Title: Intracellular Lipidation of Newly Synthesized Apolipoprotein A-I in Primary Murine Hepatocytes
Abstract: Hepatocytes, which are the main site of apolipoprotein (apo)A-I and ATP-binding cassette transporter A1 (ABCA1) expression, are also the main source of circulating high density lipoprotein. Here we have characterized the intracellular lipidation of newly synthesized apoA-I, in primary hepatocytes cultured with [3H]choline to label choline-phospholipids, low density lipoprotein-[3H]cholesterol to label the cell surface, or [3H]mevalonate to label de novo synthesized cholesterol. Phospholipidation of apoA-I is significant and most evident in endoplasmic reticulum (ER) and medial Golgi, both in the lumen and on the membrane fractions of the ER and medial Golgi. In the presence of cycloheximide, endogenous apoA-I is substantially phospholipidated intracellularly but acquires some additional lipid after export out of the cell. In cells labeled with low density lipoprotein-[3H]cholesterol, intracellular cholesterol lipidation of apoA-I is entirely absent, but the secreted apoA-I rapidly accumulates cholesterol after secretion from the cell in the media. On the other hand, de novo synthesized cholesterol can lipidate apoA-I intracellularly. We also showed the interaction between apoA-I and ABCA1 in ER and Golgi fractions. In hepatocytes lacking ABCA1, lipidation by low density lipoprotein-cholesterol was significantly reduced at the plasma membrane, phospholipidation and lipidation by de novo synthesized sterols were both reduced in Golgi compartments, whereas ER lipidation remained mostly unchanged. Therefore, the early lipidation in ER is ABCA1 independent, but in contrast, the lipidation of apoA-I in Golgi and at the plasma membrane requires ABCA1. Thus, we demonstrated that apoA-I phospholipidation starts early in the ER and is partially dependent on ABCA1, with the bulk of lipidation by phospholipids and cholesterol occurring in the Golgi and at the plasma membrane, respectively. Finally, we showed that the previously reported association of newly synthesized apoA-I and apoB (Zheng, H., Kiss, R. S., Franklin, V., Wang, M. D., Haidar, B., and Marcel, Y. L. (2005) J. Biol. Chem. 280, 21612–21621) occurs after secretion at the cell surface. Hepatocytes, which are the main site of apolipoprotein (apo)A-I and ATP-binding cassette transporter A1 (ABCA1) expression, are also the main source of circulating high density lipoprotein. Here we have characterized the intracellular lipidation of newly synthesized apoA-I, in primary hepatocytes cultured with [3H]choline to label choline-phospholipids, low density lipoprotein-[3H]cholesterol to label the cell surface, or [3H]mevalonate to label de novo synthesized cholesterol. Phospholipidation of apoA-I is significant and most evident in endoplasmic reticulum (ER) and medial Golgi, both in the lumen and on the membrane fractions of the ER and medial Golgi. In the presence of cycloheximide, endogenous apoA-I is substantially phospholipidated intracellularly but acquires some additional lipid after export out of the cell. In cells labeled with low density lipoprotein-[3H]cholesterol, intracellular cholesterol lipidation of apoA-I is entirely absent, but the secreted apoA-I rapidly accumulates cholesterol after secretion from the cell in the media. On the other hand, de novo synthesized cholesterol can lipidate apoA-I intracellularly. We also showed the interaction between apoA-I and ABCA1 in ER and Golgi fractions. In hepatocytes lacking ABCA1, lipidation by low density lipoprotein-cholesterol was significantly reduced at the plasma membrane, phospholipidation and lipidation by de novo synthesized sterols were both reduced in Golgi