Na+-stimulated H+ movement and H+-stimulated Na+ uptake were studied in basolateral (blLPM) and canalicular (cLPM) rat liver membrane vesicles. H+ movement was monitored with the fluorescent amine acridine orange; 22Na uptake was assayed by a rapid Millipore filtration technique. In blLPM, inwardly directed Na+ gradients stimulated H+ efflux and outwardly directed Na+ gradients stimulated proton influx. Outwardly directed proton gradients (pH in 5.9/pH out 7.9) stimulated initial 22Na uptake rates 5- to 10-fold over pH-equilibrated conditions (pH in 7.9/pH out 7.9). Conversely, inwardly directed proton gradients (pH in 7.9/pH out 5.9) inhibited 22Na uptake. pH-dependent 22Na uptake was inhibited by amiloride and harmaline but not by other transport inhibitors, bumetanide, furosemide, 4,4'-diisothiocyano-2,2'-disulfonic acid stilbene, 4-acetamido-4'-isothiocyanostilbene-2,2'-disulfonic acid, and acetazolamide. Lithium also inhibited H+-stimulated 22Na uptake. Although a component of pH-stimulated 22Na uptake appeared to be dependent on membrane potential, this electrogenic component was amiloride insensitive. Proton gradient-stimulated 22Na uptake in blLPM was saturable, with a Km of 5.4 mM and a Vmax of 14 nmol . min-1 . mg prot-1. In contrast, in cLPM, no Na+ gradient-stimulated proton movement and no pH-dependent Na+ uptake occurred. These findings establish an electroneutral Na-H antiport in blLPM but not cLPM in rat liver. The polarity of this exchanger supports a model of bile formation that is dependent, in part, on canalicular HCO-3 and/or OH- excretion.