the problem is clearly the water model. GB describes the electrostatic
contribution, while SA gives the hydrophobic part of the solvation free
energy. The hydrophobic surface tension parameter was fitted to exp.
hydrocarbon free energy of solvation and probably isn't very transferable
to other systems. Note that the solvation free energy is entropic in
nature and is rather complicated to model.
For other approaches see the following references which use empirical
atomic solvation parameters and got good results on Z/B DNA stabilities:
P. S. Ho et al, J. Phys. Chem. 92, 939 (1988)
T. F. Kagawa et al, Biochem. 28, 6642 (1989)
On Wed, 14 May 2003, Peter Gannett wrote:
> Amber users:
>
> We have been doing some simulations of 5'-CGCGCGCGCG-3' and a modified version (G6 from the 5' end) with respect to the relative stability of the B and Z DNA forms. We've run about 2ns of simulation on the B/Z forms of both the unmodified and modified forms (total of 4) and then have processed the trajectories (from ~200-1800 ns) with mm_pbsa. The control parms for the mm_pbsa were
>
> IGB=1
> SALTCON =0
> EXTDIEL=80
>
> SURFTEN=0.0072
> SURFOFF=0.00
> DIELC =4
> MAXCYC=5000000
> DRMS=0.0001.
>
> The final energies (delG) are:
>
> B(unmodified) = -3998.5
> Z(unmodified) = -4013.8
> B(modified) = -3954.8
> Z(modified) = 3964.2
>
> Experimentally, at 200 mM NaCl, the B form of the unmodified oligo is the only significant form while, under the same conditions, the modified form is roughly 1:1 B/Z.
>
> Obviously, the calculated results are significantly different from the experimental results. I don't understand why the Z unmodified is calculated to be the most stable form when there is no way that it is. I don't mind the order for the modified form, but given the result for the unmodified, I don't give the result much weight.
>
> Note, in case it helps, the mm_pbsa gets large differences in ELE between the B and Z forms (unmodified and modified), favoring the B form which I am willing to accept. The is counterbalanced by a large value for GB, favoring the Z form in both cases. Relatively speaking, small changes in these numbers could change the relative ordering of things. By small changes, I mean with respect to the STD reported on these values.
>
> Also, we ran the mm_pbsa with SALTCON=1. This increased the stability of the Z forms relative to the B forms, but did not change the relative ordering of the four DNAs.
>
> But, I don't think this is the whole story. Any suggestions, comments, etc? They would be appreciated.
>
> Pete Gannett
>
>
--
-------------------------------------------------------------------------
Robert Endres http://www.csm.ornl.gov/~endres/
Computer Science and Tel: (865) 574 4845
Mathematics Division FAX: (865) 574 7659
Oak Ridge National Laboratory E-mail: endresrg.ornl.gov
P.O.Box 2008
MS 6114
Oak Ridge, TN 37831-6114
Received on Wed May 14 2003 - 14:53:02 PDT