compartments, whereas ER lipidation remained mostly unchanged. Therefore, the early lipidation in ER is ABCA1 independent, but in contrast, the lipidation of apoA-I in Golgi and at the plasma membrane requires ABCA1. Thus, we demonstrated that apoA-I phospholipidation starts early in the ER and is partially dependent on ABCA1, with the bulk of lipidation by phospholipids and cholesterol occurring in the Golgi and at the plasma membrane, respectively. Finally, we showed that the previously reported association of newly synthesized apoA-I and apoB (Zheng, H., Kiss, R. S., Franklin, V., Wang, M. D., Haidar, B., and Marcel, Y. L. (2005) J. Biol. Chem. 280, 21612–21621) occurs after secretion at the cell surface. Liver and intestine, which are the major sites of expression for apoA-I (1Breslow J.L. Annu. Rev. Biochem. 1985; 54: 699-727Crossref PubMed Scopus (109) Google Scholar, 2Breslow J.L. Am. Heart J. 1987; 113: 422-427Crossref PubMed Scopus (9) Google Scholar) and ABCA1 (3Lawn R.M. Wade D.P. Garvin M.R. Wang X.B. Schwartz K. Porter J.G. Seilhamer J.J. Vaughan A.M. Oram J.F. J. Clin. Investig. 1999; 104: R25-R31Crossref PubMed Scopus (658) Google Scholar, 4Singaraja R.R. Bocher V. James E.R. Clee S.M. Zhang L.H. Leavitt B.R. Tan B. Brooks-Wilson A. Kwok A. Bissada N. Yang Y.Z. Liu G.Q. Tafuri S.R. Fievet C. Wellington C.L. Staels B. Hayden M.R. J. Biol. Chem. 2001; 276: 33969-33979Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar, 5Wellington C.L. Walker E.K. Suarez A. Kwok A. Bissada N. Singaraja R. Yang Y.Z. Zhang L.H. James E. Wilson J.E. Francone O. McManus B.M. Hayden M.R. Lab. Investig. 2002; 82: 273-283Crossref PubMed Scopus (244) Google Scholar, 6Cavelier L.B. Qiu Y. Bielicki J.K. Afzal V. Cheng J.F. Rubin E.M. J. Biol. Chem. 2001; 276: 18046-18051Abstract Full Text Full Text PDF PubMed Scopus (77) Google Scholar, 7Langmann T. Klucken J. Reil M. Liebisch G. Luciani M.F. Chimini G. Kaminski W.E. Schmitz G. Biochem. Biophys. Res. Commun. 1999; 257: 29-33Crossref PubMed Scopus (429) Google Scholar, 8Langmann T. Mauerer R. Zahn A. Moehle C. Probst M. Stremmel W. Schmitz G. Clin. Chem. 2003; 49: 230-238Crossref PubMed Scopus (236) Google Scholar), are the main contributors to high density lipoprotein (HDL) 2The abbreviations used are: HDLhigh density lipoproteinABCA1ATP-binding cassette transporter A1apoapolipoproteinERendoplasmic reticulummGolgimedial GolgidGolgidistal GolgiAd-AIadenovirus expressing apo A-IAd-Lucadenovirus expressing luciferaseGC-MSgas chromatography-mass spectrometryLDLlow density lipoproteinLDL-[3H]cholesterol[3H]cholesterol incorporated into low density lipoprotein for delivery to cellsWTwild-typeTEMEDN,N,N′,N′-tetramethylethylenediaminePBSphosphate-buffered salineTGNtrans-Golgi networkManIImannosidase II. synthesis and secretion. The recent targeted inactivation of the hepatic ABCA1 demonstrated unequivocally that the liver is indeed the major site of synthesis, accounting for 83% of the circulating HDL (9Timmins J.M. Lee J.Y. Boudyguina E. Kluckman K.D. Brunham L.R. Mulya A. Gebre A.K. Coutinho J.M. Colvin P.L. Smith T.L. Hayden M.R. Maeda N. Parks J.S. J. Clin. Investig. 2005; 115: 1333-1342Crossref PubMed Scopus (427) Google Scholar). However, it still remains unclear whether apoA-I synthesized in the hepatocyte is lipidated intracellularly and/or at the cell surface. Banerjee and Redman (10Banerjee D. Redman C.M. J. Cell Biol. 1983; 96: 651-660Crossref PubMed Scopus (30) Google Scholar, 11Banerjee D. Redman C.M. J. Cell Biol. 1984; 99: 1917-1926Crossref PubMed Scopus (27) Google Scholar) showed that in the avian hepatocyte the initial lipidation of the protein occurs in Golgi fractions but the newly formed lipoprotein particles are not immediately