DNA/RNA records for OPLS force field in GROMACS format.

1. Preface.

   We used gmx ff from GROMACS v 3.1 to study the RNA protein complex, a part of the small subunit of the ribosome (the minimal S7 protein-binding fragment of 16S RNA and protein S7 of Thermus thermophilus). After 500ps of simulation, we found that the secondary structure of the protein was completely disordered, which is in disagreement with X-ray data. This clearly shows that the charges that are used for nucleic acids by the old gmx ff are not suitable to study RNA/DNA-protein complexes.

    We decided to try other force fields of GROMACS and found that the recently developed OPLS force field does not contain complete entries for nucleic acids. Thus, we decided to add RNA/DNA entries to OPLS-AA/L force field.
2. Methods.
The following parameters were considered:
1) Bonded interactions
a) Bonds and angles: All constants for nucleic acid's bonds and angles are present in GROMACS OPLS
b) Dihedrals: The most of dihedrals already exist in OPLS ff. At present only the sugar parameters are missing. All sugars dihedrals were added from the OPLS dihedral list, which were most close to ribose.
2) Non-bonded interactions
a) The Lennard-Jones interaction: The Lennard-Jones interaction: All LJ parameters for nucleic acids are already present in GROMACS OPLS.
b) Charges: All charges were taken from Amber94 force field. These charges are atom-centered quantum-chemically derived charges (RESP: D.A. Case, T.A. Darden, T.E. Cheatham, III, C.L. Simmerling, J. Wang, R.E. Duke, R. Luo, K.M. Merz, B. Wang, D.A. Pearlman, M. Crowley, S. Brozell, V. Tsui, H. Gohlke, J.Mongan, V. Hornak, G. Cui, P. Beroza, C. Schafmeister, J.W. Caldwell, W.S. Ross, and P.A. Kollman (2004), AMBER 8, University of California, San Francisco).
2.Test.
NPT-ensemble.
To test the chosen parameters we performed several simulations:
1) MD of B-form DNA. This structure is stable in the environment of the native solution. We used B-DNA dodecamer (5'-d(CpGpCpGpApApTpTpCpGpCpG)-3') solved by X-Ray analysis (PDB ID - 1BNA).
After 2ns of MD, RMSD was less then 0.3 nm (3 A):

Legend for picture: blue is the original structure, green is the structure after 2ns of MD. Some trajectory frames in pdb format are presented here (3 Mb).

2)MD of the A-form form of the same DNA.
The A-form of DNA spontaneously transforms to the B-form in a solution.
We constructed the A-form (3DNA package) of the same DNA that was used in the first test and subjected this structure to MD in 0.1 M salt solution. We observed a complete transformation from the A-form to the B-form within 2 ns as was described in the literature.


RMSD chart shows the deviation of MD structure of the B-form DNA from that obtained by X-ray analysis.
Some trajectory frames in pdb format is presentedhere (2.8 Mb).

3) The test of Z-form stability was performed together with Istvan Kolossvary (istvan@kolossvary.hu). The resulting NPT RMSD obtained by MD deviated from that obtained by X-ray by less then 3 A.

4) RNA. We performed some simulations of RNA stem-loop structure (PDB:1ESY).
Here you can see results :

RNA fluctuations obtained by MD were in a good agreement with NMR data.

NVE-ensemble.

5) 100 ps MD in NVE ensemble was performed for DNA in the Z-form. The results of simulation demonstrated relatively good energy conservation.

6) Some tests with DNA-protein complexes were also performed. RMSD differed from that of X-ray data by less than 3 A.

That's all.
What to do:
Add records for RNA 3' and 5' ends (for DNA it has already been added).
Improve scripts.
Files to download:
ffoplsaano.tar.gzLatest GROMACS 3.2.1 OLPS force field + DNA records with 5 and 3 end residues (charges from Amber OPLS)
ffoplsaanr.tar.gzLatest GROMACS 3.2.1 OLPS force field + DNA records with 5 and 3 end residues (charges from Amber 94, RESP)
scripts.tar.gz some scripts to fix pdb files for pdb2gmx.
Any question can be asked at : "golovin_at_genebee.msu.su" and "burnnick_u_at_rambler.ru"