Simon et al. (J. Clin. Invest., 70 (1982) 401) studied cholate binding to crude liver plasma membrane vesicles and suggested that the binding may represent mainly the binding to the receptor (carrier) on the canalicular membrane. This hypothesis was supported by finding a good correlation between the number of cholate binding sites on liver plasma membrane and the maximal rate of biliary secretion (Tm) for taurocholate. We studied bile acid binding to sinusoidal and canalicular membrane vesicles isolated from rat liver by a rapid filtration technique. Scatchard analysis of the saturation kinetics showed both [3H]cholate and [3H]chenodeoxycholate bind to two classes of binding site on each membrane. However, little difference was observed between the binding to sinusoidal and canalicular membrane vesicles for each bile acid (cholate, Kd1 = 10.4 and 19.8 microM, n1 = 31.0 23.6 pmol/mg protein, Kd2 = 1.32 and 1.73 mM, n2 = 13.1 and 23.4 nmol/mg protein; and chenodeoxycholate, Kd1 = 0.207 and 0.328 microM, n1 = 36.7 and 27.4 pmol/mg protein, Kd2 = 1.16 and 2.26 mM, and n2 = 20.6 and 24.2 nmol/mg protein; numbers show the mean values sinusoidal and canalicular membrane vesicles, respectively). Chenodeoxycholate binding to sinusoidal membrane vesicles was markedly inhibited by cholate but not by Rose bengal, an organic anion dye. These studies indicate that both membranes (sinusoidal and canalicular membrane vesicles) have two kinds of binding site for bile acids, although no clear difference in the binding properties was observed between the two membranes. Consequently, the cholate binding Simon detected may represent the binding not only to canalicular membrane vesicles but also to sinusoidal membrane vesicles.