To understand interaction of molecule and protein is important but difficult step of drug discovery.
In case of SBDD, researcher often analyse structural data. But, it’s difficult to understand nonclassical interaction.
FMO is one of the tool to analyse ligand-molecular interaction more deeply than MM.
Today I read very impressive paper.
“The Fragment Molecular Orbital Method Reveals New Insight into the Chemical Nature of GPCR–Ligand Interactions”
http://pubs.acs.org/doi/abs/10.1021/acs.jcim.5b00644
The authors analysed 18 GPCRs structures with FMO.
They used GAMESS for FMO calculation, it’s key point because GAMESS can provide interaction energy as four component 1) electro statics 2) dispersion 3) charge transfer 4) exchange repulsion. This function is called PIEDA(Pair Interaction Energy Decomposition Analysis).
This means FMO can reveal contribution of interaction of fragments so this technology analyse not only classical H-bond, salt bridge but also nonclassical(difficult to identify from visual inspection) CH/pai, cation/pai, halogen/pai, etc.. interaction.
The authors revealed new interaction about some GPCRs using FMO.
And most interesting part for me was Fig 5-A.
They made a comparison between binding affinity and PIE( Pair Interaction Energy ). And the results showed good correlation but one target didn’t.
The target was CXCR4 and ligand was IT1t (PDB 3ODU). In this case ligand conformation was U shape.
So, they hypothesised that U shape is key for affinity, and calculated PIE between right hand side chain and left hand side chain of ligand.
And the PIE showed good correlation to binding affinity. Cool.
I think FMO is powerful tool for drug discovery. It can apply not only retrospective analysis but also prospective analysis i.e. scaffold hopping( or morphing ?), linker change, etc.
But, how do I prepare correct initial structure ? Docking pose is true ? X-ray structure is snapshot, is it shows true structure ?
Hmm I got lots of home work.
Is life worth living? 