Measurement of light-dependent GTPase (EC 3.1.5.1) activity in a paradigm guided by electrophysiological experiments was used to examine the involvement of a guanine nucleotide binding protein in fly phototransduction. Cell-free membrane preparations of Musca eyes responded to blue light by a 10- to 20-fold increase in GTP-hydrolyzing activity. This light-dependent GTPase had a low Km for GTP (0.5 microM) and was effectively inhibited by guanosine (5'----O3)-1-thiotriphosphate and guanosine 5'-[beta-gamma-imino]triphosphate but not by adenosine 5'-[beta-gamma-imino]triphosphate and ATP. The action spectrum of GTPase activity measured with intense light resembled closely the photoequilibrium spectrum of metarhodopsin. After illumination with blue (less than 480 nm) light, which converted rhodopsin to metarhodopsin, the GTPase remained highly active for at least 60 min in the dark. Similarly, rhodopsin-to-metarhodopsin conversion in intact cells induced a prolonged excitation in the dark, known as the prolonged depolarizing afterpotential (PDA). The persistent GTPase activity (like the PDA) was suppressed to the low basal activity of the unilluminated membranes after conversion of metarhodopsin to rhodopsin with red light (greater than 570 nm), whereas during illumination with red light, some GTPase activity was maintained. The magnitude of the persistent GTPase activity in the dark, like the PDA, depended in a supralinear manner on the amount of pigment conversion. Thus, the dependence of GTPase activity of Musca membrane preparations on photopigment conversion resembles the induction and suppression of the PDA measured in intact photoreceptors of Musca. These findings indicate that a guanine nucleotide binding protein is part of the chain of events leading to both the generation of the receptor potential and the PDA.