mature. Chisholm and colleagues (12Chisholm J.W. Burleson E.R. Shelness G.S. Parks J.S. J. Lipid Res. 2002; 43: 36-44Abstract Full Text Full Text PDF PubMed Google Scholar) reported that in HepG2 cells half of the lipidation of apoA-I occurs intracellularly and continues after secretion at the cell surface. Although in chicken livers, the newly secreted apoA-I accumulates lipids necessary to obtain the structure of a mature HDL, this is not the case in primary hepatocytes and HepG2 cells where lipidpoor apoA-I has often been found in the medium (13Melin B. Keller G. Glass C. Weinstein D.B. Steinberg D. Biochim. Biophys. Acta. 1984; 795: 574-588Crossref PubMed Scopus (19) Google Scholar, 14Thrift R.N. Forte T.M. Cahoon B.E. Shore V.G. J. Lipid Res. 1986; 27: 236-250Abstract Full Text PDF PubMed Google Scholar). high density lipoprotein ATP-binding cassette transporter A1 apolipoprotein endoplasmic reticulum medial Golgi distal Golgi adenovirus expressing apo A-I adenovirus expressing luciferase gas chromatography-mass spectrometry low density lipoprotein [3H]cholesterol incorporated into low density lipoprotein for delivery to cells wild-type N,N,N′,N′-tetramethylethylenediamine phosphate-buffered saline trans-Golgi network mannosidase II. In previous work (15Kiss R.S. McManus D.C. Franklin V. Tan W.L. McKenzie A. Chimini G. Marcel Y.L. J. Biol. Chem. 2003; 278: 10119-10127Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar), we addressed this issue indirectly, comparing the phospholipidation of endogenously synthesized apoA-I and exogenously added apoA-I. We showed that endogenous apoA-I consistently acquired more phospholipids and formed more HDL size particles than exogenous apoA-I (15Kiss R.S. McManus D.C. Franklin V. Tan W.L. McKenzie A. Chimini G. Marcel Y.L. J. Biol. Chem. 2003; 278: 10119-10127Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). The phospholipidation of newly synthesized apoA-I is for the most part dependent on ABCA1, but the transfer of cholesterol to apoA-I is controlled by more complex pathways less dependent on ABCA1 (16Zheng H. Kiss R.S. Franklin V. Wang M.D. Haidar B. Marcel Y.L. J. Biol. Chem. 2005; 280: 21612-21621Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar). Unlike apoB, which is degraded if not sufficiently lipidated, apoA-I can be secreted as poorly lipidated complexes, perhaps even as a lipid-free protein. As indicated above, endogenously synthesized apoA-I forms particles of the same size in ABCA1-deficient hepatocytes and control cells, only it forms much fewer of the large particles in the absence of ABCA1 (15Kiss R.S. McManus D.C. Franklin V. Tan W.L. McKenzie A. Chimini G. Marcel Y.L. J. Biol. Chem. 2003; 278: 10119-10127Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). Therefore apoA-I secretion is not controlled by lipid availability and the type of particle formed is dictated by the apoA-I structure and its ability to form dimers and tetramers that accommodate specific amounts of lipids. Here, we investigated the different steps of apoA-I lipidation as it traffics from the endoplasmic reticulum (ER) to the Golgi apparatus, until its secretion. We have identified early steps of phospholipidation in the ER that are independent of ABCA1 activity, and we characterized the intracellular compartments that contribute to the transfer of cholesterol to apoA-I during secretion. We also define the site of apoA-I association with apoB-containing lipoproteins. Materials—[1,2-3H]Cholesterol, [5-3H]mevalonolactone-Rs, [methyl-3H]choline chloride were obtained from PerkinElmer Life Sciences. Williams Medium E, HepatoZYME-SFM, and antibiotic-antimycotic were purchased from Invitrogen. Rabbit polyclonal anti-human apoA-I antibody was