A high throughput screening assay system for the identification of small molecule inhibitors of gsp

Abstract

Mis-sense mutations in the α-subunit of the G-protein, Gsα, cause fibrous dysplasia of bone/McCune-Albright syndrome. The biochemical outcome of these mutations is constitutively active Gsα and increased levels of cAMP. The aim of this study was to develop an assay system that would allow the identification of small molecule inhibitors specific for the mutant Gsα protein, the so-called gsp oncogene. Commercially available Chinese hamster ovary cells were stably transfected with either wild-type (WT) or mutant Gsα proteins (R201C and R201H). Stable cell lines with equivalent transfected Gsα protein expression that had relatively lower (WT) or higher (R201C and R201H) cAMP levels were generated. These cell lines were used to develop a fluorescence resonance energy transfer (FRET)-based cAMP assay in 1536-well microplate format for high throughput screening of small molecule libraries. A small molecule library of 343,768 compounds was screened to identify modulators of gsp activity. A total of 1,356 compounds with inhibitory activity were initially identified and reconfirmed when tested in concentration dose responses. Six hundred eighty-six molecules were selected for further analysis after removing cytotoxic compounds and those that were active in forskolin-induced WT cells. These molecules were grouped by potency, efficacy, and structural similarities to yield 22 clusters with more than 5 of structurally similar members and 144 singleton molecules. Seven chemotypes of the major clusters were identified for further testing and analyses.

Figure 1. Compound Identification Flow Chart.

Depicted is a flow chart of the assays, filtering, and analyses that were performed to ultimately identify the 7 chemotypes of clusters of molecules that have been selected for further study.

Figure 2. Confirmation Assay Molecules.

A 3-axis plot of the 1356 compounds identified in the confirmation assay is shown. Compounds are sorted by curve class. Red: active compounds in curve class 1 and 2. Green: weakly active compounds in curve class 3. Blue: inactive compounds in curve class 4.

Figure 3. Clustering Analysis.

Active compounds were clustered based on structural similarity to identify common chemotypes using LeadScope (Leadscope Hosted Client, Leadscope Inc., Columbus, OH). The results show a diversity of structural clusters, with 22 distinct clusters with more than 5 members. (A) Representative compounds from each of the most prominent 7 clusters are shown. (see Table S5, Cluster Analysis Compounds with Link for a complete list of the 102 molecules in the 7 clusters, their structures, IC₅₀, and active link to the complete PubChem description). Their common structural scaffolds are highlighted in red. These scaffolds are highly polar, including thiazole, triazole, and hydrozide-based derivatives. Another common structural feature is that these small molecules share a linear molecular shape, which suggests that they might compete with GTP at the active site of the G protein. (B) Inhibition-concentration curves for 7 selected compounds, one from each cluster, together with the IC₅₀ for each compound are shown.