A basis for reduced chemical library inhibition of firefly luciferase obtained from directed evolution

Abstract

We measured the "druggability" of the ATP-dependent luciferase derived from the firefly Photuris pennsylvanica that was optimized using directed evolution (Ultra-Glo, Promega). Quantitative high-throughput screening (qHTS) was used to determine IC(50)s of 198899 samples against a formulation of Ultra-Glo luciferase (Kinase-Glo). We found that only 0.1% of the Kinase-Glo inhibitors showed an IC(50) < 10 microM compared to 0.9% found from a previous qHTS against the firefly luciferase from Photinus pyralis (lucPpy). Further, the maximum affinity identified in the lucPpy qHTS was 50 nM, while for Kinase-Glo this value increased to 600 nM. Compounds with interactions stretching outside the luciferin binding pocket were largely lost with Ultra-Glo luciferase. Therefore, Ultra-Glo luciferase will show less compound interference when used as an ATP sensor compared to lucPpy. This study demonstrates the power of large-scale quantitative analysis of structure-activity relationships (>100K compounds) in addressing important questions such as a target's druggability.

Figure 1. Summary of the qHTS

a) Results of the Ultra-Glo luciferase qHTS. Concentration-response data are shown for active (red; percentage of actives noted) and inactive (blue) compounds. b). Results of the previous qHTS using the lucPpy formulation is shown for comparison. Activities for 352 compounds (green) in the lucPpy qHTS were reduced because of a partially blocked tip. Percentage of actives noted in red. c) Control inhibitor (1) titration response curves and structure of 1 are shown. d) Representation of class 1a, 1b and 2a CRCs derived from the Ultra-Glo luciferase qHTS. Data points are shown in red with the fitted CRC curves shown in blue.

Figure 2. Difference in potency distribution between lucPpy and Ultra-Glo luciferase formulations

Active compounds from the previous lucPpy qHTS and the Ultra-Glo luciferase qHTS were compared and the ΔLogAC₅₀ values were calculated. Negative ΔLogAC₅₀ values represent compounds that were less potent in the Ultra-Glo luciferase compared to the lucPpy qHTS, while positive ΔLogAC₅₀ values represent compounds that were more potent in the Ultra-Glo luciferase compared to the lucPpy qHTS.

Figure 3. Comparison of inhibitor scaffold representation between lucPpy and Ultra-Glo luciferase formulations

Center pie charts represent the amount of active (darker grey slices) and related inactive analogs (light grey area in each pie chart) for prominent luciferase inhibitor scaffolds (2–6) for the lucPpy (top pie) and Ultra-Glo luciferase (bottom pie) qHTS. The scaffolds and the associated analogs were determined as described in Auld et al. The number of lucPpy active analogs was reduced from 70 to 27 (2, 3,5-diaryl-oxadiazoles), 98 to 23 (3, benzamides), 21 to 5 (4, benzoxazoles), 89 to 6 (5, benzthiazoles), 55 to 11 (6, benzimidazoles) in the Ultra-Glo luciferase qHTS. The corresponding number of structurally related inactive analogs increased from 1,192 to 1,442 in the Ultra-Glo luciferase qHTS.

Figure 4. Comparison lucPpy and Ultra-Glo luciferase against 2-phenylbenzothiazole luciferase inhibitors at multiple substrate concentrations

Graphs of ATP variation or luciferin variation (inset) are shown. Luciferin or ATP was varied at 0.25 μM, 2 μM, 25 μM and 250 μM resulting in four sets of CRCs. In each case the constant substrate was present at 250 μM. The lucPpy data is shown as solid circles with red CRC fits and the Ultra-Glo luciferase is shown as open circles with black fitted lines. The structures of the inhibitors assayed are also shown. The top graphs show compounds with luciferin competitive behavior against both lucPpy and Ultra-Glo luciferase, while the bottom graphs demonstrate inhibitors which maintain potency at high ATP or luciferin concentrations for lucPpy but not Ultra-Glo luciferase.

Figure 5. Comparison of lucPpy and Ultra-Glo luciferase inhibition for a series of quinoline analogs

Red fits, solid circles and squares are lucPpy, and black fits, open circles and squares are Ultra-Glo luciferase for 8 (squares) or 9 (circles) assayed using K_M levels of substrates. Also shown is the activity of these two compounds using Kinase-Glo for 8 (open triangles) and 9 (upside-down open triangles). The inset shows activity in the qHTS for a series of related quinolines in the lucPpy qHTS (solid circles, red-fits) and Kinase-Glo (open circles).