any reference to data). Experimentalist I was talking to at the coffee table were in the opinion that is possible to determine the charge state of a protein with relative standard experimental effort with error +- 2(3) e, so one could get some hint from the experiment.

The only hit with google scholar using words 'Amber constant pH' is the
"Constant pH molecular dynamics in generalized Born implicit solvent"
John Mongan , David A. Case, J. Andrew McCammon
journal={Journal of Computational Chemistry},

So I'm a bit puzzled why I could not find more publications.

Terveisin Markus

On 7/13/08, Adrian Roitberg <> wrote:
> Dear Markus
> I am not sure what you mean by 'total charge of -6' in this context.
> The protonation states are pH dependent of course. You mean you run
> constant pH MD at pH = 7 (let's say) and counted charge afterwards? Are
> you sure the procedure is converged ?
> Now, since the parameters in constant pH MD are parametrized using the
> ff99 charge model (as Dave case mentioned), then you should not use the
> ff03 set at all.
> Moreover, if the difference in two for a charge comes from smal
> differences in a couple of histidines (as I suspect), then I wold be
> very careful with the results. One would be lucky to get pKa of
> histidines to within one pKa unit of experiment, and that, given the
> right conditions, would easily give you one charge per histidine.
> One last item: please go back to the original raw data for your protein
> and see if the researchers really report a charge for their system. As
> far as I can tell this an almost impossible measurement to do. There are
> other measurements that seem to correlate with total charge, but you
> must check for your own system.
> Cheers
> Adrian
> Markus Kaukonen wrote:
>> Thank's,
>> I did some testing with poplar plastocyanin and it makes clear
>> difference (~2 charge) whether one used ff03 or ff99 so I conclude
>> that one should use
>> source leaprc.ff99SB
>> set default PBRadii mbondi2
>> loadoff constph.lib
>> loadamberparams frcmod.constph
>> In this case I obtained total charge of -6 which is closer to
>> experimental one ~-8
>> Anyone else having comments on how well constant pH reproduces
>> experimental pKa values or the total charge of a protein? My
>> statistics is pretty poor here.
>> Teveisin, Markus
>> On 7/11/08, David A. Case <> wrote:
>>> On Tue, Jul 08, 2008, Markus Kaukonen wrote:
>>>> Is it ok to set up a constant pH calculation with leap commands in Amber
>>>> 10:
>>>>> source leaprc.ff03
>>>>> set default PBRadii mbondi2
>>>>> loadoff constph.lib
>>>>> loadamberparams frcmod.constph
>>>> I mean, are the force fied modifications in the manual only valid when
>>>> using
>>>> 'source leaprc.ff99' or can one use 'source leaprc.ff03'
>>> I believe that constph.lib files are based on the ff99 charges, and so
>>> would
>>> not be fully compatible with ff03.
>>>> Is the file 'frcmod.mod_phipsi.1' needed anymore for constant pH?
>>> You want to use leaprc.ff99SB. Then you don't need the mod_phipsi stuff,
>>> as
>>> it is automatically included.
>>> ...good luck...dac
>>> -----------------------------------------------------------------------
>>> The AMBER Mail Reflector
>>> To post, send mail to
>>> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
>>> to
> --
> Dr. Adrian E. Roitberg
> Associate Professor
> Quantum Theory Project
> Department of Chemistry
> Senior Editor. Journal of Physical Chemistry
> American Chemical Society
> University of Florida PHONE 352 392-6972
> P.O. Box 118435 FAX 352 392-8722
> Gainesville, FL 32611-8435 Email
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to

---home   Viinirinne 3 F 12, 02630 Espoo, FIN
---tel +358-(0)45-1242068
--- Rikos ei kannata, eika maatalous
--- Suomessa. (Paimio 1998) ---
The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)