BRD4 PROTAC Characterisation using SPR – Case Study
The bioscience division of Charnwood Molecular undertook a project on the assay development of a Surface Plasmon Resonance (SPR) assay to enable kinetic and affinity measurements of well-known BRD4-targeting PROTAC molecules. The data allows us to infer different mechanistic details of the PROTAC / BRD4 interactions.
We have previously combined a fluorescence-based thermal shift assay (FTSA) and cell-based protein degradation experiments (using our ‘Jess’ semi-automated Western blotting system) to demonstrate the binding and efficacy of several BRD4-targeting PROTAC molecules.
In this project we developed an SPR assay to measure the kinetics and affinity of compound binding to BRD4 – gaining insights into subtle differences in binding mode between similar molecules.
Using amine-coupling chemistry we could immobilise BRD4 readily to the surface of a CM5 chip and this generated a very robust surface for binding experiments. We measured the binding to BRD4 of drug-like small molecules and much larger PROTAC molecules, measuring kinetic and affinity data for both groups.
The small molecule affinities closely matched reported affinity/potency values from literature, validating our assay.
For the PROTAC molecules we saw differences in binding mode – the only differences between the PROTAC molecules being the linker regions and the ligase-targeting warhead. Both shared the same BRD4-targeting warhead. These results demonstrate how SPR can provide interesting and important characterisation of binding modes that might be assumed to be very similar.
You can see further explanation of our findings below.
BRD4 immobilised very readily to the surface of CM5 chips using the amine-coupling method.
A drastically shortened protein contact time was needed to ensure we didn’t capture too much protein.
The surface was very robust – fully stabilising around 1 hour post-immobilisation.
In the case of the BRD4-targeting warhead JQ1(+) we can measure very high quality sensorgrams for its interaction with BRD4.
At some of the higher concentrations of compound we see a deviation away from 1:1 binding and these sensorgrams have to be excluded to allow a good fit of the model to the data.
The subsequent kinetic fit (KD = 26 nM) shows good agreement with the affinity fit (KD = 24 nM).
Despite both sharing the same BRD4-targeting warhead, we see different binding modes suggested from these high quality SPR sensorgrams.
dBET6 is almost perfectly fit to by the 1:1 binding model whereas ARV-771 is a much less assured fit – and requires the removal of several higher-concentration sensorgrams to achieve this.