mechanism of secretion of biliary...

Mechanism of Secretion of Biliary Lipids

I. ROLEOF BILE CANALICULARANDMICROSOMAL

MEMBRANESIN THE SYNTHESIS ANDTRANSPORT

OF BILIARY LECITHIN ANDCHOLESTEROL

DANELH. GREGORY,Z. R. VLAHCEVIC, PETERSCHATZKI, andLEONSWELL

From the Veterans Administration Hospital and the Department of Medicine(Division of Gastroenterology), Pathology and Biochemistry, VirginiaCommonwealth University, Health Sciences Division,Richmond, Virginia 23249

A B S T R A C T The role of bile canalicular and micro-somal membranes in the synthesis and transport of bili-ary lipids was investigated by using the isolated per-fused rat liver model. Labeled lecithin precursors ([8H]-palmitic acid, [14C]linoleic acid, [8Hjcholine, and'PO4) and a cholesterol precursor ([3H]mevalonicacid) were administered with and without sodium tauro-cholate. The incorporation pattern of these labeled pre-cursors into linoleyl and arachidonyl lecithins and cho-lesterol fractions of microsomes, bile canaliculi, and bilewere examined at 30-min intervals up to 90 min.Marker enzymes and electron microscopy indicated thatisolated subfractions of plasma membranes were enrichedwith bile canaliculi (< 10% microsomal contamination).Taurocholate significantly stimulated the incorporation of'"POn, ['H]choline, ['H]palmitic acid, and ["4C]linoleicacid into linoleyl and arachidonyl lecithin with parallelincorporation curves for microsomal and bile canalicularmembranes throughout the 90-min study period. Duringthe 30-60-min period, however, these same lecithin frac-tions in bile significantly exceeded the specific activityof the membrane lecithins. The enzyme CDP-cholinediglyceride transferase was virtually absent from cana-liculi relative to microsomes, indicating that canaliculilack the capacity for de novo lecithin synthesis. In-corporation of [3H]mevalonic acid into membranous andbiliary cholesterol followed a pattern similar to that for

Presented in part at the American GastroenterologicalAssociation Meeting, San Francisco, California, May, 1974,and in part at the Bile Acid Meeting (III), Freiburg,Germany, June, 1974.

Received for publication 8 April 1974 and in revised form9 September 1974.

lecithin. These data provide evidence that (a) biliarylecithin and cholesterol are derived from a microsomalsubpool regulated by the flux of enterohepatic bile acids,(b) the role of the bile canalicular membranes with re-spect to biliary lipids is primarily transport rather thansynthesis, and (c) lecithin and cholesterol are trans-ported together from microsomes to bile. The findingsare consistent with the existence of a cytoplasmic lipidcomplex within the hepatocyte which is actively involvedin the intermembrane transport of biliary lipid.

INTRODUCTIONIt is now recognized that bile salts play an importantrole in the regulation of biliary phospholipid and cho-lesterol secretion (1-7). When the hepatic influx of bilesalts is low as reported in man and animals with an in-terrupted enterohepatic circuit (8, 9), there is an associ-ated decrease in the secretion of biliary lipids. However,the reduction in phospholipid secretion exceeds choles-terol so that the resulting bile has a high cholesterolconcentration relative to lecithin and bile salts. Thiscombination of events has been associated with cho-lesterol gallstone formation (10-12). Conversely, whenthe hepatic influx of bile salts is increased as observedafter the oral administration of bile salts (8, 13-15),biliary lipid secretion is enhanced and the resulting bilehas reduced concentrations of cholesterol and lecithinrelative to bile salts. While these observations supportthe concept of a homeostatic relationship between theenterohepatic flux of bile salts and the secretion ofbiliary lipids, the precise hepatocellular site where bilesalts exert their effect on biliary lipid metabolism hasnot been identified and data are not yet available re-

The Journal of Clinical Investigation Volume 55 January 1975-105114 105

garding the mechanism by which bile salts increasebiliary lipid secretion.

Biliary lecithin is composed mainly of the palmityl-linoleyl species which is synthesized in the liver via theCDP-choline diglyceride pathway (3, 8, 16, 17). How-ever, the predominant hepatic species of lecithin is richin arachidonic acid and is derived by transacylation oracyl exchange in a particular lecithin via the phos-phatidyl ethanolamine pathway (17-21). The studies ofVictoria, Van Golde, Hostetler, Scherphof, and VanDeenen (22) and Sarzala, Van Golde, De Kruyff, andVan Deenen (23) indicate that hepatic microsomes maybe the only hepatic membrane with CDP-choline trans-ferase activity, which is essential for de novo lecithinsynthesis. Balint, Beeler, Kyriakides, and Treble (8)have postulated the existence of a microsomal pool oflinoleyl lecithin destined for bile. On the other hand, therole of bile canalicular membranes with respect to bili-ary lecithin metabolism has not been investigated.

The present study critically examines the synthesis oflecithin and cholesterol in microsomal and canalicularmembranes during the phase of active biliary lipid se-cretion. Labeled precursors were administered to theisolated perfused rat liver to label the choline, fatty acid,and phosphate moieties of the lecithin molecule whileradioactive mevalonic acid was given as a cholesterolprecursor. Bile canalicular membranes, microsomes, andbile were then examined for radioactive labeling oflecithin and cholesterol and for CDP-choline diglyceridetransferase activity in an effort to define the role of themembranes with respect to biliary lipid metabolism.Additionally, data are presented which further elucidatethe role of taurocholate on the synthesis and transportof the biliary micelle.