purchased from Calbiochem. Monoclonal antibodies directed against human apoA-I (a combination of 4H1 (against the extreme N terminus) and 5F6 (against the central region)) were obtained as previously described (17Marcel Y.L. Jewer D. Vezina C. Milthorp P. Weech P.K. J. Lipid Res. 1987; 28: 768-777Abstract Full Text PDF PubMed Google Scholar, 18Marcel Y.L. Provost P.R. Koa H. Raffai E. Dac N.V. Fruchart J.C. Rassart E. J. Biol. Chem. 1991; 266: 3644-3653Abstract Full Text PDF PubMed Google Scholar) and biotinylated with Sulfo-NHS-Biotin from Pierce. Polyclonal anti-mouse apoB antibody was purchased from Biodesign International. Streptavidin-horseradish peroxidase conjugate and protein G-Sepharose were obtained from Amersham Biosciences. Primary anti-mouse apoB antibody was a gift from Dr. Ross Milne (University of Ottawa Heart Institute). Rabbit anti-mouse TGN38 antibody was obtained from BD Biosciences. Rabbit anti-mouse EEAI antibody was obtained from Affinity Bioreagents. Rabbit anti-mouse mannosidase II (ManII) antibody was a gift from Dr. Zemin Yao (University of Ottawa). Rabbit anti-mouse calnexin antibody was obtained from Stressgen Bioreagents. Sheep anti-mouse IgG, horseradish peroxidase-linked whole antibodies and donkey anti-rabbit IgG, horseradish peroxidase-linked whole antibodies were obtained from Amersham Biosciences. Nycodenz gradient maker, fibronectin, cycloheximide, stigmasterol, Sil-A derivatization chemical, and collagenase were obtained from Sigma. Complete protease inhibitor mixture was obtained from Roche. Acrylamide-Bis solution and SDS were obtained from Bio-Rad. Eco-Lite scintillation liquid and TEMED were obtained from ICN Biomedicals. SuperSignal chemiluminescent substrate was obtained from Pierce. Cell Culture—Primary mouse hepatocytes were isolated from C57BL/6 ABCA1+/+ and C57BL/6 ABCA1–/– mice (a kind gift from Dr. Edward M. Rubin, DOE Joint Genome Institute, Berkeley, CA) by collagenase liver perfusion. The cells were plated on fibronectin-coated 10-cm plates and grown in Williams media (Invitrogen). [3H]Choline (10 μCi/ml) and [3H]mevalonate (15 μCi/ml) were directly added to loading media (hepatozyme media with 1% l-glutamine and 1% antimycotic antibiotic reagent). [3H]Cholesterol (10 μCi/ml) was dried under nitrogen into a thin film, then solubilized in a small volume of ethanol. LDL was added to the [3H]cholesterol and equilibrated with the LDL for one-half hour before adding to loading media. Five hours after the initial seeding, cells were incubated with loading media labeled with [3H]choline, [3H]mevalonate, or [3H]cholesterol. The next day (24 h later) the labeled media was removed and the cells were infected for 1 h with a recombinant adenovirus encoding apoA-I (Ad-AI) or control luciferase (Ad-Luc) at a multiplicity of infection of 75:1 plaque-forming units per cell in Williams Medium E without serum (19McManus D.C. Scott B.R. Frank P.G. Franklin V. Schultz J.R. Marcel Y.L. J. Biol. Chem. 2000; 275: 5043-5051Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). After the 1-h infection, the adenovirus containing media was removed and the radioactive media returned for another 24 h. Cellular Fractionation and Preparation of Nycodenz Gradient—The subcellular fractionation protocol was essentially as described in Tran et al. (20Tran K. Thorne-Tjomsland G. DeLong C.J. Cui Z. Shan J. Burton L. Jamieson J.C. Yao Z.M. J. Biol. Chem. 2002; 277: 31187-31200Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). At the end of the labeling period, the cells were washed twice with Williams media. At time 0, cycloheximide in hepatozyme media was added to cells (3.55 μl/ml of 100 mm stock). At the appropriate time point (0, 20, 40, and 60 min), the cells