Investigating the influence of tracer kinetics on competition-kinetic association assays; identifying the optimal conditions for kinetic fragment screening

D. A. Sykes, S. J. Charlton. Cell Signalling Group, The University of Nottingham, Nottingham, UNITED KINGDOM.

Introduction

The importance of optimising drug-binding kinetics has led to an increase in the development and utilisation of assay-systems for measuring the kinetics of unlabelled compounds. One popular approach is the competition-association kinetic binding approach, first described by Motulsky and Mahan1. It is now accepted that a tracers kinetic characteristics can greatly effect the reliability of estimated kinetic parameters,2 an obstacle to successfully introducing kinetic assays earlier in the drug discovery screening-cascade. Using a simulation approach we have identified the optimal tracer characteristics for determining the kinetics of unlabeled ligands typically encountered during the different stages of a drug discovery program (i.e. rapidly-dissociating eg. koff = 100min-1 low-affinity “hits” through to slowly-dissociating eg. koff = 0.01min-1 high-affinity “candidates”).

Method

Monte Carlo simulations (200 per condition with an associated error of 1 SD) using an association kinetic binding model were performed in GraphPad Prism 6.0, with four model tracers, with off-rates ranging from 10-0.01 min-1. For simulation purposes assay read start-time was fixed at either 1 sec to mimic online addition of membranes via injectors, or 30 sec to mimic the delay in time to read following offline addition. Read interval-time (i.e. the time between well-reads) was varied between 1-60 secs. Further simulations were performed using the competition-association kinetic binding model to assess our ability to determine the kinetics of unlabeled compounds in competition with the model tracers.

Results

For more rapidly dissociating unlabeled ligands (eg. koff = 100min-1) the key to obtaining accurate kinetic parameters is to employ a tracer with a relatively fast off-rate (eg. 10min-1), utilizing online addition and a short read interval-time. Table 1 compares kinetic data obtained using online and offline addition protocols. Online addition also proved crucial for accurate parameter estimation of the most rapidly dissociating tracer examined (10 min-1). The potential to impose strict timing constraints is largely governed by sample injection capability and the method of detection employed (eg. TR-FRET versus radiometric).

Table 1. Summary of kinetic input and output parameters for a fragment-like compound using a rapidly-dissociating tracer (kon 3E7M-1min-1 -and koff 10min-1) at a fixed read interval-time of 5sec.

Conclusion

The insight into tracer binding presented has consequences for experimental design strategy and provides a framework for the identification and testing of tracers necessary for profiling rapidly dissociating low-affinity competitors, e.g. fragments.

References

1. Motulsky and Mahan (1984). Mol Pharmacol. 25(1):1-9.

2. Klein-Herenbrink C et al (2016). Nature Communications 7, 10842.