METHODSLiver perfusion. Male Wistar rats weighing approxi-

mately 400 g maintained on Purina rat chow and tap H20ad libitum were used as liver donors and as a blood source.The common bile duct was cannulated with no. 10 PEtubing and the liver isolated surgically by a modificationof a procedure described earlier (3, 24). A glass cannulawas inserted into the portal vein and immediately attachedto a reservoir containing an oxygenated solution of 5% redblood cells in saline. A glass outflow cannula was theninserted into the vena cava. This procedure reduced thecritical period of liver anoxia to less than 1 min. The liverwas then removed, attached to the perfusion system, andperfused with Krebs-Ringer bicarbonate (pH 7.4 medium)containing 3% bovine albumin, 0.1% glucose, 0.3 ml heparin,and 25% washed rat red blood cells. The total perfusionvolume was 100 ml and was continuously recirculated. Theplatform on which the liver rested was adjusted to give ahydrostatic head of pressure of 21 cm water. The flow ratethrough the liver was 60-75 ml per min. During the entireperfusion period the liver maintained a good appearanceand bile production averaged 1-11 ml/h. An initial 20-minperfusion period was used to establish uniform blood andbile flow rates before the start of the experiments. Two

types of liver perfusion experiments were carried out. Inone series of experiments a priming dose of 2 mg sodiumtaurocholate (Calbiochem, San Diego, Calif.) was added tothe perfusate and then sodium taurocholate (10 mg/h, 1.6mg/ml) was infused at a constant rate throughout theduration of the experiment. Previous studies have shownthat the isolated perfused rat liver is in a steady state withrespect to bile salt secretion of an infusion rate of 10 mg/h(3). In the other series of liver perfusions, saline wassubstituted for taurocholate. After a 90-min taurocholate orsaline infusion period, the labeled precursors were added tothe perfusion reservoir in 10 ml of perfusion medium. Theprecursors were 200 uCi of [9,10-3H]palmitic acid, 50 AGCiof [1-'4C]linoleic acid, 50 AGCi of [methyl-3H]choline, 500,uCi of [3P]phosphate, and 100 mCi of [5-3H]mevalonicacid. The labeled compounds were obtained from NewEngland Nuclear, Boston, Mass. After the addition of thelabeled precursors the livers were perfused for an addi-tional 30, 60, or 90 min. The saline control livers (withouttaurocholate) were perfused for 60 min after the pre-cursors were added. Bile samples were collected at 30-minintervals throughout the experiments before and after theaddition of the labeled compounds. At the end of the per-fusion period, the livers were weighed and immediatelyplaced in chilled 1 mMbicarbonate buffer (pH 7.5) main-tained at 20C.

Membrane isolation. Bile canalicular membranes wereisolated according to the procedure of Song, Rubin, Rifkind,and Kappas (25). Livers (weighing 10-12 g) were addedto 10 vol of 1 mMbicarbonate and then homogenized witha Dounce homogenizer. The homogenate was centrifugedat 1,500 g for 10 min. Residue from this centrifugationcontained debris, nuclei, and plasma membranes while thesupernatant solution included mitochondria and microsomes.Sucrose was added to the supernatant solution to give aconcentration of 0.25 M. This solution was then fractionatedby ultracentrifugation into mitochondria (10,000 g for 30min) and microsomes (104,000 g for 60 min). Microsomeswere washed once with 0.25 M sucrose and centrifuged at104,000 g for 30 min. Bile canalicular membranes were iso-lated from the residue of the first 1,500-g centrifugation ona discontinuous sucrose density gradient as outlined by Songet al. (25). The canalicular membranes (tight junction com-plexes) appeared in the upper 1.16 sucrose density layerand the remainder of the liver plasma membranes werelocalized in the 1.18 sucrose density fraction. Criteria forascertaining the purity of the membrane fractions includedthe marker enzymes glucose-6-phosphatase (26), 5'-nucleo-tidase (27), and Mg++-ATPase (28). The bile canalicularpreparations were also examined by electron microscopy.Morphologic criteria for identifying intact bile canaliculiincluded the presence of adj acent plasma membranes ad-joined by tight junctional complexes and associated withmicrovilli.

Other methods. The protein content of the membranefractions was determined by the method of Lowry, Rose-brough, Farr, and Randall (29). The bile and liver mem-brane fractions were extracted with 20 vol of 2: 1 chloro-form-methanol and partitioned with 0.2 vol of water (30).Free cholesterol and phospholipids were isolated by silicicacid column chromatography (31) and the free cholesterolfraction assayed for mass by gas liquid chromatography(32). The cholesterol 3H activity was determined by add-ing carrier-free cholesterol and precipitating the mixture asthe digitonide. After dissolving the digitonide in methanol

106 D. H. Gregory, Z. R. Vlahcevik, P. Schatzki, and L. Swell

the 'H activity of the digitonide was determined by liquidscintillation counting.

