Introduction: Transport across the hepatocyte canalicular membrane

THE HEPATOCYTE, the parenchymal cell of the liver, comprises about 90% of the cellular volume of this organ (1). The polarity of the hepatocyte is maintained by three distinct plasma membrane domains that can be recognized morphologically and by functional analyses (2-5). The highly specialized canalicular (or apical) membrane of the hepatocyte is rich in microvilli and comprises 10-15% of the membrane surface area of the hepatocyte; the smooth lateral hepatocyte membrane accounts for 15% of the surface area, and the microvillus-rich sinusoidal (or basal) membrane represents at least 70% of the hepatocyte surface (3). Many endogenous as well as xenobiotic substances are taken up into the hepatocyte by transport systems in the sinusoidal membrane (recently reviewed in ref 6), in part metabolized by monooxygenases and transferases in the hepatocytes, and subsequently secreted into bile by primary-active ATP-dependent export pumps located in the canalicular membrane (recently reviewed in ref 7). The hepatocyte canalicular membrane contains a unique set of these ATP-dependent export pumps, which enable transport against the concentration gradient of many substances, including bile salts and bilirubin glucuronides, between the hepatocyte and the fluid in the bile canaliculus. The unidirectional ATP-driven bile salt secretion across the canalicular membrane represents the rate-determining step in overall bile salt transport from the blood into bile and is a major driving force for the flow of biliary fluid (8). In addition, the propulsion of bile is promoted by the actin-myosin-containing system of the bile canaliculus (9). The concept of the forces driving the secretion of biliary fluid across the hepatocyte canalicular membrane has changed in recent years (6-8, 10). The electrochemical gradient across this membrane domain, which amounts to about -35 mV (11), was formerly considered the decisive driving force for the secretion of negatively charged compounds into the biliary tract (12-14). It has become evident, however, that this electrochemical gradient is insufficient to account for the concentration gradient for compounds, such as bile salts (11) and cysteinyl leukotrienes (15, 16), which can be at least 100fold. The elucidation of unidirectional ATP-dependent transporters, termed export pumps, has recently provided a major advance in hepatobiliary transport biology (7, 10). The first such export pump to be localized (17) and functionally characterized (18) in the hepatocyte canalicular membrane was the product of the multidrug resistance (MDR1) gene, also termed MDR1 P-glycoprotein. Subsequently, the ATP-dependent export pump for conjugates of lipophilic compounds with glutathione, glucuronate, or sulfate was functionally characterized in the canalicular membrane (19-2 1) and, more recently, cloned and sequenced (22, 23). In 1991, the ATP-dependent transport of bile salts across the canalicular membrane was discovered (24-26). The product of the murinemdr2 gene, corresponding to the human MDR2/3 gene, was recognized as the major P-glycoprotein in the canalicular membrane (27) functioning in the translocation (28, 29) and biliary elimination (30) of phospholipids. The ATP-dependent export pumps in the canalicular membrane are members of the ATP-hinding cassette (ABC)2 family of transporters and share common structural and functional characteristics, at least as far as the cloned and sequenced canalicular ABC transporters are concerned (Table 1). The relation between structure and function of ABC transporters has been most extensively studied for the MDR1 P-glycoprotein (3 1-33). The human MDRJ gene encodes an integral membrane glycoprotein of about 170 kDa composed of two homologous halves. Each half is predicted to contain six transmemhrane segments and to have an intracellular ATP-binding/utilizing domain. These ATP-binding domains are highly conserved among different ATP-dependent export pumps. Hydrolysis of the y-phosphate of ATP is essential for substrate transport. Extensive point mutational analyses, deletion and insertion analyses, and generation of chimeric molecules have led to the prediction that transmembrane domains 5, 6, and 11, 12 are major substrate binding sites that form