Abstract It has been proposed that the phosphorylation of membrane proteins is the mechanism by which postsynaptic membrane permeability is controlled at catecholaminergic synapses. Part of the evidence in favour of this interpretation is the rapid turnover of protein-bound phosphate observed in isolated synaptosomal plasma membranes in vitro. In the present study the rate of incorporation of phosphate from [γ- 32P]adenosine 5′-triphosphate into the protein of synaptic membranes isolated from rat brain was found to be dependent upon the concentration of adenosine triphosphate. At 10 μ m, maximum phosphate incorporation occurred after 30 s but at 1 μ m the maximum was not reached until 5 min. The net amount of phosphate incorporated also rose with increasing adenosine 5'-triphosphate concentration, being lowest at 10 μ m and rising to reach a maximum at between 1 and 5 m m adenosine triphosphate. When the phosphorylation of individual synaptic membrane proteins was followed by gel electrophoresis and autoradiography even more pronounced differences were found between the two concentrations of adenosine triphosphate: at 10 μ m there was a rapid phosphorylation and a striking stimulation by cyclic adenosine monophosphate whereas at 1 m m the time-course of turnover was prolonged and stimulation by cyclic adenosine monophosphate was more modest. At both concentrations of adenosine triphosphate the phosphorylation of synaptic membrane proteins was not completely reversible in vitro. In both cases more than 50% of the incorporated phosphate was still bound to membrane protein after dephosphorylation had ceased. It seems probable that results obtained with micromolar concentrations of adenosine triphosphate are spurious, the phosphorylation process being prematurely terminated by the consumption of the substrate by non-kinase membrane adenosine 5′-triphosphatase activity. The longer time-course found with adequate levels of adenosine triphosphate (millimolar concentrations) is not incompatible with a transmission-dependent role for synaptic membrane phosphorylation. However, it does suggest that the phenomenon may be part of a more general metabolic response to afferent stimulation than has hitherto been proposed.