Phospholipid fractions were separated into the majorphospholipid classes by thin-layer chromatography on silicagel H impregnated with 1.0 mMsodium carbonate (33)with a solvent system of chloroform-methanol-acetic acid-water (50: 25: 8: 4). Phospholipid standards were bandedon either side of the plate. The thin-layer plate wassprayed with 2,7-dichlorofluorescein and the phospholipidzones were visualized under ultraviolet light. The phospha-tidylcholine zone was scraped from the plate and elutedthree times from the silica gel with 5 ml of 50: 39: 1:10chloroform-methanol-acetic acid-water. The eluants werepooled and washed successively with 6 ml of water, 6 mlof NH40H, and 6 ml of 50% methanol as described byArvidson (34). This washing procedure removed the 2,7-dichlorofluorescein dye and the phosphatidylcholine fractionwas retained in the chloroform phase. The phosphatidyl-choline fraction was then further fractionated into the majorlecithin species by thin-layer chromatography on AgNO3-impregnated plates and the system developed with 60: 30:5chloroform-methanol-water (34). Linoleyl- and arachidonyl-rich lecithin fractions were eluted three times from thesilica gel with 50: 39: 1: 10 chloroform-methanol-aceticacid-water. The eluants were pooled and washed successivelywith 6 ml water, 6 ml NH40H, and 6 ml 50% methanol(containing 0.5 NaCl). This washing procedure removedthe 2,7-dichlorofluorescein dye and small amounts of residualsilver ion. The thin-layer argentation procedure resolvedthe phosphatidylcholine fraction into three major zonesbased on the degree of fatty acid unsaturation of the phos-phatidylcholines. These zones were rich in linoleic, arachi-donic, and long-chain polyunsaturated fatty acids. Gas liquidchromatographic analysis (32) of these fractions indicatedthat the arachidonyl lecithin fraction was devoid of linoleicand other polyunsaturated fatty acids and that arachidonicacid was absent from the linoleyl fraction. The fractionrich in linoleyl lecithin contained 8-10% monounsaturatedfatty acids. Phospholipid phosphorus was determined onthe fractions by the method of Bartlett (35).

Aliquots of the fractions containing linoleyl lecithin weresubjected to mild alkaline hydrolysis with 5% KOH at75°C for 2 h to remove the fatty acid portion of the mole-cule. The hydrolysis mixture was then acidified to pH 1with 7% HCI and extracted three times with diethyl ether.The labeled fatty acids (['Hjpalmitic and ["C] linoleic)were extracted into the diethyl ether phase; the aqueouslayer contained labeled glycerylphosphoryl choline (['2P]-and ['H]choline). The radioactivity of these fractions wasthen determined by liquid scintillation counting. The hydro-lytic and isotopic separation techniques were validated byseveral types of experiments. Labeled phosphatidylcholinewas prepared biosynthetically by administering labeled pre-cursors ( [3Hpalmitic and [14C] linoleic or ['HIcholine) toisolated perfused rat livers and then isolating the micro-somal fraction rich in linoleate. This gave two types oflabeled phosphatidylcholines: one with [3H]palmitic and['4Cllinoleic and the other with ['Hlcholine. Gas liquidradiochromatography (32) of the methyl fatty acid estersobtained from the phosphatidylcholine fraction with thelabeled fatty acids indicated that virtually all (97%) ofthe 'H and "C activities were in palmitic and linoleic acids,respectively. The phosphatidylcholine fraction labeled withfatty acids was also used to check the hydrolytic procedurefor quantitative recovery of the labeled fatty acids and foroverlap of radioactivity between the diethyl ether and aque-

ous layers. Complete recovery of the radioactive fatty acidsin the diethyl ether fraction was obtained and no fatty acidradioactivity was found in the aqueous layer. Also thin-layer chromatographs of the lipids in the diethyl ether phaseshowed only free fatty acids and no radioactive phospha-tidylcholine. The phosphatidylcholine fraction labeled with[3H] choline was used to check for recovery of glyceryl-phosphoryl choline radioactivity into the aqueous layer.After hydrolysis all of the 'H activity was found in theaqueous layer and none in the diethyl ether phase. Radio-active recovery experiments were also carried out singlyand with mixtures Of '2PO4 and [3H]choline or with ['H]-palmitic and [14C] linoleic acids. The results also indicatedquantitative recovery and separation of the isotopic labels.The radioactivity of the fatty acids (3H and 14C) andglycerylphosphoryl choline ('P and 3H) moieties of thephosphatidylcholine fraction was determined by liquid scin-tillation counting (Isocap 300, Searle Analytic Inc., DesPlaines, Ill.). External standardization ("'Ba) was used tocorrect for quench. The fatty acid fraction was countedin 10 ml of Liquifluor and the glycerylphosphoryl cholinefraction (1 ml) in 15 ml Aquasol (New England Nuclear).

CDP-choline diglyceride transferase activity was deter-mined according to the method of De Kruyff, Van Golde,and Van Deenen (21). The 1,2-diglyceride substrate wasprepared from liver lecithin treated with phospholipase C(Bacillwu cereus). The enzyme digest mixture contained200 nmol of CDP-choline ("C-methyl, New England Nu-clear), 18 Amol Mg", 8 usmol glutathione, 0.4 mg Tween20, 2 mg of 1,2-diglyceride, and 0.5-1 mg of membraneprotein in tris buffer (pH 7.2) and was brought to a finalvolume of 1 ml. Incubation time was 20 min at 370C.10 vol of 2: 1 chloroform-methanol was used to stop thereaction. The phospholipid fraction was isolated by silicicacid column chromatography (30) and assayed for 14Cactivity.

CALCULATIONSThe specific activities of the phosphatidylcholines were cal-culated from the radioactivities ('[P2P, ['HIcholine, ['H]-palmitic, ["C]linoleic) and the mass obtained from phos-pholipid phosphorous. The values are expressed as disinte-grations per minute per micromole phosphatidylcholine.The microsomal and bile canalicular membrane fractionsrich in linoleyl lecithin were found to have 10% less pal-mitic and 10% more stearic acids than biliary linoleyl leci-thin. The specific activities of the linoleyl lecithin-rich frac-tion with respect to ['HIpalmitic was therefore correctedto a standard 50% palmitic acid to reflect comparability be-tween the linoleyl lecithin fractions of the membranes andbile. This correction was valid since all of the fatty acid'H activity was associated with palmitic acid.

Data were analyzed by the paired t test to determine thesignificance of isotope incorporation into the various lipidand membrane fractions by the same liver. The group ttest was used for comparing data from different group oflivers (i.e., taurocholate-perfused livers vs. saline controls)(36).