were washed twice with cold PBS and then scraped with a plastic scraper. The collected cells were then spun down at 2,000 × g for 5 min at 15 °C. The pellet was resuspended in membrane solubilization buffer (10 mm Tris-HCl, pH 7.4, 250 mm sucrose in ddH2O, Complete protease inhibitor mixture, 5 μm EDTA, pH 8) and homogenized using a ball bearing homogenization apparatus. The cell homogenate was then transferred to 15-ml glass Corex tubes and centrifuged at 9000 × g for 10 min at 5 °C, and the supernatant was kept. Nycodenz stock solution (27.6% Nycodenz in 10 mm Tris-HCl, pH 7.4, 3 mm KCl, 1 mm EDTA) and saline buffer were used to prepare four Nycodenz solutions of increasing percent concentrations (10, 14.66, 19.33, and 24%) and 2.5 ml of each were loaded from the bottom of the tube (Beckman Polyallomer Centrifuge Tubes) in decreasing percentage order. The tubes were then sealed with a piece of parafilm, and a linear gradient was formed by placing the tubes horizontally for 45 min at room temperature. The tubes were then centrifuged at 37,000 × g for 4 h at 15 °C (Beckman L8–70 M Ultracentrifuge). The supernatant was layered on top of the created Nycodenz gradient and centrifuged at 37,000 × g, for 1.5 h at 15 °C (SW41 rotor). Following the spin, each tube was fractionated into 15 aliquots. These aliquots were stored at 4 °C for Western blotting and determination of radioactive content. Separation of Lumenal and Membrane-bound Fractions—To separate microsomal lumen from membranes, the 15 aliquots were grouped into 3 microsomal fractions (aliquots 1–3 distal Golgi fraction, 4–8 medial Golgi fraction, and 9–15 ER fraction). 1.5 ml were removed from each fraction and 1.5 ml of 0.2 m Na2CO3, pH 12.4, were added to each condition. The samples were allowed to mix on a rotator for 30 min at room temperature. The samples were transferred to open-end tubes (Beckman Polyallomer Thick Wall 3.2 ml tubes) and centrifuged at 70,000 × g for 30 min at 15 °C (TLA 100.4/TLA110). The supernatant, representing the lumenal content of each intracellular compartment, was collected and the pH adjusted to 8 with 75 μl of 2.5 n HCl. The pellet representing the membrane fraction was resuspended in 250 μl of membrane solubilization buffer. Immunoprecipitation—Immunoprecipitation was preformed on subcellular and media samples collected during the fractionation procedure. 90 μl of anti-apoA-I antibody (Calbiochem) was added to 3 ml of sample. The samples were incubated at 4 °C overnight with continuous mixing. The next day, 100 μl of protein G-Sepharose (Amersham Biosciences) was added to the samples and the samples were incubated overnight at 4 °C. The next day, samples were centrifuged at 3,000 × g at 10 °C for 10 min (Sorvall RT 6000D). The supernatant was discarded and the pellet washed 3 times. When immunoprecipitating apoB with rabbit anti-apoB antibody (Biodesign) the same procedure was followed except that protein A-Sepharose (Amersham Biosciences) was used. When apoA-I was immunoprecipitated from a membrane sample, the sample was treated with 1% Triton X-100 for 30 min at 4 °C. The solubilized sample was then centrifuged at 75,000 × g for 30 min at 15 °C (TLA 100.4/TLA110) and the supernatant loaded on a gel. Western Blotting—Protein samples from aliquots and fractions were separated on a 3–15% SDS-PAGE and then transferred to a nitrocellulose membrane (Bio-Rad). The membrane was blocked (5% skim milk in 0.5% Tween PBS) for 30 min and washed three times for 15 min in 0.5% Tween PBS. Primary mouse anti-human apoA-I antibodies 5F6 and 4H1 (University of Ottawa Heart Institute) were added at 1:1000 dilution. Primary anti-mouse apoB antibody (a gift from Dr. Ross Milne) was added at 1:1000 dilution. Rabbit