I am bit new to bioinformatics and I am working on protein-protein docking studies. I have modelled a protein structure and 9 of its point mutants using homology modelling. The template has a sequence identity of 26% and the length of the modelled structure has 75 amino acids. When I subjected all the 10 structures to Molecular dynamics, the wild type RMSD was very unstable comparing to its mutants. The wild type structure was not stabilised even at 35ns, but the mutants were stabilised within 15ns.
But when we checked the stability of the docked complexes, the wild type complex was much stable as comparing with few mutant complexes. why the wild type structure was not stable, when it is alone?
Why the wild type protein structure is unstable comparing to its point mutants?
A protein sequence can have evolved to exhibit a greater stability when in a complex. This is not so uncommon as favorable binding energy lowers the energy for the complex compared to the unbound states. There are examples of inherently unfolded proteins adopting a unique structure only in the presence of a complex. Many possibilities exist.
Explaining the behavior of your simulations is not so simple. What program are you using for your simulations? Are you using explicit solvent? What are the rmsd differences? What evidence do you have that your modeled wt protein has a "reasonable" or accurate structure to start with?
MD simulation is not a crystal ball - at least not yet anyway.
As Kurt suggested, you should verify your simulation results by using different force fields. Well you have separately modeled the mutant proteins,you have to consider few points here.
1. Mutant models can be generated form the wild protein after altering the desired residue, why to model separately?
2.Check all the mutant models, comapre with the wild model and verify there should not be much structural deviation, because different starting conformation will give you different trajecotry.
3.26% identity is not good to develop reliable homology models. Use threading or ab initio modeling methods.
Thank you Dr. Kurt. I am using Amber program with explicit solvent model. The average RMSD value of wild type protein ranges between 5 to 10 nm and the mutants ranges between 5 to 7.5 nm. The quality of the model was checked using Ramachandran's plot., which shows 98% of the residues are in the favourable region.
Thank you vikash. The mutants were generated from the wild type structure only. The structures were compared and there were not much structural deviations.
Hi again:
The force field will help to insure that violations are minimized and distributed. I am not surprised by the large % in allowed regions in the Ramachandran analysis - even with questionable simulations.
I am concerned with the large RMSD values. Are you starting from a high-quality X-ray structure or NMR structure? How do you select the starting structure if there are more than one in the asymmetric unit or ensemble? Are you minimizing before starting the simulation? Are your reported RMSD values of MD simulations of the complex or the starting protein structure? What is the ligand?
As @Vikash mentions, the low sequence identity is definitely in the grey area for homology modeling. I would say that before docking and creating mutants you need to focus on obtaining a MD protocol that gives you reliable behavior with the starting protein structure alone.
The template is a X-ray structure. We searched multiple databases like Blast, Pfam, ITasser, and Phyre2 for the template structure. Further we finalised the template structure using RMSD values from PDBefold. At the end we managed to get only 26% sequence identity.
Yes. We do minimisation before simulation. RMSD is for the starting structure. The complex become stable at 1ns itself.
Thank you martins. If the kinetic energy and potential energy are stabilised, still we need to look for other system?
Also the point mutants become stable as compared to WT, which means the system is working right?
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics