Simulated annealing of protein-DNA complex

User discussions
,
1

GROMACS version: 2024.2
GROMACS modification: No

Dear all,

I am trying to set up an atomistic simulation system containing a protein bound to a 19-base stretch of double-stranded DNA. The protein contains 3 zinc-finger motifs. Following published work on similar complexes, I have chosen to use the Amber14sb force-field with additional parameters for the bound zinc atoms from Macchiagodena et al, 2019.

My starting model is a “manually” generated composite complex, consisting of a protein homology model, which I have positioned relative to the DNA by superposition onto the experimental template structure bound to DNA. The DNA was then mutated in silico to the known cognate sequence of my protein.

Due to the amount of manual tampering with the starting model, I thought it would be a good idea to do some simulated annealing during equilibration, in order to let the protein and DNA find a low-energy starting configuration before my production runs.

Unfortunately, I haven’t been able to find any detailed protocol for SA of protein/DNA (nor protein/protein) complexes in the literature, so I am hoping for some hints from the GROMACS community.

My current plan is to do this in 4 steps:

  1. An energy minimization basically following Justin Lemkul’s Lysozyme Tutorial

  2. a 150 ps NVT equilibration run with position restraints on protein, zinc and DNA to equilibrate the solvent around the complex and set the temperature to 310.15 K

  3. Still within an NVT setting, 3 cycles of simulated annealing from 310.15 K to 410.15 K and back to 310.15 K, changing the temperature by 10 K every 500 ps.

  4. Another 500 ps (?) of NVT at the final temperature of 310.15 K to let everything settle.

  5. NPT equilibration following the Lysozyme tutorial.

This plan is based on both literature and my “gut feeling” from an X-ray crystallography background.

My questions are:

  • Does this sequence of steps make sense?
  • Is the upper temperature limit of 410 K fine for the purpose?
  • Is it common (and advisable) to do several cycles of SA?
  • Most importantly: what to do about position restraints? Should I keep both protein and DNA atoms restrained to keep them from unfolding/melting? Can the two molecules then move relative to each other during the SA - kind of like two rigid bodies if I keep them in 2 separate t-coupling groups? Or shall I only restrain the DNA and leave the protein unrestrained? But how do I then prevent the zinc atoms to dissociate from their binding sites at high temperatures? Will (should?) the protein unfold and then refold in silico? Or should I rather go for a milder temperature and just warm it up to, say, 350 K?

Sorry if these are basic questions, I am relatively new to MD simulations and truly appreciate any input or hints on relevant literature on this.

Many thanks for sharing your thoughts!

Maike