RESULTSEnzyme activities and electron microscopy of the

liver cell membranes. The relative purity of each mem-brane fraction was determined by marker enyzme ac-tivity and electron microscopy. As shown in Table I,5'-nucleotidase and Mg++-dependent ATPase were found

Bile Canalicular Membranes and Biliary Lipid Formation 107

TABLE I

Enzyme Activity Units* of Liver Cell Membranes

Activity

BiliaryEnzymes Microsomes canaliculi

Glucose-6-phosphatase 83 4 16 9± 15'-Nucleotidase 50±8 516±t125Mg+-ATPase 374±5 444496CDP-CDT

Preparation (1) 1.46 0.008(2) 2.52 0.007

* Represents the average ±SD of cell preparation. Eachpreparation was a pool of three livers with all assays per-formed in duplicate. Enzymatic activity for glucose-6-phos-phatase, 5'-nucleotidase, and Mg++-ATPase is expressed asnanomoles P per milligram protein per minute. Enzymaticactivity for CDP-choline 1,2-diglyceride transferase (CDT)expressed as nanomoles CDP-choline incorporated intolecithin per milligram protein per minute.

almost exclusively within the canalicular membranefraction. As previously reported (25) these enzymeswere concentrated primarily in the plasma membranesof the hepatocyte. The specific activity of glucose-6-phos-phatase in microsomes was more than nine times higherthan in bile canaliculi indicating that there was verylittle contamination of the canalicular preparation withmicrosomal membranes. Also shown in Table I is thefractional distribution of CDP-choline diglyceride trans-ferase activity. Relative to microsomes, this enzyme wasvirtually nonexistent in bile canalicular membrane prepa-rations. Comparison of the relative distribution patternsfor CDP-choline diglyceride transferase with glucose-6-phosphatase (Table I) shows a higher relative activityof the latter in bile canaliculi. It would thus appear fromthese data that CDP-choline diglyceride transferase maybe a more sensitive indicator for ascertaining the purityof canalicular membrane preparations. While thesemarker enzymes are helpful in distinguishing plasmafrom microsomal membranes, they are not specific fordistinguishing the portion of plasma membranes whichform bile canaliculi. Fig. la is a representative low-power electron photomicrograph of a typical bile cana-licular preparation which contains a virtually homogene-ous collection of intact bile canalicular membranes re-sembling their original in situ morphology in hepato-cytes. Fig. lb is a higher-power magnification of atypical intact bile canaliculus and clearly demonstratesmembranous layers surrounded by characteristic fibril-lar cytoplasm joined together to form a classical tightjunction which then separates and extends into a micro-villous pattern.

Lipid synthesis and transport. To establish the op-timal conditions for studying the synthesis and secretion

of biliary lecithin and cholesterol, sodium taurocholatewas infused into the liver system for 90 min before theaddition of the labeled precursors; the taurocholate in-fusion was then continued throughout the study periods.Without taurocholate, biliary lipid secretion was ex-tremely low (phospholipid 0.2 /Amol/h and cholesterol0.05 Azmol/h). Addition of taurocholate to the systemresulted in the immediate appearance of phospholipid andcholesterol into bile. During a 90-min perfusion periodthe average secretion rate was 2 ,mol/h for phospholipidand 0.2 ;imol/h for cholesterol. These data confirm ourearlier findings (3) and emphasize the necessity of bilesalts for biliary lipid transport.

Fig. 2 shows the incorporation of ['H]palmitic and["C]linoleic acids into linoleyl lecithin. Incorporation of["C]linoleic and ['H]palmitic acids into linoleyl lecithinof the membranes was very rapid. The highest recordedspecific activity was obtained in approximately 30 min.At that time the specific activity relationships with re-spect to [8H]palmitic indicated that the linoleyl lecithinof microsomes had a significantly (P < 0.01) higherspecific activity than linoleyl lecithin isolated from bilecanalicular membranes. There were no significant dif-ferences in the ["C]linoleic specific activities of linoleyllecithin at 30 min. During the second study period (30-60 min) specific activities of ["C]- and [8H]-fatty acidsfrom biliary linoleyl lecithin exceeded that of both mi-crosomal and biliary canalicular membranes. These spe-cific activity differences between bile and the membranefractions were most pronounced by the last studyperiod (60-90 min) for [3H]palmitic (P <0.01) andduring the 30-60-min biliary collection period for ['C]-linoleic (P <0.01) at which time there was a signifi-cantly higher specific activity in bile than in themembranes.

Fig. 3 shows the incorporation of [3H]choline into thelinoleyl and arachidonyl lecithins of bile and the mem-brane fractions. The ['H]choline was incorporated intolinoleyl lecithin of microsomes, canaliculi, and bile fromthe same source more rapidly than it was into arachi-donyl lecithin. There was a close parallel between thespecific activity relationships of the lecithins from bilecanaliculi and microsomal membranes throughout the90-min period, with microsomal lecithin activity slightlyexceeding the specific activity of the biliary canalicularlecithin. The very rapid attainment of similar lecithinspecific activities in each of these membranes indicatesa very rapid transfer of biliary lecithin from microsomalto canalicular membranes. The incorporation of [8H]-choline into the linoleyl and arachidonyl lecithins of bilewas markedly greater than the incorporation into themembranous fractions. By the 30-60-min period the spe-cific activity of the biliary lecithins significantly (P <0.01) exceeded the specific activity of the membrane


FIGURE 1 Representative electron micrographs of bile canaliculi-enriched liver fractionobtained by discontinuous gradients of sucrose. (a) X 10,000. (b) X 48,000. Intact bile cana-liculi (bc) with microvilli (mv) and tight junctional complexes (tj) from adjoining plasmamembranes (pm). Fragmented plasma and bile canalicular membranes (fm).

lecithins. This difference was even more significant dur-ing the final study period (60-90 min) (P < 0.001).

