All-Atom MD Simulation Setup for DNA (AMBER)

System setup with AMBER

• Generate the DNA structure with a tool like NAB or Chimera.
• Generate the ligand structure with GaussView or Avogadro2.
• Run antechamber to convert the ligand into an AMBER-suitable format (prepi or mol2 — mol2 is preferred), then parmchk2 to generate the matching force field modifications:

antechamber -fi pdb -fo prepi -i ligand.pdb -o ligand.prepi -rn LIG -c bcc -at gaff2
parmchk2 -f prepi -i ligand.prepi -o ligand.frcmod

• Combine the DNA, ligand, and protein with tleap, then run it:

tleap -s -f tleap.in > out.txt

Force field choices

For the DNA strands, the standard AMBER force fields to pick from:

ff99bsc0    legacy bsc0 correction
OL15        bsc0 + OL1 (ε/ζ) + OL4 (χ)
bsc1        parmbsc1
OL21        OL15 + OL21 (α/γ)       <- current default

Pair the DNA force field with ff14SB for protein and a water leaprc — leaprc.water.tip3p or leaprc.water.opc.

Building the system in tleap

The canonical build for a plain DNA duplex:

source leaprc.DNA.OL15      # or leaprc.DNA.bsc1 / leaprc.DNA.OL21
source leaprc.water.tip3p   # or leaprc.water.opc
dna = loadpdb dna.pdb
addions dna Na+ 0           # neutralize
solvateoct dna TIP3PBOX 8.0
saveamberparm dna dna.prmtop dna.rst7

During equilibration, hold the DNA geometry in place with positional restraints — start with &cntrl ntr=1, restraintmask=':1-24', restraint_wt=25.0, then relax the weight down to roughly 0.5–1.0 kcal/mol·Å⁻² for the equilibration runs. Run pmemd with -ref heat.rst so the reference coordinates come from your heating-stage output.

Homology modeling for a missing residue

The easiest way to install MODELLER is through conda:

mamba install -c salilab modeller

• Add the MODELLER license key as instructed by the installer.
• Note the executable name — it follows the pattern modVersion (for me, mod10.5).
• MODELLER can be run manually per its docs, but Chimera has a built-in function with a much nicer visualizer for this.
• Load the structure with the missing residues (I used 5L2I) and let Chimera drive MODELLER from there.

A note on cysteine naming

CYS is a normal cysteine residue; the deprotonated (negatively charged) state is written CYM, and a cysteine in a disulfide bond (or other special bond) is written CYX. [1]

References

1. AMBER Advanced Tutorial 1