Determining whether bespoke tracers work with ALK5 in nanoBRET target engagement assay

Special thanks to:

David Drewry – Helped with designing the nanoBRET tracers

M4K pharma OICR chemist team – Synthesised the M4K1046-linker derivatives

Carrow Wells (UNC) – Conjugated the M4K1046 to nanoBRET fluorophore

Background:

I have always wanted to establish nanoBRET target engagement assay for ALK5. Relative to dual luciferase promoter assay and immunofluorescent staining, nanoBRET is many times faster. With a robust ALK5 nanoBRET assay, I will be able to rapidly screen for cellular off-target activity.

Since none of the commercial nanoBRET tracers worked with ALK5, we attempted to generate our own tracers. We chose to create these bespoke tracers based on M4K1046 because it has cellular IC50 of ~50nM for ALK5. David Drewry has helped to design linkers that will attach fluorophores (nanoBRET energy acceptor) to the solvent-facing end of M4K1046. Based on known structures, additional bulk in this region should not hinder the binding of the compounds to ALK5. Two versions of M4K1046 (with different linker length) were synthesised by the M4K pharma chemist team in OICR. They were subsequently sent to Carrow Wells for conjugation to nanoBRET fluorophore.

Experimental design:

I wanted to determine the binding of these tracers to ALK5-nanoluciferase fusion. Therefore, I incubated increasing concentrations of tracer with HEK293 cells expressing ALK5-nanoluciferase fusion protein. If the tracers can bind nicely to the ATP pocket of ALK5, incubation with increasing concentrations of tracer will result in increasing BRET signal (wavelength = 610 nm). The resulting magnitude of BRET and EC50 estimated from the curve are good indicators to whether these tracers can be used for nanoBRET assay with ALK5.

To identify any background signal, I have included a replicate of the above experiment with 10,000nM of parent M4K1046. These unlabelled compound will saturate ALK5 ATP pockets and prevent the binding of nanoBRET tracers. All signal from this set of experiment is contributed by non-specific background.

To be eliminate doubts of any technical or reagent issues with the nanoBRET experiment itself, I have replicated both experiments above, substituting ALK5 with ALK2.

Results:

Conclusion:

Both bespoke tracers cannot be used for ALK5 nanoBRET. Both of them did not bind to ALK5. The nanoBRET experiment itself was a success since both tracers worked well with ALK2 (over 20 folds assay windows). The background noise was low for both tracers. ALK5 seems to intrinsically not tolerate bulky additions on the solvent end of M4K1046. This might be additional avenue for improving ALK2 vs ALK5 selectivity. At the very least, we have gained this additional information from these efforts.

For additional experimental details, please refer to my Zenodo post.