Modeling structures of novel biological targets

Understanding more about the protein and the binding site makes it possible to carry out a virtual screen. It also becomes easier to predict the bioactive conformation. Furthermore, having a model of the biological target makes it is easier to optimize and make changes to a compound structure since you have more insight into the mechanism of the compound and the spatial restraints on possible changes.

Homology modeling involves taking all of the amino acids in a sequence then swapping the residues from the sequence you have the structure of, to the sequence for which you don’t have the structure. This is a complex task that is carried out in 3D modeling space.  The process works well for similar sequences, but is not successful where there is a large difference between sequences.

Using modeling to predict the binding site

One of our customers had identified a novel target. They had crystal structures of a similar protein in humans and rats, but not of the version that they were interested in. In the absence of a crystal structure, the only way to get to the protein of interest was by modeling.

Cresset Discovery scientists modeled two proteins using the crystal structures of the related human protein. This involved comparing the sequences, carrying out sequence alignment and checking where they were different. The models were built and minimized using the XED force field, Cresset’s own unique force field that is ideal for protein work.

Modeling work revealed amino acid differences that the customer could do some experiments on. Cresset Discovery Services also predicted a few residues that they might want to mutate since they had an effect on the binding. This work confirmed that it was highly likely that they had correctly identified the inhibitor binding site, having predicted it from modeling.

Predicting a flavors and fragrance binding mechanism

A flavors and fragrances customer wanted to know more about the binding mechanism for a GPCR. They didn’t have the structure, but they did have some related crystal structures. Cresset Discovery was able to use these to carry out homology modeling using the XED force field to build a model of the chemical system.

The system turned out to be far more complicated that the customer had suspected. The insights that they gained from this model were born out by experiment, helping them to understand the way the binding mechanism worked. This made it far easier for them to make the right changes in order to optimize the effect of their compound.

Modeling the binding of an anti-malarial inhibitor

An academic group based in the UK had identified a protein synthesis inhibitor with potential as an anti-malarial.

Plasmodium is a malarial parasite that lives in mosquitos and infects humans. It carries a protein that enables the malaria virus to reproduce. The compound of interest is a plasmodium protein translation inhibitor that blocks the synthesis of the malaria virus.

Cresset Discovery built a homology model from a yeast molecule. From this, it was possible to work out where the molecule binds and what the shape is. This gave the basic information required to carry out a virtual screen.

Protein modeling

The three diverse projects above show how the XED force field applied to homology modeling can be highly effective for protein modeling, leading to insights and breakthroughs and making it possible to move projects on to the next step.