anti-mouse TGN38 antibody (BD Biosciences) was added at 1:1000 dilution. Rabbit anti-mouse EEAI antibody (Affinity Bioreagents) was added at 1:1000 dilution. Rabbit anti-mouse ManII antibody (a gift from Dr. Zemin Yao) was added at 1:2000 dilution. Rabbit anti-mouse calnexin antibody (Stressgen Bioreagents) was added at 1:3000 dilution. After an overnight incubation at 4 °C, the membrane was washed three times at 15 min in 0.5% Tween-PBS. Secondary antibodies were added at 1:5000 dilution and the membrane was incubated for 1 h at room temperature. For apoA-I and apoB, sheep anti-mouse IgG horseradish peroxidase-linked whole antibodies were used (Amersham Biosciences). For all the remaining proteins, donkey anti-rabbit IgG horseradish peroxidase-linked whole antibodies were used (Amersham Biosciences). The membrane was washed 3 times for 15 min in 1× PBS with 0.5% Tween. The membrane was then washed with SuperSignal chemiluminescent substrate (Pierce) for 5 min and exposed to film. Lipid Extraction and TLC Analysis—Lipids associated with immunoprecipitated samples (1 ml) were extracted by the method of Bligh and Dyer (21Bligh E.G. Dyer W.J. Can. J. Biochem. 1959; 37: 911-917Crossref PubMed Scopus (43133) Google Scholar). To determine the amount of [3H]choline incorporated into sphingomyelin and phosphatidylcholine, phospholipids were separated by thin layer chromatography (TLC) using a polar solvent system containing CHCl3:methanol:acetic acid:formic acid:H2O (105:45:18:6:3, v/v). The lipids on the TLC plate were stained with iodine, and the appropriate bands scraped. The associated radioactivity was determined by scintillation counting. To determine the amount of [3H]mevalonate incorporated into cholesterol versus cholesteryl esters, the non-polar solvent system hexane:diethyl ether:acetic acid (105:45:1.5) was used. Gas Chromatography Analysis—For GC analysis, samples from the immunoprecipitation of apoA-I and media samples were resuspended in PBS and the lipids extracted (21Bligh E.G. Dyer W.J. Can. J. Biochem. 1959; 37: 911-917Crossref PubMed Scopus (43133) Google Scholar). The lipid extract was divided in two. One sample for unesterified cholesterol determination was dried under nitrogen and stored at –20 °C. The other sample for total cholesterol determination was hydrolyzed with alcoholic KOH at 60 °C for 1. Cholesterol was extracted with hexane and evaporated under nitrogen and frozen at –20 °C. Stigmasterol was added as an internal standard and the dried samples were treated with Sil-A derivatization chemical (Sigma) prior to GC-MS analysis. Isolation of LDL—LDL was isolated from a peripheral blood sample by sequential density ultracentrifugation. The concentration of LDL was calculated using the Markwell Lowry protein assay. Primary mouse hepatocytes were isolated from C57BL/6 ABCA1+/+ mice by collagenase liver perfusion, and cultured as described under “Experimental Procedures.” The cells were lysed and subcellular fractions isolated by Nycodenz gradient centrifugation. Antibodies against specific intracellular marker proteins were used to identify the various fractions (Fig. 1). The protein markers EEAI and TGN38 identify early endosomes and distal Golgi (dGolgi) microsomes, respectively. ManII and calnexin are markers of medial Golgi (mGolgi) and ER microsomes, respectively. The first three fractions were pooled to represent dGolgi with some early endosome vesicles. Fractions 4–6 and 9–15 were pooled to represent mGolgi and ER vesicles, respectively. Steady State ApoA-I Lipidation with [3H]Choline-phospholipids or LDL-derived [3H]Cholesterol and ApoA-I-mediated Export of Cholesterol—To determine whether phospholipids and cholesterol become associated with apoA-I in ER