The purpose of this study was to evaluate the effects of a novel neuroactive steroid, Co 2-6749 (GMA-839; WAY-141839; 3alpha, 21-dihydroxy-3beta-trifluoromethyl-19-nor-5beta-pregnan-20-one), on gamma-aminobutyric acid(A) receptors in vitro and to define its anxiolytic-like effects and side effect profile in vivo. Co 2-6749 fully inhibited [(35)S]t-butylbicyclophosphorothionate binding in rat brain cortical membranes with an IC(50) value of 230 nM and in human gamma-aminobutyric acid(A) receptor subunit combinations of alpha1beta2gamma2L, alpha2beta2gamma2L, alpha3beta2gamma2L, alpha4beta3gamma2L, alpha5beta2gamma2L, and alpha6beta3gamma2L receptors (IC(50) values of 200, 200, 96, 2300, 210, and 2000 nM). Rats were trained in a Geller-Seifter operant conflict paradigm. Co 2-6749 caused a dose-related increase in punished responding with a minimum effective dose of 1.6 mg/kg, p.o., a wide therapeutic index relative to a decrease in unpunished responding and relative to ataxia, and no tolerance. Additionally, ethanol caused less than a 2-fold shift to the left in the dose-response function of Co 2-6749 in the rotorod procedure in rats. In a pigeon conflict paradigm, punished responding was maximally increased to 784% of vehicle control by 30 mg/kg, p.o., with a 2-h duration and no effect on unpunished responding at this dose. Similarly, punished responding in squirrel monkeys was maximally increased to 1774% of control by 10 mg/kg, p.o., with no effect on unpunished responding at this dose. With robust anxiolytic-like activity across species, a large separation between anxiolytic-like effects and sedation/ataxia, a minimal interaction with ethanol, a lack of tolerance, and apparent oral bioavailability, Co 2-6749 makes an ideal candidate for development as a novel anxiolytic drug.