Fig. 4 illustrates the incorporation of [3P]phosphateinto linoleyl and arachidonyl lecithins of the membranes

and bile. There was a greater incorporation of [NP]-phosphate into microsomal linoleyl lecithin than arachi-donyl lecithin. The pattern of incorporation of 'P intothe lecithin of the membranes and bile was markedly


Minutes

FIGURE 2 Incorporation of ['H]palmitic and ["C]linoleicacid into linoleyl lecithin of microsomes, bile canaliculi, andbile. Each point represents 7-9 perfused livers (±SE) pertime period. Bile samples are plotted at the midpoint ofeach collection period.

different than for the incorporation of the labeled fattyacids (Fig. 2) and choline (Fig. 3). Within 30 min ofadministering the 'P precursors, the specific activitiesof the lecithin from the bile canalicular membranes sig-nificantly (P < 0.02) exceeded those from the micro-somal fractions. At this time the linoleyl lecithin specificactivity in bile was also higher than that in the micro-somal fraction, but was less than that in the bile cana-licular membrane fraction. The specific activities of mi-crosomal 'sP in linoleyl and arachidonyl lecithins weresimilar at 30 min. After the rapid initial rise in the in-corporation of s'P into bile canaliculi there was a 30-minplateau period followed by a second rise in the incorpora-tion of 3P into both linoleyl and arachidonyl lecithin ofthe membranes. By the final study period (60-90 min)the 8'P specific activities of biliary linoleyl lecithin ex-ceeded the 8P specific activities of the membranefractions.

90 0Minutes

FIGURE 3 Incorporation of ['H]choline into linoleyl andarachidonyl lecithin of microsomes, bile canaliculi, and bile.Each point represents 7-9 perfused livers (+SE) per timeperiod. Bile samples are plotted at the midpoint of eachcollection period.

FIGURE 4 Incorporation of 8'PO4 into linoleyl and arachi-donyl lecithin of microsomes, bile canaliculi, and bile. Eachpoint represents 7-9 perfused livers (±SE) per time period.Bile samples are plotted at the midpoint of each collectionperiod.

The marked effect of taurocholate on microsomal leci-thin synthesis is shown in Fig. 5. Taurocholate signifi-cantly stimulated the incorporation Of "PO4 and ['H]-choline into the linoleyl (P < 0.05) and arachidonyl(P < 0.025, 0.05) lecithin species at 60 min as com-pared to nontaurocholate-perfused controls (saline).

The incorporation of ['H]mevalonic acid into choles-terol (Fig. 6) shows that the microsomal and biliarycanalicular incorporation curves were virtually identicaland indicate a very rapid transfer of the newly synthe-sized cholesterol from microsomes to biliary canaliculiin the presence of taurocholate. The cholesterol specificactivity of the biliary fraction was the same as that forthe membrane fractions during the first two secretion pe-riods (0-60 min), but during the last secretory period(60-90 min) the bile activity significantly (P < 0.02)exceeded that in both microsomes and biliary canaliculi.

Linoleyl Lecithin Arachidonyl Lecithin

48 Without Taurocholote 3With Taurocholote I

CL _ CLo0 32 - 300 a.

P<o005 P<oo5 P<0.025 P<0.05 T

E :t116 j 11 111 115°+0"

-2l

8~~~~~~~~~~~~~~~~~~~EC

32PO4 ['H]Choline '0PO. ['H] CholineMicrosomes

FIGURE 5 Effect of sodium taurocholate on the incorpora-tion of "PO4 and ['H choline into microsomal arachidonyland linoleyl lecithin at 60 min compared to controls (sa-line). Each bar represents eight taurocholate perfusions(+SE) or five saline controls.


DISCUSSION

Data have been presented which elucidate the possiblerole of the bile canalicular and microsomal membranesin the formation of biliary lecithin and cholesterol. Anessential prerequisite for this study was the availabilityof a method for isolating a liver cell fraction rich inthese bile canalicular membranes. The procedure ofSong et al. (25) yielded a preparation enriched withbiliary canaliculi (tight junction complexes) as evi-denced by marker enzymes and electron microscopy. Theabsence of CDP-choline diglyceride transferase activityin the canalicular membranes and the high concentrationof this enzyme in microsomes further attest to the verylow contamination of the canalicular preparation withendoplasmic reticulum. Another critical factor for study-ing hepatic lipid synthesis and biliary transport was theavailability of a model (the isolated perfused rat liver)which was relatively inactive with respect to biliary lipidsecretion until the system was perfused with sodiumtaurocholate (1-4).

Several lines of evidence from the present reportsuggest that bile canalicular membranes do not synthe-size biliary lecithin. First, CDP-choline diglyceridetransferase activity was not found in bile canaliculi.This enzyme is essential for the de novo synthesis oflecithin. The isotopic data can also be interpreted to in-dicate a lack of participation by the canalicular mem-brane in the synthesis of biliary lecithin. If the canalicu-lar membranes were the source of biliary lecithin, onewould expect that the linoleyl lecithin fraction of thismembrane should have a very rapid turnover during theactive phase of biliary lecithin secretion. Additionally,since linoleyl lecithin from the canalicular membranes israpidly flowing down the canaliculus during this activephase of biliary lipid secretion, the specific activity of thelinoleyl lecithin fraction in this membrane should be ap-proximately equivalent to bile. Therefore, the signifi-cantly higher specific activity of linoleyl lecithin in bilecompared to this same lecithin species in the bilecanalicular membrane indicates that linoleyl lecithinsynthesis probably did not occur in this membrane. Al-ternatively, the bile canalicular membrane could be thesite of biliary lecithin synthesis if the lecithin destinedfor bile was rapidly secreted without prior mixing in thecanalicular linoleyl lecithin pool. In this manner a frac-tion of the newly synthesized biliary lecithin could berapidly released into bile while another portion couldbe incorporated into a more slowly turning over struc-tural pool of membranous canalicular linoleyl lecithin.However, the inability to demonstrate CDP-choline di-glyceride transferase in the bile canalicular membraneprovides strong evidence that this is an unlikely pos-sibility. Although linoleyl lecithin is the major biliarylecithin, arachidonyl lecithin constitutes 10% of the bili-