and Golgi compartments, primary mouse hepatocytes were labeled with [3H]choline or LDL-[3H]cholesterol, infected with either Ad-Luc or adenovirus expressing apoA-I, and labeling was pursued by continued culture in labeling medium. The cells were then washed, homogenized, and subcellular fractions isolated. ApoA-I was immunoprecipitated and the associated radioactive lipid measured. We observed very significant acquisition of choline-labeled lipids by apoA-I in the ER compartment (Fig. 2). The presence of phospholipidated apoA-I was also evident in the mGolgi, although radioactivity was consistently lower in Golgi compared with ER. Under the same conditions, when the cells were labeled with [3H]cholesterol derived from LDL, lipidation with cholesterol was almost completely absent. We have shown previously that cholesterol delivered with LDL preferentially labels the cell surface pool, including the recycling endosome compartment (16Zheng H. Kiss R.S. Franklin V. Wang M.D. Haidar B. Marcel Y.L. J. Biol. Chem. 2005; 280: 21612-21621Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar). Therefore the absence of cholesterol labeling of ER and Golgi by exogenous cholesterol explains these negative results. ApoA-I Protein Localization following Cycloheximide Treatment— Pulse-chase experiments were performed, where the cells were cultured as described above and at time 0 treated with cycloheximide to inhibit any new production of apoA-I (Fig. 3). The cells were collected at 0, 20, and 40 min and subcellular fractions isolated as described. ApoA-I was located both on the membrane fractions and in the lumen of ER, mGolgi, and dGolgi for the first hour following cycloheximide addition and accumulated in the media (Fig. 3A). Densitometric quantification of Western blot bands show that initially, apoA-I is located in all three intracellular compartments, mainly in the dGolgi and mGolgi, whereas the ER seems to contain the smallest amount, a result compatible with a previous report of the rapid transit of apoA-I (10Banerjee D. Redman C.M. J. Cell Biol. 1983; 96: 651-660Crossref PubMed Scopus (30) Google Scholar). With time the ER compartment emptied and apoA-I transits into mGolgi and dGolgi. After 40 min, apoA-I leaves the mGolgi to accumulate in the media. The subcellular fractions were treated with sodium carbonate to release the proteins residing inside the ER and Golgi vesicles and centrifuged to separate lumen and membrane compartments. Western blotting (Fig. 3B) shows at 0 and 20 min the presence of a large amount of membrane-bound apoA-I in the mGolgi fraction, which contrast with the low level of membrane-bound apoA-I in dGolgi. At 40 min in the dGolgi, we observed only the presence of lumenal apoA-I, which suggests that the progressive lipidation of apoA-I caused the release of apoA-I as a soluble complex. Intracellular Lipidation of ApoA-I in ABCA1 WT Cells after Cycloheximide Treatment—Hepatocytes from C57BL/6 ABCA1+/+ mice were labeled with [3H]choline, [3H]cholesterol, or [3H]mevalonate and infected with Ad-human-AI as described above. Phospholipid and cholesterol lipidation of apoA-I was assessed after addition of cycloheximide for 0, 20, and 40 min. In [3H]choline-labeled cells, immunoprecipitation of apoA-I in the different subcellular fractions shows that the phospholipidation of apoA-I starts in the ER, and continues through the Golgi leading to the secretion of phospholipidated apoA-I (Fig. 4A). Clearly, apoA-I binds phospholipids during or soon after its translation, and ER apoA-I, despite its low concentration, retains a high level of labeled phospholipids throughout the chase. ApoA-I remained significantly lipidated during transit through the mGolgi, but its clearance