Just adding to this reply, which I essentially agree with.
First shell ligands of magnesium like hydrating waters
exchange on a scale of ca 1 microsecond (experiment, I have no idea
about force field timescale).
Divalents cannot be well described by pair additive force
fields as the total amount of polarization/charge transfer
nonadditive contributions just in the first ligand
shell is for Mg2+ ca 70 kcal/mol. These are missing in simulations
although there is a mutual compansation of errors, fortunately.
One of the problems is lack of polarization/CT towards the first
shell, first shell water molecules are highly polarised and their
H-bonding properties are quite different from bulk water molecules
(in a real system or QM description, "red shift" O-H streches can go
to 0.04A, I would tell some of these could form low-barrier H-bond bridges).
Divalents sample poorly, so if Mg2+ is improperly placed at the
start it may behave like unguided missile in the simulation
with adverse impact on the solute structure once it hits in a wrong
place.
Thus, my suggestion is to be very carefull with
divalents and always compare such Mg2+ simulations
wiht simulations without divalents.
We have briefly commented on this for example in
Long-residency hydration, cation binding and dynamics
of Loop E/Helix IV rRNA - L25 protein komplex.
Biophys. J. 87, 2004, 3397-3412
or in very recent release of electronic encyclopedia of
computational chemistry, Schleyer ..., Eds.
(Metal - Nucleic Acids Interactions item).
See also, e.g., Petrov AS, Lamm G, Pack GR BIOPOLYMERS 77 137-154 2005
I also agree with the comment on Cl-, just adding,
anions are highly polarisable (in contrast to monovalent
cations), so somewhat outside the applicability
of pair-additive description. This may lead to a range of
troubles, including ion-pair clustering in the production
phase after successfull equilibration.
best wishes Jiri
-------------------------------------------------------
Jiri Sponer
Institute of Biophysics
Academy of Sciences of the Czech Republic
Kralovopolska 135
CZ-61265 Brno
Czech Republic
e-mail: sponer.ncbr.chemi.muni.cz
fax: 420 5412 12179
phone: 420 5415 17133
http://www.ibp.cz/labs/LSDNA/
--------------------------------------------------------
[ Charset UTF-8 unsupported, converting... ]
> Hi according to our experience, if some negatively charged atom is close
> enough to Mg2+ (for example, the Mg2+ ions in polymerases or in kinases
> and the COO- group of the nearby amino acid residues), this distance is
> unlikely to change greatly during simulation. By close enough, I mean
> <=2.8 angstroms. By "unlikely to change greatly", i mean you will unlikely
> see the distance go beyond 3 angstroms. there are several reasons i could
> think of:
>
> 1. the interaction between the Mg2+ and your Cl- is very strong and there
> is a limited number of solvent molecules nearby to penetrate.
>
> 2. most likely you assigned a partial charge of +2 to the Mg2+ and -1 to
> the Cl-. This does not necessarily reflects the real situation. There have
> been reports that charge associated with Mg ions is "dampened" by the
> solvent to something like 1.4 instead of 2.0. Using charge = +2 for Mg
> might account for the unexpected stability of interactions between Mg and
> negatively charged atoms/ions. However, there is probably no easy way
> to use a non-integer charge value for ions. We have thought about
> associating Mg2+ with a neighboring residue/entity and treat them as ONE
> customized (non-standard) residue and compute the RESP charges. This will
> likely "soften" the charge on Mg a bit.
>
> 3. The added Cl- ion might be too "artificial". i.e. usually Mg2+ at
> binding site plays some critical roles in stablizing the bound ligand. I
> am not sure how "physiological" it is to have this Cl- so close to the
> Mg2+ and what kind of impact it will have on the binding interactions.
> This has to be considered carefully...It is probably safer to add a couple
> water molecules around Mg2+ first (thus Cl- cannot be that close to Mg2+)
> and then neutralize your system by adding Cl-. Adding water is
> likely "softer" than adding Cl- directly.
>
> just my 2 cents.
>
> Lihua
>
> On Sat, 9 Jul 2005, Ye Mei wrote:
>
> > Date: Sat, 9 Jul 2005 14:54:16 +0800
> > From: Ye Mei <ymei.itcc.nju.edu.cn>
> > Reply-To: amber.scripps.edu
> > To: amber mailing list <amber.scripps.edu>
> > Subject: AMBER: simulation with Mg2+ ions and counter ions
> >
> > Dear Amber users
> >
> > Please help to check whether my simulation is reasonable.
> > I am trying to study a certain ligand binding to a protein with Mg2+ ions. First, I docked this ligand to the binding site using autodock. In the docking procedure, I included neither the Mg2+ ion, counter ions nor the water molecules near the binding site. Then I took some binding modes and add Mg2+ to the protein and locate it where it was in the pdb file. 2 CL- ions were then added to this complex to neutralize the whole system. One of the CL- ion is binded to this Mg2+ ion with distance less than 3A. Periodic water box is added to this complex with 8 A between the protein and the unit boundary. Minimization and 2ns MD simulation under 300K were applied to this system. But I found that the CL- ion that is bonded to Mg2+ has not been pushed away by water molecules, but is still there. I guess maybe the simulation time is not long enough, though the whole system has seemed to in equilibrium for a very long time. But my computer resource is very limited, and I cannot giv!
e !
an!
> > ot!
> > her ns simulation for each binding mode.
> > So, can I just stop here, and calculate the binding free energy? If not proper, is there any easy way for me to correct it?
> >
> >
> > Best regards,
> >
> > Ye Mei
> > ymei.itcc.nju.edu.cn
> > Institute of Theoretical and Computational Chemistry
> > Key Laboratory of Mesoscopic Chemistry
> > School of Chemistry and Chemical Engineering
> > Nanjing University
> > Nanjing 210093
> > P.R.China
> > 2005-07-09
> >
> > -----------------------------------------------------------------------
> > The AMBER Mail Reflector
> > To post, send mail to amber.scripps.edu
> > To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
> >
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sat Jul 09 2005 - 09:53:01 PDT