5 6007

0

0

,( 400E

E

Minutes

FIGURE 6 Incorporation of [3Hjmevalonic acid into choles-terol of microsomes, bile canaliculi, and bile. Each pointrepresents 7-9 perfused livers (±SE) per time period. Bilesamples are plotted at the midpoint of each collection period.

ary lecithin fraction. The incorporation of [3H]cholineinto arachidonyl lecithin of bile canalicular and micro-somal membranes paralleled linoleyl lecithin and sug-gests that both of these biliary lecithins were derivedfrom microsomes. While these data show that bile cana-licular membranes do not actively participate in biliarylecithin synthesis, it is clear that newly synthesizedlecithin is rapidly transferred from microsomes to thecanalicular membranes before being released into bile.Thus, it seems that the major role of the bile canalicularmembrane with respect to biliary lipid metabolism istransport rather than synthesis.

Although these studies indicate that microsomes arethe site of biliary lecithin synthesis, the findings alsoshow that the hepatic metabolism of linoleyl lecithin is acomplex and heterogeneous process. The more rapidincorporation of [3H]choline and ["4C]linoleic into bili-ary linoleyl and arachidonyl lecithins rather than micro-somal and canalicular membrane lecithins provide evi-dence for the existence of a very rapidly turning oversubpool of these lecithins that are secreted in bile.Further evidence of a microsomal subpool of biliarylecithin is the incorporation o£f 'P into bile and canalicu-lar linoleyl and arachidonyl lecithin. At 30 min the spe-cific activity of 'PO4 in the canalicular membrane ex-ceeded microsomes and bile for both linoleyl and arachi-donyl lecithin and was disassociated from the incorpora-tion of [8H]choline. The much higher specific activityof 'PO4 in bile canalicular membranes compared to mi-crosomes could result from incomplete equilibration of'PO4 with a large ATP pool if the newly synthesizedlinoleyl lecithin destined for bile was selectively trans-ferred to the bile canalicular membrane. Alternatively,

Bile Canalicular Membranes and Biliary Lipid Formation ill

transformation of linoleyl lecithin involving high-energyphosphate could occur within the bile canalicular mem-brane. By 90 min the "PO4 had equilibrated within theATP pool and the specific activity relationship indicateda homogeneous incorporation pattern of 'PO4 into bothlecithin species of bile. Balint et al. (8, 17) comparedthe specific activities of hepatic and biliary lecithin inbile fistula rats after administration of labeled lecithinprecursors. They observed that biliary linoleyl lecithinhad a higher specific activity than hepatic linoleyl leci-thin and also postulated the existence of hepatic subpoolof biliary lecithin.

The findings of the present study indicate that tauro-cholate has an essential role in the transport of lipidsinto bile, and at the same time probably regulates therate of synthesis of biliary lecithin. The incorporationof ['H]choline and "PO4 into microsomal lecithin wasmarkedly greater in taurocholate-perfused livers ascompared to the saline-infused control livers. The ef-fect of taurocholate on increasing lecithin synthesis wasvery rapid, and was demonstrable within 90 min of tauro-cholate infusion. Infusion of taurocholate for 90 minbefore the addition of the labeled precursors was es-sential for stimulating lecithin synthesis and transport.The stimulatory effect of taurocholate was common toboth arachidonyl and linoleyl lecithin but was more pro-nounced on the linoleyl species. Balint et al. (8, 17)have previously reported a similar effect of taurocholateon linoleyl lecithin synthesis and attributed their findingsto a specific effect on the CDP-choline pathway. Fromthe available evidence (8, 17, 18, 37) it appears thatpalmityl linoleyl lecithin is the principal lecithin syn-thesized via this pathway. Arachidonyl lecithin is prob-ably derived from a combination of several pathways,including deacylation and acylation reactions involvingthe palmityl linoleyl species and the transmethylation ofphosphatidyl ethanolamine (37-39). Since arachidonyllecithin is probably not synthesized de novo, the ob-served increase in labeling of the phosphoryl cholinemoiety of that species in the presence of taurocholatemost likely resulted from transacylation with the morehighly labeled palmityl linoleyl species.

In addition to stimulating microsomal lecithin synthe-sis, taurocholate also appeared to facilitate the intra-cellular transport of biliary lipids as evidence by therapid transfer of labeled microsomal linoleyl lecithin tobile via the bile canalicular membrane. Thus bile saltsprobably have a dual role in the metabolism of biliarylipids which involves both synthesis and transport. Al-though intracellular cytoplasmic proteins have beenidentified (40, 41) which appear to participate in thetransfer of lecithins from microsomes to mitochondrialand plasma membranes, no attempt was made in thepresent study to isolate such a carrier protein. However,

the rapid exchange of lecithin observed between micro-somes, bile canaliculi, and bile make such a transfermechanism a distinct possibility. Whether the primaryeffect of bile salts is on synthesis or transport is notpresently known. It is possible that microsomal lecithincould be depleted as a result of increased transport. Ifthis were the case, microsomal linoleyl lecithin synthesiscould be initiated as a consequence of a negative feed-back effect on a rate-limiting step in the linoleyl lecithinbiosynthetic pathway such as the CDP-choline diglycer-ide transferase enzyme.