from the dGolgi (Fig. 3B) may explain the low level of apoA-I-associated phospholipids in that fraction (Fig. 4A). After 20 min, it is evident that some of the intracellularly phospholipidated apoA-I left the cell and could be found in the media, where it continues to accumulate. The distribution of [3H]choline label in the phospholipids associated with apoA-I was analyzed by TLC. The lumenal fraction was isolated from the ER and Golgi compartments, and the lipids associated with immunoprecipitated apoA-I were extracted and separated by TLC. Phosphatidylcholine associated with apoA-I early in the ER lumen and peaked at 20 min (Fig. 5). Some sphingomyelin radioactivity started to associate with apoA-I early in the ER, but the majority was seen in the mGolgi (Fig. 5), where sphingomyelin is synthesized. Very little phosphatidylcholine or sphingomyelin radioactivity was measured in dGolgi. The specificity of cholesterol binding by apoA-I was strikingly different when the cells were labeled with LDL-[3H]cholesterol (Fig. 4B), a protocol that preferentially labels the cell surface compartment (16Zheng H. Kiss R.S. Franklin V. Wang M.D. Haidar B. Marcel Y.L. J. Biol. Chem. 2005; 280: 21612-21621Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar). At all three time points, intracellular lipidation was almost non-existent. However, secreted apoA-I in the media was able to accumulate a large amount of cholesterol at the plasma membrane. This lipidation was significant at time 0 and continued to increase over the next 40 min, reflecting the transfer of cholesterol to secreted apoA-I at the level of plasma membrane and recycling endosomes. The absence of cholesterol-labeled apoA-I in the dGolgi fraction suggest that secreted apoA-I does not recycle to this fraction. To ascertain whether or not apoA-I can acquire cholesterol in the early stages of synthesis and transport, the cells were cultured with [3H]mevalonate to label de novo synthesized cholesterol, which originates in the ER and transits along the pathway of apoA-I transport. The results showed that de novo synthesized cholesterol does lipidate apoA-I early in ER and remain associated to apoA-I in Golgi compartments as well as after the secretion of the lipoprotein particle out of the cell (Fig. 4C). The amount of de novo synthesized cholesterol bound to apoA-I in ER remains constant throughout the secretion pathway. This lipidation is significant and complements the acquisition of cholesterol by apoA-I that occurs at the cell surface. TLC analysis of the lipids immunoprecipitated with apoA-I showed that [3H]mevalonate was almost exclusively found in the cholesterol fractions with very little if any with cholesteryl esters (data not shown). It is clear that hepatic apoA-I secretion is accompanied by the export of cholesterol and phospholipids. To provide a quantitative evaluation, we measured by GC-MS the net mass of cholesterol associated with the immunoprecipitated apoA-I after a 3-h time point. In the total intracellular lumenal compartments, we found a total of 11.7 ng of cholesterol/μg of total secreted apoA-I. In all the membrane compartments, we found 2.5 ng of cholesterol/μg of secreted apoA-I. In the media, we measured 6.7 ng of cholesterol/μg of secreted apoA-I. Intracellular Lipidation of ApoA-I in ABCA1 Knock-out Cells—The importance of hepatic ABCA1 in the lipidation of newly synthesized apoA-I has been documented by us and others (9Timmins J.M. Lee J.Y. Boudyguina E. Kluckman K.D. Brunham L.R. Mulya A. Gebre A.K. Coutinho J.M. Colvin P.L. Smith T.L. Hayden M.R. Maeda N. Parks J.S. J. Clin. Investig. 2005; 115: 1333-1342Crossref PubMed Scopus (427) Google Scholar, 15Kiss R.S. McManus D.C. Franklin V. Tan W.L. McKenzie A