Here, we demonstrate the benefit of considering cellular substrate affinity when developing NanoBRET target engagement assays.
Aim
We wanted to assess target engagement of ATP-competitive binders to a family of three homologous ATPase proteins via whole cell NanoBRET assays
How does the NanoBRET target engagement technology work?
The technology uses Bioluminescence Resonance Energy Transfer (BRET) to measure compound engagement with the exogenous ATPase target. The cDNA encoding the full-length target ATPase, with the addition of a NanoLuc tag at the N-terminal, are transiently transfected into cells, typically HEK-293T, for 24-hours. When expressed, the protein is linked to the luminescence enzyme (NanoLuc). In the presence of a cell permeable NanoLuc-substrate, photons are generated and emitted at 460 nm (luminescence).
When cells are treated with a cell-permeable fluorescently tagged tracer, binding of the tracer to the ATP binding site of the target protein brings the tracer in close proximity to NanoLuc (Luciferin). This generates photons and excites the fluorescent tag, resulting in fluorescence emission at 600+ nm and a BRET signal. Compound engagement is measured in a competitive format of tracer versus compound, typically at equilibrium. Binding of the test compound results in a loss of NanoBRET signal between the target protein and the tracer inside living, intact cells (Figure 1). Hence the compounds compete in a cellular environment of physiological ATP concentration and pH.
What We Did...
Two competent cell NanoBRET assays were successfully developed for two members of the ATPase protein family.
Addition of a tracer, an ATP-competitive molecule linked to a fluorophore that fluoresces when excited by the NanoLuc-Substrate interaction, specifically generated a robust BRET ratio signal. A good signal window could be achieved for ATPase-I and -II but not for ATPase-III.
Despite various optimization attempts, ATPase-III failed to show any significant increase in BRET ratio upon the addition of the tracer (Figure 2).
For ATPase-I, compounds were able to displace the tracer and, expectedly, reduce BRET ratios in a concentration dependent manner. Compounds could be ranked in order of potency (Figure 3).
Challenge & troubleshooting
To troubleshoot the lack of signal in the ATPase-III assay, we assessed the separate components that make up the BRET ratio: the fluorescence and the luminescence. We observed minimal luminescence counts for ATPase-III. This suggested either lack of expression of the tagged protein or that the NanoLuc substrate couldn’t access the NanoLuc enzyme. As such the tracer was not excited, and consequently no BRET could be detected.
ATPase-III is reported to bind as heterodimers with ATPase-I, which could be responsible for the lack of luminescence. For example, without its dimer partner, ATPase-III may be unstable or form homodimers, somehow rendering the NanoLuc inaccessible to its substrate. Due to the dual affinity of the tracer for both ATPases -I and -III, we developed a co-expression system. ATPase-III-NanoLuc was expressed alongside an untagged ATPase-I, where we could observe strong luminescence. We, however, failed to detect any significant increase in BRET ratio with the addition of tracer. This suggested there was an issue with the cellular affinity of the tracer – despite its in vitro biochemical characterization and acceptable affinity against both targets. The same tracer is used for ATPase-I and -II ruling out the notion that the tracer is unable to enter cells.
We then considered that the ATPases may have different affinities for the endogenous substrate ATP, and this may affect tracer binding. With high affinity for cellular ATP, ATPase-III would then effectively behave as a sink for endogenous ATP.
To test this theory, digitonin was used as a mild detergent to permeabilize cellular membranes and liberate cellular ATP, including from the active site. This allowed tracer binding and subsequent compound displacement to produce dose-response curves.
The addition of digitonin to the cells effectively enabled tracer binding. Compounds could then displace the tracer to reduce BRET in a concentration-dependent manner. We could therefore achieve a target engagement assay for ATPase-III, in which we could screen compounds (Figure 4).
This permeabilized assay set up enabled compound ranking but did not establish why digitonin enabled BRET to occur.
Examining the relationship between cellular ATP and exogenous target affinity:
Cell membrane permeabilization, using digitonin, leads to the depletion of cytosolic cellular components, including endogenous ATP. As such we reasoned that it could thereby liberate endogenous ATP from the active site to enable either compound and/or tracer binding and thus their competition for the active site.
To examine this, we spiked ATP into the permeabilised NanoBRET assay.
Spiking of ATP at 1mM prevented tracer binding (red line drop to black line) and thus compound displacement of tracer (blue line to green line) (Figure 5).
Investigating the ATP affinity of each ATPase
We titrated ATP into the permeabilized assay set up for both ATPase I and III and for the heterodimer ATPase-I:ATPase-III to determine an EC50 value for ATP in all conditions.
Cells were transfected with the NanoLuc-tagged ATPases singularly and then as a pair, to form the heterodimer. Cells were then subjected to the addition of digitonin to permeabilize them, in addition to the tracer at a set concentration. A titration series of ATP was then spiked into the cells with BRET as the readout. In theory, the digitonin is used to liberate “native” intracellular ATP and then the addition of exogenous ATP at different concentrations should prevent tracer binding in a concentration dependent manner, to allow us to determine an EC50 for ATP.
This allowed us to determine an intracellular ATP EC50 for the ATPases, singularly and as a pair.
These experiments demonstrated that ATPase-III alone had a much greater affinity for ATP with an EC50 of 0.265 µM but an EC50 of 1.04 µM when paired with its heterodimer partner. ATPase-I alone had a EC50 of 15 mM for ATP. This suggests that ATPase-III may be constitutively ATP-bound, with high affinity in its endogenous state. This may therefore explain why a whole cell assay could not be successfully developed for this target unless cellular ATP could be depleted following cell permeabilization (Figure 6).
Figure 6. The effect of ATP on the binding of the tracer in different permeabilized cell contexts – NanoLuc-tagged ATPase-I (red line), ATPase-III (black line), and ATPase-I:ATPase-III (blue line) were transiently expressed in HEK293T cells. Digitonin at a set concentration was used to liberate the native endogenous ATP in the cell. Tracer was added at a set concentration. ATP was then titrated into the cells across a 11-point titration series. The BRET signal was then determined. A EC50 value for ATP was also determined from each curve.
Investigating the ATP affinity of each ATPase
Using the EC50 values of ATP for each ATPase, we can determine a value for ATP occupancy at different ATP concentrations for each ATPase (Figure 7).
This enabled us to determine that ATPase-III, when alone or paired with ATPase-I in a heterodimer, reaches a higher percent occupancy at lower concentrations than that of ATPase-I alone. This could be because of the differing EC50 values above. This demonstrates that ATPase-III has a greater affinity for cellular ATP and is likely to be constitutively ATP-bound.
High affinity, together with low turnover, may both be responsible for the lack of success in establishing a competent cell NanoBRET assay for target ATPase-III.
Determining the Kd of ATP for each ATPase:
Based on a typical Kd calculation, using the IC50 from the ATP titration series, the EC50 value from a tracer dilution series and the concentration of tracer at that EC50 value, we could calculate the Kd of ATP for each ATPase (calculation below).
This determined a Kd of 0.254 µM for ATPase-III and 467 µM for ATPase-I.
Summary
These studies led to the successful development of two competent cell target engagement assays against two targets from a homologous ATPase family. In addition, a permeabilized cell target engagement assay was developed against a third novel ATPase target. Furthermore, we revealed an understanding of relative endogenous ATP affinities of the three homologous ATPases informing their likely druggability.