As noted by others (3, 6, 7), taurocholate was alsoshown to have a marked effect on biliary cholesterolsecretion. This effect was similar to that observed forlecithin as the ['H]cholesterol rapidly equilibrated be-tween microsomal and canalicular membranes. After theinitial washout (low specific activity) of ['H]cholesterolinto bile there was a progressive increase in biliary ['H] -cholesterol specific activity which by 90 min exceededboth microsomal and canalicular membranes. This phe-nomenon suggests the presence of a microsomal subpoolof cholesterol destined for bile. The apparent dependenceof both cholesterol and lecithin transport on the avail-ability of bile salts provides further support for theexistence of an intracellular lipoprotein complex thatis involved in the transport of lipids from microsomes tobile. Soluble proteins have been isolated from the super-nate of rat liver homogenates with binding affinity forbile acids (42) and cholesterol precursors (43). Such acomplex could originate in the microsomes and be depen-dent on the concentration of taurocholate entering theliver cell for its specific composition and rate of forma-tion. After coupling with carrier protein this complexcould be transported to other intracellular membranes,including the bile canalicular membrane. Alternatively,the biliary lipids could be transported independently tothe canalicular membrane or to an assembly site suchas the Golgi body and then subsequently to the bilecanaliculi. Whether the biliary lipids are independentlyand directly transported to canalicular membranes orarrive together as a lipoprotein macromolecular complexis not yet known. It also remains to be determined howthe biliary lipids are transported across the canalicularmembrane. Since it has been shown that liver plasmamembranes have the enzymatic capacity for the trans-acylation of lecithins (22), it is possible that microsomallecithin destined for bile could be altered in the bilecanalicular membranes before its appearance in bile.

ACKNOWLEDGMENTSThis work was supported in part by U. S. Public HealthService Research Grant 1ROl-AM-14668-02 from the Na-tional Institute of Arthritis and Metabolic Diseases, Na-tional Institutes of Health, U. S. Public Health Service,and funded research project (MRIS 3277) from the Vet-erans Administration.


REFERENCES1. Entenman, C., R. J. Holloway, M. L. Albright, and G.

F. Leong. 1968. Bile acids and lipid metabolism. I.Stimulation of bile lipid excretion by various bile acids.Proc. Soc. Exp. Biol. Med. 127: 1003-1006.

2. Kay, R. E., and C. Entenman. 1961. Stimulation oftaurocholic acid synthesis and biliary excretion of lipids.Am. J. Physiol. 200: 855-859.

3. Swell, L., C. Entenman, G. F. Leong, and R. J. Hollo-way. 1968. Bile acids and lipid metabolism. IV. Influenceof bile acids on biliary and liver organelle phospholipidsand cholesterol. Am. J. Physiol. 215: 1390-1396.

4. Swell, L., C. C. Bell, Jr., and C. Entenman. 1968. Bileacids and lipid metabolism. III. Influence of bile acidson phospholipids in liver and bile of the isolated per-fused dog liver. Biochim. Biophys. Acta. 164: 278-284.

5. Young, D. L., and K. C. Hanson. 1972. Effect of bilesalts on hepatic phosphatidylcholine synthesis and trans-port into rat bile. J. Lipid Res. 13: 244-252.

6. Nilsson, S., and T. Schersten. 1969. Importance of bileacids for phospholipid secretion into human hepaticbile. Gastroenterology. 57: 525-532.

7. Wheeler, H. O., and K. K. King. 1972. Biliary excre-tion of lecithin and cholesterol in the dog. J. Clin. In-vest. 51: 1337-1350.

8. Balint, J. A., D. A. Beeler, E. C. Kyriakides, and D.H. Treble. 1971. The effect of bile salt upon lecithinsynthesis. J. Lab. Clin. Med. 77: 122-133.

9. Thureborn, E. 1962. Human hepatic bile. Compositionchanges due to altered enterohepatic circulation. ActaChir. Scand. Suppl. 303.

10. Bell, C. C., Jr., Z. R. Vlahcevic, J. Prazich, and L.Swell. 1973. Evidence that a diminished bile acid poolprecedes the formation of cholesterol gallstones in man.Surg. Gynecol. Obst. 136: 961-965.

11. Vlahcevic, Z. R., C. C. Bell, Jr., I. Buhac, J. T. Farrar,and L. Swell. 1970. Diminished bile acid pool size inpatients with gallstones. Gastroenterology. 59: 165-173.

12. Admirand, W. H., and D. M. Small. 1968. The physi-cochemical basis of cholesterol gallstone formation inman. J. Clin. Invest. 47: 1043-1052.

13. Bell, C. C., Jr., Z. R. Vlahcevic, and L. Swell. 1971.Alterations in the lipids of human hepatic bile after theoral administration of bile salts. Surg. Gynecol. Obst.132: 36-42.

14. Danzinger, R. G., A. F. Hofmann, J. L. Thistle, andL. J. Schoenfield. 1973. Effect of oral chenodeoxycholicacid on bile acid kinetics and biliary lipid compositionin women with cholelithiasis. J. Clin. Invest. 52: 2809-2821.

15. Swell, L., C. C. Bell, Jr., and Z. R. Vlahcevic. 1971.Relationship of bile acid pool size to biliary lipid ex-cretion and the formation of lithogenic bile in man.Gastroenterology. 61: 716-722.

16. Balint, J. A., E. C. Kyriakides, H. L. Spitzer, and E.S. Morrison. 1965. Lecithin fatty acid composition inbile and plasma of man, dogs, rats, and oxen. J. LipidRes. 6: 96-99.

