A virtual screening study towards novel noncompetitive antagonists of the metabotropic glutamate receptor 1 (mGluR1) is described. Alignment-free topological pharmacophore descriptors (CATS) were used to encode the screening compounds. All virtual hits were characterized with respect to their allosteric antagonistic effect on mGluR1 in both functional and binding assays. Exceptionally high hit rates of up to 26 % were achieved, confirming the applicability of this virtual screening concept. Most of the compounds were found to be moderately active, however, one potent and subtype selective mGluR1 antagonist, 13 (IC(50): 0.362 microM, SEM +/-0.031; K(i): 0.753 microM, SEM +/-0.048), based on a coumarine scaffold was discovered. In a following activity optimization program a series of coumarine derivatives was synthesized. This led to the discovery of potent (60, IC(50): 0.058 microM, SEM +/-0.008; K(i): 0.293 microM, SEM +/-0.022) and subtype selective (rmGluR5 IC(50): 28.6 microM) mGluR1 antagonists. From our homology model of mGluR1 we derived a potential binding mode within the allosteric transmembrane region. Potential interacting patterns are proposed considering the difference of the binding pockets between rat and human receptors. The study demonstrates the applicability of ligand-based virtual screening for noncompetitive antagonists of a G-protein coupled receptor, resulting in novel, potent, and selective agents.