Protein Protein Interaction Support

[Titorat]

Hello,
Does BindFlow support Protein Protein affinity Prediction?
Sorry if this is already answered in the Docs, still trying to find my way around the docs.

Thanks a lot,

[ale94mleon]

Hi @Titorat.

Thank you for your comment!

Currently, BindFlow is designed for studying small molecule binding to receptors, which may be either protein-based or non-protein-based.

The calculation is done using either FEP or MM(PB/GB)SA.

Using FEP for protein-protein is unfeasible as the simulation requires to entirely disappear one of the protein during the simulation and such perturbations are intractable (or at the very least extremely expensive). However, if one of the protein (the binder) is a small/mediums size peptide is doable.

In the other hand, you might use the most approximate MM(PB/GB)SA for protein/protein binding free energy calculation (see gmx_MMPBSA protein-protein tutorial). But we have not tested this yet. Be aware that if you have a couple of protein/protein pairs you might not need BindFlow, you may use gmx_MMPBSA directly. Only you will benefit from BindFlow automation in the hypothetical case of having from hundred to thousand of protein/protein pairs.

[Titorat]

Thanks a lot for the response @ale94mleon

Would just like to note that I believe FEP for Protein-Protein interactions has been mentioned a couple of times in the literature

https://pmc.ncbi.nlm.nih.gov/articles/PMC9241073/
https://pmc.ncbi.nlm.nih.gov/articles/PMC5383735/#ab0005

Thanks agiain,

[ale94mleon]

Hi @Titorat, thanks for the papers.

Maybe a small correction on my initial answer, BindFlow use Absolute Binding Free Energy (ABFE). The papers you mentioned calculate the effect of mutations at the protein-protein interface, this is Relative Binding Free Energy (RBFE). RBFE perturbations involve typically very small perturbations: transform one amino acid into another one using an alchemical route. Those studies do not calculate the effect of one protein binding to another one, but how changes in the interaction interface affect the binding relative to, let say, the wild type protein-protein complex. These studies are highly valuable for the biological insights they provide; however, BindFlow is not designed to perform them at this time, it might be in the future 😉.

BindFlow (and the major ABFE methods) use a thermodynamic cycle that "disappear" the whole binder in solvent, this means turn off all interactions, and then reappear those interactions in the presence of the receptor. This is, for small to medium peptide accessible, however, for a full protein domain I am not aware of someone tried this as this would imply very costly simulations.