17. Balint, J. A., D. A. Beeler, D. H. Treble, and H. L.Spitzer. 1967. Studies in the biosynthesis of hepaticand biliary lecithins. J. Lipid Res. 8: 486-493.

18. Bremer, J., and D. M. Greenberg. 1961. Methyl trans-ferring enzyme system of microsomes in the biosynthe-sis of lecithins (phosphatidylcholine). Biochim. Bio-phys. Acta. 46: 205-216.

19. Kanoh, H. 1969. Biosynthesis of molecular species ofphosphatidyl choline and phosphatidyl ethanolamine fromradioactive precursors in rat liver slices. Biochim. Bio-phys. Acta. 176: 756-763.

20. Sundler, R., G. Arvidson, and B. Akesson. 1972. Path-ways for the incorporation of choline into rat liverphosphatidylcholines in vivo. Biochim. Biophys. Acta.280: 559-568.

21. De Kruyff, B., L. M. G. Van Golde, and L. L. M.Van Deenen. 1970. Utilization of diacylglycerol speciesby cholinephosphotransferase, ethanolaminephosphotrans-ferase and diacylglycerol acyltransferase in rat livermicrosomes. Biochim. Biophys. Acta. 210: 425-435.

22. Victoria, E. J., L. M. G. Van Golde, K. Y. Hostetler,G. L. Scherphof, and L. L. M. Van Deenen. Somestudies on the metabolism of phospholipids in plasmamembranes from rat liver. Biochim. Biophys. Acta.239: 443-457.

23. Sarzala, M. G., L. M. G. Van Golde, B. De Kruyff,and L. L. M. Van Deenen. 1970. The intramitochon-drial distribution of some enzymes involved in the bio-synthesis of rat-liver phospholipids. Biochim. Biophys.Acta. 202: 106-119.

24. Swell, L., and M. D. Law. 1971. Release of lipoproteincholesterol esters by the isolated perfused rat liver.Biochim. Biophys. Acta. 231: 302-313.

25. Song, C. S., W. Rubin, A. B. Rifkind, and A. Kappas.1969. Plasma membranes of the rat liver. Isolationand enzymatic characterization of a fraction rich inbile canaliculi. J. Cell Biol. 41: 124-132.

26. Schwartz, M. K., and 0. Bodansky. 1961. I. Glycolyticand related enzymes. Methods Med. Res. 9: 5-23.

27. Widnell, C. C., and J. C. Unkeless. 1966. Partial puri-fication of a lipoprotein with 5'-nucleotidase activityfrom membranes of rat liver cells. Proc. Natl. Acad.Sci. U. S. A. 61: 1050-1057.

28. Emmelot, P., and C. J. Bos. 1966. Studies on plasmamembranes. III. Mga+-ATPase, (Na+-K+-Mg'e)-ATPaseand 5'nucleotidase activity of plasma membranes iso-lated from rat liver. Biochim. Biophys. Acta. 120:369-382.

29. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, andR. J. Randall. 1951. Protein measurement with theFolin phenol reagent. J. Biol. Chem. 193: 265-275.

30. Folch, J., M. Lees, and G. H. Sloane-Stanley. 1957. Asimple method for the isolation and purification oftotal lipids from animal tissues. J. Biol. Chem. 226:497-509.

31. Hirsch, J., and E. H. Ahrens, Jr. 1958. The separationof complex lipide mixtures by the use of silicic acidchromatography. J. Biol. Chem. 233: 311-320.

32. Swell, L. 1968. Separation, identification, and estimationof radioactive and nonradioactive lipids. In Gas Chro-matography. Grune & Stratton, Inc., New York. 97-109.

33. Skipski, V. P., R. F. Peterson, and M. Barclay. 1964.Quantitative analysis of phospholipids by thin-layerchromatography. Biochem. J. 90: 374-378.

34. Arvidson, G. A. E. 1968. Structural and metabolicheterogeneity of rat liver glycerophosphatides. Eur. J.Biochem. 4: 478-486.

35. Bartlett, G. R. 1959. Phosphorus assay in columnchromatography. J. Biol. Chem. 234: 466-468.

36. Snedecor, G. W., and W. G. Cochran. 1967. StatisticalMethods. The Iowa State University Press, Ames,Iowa. 6th edition.


37. Trewhella, M. A., and F. D. Collins. 1973. Pathwaysof phosphatidylcholine biosynthesis in rat liver. Bio-chim. Biophys. Acta. 296: 51-61.

38. Lee, T. C., and F. Snyder. 1973. Phospholipid metabo-lism in rat liver endoplasmic reticulum structural anal-yses, turnover studies and enzymic activities. Biochim.Biophys. Acta. 291: 71-82.

39. Wykle, R. L., M. L. Blank, and F. Snyder. 1973. Theenzymic incorporation of arachidonic acid into ether-containing choline and ethanolamine phosphoglyceridesby deacylation-acylation reactions. Biochim. Biophys.Acta. 326: 26-33.

40. Zilversmit, D. B. 1971. Exchange of phospholipid

classes between liver microsomes and plasma: com-parison of rat, rabbit and guinea pig. J. Lipid Res. 12:3642.

41. Akiyama, M., and T. Sakagami. 1969. Exchange ofmitochondrial lecithin and cephalin with those in ratliver microsomes. Biochim. Biophys. Acta. 187: 105-112.

42. Hanson, R. F., K. McCoy, and M. E. Dempsey. 1973.The role of a carrier protein in bile acid biosynthesis.Gastroenterology. 64: 154. (Abstr.)

43. Scallen, T. J., M. W. Schuster, and A. K. Dhar. 1971.Evidence for a noncatalytic carrier protein in choles-terol biosynthesis. J. Biol. Chem. 246: 224-230.


mechanism of secretion of biliary...

Documents