Unidirectional influx and efflux of choline into the syncytiotrophoblast were investigated from both maternal and fetal circulations of the perfused guinea-pig placenta by using a single-circulation paired-tracer (extracellular reference and test substrate) dilution technique. Cellular uptake of [3H]choline at 0.05 mM was (mean percentage +/- S.E. of mean, n = 14 placentae) 51 +/- 2 and 49 +/- 2, on maternal and fetal sides, respectively. Kinetics of unidirectional influx (0.05-4.0 mM-choline) indicated the existence of saturable and non-saturable components on both sides: on maternal and fetal interfaces the Km (mM) values were respectively, 0.12 and 0.13, the Vmax (mumol min-1 g-1) values, 0.08 and 0.07 and the apparent linear transfer constants (min-1 g-1) 0.11 and 0.12. Efflux of [3H]choline from the placenta back into the ipsilateral circulation (backflux) was generally fast (20-60% in 5-6 min) and asymmetric with the fetal: maternal ratio usually above unity. Transplacental specific choline transfer in the dually perfused placenta, when observed, was small (less than 10% of the injected dose) following tracer injections in either direction based on the 5-6 min collection of the contralateral circulation (at 0.05 mM-choline). Placental retention of [3H]choline at the end of the 5-6 min period was about 25% of the injected dose when the tracers were injected from either circulation. Analogues of choline such as hemicholinium-3, thiamine, ethanolamine and N,N-dimethylethanolamine inhibited choline unidirectional influx, whereas betaine and acetate had no effect. The absence of the normal sodium gradient (perfusate sodium was replaced by Tris or by lithium) did not inhibit choline transport. The metabolic inhibitors dinitrophenol (1.0 mM) and potassium cyanide (1.0 mM) were essentially ineffective (up to 40 min perfusion). The sulphydryl reagent N-ethylmaleimide did not appear to inhibit the influx, in comparison with its effect on [3H]choline backflux which was greatly accelerated, resulting in a dramatic reduction in placental net uptake of the label. Our findings show that choline transport into the placenta is a rapid carrier-mediated process occurring at both maternal and fetal sides of the trophoblast, at physiological blood concentrations. This cellular uptake is possibly related to the synthesis of acetylcholine, which is known to occur in human placental tissue. Specific transplacental transfer of choline was a very slow process under the conditions of our experiments and this contrasted with the observed fast and high uptake into the trophoblast.